TTXr J Neurophysiol 1998 Scholz 1746 54

7/31/2019 TTXr J Neurophysiol 1998 Scholz 1746 54

1/10

79:1746-1754, 1998.J NeurophysiolAndreas Scholz, Noboru Kuboyama, Gunter Hempelmann and Werner VogelDRG NeuronsLidocaine and Bupivacaine Reduce Firing Frequency in

Currents by+Complex Blockade of TTX-Resistant Na

You might find this additional info useful...

28 articles, 14 of which can be accessed free at:This article cites

http://jn.physiology.org/content/79/4/1746.full.html#ref-list-1

25 other HighWire hosted articles, the first 5 are:This article has been cited by

[PDF][Full Text][Abstract]

, June , 2007; 104 (6): 1434-1439.Anesth AnalgKoji Hara and Takeyoshi Sata

Acid Receptors Expressed in Xenopus Oocytes-AminobutyricThe Effects of the Local Anesthetics Lidocaine and Procaine on Glycine and


, July 1, 2007; 98 (1): 467-477.J NeurophysiolJuan Zhao, Rahima Ziane, Aurlien Chatelier, Michael E. O'Leary and Mohamed Chahinein Mammalian Cells

1.8 Sodium ChannelsvLidocaine Promotes the Trafficking and Functional Expression of Na

[PDF][Full Text][Abstract], February 1, 2008; 99 (2): 617-628.J Neurophysiol

Vitor Pinto, Victor A. Derkach and Boris V. Safronovand Synaptic Transmission

- and C-Fiber ConductionRole of TTX-Sensitive and TTX-Resistant Sodium Channels in A

[PDF][Full Text][Abstract], July , 2008; 326 (1): 270-277.J Pharmacol Exp Ther

Takafumi Horishita and R. Adron HarrisOocytesXenopusChannels Expressed in+-Alcohols Inhibit Voltage-Gated Nan


, September 1, 2009; 587 (17): 4249-4264.J PhysiolDesdemona Fricker, Cline Dinocourt, Emmanuel Eugne, John Wood and Richard Miles

current+Pyramidal cells of rodent presubiculum express a tetrodotoxin-insensitive Na

including high resolution figures, can be found at:Updated information and services

http://jn.physiology.org/content/79/4/1746.full.html

can be found at:Journal of NeurophysiologyaboutAdditional material and information

http://www.the-aps.org/publications/jn

This infomation is current as of March 9, 2012.

Physiological Society. ISSN: 0022-3077, ESSN: 1522-1598. Visit our website at http://www.the-aps.org/.by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright 1998 by the American

publishes original articles on the function of the nervous system. It is published 12 times a year (monthly)Journal of Neurophysiology
http://jn.physiology.org/content/79/4/1746.full.html#ref-list-1http://www.anesthesia-analgesia.org/content/104/6/1434.full.pdfhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.pdfhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.abstract.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.pdfhttp://www.anesthesia-analgesia.org/content/104/6/1434.abstract.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.htmlhttp://jn.physiology.org/content/98/1/467.full.pdfhttp://jn.physiology.org/content/98/1/467.full.pdfhttp://jn.physiology.org/content/98/1/467.full.htmlhttp://jn.physiology.org/content/98/1/467.abstract.htmlhttp://jn.physiology.org/content/98/1/467.full.htmlhttp://jn.physiology.org/content/98/1/467.full.pdfhttp://jn.physiology.org/content/98/1/467.abstract.htmlhttp://jn.physiology.org/content/98/1/467.full.htmlhttp://jn.physiology.org/content/99/2/617.full.pdfhttp://jn.physiology.org/content/99/2/617.full.pdfhttp://jn.physiology.org/content/99/2/617.full.htmlhttp://jn.physiology.org/content/99/2/617.abstract.htmlhttp://jn.physiology.org/content/99/2/617.full.htmlhttp://jn.physiology.org/content/99/2/617.full.pdfhttp://jn.physiology.org/content/99/2/617.abstract.htmlhttp://jn.physiology.org/content/99/2/617.full.htmlhttp://jpet.aspetjournals.org/content/326/1/270.full.pdfhttp://jpet.aspetjournals.org/content/326/1/270.full.pdfhttp://jpet.aspetjournals.org/content/326/1/270.full.htmlhttp://jpet.aspetjournals.org/content/326/1/270.abstract.htmlhttp://jpet.aspetjournals.org/content/326/1/270.full.pdfhttp://jpet.aspetjournals.org/content/326/1/270.abstract.htmlhttp://jpet.aspetjournals.org/content/326/1/270.full.htmlhttp://jpet.aspetjournals.org/content/326/1/270.full.pdfhttp://jp.physoc.org/content/587/17/4249.full.pdfhttp://jp.physoc.org/content/587/17/4249.full.pdfhttp://jp.physoc.org/content/587/17/4249.full.htmlhttp://jp.physoc.org/content/587/17/4249.abstract.htmlhttp://jp.physoc.org/content/587/17/4249.abstract.htmlhttp://jp.physoc.org/content/587/17/4249.full.htmlhttp://jp.physoc.org/content/587/17/4249.full.pdfhttp://jp.physoc.org/content/587/17/4249.abstract.htmlhttp://jn.physiology.org/content/79/4/1746.full.htmlhttp://jn.physiology.org/content/79/4/1746.full.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.pdfhttp://www.anesthesia-analgesia.org/content/104/6/1434.full.htmlhttp://www.anesthesia-analgesia.org/content/104/6/1434.abstract.htmlhttp://jn.physiology.org/content/98/1/467.full.pdfhttp://jn.physiology.org/content/98/1/467.full.htmlhttp://jn.physiology.org/content/98/1/467.abstract.htmlhttp://jn.physiology.org/content/99/2/617.full.pdfhttp://jn.physiology.org/content/99/2/617.full.htmlhttp://jn.physiology.org/content/99/2/617.abstract.htmlhttp://jpet.aspetjournals.org/content/326/1/270.full.pdfhttp://jpet.aspetjournals.org/content/326/1/270.full.htmlhttp://jpet.aspetjournals.org/content/326/1/270.abstract.htmlhttp://jp.physoc.org/content/587/17/4249.full.pdfhttp://jp.physoc.org/content/587/17/4249.full.htmlhttp://jp.physoc.org/content/587/17/4249.abstract.htmlhttp://jn.physiology.org/content/79/4/1746.full.html#ref-list-1


