Differential actions of brevetoxin on phrenic nerve and diaphragm muscle in the rat

Toxlrn~, VoL 31, No. 4, pp. 439-470, 1993 . Aimd in Omt &iLie.

Per~mon Aer 1.1d

DIFFERENTIAL ACTIONS OF BREVETOXIN ON PHRENIC NERVE AND DIAPHRAGM MUSCLE IN THE RAT* SHARAD S . DFSHPANDE, MICHAEL ADLER

and

ROBERT

E.

SHERII)AN

Neurotoxicology Branch, Pathophysiology Division, USAMRICD, Aberdeen Proving Ground MD 21010-5425, U.S .A. (Received

31 July 1992;

accepted 7 October

1992)

and R. E. SHERIDAN. Differential actions of brevetoxin on phrellic nerve and diaphragm muscle in the rat. Toxicon 31, 459470, 1993 .-The mechanism of inhibition of skeletal muscle function by brevetoxin (PbTX-3) was examined in vitro in the rat phhenic nerve-diaphragm preparation. PbTX-3 in low concentrations ( < 0.06 ~M) preferentially blocked conduction in the phrenic nerve without altering the resting membrane potential of the muscle fibers. Endplate potential failure occurred in an all-ornone fashion in the presence of PbTX-3 ( > 0.06 ~M). An increase in the frequency of miniature endplate potentials resulting from nerve terminal depolarization was observed only after endplate potential failure. Higher concentrations of toxin ( > 0.31cM) depressed directly-elicited muscle twitches and produced significant muscle membrane depolarization. Tetrodotoxin was effective in reversing membrane depolarization and alterations in MEPP frequency caused by PbTX-3. These findings suggest that diaphragmatic failure in PbTX-3 is primarily caused by a block of impulse conduction in the phrenic nerve due to a higher sensitivity of nerve than muscle membrane to the toxin. S . S . DE3HPANDE, M . ADLER

INTRODUCTION

(PbTXs) are cyclic polyether compounds produced by the marine dinoflagellate Ptychodist"us brevis. The lipid-soluble neurotoxins have been shown to impair nerve and skeletal muscle function (GALLACI~t and St~m~Ntcx-GALLA~I~R, 1980, 1985; SHIIVNICR-GALLACiHER, 1980; BADEN et al., 1984; ATCHISON et al., 1986), produce airway smooth muscle contraction (RICHARDS et al., 1990) and cause cardiac arrhythmia (RODOER3 et al., 1984). Crude extracts of PbTX induced repetitive discharges of action potentials in squid axon membranes (WBSTERFIEL.D et al., 1977 ; PARl~r1TIBR et al., 1978). Among the toxins isolated from the parent source and purified by HPLC, PbTX-2 and PbTX-3 have been studied in greater detail for their mechanism of action (BADEN et al., 1984; ATCHL90N et al., 1986). The main target of these toxins is the sodium channel of excitable membranes. The PbTXs bind to a specific site on the sodium channel complex to BREVE7b7QNS

'Opinions or aseertiona contained hetsin are the private views of the author end aro not to be construed es o~ciel or ad reflecting the views of the Army or the Department of Defe~e . In conducting the research described in this npoR, the inveatigaton adhered to the Guidejor the Care and Use of laboratory .lninrols, National Institutes of Health publication 85-23 . 459

460

S . S . DESHPANDE et al.

