Original Article

Photo-Control of Excitation Wavesin Cardiomyocyte Tissue Culture

Nobuyuki Magome, Ph.D.,1 Giedrius Kanaporis, Ph.D.,2 Nicolas Moisan, Ph.D.,1

Koichiro Tanaka, Ph.D.,1 and Konstantin Agladze, Ph.D.1

Azobenzene photoswitches were recently reported to control the activity of neural cells and heart beat in leeches.Here, we report photocontrol of excitation of cultured cardiomyocytes that have been made light sensitive byusing the addition of azobenzene trimethylammonium bromide (AzoTAB). The trans-isomer of AzoTAB re-versibly suppresses spontaneous activity and propagation of excitation waves, whereas the cis-isomer has nodetectable effect on the electrical properties of cardiomyocytes. Photoisomerization of AzoTAB was achieved byswitching the illumination wavelength, l, from *440 nm (trans-isomer) to *350 nm (cis-isomer). Simultaneousirradiation at two wavelengths with properly chosen intensities allowed for dynamic control of the cis-isomer/trans-isomer ratio and the level of excitability from normal to fully unexcitable. Experiments were conducted byusing AzoTAB-treated confluent monolayers of neonatal rat cardiomyocytes. Excitation waves were monitoredby using the Ca2 + -sensitive fluorescent dye Fluo-4. By projecting two-wavelength illumination patterns ontootherwise uniform cell layers, we were able to create excitable networks with the desired topology, dimensions,and functional properties. The present article discusses potential applications of this technique for the analysis ofcomplex patterns of electrical excitation and cardiac arrhythmias.

Introduction

Cardiomyocyte tissue culture has been extensivelyused to study the fundamental mechanisms of electrical

propagation in cardiac cell networks and to examine theorigin of deadly arrhythmias.1–7 The desired network ge-ometry is usually achieved by patterned plating or surgicalremoval of unnecessary cells8,9; however, the resulting net-work geometry is static and cannot be readily modifiedduring the course of an experiment. Here, we explored thepossibility of creating dynamic networks with topology andlocal functional properties that could be changed in ‘‘realtime’’ using photoswitches. Specifically, we focused on theazobenzene10,11 derivatives that have recently been reportedto block acetylcholine receptors12–14 and potassium channelsin a light-dependent manner. These photoswitches have alsobeen shown to control patterns of neural activity.15–17

In this study, we demonstrated that monolayers ofneonatal cardiomyocytes treated with azobenzene tri-methylammonium bromide (AzoTAB)18–22 can be renderedunexcitable when illuminated by 440-nm light and can fullyrecover their excitability when the illumination wavelengthis switched to 350 nm. Using simultaneous illuminationat two wavelengths with properly adjusted intensities, wewere able to control the propagation velocity of the action

potentials (APs) and the rate of spontaneous firing of cardi-omyocytes.18 Using structured illumination, we could dy-namically pattern the excitable network by directing excitationwaves over a defined path and creating pacemakers at desiredlocations. In contrast to the effects in neurons, our data sug-gested that AzoTAB-induced photosensitivity in cardiomyo-cytes involves sodium and calcium channels.

Materials and Methods

Preparation of cardiomyocyte cultures

Hearts isolated from neonatal 1-day-old Wistar rats wereminced and dispersed in a collagenase type 1 solution. Toremove the fibroblasts after centrifugation, the isolated cellswere suspended in Dulbecco’s modified Eagle’s medium(with 10% fetal bovine serum, 1% penicillin streptomycin,and 1% kanamycin) and kept in a tissue culture dish for 1 h.Nonadherent cells were collected and plated on a glass-based dish (27-mm j glass window) coated with humanplasma fibronectin (Gibco) and incubated at 37�C underhumidified 5% CO2 conditions for 1 day. After incubation,the medium was replaced with a minimum essential me-dium containing 10% calf serum, 1% penicillin-streptomycin,and 1% kanamycin; and the incubation continued for 5–8days until confluent monolayers were formed.

1Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Kyoto, Japan.2Institute of Cardiology, Lithuanian University of Health Sciences, Kaunas, Lithuania.