2/10

Complex Blockade of TTX-Resistant Na/ Currents by Lidocaine and

Bupivacaine Reduce Firing Frequency in DRG Neurons

ANDREAS SCHOLZ, 1 NOBORU KUBOYAMA,2 GUNTER HEMPELMANN,3 AND WERNER VOGEL11Physiologisches Institut, Justus-Liebig-Universitat Giessen, D-35392 Giessen, Germany; 2Department of Pharmacolog

Nihon University, Chiba 271, Japan; and 3Abteilung fur Anaesthesiologie und Operative Intensivmedizin,

D-35398 Giessen, Germany

Scholz, Andreas, Noboru Kuboyama, Gunter Hempelmann, of neurons may be due to their equipment with differeand Werner Vogel. Complex blockade of TTX-resistant Na/ cur- types of ion channels that are differently sensitive to locrents by lidocaine and bupivacaine reduce firing frequency in DRG anesthetics. Of the two groups of Na/ conductances, senneurons. J. Neurophysiol. 79: 17461754, 1998. Mechanisms of tive and resistant to TTX, the latter has been observed blockade of tetrodotoxin-resistant (TTXr) Na/ channels by local

dorsal root ganglion cells (Elliott and Elliott 1993; Kostyuanesthetics in comparison with the sensitivity of tetrodotoxin-sensi-

et al. 1981; Ogata and Tatebayashi 1993; Roy and Narahastive (TTXs) Na/ channels were studied by means of the patch-1992), in C fibers contributing to compound action poteclamp technique in neurons of dorsal root ganglions (DRG) of rat.tials (Jeftinija 1994) and in human peripheral nerve (Qua

Half-maximum inhibitory concentration (IC50 ) for the tonic block hoff et al. 1995). In myelinated fibers of peripheral nervof TTXr Na/ currents by lidocaine was 210 mmol/ l, whereas TTXsNa/ currents showed five times lower IC50 of 42 mmol/l. Bupiva- only TTXs Na

/ channels could be studied directly with tcaine blocked TTXr and TTXs Na/ currents more potently with patch-clamp technique (Scholz et al. 1993), but it is nIC50 of 32 and 13 mmol/l, respectively. In the inactivated state, possible to use this method for investigation of thin unmTTXr Na/ channel block by lidocaine showed higher sensitivities elinated fibers, which also contain TTX-resistant ( TTX(IC50 60 mmol/l) than in the resting state underlying tonic block- Na/ channels (Quasthoff et al. 1995). Both Na / channade. The time constant t1 of recovery of TTXr Na

/ channels fromtypes, however, can be studied at the somata of those senso

inactivation at080 mV was slowed from 2 to 5 ms after addition offibers. About the blockade of TTXr Na/ currents by loc10 mmol/l bupivacaine, whereas the t2 value of500 ms remainedanesthetics, however, little is known so far (Akopian et aunchanged. The use-dependent block of TTXr Na/ channels led1996; Roy and Narahashi 1992). The local anesthetics lidto a progressive reduction of current amplitudes with increasingcaine and bupivacaine, both frequently used for spinal anefrequency of stimulation, which was 53% block at 20 Hz in 10

mmol/l bupivacaine and 81% in 100 mmol lidocaine. The functional thesia, have been investigated in their blocking effects importance of the use-dependent block was confirmed in current- TTXr Na/ currents. Two preparations were used in oclamp experiments where 30 mmol/l of lidocaine or bupivacaine study. Pharmacological and voltage-clamp experiments wedid not suppress the single action potential but clearly reduced the performed with enzymatically isolated dorsal root gangliofiring frequency of action potentials again with stronger potency (DRG) neurons with the advantage of easy solution eof bupivacaine. Because it was found that TTXr Na / channels

change and proper space-clamp condition. Current-clampredominantly occur in smaller sensory neurons, their blockade

experiments were performed with Ad- and C-type neuromight underlie the suppression of the sensation of pain. Differentvisually identified in thin DRG slices ( Safronov et al. 1996sensitivities and varying proportions of TTXr and TTXs Na/ chan-avoiding enzymatic treatment and with the advantage nels could explain the known differential block in spinal anesthesia.good physiological condition.We suggest that the frequency reduction at low local anesthetic

concentrations may explain the phenomenon of paresthesia where Some of these results have been presented to the Germsensory information are suppressed gradually during spinal anes- Physiological Society (Scholz et al. 1996).thesia.

M ET H O D S

I NT R O D U C TI O N Materials

Young Wistar rats (819 days) of both sexes were killed, tThe mechanism underlying spinal and regional anesthesiavertebral column was removed quickly and opened in a coolby local anesthetics generally is explained by a blockadedish (24C) containing the preparation solution. DRGs (102of tetrodotoxin-sensitive (TTXs) Na/ channels. Less clear,from lower thoracic (Th8) and lumbar levels were dissectehowever, is the mechanism of differential anesthesia, a clini-Ganglia were transferred into a petri dish containing preparatically observed phenomenon of different onsets of suppres-solution with 0.71.5 mg/ml trypsin (type I, Sigma) and 46 m

sions of various sensory and motor functions ( Raymond andml collagenase (type II Worthington, Biochrome) and incubat

Gissen 1987). Explanations have been suggested as varying for 2535 min at 37C in a shaking closed chamber (110 min0access of the drug molecules to their receptors at the ion with 5% CO2-95% O2 . After washing four to five times at roochannels that may be subject to varying caliber and myelina- temperature in preparation solution, the softened ganglia were d

sociated by gentle titration with a small pipette in a petri dish ttion of the nerve fibers. Furthermore, differential blockade

1746 0022-3077/98 $5.00 Copyright 1998 The American Physiological Society


3/10

LOCAL ANESTHETIC EFFECTS ON NA/ CHANNELS IN DRG NEURONS 17

bottom of which was covered by silicone elastomer (Sylgard Blue, 150, Clark Electromedical Instruments) using a David Kopf clotwo-stage vertical puller (Physiologisches Institut) and fire-pDow Corning). The cells were stored for 15 h in preparation

solution bubbled with a 95% O2-5% CO2 mixture and humidified ished on a microforge ( workshop Physiologisches Institut) tofinal resistance of 0.62.5 MV when filled with internal solutioatmosphere at 24C and finally were transferred by a small pipette

into the experimental petri dish. Whole cell recordings were performed as described by Hamillal. (1981). Additionally series resistance was compensated (5In current-clamp experiments, thin 150-mm slices of DRGs

(without enzymatic treatment) from young Wistar rats (6 9 days ) 80%) . Experiments were rejected when the voltage error exceed5 mV after compensation of the series resistance. Voltage-depehave been used as described in more detail by Safronov et al.