induce persistent activation leading to increased Na+ flux and depolarizstion of the excitable cells in the resting state (PoLt et al., 1986). Studies aimed at investigating mechanism of membrane depolarization induced by PbTX-2 and PbTX-3 have suggested that these toxins modify both activation and inactivation of sodium channels (HuArta et al., 1984; ATC>:nsox et al., 1986). In a recent investigation with NG108-15 neuroblastomaglioma cells, single channel and whole-cell current recordings demonstrated that PbTX-3 shifted activation of sodium channels to more negative membrane potentials without any change in inactivation kinetics ($I~tIDAN and Ant~t, 1989, 1990). At the mammalian neuromuscular junction, PbTX-3 produces membrane depolarization and a large increase in the frequency of miniature endplate potentials (MEPPs) that can be antagonized by tetrodotoxin (Trx) (GALLAGI~t and SHtxlvtctc-GwLLAat;a~t, 1980, 1985 ; Arc~oN er al., 1986). Recent experiments (TSAt et al., 1991) with brevetoxin-B (synonymous with PbTX-2 according to taxonomic reclassification, see Poet et al., 198 demonstrated that compound action potentials of the phrenic nerve could still be recorded under conditions where complete neuromuscular blockade should have occurred. In addition, the toxin produced an initial increase followed by a decrease of antidromic firing in the phrenic nerve. Based on the perineural waveform recordings from the nerve terminal regions of mouse triangularis sterei muscles, a report from the same group of investigators (TsAt and G~x, 1991) concluded that PbTX-2 affects sodium, potassium and calcium currents in the nerve terminal resulting in the failure of neuromuscular transmission . The purpose of the present investigation was to examine PbTX-3 for its ability to block conduction in the rat phrenic nerve and to determine the onset and degree of membrane depolarization and neuromuscular failure in the diaphragm muscle . Since a specific pharmacologic antidote for PbTX-3 poisoning is not known, we attempted to antagonize toxin-induced neuromuscular blockade with pharmacologic agents . The data suggest that PbTX-3 has a profound blocking action on nerve conduction at concentrations that produce no depolarization of the muscle membranes and no apparent depolarization of the nerve terminal . MATERIALS AND METHODS .lninrals and tissue rcmova! Male Sprague-Dawky rata (200-300 g) were used in these experiments . Diaphragm musdes together with the phrenic nerves were rapidly removed under constant flow of phy:;ological solution from rats sacrifloed by decapitation after being rendered unconscious in a ly0= chamber. The tissues were kept in bathing solution and bubbled with a 95X O~/SY° tJ0= mixture. All experiments were perfonmed at room tempetaturo (21-25°~ . Contraction studks Hemidiaphragats with phrenic nerve were swpended in a chamber (15 ml capacity) for recording wntractions in response to nerve stimulation (0 . I msx pulses at 0.1 Hz) alternated with those elicited by direct dimulation of the muscle (3.0 msoc pulses at 0.l Hz) . The intercity of stimuli was adjusted to 4 x the threshold to evoke maximal muscle responses . Isometric oontractiona were rocorded using a strain gauge (crass FT-03) and a chart rotorder (Gould, Model 2800) . RtcoroYng of nerve condtection A length of phrenic nerve (25-30 mm) was placed in a nerve chamber (2 ml capacity) and auperfused with oxygenated (95% O~/SY° 1y0~ bathing solution at room temperaturo using a micro-perfLsion pump. The entiro length of the phrenic nerve lay in the solution, except for the proximal and distal cut e~s which were drawn into suction electrodes. The tips of the stainless steel wire in the suction electrodes made contact with the nerve. An indifferent electrode made contact with the bath fluid. The nerve was stimulated onx every 2 min by aupramaximal pulses of 0.1 cosec duration deliverod from a stimulator and ttiggered by a pulse protocol pengram

PbTX-3 Action on Rat Diaphragm

46 1

using pClamp software (Axon Instruments Inc.) . Compound action potentials were amplified by an AC amplifier (WPI Model DAM-8) and displayed on a storage oscilloscope (Tektronix model SI 10). The signals wero digitized and stored on a Zenith data systems computer for offline analysis of the peak amplitude (mV) and duration (meet) of the responses using version 5.0 of the pClamp software. ekctropkyrtology Left hemidiaphregtna with phrenic nerve were mounted in a chamber (6 ml capacity) for electrophysiological recording of the resting membrane potential, MEPPn and endplate potentials (EPPs). The muscles wero perfused with oxygenated physiobgical solution and all recordings were made at room temperature. Intracellular recordings of resting membrane potentials and MEPPs from the surface fibers of diaphragm muscleswere made according to methodsdescribed earlier (DESrrPerana et d., 1976). The recording micropipettes wero filled with 3 M KC1 and had resistanoes of 10-1 S wegohms. The time oonataat of the recording circuit with a 15 me~hms microdectrnde was approximately 40 peat. The membrane potential vahres wero obtained from a micxoelectrode amplifier (WPI Madd K-S 700) . MEPPs wero displayed on an oscilloscope screen and recorded simultaneously on a Neurocorder videotape unit (Neuro Data Instruments Corp., Madd DR~84) and a chart recorder (Gould model 2200 S). In experiments where EPPs wero to be reoordod, muscle contractions wero prevented either by bathing muscles in physiological solution containing 8 mM Mg2+ and 1 mM Cap+ or by using cut muscle preparations (B~rtareu and Ln ~ ~^ , 1968). The latter meWod allowed ua to examine PbTX-3 effects at the neuromuscular jrmction without altering Mg=+ or Ca=+ in the bathing solution . The distal end of the phranic nerve was àimulated with sgnaro pulses (0 .1 Hz) of supramaximal strength sad of 0.1 cosec duration using a suction electrode. The evoked EPPs ware recorded inhaeellularly from the endplate region of the surface fibers. The methods used to locate endplate ragions and record F.PPa were eaeeatially a®ilar to those described earlier (F~rr and IG~ 1951). Merck

Drrtgr and solutiarrs

The bathing solution had the following composition (in mM): NaCI, 135; KCI, 5.0; CaC1r 1.8 ; MgClr 2.0; Na=IiPO 1 .0; NaHCOr IS and glucose, 6 (pH 7.4 when equilibrated with 95'x. 02+5'/e COQ. Tetrodotoxin (TI7{) was purchased from the Sigma Chemical Co. (St. Louis, MO, U.S .A.). Purified lyophilized PbTX-3 (Pou et al., 1986) was obtained from the U.S. Army Medical Research Institute for Infectious Dbeases (Ft. Derrick, MD, U.S.A.). The stock solution (1 mM) of PbTX-3 was made in a mixturo of chloroform, methanol and absolute ethanol (1 : 1 :S) and stored at -IS°C. Working oonoentrations of the toxin wero made in bathing solution prior to use. All toxin concentrations aro expressed as final concentrations bathing the nerve or muscle . Dotes mwlysLr