TISSUE ENGINEERING: Part AVolume 00, Number 00, 2011ª Mary Ann Liebert, Inc.DOI: 10.1089/ten.tea.2010.0745

1

Monitoring and control of excitation

Excitation waves were monitored by using the Ca2 + -sensitive fluorescent dye Fluo-4 (Invitrogen). The dye wasadded at a final concentration of 10 mM into standard Tyr-ode’s solution (Sigma) for 60 min before the onset of theexperiments. Images were acquired using a CMOS camera(pco.1200hs; PCO AG) equipped with an image intensifierunit (C8600; Hamamatsu) and connected to a macro-viewMVX10 Olympus microscope. Data were acquired at 50frames per second using 480 · 480-pixel resolution. Thefluorescent dye was excited at l = 490 nm using the micro-scope’s light source unit outfitted with a mercury lamp and ablue bandpass filter. The same blue light source was used toinduce the trans-conformation of AzoTAB. To induce thetrans- to cis-transition, we utilized a 365-nm UV-LED module(LC-L2; Hamamatsu). The power densities of blue and UVlights at the culture level were 3.5 and 7 mW/cm2, respec-tively. The images were analyzed by using an ImageJ imageprocessing system (NIH). In several experiments, we usedelectrical stimulation to induce an excitation wave. Rectan-gular pulses with amplitudes ranging from 1.5 to 4.0 V and aduration of 20 ms were delivered via a 1-mm nonpolarizingplatinum electrode.

Spectral measurements

The absorption spectra of AzoTAB isomers were recordedin Tyrode’s solution using a photospectrometer (U0080D;Hitachi). Before spectra recordings, the samples were illu-minated for 1 min using either 490 nm blue or 365 nm UVlight. To measure AzoTAB binding to the cell surface, weused a lazer Raman spectrometer (Nanofinder 30; TokyoInstruments) with a 532-nm lazer. The laser light was tightlyfocused on the cell surface to maximize the signal. BeforeRaman spectra measurements, all cell culture samples weredried under a mild vacuum condition.

Measurement of APs

For electrical recordings, glass coverslips containing ad-herent neonatal rat cardiomyocytes were transferred to anexperimental chamber mounted on the stage of an invertedmicroscope (CKX-41; Olympus). The chamber was perfusedin a bath solution (pH 7.4) containing millimolar concentra-tions of the following chemicals: NaCl, 140; MgCl2, 1; KCl, 5;CaCl2, 2; HEPES, 5; and glucose, 10. The patch pipettes werefilled with a solution containing millimolar concentrations ofthe following chemicals: KCl, 120; NaCl, 10; MgATP, 3;HEPES, 5 (pH 7.2); and EGTA, 10 (pH 7.2). Patch pipetteswere pulled from glass capillaries (code GC150F-10; HarvardApparatus) by using a horizontal puller (P-97; SutterInstruments). The resistance of the filled pipettes measured2–5 MO. A whole-cell current clamp method was used tocontrol and record the membrane potential, and APs insingle cardiomyocytes were generated by using electricalstimulation (2 ms; 10 mV) via the patch electrode at intervalsof 2 s. All recordings were performed at room temperature(*22�C).

Voltage and current signals were recorded using patchclamp amplifiers (Axopatch 200B; Molecular Devices). Thecurrent signals were digitized using a 16 bit A/D converter(Digidata 1440A; Molecular Devices) and stored on a per-

sonal computer. Data acquisition and analysis were per-formed using pClamp10 software (Molecular Devices).

Results

Light-induced isomerization of AzoTAB

The AzoTAB has two isoforms, which are schematicallyillustrated in Figure 1A.21 The absorption spectrum of thetrans-isomer (blue line) has a pronounced peak at 365 nm(Fig. 1B). Illumination of this isomer at its absorbancewavelength causes the transition to the cis-conformation,which is characterized by the spectrum shown in purple (Fig.1B). The spectrum of the cis-isomer has an additional peaknear 450 nm. Illumination at this wavelength causes cis-to-trans isomerization. The trans-isomer conformation is ener-getically more stable than the cis-isomer, and the cis-isomerspontaneously reverses to a trans-conformation in the ab-sence of UV illumination. In addition, cis- and trans-isomershave different effects on the excitability of cardiomyocytes,which enables photocontrol of electrical excitation and dy-namic patterning of excitable networks in confluent mono-layers of cardiomyocytes.