(1996). dent currents were corrected off-line for unspecific leak curre

and transients using averaged recordings evoked by hyperpolaring impulses. All values presented are means { SE.Solutions

Preparation solution: mixture of 50% F12 and 50% DulbeccosR ES U L T Smodified Eagles medium (D 8900, Sigma) with 1.8 g/l sodium

bicarbonate, bubbled with a 95% O2-5% CO2 mixture, 100 U peni- Na/ currents were separated on the basis of their differecillin, and 0.1 mg/ml streptomycin (P 0781 Sigma), adjusted to

sensitivities to TTX (Elliott and Elliott 1993; Kostyuk et apH 7.4 with NaOH.

1981). TTXr Na/ currents were recorded in the presenThe Ringer bicarbonate solution contained (slice preparation, inof 300 nmol/l TTX. Na/ currents were considered to mmol/l) 115 NaCl, 5.6 KCl, 2 CaCl2 , 1 MgCl2 , 11 glucose, 1TTX sensitive if they demonstrated fast kinetics and NaH2PO4 , and 25 NaHCO3 (pH 7.4 by bubbling with a 95% O 2-95% of them could be blocked by 100 or 200 nmol/l TT5% CO2 mixture), final Na 141.

The Ringer-tetraethylammonium (TEA; /TTX) solution con- at the end of the experiment. K/ currents were suppresstained (in mmol/l) 134 NaCl, 4.5 KCl, 20 TEA, 0.1 CaCl2 , 5 by internal Cs/ (Ogata and Tatebayashi 1993) and externMgCl2 , and 10 N-[2-hydroxyethyl]-piperazine-N[2-ethanesul- TEA (Safronov et al. 1996). Currents through six differe

fonic acid] (HEPES), adjusted with NaOH to pH 7.4, final Na types of Ca 2/ channels known so far (Mintz et al. 199144, 0.0003 TTX (Latoxan). Scholz et al. 1997) were abolished by Mg 2/ ions added

The Ringer 42 Na solution (/TTX) was as the Ringer-TEAinternal and external solutions (Hess et al. 1986) and

solution except that NaCl was 32 mmol/ l and cholineCl (102some experiments by F0 ions in internal solutions (Kostymmol/ l) was added; final Na 42.et al. 1977). A low Ca 2/ concentration in the external solThe high-K internal solution contained (in mmol/l) 144.4 KCl,tions of 100 mmol/l prevented the Ca 2/ channels from b1 MgCl2 , 5 NaCl, 10 HEPES, and 3 ethylene glycol-bis(b-amino-coming Na/ conducting (Carbone and Lux 1988; Hess ethyl ether)N,N,N,N,-tetraacetic acid (EGTA), adjusted with

KOH to pH 7.3, final K 155. al. 1986) .The CsCl internal solution contained (in mmol/l) 105 CsCl, 10 In general, with increasing cell diameter, we observed

NaCl, 2 MgCl2 , 10 EGTA, 4 Na2-adenosine 5-triphosphate decreasing proportion of TTXr Na/ channels, although(ATP), and 10 HEPES, adjusted with CsOH to pH 7.2, final Cs clear correlation between current proportion and neuron d135. ameter could not be established. In neurons 40 mm, TTX

The CsF internal solution contained (in mmol/l) 40 CsF, 67.5Na/ currents prevailed, and in some cases current amplitu

CsCl, 6 NaCl, 2 MgCl2 , 10 EGTA, 4 Na2-ATP, and 10 HEPES, exceeded 10 nA reaching the limits of the amplifier. Ther

adjusted with CsOH to pH 7.2, final Cs 135, final Na 14. fore Na/ ions in the external solution were substituted paLidocaine was dissolved and stored as stock solution (1 mol/l)tially by choline/ ions, and Na/ concentration was reducin aqua bidest and bupivacaine (0.1 mol/l) in dimethyl sulfoxide.to 42 mmol/l in some of our experiments. Cells 25 mAll reagents were purchased from Sigma.mainly displayed TTXr Na/ currents. In 85% of mediusized cells, there was a mixture of TTXs and TTXr NaApplication systemcurrents in varying proportions with 80% from one typ

Control solution and local anesthetics containing solutions were but the remaining 15% of these cells showed only one tyapplied by a modified six-barrel perfusion system. Continuous and of Na/ current.equal flows out of the six glass barrels were produced by a syringepump delivering pressure through a tubing to the six syringes. A

Blockade of peak Na/ currentscomplete exchange of solutions was accomplished in a few secondsby moving the tip of the pipette with the neuron from one barrel

Medium- and large-sized cells, 14 of 134 cells in thinto another one.study, displayed TTXs Na/ currents only that showed typcal fast activation with a time to peak of 0.65 { 0.25 m

Electrophysiology (mean { SE, n 14) and fast inactivation with a timVoltage-clamp experiments were carried out on isolated cells constant of 0.84 { 0.32 ms at 0 mV (Fig. 1, Cand D) . The

using an EPC 7 (List) patch-clamp amplifier. Current-clamp exper- currents showed a reversal potential close to the calculatiments were done using a thin slice preparation with an Axopatch equilibrium potential of Na/ ions and were abolished com200A ( Axon Instruments) patch-clamp amplifier. Data were digi-

pletely at the end of the experiment by 100 or 200 nmotized with a 12-bit AD/DA converter TL/1 (Scientific Solutions)

TTX and were thus identified as TTXs Na / currents. and filtered at least at two times lower frequency with an eight-complete exchange of the external solution was accompole, low-pass Bessel filter (Physiologisches Institut). Voltage andplished by moving the cell under investigation from ocurrent steps and data acquisition were controlled on-line by abarrel into the next (see METHODS). The local anesthetiPC/AT computer with pCLAMP software (Axon Instruments).