Statiadcal analysis between the means of values obtained for various troatments was performed using unpaired t-teats (two-taikd) . P values < 0.05 wero considered significant. RESULTS

EtFxts of PbTX-3 on the twitch tension reeordod from the diaphragm muscles are shown in Fig. 1 . Application of 0.005 pM PbTX-3 led to a reduction of nerve~licited twitches after an initial lag of 1 .5 min followed by complete blockade within 6 min. Rocovery was observed upon washing, which required 14 min in the experiment illustrated here . When the dose of toxin was increased, the rate of inhibition of indirectly-elicitod twitches was faster, although the lag time for onset of block was not altered. The time required for recovery also depended on the dose of PbTX-3 and was slower attar exposure to higher toxin concentrations . Nerve-elicited twitches were selectively blocked by the toxin, leaving twitches indutxd by direct muscle stimulation unatFected (Fig. 1). Concentrations of PbTX-3 in excess of 0.5 ~tM were required for the depression of directly elicited twitches . Concentration-response curves for compound action potentials of the phrenic nerve and the mean resting mtmbrane potential recorded from surface fibers of diaphragm muscles are shown in Fig. 2. The nerve compound action potential was almost completely blocked in the presence of 0.2 uM PbTX-3 with an tC~ value of 0.06 ~eM. A significant

462

S. S. DESHPANDE e~ al.

0nM PbTX-3

10 nM PbTx-a

wosn

20nM PbTX-3

000 nAl PbTX-3

~IOp

Ftc. I . Se~cnve awcx of rmtve-evox~ wsc~ ooNnucnoNS av PbTX-3 . The muscle was stimulated directly (3 cosec duration at 0.1 Hz, larger responses) alternated by nerve-evoked twiuhes (0 .1 cosec duration at 0.1 Hz, smaller responses) with supramaximal stimuli. After control recording, exposure of muscle to 5 nM PbTX-3 (at arrow) produced a gradual block of nerve-elicited twitches (onset 1 .4 min) leading to a oompleu block at 5.4 min after toxin application. A compleu recovery (right panel, óret row) was obtained 14 min after wash . Filled circles denou the points at which additional washes were performed. The responses of the muscle to 10 and 20 nM 16TX-3 are shown in the second and third row, respectively. The onset times for 10 and 20 nM PbTX-3 were 1 .1 and 1 .2 min, respectively. A complete block occurred at 4.4 min (10 nM) and 3.6 min (20 nM). The muscle showed recovery from 20 nM 1bTX-3 at 21 .3 min after wash. Nou lack of effect on directly stimulated muscle Sbers even after exposure to 1bTX-3 as high as 0.5 pM (arrow, bottom right panel).

(P < 0.0001 with respect to control) muscle membrane depolarization occurred after application of PbTX-3 only at concentrations in excess of 0.3 pM. The mean membrane potentials in control and toxin-treated muscles (1.O~M) were -72 f 1 .3 mV (n = 44 fibers, 5 muscles) and -61 f 1 .3 mV (n = 70 fibers, 7 muscles) respectively (Fig. 2). The time course of PbTX-3 (0.3 pM) eßect on the nerve and muscle is shown in Fig. 3. The phrenic nerve compound action potential was completely blocked at 20 min. Continuous recording in this experiment showed that up to 40 min, membrane depolarization in the muscle fibers had not occurred. Similarly recording from the endplate region did not show any alteration in the frequency of MEPPs up to this time. To determine ifthe PbTX-3-induced depression of neuromuscular transmission was due to a graded reduction of quantal content of the EPP, two types of muscle preparations were used. In one, the muscle was initially exposed to a high Mgt+ (8 mM) and low Cat+ (1 .0 mM) bathing solution to block muscle contractions . EPPs were recorded under control conditions and the muscle was exposed to PbTX-3 (0.1 ~M). Within 10 min (range 6-15 min) an abrupt cessation of EPPs in response to nerve stimulation occurred without any alteration in MEPP frequency. The presence of 8 mM Mg2 + largely prevented the expected increase in the MEPP frequency even up to 1 hr of perfusion with toxin. As seen from Table 1, in Mgt+ blocked preparations the MEPP frequency in the presence of PbTX-3 was not significantly dißerent from that obtained for high Mgt+ alone (P > 0.1).

PbTX-3 Action on Rat Diaphragm

46 3

-50 ,~ 0 c a+

c

0 .8

a m v -60 ~

0 .6

rt V ar+

Y

m 07

a

0 .2 0 .0

Fla . 2 .

