869

Evidence for Two Components of SodiumChannel Block by Lidocaine in Isolated

Cardiac MyocytesC.W. Clarkson, C.H. Follmer, R.E. Ten Eick, L.M. Hondeghem, and J.Z. Yeh

The effects of lidocaine on sodium current in cardiac myocytes isolated from cat and guinea pigwere investigated using the whole-cell variation of the patch-damp technique. Lidocaine (43-200 (iM) reduced sodium current during repetitive depolarizing pulses in a use-dependentmanner. To clarify the nature of the use-dependent block, we characterized the time course ofblock development using a two-pulse protocol. Two distinct phases of block development werefound: a rapid phase ( T = 1-6 msec) having a time course concurrent with the time course ofchannel activation, and a slower phase ( T = 100-900 msec), which developed after channelsinactivated. The amplitude of the block during the rapid phase of development was a steepfunction of transmembrane voltage over the range of - 70 to +20 mV. The voltage-dependencewas similar to that for sodium channel activation (sodium conductance) but was too steep to beattributed solely to the passive movement of a singly charged molecule under the influence of thetransmembrane voltage gradient. These results suggest that use-dependent block of sodiumchannels in cardiac tissue may result from an interaction of lidocaine with sodium channels inthe activated as well as the inactivated channel states. Possible mechanisms underlying the fastcomponent of block are discussed. (Circulation Research 1988;63:869-878)

Local anesthetic drugs appear to block sodiumchannels in nerve, skeletal muscle, andcardiac tissue in a similar manner. In addi-

tion to producing a resting or "tonic" block,lidocaine-induced block of sodium channels in allthree tissue types has also been shown to be use- orrate-dependent.'-5 Measurements of sodium cur-rent (INi) in both nerve4 and skeletal muscle3 sug-gest that lidocaine's use-dependent block results inpart from interaction of the drug with both activatedand inactivated channel states. Previous studiesbased upon measurements of maximum upstrokevelocity (V,^) of cardiac action potentials have alsosuggested that lidocaine interacts with both acti-vated (i.e., open) and inactivated (i.e., closed) statesof the cardiac sodium channel.67 However, evi-

From the Department of Pharmacology (C.H.F., R.E.T., andJ.Z.Y.), Northwestern University, Chicago, Illinois, the Depart-ment of Pharmacology (L.M.H.), University of California, SanFrancisco, San Francisco, California, and the Department ofPharmacology (C.W.C.), Tulane University Medical School,New Orleans, Louisiana.

Supported in part by National Institutes of Health grantsHL-21672, HL-27026, HL-36096, and HL-32577 and the ChicagoHeart Association. L.M.H. is an Established Investigator of theAmerican Heart Association.

Address for correspondence: R.E. Ten Eick, Department ofPharmacology, School of Medicine, Northwestern University,303 E. Chicago Avenue, Chicago, IL 60611.

Received January 23, 1986; accepted May 23, 1988.

dence consistent with open channel block has notbeen convincingly demonstrated in recent studies inwhich the effect of lidocaine on cardiac INo wasdirectly measured.3-8

The present study with isolated cardiac myocyteswas designed to assess the time course of lidocaineblock of the cardiac IN, and to determine whetherlidocaine block has a component that would beconsistent with block of activated sodium channels.Experiments were performed under voltage clampconditions on single cardiac myocytes using thewhole-cell variation of the patch clamp technique.9

We found that the use-dependent block of cardiacsodium channels in cat and guinea pig myocytes canbe divided into two components: one closely asso-ciated with sodium channel activation in both timecourse and voltage-dependence, and a second thatdevelops following channel inactivation. We havepublished some of our results in preliminary form.10

Materials and MethodsIsolation of Cardiac Myocytes

The method for cell isolation used in this studywas essentially identical to that described by Silveret al." Briefly, hearts from 2-4-kg adult mongrelcats or 200-500-g guinea pigs, of either sex, wereremoved under pentobarbital anesthesia, rinsed ofblood with saline solution (0.9%), and mounted

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

870 Circulation Research Vol 63, No 5, November 1988

immediately on the cannula of a perfusion appara-tus for retrograde perfusion through the aorta(Langendorff perfusion). The heart was initiallyperfused with a modified, nominally Ca2+-free Krebs-Henseleit solution of the following millimolar com-position: NaCl 130, KC1 4.8, MgSO4 1.2, andNaH2PO< 0.33, and warmed to 37° C, equilibratedwith 95% O2-5% CO2, and titrated to pH 7.3 withhydrochloric acid. Once the initial perfusate wasfree of blood, the heart was perfused with recircu-lating Krebs-Henseleit supplemented with 0.15%collagenase (type II, Cooper Biomedical, Malverne,Pennsylvania). After the heart had become ade-quately digested (20-30 minutes for guinea pig; 40-50 minutes for cat), the atria and ventricles were cutaway from each other, minced with scissors, andincubated at 37° C for 10 minutes in the collagenasesolution. Harvesting of atrial and ventricularmyocytes was the same as that described by Silveret al." Cells were incubated in Eagle's MinimumEssential Medium with Earle's Salts containingL-glutamine and supplemented with 0.15 /ug/ml gen-tamicin sulfate. Cells were kept in an incubatormaintained at 37° C and equilibrated with 5% CO2until they were used (on the day of isolation) inelectrophysiological experiments.

At the beginning of each experiment, cells weretransferred from the storage solution to a shallowchamber perfused with a standard external physio-logical solution (described below). The chamberwas mounted on the stage of an inverted micro-scope to permit visual observation of both cells andthe tip of the suction pipette. A thermistor wasplaced in the chamber near the suction pipette andthe solution temperature was kept constant(±0.5° C). Experiments were performed at temper-atures between 14° and 17° C.