Photocontrol of wave propagation

Figure 2 shows the stimulation threshold as a function ofthe concentration of AzoTAB under constant blue light illu-mination. Increasing the AzoTAB concentration from 0.1 to0.3 mM causes a threshold increase from 1.5 to *3 V. Ourexperiments show that concentrations of the trans-isomer ofAzoTAB greater than 0.3 mM increase the excitation thresh-old. In addition, concentrations more than 0.5–0.6 mM cause

FIG. 1. Photoisomerization of AzoTAB. (A) Schematic il-lustration of isomerization of AzoTAB: trans- (left) and cis-(right) isomers of AzoTAB. (B) Absorbance spectrum of0.05 mM AzoTAB in Tyrode’s solution. AzoTAB, azobenzenetrimethylammonium bromide. Color images available onlineat www.liebertonline.com/tea

2 MAGOME ET AL.

complete blockade of AP propagation, so it was not possi-ble to measure the excitation threshold (or it may be re-garded as approaching infinite value). The spontaneousactivity of cardiomyocytes is generally suppressed at lowerconcentrations of AzoTAB (0.1–0.2 mM) compared with theconcentrations required to block the wave propagation. In-terestingly, concentration-dependent changes of the excita-tion threshold for AzoTAB are almost identical to thoseobserved with the local anesthetic Lidocaine (for concentra-tions up to 0.15 mM). In contrast, application of the potas-sium channel blocker tetraethylammonium chloride atconcentrations up to 1 mM did not change the thresholdamplitude. Since the initiation of an AP in cardiac cells isdue to sodium and calcium currents, this is strong evidencethat the trans-isomer of AzoTAB affects sodium and/or cal-cium channels. Unlike the trans-isomer, the cis-isomer ofAzoTAB does not significantly affect the excitability of car-diac cells. Illumination of AzoTAB-treated cultures with UV(7.0 mW/cm2) was sufficient to restore the propagationthreshold for the whole range of tested AzoTAB concentra-tions (0.1–0.6 mM). Importantly, the normalization of excit-ability under UV illumination can occur in the presence ofsimultaneous blue illumination at 3.5 mW/cm2.

Figure 3 shows AzoTAB-mediated light-dependent sup-pression of excitation waves in the cardiomyocyte mono-layer. AzoTAB was added to the incubating Tyrode’ssolution to a final concentration of 0.6 mM. In this experi-ment, the whole Petri dish was illuminated with blue light.In addition, one half of the dish was exposed to UV light;whereas the other half was shielded from UV light using aUV-cutoff filter (Fig. 3A). Spontaneous activity and wavepropagation were only present in the UV-exposed half of themonolayer. Indeed, waves failed to enter the shielded regionof the tissue (upper row, Fig. 3B). Interestingly, this patternchanged dramatically when the UV filter was removed(lower row, Fig. 3B). After 7.5 s, the excitation waves startedpenetrating the previously shielded area, and they eventu-ally encompassed the entire tissue layer. The detailed dy-

namics of this process can be seen in the spatiotemporal plot(Fig. 3C) generated based on the vertical scan line positionedat the center of the frame.

We were able to repeat the effect illustrated in Figure 3several times by moving the UV filter in and out. Within oneexperiment, it was possible to observe up to 5–7 cycles ofsuppression and recovery. Figure 4 shows the dependence ofspontaneous excitation propagation as a function of time inone representative experiment in which the UV illuminationwas periodically turned on and off. The duration of eachcycle was *40 s (20 s of UV illumination followed by 20 swithout UV illumination). The electrical propagation wasrecovered after UV illumination had been turned on, therecovery took 8 s in the first cycle and then occurred 4 s afterUV illumination was turned on in the subsequent cycles.Excitation disappeared after 5–7 s when UV illumination wasturned off, but there was no noticeable decreasing or in-creasing trend. The transitions themselves were very abrupt.After switching to the excitable state, the propagation speedalmost instantly reached 60–80 mm/s with relatively smalland irregular subsequent changes.

Dynamic control of network topology

We were able to dynamically define the network topologyand its dimensions through the use of elaborate masks forUV illumination patterning. Figure 5 shows one experimentin which UV illumination was patterned in the form of thekanji character dai, which means ‘‘big’’ or ‘‘great’’ (it is thefirst part of the Japanese word daigaku or university). Pro-jecting this pattern onto the confluent monolayer of cardiacmyocytes produced an excitable network that had the shapeof the dai character (the rest of monolayer was left unexcited).The excitation originated in the bottom left corner andgradually filled the rest of the excitable domain (red arrowsindicate the direction of propagation), which revealed theoverall shape of the domain.