Electrodes were fabricated from borosilicate glass tubings ( GC lidocaine and bupivacaine reversibly blocked TTXs Na/ cu


4/10

A. SCHOLZ, N. KUBOYAMA, G. HEMPELMANN, AND W. VOGEL1748

FIG . 1. Tetrodotoxin-sensitive (TTXs) and -resistant (TTXr ) Na/ channels blocked by lidocaine and bupivacaine, respec-tively. Whole cell currents recorded at 0 mV from 4 different dorsal root ganglion (DRG) neurons of rat. A : fits to thedecaying parts of TTXr Na/ currents (rrr) showed a t of 3.6 ms in Ringer-TTX solution and 3.2 ms in the presence of300 mmol/l lidocaine. Time to peak was 1.7 and 1.5 ms, respectively. Pipette resistance: 0.9 MV. B : TTXr Na/ currentdecayed with t of 1.3 ms in Ringer-TTX and 1.2 ms in the presence of 50 mmol/l bupivacaine. Time to peak was 0.7 and0.6 ms, respectively. Pipette resistance: 0.6 MV. C: TTXs Na/ currents decayed faster with t of 0.4 ms in Ringer solutionand 0.5 ms in the presence of 50 mmol/l lidocaine. Time to peak was 0.3 and 0.4 ms, respectively. Currents were blockedcompletely by 200 nmol/l TTX. Pipette resistance: 0.7 MV, pipette solution: CsF, bath solution: Ringer with reduced 42mmol/l Na/ . D : another neuron displayed TTXs Na/ currents decaying with t of 0.6 ms and 0.7 ms in control and in thepresence of 10 mmol/l bubivacaine. Time to peak was 0.4 and 0.5 ms, respectively. Less than 2% of the current remainedin the presence of 200 nmol/l TTX. Pipette resistance: 1.4 MV. Pulse protocol is shown above the original recordings,durations of prepulses to 0110 mV were 50 ms for TTXs Na/ channels and 100 ms for TTXr Na/ channels. , zerolevel. Holding potential was 080 mV. Current traces shown are corrected for leakage and capacity currents as described inMETHODS. Except mentioned, pipette solution, CsCl; bath solution: Ringer or Ringer-TTX; temperature: 22 24C.

rents without modifying kinetics of activation and inactiva- caine demonstrated a higher affinity to the TTXs Na/ chanel with an IC50 of 12.5 { 1.4 mmol/l (Fig. 2B ) than to ttion significantly as shown in Fig. 1, C and D .

In contrast, TTXr Na/ channels activated at more posi- TTXr Na/ channel with an IC50 of 31.7 { 5.2 mmol/l. Ththe relation of potencies of bupivacaine and lidocaine wetive potentials (E030 mV) with two times slower kinet-

ics (time to peak 1.62 { 0.64 ms ) and inactivated with a higher for TTXr Na/ channels with 6.6 than for TTXs Nchannels with 3.4. For action of both local anesthetics otime constant of 3.5 { 1.8 ms at 0 mV, showing wide

variations ( n 29) . These currents were not impaired by both Na/ channel types, the Hill coefficients were near onindicating an 1:1 stoichiometry of interaction.300 nmol/l TTX. As shown by the recordings in Fig. 1,

A and B, lidocaine and bupivacaine reduced the amplitudeof the TTXr Na/ current. Here, and in another 23 experi- Blockade of TTXr Na/ channels in inactivated and restinments, inactivation kinetics were not changed significantly, state by lidocaineand therefore a slow open channel block could be ex-

cluded. Kinetics of wash-in and -out were faster than the It has been shown that TTXs Na/ channels in the inacvated state are more sensitive to lidocaine than in the restimanoeuvre of switching from one barrel to another (see

METHODS) . state ( Hille 1977; Ragsdale et al. 1994 ) . Because this chanel characteristic has some impact on its physiological funHalf-maximal inhibitory concentrations (IC50 ) for tonic

blockade were derived from fits to data plotted as relative tion concerning conducting capacity, we have tested whethdifferent states of TTXr Na/ channels are also differentblock of peak Na/-current (INa,peak) versus concentration of

local anesthetics (Fig. 2) . With an IC50 of 42.3 { 5.4 sensitive to this local anesthetic.Data for drug binding in the resting state were taken frommol/ l, the TTXs Na/ current was five times more sensitive

to lidocaine than the TTXr Na/ current with an IC50 of experiments in Fig. 2 that had been obtained without inacvating prepulse. The protocol for investigation of chann210 { 31 mmol/l (Fig. 2A ). The more lipophilic bupiva-


5/10


FIG . 3. TTXr Na/ channels are more sensitive to lidocaine in the inacvated than in the resting state. Sensitivity of TTXr Na / channels in inactivated state was investigated using a 2.5 s prepulse to 035 mV followby a short gap of 5 ms to 0100 mV and a subsequent test impulse to /mV. Concentration-block data and curve fit were obtained as in Fig. 2. ICvalues were 59.5 { 11.7 mmol/l lidocaine for inactivated TTXr Na/ chnels (, 4 cells). Corresponding data for the blockade in the resting stare taken from Fig. 2A and are given as . Bath solution: RingTTX; pipette solution: CsF.

mmol/l, 3.6 times lower than the blockade of channels the resting state with 210 mmol/l.

Modulation of recovery from inactivation

It was shown in preceding sections that the time to peof Na/ current and the time constant of inactivation rflecting the transitions from closed to open and from op

FIG . 2. Concentration-block-curves for lidocaine and bupivacaine. A :to inactivated states remained unaffected in the presence

relative block was calculated from peak Na/ currents in the presence oflocal anesthetics. As the next step in investigating the effeclidocaine related to those under control conditions at depolarizations to /10

mV. Data points were fitted by Hill equation with IC50 values of 42.3 { of local anesthetics on gating of TTXr Na/ channels, w