Solutions and DrugsThe experimental chamber was perfused with a

low-sodium, N-2-hydroxyethylenepiperazine-/V'-2-ethanesulfonic acid (HEPES)-buffered salt solutionof the following millimolar composition: NaCl 25,tetramethylammonium chloride 115, CaCl2 1.8, CsCl5, MgCl2 1.2, CoCl2 1.0, D-glucose 11, and HEPES20. In some experiments, the external Na+ concen-tration was raised to 50 mM by equimolar substitu-tion for tetramethylammonium chloride. The solu-tion was titrated to a pH of 7.3 ±0.01 with 1 Mtetramethylammonium hydroxide. The solution insidethe suction pipette had the following millimolar com-position: CsF 145, NaF 5.6, and HEPES 5 and wastitrated to a pH of 7.2 ±0.02 with 1 M CsOH. Afteraddition of lidocaine to the external solution, anequilibration period of 5-15 minutes was allowed forthe effects of lidocaine on IN, to approach steadystate. Peak 1^ amplitudes were measured before andafter adding drug. When washout of lidocaine waspossible, IN, recovered completely.

Electrophysiological TechniquesThe voltage clamp circuit used to record mem-

brane currents from cat myocytes was the same asthat designed by M. Yoshii and described else-where.12 An Axoclamp voltage clamp amplifier(Axon Instruments, Burlingame, California) wasused for experiments using guinea pig myocytes.Suction pipettes were made using a vertical pipettepuller (Kopf 700D, Tujunga, California) and werefire polished. When filled with internal solution,pipettes had tip resistances of about 0.3 to 0.5 Mil.On occasion, electrodes with slightly higher resis-tances were used to patch-clamp small atrial cells.

Capacitative transients evoked by 10 mV hyper-polarizing pulses were well-described by single expo-nential functions and had mean time constants of110 ±10 /isec in cat atrial myocytes (n = 9) and85 ±8 /isec in guinea pig ventricular myocytes(n= 13). Mean cell capacitance was estimated fromthe equation C = Q/AV, where Q is the total chargedisplaced during a voltage step, AV. Q was derivedfrom integration of the capacitative transient. Meanvalues of cell capacitance derived by this methodwere 78 ± 10 pF for cat atrial myocytes and 66 ±4pF for guinea pig ventricular myocytes.

The total series resistance for the connectivepathway between the pipette and cell membranewas calculated from estimates of cell capacitanceand the time constant of the capacitative current'sdecay (TC = R, x CJ . The mean series resistance was1.5 ±0.1 Mft for cat atrial myocytes (n = 9), and1.3 ±0.1 MH for guinea pig ventricular myocytes{n = 13). Compensation for this resistance was doneempirically to speed the decay of the capacitativetransient by applying a level of electronic seriesresistance compensation just short of producingoscillation.

Command pulses and data acquisition were con-trolled by either a DEC PDP-11/23 computer or anIBM PC/AT. The membrane current (i.e., INa) sig-nal was filtered at 10 kHz, fed into an A/D converter(sample interval 30-100 ^sec), and the digitizeddata were stored on floppy disks. Since leak cur-rents were relatively small and can be a nonlinearfunction of voltage, no electronic compensation forvoltage-dependent leak current was used.

Evaluation of MethodTo accurately measure the effects of drugs on

cardiac INt, certain requirements must be met.First, all other ionic currents that could contami-nate measurement of INa must be eliminated. Sec-ond, errors resulting from current flowing across aresidual series resistance (R,) must be reduced toacceptable values. Isolation of INa current fromother ionic currents was effectively accomplishedusing Cs+ to replace K+, and by placing Co2+ on theoutside and F~ on the inside of the membrane (see"Solutions and Drugs") to block current throughcalcium channels. Under these conditions, no evi-dence for transient potassium or calcium currents

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

Clarkson et al Lidocaine Block of Sodium Channels 871

was observed at any test potential.13 Also, at theconcentration used, Co2+ did not alter the timecourse or voltage-dependence of INa when examinedin the presence of 1.8 mM Ca2+.

Voltage errors related to the flow of INa across theseries resistance were minimized by a combinationof using low resistance electrodes (300-500 KXI),reducing peak current to values less than 15 nAusing 25-50 mM external sodium concentration andlow temperature (14-17° C), and by electronic seriesresistance compensation. To test for the presenceof uncompensated series resistance, we defined thepeak sodium conductance (GNa) versus Em relationat holding potentials of -140 and -90 mV. If asignificant uncompensated series resistance errorexists, reduction of peak INa amplitude by partialinactivation should cause both a depolarizing shiftof the GN, versus Em relation, as well as a decreasein its slope. We calculated GN, using a Hodgkin-Huxley model (where GNa = INa/ [Em-ENa]). The rela-tion between GNa and Em was approximated by aleast-squares fit to Equation 1:

N. = GMai/{l+exp[(Emid-EJ/S]} (1)

where GM M is the maximal value of GNa, Ema is thevoltage at which GNa is half maximal, and S is aslope factor.

In ventricular myocytes, the peak IN, amplitudeduring pulses to - 20 mV from a holding potential of-140 mV was 10.1+2.1 nA (mean±SEM; n = 6).Reducing the holding potential to - 90 mV reducedmaximum GNa by 44 ± 4%, but produced only a verysmall shift (1.6±0.5 mV) of E ^ (from -36.2 to- 34.6 mV), and a reduction of the slope factor from6.9±0.7 mV to 6.2 + 0.8 mV, indicating a slightsteepening of the slope. These small changes sug-gest that the measurement error due to uncompen-sated series resistance was within acceptable limits.Similar results were obtained after inhibiting approx-imately two thirds of INa with tetrodotoxin.1314

To evaluate the adequacy of voltage clamp con-trol we estimated the length constant (A) at the timeof peak sodium conductance. According to Belles etal,15 the length constant can be approximated by thefollowing equation:

= V(2-L-q-Ri n /Ri) (2)

where L is cell length, q is cross sectional area ofthe cell, Ru, is average input resistance, and Rj isspecific resistivity of the cell interior. Rjn at peaksodium conductance, calculated from the positiveslope of the current voltage relation, was 9.9±1.5x10* fl (mean±SEM; n = 10). L and q wereassumed to be 98 fxm and q = 314/xm2l6and Rj = 605Hem.17 With these values, the estimated lengthconstant at peak sodium conductance is 317 ^m,more than five times the distance from the pipettetip to the end of a typical myocyte when the pipetteis placed centrally along the cell length.