EPPEGT of

3

10 -°

10-~

10 -°

PbTX-3 Concentration (11)

10~

PbTX-3

ON TEII? COMPOUND ACTION P078N17AL IN TES PHRENIC NFRVE AND ON THC RFSIßdO 1~AANE POT~1'r1AL ad TES DIAPHRAGM MUSCLE .

The compound action potential (CAP) value (filled triangles) for each concentration is expressed as a ratio of peak amplitude in toxin over control amplitude. The Ic,s for block of CAP was 0.06 ~M . Note muscle membrane depoleriTation (filled circles) at PbTX-3 concentrations > 0.3 pM (P < 0.0001 with respect to control value for 1.0, 3.0 and 10 ~IuI PbTX-3). The values for peak CAP and membrane potential were obtained when maximal responses for a given concentration for PbTX-3 were achieved (usually within 20 to 30 min). Bars indicate S.E. of individual values recorded from 5 to 7 muscles for each concentration.

The other type of preparation used for EPP recording was the transversely cut muscle . Exposure of these muscles to PbTX-3 (0.1 ~M) produced an abrupt cessation of EPPs at 9 min (range 5-12 min) . The records obtained from a representative experiment are shown in Fig. 4. In this experiment the EPP was blocked completely at 6 min when the MEPP

E m v .. á E e n. V

TIME (mln) FIG . 3. A rrPICAL

exP>aenr~rr SHOWIIVO THe Tn~ oovRSE oP NE~IVE AND MUSCLE.

PbTX-3 (0 .3 pM)

ACr10N

ox

THE

After control recordings (value shown at 0 time) muscles were auperfueed with PbTX-3 . Nou complete block of CAP at 20 min. No depolariation of the muscle membrane nor increase in the MEPP frequency was observed up to 40 min etter toxin addition.

S . S. DESHPANDE et aJ. T~atB 1 . EPrsc,-r o~ PbTX-3 ox rra~ rteat2~rcr or MEPPs w ~ aerH® na t~osestnt. exn xtax Mgr+ (8 mM, Ca'+ 1 mM) rx~ot oacu. sottmoN Condition Control solution PbTX-3 (0.1 pM) at EPP/AP blockt 30 min in PbTX-3 (range 2040 min) Wash ( > 40 min)$ Iügh Mg:+ solution alone PbTX-3 (0.1 pM) in high Mgr+ solution$ (3011 min)

MEPP frequency (Hz) 0 .8710 .01 0 .9010 .12 (7) 41 .83 t 5 .27 1 .0616 .00 (3) L19í0 .47 (5) 6.10 f 2 .70 (S)"

Values are mean f 3 .E. Numbers in parentheses represent number of cells in three muscles. 'P < O.OS with respect to walrol value; "P > 0 .1 with speck to high Mg'+ solution alone . tEPP or nerve-elicited action potentials were blocked at about 9 min (range: 5-12) after perfusion with PbTX-3 . $Reduction in the MEPP frequency was noticeable as early as S min after wash; however, rooovdy of EPP/AP oocarred after about 40 min . ~Mg:+ effectively prevented an increase is MEPP frequency induced by PbTX-3 . The increase in MEPP frequency after exposuro to PbTX-3 (0.1 pM) in normal physiological solution subsided when the perfusion medium was switched to one with high Mg=+ and toxin; the enhanced transmitter release was blocked as early as S min.

frequency was comparable to that seen under control conditions. Spontaneous transmitter release was increased only after 30 min exposure to PbTX-3. Upon washing, the MEPP frequency underwent a significant recovery within 3 min, whereas the nerve-elicited muscle response (shown by the action potential) could be recorded only after 45 min of wash. The mean MEPP frequency of 0.90 Hz observed at a time when all nerve-elicited activity was abolished (about 9 min) was not significantly different from that shown for control conditions (0.87 Hz, Table 1). In general, the recovery in amplitude of the compound action potential of the phrenic nerve or reversal from neuromuscular depression was partial and inconsistent after wash, especially following incubation with high PbTX-3 concentrations ( > 0.5 ~cM) or prolonged exposure times. In an effort to antagonize PbTX-3-induced neuromuscular depression, we tested Mgt+, TTX, lidocaine and 4-aminopyridine (4-AP) for their ability to reverse the block. Our approach consisted of recording nerve~licited muscle contractions or measuring membrane potentials and MEPP frequencies. In two experiments where contractile responses were recorded, the muscles were preincubated with 4 to 6 mM Mg2* for 30 min before PbTX-3 (0.03 pM) application. The toxin produced total blockade of nerve-elicited twitches leaving directly stimulated muscle contractions unaffected . In one experiment, PbTX-3 (0.3 pM) depolarized muscle membranes from a control value of -85 t 1 .6 mV (n = 16 fibers) to - 55.8 t 2.3 mV (n = 20 fibers, P < 0.001). Perfusion of the muscle with 4 mM Mg2 * in the presence of PbTX-3 (0.3 pM) resulted in no beneficial effect; the resting potential in the presence of both PbTX-3 and Mgt* was -61 .3f 1.7 mV (n = 19 fibers, P > 0.1 with respect to PbTX-3 alone) . Similarly, Cat* (fí mM), the local anesthetic lidocaine (up to 0.1 mM) and 4-AP (0.03 mM) could not protect from, or reverse neuromuscular depression produced by, PbTX-3. In contrast, TTX was very effective in antagonizing PbTX-3-induced membrane depolarization (Fig. ~ and alterations in MEPP frequency, but as expected, did not restore the EPP or nerve compotmd action potential (not shown). The exposure of muscles to TTX (1 ~M) produced a complete recovery of