5.4 mmol/l for TTXs (, 6 cells) and 210 { 31 mmol/l for TTXr Na/ studied the modulation of transition of the channel from tchannels (q, 8 cells) . Hill coefficients were 1.06 and 1.02, respectively. B : inactivated to the resting state, i.e., on its recovery froblockade by bupivacaine was obtained as described in A. IC50 values for inactivation. In these experiments, DRG neurons were stimTTXs (4 cells) and for TTXr Na/ currents (5 cells) were 12.5 { 1.4 and

lated with a double-pulse protocol at 080 mV with interva31.7 { 5.2 mmol/l, respectively. The Hill coefficients were 1.05 and 0.96,respectively. Frequency of stimulation was 0.33 Hz for TTXs and 0.25 Hz varying between 1 ms and 2 s ( Fig. 4) . The recovery coufor TTXr Na/ currents. Error bars give SE. be fitted adequately with two exponential functions, a fa

one with t1 2.1 ms comprising 75% of the amplitude, ablock in the inactivated state shown in Fig. 3 consisted of a slow one with t2 514 ms (25%). In the presence of a 2.5-s prepulse to 035 mV (Epre ), a 5-ms hyperpolarizing mmol/l bupivacaine, the fast component t1 was slowed step to 0100 mV (Ehyp ), and a 30-ms test pulse to /10 mV 5.4 ms, whereas the slow component was nearly as w(Etest ) . Epre inactivated most of the channels but caused little unaffected as the relative amount of the components wactivation (Elliott and Elliott 1993) and allowed channels nearly unchanged. Thus a major portion of current recoverto reach an equilibrium for binding of local anesthetics to slower from inactivation already at a concentration of bupthe inactivated state. Ehyp allowed recovery from inactiva- vacaine as low as 10 mmol/l (IC50 32 mmol/l); this wtion, but it was chosen to be short to prevent drug dissocia- seen in five experiments. This could limit conduction abition from most of the blocked channels. The following test ties of a neuron at high firing frequencies as the followinpulse activated unblocked channels having recovered from experiments will show.inactivation. The protocol was applied every 10 s. In controlsolution, INa,peak evoked by the pulse protocol for the inacti- Frequency-dependent blockade by bupivacaine andvated state was 5060% compared with the current evoked lidocaineby the pulse protocol for the resting state. This reductionwas due to slow inactivation during the long impulse from To test whether the TTXr Na/ channel showed a us

dependent (phasic) block as did the TTXs Na/ channwhich channels did not recover during the short pulse to0100 mV (Chandler and Meves 1970). (Hille 1977; Ragsdale et al. 1994; Roy and Narahashi 1992

repetitive impulses to /10 mV were applied at differeThus INa,peak of TTXr channels obtained in control solutionand in different concentrations of lidocaine were plotted as frequencies. The amplitude of currents evoked by the n

impulse was normalized to that of the current evoked by trelative block in Fig. 3 revealing an IC50 of 59.5 { 11.7


6/10


by 10 mmol/l bupivacaine depending on frequency is showin Fig. 5C and revealed the increasing phasic block fro8.6% at 0.2 Hz to a considerable amount of 52.6% at 2Hz. In presence of 100 mmol/l lidocaine, the phasic bloranged from 5.6% at 0.4 Hz to 81.2% at 20 Hz (Fig. 6 CIt adds to the tonic blockade of the channels (shown earlieFigs. 2 and 3).

Modulation of firing behavior in a slice preparation

The effect of local anesthetics on action potentials (Awas tested on neurons in a thin slice preparation. This prepration provided access to small cells identified as Ad-and type neurons (Safronov et al. 1996) based on soma diametand duration of AP ( Harper and Lawson 1985b) . The resti

FIG . 4. Modulation of time of recovery from inactivation by bupiva-potentials were below 060 mV and the input resistanccaine. Time constants for recovery from inactivation of TTXr Na/ channels

were measured with a double pulse protocol. A first pulse (P1) for 50 ms were as high as 0.73 GV. These neurons mostly containto /10 mV caused complete inactivation, and INa,P2 evoked by the test pulse TTXr Na/ channels. Figure 7A demonstrates the blocki(P2) to /10 mV after variable intervals was compared with INa,P1 of the effects of 30 and 300 mmol/l bupivacaine on trains of actiosame episode. Resulting data points could be fitted with a double exponen-

potentials elicited by injection of a sustained depolarizintial function giving a major fast component with t1 of 2.1 ms (75%) anda slow component with t2 of 514 ms (25%) under control condition () . current. In control the neuron responded with a train of acti

In the presence of 10 mmol/ l bupivacaine (q), the kinetics oft1 was slowed potentials showing some adaptation within the first 500 mto 5.4 ms (71%), whereas t2 was unaffected with 568 ms ( 29%). Inset: Firing frequency at the beginning of the train calculated frodata graphed of the fast component. Pulse protocol was repeated with 0.1

the time interval between the first two action potentials wHz at a holding potential of080 mV. Bath solution: Ringer TTX; pipette35 Hz. The peak potential of the first AP was /49 mV asolution: CsCl; temperature: 22C.decayed slightly to /44 mV at the end of the train.

After changing to an external solution containing addfirst impulse. In contrast to TTXs Na/ channels ( Hille 1977;tional 30 mmol/ l bupivacaine, the same stimulus elicitSchmidtmayer and Ulbricht 1980), TTXr Na/ channels al-only two action potentials showing a decrease of the peready in control solution showed a reduction in amplitudepotentials from /41 to /31 mV and a reduced calculatthat depended on frequency of stimulation (Figs. 5A andfiring frequency of 22 Hz, which can be seen in detail 6A, ). This reduction could be explained by the slowFig. 7C. In the presence of 300 mmol/l bupivacaine, trecovery of TTXr Na/ channels from inactivation (t2 514neuron did not respond with action potentials any more anms, see Fig. 4). The effect became stronger with increasingwas unexcitable also with stronger stimuli. After wash out frequency of stimulation from 0.4 to 5 and 20 Hz. In thethe local anesthetic the blocking effects were fairly reversibpresence of 10 mmol/l bupivacaine, an additional reduction

showing a train of six action potentials, a peak potential of the current due to the development of the use-dependent/51 mV and a calculated firing frequency of 28 Hz at thblock could be observed with faster kinetics. The use-depen-beginning.dent block was measured as a ratio between the normalized