Thus, we conclude that problems concerning INaresolution, voltage errors due to series resistance,

and nonuniformity are within acceptable limits underour experimental conditions and should not pre-clude interpretation of the data.

Statistics and Data AnalysisFits of experimental data to exponential equa-

tions were performed with a nonlinear least-squaresfitting algorithm. Paired comparisons between datapoints were done using Student's / test. Grouped ormultiple comparisons were made using an analysisof variance with a Scheffe test for critical differ-ence. Differences were considered significant atp<0.05. All results are expressed as mean±SEM.

ResultsTonic Block

Lidocaine had three distinct effects on the steady-state sodium current availability curve. It reducedthe maximum available current, increased the slopefactor (k), and shifted the voltage-dependence of INato more-negative potentials. Figure 1 shows a rep-resentative example of the effects of lidocaine onthe steady-state availability curve for INa in cat atrialcells. In the example shown, the maximum avail-able current was reduced 14% by lidocaine. Thisblock, produced at very negative potentials and atlow stimulus rates, was defined as "tonic block." Incat atrial myocytes, 200 jiM lidocaine reduced INaby 14 ± 1.3% (mean ± SEM; n = 5). A similar level oftonic block was also observed in guinea pig ventric-ular myocytes exposed to 200 .̂M lidocaine(23 ±3.5%, n = 4). As illustrated in Figure 1, typi-

controJ

-i 1 1

-150 -130 -110 -50-70Prepulse Potential (mV)

FIGURE 1. Effects of lidocaine on the steady-state sodiumcurrent (INJ availability curve measured for cat atrialcell. Peak Na* current elicited by a test pulse to —20 mVis plotted as a function of the potential during a 10-second prepulse. Both under control conditions (trian-gles) and during application of 200 fiM lidocaine (squares),the data were well-fitted to lines derived from simpleBoltzmann equations (INa=Imaj/{l+exp[(Em-EmUI)/S]},where /mal is maximal INa, Em is the prepulse potential,Emid is the membrane voltage at which INa is half-maximal, and S is a slope factor). During control, 1^,Emid, and S were 6.5 nA, -86 mV, and 5.3 mV per e-foldchange in INa, respectively; during exposure to lidocaine,the values were 5.7 nA, —96mV, and 7.3 mV per t-foldchange, respectively. Temperature, 14° C; extracellularsodium concentration, 50 mM.

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

872 Circulation Research Vol 63, No 5, November 1988

cally under control conditions, the IN, availabilitycurve was maximal (i.e., hw= 1) at — 120 mV, halfmaximal at — 8 6 mV, and 0 at — 5 0 mV. Thesolid line representing a simple Boltzmann equationhas a slope factor of 5.3 mV and adequately describesthe data. In the presence of 200 /xM lidocaine thepotential at which INa was half maximal shifted to~ —96 mV, that is, 10 mV more negative than thecontrol. The slope factor was increased to 7.3 mV inthe presence of lidocaine. On average, 200 /nMlidocaine shifted the steady-state availability curvein the negative potential direction 10 ±1 mV andincreased the slope factor by 1.7 ±0.3 mV (5.7 ±0.4to 7.5 ±0.3; n = 5) in cat atrial myocytes.

Use-Dependent BlockIn addition to producing tonic block of INa, lido-

caine suppressed INa in a use-dependent mannerwhen cells were repetitively stimulated at ratesfaster than 0.5 Hz. Use-dependent block is illus-trated in Figure 2. In the absence of drug, repetitivedepolarization to -20 mV for 20 msec at 2-5 Hzproduced very little decrease in INa (Figure 2B).However, in the presence of 100 fiM lidocaine,there was a substantial decrease in IN, amplitudefollowing a single 20-msec pulse, and peak IN>amplitude decreased further with each pulse until asteady-state level was approached within 12 pulses(Figures 2B and 2C). This block, which is over andabove the tonic block, is referred to as "use-dependent block." An increase in the stimulus rateenhanced use-dependent block (Figure 2C). Theamplitude of use-dependent block at 5 Hz producedby 100-200 /JM lidocaine ranged from 50% to 79%in guinea pig ventricular myocytes (n = 5).

Two Phases of Lidocaine Block DevelopmentCumulative reduction of peak INa during repeti-

tive pulsing is usually thought to result from thedrug blocking sodium channels in either the acti-vated or the inactivated states, or from blockingchannels in both states. To determine in whichstates sodium channels can be blocked by lidocaineduring a depolarizing step, the time course of blockdevelopment was characterized with the two-pulseprotocol shown in the inset of Figure 3A. The blockproduced by conditioning pulses having selecteddurations ranging from 1 msec to 10 seconds wasdetermined by a test pulse applied after a short (500msec) recovery interval. At the holding potential used(—120 mV), the 500 msec recovery interval wassufficient for the drug-free channels to recover fromthe inactivation produced by the conditioning pulse.13

The pulse sequence was applied at 30-second inter-vals to obviate cumulative block by lidocaine and toallow full recovery from slow inactivation.13

In guinea pig ventricular myocytes, under controlconditions, test INo was not appreciably decreasedeven after a 1-second conditioning prepulse (Figure3). In contrast, in the presence of 172 .̂M lidocaine,test INa was decreased markedly even after short