Creation of an artificial pacemaker

Simultaneous deployment of blue and UV lights and ad-justment of the irradiation power for each wavelength en-abled the complete extinction of excitation in part of thetissue, as shown above. By changing the power ratio, it wasalso possible to create excitation sources in desired locationswhile keeping the entire tissue in an excitable state. Figure 6shows the creation of a pacemaker by local UV irradiation. Inthis experiment, the cultured cells were treated with 0.1 mMAzoTAB (much lower than the 0.5 mM concentration usedfor complete blockage of excitation) and continuously illu-minated with the blue light, which was sufficient to suppressany spontaneous activity without affecting propagation ofthe excitation waves. Subsequently, a spot of UV light wasprojected onto the small area of the tissue culture, whereasthe entire tissue continued to be illuminated with the bluelight (Fig. 6A). As shown in the sequential snapshots of theexcitation wave in Figure 6B, the UV light, indicated as whiteand black circles, locally removed the suppression of spon-taneous activity in the selected spot (i) and converted it into apacemaker that was capable of generating excitation waves.After repositioning the UV beam from location (i) to (ii), thepacemaker stopped and moved to the new location (ii),which was determined by the UV-illuminated spot.

FIG. 2. Dependence of the cardiomyocyte excitationthreshold on AzoTAB, Lidocaine, and TEA concentration.We plotted the critical values of electrical potential to initiateexcitation waves under blue light without UV light. Anelectrical pulse was applied for 20 ms with a platinum elec-trode, and we measured the potential to initiate a propa-gating wave. Although tetraethylammonium chloride (TEA)does not show an increasing threshold, Lidocaine and Azo-TAB show the same trend.

PHOTO-CONTROL OF EXCITABILITY OF CARDIOMYOCYTE 3

Binding AzoTAB isomers to the excitable membranesof cardiomyocytes

Our data suggest significant differences in binding affini-ties of AzoTAB isomers to the excitable membranes of car-diomyocytes. The trans-isomer, which causes suppression ofexcitability, is more strongly bound to the membrane thanthe cis-isomer. An important piece of evidence supportingthis hypothesis is provided by measurements of the washoutkinetics of cis- and trans-AzoTAB isomers. The trans-isomerof AzoTAB can be washed out within several minutes,whereas the cis-isomer washout kinetics were much faster(within seconds). Both cis- and trans-washout kinetics weremeasured after we had applied 0.5 mM AzoTAB in the trans-conformation, which caused complete propagation block.The depletion of membrane-bound AzoTAB was assessedfrom the degree of recovery of conduction velocity. In thecase of the cis-isomer, cells were irradiated using UV lightbefore and during the washout, which induced trans- to cis-transformation.

Figure 7 shows an experiment in which we evaluated thewashout kinetics of the cis-isomer. After 60 s of UV illumi-nation followed by 30 s of washout under UV, the UV illu-mination was turned off and replaced by blue light. Theillumination switch did not cause propagation block, whichwas typical for this transition in the presence of AzoTAB (seeFig. 4). The simplest interpretation of this result is that thecis-isomer does not strongly bind to the membrane. UnderUV light, AzoTAB molecules acquire the cis-conformation,detach from the membrane, and are rapidly removed fromthe tissue culture during washout.

Raman spectra measurements23,24 showed preferentialbinding of the trans-form of residual AzoTAB molecules tothe cell membrane (i.e., there was a lack of binding for the cis-form) after washout (Fig. 8) We used two washout protocols:trans-isomer washout and cis-isomer washout, which werevery similar to the washout protocols just described (Fig. 7).In both cases, the cultures were treated with 0.5 mM AzoTABin Tyrode’s solution and irradiated with blue light for 60 sto establish identical initial conditions. The trans-isomer

FIG. 3. Propagation of exci-tation waves in the photo-sensitive cardiomyocytetissue culture. (A) A sche-matic illustration of a par-tially illuminated system ofdual lights (left) and photo-graphs taken when lightswere irradiated (right). Thehalf of the culture dish thatwas covered by the UV-cutofffilter was transparent to visi-ble light. (B) This panel showsthe fluorescence images ofpropagating excitation waves.The elapsed time from thestart of UV irradiation is in-dicated at the upper right ofeach figure frame. Upper row:waves are only observed inthe lower part of the frameswhere tissue was exposed toUV light. Lower row: after thecutoff UV shield is removed,the excitation waves enter thepreviously shielded area. (C)Spatiotemporal plot of prop-agating waves at positionsindicated by white arrow.[AzoTAB] = 0.6 mM.

4 MAGOME ET AL.

washout was conducted under 15 mW/cm2 blue light illu-mination. The culture was rinsed thrice with AzoTAB-freeTyrode’s solution for 1 min and dried under vacuum con-ditions. The cis-isomer washout protocol included irradiationusing 15 mW/cm2 of UV light for 60 s before the washout.Raman spectra of cells that were not exposed to AzoTAB andAzoTAB solution (in the absence of cells) were used asnegative and positive controls, respectively.