PbTX-3 Action on Rat i7iaphragm

MEPP

EPP CON

PbTX-9 (100 nM)

WASH 3 mln. .{..--.~-~

AP

IÉ 0 a

40 moon

400 mpo

Fra 4. Aaatrrr cmse~noN or EPPs nv ~ rawc~ oe PbTX-3. EPPs were recorded in isolated diaphragm muscles . Muscle contractions were eliminated by transverse wts of fibers on either side of the endplate region. Note the presence of MEPPs in control (t70N)t trace . Six minutes after toxin perfusion, blade of the EPP occurred suddenly, without graded reductions in quantal content. At thin time the amplitude and frequency of MEPPs were within the normal range . At 30 min when MEPP frequency had increased, no nerve-elicited activity was evident. After a 45 min wash, the muscle generated an action potential (AP, pear off scale) after nerve stimulation and the MEPP frequency recovered to control levels .

465

466

S. S. DESHPANDE et al. -ao E .. -so 0

c

ó o0 c

-SO -40

-30 0 w -20

E

0 -10 0

CON

0.1

0.3

1.0

3.0

10 .0 WASH

PbTX-3 Concontratlon (~11)

Tr1I

FIG. S. THE RL'9IJLTS OF AN EXPERISQ~iT UE110N3171ATING THE EFPECI7VE~Q~ OF TTX IN REVER~NCi

~AxE nEroI.AnIZers~IV Ixnuc~ ar PbTX-3 . The resting membrane potential was not altered with PbTX-3 superfusion through 0.3 ~I . Note We significant depolarization (asterisks) in the presence of 1 .0 (P < 0.04), 3.0 (P < 0.0001) and 10.0 (P < 0.001) ~I PbT7C-3. Prolonged wash for 45 to 60 min alter 10 pM toxin led to a partial but significant recovery (P < 0.01) with respect to the mean value for 10 pM. Resting potentials after wash were significantly (P < 0.001) lower than control. Superfusion with TTX (1 uM) restored membrane potentials to control within 30 min. Hare indicate S.E. of 10 to 20 individual values for each treatment.

excessive spontaneous transmitter release caused by PbTX-3 within 20 min. TTX applied prior to or after PbTX-3 was equally effective in protecting muscles from depolarization and from the elevation of MEPP frequency. The dramatic action of TTX in reversing the membrane potential changes following PbTX-3 is shown in Fig. 5. Significant membrane depolarization was evident in this experiment following superfusion with 1 .0 pM PbTX-3. The muscle was depolarized by 20 and 24 mV at 3 and 10 PM toxin concentrations, respectively . A 20 min wash with physiological solution produced a small ( ~ 8 mV) but significant recovery . Bathing the muscle subsequently in TTX (1 pM) restored the membrane potential to control levels . DISCUSSION

PbTX-3 in concentrations < 0.06 pM produced a selective block of phrenic nerve compound action potentials . Higher concentrations ( > 0.3 pM) were required to depolarize the muscle membrane, while an increase in the frequency of MEPPs as a result of depolarization of the presynaptic nerve terminal was seen at an intermediate concentration ( > 0.1 ~M) of toxin. The contractions elicited by direct stimulation of muscle were not depressed by concentrations of the toxin up to 0.5 pM (Fig. 1). All of the above observations, with the exception of phrenic nerve conduction block, are qualitatively similar and consistent with earlier reports on the mammalian skeletal muscle using crude toxin extracted from PtychoaKscus brevis (GALLA~ and Sfmvrnc>~-GAC.t .Aat~t, 1980, 1985 ; S>:mvrncx-GwLLAt3I~, 1980) and purified PbTX-3 (ATC~nsoN et al., 1986). The results presented here reveal an important finding about PbTX-3. The time course and concentration-response study of PbTX-3 on the phrenic nerve and the diaphragm muscle clearly indicate that the toxin in concentrations less than 0.2 ~M blocked compound action potentials of the phrenic nerve and that such concentrations of toxin did