-20 mV

-140 mV

20 ms

B

"n = 15

100 pM Lidocaine5 Hz

<

Control.e » > » « O O 0 O Q O O 0 S O 2 & 5 Hz

Lidocaine:o n o o D B O P D o n i.o Hz

ODDODDDOD 2.0 Hz

3.3 HzSHDDODDO 5.0 Hz

0 4 8 12 16Pulse Number

FIGURE 2. Tonic and use-dependent inhibition of thepeak sodium current (INa) by lidocaine in a guinea pigventricular myocyte. A: Pulse protocol. A train of depo-larizing pulses of 20 msec duration to —20 mV wasapplied at different stimulation rates after a rest of >20seconds. B: Superimposed INa records obtained during atrain of pulses applied at 5 Hz before and during externalexposure to 100 fiM lidocaine. There was no noticeabledecrease in peak current amplitude during control condi-tions, whereas in the presence of lidocaine there was ause-dependent reduction to an apparent steady-statelevel. Lidocaine also produced a small reduction of peakcurrent during the first pulse (referred to as "tonicblock"). C: Shows the relation between peak lNa ampli-tude and number of pulses applied at different stimulationrates under control conditions (circles) and in the pres-ence of 100 \iM lidocaine (squares). The degree ofuse-dependent block increased as the stimulation ratewas increased. Temperature, 16° C; extracellular calciumconcentration, 25 mM. Preparation 9104.

conditioning pulses (1-10 msec). Further increase inprepulse duration enhanced the degree of blockproduced by lidocaine. Two distinct phases of blockonset that can be fit by a sum of two exponentialswere observed, a rapid phase with a time constantof 3 msec accounting for approximately 37% of theblock and a slow phase (T=351 msec) accountingfor the remainder. The rapid phase of block occurredduring the period (0-10 msec at 16° C) when mostsodium channels were, or had been, in the activatedstate, whereas the slow phase of block did notbecome appreciable until most channels had becomeinactivated. Qualitatively similar results wereobtained over a wide range of drug concentrationsin guinea pig ventricular myocytes (Table 1), as wellas in cat atrial myocytes (Figure 3B).

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

Clarkson et al Lidocaine Block of Sodium Channels 873

TABLE 1. Two Components of Lidocaine Block Development

0.2 0.4 0.0 0.8 1.0Prepulse Duration (sec)

1.0

0.8

0.8-

0.4 •

0.2

0.0

\

-101--130 -1

1

f Tt"f..t" 3

^ ^

10.1

iT OOOal

—i y

ms

\

\

- 3 0

control

lidocaine (20C

350ms

.001 10.0.01 0.1 1.0Prepulse Duration (sec)

FIGURE 3. An increase in conditioning pulse durationrevealed that lidocaine's use-dependent block of sodiumcurrent f/jvj developed in two phases. The sodium chan-nels were conditioned by prepulses having selected dura-tions at +20 mV (Panel A) or -30 mV (Panel B). Theextent of block produced by the conditioning prepulseswas assayed by test pulses applied after a 500-msecrecovery interval. The interval between each trial of thepulse protocol was 30 seconds. The peak amplitude of thelNa during the test pulse was normalized to the valueobtained without a conditioning prepulse. Under controlconditions there was little noticeable change in peakcurrent amplitude following conditioning prepulses of upto 1 second duration. In the presence of lidocaine, peakcurrent was inhibited in two distinct phases having con-stants of 3 and 351 msec. Temperature, 14° C; extracel-lular sodium concentration, 50 mM. A: Guinea pig ven-tricular myocyte, plotted with a linear time scale;Preparation 4221. B: Cat atrial myocyte plotted withlogarithmic time scale.

Voltage-Dependence for the Rapid Phase ofLidocaine Block

Previous studies have shown that the slow phaseof block is not strongly voltage-dependent at volt-ages in which sodium channels are fully inactivated(e.g., at membrane voltages positive to -50 mV).5-6

However, the voltage-dependence of the rapid phasewas not defined during these studies. Therefore, thevoltage-dependence of the previously unresolvedrapid phase of block was determined using theprotocol shown in Figure 4A. A train of 2-msecpulses of selected amplitudes (EJ were applied at 3

LidocaineExperiment concentration Ec Tf Amplr'number (MM) (mV) (msec) (msec)

Ampl,*

Cat atrial myocyte

4010

Guinea pig

6101

4221

6101

4221

4111

8251

8252

9102

9104

1014

200ventriculai

258172868643

200200200100100

-30• myocytes

+ 20

+ 20

+ 20

+ 20

+ 20

-20-20-20-20-20

2.9

2.02.66.23.03.61.55.83.03.34.1

39.0

42.7

37.0

9.65.713.5

22.5

28.8

31.5

16.7

20.6

350

177351434696902122193164263241

61.0

57.3

63.0

90.4

94.3

86.5

77.5

71.2

68.5

83.3

79.4

•Percent of total time-dependent block produced during aconditioning pulse at Ec attributed to either the rapid or slowphase (i.e., %Amplf+%Ampl,= 100). Calculated from a two-pulse protocol.

Hz to produce a steady-state level of INa block. Theamount of block produced by the train of pulses tovoltage Ec was then assessed with a test pulse to- 20 mV delivered 300 msec after the conditioningtrain. As illustrated, block was a steep function ofmembrane voltage over a range of conditioningpotentials from -70 to ~ 0 mV. The block exhibitedvoltage-dependence at potentials positive to — 80 mV,had a midpoint near —44 mV, and saturated atapproximately 20 mV. The line drawn through thedata points describing the voltage-dependence oflidocaine-induced block represents an equation thathas been used previously by Cahalan and co-workers,1819 and in a different form by Strichartz,1 todescribe the open-channel block of Na+ channels innerve preparations by local anesthetics:

{(l-B)/[l+exp(Ec-Em i d)/S]} + B (3)

where B is the steady-state level of INa remainingafter the series of prepulses, Ec is the membranepotential during the conditioning pulse, E ^ is themembrane potential at which drug-induced block ishalf-maximal, and S is the slope of the voltage-dependent block.18 This model assumes that chan-nel block results from the passive movement of amonovalent cation under the influence of the mem-brane field to a blocking site within the sodiumchannel. According to this model, any voltage-dependence derived from the opening and closing ofsodium channels is ignored, and the slope factor (S)is equal to RT/ZSF, where F, R, and T have theirusual meaning, Z is the drug molecule's valency,and S is the fractional distance (0 to 1) of the bindingsite across the membrane when measured from theinside. Under our experimental conditions, RT/F = ~24 mV and Z = l . Therefore, the minimumpossible value for the slope factor is 24 mV (when5=1). In our experiments, the slope factor S for