Figure 8 compares Raman spectra of the residual amountof AzoTAB after washout in the trans- and cis-forms. Ra-man peaks of AzoTAB clearly decreased in intensity in cellsirradiated with UV light compared with cells irradiatedwith blue light. To quantify these differences, we normal-ized the spectra using the intensity of the 1638 cm - 1 Tyr-ode’s solution line. The comparison of normalized spectrashowed that the residual AzoTAB in the case of the cis-washout protocol was 70%–100% lower than the case of thetrans-washout protocol. The spectral analysis confirmed thatthe cis-isomer of AzoTAB was easily removed from the cellsduring incubation of the cell layer in AzoTAB-free Tyrode’ssolution. The Raman spectrum of the AzoTAB moleculedissolved in Tyrode’s solution (0.5 mM) was compared withthe spectra of AzoTAB pure powder. We observed fourintense lines at 1152, 1159, 1195, and 1415 cm - 1 in both thepowder and solution spectra. To verify that we could usethese lines to track the molecule in either the cis or transconformation, we checked the effect of alternately irradiat-ing a solution sample using blue and UV lights for 60-spulses. Although the color of the solution clearly changesafter each exposure from pale yellow under blue light todark yellow under UV light, no change in the Ramanspectrum was observed. We also noticed a broad and in-tense line at 1638 cm - 1, which was a result of the Tyrode’ssolution.

Ionic currents involved in AzoTAB-mediated effects

To gain insight into the AzoTAB-mediated changes inexcitability, we recorded APs in isolated AzoTAB-treatedcardiomyocytes under various illumination conditions.Figure 9 shows AzoTAB-induced changes in APs recorded

FIG. 5. Light-controlledwave propagation path incardiomyocyte tissue culture.(A; Left) A schematic illus-tration of ‘‘partial excitation’’by masking UV light. Blueand UV lights were illumi-nated from the top and bot-tom, respectively. (B)Propagation of spontaneousexcitation wave in the area ofUV illumination. The excitedregion is shown by pseudo-color depending on fluores-cence intensity. Elapsed timesfrom the initiation of wavesare indicated in the pictures.(C) This panel show theoverlapping image of the po-sition of the wave frontshown in B. Red arrows in-dicate the direction of propa-gation. [AzoTAB] = 0.7 mM.

FIG. 4. Reversible excitability by alternative light illumi-nation. Solid circles show the speeds of the propagatingwaves generated by spontaneous activity (without anystimulation). Blue light (4 mW) was permanently illuminatedduring the experiment, and 6 mW of UV was irradiated atthe violet region. The shaded portion indicates the period ofspontaneous wave generation. The waves were initiatedunder UV irradiation and suppressed under blue light usingan induction period of several seconds. [AzoTAB] = 0.5 mM.Color images available online at www.liebertonline.com/tea

PHOTO-CONTROL OF EXCITABILITY OF CARDIOMYOCYTE 5

from a single cardiomyocyte paced at a constant frequencyof 0.5 Hz using the whole-cell current clamp method. Incontrol conditions, we recorded APs with upstrokes from- 70 mV resting potential to + 40 mV with an *500 ms APduration (black line). Application of AzoTAB (0.5 mM)caused complex changes in AP morphology. Within 2 minafter AzoTAB administration, AP duration decreased to*100 ms, which was accompanied by a substantial delay inupstroke formation indicated by a reduction of its slope(orange line) and a reduction of its slope. Further exposureto AzoTAB led to a complete suppression of the AP (redline). All these effects of AzoTAB were readily reversible byUV radiation for 1.5 to 2 min (30 mW/cm2, green line).When UV irradiation was initiated 11 min after the appli-cation of AzoTAB, the AP recovered within 30 s; however,this recovery was suppressed by the termination of UVexposure.

Discussion

Sensitization of voltage-gated channels to light25–27 provi-des a powerful tool for precisely controlling tissue excitabilityat a single-cell level.28 Among the prospective methods forlight sensitization (e.g., optogenetics29–32 and use of cagedmolecules33), studies suggested the application of photo-switches, such as derivatives of azobenzene.10,11 Azobenzenecompounds have been shown to reversibly block acetylcholinereceptors,12–14 to block/unblock potassium channels byswitching the irradiation wavelength,6 and to control patternsof neural activity.15,16

The suggested mechanism of light-dependent blockade ofion channels is based on the concept that photoisomerizablemolecules of azobenzene derivatives consist of an azo-benzene group flanked by a covalent attachment group anda group able to ‘‘plug’’ and block open ion channels

FIG. 6. Photo positioningof the artificial pacemaker.(A) A schematic illustrationof the system. The UV is il-luminated to right side of theculture through a hole (4 mmin diameter) from the bottomside. The hole, which wasinitially positioned at theright side, was shifted to theleft side during the course ofthe experiment. Blue lightwas continuously illuminat-ing the culture from the top.(B) Photographs of propa-gating waves from the UV-created wave sources. (C)Overlapping images of theposition of the wave frontshown in B. White arrowsshow the direction of prop-agation. [AzoTAB] = 0.1 mM.