Pb17{-3 Action on Rat Diaphra~

467

not depolarize the muscle membrane . The presynaptic effect exhibited as an increase in the frequency of MEPPs appears much later, approximately 40 min after exposure of the muscles to PbTX-3. In an earlier investigation (GALLAGi~R and SFmvrncic-Gwt.~c~t, 1980) it was reported that, when crude PbTX was applied to the phrenic nerve~íiaphragm preparation, membrane depolarization occurred within 3 min of application of toxin while blockade of the EPP was observed a few minutes later. The amplitude of the antidromic compound phrenic nerve action potential was not affected. These results are in contrast to those reported here . Two factors could account for the discrepancy. First, we usod pure PbTX-3, which in comparison to crude toxin could have selective action on the neuronal sodium channels even in very low concentrations. Second, if the recording electrodes for the phrenic nerve in the previous study (GALLAGHBit and S~trticic-GAr i.~Gt~t, 1980) were in a pool of mineral oil, it is likely that high lipid solubility of PbTX-3 caused some fraction of the toxin to be retained in the mineral oil compartment and therefore become unavailable to the nerve. A recent investigation (Ts~ et al., 1991) with PbTX-2 showed that at a concentration of 0.11 ~M, nove-elicited contractions of the mouse diaphragm were completely blocked within 10 min while amplitudes of compound action potentials recorded from the phrenic nerve were reduced only by approximately 20% of the control value during the 10 min recording period . These results impliod that sodium channels in the nerve terminal and muscle membrane were the primary target for PbTX-2. Our results with PbTX-3 strongly suggest that the primary cause of failure of neuromuscular transmission in the diaphragm muscle is conduction block in the phrenic nerve. If the toxin caused selective depolarization of the nerve terminal, EPPs recorded from muscles bathed in high MgZ+ solution or from cut muscle preparations would have shown a gradual decline in amplitude instead of an abrupt cessation in the all-or-none fashion seen here. These results are consistent with the notion that a block in nerve conduction is the primary cause for neuromuscular failure. Nerve terminal depolarization (indicated by an increase in MEPP frequency) and muscle membrane depolarization occurred much later in the sequence of events and these effects showed a different dose-dependence . For a complete analysis of the differential actions of PbTX-3 on skeletal muscle function, it would be of interest to determine the kinetics and concentration-response profile for the PbTX-3-induced depolarization at its three principal sites of action : the myelinated nerve fiber, the nerve terminal and the muscle membrane . Unfortunately, the small size and unfavorable geometry of the nerve axon and terminal preclude direct micrcelectrode measurements . However, an approximate value for the nerve terminal depolarization can be calculated from an established relationship between presynaptic membrane potential and MEPP frequency (LII.EV, 1956; Iü~tz, 1962). The change in membrane potential of the presynaptic membrane in the presence of PbTX-3 can then be deduced from the increase in MEPP frequency. In the present study PbTX-3 (0.1 uM) increased MEPP frequency from about 1 Iiz to 42 Hz, which implies a depolarization of approximately 20 mV at the presynaptic membrane . This depolarization is very close to the maximum depolarization (approximately 29 mV) seen in the muscle fibers exposed to PbTX-3. Using the voltage clamp technique, previous studies with crayfish (Hu~ta et al., 1984) and squid (ATCtnsoN et al., 1986) giant axons have attributed PbTX-3-induced membrane depolarization to the modification of a fraction of sodium channels which open at potentials more negative than normal and which inactivate at a slower rate in the presence of toxin. A more recent study using clonal NG108-15 cells (Sm~tm~x and Ange, 1989) has shown that PbTX-3 merely shifts activation of sodium channels to potentials 6 to