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

874 Circulation Research Vol 63, No 5, November 1988

condition

-140 mV-- 2ms -

test

10ms

0.3 s-20 mV

3 Hzn = 15

BControl 200 uM Lidocaine

~ 0.0L,-140-100 -60 -20 20 60

Ec (mV)

FIGURE 4. Comparison of the voltage-dependence ofuse-dependent block with that of sodium channel activa-tion (GNJ in a guinea pig ventricular myocyte. A: Use-dependent block was produced by a train of 15 condition-ing pulses of 2 msec duration at 3 Hz to selected membranepotentials (Ec). INa available after the conditioning trainwas determined using a test pulse to —20 mV after a fixedrecovery interval of 300 msec. The interval between eachrun of the pulse protocol was 20 seconds. B: SuperimposedlNa recorded during test pulses under control conditions(left panel), and in the presence of 200 fiM lidocaine (rightpanel). Test currents were constant in control, butdecreased progressively in lidocaine as the amplitude ofthe conditioning pulses were increased from —140 to +40mV. C: Shows the relation between peak INa and condi-tioning potential. Peak current was normalized to its valueafter a long rest at Vh=-140 mV. O, control; • , 200 fiMlidocaine. The relation between membrane potential andGut, for the same cell is also shown for comparison. GNc

was calculated using both peak INa (unfilled diamonds) andthe integral of lNa • dt over a 2-msec interval (filleddiamonds). Temperature, 16° C; extracellular sodium con-centration, 25 mM. Preparation 8251.

lidocaine block ranged from 9.3 to 16.1 mV pere-fold increase in block (Table 2). Thus, the steep-ness for voltage-dependent block by lidocaine wasgreater than that predicted for passive movement ofa monovalent cation within the membrane field.

To determine whether the rapid component ofblock had a voltage-dependence similar to channelopening, we compared the voltage-dependence ofchannel block with that of GNl. GN, was calculatedfrom either peak INl, or the integral of IN, • dt duringthe initial 2 msec of depolarizing pulses (i.e., for the

same pulse duration used to define channel block)(Figure 4, Table 2). As shown in Table 2, the bestagreement between channel block and channel con-ductance was obtained using the time integral ofGNa. However, while the voltage-dependencies ofchannel block and of the integrated GN, had similarvoltage midpoints, the slopes for the two relationswere significantly different (p <0.05). For compar-ison, the voltage-dependence of GN, was also definedusing a 50 msec integral of IN, elicited at each of theseveral different voltages. However, the disparitybetween the characteristics of GNa as defined using50 msec integrals (E^ = -51.0± 3.5 mV; slope = 2.7±0.2 mV) and those of lidocaine-induced block( E ^ = -34.7±5.5 mV; slope= 13.7±0.7 mV) wasnot lessened (« = 8).

Effect of Lidocaine on Sodium CurrentTime Course

As illustrated in Figure 5A, lidocaine (100-200fiM) produced a barely perceptible change in thetime course of IN> decay in cells with intact inactiva-tion. Under control conditions, the decay of IN,following peak inward current at - 20 mV could bewell described as the sum of a fast and a slowexponential plus a small constant (Figure 5A, Table3). Exponential fits to the time-course of INa decayafter a 5-20 minute exposure to 200 /iM lidocaineindicated that the rapid and slow time constants wereslightly reduced (by 9 ±9% and 18 ±12%, respec-tively), but this reduction was not significant (p >0.1).The relative fractions of current decay attributed tothe fast and slow components were also not signifi-cantly altered by 200 piM lidocaine (p >0.1). Similarresults were also observed in two additional cellsexposed to 100 /iM lidocaine (Table 3).

Many putative open-channel blockers (includinglidocaine) have been reported to produce relativelysmall effects on lNa time course in nerve31*-21 andcardiac preparations3-8-22 when INa inactivation isintact. In contrast, open-channel blockers havebeen demonstrated to exert very marked kineticeffects in nerve cells in which inactivation has been"removed" (i.e., modified) by pretreatment withchemical reagents or enzymes.19-20 Therefore, weexamined the effects of lidocaine on IN> of myocytesthat had their inactivation process modified byexposure to a-chymotrypsin (0.7 to 0.8 mg/ml,applied intracellularly by addition to the pipettesolution).23 After treatment with a-chymotrypsinfor 30-60 minutes, there was little, if any, decay ofINl during 20 msec depolarizing pulses (Figure 5B).Nevertheless, subsequent exposure to lidocaineinduced a time-dependent decay in IN>. This findingwas obtained in five out of five cells exposed to 100or 200 /iM lidocaine and is consistent with thenotion that lidocaine can block sodium channelswhen they are in an open, or activated, state.

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

Clarkson et al Lidocaine Block of Sodium Channels 875

TABLE 2. Voltage Dependence of Rapid Channel Block and G*.

Experiment

Preparationvh(mV)

A. Cat atrial myocytes

4010

4910

B. Guinea pig

4111

6101

4221

9104

1014

9102

8251

8252

-100

- 1 0 0

[Lidocaine](MM)

2020020200

ventricular myocytes

-120

- 1 2 0

-120

- 1 4 0

- 1 4 0

- 1 4 0

- 1 4 0

-140

438686100100200200200

Mean

±SEM

Lidocaine

En*(mV)

-28 .3

-30 .3

-16 .6

-15.1

-39 .0

-10 .7

-17 .5

-34 .3

-49 .5

-51 .1

-44.1

-31 .5

-34 .7

±5.5

block*

slope(mV)

16.1

12.3

9.311.6

14.1

15.5

15.2

14.1

14.9

13.5

12.4

9.9

13.7

±0.7

B

0.69

0.33

0.61

0.38

0.74

0.82

0.81

0.73

0.60

0.39

0.62

0.74

PeakG

Enid(mV)