6 MAGOME ET AL.

(quaternary ammonium group).10,15 According to thisscheme, the covalent attachment group binds with thechannel complex protein; whereas the ‘‘plug’’ blocks thechannel. While undergoing a trans- to cis-transition, themolecule bends and pulls the ‘‘plug’’ from the channel.10,15

Thus, the photoswitching molecule should exhibit specificityin both binding to the channel protein complex and pluggingthe corresponding ion channel. The length of the moleculeshould also fit precisely in the ion channel. In addition, co-valent binding should result in a relatively strong connectionof the azobenzene complex to the surface of the cell.

Our data demonstrated that the mechanism of AzoTABinteraction with ion channels for cardiac cells might besignificantly different than other cell types. The trans-configuration of AzoTAB elicited the maximal effect, whichgreatly affected (a) spontaneous electrical activity in cardiactissue, (b) the propagation speed of excitation, with amodest effect on the maximal captured frequency in thetissue (data not shown), and (c) the delayed upstroke shapeof the AP, which indicates suppression of fast sodium chan-nels. These data suggest that AzoTAB may affect voltage-gated fast sodium channels and/or calcium channels incardiomyocytes; thus, AzoTAB could serve as a reversible,light-controlled excitation inhibitor. We suspected that thecis-isomer of AzoTAB weakly binds to the ion channels,because it is easily washed out. In the trans-form, theAzoTAB molecule is less polarized and might bind to the

membrane via hydrophobic interactions; whereas the cis-form is suspended in the surrounding solution. We proposethe following hypothesis of nonspecific AzoTAB bindingwith the surface of the cells: in the trans-form, AzoTABnonspecifically attaches to the cell membrane and globallyinterferes with the functioning of ion channels, whereas thecis-form of AzoTAB leaves the membrane, which restoresion channel activity. Interestingly, a previous study reportedthat the interfacial tension of AzoTAB in water changesdepending on the conformation of AzoTAB.20 Since Azo-TAB molecules adsorb on the cell surface, changes in in-terfacial tension may affect cell activity, earlier it was shownthat changes in interfacial tension may affect functioning ofvoltage-gated ion channels.34,35 Further data on the mech-anism of AzoTAB-mediated light sensitivity of cardiomyo-cytes can be obtained from single-cell experiments.

Conclusion

Our experiments demonstrated the possibility of sensitiz-ing cardiomyocytes to light illumination using AzoTAB. In alight-sensitive tissue, we could execute dynamic patterningof the excitable network. We could reversibly suppress ex-citation waves, create a desired patterned of excitation, andproduce artificial pacemakers. Although AzoTAB is not di-rectly suitable for biomedical applications, one immedi-ate potential application of our method is the modeling of

FIG. 7. Wave propagation during theaddition of AzoTAB and during thewashout. (A) A typical example of a se-quence of speeds of the propagatingwaves generated by spontaneous activity(solid circles). Samples were rinsed thricein AzoTAB-free Tyrode’s solution underUV during the washout period. (B) Thispanel shows the averages and standarddeviations of the propagation speeds forthe 16 samples. At least 10 waves in eachsample were measured. [AzoTAB] = 0.5mM. Power density of UV and blue lightswere 7 and 3.5 mW/cm2, respectively.Color images available online at www.liebertonline.com/tea

PHOTO-CONTROL OF EXCITABILITY OF CARDIOMYOCYTE 7

various geometrical and time-dependent excitation patternsin cardiac tissue by using light instead of the complex andunstable process of patterned plating of cardiac cells. Light-controlled cardiomyocytes may also be used in artificial ex-citable networks for lab-on-a-chip systems.

Acknowledgments

The authors thank Prof. K. Yoshikawa, Kyoto University,and Dr. D. Baigl, Ecole Normale Superieure, for their helpfuldiscussions. They would also like to thank Prof. S. Murata,Dr. A. Zinchenko, and Mr. Y. Kuroda, Nagoya University,for the synthesis of the azo compound. This work is partiallysupported by KAKENHI-21340120 ( Japan Society for thePromotion of Sciences).

