Recent evidence suggests that changes in membrane potential influence the proliferation and differentiation of keratinocytes. To further elucidate the role of changes in membrane potential for their biological fate, the electrical behavior of keratinocytes needs to be studied under complex conditions such as multilayered cultures. However, electrophysiological recordings from cells in the various layers of a complex culture would be extremely difficult. Given the high spatial resolution of confocal imaging and the availability of novel voltage-sensitive dyes, we combined these methods in an attempt to develop a viable alternative for recording membrane potentials in more complex tissue systems. As a first step, we used confocal ratiometric imaging of fluorescence resonance energy transfer (FRET)-based voltage-sensitive dyes. We then validated this approach by comparing the optically recorded voltage signals in HaCaT keratinocytes with the electrophysiological signals obtained by whole cell recordings of the same preparation. We demonstrate 1) that optical recordings allow precise multisite measurements of voltage changes evoked by the extracellular signaling molecules ATP and bradykinin and 2) that responsiveness to ATP differs in various layers of cultured keratinocytes.
- fluorescence resonance energy transfer
- voltage-sensitive dyes
- laser scanning microscopy
- violet diode laser
the extracellular signaling molecules ATP and bradykinin mobilize intracellular Ca2+ (22) and induce a biphasic change of the membrane potential in a monolayer of confluent HaCaT keratinocytes due to the activation of Ca2+-dependent Cl−channels, cation channels, and K+ channels (15). The voltage response consists of an initial transient depolarization followed by a prominent and long-lasting hyperpolarization 30–40 mV below resting membrane potential. Given the increasing evidence that changes in membrane potential are intimately involved in the proliferation and differentiation of keratinocytes (15, 17-19, 27), the pronounced electrophysiological response to ATP and bradykinin might also bear significance on the cellular biology of keratinocytes. To better understand how alterations of membrane potential influence mitogenic activity and differentiation, it is essential to determine the electrical behavior under more complex culture conditions that allow simultaneous recordings from cells at different states of their life cycle.
HaCaT keratinocytes retain the capacity to differentiate partially in vitro. This differentiation is accompanied by the formation of multiple cell layers and the expression of defined patterns of keratins (e.g., keratin K10) similar to those observed in primary epidermal cells (2, 8, 23). With respect to electrophysiological experimentation, however, the multilayer of a complex culture poses a serious technical problem. Whereas cells in a monolayer are easily accessible to patch-clamp recordings, cells grown in a more complex environment are not. Furthermore, whole cell patch-clamp recordings are necessarily single-site measurements yielding information from just one cell in the dish. Therefore, the patch-clamp technique is not a suitable approach to measure and compare simultaneously the electrical behavior of different cells in a complex culture.
Voltage-sensitive fluorescent dyes have been known for many years as useful, noninvasive tools to record voltage changes in several cells at the same time (13, 24). However, poor voltage sensitivity or slow response times are significant limitations for many of these dyes. Recently, improved two-component fluorescent indicators of cell membrane potential have been described that use fluorescence resonance energy transfer (FRET) (10-12). In this system, a FRET donor dye is present in the outer leaflet of the cell membrane, while the voltage-sensing FRET acceptor dye moves within the cell membrane depending on the membrane potential. The excitation wavelength required to activate FRET in such pairs of voltage-sensitive dyes is around 405 nm. This wavelength is not available in conventional laser scanning microscopes. However, a violet diode laser with the necessary excitation wavelength has recently become commercially available (9). Here we describe the extension of a conventional confocal laser scanning system with a violet diode laser and the application of such an imaging system for confocal, FRET-based voltage measurements in cultured human keratinocytes. We demonstrate that this technique allows simultaneous multisite recordings of ATP- and bradykinin-induced changes in membrane potential from cultured human keratinocytes. Additionally, by confocal optical sectioning through multilayered HaCaT cultures, we show that the responsiveness to ATP differs dramatically between proliferating cells and cells undergoing partial differentiation.
MATERIALS AND METHODS
HaCaT keratinocytes were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (GIBCO BRL, Grand Island, NY) and antibiotics (100 U/ml penicillin/streptomycin). Final Ca2+ concentration was 1.8 mM. Electrophysiological measurements and optical recordings were performed as soon as the cells reached 70–100% confluence, usually 2–3 days after plating. Multilayered HaCaT cultures were obtained by growing confluent cultures for an additional 4 days.
