Regional Electroporation of Single Cardiac Myocytes in a Focused

Anal. Chem. 2010, 82, 585–592

Regional Electroporation of Single Cardiac Myocytes in a Focused Electric Field Norbert Klauke,*,† Godfrey Smith,‡ and Jonathan M. Cooper† Department of Electronics, University of Glasgow, Glasgow G12 8LT, and Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ There is now a significant interest in being able to locate single cells within geometrically defined regions of a microfluidic chip and to gain intracellular access through the local electroporation of the cell membrane. This paper describes the microfabrication of electroporation devices which can enable the regional electroporation of adult ventricular myocytes, in order to lower the local electrical resistance of the cell membrane. Initially three different devices, designed to suit the characteristic geometry of the cardiomyocyte, were investigated (all three designs serve to focus the electric field to selected regions of the cell). We demonstrate that one of these three devices revealed the sequence of cellular responses to field strengths of increasing magnitudes, namely, cell contraction, hypercontraction, and lysis. This same device required a reduced threshold voltage for each of these events, including in particular membrane breakdown. We were not only able to show the gradual regional increase in the electric conductivity of the cell membrane but were also able to avoid changes in the local intra- and extracellular pH (by preventing the local generation of protons at the electrode surface, as a consequence of the reduced threshold voltage). The paper provides evidence for new strategies for achieving robust and reproducible regional electroporation, a technique which, in future, may be used for the insertion of large molecular weight molecules (including genes) as well as for on-chip voltage clamping of the primary adult cardiomyocyte. Manipulation of the membrane permeability through conventional electroporation for nonexcitable cells has been the subject of considerable technological interest for the last 30 years1 and has most recently resulted in the implementation of single-cell electroporation in microfluidic chips2 and in the selective electroporation of intracellular membranes, e.g., mitochondria or endoplasmic reticulum, using nanosecond pulses.3 Conventional electroporation with >1 ms pulse duration and 50 V/cm. The incapacity of these cells (1-5% of the cell preparation) to respond to the electrical stimulation might be inflicted through metabolic changes during the isolation process like intracellular pH shift or lactate accumulation.21 In some experiments, the electroporated cell surface was superfused with an intracellular mock solution of (mM): 100 KCl, 5 Na2ATP, 10 disodium creatine phosphate, 5.5 MgCl2, 25 Hepes, 0.05 K2EGTA, pH 7.0, 20-21 °C.22 Unless otherwise stated, all chemicals were obtained from Sigma-Aldrich (U.K.). Fabrication Protocol. The fabrication of the electrodes and their alignment to the microfluidic channels was as previously described.23 The electrode design was transferred into a thin layer of spun photoresist (S1818, Shipley, U.K.; 1.8 µm thick film at 4000 rpm) on a glass substrate (coverslip no. 1, ca. 170 µm thick) to provide a transparent substrate suitable for inverted light microscopy with a high NA (>1.2) objective lens. The microelectrodes (formed from either gold or platinum) were deposited onto the patterned photoresist using electron beam assisted metal evaporation (Plassys MEB e-beam evaporator, PLASSYS-BESTEK, France). After lifting-off the photoresist, the coverslips with the metal microelectrode pattern were spin-coated with a positive photoresist (12 µm thick film at 1200 rpm; AZ4562, Clariant, Switzerland) and baked at 90° for 30 min. The mask with the microchannel pattern was aligned to the electrode array on the coverslip and exposed for ca. 25 s at 7.2 mW/cm2 on the mask aligner (MA6, Karl Suess, Germany). When fabricating planar microstructures, used in this paper to study the geometric arrangement of microelectrodes in relation to the orientation of the cardiomyocyte, pattern transfer was always by hard-contact exposure. The microchannel pattern was developed in alkaline solution (AZ400K, ca. 5 min) and provided a “sacrificial” master for the replica molding of poly(dimethylsiloxane) (PDMS). A dilute PDMS solution (1:4 in hexane) was spun onto the glass substrate and oven-cured at 90 °C, prior to the dissolution of the “sacrificial” master in acetone. The microfluidic channels were then first wetted with ethanol, which was subsequently replaced with buffer solution (base Krebs solution, see above). This protocol provided well-defined microchannels precisely aligned to the electrodes which were free of residues of photopolymer. In addition, this “sacrificial” technique resulted in a highquality seal between the glass, the integrated microelectrodes, and the PDMS (when the master was removed). The technique produced microfluidic channels in thin (515 nm. Regional fluorescence changes were measured in selected regions of interest (ROIs 1-5) as indicated in Figure (24) Klauke, N.; Smith, G. L.; Cooper, J. M. Biophys. J. 2006, 91, 2543–2551.