-44 .0

-49 .9

-42 .5

-46 .0

-44 .0

-55 .8

- 4 8 . 9

-38 .2

-46 .2

±2.0

N.t

slope(mV)

6.34.33.96.65.18.37.19.4

6.4*

±0.7

C'

J(E-nrid

(mV)

-22 .4

-30 .2

- 1 9 . 3

-37 .7

-34 .4

- 4 0 . 9

-36 .3

- 2 9 . 2

- 3 1 . 3

±2.8

! msec

.G^dtt

slope(mV)

9.17.57.65.96.68.77.28.2

7.6*±0.4

*Block was determined following a train of 15 conditioning pulses of 2 msec duration applied at 2 Hz (cat) or 3 Hz (guinea pig) tovarious levels of Em. The recovery internal between the conditioning train and test pulse was 300 msec for guinea pig and 725 msec forcat. Data were fit with Equation 3 using the method of least squares.

tG N , was calculated from Equation 1 using either peak IN, (peak GNJ or IN, integrated over the initial 2 msec (G^dt) of depolarizingpulses to voltages between - 7 0 and +30 mV. E ^ , voltage at which GN, was half-maximal; Vh, holding potential; slope, the value forS in Equation 3 is the slope of the voltage-dependence for lidocaine-induced block of INa.

tSlope was significantly different than that for channel block, p<0.05.

DiscussionIn 1977, Hondeghem and Katzung24 developed a

set of differential equations for a model derivedfrom the modulated receptor hypothesis. Whenavailable data on the effects of lidocaine on maxi-mum upstroke velocity were fit to the model, itpredicted that lidocaine has a low affinity for sodiumchannels in the closed, rested state (AP

d>l mM) anda relatively high affinity for channels in both acti-vated (open) and inactivated (closed) states (K^ = 30-40 nM). Based upon the association and dissocia-tion rate constants estimated from the model for thethree primary channel states, the lidocaine-induceduse-dependent block that developed during trains ofventricular action potentials was predicted to con-sist of two kinetically different components: a rapidcomponent related to binding of lidocaine to acti-vated (open) channels, and a second slower compo-nent related to lidocaine binding to inactivated(closed) channels.23 Direct evidence for the pres-ence of two components was subsequently obtainedfrom measurement of V ^ of action potentialupstrokes in guinea pig ventricular muscle whenaction potential plateau duration was controlled byvoltage clamp.67

In contrast to the studies using V^,,6-7 recentstudies investigating the effect of lidocaine directlyon IN, during exposure to cool temperature and lowexternal sodium concentration provided strong evi-

dence that lidocaine can bind to inactivated chan-nels, but little5 or no evidence8 that lidocaine canbind to channels in an activated, or open, state.Reasoning that the voltage clamp protocols used inprevious studies were not particularly well adaptedfor detection of open-channel block, in the presentstudy we have reexamined the effects of lidocaineon IN, in voltage clamped cardiac myocytes andfound evidence for an interaction between lidocaineand sodium channels while in or going to an openstate. During a depolarizing pulse, we observed twocomponents of block development (Figure 3). Onecomponent developed rapidly with a time constantof a few milliseconds at potentials positive to thresh-old for channel opening. This rapid component wasthen followed by a slower component having a timeconstant 100-fold larger and which is essentiallyidentical to that previously described by others incooled Purkinje fibers3 and isolated rat myocytes.8

Role of the Membrane Field Acting on theCationic Form

Lidocaine (100-200 p-M) induced a time-dependent decay of IN, in cells whose inactivationhad been slowed or removed by internal applicationof a-chymotrypsin (Figure 5B). This effect of lido-caine on cardiac INa is similar to the reported effectsof a variety of tertiary and quaternary amine localanesthetics documented to block open channels in

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

876 Circulation Research Vol 63, No 5, November 1988

nerve fibers.19-20 Block of open sodium channels innerve fibers has also been shown to be stronglyvoltage-dependent.118-20 This voltage-dependenceof open channel block has been proposed to resultfrom the voltage-sensitive movement of the cationicform of the drug between the cytoplasm and areceptor site which is located partway across thetransmembrane voltage gradient.1819

To determine whether the rapid component oflidocaine block in cardiac tissue could be attributedto a similar mechanism, the voltage-dependence ofthe rapid component of block was defined andfound in the present study to be a steep function ofmembrane voltage between —70 and 0 mV, saturat-ing at ~ + 20 mV (Figure 4, Table 2). However, thedata could not be fit well to an equation describingpassive movement of a cation within a voltagegradient19 unless either an ionic charge greater thanunity or a ratio of drug-channel binding greater than1:1 was assumed. Since lidocaine cannot bear morethan one positive charge, and the available evidencesupports 1:1 stoichiometry for lidocaine block ofsodium channels,525 it seems unlikely that theobserved steepness in the relation between mem-brane voltage and channel block can be attributedto simple passive movement of charged lidocainemolecules across a portion of the transmembraneelectric field. The observed sharp steepness involtage-dependent block by lidocaine is in contrastwith the more shallow voltage-dependence previ-ously reported for etidocaine and tetracaine innerve19 (however, see Reference 21).

Voltage-Dependence of Rapid Component MayResult From Channel Gating

Since the voltage-dependence of channel blockwas too steep to be simply attributed to voltage-dependent movement of the charged form of lido-caine into sodium channels, we explored the possi-bility that the voltage-dependence of block resultsindirectly from state-dependent drug-binding to theopen state of the channels in a manner similar tothat described by Yeh and Ten Eick26 for the blockof INa in squid giant axon by derivatives of disopyr-amide. To evaluate this possibility, we comparedthe voltage-dependence of sodium channel blockwith that for channel opening (as defined peak GN,or by integrated GNl) and found them to be similar,but not identical. Both channel block and the inte-gral of GNa could be described as sigmoid functionsof membrane voltage that develop over the range of- 70 to + 20 mV and have similar midpoints (Table2, Figure 4). This result suggests that the activationgate for the sodium channel plays an important rolein the development and modulation of use-dependentblock of INl. The voltage dependence of channelblock, however, was consistently less steep than thatderived for GN> (Table 2). The meaning of thisdifference is not clear. It is conceivable that otherfactors, such as a competitive interaction betweenlidocaine and monovalent cations (e.g., Na+) for a