HaCaT cell membrane potential was recorded by using the nystatin perforated-patch variation of the whole cell configuration as described previously (15). Before recording, the culture medium was replaced with standard bath solution containing (in mM) 130 NaCl, 3 KCl, 2 CaCl2, 2 MgCl2, 25 HEPES/NaHEPES, and 10d-glucose, pH 7.4 (21–24°C). Patch pipettes were fabricated from borosilicate glass by using a two-stage pull protocol on a horizontal puller (DMZ, Zeitz, Germany) and were filled with a solution containing (in mM) 135 K-gluconate, 10 KCl, 1.6 Na2HPO4, 0.4 NaH2PO4, 0.73 CaCl2, 1.03 MgCl2, 1 EGTA, 14 HEPES/NaHEPES, and 100 mg/l nystatin, pH 7.2. After formation of a seal in the gigaohm range (typically 1.5–2 GΩ), the amplifier was switched to current-clamp mode and membrane potentials attained stable values within 3–5 min. Electrophysiological signals were recorded, amplified, and digitized with the use of an Axopatch 200 amplifier in conjunction with a TL-1 Labmaster interface and AXOTAPE software (Axon Instruments, Burlingame, CA). Data from electrophysiological and optical measurements are given as means ± SE. Statistical analysis was performed by means of Student's t-test.
The optical setup consisted of an upright microscope (Olympus BX50WIF; objectives: Olympus UMPlanFl ×20W, Olympus LUMPlanFl ×40W, Olympus LUMPlanFl ×60W, and Zeiss Achroplan ×63W Ph3) to which a confocal laser-scanning head (Bio-Rad MRC-1024) and an independent custom-made photometric system (26) were attached. For photometric measurements using the FRET dyes, cells were illuminated with brief pulses (7 ms) at 400 nm from a xenon light source at intervals between 1 and 5 s. Two photodiodes collected fluorescent light emitted from the cell layer at wavelengths >550 nm (long-pass OG550) and at 450 ± 29 nm (450AF58; Omega Optical, Brattleboro, VT) and produced voltage signals that were proportional to fluorescence emissions. Signals were linearly amplified before PC-based data digitization (DT2812; Data Translation, Marlboro, MA).
In confocal imaging experiments, excitation light at 408 nm from a 30-mW violet diode laser (DL 100; TOPTICA Photonics, Martinsried, Germany) was directed via a single-mode optical fiber into the scanning unit of the Bio-Rad MRC-1024 confocal laser scanning system. Beam splitters at 440 nm (440 DCLP; Chroma Technology, Brattleboro, VT) and at 505 nm (505 DRLPxR; Omega Optical), respectively, were used in the first and second position of dichroic mirrors. Emitted light was collected by using a long-pass emission filter with cut-on wavelength of 550 nm (OG550) at photomultiplier (PMT) 1 and a band-pass filter at 450 ± 29 nm (450AF58) at PMT 2. Pinholes were 6 at PMT 1 and 5 at PMT 2 unless stated otherwise. The analysis of confocal images was performed on selected regions of interest using the imaging and time course software of the Bio-Rad MRC-1024 system. Acquisition of time series was usually done with sampling intervals between 10 and 20 s without averaging. To achieve sampling intervals of 2 s, the image size was reduced from 512 × 512 to 256 × 256 pixels. To enable optical sectioning, pinholes were set to 3.5 for PMT 1 and 2.2 for PMT 2. If optical sectioning was applied, the distance between sections was 7 μm. Confocal image acquisition using Ca2+-sensitive dyes was carried out with filter sets as previously described (20). Briefly, excitation was at 488 nm (argon laser), and emission band-pass filters were 660 ± 50 nm at PMT 1 and 530 ± 30 nm at PMT 2.