Figure 1. (A) Parallel electric field with the cardiomyocyte’s long axis parallel to the electrical current direction (electric field indicated through dotted lines). The sketch in panel A, part i, depicts the confined extracellular volume (darker gray area filled with electrolyte, lighter gray area without electrolyte) stretching just beyond the electrodes. The right border of the extracellular volume is visible on the micrograph (panel A, part ii; arrowheads); the left border is hidden by the opaque gold electrode (as viewed through the inverted microscope). The confined extracellular volume between the two electrodes was ca. 80 pL. Bar ) 40 µm. (B) Orthogonal E-field with the cardiomocyte’s long axis perpendicular to the electrical current direction. The cardiomyocyte was positioned between the electrodes so that only the central part was exposed to the electric field. The extracellular volume was confined to ca. 80 pL (the borders of the electrolyte are visible on the micrograph in panel B, part ii; arrowheads). Bar ) 40 µm. (C) Constricted parallel electric field with the ends of the cardiomyocyte electrically insulated from each other through a central fluidic sealant (mineral oil, seal resistance ca. 50 MΩ, ref 23). The central constriction reduces the electrolyte to a thin layer between the cell surface and the surrounding oil (Figure 3), thus focusing the electrical current onto the cell surface. The sketch indicates the position of five regions of interest (ROIs 1-5) as they were used for evaluation of the regional fluorescent intensity throughout the paper. Bar ) 40 µm. (D) Comparison of the threshold for electroporation (indicated through hypercontraction) for the different geometries depicted in panels A-C.

1C, with two extracellular regions of interest (ROI 1 and ROI 5) to measure the fluorescence intensity in the anodal and cathodal chamber, respectively, and three intracellular regions of interest Analytical Chemistry, Vol. 82, No. 2, January 15, 2010

587

gap between the two chambers and allowing for the passage of electrical current between the electrodes. In contrast, hypercontracted cells which changed from rod shape to ball shape after electroporation were not able to maintain the electrical connection between the electrodes because the ball-shaped cell slid away from the hydrophobic gap into one of the chambers. The failure to contract and hypercontract was caused through the loss of excitation either during continual electrical stimulation (Figures 5 and 6) or through the cell isolation process (see above). RESULTS

Figure 2. Central fluidic seal was used to separate the buffer solution in the extracellular space around the cell ends. No fluorescence was detected in the right end pool after a fluorescent dye (70 kDa SNARF-1 dextran) was loaded into the left end pool. The example depicted in the figure is representative for seven cells (panel C). (A) Transmitted light image of the cardiomyocyte across the oil gap and (B) fluorescent confocal image of the same cell. Note the two thin lines of fluorescence on the confocal image running along the cell surface toward the right end pool (arrowheads in panel B), indicating the electrolyte-filled residual extracellular space between the cell surface and the sealant (oil) which provided a limited path for the extracellular dye diffusion between the two end pools. Bar ) 40 µm.