Lidocaine (100 uM)

"2ms Control ormalizedLidocaine

FIGURE 5. Effect of lidocaine on INa time course in"normal" (A) and "'inactivation-modified" (B) guineapig ventricular myocytes. Sodium currents were evokedduring test pulses from -140 to -20 mV (see inset) aftera long (>20 seconds) rest. A: Shows fits ofINa time coursein a myocyte with intact inactivation before and 12minutes after exposure to 100 fiM lidocaine. The smoothcurves indicate least-squares fit to the decay phases forpoints between the vertical bars. The equation of fit was

INa = A ^ e ' " ^ + A ^ - * - ' + A.

where Afail, A^^,, and A* represent the amplitudes of thefast, slow, and steady-state components, respectively.For control: A!all= — ll.6 nA, Attm,= -3.47 nA, Ax=—0.3nA> Tfast—1-2 msec, and T^OW=5.5 msec. After lidocaine:Afall=-10.8 nA, Alhw=-2.77 nA, A»=-0.2 nA, Tfasl=1.2msec, and Tjhw=4.8 msec. Preparation 1014. B: Effect of100 uM lidocaine on INa time course in a cell pretreatedwith a-chymotrypsin (0.7 mglml). Lidocaine induced atime-dependent decay of lNc that was not present in thecontrol record. The normalized lidocaine trace was mul-tiplied by a scaling factor such that the activation phaseof the current superimposed with the control trace atearly time intervals. The holding (leak) current for thecurrents shown in Panel B was subtracted by computer.Difference current was obtained by subtracting the normal-ized lidocaine trace from the control trace. Preparation9301. Temperature, 16° C; extracellular sodium concentra-tion, 25 mM. AID sampling rate, 20 kHz (1/50 fjsec).

common binding site, may contribute to the observeddifference between the voltage-dependence of chan-nel gating and channel block. "8.21.22.26.27

Implications of Rapid Open Channel BlockThe results of this study support the notion that

lidocaine can produce a rapidly developing time-dependent block of channels that can reduce thewhole cell IN, measurably within a very few milli-seconds after the onset of INl activation. Since thetime-to-peak for INa in cooled preparations is approx-

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

Clarkson et al Lidocaine Block of Sodium Channels 877

TABLE 3. Effect of Lldocaine on the Rate of IN, Decay at - 20 mV*

Experiment

1014

9104

9102

9103

8251

8252

[Lidocaine](MM)

0100

0100

0200

0200

0200

0200

Afa,(%)

75.4

78.5

76.0

75.3

74.2

73.5

87.0

85.4

77.0

82.7

75.4

82.3

(msec)

1.21.21.51.31.51.11.51.51.31.11.51.6

A,**

(%)

22.5

20.1

22.2

22.9

24.5

25.5

12.0

13.5

21.1

16.4

22.7

16.5

TjJow

(msec)

5.54.85.85.16.54.15.54.66.24.56.57.1

A.(%)

2.11.41.81.81.41.01.01.11.90.91.91.3

•Values represent estimates of the amplitudes and time constants for the components of IN, decay at - 20 mV forcontrol conditions, and after exposure to lidocaine for 5-20 minutes (steady-state effect). The parameters werecalculated by a least-squares fitting method shown in Figure 5. Pulses were elicited after a long >20 second rest atthe holding potential ( -140 mV).

imately 1 to 3 msec at -20 mV, a significantfraction of the block occurring during the first pulsefollowing a long rest period (i.e., the block termed"tonic block") appears to result from a rapidlydeveloped block of open channels. Therefore, espe-cially at strongly depolarized potentials and highdrug concentrations, tonic block may not be anaccurate reflection of drug binding to channels thatare in the closed (i.e., rested) state since in parttonic block can result from open channel block.

The time course for the rapidly developing com-ponent of block (see Figure 3) falls within the periodduring which channel transitions between the closed(rested), open and inactivated states are occurring,and the fraction of channels in the open state is notconstant. Therefore, the onset rate for block devel-opment, characterized by applying conditioningdepolarizing pulses, should not be regarded as anaccurate reflection of the rate of drug-to-open-channel binding.

Although not necessarily anticipated in everyinstance, open channel block also may acceleratethe decay phase of the IN,-time course. Table 3shows that in the presence of lidocaine the slowerinactivation time constant decreased in five out ofsix instances. The decrease, however, was notstatistically significant (p>0.l). Figure 5 also sug-gests that lidocaine may modestly accelerate theinactivation time course even when inactivation isintact. The basis of any acceleration would be quitecomplex, however. INl decay in the presence oflidocaine would be influenced by several factors,among which would be the usual normal (i.e., in theabsence of lidocaine) time course of IN, inactiva-tion, the rate of development of open-channel block,and any effects with a potential for altering IN,inactivation gating kinetics directly. Therefore, it isunlikely that the difference between the time coursesof IN, decay when in the presence and absence oflidocaine will reflect the true drug-to-channel bind-

ing kinetics. Despite the quantitative uncertainties,the finding that lidocaine accelerated the time-dependent decay of IN> in cells whose sodiuminactivation had been modified (by exposure intra-cellularly to a-chymotrypsin), is consistent with thenotion that the change in time-course involved openchannel block, and represents a second line ofevidence supporting this view.

Use-Dependence Involves Both Activation andInactivation Gating

In summary, the present data obtained usingvoltage clamped isolated guinea pig ventricular andcat atrial myocytes supports earlier findings suggest-ing that use-dependent block of cardiac INa pro-duced by lidocaine can involve binding to inacti-vated channels.58 In addition, the data provideevidence suggesting that binding with fairly rapidkinetics of drug to an open channel conformationcan also contribute significantly to the production ofuse-dependent block. The data also suggest thatlidocaine-induced tonic block of IN, may not be areliable measure of drug-affinity for channels in theclosed, rested state because at least some portion ofthe tonic block should result from lidocaine bindingto channels in the open state so quickly that peakINa, even during the first pulse following a long rest,is reduced compared with control. The rapidity ofdrug binding to the sodium channels may play animportant role in lidocaine's ability to depress con-duction under conditions that can substantiallyshorten the duration of cardiac action potentials,such as hypoxia and ischemia.