Disclosure Statement

No competing financial interests exist.

References

1. Agladze, K., Kay, M.W., Krinsky, V., and Sarvazyan, N.Interaction between spiral and paced waves in cardiac tis-sue. Am J Physiol Heart Circ Physiol 293, H503, 2007.

2. Vaquero, M., Calvo, D., and Jalife, J. Cardiac fibrillation:from ion channels to rotors in the human heart. HeartRhythm 5, 872, 2008.

3. Carmeliet, E., Fozzard, H.A., Hiraoka, M., Janse, M.J., Oga-wa, S., Roden, D.M., Rosen, M.R., Rudy, Y., Schwartz, P.J.,Matteo, P.S., Antzelevitch, C., Boyden, P.A., Catterall, W.A.,Fishman, G.I., George, A.L., Izumo, S., Jalife, J., January,C.T., Kleber, A.G., Marban, E., Marks, A.R., Spooner, P.M.,Waldo, A.L., Weiss, J.M., and Zipes, D.P. New approaches toantiarrhythmic therapy: emerging therapeutic applicationsof the cell biology of cardiac arrhythmias. CardiovascularRes 52, 345, 2001.

4. Jalife, J. Ventricular fibrillation: mechanisms of initiation andmaintenance. Annu Rev Physiol 62, 25, 2000.

5. Cabo, C., Pertsov, A.M., Baxter, P.W., Davidenko, J.M.,Gray, R.A., and Jalife, J. Wave-front curvature as a cause ofslow conduction and block in isolated cardiac-muscle. CircRes 75, 1014, 1994.

6. Tung, L., and Cysyk, J. Imaging fibrillation/defibrillation ina dish. J Electrocardiol 40, S62, 2007.

7. Fast, V.G., and Ideker, R.E. Simultaneous optical mapping oftransmembrane potential and intracellular calcium in myo-cyte cultures. J Cardiovasc Electrophysiol 11, 547, 2000.

8. Isomura, A., Horning, M., Agladze, K., and Yoshikawa, K.Eliminating spiral waves pinned to an anatomical obstacle incardiac myocytes by high-frequency stimuli. Phy Rev E 78,

066216, 2008.9. Rohr, S., and Salzberg, B.M. Multiple-site optical recording

of transmembrane voltage (MSORTV) in patterned growthheart cell cultures: assessing electrical behavior, with mi-crosecond resolution, on a cellular and subcellular scale.Biophys J 67, 1301, 1994.

10. Banghart, M.R., Mourot, A., Fortin, D.L., Yao, J.Z., Kramer,R.H., and Trauner, D. Photochromic blockers of voltage-gated potassium channels. Angew Chem Int Ed 48, 9097,2009.

11. Sjulson, L., and Miesenbock, G. Photocontrol of neural ac-tivity: biophysical mechanisms and performance in vivo.Chem Rev 108, 1588, 2008.

12. Bartles, E., Wassermann, N.H., and Erlanger, B.F. Photo-chromic activators of the acetylcholine receptor. Proc NatAcad Sci USA 68, 1820, 1971.

13. Krouse, M.E., Lester, H.A., Wassermann, N.H., and Erlan-ger, B.F. Rates and equilibria for a photoisomerizableantagonist at the acetylcholine receptor of ElectrophorusElectroplaques. J Gen Physiol 86, 235, 1985.

FIG. 8. Raman spectra of the monolayer surface after AzoTABwashout. The spectrum of AzoTAB (2 mM) in Tyrode’s solution(orange) and the spectrum of untreated cultures (black) were usedas positive and negative controls, respectively. Blue and purplelines show the spectra after different washout protocols. The blueline corresponds to the case in which the washout was conductedunder blue light (trans-isomer washout). The purple line corre-sponds to the case when we used a UV illumination protocol (cis-isomerwashout).Theamplitude of AzoTAB-specific spectral linesin the case of cis-isomer washout (purple) was significantly lowercompared with the trans-isomer washout (blue). Color imagesavailable online at www.liebertonline.com/tea

FIG. 9. Effect of AzoTAB on AP formation in rat neonatalmyocytes. Superimposed representative traces of APs re-corded within the same rat neonatal myocyte in control con-ditions and during exposure to AzoTAB. AP, action potential.

8 MAGOME ET AL.

14. Nargeot, J., Lester, H.A., Birdsall, N.J.M., Stockton, J., Was-sermann, N.H., and Erlanger, B.F. A photoisomerizablemuscarinic antagonist. J Gen Physiol 79, 657, 1982.