The blocking effects of 30 and 300 mmol/l lidocaine amplitudes of currents evoked by the 20th impulses in thetrains of action potentials of another neuron are shown presence and absence of bupivacaine. From these data, theFig. 7B . In control the calculated firing frequency at tavailability of TTXr Na/ currents as a function of stimula-beginning of the train was 35 Hz. The peak potential of ttion frequency is shown in Fig. 5B . It was calculated byfirst AP was /46 mV and decayed to /38 mV at the enormalizing the current of the 20th pulse (INa,P20 ) to that ofof the train. In presence of 30 mmol/l lidocaine, the pethe first pulse (INa,P1 ). Connecting lines were drawn by eyepotential was unchanged and the calculated firing frequenrevealing half-maximal availabilities of TTXr Na/ currentat the beginning was reduced to 21 Hz. The shape of tat frequencies of 4.9 and 2.1 Hz in control and in the pres-first AP was nearly unchanged compared with control, whience of 10 mmol/l bupivacaine, respectively. The effect ofcan be seen on the superimposed recordings in Fig. 7D . T100 mmol/l lidocaine on the relative reduction of peak cur-number of APs during the train was reduced from 12 to rent amplitude is shown in Fig. 6A. Again, kinetics of reduc-In contrast to the result with 300 mmol/l bupivacaine, in ttion of the relative amplitude was faster in presence of thepresence of 300 mmol/ l lidocaine, a single AP could local anesthetic than in control. The more pronounced use-elicited with a peak potential of/29 mV. At a concentratidependent effect of lidocaine also could be seen in Fig. 6Bof 1 mmol/l lidocaine, it was impossible to elicit an Awhere half-maximal availability of TTXr Na/ currents is(data not shown). After wash out of lidocaine a train shifted from 7.7 Hz in control to 2.2 Hz in the presence ofnine APs reappeared with a peak potential of/49 mV a100 mmol/ l lidocaine.a calculated firing frequency of 38 Hz at the beginning. Thus the frequency dependent decay of the TTXr Na/

four other neurons displaying trains of APs, there wascurrent functions like a filter. A fairly low concentration ofsimilar reduction observed in calculated firing frequency aa local anesthetic shifts this decay to a lower frequency

value. The relative phasic blockade of TTXr Na/ currents number of APs in presence of lidocaine.


7/10


D IS C U S SI ON expressing80% and a smaller population of small diametneurons expressing mainly TTXr Na/ currents were fou

It is shown that native TTXr and TTXs Na/ channels in in older animals (1519 days). Roy and Narahashi (199rat DRG neurons are blocked by lidocaine and more potently observed only 10% neurons expressing TTXr Na/ curreby bupivacaine, both frequently used for spinal anesthesia. in rats 12 days of age.The distribution of TTXr and TTXs Na/ currents was found The sensitivities of tonic blockade of the TTXs comparto correlate with the diameter of the neurons although there with the TTXr Na/ channels were higher by factors ofwas no close correlation especially for medium-sized neu- for lidocaine and 2.5 for bupivacaine (Fig. 2). Our IC

rons. The expression of TTXr Na/

currents in 60% of all value of 42 mmol/l for lidocaine is somewhat lower thinvestigated neurons, a great number of medium-sized cells reported for TTXs Na/ currents of amphibian node of Ra

vier ( Hille 1977; Schmidtmayer and Ulbricht 1980) bcomparable to the result of Roy and Narahashi (1992). Fa cloned TTXr Na/ channel, the sensitivity to lidocaine wsix times lower (Akopian et al. 1996) than the one founin this study. The difference might be explained by differeb subunits present at the Na/ channel protein or by follicutissues in Xenopus oocytes, which reduce drug effects ion channels (Madeja et al. 1997). In general, the IC50 valufound here for both types of Na/ channels were below tconcentrations used in whole nerve preparations to block tconduction (Gissen et al. 1980; Wildsmith et al. 1989). Thcan be due to diffusion barriers existing in nerve and

surrounding tissues (Raymond and Gissen 1987). Anothreason could be the safety factor still allowing to elicit actiopotentials with a reduced number of Na/ channels. Concetrations of bupivacaine measured in cerebrospinal fluid duing spinal anesthesia in man were between 50 and 300 mmol (Dennhardt and Konder 1983). Under our conditions theconcentrations blocked 6095% of the TTXr Na / curren(Fig. 2) and were effective enough to suppress generatiof TTXr action potentials (Fig. 7).

During local anesthesia of peripheral nerves, it is generexperience that fast conducted sensory fibers are blockedlower concentrations, and other slow conducting sensofibers, especially pain transmission, are suppressed at highconcentrations (Gissen et al. 1980; Wildsmith et al. 1989

An obvious explanation could be the presence of TTXr Nachannels underlying TTXr action potentials in small fibeof spinal and peripheral nerve (Jeftinija 1994; Quasthoffal. 1995). These channels activating more slowly and ahigher threshold may be another reason for slower condution velocities in these fibers besides small diameter. ThTTXr Na/ channels exist in both somata of small DRneurons and axons of small peripheral nerve fibers. Intere

FIG . 5. Use-dependent block of TTXr Na/ channels by bupivacaine.to demonstrate a use-dependent block of TTXr Na/ channels, a pulse procol with repetitive impulses to /10 mV (10 ms long) was used at differstimulating frequencies. Amplitudes of the currents ( after correction leakage and capacity currents) were normalized to the current amplituof the first impulse under control condition () and in the presence of mmol/l bupivacaine (q). In this plot, the tonic component of the block the local anesthetic has been eliminated by normalization and only the udependent component is demonstrated. B : relative amplitudes of TTXr Ncurrent under steady-state condition at 20th impulse of experiment in Awere plotted as depending on frequency of stimulation. Data points wefitted by eye. In the presence of 10 mmol/ l bupivacaine, the curve wshifted from 4.9 Hz under control conditions to 2.1 Hz measured at hamaximal inhibition. C: relative phasic blockade of TTXr Na/ currents 10 mmol/l bupivacaine depending on frequency. Additional relative phablock was measured at the 20th impulse as relative reduction of the amptudes of TTXr Na/ currents in A ( ). Bath solution: Ringer TTX; pipesolution: CsF; temperature: 24C.