The coumarin lipid CC2-DMPE (FRET donor) and the oxonol DiSBAC4(3) (FRET acceptor) were purchased from Aurora Biosciences (San Diego, CA). Storage and loading of these dyes were performed according to the manufacturer's instructions. Aliquots of stock solutions (coumarin with a concentration of 1 mM; oxonol with a concentration of 2 mM) were prepared by the addition of an appropriate volume of dry DMSO. Cell cultures were stained sequentially at room temperature, with coumarin lipid loaded first. This was followed by washing and then oxonol loading in HEPES-buffered solution (HBS), which contained (in mM) 118 NaCl, 3 KCl, 1 MgCl2, 1.5 CaCl2, 20 Na-gluconate, 6 HEPES, and 10d-glucose, pH 7.4. CC2-DMPE was loaded by mixing 30 μl of the 1 mM stock solution with 15 μl of a 10% Pluronic-127/DMSO solution (Molecular Probes, Eugene, OR), and this was added to 3 ml of HBS for 20 min (10 μM final concentration). DiSBAC4(3) was loaded by mixing 9 μl of the 2 mM stock solution with 18 μl of a 10% Pluronic-127/DMSO solution, and this was added to 3 ml of HBS for 25 min (6 μM final concentration). Staining was prolonged to 1 h if multilayered cultures were stained. After staining, excess dye was removed by replacing the extracellular solution with HBS. Loading of Ca2+-sensitive dyes was carried out as previously described (20).
Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and fixed in acetone-methanol 1:1 for 20 min at −20°C. After blocking of nonspecific binding sites with 3% bovine serum albumin (BSA)/PBS, cells were incubated overnight with a 1:500 solution of a keratin K10-specific antibody (DAKO, Carpinteria, CA) in 3% BSA/PBS at 4°C and rinsed three times with PBS and once with 3% BSA/PBS. After a 2-h incubation with a FITC-labeled anti-mouse antibody at room temperature, cells were washed three times with PBS. Fura red (5 μM) was added to the culture medium 1 h before fixation to counterstain the cells for confocal fluorescence microscopy (filter sets were identical to those used for Ca2+ imaging). To enable optical sectioning, pinholes were set at 3.7 for both PMTs. The distance between optical sections was 7 μm.
Drugs and application system.
In both electrophysiological and optical recordings, drugs were applied by a remote-controlled Y-tube system that allowed rapid solution changes (21). In some of the confocal experiments, ATP and bradykinin were administered slowly (5 ml/min) via the bath perfusion system. 1-EBIO was purchased from Biotrend (Cologne, Germany); ATP, bradykinin, and tetraethylammonium-Cl (TEA-Cl) were purchased from Sigma (Taufkirchen, Germany).
To examine the distribution of the fluorescent probes, confocal images were taken after loading of HaCaT cells with the voltage-sensitive dyes CC2-DMPE and DiSBAC4(3). For comparison, other cultures were stained with the Ca2+-sensitive fluorescent dyes calcium green and fura red. As shown in Fig.1, the voltage-sensitive dyes led to predominant staining of the cell membranes, whereas loading with Ca2+-sensitive fluorescent probes led to cytoplasmic fluorescence.
Temporal resolution and sensitivity of FRET-based voltage-sensitive dyes.
To determine the temporal resolution and sensitivity of the optical voltage measurements, we compared optically recorded voltage responses to ATP with those obtained by means of the patch-clamp technique. Both the optical and electrophysiological recordings were performed on HaCaT cells grown to a confluent monolayer. As shown previously (15), ATP (100 μM) induced a biphasic change of the membrane potential in electrophysiological recordings, which consisted of a brief depolarization followed by a prolonged hyperpolarization (Fig. 2 A). In experiments with the voltage-sensitive dyes, ATP (100 μM) was applied to HaCaT cells via the same fast application method (Y-tube), and the emission ratio at 450 and 550 nm was continuously recorded with the use of the photometric optical setup (Fig. 2, B–D). Application of ATP (100 μM) for 90 s resulted in opposite changes in the emitted light from the FRET donor CC2-DMPE (Fig. 2 C) and the FRET acceptor DiSBAC4(3) (Fig. 2 D). In the emission ratio (Fig. 2 B), the initial increase indicates membrane depolarization, whereas the subsequent, longer lasting decrease indicates membrane hyperpolarization.