(ROIs 2-4) to measure the fluorescence intensity in the anodefacing cell end (ROI 2), in the cathode-facing cell end (ROI 4), and in the cell center (ROI 3). The quality of the hydrophobic gap with respect to its barrier function to prevent solution exchange between the two chambers was tested by measuring the diffusion of seminaphthorhodafluor (SNARF-1) dextran (70 kDa, excitation wavelength 543 nm/ emission wavelength >580 nm, Invitrogen, Paisley, U.K.) on confocal z-stacks (Zeiss 510, Germany) of the gap region bridged by the cardiomyocyte. The SNARF-1 dye (ca. 5 µM) was pipetted into one chamber using the previously described dual-perfusion pipet,25 and the fluorescent change in the other chamber was monitored for 30 min. The intracellular pH was imaged in cardiomyocytes loaded with 2′,7′-bis(2-carboxyethyl)-5-(and-6)carboxyfluorescein (BCECF, excitation wavelength 490 nm/ emission wavelength >515 nm, Invitrogen, Paisley, U.K.) during electrical stimulation and electroporation. Intracellular pH measurements at field strength sufficient for cell lysis (>200 V/cm) were performed in cardiomyocytes selected for their incapacity to hypercontract at this high field strength (see below). Extracellular BCECF during cell lysis (see below) was measured at its isosbestic excitation (440 nm). Cell Lysis. Regional electroporation with field strength of >200 V/cm (above the threshold for hypercontraction) was investigated in cells which were unable to hypercontract during regional electroporation but maintained their rod shape during the course of the experiment, thus continuously bridging the hydrophobic (25) Klauke, N.; Smith, G. L.; Cooper, J. M. Anal. Chem. 2007, 79, 4581–4587.

588

Analytical Chemistry, Vol. 82, No. 2, January 15, 2010

During electrophysiological experiments, while recording extracellular potentials, changing the ion-permeation properties of the cell membrane through electroporation at low field strength (ca. 50 V/cm) provides a mechanism for fast current injection into the intracellular space via a low electrical resistance pathway. In future, such a technique will also be used for successfully clamping the membrane potential at a set voltage, in order to study ion-channel activity. This paper presents results which show, for the first time, how the regional sarcolemmal (membrane) conductivity changes as a result of the disruption of the sarcolemma using electroporation at low field strengths (50 V/cm), on a chip. We demonstrate that, by the local increase in the conductivity of the sarcolemma, we are able to differentially record the extracellular potential of the single cardiomyocyte, e.g., the potential difference between the undisrupted the plasma membrane and the electroporated membrane, using microfabricated extracellular metal film electrodes, within a microfluidic channel (Figure 3). We also show, that at higher field strength (>200 V/cm), electroporation allowed dye (BCECF) efflux from cardiomyocytes selected for their incapacity to hypercontract at this high field strength (see the Experimental Section), indicating that the pores formed in the membrane are not only large but also irreversible (Figure 4). Such a technique may in future be generally used to manipulate the cellular content, including either the capture of intracellular proteins or the influx of large molecules, including sensors and genetic material, into the cell. In Figure 1 we show three different cell-electrode configurations (Figure 1A-C), which we fabricated in order to investigate three different methods to regionally focus the electric field onto a defined area of the cell surface. In Figure 1, parts A and B, we positioned the axis of the cell either parallel (A) or orthogonal (B) to the electric field between a pair of planar integrated gold microelectrodes. In Figure 1C the cell was positioned within a hydrophobic fluidic sealant (mineral oil), which served to fluidically isolate the two ends of the cell, Figure 2 (and consequently to electrically separate these regions). In each of the three cases, the cellular response to a series of applied electric fields, between 10 and ca. 400 V/cm, was monitored with an optical microscope. The parameters recorded included cell contraction and electroporation (demonstrated by local Ca2+ transients), localized changes in intracellular and extracellular pH, cell hypercontraction, and finally dye efflux (indicating the formation of large and lasting pores). The threshold voltages for electroporation (as indicated through the local artificial irreversible Ca2+ influx and subsequent hypercontracture upon a single monophasic pulse of 4 ms

Figure 3. Extracellular potential of a single adult ventricular myocyte at the beginning (trace A) and after 10 min (trace B) and 30 min (trace C) of continuous electrical stimulation for ca. 30 min at 1.0 Hz using the constricted electric field (average of 20 traces each). At the beginning of the experiment the evoked action potential (trace A, 0 min) propagates fast along the entire cell, and the differential amplifier recorded a plateau phase of low amplitude (**) starting with a major upward deflection (*, common for all three traces) marking the depolarization and a minor downward deflection (***) marking the repolarization. The stimulus artifact has been subtracted. Traces B and C (10 and 30 min): the amplitude of the plateau phase gradually increased from 0.58 ( 0.35 mV (0 min) over 1.81 ( 0.99 mV (10 min) to 4.10 ( 1.55 mV (30 min, n ) 5, inset), indicating the increase in potential difference between the two cell ends separated through the lithographic seal. The asymmetric injury occurring during the ongoing electrical stimulation (ref 17) possibly depolarized the anodefacing cell end more compared to the cathode-facing end, thus increasing the electrical potential difference between the two ends.