8/10


ingly, they both are parts of Ad- and C-type DRG neuro(Harper and Lawson 1985a,b) . The lower sensitivity TTXr Na/ channels to lidocaine and bupivacaine comparwith TTXs Na/ channels gives an additional explanation why at higher concentrations of both local anesthetics onC fibers of the nerve were conducting ( Gissen et al. 1980These authors showed that the difference in sensitivities wtwo times more pronounced with lidocaine than with bupiv

caine; this is in good agreement with our data.The higher sensitivity of TTXr Na/ channels to lidocai

in the inactivated than in the resting state ( Fig. 3 ) is probabdue to the same mechanism as already published for TTXNa/ channels. A molecular model by Hille (1977) suggesttwo different access routes to the binding sites, and a moleclar counterpart was shown by Ragsdale et al. (1994). cloned TTXs Na/ channels, they demonstrated two differebinding regions for local anesthetics, one of them causinguse-dependent blockade. The higher affinity of lidocaine the inactivated than to the resting state of the TTXr Nachannel may in reality be much higher than the one observin our studyfactor 4 (Fig. 3) because measurement the real affinity of a local anesthetic to the inactivated cha

nel required a compromise: theoretically, the portion of inativated Na/ channels not blocked by the local anestheshould be measured with a test pulse at the end of a loinactivating prepulse. However, a short period for recovefrom inactivation at a negative potential has to precede ttest pulse to obtain a measurable amount of Na/ curreWith increasing duration of the period preceding the tepulse, an increasing fraction of previously blocked channwill be unblocked, leading to an underestimation of the sentivity of the channel. Practically, we applied a prepulse 5 ms assuming that the time constant of dissociation of tdrug from its receptor would be longer as described for TTXNa/ channels (Schmidtmayer and Ulbricht 1980; Ulbric1981). For cloned TTXs Na/ channels blocked by etid

caine, this factor of sensitivities of inactivated to restichannels was 100 (Ragsdale et al. 1994).

Our observation of a slowed fast time constant of recove(Fig. 4) differs from the reported lidocaine-induced slowrecovering fraction (t, 12 s) and additionally from tunchanged time constant of the normally recovering fractiof Na/ channels (Bean at al. 1983). Therefore our findinmight be explained by local anesthetic molecules bindinginner parts of the channels. This local anesthetic molecumay impede the recovery from inactivation because the inativation gate cannot be released until the blocking molecuhas dissociated from the receptor. The time needed for thprocess is reflected by an increase in the fast time constaby a factor of three. Almost the complete fraction of the faFIG . 6. Use-dependent block of TTXr Na/ channels by lidocaine.

Same type of experiment as in Fig. 5 at another neuron except Ringer- time constant is affected already at a low concentration TTX solution containing 100 mmol/l lidocaine. A : relative peak current bupivacaine. This lends further support to the argument amplitude at stimulating frequencies of 0.4, 5, and 20 Hz in control higher sensitivity of the inactivated state of the channel condition () and in the presence of 100 mmol/ l lidocaine (q ) . B : relative

local anesthetics.peak current amplitudes of TTXr Na/ current from the 20th impulseA low concentration of local anesthetics blocked trainsmarked in A ( ) were plotted as depending on frequency of stimulation.

In the presence of 100 mmol/l lidocaine, the curve was shifted from 7.7 action potentials at higher frequency in C fibers better thHz under control conditions to 2.2 Hz measured at half-maximal inhibi- at lower frequencies (Wildsmith et al. 1989). This observtion. C: relative phasic blockade of TTXr Na/ currents by 100 mmol/l

tion can be explained by a use-dependent blockade as selidocaine at the 20th impulse ( in A) depending on frequency. Bath

in our experiments with TTXr Na/ currents (Figs. 5 asolution: Ringer TTX; pipette solution: CsCl; pipette resistance: 1 MV;temperature: 23C. 6) . TTXr Na/ currents even without local anesthetics ( Fig


9/10


FIG . 7. A : blockade by bupivacaine of action potentials of an Ad- or C-type neuron in a slice preparation. Restingpotentials was between 061 and 064 mV and was not influenced by application of the local anesthetic. Cell diameter: 23

mm. B : blocking effects by lidocaine of action potentials of another neuron in a slice preparation. Cell diameter: 21 mm.Protocol of current injected under current clamp is given on the top row for both neurons. C and D : part of traces indicatedby boxes and symbols in A and B of action potential recordings in control and in presence of 30 mmol/l local anestheticsare shown superimposed on an expanded time scale. Symbols are corresponding to the registrations in A and B . Bath solution:Ringer bicarbonate with TTX; pipette solution: high-K; temperature: 23 C.

5A and 6A ) at frequencies of 0.4 and more clearly at 5 and only a reduced sensory input but no complete insensitivis reported. This behavior becomes also evident in our obse20 Hz showed a reduction of peak amplitudes. It may be

due to the slow recovery from inactivation (t2 514 ms; vations with trains of action potentials elicited under curreclamp conditions (Fig. 7) showing this decrease of frFig. 4) . Roy and Narahashi ( 1992) reported a lower reduc-

tion of amplitude at all frequencies, the remaining relative quency and reduction of number and amplitude of actipotentials even at a low concentration of bupivacaine. Singamplitude of TTXr Na/ currents in the presence of lidocaine

was up to three times larger. A possible explanation of the action potentials are blocked on the basis of a resting blocwhereas during trains of APs, some additional blockade mobserved differences might be a heterogeneity in kinetics of

TTXr Na/ channels ( Elliott and Elliott 1993; Ogata and develop due to use-dependent blocking mechanisms of TTNa/ channels.Tatebayashi 1993).

The faster kinetics of reduction of the relative peak ampli- This effects also may explain an increase of the threshofor sensory input. The clinical observation of an incompletudes in the presence of local anesthetic (Figs. 5A and 6A )

could be explained by adding up the two mechanisms in anesthesia named Wedensky block may be explained bsuch a reduction of frequency leading to an increase of tcontrol and the use-dependent effect of the local anesthetic

on the channel. threshold for sensory input.We suggest that the blockade of TTXr Na channels bThe shift in E50 for the frequency-dependent reduction of

TTXr Na/ currents in the presence of local anesthetics ( Figs. local anesthetics has to be considered more seriously asmechanism of pain suppression in addition to blockade 5B and 6B ) may work like a filter for higher frequencies.

It may explain why at the beginning of spinal anesthesia TTXs Na channels. These mechanisms described above m


10/10


ganglion neurones with different peripheral nerve conduction velocitigive an additional explanation to the well known differentialJ. Physiol. (Lond.) 359: 4763, 1985b.block observed during spinal anesthesia for different classes

HESS, P., LANSMAN, J. B., AND TSIEN, R. W. Calcium channel selectivof neurons differently equipped with TTXr and TTXs Na/ for divalent and monovalent cations. Voltage and concentration depechannels. dence of single channel current in ventricular heart cells. J. Gen. Phys

88: 293319, 1986.Future experiments will have to show whether other typesHILLE, B. Local anesthetics: hydrophilic and hydrophobic pathways for tof ion channels are involved in conduction block by local

drug-receptor reaction. J. Gen. Physiol. 69: 497515, 1977.anesthetics. JEFTINIJA, S. The role of tetrodotoxin-resistant sodium channels of sm

primary afferent fibers. Brain Res. 639: 125134, 1994.