Although these recordings revealed a good correlation between the shape and the time course of the optical and the electrophysiological signal, the ATP-induced voltage changes were too slow to assess the temporal resolution of the dyes. To induce rapid changes of the membrane potential in the hyperpolarizing and depolarizing direction, HaCaT cells were first exposed to 1-EBIO (300 μM), a direct opener of epithelial Ca2+-activated K+ channels (6), and then depolarized with 30 mM extracellular K+, again using the fast drug application system. In the electrophysiological experiments, the 1-EBIO-induced hyperpolarizing shift and the high K+-induced depolarizing shift were each completed within a few seconds (Fig.3 A). Very similar kinetics were obtained when the voltage changes were optically recorded by means of photometry (Fig. 3 B) or confocal imaging (Fig.3 C), suggesting that the temporal resolution of the dyes is sufficiently high to monitor accurately the fastest physiological voltage changes that might occur in our preparation.
The experiments of Fig. 3 also served to determine the voltage sensitivity of the dyes. In electrophysiological recordings, 1-EBIO (300 μM) hyperpolarized HaCaT cells by 33.3 ± 2.4 mV (n = 7, Fig. 4 A,left). This corresponded to a decrease in the normalized emission ratio of −5.57 ± 0.33% (n = 28 applications, 15 culture dishes) in photometric experiments (Fig. 4 B, left) and −5.45 ± 0.51% (70 regions of interest, 6 culture dishes) in confocal experiments (Fig. 4 C, left), respectively. From these data, the relationship between the shift in membrane potential and the change in the emission ratio was calculated. On average, hyperpolarization by 1 mV decreased the emission ratio by −0.167% in photometric experiments and by −0.164% in confocal imaging experiments (Fig. 4, B and C, left). Likewise, membrane depolarization by 1 mV increased the emission ratio by 0.140% in photometric experiments and by 0.135% in confocal imaging experiments (Fig. 4, B and C,right).
Confocal imaging with FRET-based voltage-sensitive dyes.
In conjunction with confocal imaging, the dyes were successfully employed to record voltage responses to bath-applied ATP (n = 15) or bradykinin (n = 5) simultaneously from several cells (Fig. 5,A–C). Again, experiments were performed on HaCaT cells grown to confluence. Six regions of interest containing cell membranes and a background field in the center of a cell were selected for measurements of changes in the intensity of emitted light from CC2-DMPE (450 nm), from DiSBAC4(3) (550 nm), and in the emission ratio. Individual regions of interest did not respond uniformly, but the decrease in the emission ratio indicated a hyperpolarizing action of the two compounds in all cells examined. Changes in emission ratio ranged between −5 and −10%, corresponding to hyperpolarizing voltage shifts between −30 and −60 mV (see Temporal resolution and sensitivity of FRET-based voltage-sensitive dyes). Demonstrating the specificity of the optical signals, the emission ratio from the background field, which did not include a cell membrane, remained unresponsive to bradykinin and ATP.
HaCaT cells undergo partial differentiation in long-term cultures of confluent monolayers (23). The formation of additional cell layers on the basal layer is accompanied by expression of keratin K10, an early differentiation marker. Before comparing the voltage responses to ATP in basal and suprabasal cells, we used confocal laser scanning microscopy to establish the differential expression of K10 in multilayer cultures at 4 days postconfluence. In Fig. 6,A–C, the immunolocalization of K10 (green fluorescence) is illustrated in a section of the dish where a multilayer of basal and suprabasal cells is next to a monolayer of basal cells. As expected, K10-positive cells were predominantly present in the upper layers, whereas cells of the basal layer attached to the culture dish rarely expressed K10 (Fig. 6,A–C). Postconfluence, multilayered HaCaT cell cultures were loaded with CC2-DMPE and DiSBAC4(3) and were superfused with ATP (200 μM) for 2 min. As depicted in Fig. 6,D1–F1, basal and suprabasal cells displayed a striking difference in their ATP responsiveness: K10-positive cells responded poorly to ATP, whereas cells of the basal layer proved highly sensitive to ATP, as did confluent cells in a monolayer (see above).