duration) in the three geometric configurations were measured as follows: for parallel cell orientation, 8.1 ± 1.3 V (404 ± 66 V/cm), Figure 1A; for the orthogonal cell orientation, 2.4 ± 0.4 V (119 ± 20 V/cm), Figure 1B; finally, for the constricted geometry, 1.4 ± 0.3 V (68 ± 16 V/cm), Figure 1C, see below. These results (summarized in Figure 1D) clearly show that the hydrophobic gap structure provides a route by which the membrane was disrupted at an absolute voltage close to the level at which electrolysis occurs (typically ca. 1.2 V). The significance of this is that it was possible to induce electroporation for current injection without causing local perturbation to the pH of the base Krebs solution surrounding the cell. The reduction in the threshold potential arises simply as a consequence of the high extracellular resistance across the soft-lithographic seal, formed by sealant estimated to be ca. 50 MΩ,24 resulting in high field densities at the edge of the fluidic seal. This locally high field, which resides at the interface between the microfluidic chamber and the seal, is caused by the sharp potential drop at this point as result of the abrupt difference in conductivity between the buffer solution and the oil. This is illustrated in Table 1, where it can be seen that the [Ca2+]i in this region of the cell was almost twice as high as rest of the cell (83% ± 42% compared to 47% ± 21%) ca. 100 ms after a single electroporation pulse. In the configurations depicted in Figure 1, parts A and B, when using base Krebs solution, pH 7.4, the voltage pulses at supraelectrolysis thresholds (8.1 ± 1.3 and 2.4 ± 0.4 V, Figure 1D), which were required to cause abnormal Ca2+ influx with subsequent calcium waves and hypercontracture after electroporation, also

Figure 4. Regional Ca2+ transients evoked through regional electroporation. A Fluo-3 loaded cardiomyocyte was electrically stimulated for >3 min across the insulating central gap which generated the high electrical resistance (ca. 50 MΩ) between the cell ends. The global [Ca2+] was recorded and subsequently analyzed in three regions of the cell: in the two cell ends (ROIs 2 and 4) and the cell center (ROI 3) as indicated on the sketch in Figure 1C. Panels A and B show the time courses of the regional [Ca2+]i in the cell ends (ROIs 2 and 4) during the increase in field strength (A) and during the frequency change between 0.5 and 1.0 Hz (B). Note the gradual increase of the amplitude of the local Ca2+ transients in ROI 2 along with increasing stimulus strength. The [Ca2+]i in the opposite cell end (ROI 4) remained unchanged. The local Ca2+ transients followed the change in stimulation frequency indicating that they were triggered by the electrical stimulus (panel B). The amplitude of the local Ca2+ transient in ROI 2 was ca. 10% of the amplitude of the global Ca2+ transient evoked in the same cardiomyocyte upon a minor increase in the stimulus strength (data not shown).

Table 1. Regional Inhomogeneities in the [Ca2+]i during Hypercontracture (n ) 5)a percentage of [Ca2+]i increase left of constriction (ROI 2)

percentage of [Ca2+]i increase right of constriction (ROI 4)

47 ± 21

83 ± 42

a Fluo-3 loaded ventricular myocytes were electroporated in the constricted electric field with a single pulse of a field strength of ca. 70 V/cm, sufficient for hypercontracture, and the [Ca2+]i was imaged as Fluo-3 fluorescence change. The electroporation-induced irreversible Ca2+ influx caused the irreversible rise in intracellular Fluo-3 fluorescence, which appeared first in ROI 4 (see Figure 1C) next to the constriction of the electric field. The data in Table 1 were recorded 100 ms after the electroporation pulse.