15. Fortin, D.L., Banghart, M.R., Dunn, T.W., Borges, K., Wa-genaar, D.A., Gaudry, Q., Karakossian, M.H., Otis, T.S.,Kristan, W.B., Trauner, D., and Kramer, R.H. Photochemicalcontrol of endogenous ion channels and cellular excitability.Nat Methods 5, 331, 2008.

16. Banghart, M., Borges, K., Isacoff, E., Trauner, D., and Kra-mer, R.H. Light-activated ion channels for remote control ofneuronal fireing. Nat Neurosci 7, 1381, 2004.

17. Lee, C.T., Jr., Smith, K.A., and Hatton, T.A. Photocontrol ofprotein folding: the interaction of photosensitive surfactantswith bovine serum albumin. Biochemistry 44, 524, 2005.

18. Magome, N., and Agladze, K. Patterning and excitability controlin cardiomyocyte tissue culture. Physica D 239, 1560, 2010.

19. Estevez-Torres, A., Crozatier, C., Diguet, A., Hara, T., Saito,H., Yoshikawa, K., and Baigl, D. Sequence-independent andreversible photocontrol of transcription/expression systemsusing a photosensitive nucleic acid binder. Proc Natl AcadSci USA 106, 12219, 2009.

20. Diguet, A., Guillermic, R.-M., Magome, N., Saint-Jalmes, A.,Chen, Y., Yoshikawa, K., and Baigl, D. Photomanipulation ofa droplet by the chromocapillary effect. Angew Chem Int Ed48, 9281, 2009.

21. Le Ny, A.L., and Lee, C.T., Jr. Photoreversible DNA con-densation using light-responsive surfactants. J Am ChemSoc 128, 6400, 2006.

22. Hayashita, T., Kurosawa, T., Miyata, T., Tanaka, K., andIgawa, M. Effect of structural variation within cationic azo-surfactant upon photoresponsive function in aqueous solu-tion. Colloid Polym Sci 272, 1611, 1994.

23. Trotter, P.J. Azo dye tautomeric structure determined bylaser-raman spectroscopy. Appl Spectrosc 31, 30, 1977.

24. Matazo, D.R.C., Ando, R.A., Borin, A.C., and Santos, P.S.Azo-hydrazone tautomerism in pronated aminoazo-benzenes: resonance raman spectroscopy and quantum-chemical calculations. J Phys Chem A 112, 4437, 2008.

25. Banghart, M.R., Volgraf, M., and Trauner, D. Engineeringlight-gated ion channels. Biochemistry 45, 15129, 2006.

26. Gorostiza, P., and Isacoff, E. Optical switches and triggersfor the manipulation of ion channels and pores. Mol Bio-systems 3, 686, 2007.

27. Gorostiza, P., and Isacoff, E.Y. Nanoengineering ion chan-nels for optical control. Physiologica 23, 238, 2008.

28. Gorostiza, P., and Isacoff, E.Y. Optical switches for remoteand noninvasive control of cell signaling. Science 322,

395, 2008.29. Zhang, Y., and Oertner, T.G. Optical induction of synaptic

plasticity using a light-sensitive channel. Nat Methods 4,

139, 2007.30. Fiala, A., Suska, A., and Schluter, O.M. Optogenetic ap-

proaches in neuroscience. Curr Biol 20, R897, 2010.31. Knollmann, B.C. Pacing lightly: optogenetics gets to the

heart. Nat Method 7, 889, 2010.32. Arrenberg, A.B., Stainier, D.Y.R., Baier, H., and Huisken, J.

Optogenetic control of cardiac function. Science 330,

971, 2010.33. Mayer, G., and Heckel, A. Biologically active molecules with

a ‘‘light switch.’’ Ang Chem Int Ed 45, 4900, 2006.34. Tomohiko, N., Kamiyama, N., and Kashiwayanagi, M.

Modulation of voltage-gated ion channels on SH-SY5Yneuroblastoma by non-ionic surfactant, Cremophor EL. BiolPharm Bull 33, 2013, 2010.

35. Tillman, T.S., and Cascio, M. Effect of membrane lipids onion channel structure and function. Cell Biochem Biophys38, 161, 2003.

Address correspondence to:Konstantin Agladze, Ph.D.

Institute for Integrated Cell-Material SciencesKyoto University

606-8501 KyotoJapan

E-mail: k.agladze@fy8.ecs.kyoto-u.ac.jp

Received: December 24, 2010Accepted: June 15, 2011

Online Publication Date: August 11, 2011

PHOTO-CONTROL OF EXCITABILITY OF CARDIOMYOCYTE 9