Impaired diffusion of the dyes or ATP is unlikely to account for the layer-specific efficacy of ATP, because a diffusion gradient, if present, should produce the opposite effect, eliminating ATP responses in the lower but not in the upper layers, which are directly exposed to the test solutions. We had to exclude, however, that the diminished effect of ATP in K10-positive cells arises from variations in staining properties or dye behavior that might occur as a function of cell age or depth in culture. Therefore, we performed a series of experiments in which we recorded the optical response to ATP in the absence or presence of the K+ channel blocker tetraethylammonium (TEA; 5 mM), which, however, does not inhibit ATP-activated human intermediate-conductance Ca2+-activated K+(hIK1) channels (15). TEA should increase the electrical resistance of the membrane, thereby augmenting the impact of even a small ATP-induced K+ current on membrane voltage. This should allow us to uncover any residual ATP-induced K+current, if present in K10-positive cells. Application of TEA indeed revealed an optical response to ATP in K10-positive cells (Fig.6 F2). By contrast, TEA had no appreciable effect on the ATP response in basal cells (compare Fig. 6 F1, right vs. 6F2, right), suggesting that the ATP-induced hyperpolarization was already maximal in the absence of TEA. The differential effect of TEA in K10-negative vs. K10-positive cells dispels concerns that the disappearance of the hyperpolarizing action of ATP with differentiation results from an altered sensitivity of the indicator system. Instead, our data suggest that age-dependent changes in the electrical membrane properties and/or in the efficacy of the signaling system (see following paragraph) render the prominent ATP response of proliferation-competent cells biologically insignificant in differentiating cells.
Because bradykinin is coupled to the same effector system and displays a very similar action on the resting membrane potential of K10-negative cells as ATP (Fig. 5), we wondered whether the hyperpolarizing response to bradykinin would also decrease with differentiation. We therefore performed a series of experiments in which we examined the optical response to both bradykinin (5 μM) and ATP (200 μM) in the same multilayered cultures. The representative traces from a single experiment (Fig. 7 A) and the histogram (Fig. 7 B) indicate that in the same cultures in which ATP displayed a significant layer-specific response, bradykinin did not. The response to bradykinin showed a tendency to be smaller in suprabasal cells, but this did not reach statistical significance.
Many attempts have been made to improve optical imaging utilizing voltage-sensitive dyes (e.g., 3, 4, 25). A promising recent development are two-component fluorescent indicators that use FRET (10-12). The advantage of FRET-based voltage-sensitive dyes (28) is a ratiometric output that can be considerably more sensitive than previous indicators. To our knowledge, the combination of FRET-based optical voltage recordings with imaging techniques has so far only been realized with a charge-coupled device camera system (5). This approach has a high temporal resolution, but difficulties arise with optical sectioning in thick preparations and dual-emission wavelength imaging. In an attempt to overcome these problems, we describe in the present study a confocal laser scanning system equipped with a novel violet laser diode that provides the excitation wavelength of 408 nm required for FRET-based voltage-sensitive dyes (9).
We obtained a sensitivity of 15.4% ratio at 100 mV in photometric experiments and 15.0% ratio at 100 mV in confocal imaging experiments (Figs. 3 and 4). In previous studies, fluorescence ratio changes of between 4 and 34% for a 100-mV depolarization were observed for fibroblasts, astrocytoma cells, cardiac myocytes, and neuroblastoma cells (11). Unlike the situation in whole cell recordings, we observed baseline drifts in nearly every optical measurement of membrane potential. This may have been caused by differential bleaching or leakage of the dyes. However, drift did not influence the relative change of emission ratios (ΔF/Fo) if appropriate time frames were chosen. In confocal imaging experiments, we observed that noise levels decreased with an increase in the number of pixels for a certain region of interest (Fig. 5). We therefore conclude that averaging of several image frames per time point might substantially enhance the quality of confocal recordings. However, averaging will lead to more bleaching of the dye and photodamage of the specimen. In accordance with previous data (1, 10), we found that the temporal resolution of the dye pair CC2-DMPE/DiSBAC4(3) is below 1 s (Fig.3), which is sufficient to detect changes in membrane potential induced by ATP, bradykinin, and 1-EBIO in HaCaT cells.