Acknowledgments

We thank Janet Henderson, Victoria James-Houff, and Marisa Ten Eick for their excellentsecretarial assistance, patience, and good humor.

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

878 Circulation Research Vol 63, No 5, November 1988

References1. Strichartz GR: The inhibition of sodium currents in myeli-

nated nerve by quaternary derivatives of lidocaine. J GenPhysiol 1973;62:37-57

2. Hille B: Local anesthetics: Hydrophilic and hydrophobicpathways for the drug-receptor reaction. J Gen Physiol1977;69:497-515

3. Schwarz W, Palade PT, Hille B: Local anesthetics. Effect ofpH on use-dependent block of sodium channels in frogmuscle. Biophys J 1977,20:343-368

4. Courtney KR, Kendig JJ, Cohen EN: The rates of interac-tion of local anesthetics with sodium channels in nerve. JPharmacol Exp Ther 1978;207:594-604

5. Bean BP, Cohen CJ, Tsien RW: Lidocaine block of cardiacsodium channels. J Gen Physiol 1983;81:613-642

6. Matsubara T, Clarkson C, Hondeghem L: Lidocaine blocksopen and inactivated cardiac sodium channels. NaunynSchmiedebergs Arch Pharmacol 1987;336:224-231

7. Clarkson CW, Hondeghem LM: Mechanism for bupivacainedepression of cardiac conduction: Fast block of sodiumchannels during the action potential with slow recovery fromblock during diastole. Anesthesiology 1985;62:396-405

8. Sanchez-Chapula J, Tsuda Y, Josephson IR: Voltage anduse-dependent effects of lidocaine on sodium current in ratsingle ventricular cells. Circ Res 1983^2:557-565

9. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ:Improved patch-clamp techniques for high-resolution cur-rent recording from cells and cell-free membrane patches.PflugersArch 1981;391:85-100

10. Clarkson CW, Follmer CH, Yeh JZ, Ten Eick RE, Hondeg-hem LM: Evidence for two components of sodium channelblock by lidocaine in single isolated cardiac myocytes(abstract). Circulation 1985;72(suppl III):III-38

11. Silver LH, Hemwall EL, Marino TA, Houser SR: Isolationand morphology of calcium-tolerant feline ventricularmyocytes. Am J Physiol 1983;245:J891-H8%

12. Narahashi T, Tsunoo A, Yoshii M: Characterization of twotypes of Ca++ channels in neuroblastoma cells. J Physiol1987;383:23l-249

13. Follmer CH, Ten Eick RE, Yeh JZ: Sodium current kineticsin atrial myocytes of adult cats. J Physiol 1987^84:169-197

14. Marty A, Neher E: Tight-seal whole-cell recording, in Sak-man B, Neher E (eds): Single-Channel Recording. NewYork, Plenum Press, Publishers, 1983, pp 107-122

15. Belles B, Hescheler J, Trube G: Changes of membranecurrents in cardiac cells induced by long whole-cell record-ings and tolbutamide. Pflugers Arch 1987,409:582-588

16. Isenberg G, Klockner U: Calcium tolerant myocytes pre-pared by preincubation in a "KB medium." Pflugers Arch1982^95:6-18

17. Brown M, Lee KS, Powell T: Voltage clamp and internalperfusion of single rat heart muscle cells. J Physiol 1981;318:455-477

18. Cahalan MD, Aimers W: Interactions between quaternarylidocaine, the channel gates, and tetrodotoxin. Biophys J1979;27:39-56

19. Cahalan MD: Local anesthetic block of sodium channels innormal and pronase-treated squid giant axons. Biophys J1978:23:285-310

20. Yeh JZ: Dynamics of 9-aminoacridine block of sodiumchannels in squid axons. J Gen Physiol 1979;73:1—21

21. Wang GK, Brodwick MS, Eaton DC, Strichartz GR: Inhibi-tion of sodium current by local anesthetics in chloramine-T-treated squid axons. The role of channel activation. J GenPhysiol 1987:89:645-667

22. Alpert LA, Makielski JC, Hanck DA, Fozzard HA: Lidocaineblock of sodium current in single cardiac canine Purkinje cells(abstract). Circulation 1987;76(suppl IV):IV-149

23. Gonoi T, Hille B: Gating of sodium channels. Inactivationmodifiers discriminate between models. J Gen Physiol1987:89:253-274

24. Hondeghem LM, Katzung BG: Time- and voltage-dependent interactions of antiarrhythmic drugs with cardiacsodium channels. Biochim Biophys Ada 1977;472:373-398

25. Sanchez-Chapula J: Interaction of lidocaine and benzocainein depressing V,,,, of ventricular action potentials. J Mol CellCardiol 1985;17:495-503

26. Yeh JZ, Ten Eick RE: Molecular and structural basis ofresting and use-dependent block of sodium current definedusing disopyramide analogues. Biophys J 1987:51:123-135

27. Yeh JZ, Oxford GS: Interactions of monovalent cations withsodium channels in squid axon. II. Modification of pharma-cological inactivation gating. J Gen Physiol 1985:85:603-620

KEY WORDS • lidocaine • sodium channelvoltage clamp • antiarrhythmic drugs

myocytes

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from

C W Clarkson, C H Follmer, R E Ten Eick, L M Hondeghem and J Z Yehmyocytes.

Evidence for two components of sodium channel block by lidocaine in isolated cardiac

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 1988 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.63.5.8691988;63:869-878Circ Res. 

http://circres.ahajournals.org/content/63/5/869World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Research Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on May 10, 2018

http://circres.ahajournals.org/D

ownloaded from