caused pH transient shifts of between 1 and 2 pH units at the cathode (pH up to 9.0) and anode (pH ca. 6.0). As a consequence, all subsequent electroporation measurements were made in a confined extracellular volume of ca. 500 pL within the device containing the seal, Figure 1C, which allowed electroporation (as indicated through hypercontracture) at voltage pulses near the threshold for electrolysis (1.4 ± 0.3 V, Figure 1D), preventing any extracellular pH change caused through electrolysis. The analysis of the amplitude of the stimulus artifact on the recordings of the extracellular potential (Figure 3) showed a potential difference of ca. 48 ± 3 mV (n ) 5) between the two microelectrodes separated by the constriction, which was sufficient for electroporation and irreversible Ca2+ influx to occur with subsequent hypercontracture. This value could be directly Analytical Chemistry, Vol. 82, No. 2, January 15, 2010

589

compared to the corresponding potential difference of ca. 44 ± 2 mV (n ) 5), which was sufficient for electrical stimulation indicated through the transient cell-shortening and intracellular Ca2+ transients (see Supporting Information Figure S-1 for waveforms of the concomitant extracellular potential). The microfluidic configuration in Figure 1C also enables a mock intracellular solution to be introduced at the site of local electroporation. The extracellular solution inside the chamber was changed using the dual-perfusion pipet as described previously.25 Figure 2 shows an example of such a manipulation where the mock solution contains a fluorophore (70 kDa SNARF-1 dextran). No leakage of the dye was observed for 30 min (Figure 2C). The arrowheads in Figure 2B show clearly the position of the sealant gap. The high extracellular resistance in Figure 1C also enables the measurement of extracellular potentials during electrical stimulation of the cell, causing contraction and electroporation (evidence by an increase in the extracellular potential with time). Parts A-C of Figure 3 are example traces representative of typical measurements, presented in the inset of Figure 3 (n ) 5), and show the increasing level of electroporation with increasing amplitude of the extracellular potential, reaching a plateau phase in trace C. This point is further illustrated by Figure 4, which shows how local electroporation induces local Ca2+ influx without propagation of the Ca2+ transient across the entire sarcoplasm. For example, Figure 4, parts A and B, shows the measurement of the intracellular Ca2+ transients during voltage pulses close to the electroporation threshold in regions ROI 2 and ROI 4 either side of the sealant gap (ROIs as indicated in Figure 1C). Ca2+ transients occurred in the anode-facing cell end (ROI 2) but not in the cathode-facing cell end (ROI 4). At the relatively low field strengths of ca. 40 V/cm, seen in Figure 4, parts A and B, the process of regional electroporation is fully reversible and creates small pathways for increased ion flux across the cell membrane (although at these magnitudes of field strength, no dye efflux was observed). In order to further investigate the process of electroporation, we were able to monitor the regional efflux of a fluorescent dye as a marker for membrane integrity during electrical stimulation of cardiomyocytes. These experiments were performed on cells which initially contracted during continual field stimulation but then gradually lost excitability26 and failed to hypercontract at the higher field strength of ∼150 V/cm, as described in the Experimental Section. A single dye-loaded cell was retained in a lowvolume microchamber (ca. 500 pL), and even very small amounts of dye efflux could be measured, due to minimal dilution (Figure 5). In our experiments, we used the fluorescent pH indicator BCECF, which was loaded into the cardiomyocyte. The efflux of the dye was monitored at its isosbestic point (440 nm excitation, 520 nm emission, pH-insensitive). Figure 5B shows the absence of dye efflux during the application of a train of stimulation pulses, at the weak field strength of ca. 33 V/cm. These pulses of low-voltage amplitude caused normal cell contraction at 0.3 Hz, over a period of at least 10 min. The cell then gradually lost excitability, as indicated through contraction with declining shortening amplitude, and (26) Sharma, V.; Susil, R. C.; Tung, L. Biophys. J. 2005, 88, 3038–3049.

590

Analytical Chemistry, Vol. 82, No. 2, January 15, 2010

Figure 5. Local dye efflux after regional electroporation in the constricted electric field. For the dye-efflux experiment, a cardiomyocyte was loaded with BCECF and first electrically stimulated with monophasic rectangular 4 ms pulses at 0.3 Hz and