The use of FRET-based, voltage-sensitive dyes in conjunction with confocal imaging has two distinct advantages over conventional patch-clamp recordings. First, in monolayers, in which individual cells are also accessible to whole cell recording, confocal imaging allows simultaneous multisite recordings. Our data indicate that virtually no cell in the dish remains unstained. Given the high temporal resolution and sensitivity of the indicator system, as demonstrated in the present study, reliable data on drug-induced voltage changes can be obtained more easily, and at a much higher throughput rate, compared with conventional whole cell patch-clamp recording. Second, and more importantly, confocal imaging allows simultaneous voltage recordings from different cells in complex multilayer cultures. This application, which is unique to the technique presented here, is of particular biological significance, because it offers the opportunity to address long-standing questions on the relationship among membrane potential, channel activity, and elementary cellular processes such as proliferation and differentiation.
We demonstrate the feasibility of this new approach by comparing voltage responses to the extracellular signaling molecule ATP in basal, K10-negative cells and in suprabasal, K10-positive cells. In this multilayer culture of postconfluence HaCaT cells, robust ATP responses were confined to K10-negative cells, whereas the hyperpolarizing action of ATP disappeared as the cells began to stain positive for K10. Interestingly, a hyperpolarizing ATP response could be unmasked in K10-positive cells, when the purinergic agonist was applied in the presence of the K+ channel blocker TEA. This finding has both a technical and a biological implication. Methodologically, it strongly argues against an altered sensitivity of the indicator system as the underlying cause of the declining ATP responsiveness in K10-positive cells. Functionally, it suggests that differentiating keratinocytes have progressed into a state in which the hyperpolarizing effect of ATP becomes biologically meaningless, although the elements of the ATP-induced signaling pathway are apparently retained in residual form. In support of this notion, we have recently demonstrated that the mRNA level of the P2Y2 receptor, which binds extracellular ATP, is downregulated in differentiating HaCaT cells (15). This agrees well with in situ hybridization studies of human skin sections, which revealed a striking pattern of localization of P2Y2 receptor transcripts to the basal layer of the epidermis, the site of cell proliferation (7). Although we cannot directly infer from mRNA data the number of functional membrane proteins, it seems reasonable to assume that the K+ current generated by ATP in K10-positive cells should be smaller than in proliferation-competent cells. With much less K+ current generated, the hyperpolarization of the membrane potential should be significantly smaller, as observed in our experiments. Only after TEA had enhanced the electrical resistance of the membrane potential is the same small K+ current capable of producing a considerable hyperpolarization of the membrane potential, as predicted by Ohm's law and confirmed in our experiment.
Unlike ATP, bradykinin did not lose its hyperpolarizing effect with keratinocyte differentiation. Bradykinin is mainly liberated during inflammatory processes of the skin, and our data suggest that membrane hyperpolarization might be an important signal transduction pathway of bradykinin in both proliferation-competent and differentiating keratinocytes. By contrast, the action of ATP appears to be confined to basal, proliferation-competent cells. This agrees well with the main biological function attributed to ATP in human epidermis: ATP has been shown to stimulate proliferation and inhibit differentiation of human keratinocytes (7, 22). Furthermore, several lines of evidence suggest that the hIK1 channel that serves as effector of the hyperpolarizing ATP response plays an essential role in promoting mitogenic activity. For example, pharmacological downregulation of IK channels was found to suppress proliferation of HaCaT keratinocytes (16). Along the same line, expression of IK channels was found to disappear with differentiation in various cell types (reviewed in Ref. 14). By measuring directly ATP-induced voltage responses in cells of different age in the same preparation, we are now capable of monitoring the functional expression of an effector system that appears to be crucially involved in regulating the balance between keratinocyte proliferation and differentiation. Thus our method provides a novel approach to analyze the intricate relationship between bioelectrical and cell biological events in cells or complex tissue systems that are not accessible to conventional electrophysiological recording techniques.
We thank Christa Müller for technical assistance and Prof. Gerrit ten Bruggencate for help in the initial phase of this project. HaCaT keratinocytes were kindly provided by Dr. Petra Boukamp (German Cancer Research Institute, Heidelberg, Germany).
Financial support was provided by the Deutsche Forschungsgemeinschaft (GR 801/2-1, SFB 391/TPA1, and a Heisenberg-Fellowship to C. Alzheimer).
Address for reprint requests and other correspondence: R. Burgstahler, Dept. of Physiology, Univ. of Munich, Pettenkoferstr. 12, 80336 Munich, Germany (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published December 21, 2002;10.1152/ajpcell.00053.2002
- Copyright © 2003 the American Physiological Society