468

S. S . DESHPANDE et d.

8 mV more negative than normal without any alteration in inactivation kinetics. Based on these findings and on the affinity of PbTX-3 for its binding site on the sodium channel it would appear that steady-state shifts in the activation voltage of a few millivolts in the negative direction (towards the resting potential) could result in a large depolarization that would significantly reduce the number of channels available for impulse conduction . Since the relationship between toxin binding and steady-state depolarization depends on the ratio of sodium channel density to membrane leak currents (SI~RIDAN and ADI~It, 1990), the PbTX-3 sensitivity of nerve and muscle will be influenced by the distribution of sodium channels in these tissues. Differences in sodium channel density will alter the apparent effectiveness of the toxin even for equal toxin affinities. Apart from the differences in the sodium channel characteristics of the nerve and muscle membranes, the fact that PbTX-3 causes block of conduction in the nerves at low concentrations has significant bearing on the poisoning by red-tide toxins in general. In contrast to saxítoxin and TTX, PbTX-3 is lipophilic, thus allowing rapid passage of PbTX-3 through biological membranes and diffusion barriers, particularly in myeGnatod nerves and CNS neurones . In vivo studies in cats have shown that cardiovascular and respiratory irngularities occur after PbTX-3 (Botetsox et al., 198 . These actions were attributed to reflex activation of the BezoldJarisch reflex via stimulation of vagal afferents. They may also be of central origin through direct effects of toxin on the cardiorespiratory center (Botetsox et al., 1985 ; Joxxsox et al., 1985). Respiratory failure appears to be the primary cause for lethality in guinea-pigs receiving slow intravenous infusions of PbTX-3 (FtiAxz and LECLAtxE, 1989). Recent work in this laboratory has shown that, in guinea-pig hippocampal slices, bath-applied PbTX-3 produced a concentration-dependent (tcm = 0.08 pM) depression of orthodromicmlly evoked population spikes from the neurones in area CA1 (Apt.AtvD et al., 1993). These observations further indicate the importance of neuronal sodium channels in PbTX-3 toxicity . The only effective antagonist of PbTX-3 action on membrane depolarization was TTX as observed here (Fig . 5) and reported earlier by others (GALLAGtn?tt and St-mvxtcx-GAt t.Aat~tt, 1985 ; HuAxc et al., 1984 ; ATCxrsox et al., 1986). We had hoped that preincubation of muscles with 4~ mM Mgt+ would protect them from PbTX-3 induood membrane depolarization and neuromuscular block . The rationale for these experiments was based on findings that the divalent canons raise the threshold for electrical excitation of nerve and muscle. This stabilizing effect is due to the production of local electric fields near the membrane interface which bias voltage sensors within the membrane (HILLS et aL, 197 . Mgt+ would thus be expected to antagonize the PbTX-3-induced shift in channel activation to more negative membrane potentials . Contrary to expectations MgZ+ failed to show any antagonism of PbTX-3 action on muscle contractions or EPPs . The divalent cation was, however, very effective in preventing PbTX-3 induced increases in spontaneous transmitter release. Mgt+ also caused a prompt reversal of the elevated MEPP frequencies when muscles were first exposed to PbTX-3 in control solution and subsequently switched to a solution containing the same concentration of toxin in high Mgt+ solution (Table 1) . Given that EPPs are not restored, it seems likely that Mgt+ directly or indirectly inhibits influx through voltagesensitive Ca'+ channels (DODGE and Rwxwxusto~, 1967; Mtn.LEtt and Fnvx~srsttv, 1974). It is apparent that because of the highly complex nature of the sodium channel, it may be difficult to evolve specific therapeutic measures until we understand more about specific binding sites for various toxins . In fact the PbT'Xs may be binding to a new site on the sodium channel (designated site 5 by $HARREY et al., 1987). In their study with voltage-