KOSTYUK, P. G., KRISHTAL, O. A., AND PIDOPLICHKO, V. I. AsymmetriWe acknowledge the excellent technical assistance of B. Agari and O. displacement currents in nerve cell membrane and effect of internal flu

Becker and thank Drs. B. Safronov, M. Bra u, and E. Habermann for critical ride. Nature 267: 7072, 1977.discussions and for reading the manuscript. KOSTYUK, P. G., VESELOVSKY, N. S., AND TSYNDRENKO, A. Y. Ionic c

This work was supported by Japanese Exchange Program (N. Kubo- rents in the somatic membrane of rat dorsal root ganglion neuronsyama). Sodium currents. Neuroscience 6: 24232430, 1981.

Address for reprint requests: A. Scholz, Physiologisches Institut, Justus- MADEJA, M., MUSSHOFF, U., AND SPECKMANN, E. J. Follicular tissues reduLiebig-Universitaet Giessen, Aulweg 129, D-35392 Giessen, Germany. drug effects on ion channels in oocytes of Xenopus laevis. Eur. J. N

rosci. 9: 599604, 1997.E-mail: [email protected], I. M., ADAMS, M. E., AND BEAN, B. P. P-type calcium channels

rat central and peripheral neurons. Neuron 9: 8595, 1992.Received 22 September 1997; accepted in final form 2 December 1997.OGATA, N. AND TATEBAYASHI, H. Kinetic analysis of two types of N

channels in rat dorsal root ganglia. J. Physiol. ( Lond.) 466: 937, 19QUASTHOFF, S., GROSSKREUTZ, J., SCHRODER, J. M., SCHNEIDER, U., A

REFERENCESGRAFE, P. Calcium potentials and tetrodotoxin-resistant sodium potentiin unmyelinated C fibres of biopsied human sural nerve. Neuroscien

AKOPIAN, A. N., SIVILOTTI, L., AND WOOD, J. N. A tetrodotoxin-resistant 69: 955965, 1995.

voltage-gated sodium channel expressed by sensory neurons. Nature 379: RAGSDALE, D. S., MC PHEE, J. C., SCHEUER, T., AND CATTERALL, W.257262, 1996. Molecular determinants of state-dependent block of Na/ channels BEAN, B. P., COHEN, C. J., AND TSIEN, R. W. Lidocaine block of cardiac local anesthetics. Science 265: 1724 1728, 1994.

sodium channels. J. Gen. Physiol. 81: 613642, 1983. RAYMOND, S. A. AND GISSEN, A. J. Mechanisms of differential nerve bloCARBONE, E. AND LUX, H. D. Omega-conotoxin blockade distinguishes Ca In: Local Anesthetics, edited by G. R. Strichartz. Berlin: Springer-Verl

from Na permeable states in neuronal calcium channels. Pflugers Arch. 1987, p. 95165.413: 1422, 1988. ROY, M. L. AND NARAHASHI, T. Differential properties of tetrodotox

CHANDLER, W. K. AND MEVES, H. Slow changes in membrane permeability sensitive and tetrodotoxin-resistant sodium channels in rat dorsal rand long-lasting action potentials in axons perfused with fluoride solu- ganglion neurons. J. Neurosci. 12: 21042111, 1992.tions. J. Physiol. (Lond.) 211: 707728, 1970. SAFRONOV, B. V., BISCHOFF, U., AND VOGEL, W. Single voltage gated

DENNHARDT, R. AND KONDER, H. Blood and cerebrospinal fluid levels of channels and their functions in small dorsal root ganglion neuronesbupivacaine in spinal anesthesia (in German). Reg. Anaesth. 6: 7275, rat. J. Physiol. (Lond.) 493: 393408, 1996.1983. SCHMIDTMAYER, J. AND ULBRICHT, W. Interaction of lidocaine a

ELLIOTT, A. A. AND ELLIOTT, J. R. Characterization of TTX-sensitive and benzocaine in blocking sodium channels. Pflugers Arch. 387: 47TTX-resistant sodium currents in small cells from adult rat dorsal root 1980.ganglia. J. Physiol. (Lond.) 463: 3956, 1993. SCHOLZ, A., APPEL, N., AND VOGEL, W. R- and Q-type Ca-channels in

GISSEN, A. J., COVINO, B. G., AND GREGUS, J. Differential sensitivities of dorsal root ganglion neurones. Pflugers Arch. 433: R143, 1997.

mammalian nerve fibers to local anesthetic agents. Anesthesiology 53: SCHOLZ, A., KUBOYAMA, N., HEMPELMANN, G., AND VOGEL, W. Effect467474, 1980. lidocaine and bupivacaine on TTXr and TTXs Na/-channels in rat dorHAMILL, O. P., MARTY, A., NEHER, E., SAKMANN, B., AND SIGWORTH, F. J. root ganglion neurones (Abstract). Pflugers Arch. 431: R83, 1996.

Improved patch-clamp techniques for high-resolution current recording SCHOLZ, A., REID, G., VOGEL, W., AND BOSTOCK, H. Ion channels in humfrom cells and cell-free membrane patches. Pflugers Arch. 391: 85100, axons. J. Neurophysiol. 70: 12741279, 1993.1981. ULBRICHT, W. Kinetics of drug action and equilibrium results at the no

HARPER, A. A. AND LAWSON, S. N. Conduction velocity is related to mor- of Ranvier. Physiol. Rev. 61: 785828, 1981.phological cell type in rat dorsal root ganglion neurones. J. Physiol. WILDSMITH, J. A., BROWN, D. T., PAUL, D., AND JOHNSON, S. Structu(Lond.) 359: 3146, 1985a. activity relationships in differential nerve block at high and low frequen

stimulation. Br. J. Anaesth. 63: 444452, 1989.HARPER, A. A. AND LAWSON, S. N. Electrical properties of rat dorsal root

Documents

TTXr J Neurophysiol 1998 Scholz 1746 54