PbTX-3 Action on Rat Diaphragm

469

sensitive sodium channels in rat brain synaptosomes, PbTX-2 enhaneod the activation of sodium channels by veratridine, aconitine, and batrachotoxin acting at site 2 of the sodium channel without inhibiting binding of saxitoxin or scorpion venom, which bind at site 1 and 3, respectively . These results suggested allosteric modulation of neurotoxin binding by PbTX-2. By contrast, PbTX-2 blocked specific binding of PbTX-3 on site 5 of the sodium channel (POLI et al., 1986), suggesting that these forms of PbTX share a common well-defined binding site. So far only antibody to PbTX-2 has shown prophylactic and therapeutic efïectiveness in rats infused with PbTX-2 ('I~1PLgroN et al., 1959). In conclusion, the present study clearly demonstrates the selectivity with which PbTX-3 sí%cts neuronal membranes. The sodium channels in axons and CNS neurones appear to be more sensitive to the depolarizing action of PbTX-3 than those from muscle membranes (APLAND et al., 1993). REFERENCES An.erro, J. P., Ant~t, M. and St~tm~uv, R. E. (1993) Brevetoxin degreases synaptic transmission in guinea pig hippocampal slices . Brain Res. Bull. 30. A~rc,~nsotv, W. D., Lt>s~, V. S., Neaetr~stu, T. and Voc~et., S. M. (1986) Nerve membrane sodium channels as the target site of brevetoxins at neuromuscular junctions. Br . J. Pharmac . 89, 731-738. Benetv, D. G., Bn~rrezr, G., Decrtm, S. J., FOt.uFS, F. F. and Lemur, I. (1984) Neuromuscular blocking action of two brovetoxinn from the Florida red tide organism PtychodLrcrv brew. Toxicon 22, 75-84. 9easrnn, J. A. B. and i n~1PJ~, G. (1968) Transversely cut diaphragm preparation from rat. Arch. /nt. Pharnwcodyn. 175, 373-390. Hoarsox, H. L., McC~trrtrv, L. E. and Et.us, S. (1985) Neurological analysis of respiratory, cardiovascular and neuromuscular effects of brevetoxin in cats. Toxicon 23, 517-524. D~rrrwNne, S. S., Ar.suQueaQue, E. X. and Gurx, L. (1976) Neurotrophic regulation of projunctional and postjunctional membrane at the mammalian motor endplate . Fxp. Newel. 53, I51-165. Dome, F. A., Jr and ~, R. (1%7) Co-operative action of calcium ions in transmitter rolcase at the neuromuscular junction. J. Pkysiol., Loral. 193, 419-432. FeTT, P. and ICerz, B. (1951) An analysis of the endplate potential recorded with an intracellular electrode. J. PhysioL, Lond. 115, 320-370. FaNVZ, D. R. and LECutas, R. D. (1989) Respiratory effects of brevetoxin and saxitoxin in awake guinea pigs. Textron 27, 64754. Ger.r.~ar~a, J. P. and Sxnvrnctc-G+i.uatn=.a, P. (1980) Effect of Gymnodinium breve toxin in the rat phrenic nerve diaphragm preparation. Br. J. Pharnrac. 69, 367-372. Gxt.uarma, J. P. and Swnrtics:-Gw>~a~t, P. (1985) Effects of crude brovetoxin membrane potential and spontaneous or evoked end-plate potentials in rat hemidiaphragm. Textron 23, 489-d%. Htu~, B., WoonttuLt., A. M. and Srrertno, B. I. (1975) Negative surface charge near sodium channels of nerve: divalent ions, monovalent ions, and pH . Phil. Trans. R. Soc. Lend. B 270, 301-318. Huena, J. M. C., Wu, C. H. and BeuErr, D. G. (1984) Depolarizing action of a red tide dino8agellate brevetoxin on axonal membranes. J. Pharnwc . exp. Ther. 229, 615fi21 . JotuvsoN, G. L., Srutrs, J. J. and Et,us, S. (1985) Cardiovascular effects of brevetoxin in dogs . Textron 23, 505-51 S. ICerz, B. (I%2) The transmission of impulsen from nerve to muscle, and the subcellular unit of synaptic action . Proc. R. Soc. B 155, 455-479. Lu~sr, A. W. (1956) The effects of presynaptic polari7stion on the spontaneous activity of the mammalian neuromuscular junction . J. Physiol., Load. 134, 42743 . Mut.t,He, R. U. and FttattestnN, A. (1974) The electrostatic basin of Mg* * inhibition of transmitter release . Proc. nacre. Acad. Sci. U.S.A . 71, 923-926. Pea~vrrFrt, J. L., Nea~rr~n, T., Wnsox, W. A., Tamfi, N. M., SADAOOPA-RAIIANUJAII, V. M., Rte, M. and Rw, S. M. (1978) Electrophysiological and biochemical characteristics of Gynrnodireirare brave toxins. Toxlcon 16, 235-244. Pot.t, M. A., MHtvne, T. J. and Benw, D. G. (1986) Brevetoxins, unique activators of voltage-sensitive sodium channels, bind to specißc sites in rat brain synaptosornes. MoJ. Pharneac. 30, 129-135. RrcFUnna, I. S., Ktttauarn, A. P., Bapotts, S. M. and PreRC~, R. (1990) Florida rod-tide toxins (brevetoxins) produce depolarvation of airway smooth muscle. Textron 2g, 1105-11 I1 . Itooa®rs, R. L., Cxou, H. N., Tt~ou, K., Artrßu, Y. and Srmazu, Y. (1984) Positive isotropic and toxic effects of brevetoxin-B on rat and guinea pig heart. Toxicol. appl. Pharnurc. 76, 296-303.

470

S. S. DESHPANDE rt al.

Stient~r, R. G., Jove, E., Couawtro, F., Heues, D. G. and Car~~"~.~, W. A. (1987) Alloeteric modiilation of neurotoxin binding to voltago-senaitive sodium duumela by Ptychoayscvs brrvir toxin 2. Mol. Phanrwc. 31, 273-278. St~emex, R E. and Auufle, M. (1989) The actions of rad tide toxin from Ptydradlacen brcvir on single sodium channels in mammalian auroblaatoma cells. FEBS Left . 247, 448-452. Srmwuex, R. E. and Auuse, M. (1990) Functional effects of sodium channel agoniata. Proc. l2tb A. Int. Coq/: IEEE Eng Afrd BtoJ. Soc. 12, 1657-1659. S~nxtacr-Geu.eat~t, P. (1980) Paeaible mxhanisma of action of Gy~ro~o~ábehon burr toxin at the mammalian neuromuscular junction. Br. J. Pharmae . 69, 373-378. T~e,irr+ox, C. B., Pou, M. A. and $OLOW, R (1989) Prophylactic and therapeutic use of an antibrevetoxin (PbTx-2) antibody in conscious rats . Toxtrnn 27, 1389-1395. Tsee, M. C. and Cf~e, M. L. (1991) Effects of brevetoxin-B on motor nerve terminals of mouse skeletal mouse. Br. J. Póarnwc . 103, 1126-1128. Tsee, M. C., Cxou, H. N. and Car, M. L. (1991) Effect of br+evetoxin-H on the neuromuscular transmission of the mouse diaphragm. J. Formarmi mrd. dswc. 90, 431-436. Wrs~ermtu, M., Mooaea, J. W., ICnr, Y. S. and Peutzt .e, G. M. (1977) How Gynueoa3beGon bmr red tide toxins) produoea repetitive firing in squid axons. .bn. J. PFystol. Z32, C23-C29.