Anesthesiology 2005; 102:353–63 © 2005 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Long-term Exposure to Local but Not Inhalation Anesthetics Affects Neurite Regeneration and Formation between Identified Lymnaea Shin Onizuka, M.D.,* Mayumi Takasaki, M.D., Ph.D.,† Naweed I. Syed, Ph.D.‡

Background: General and local anesthetics are used in vari- MOST surgical procedures require either general or local ous combinations during surgical procedures to repair dam- anesthetic treatments, which last from a few minutes to aged tissues and organs, which in almost all instances involve several hours. Although surgical interventions necessi- functions. Because synaptic transmission re- tate the use of various anesthetic agents, their choices covers rapidly from various inhalation anesthetics, it is gener- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 ally assumed that their effects on nerve regeneration and syn- should be based on the ones with the least deleterious apse formation that precede injury or surgery may not be as effects on neuronal function, nerve regeneration, and detrimental as that of their local counterparts. However, a di- synaptic repair. This information is important in light of rect comparison of most commonly used inhalation (sevoflu- better choices vis-à-vis various agents that are available to rane, isoflurane) and local anesthetics (lidocaine, bupivacaine), vis-à-vis their effects on synapse transmission, neurite regener- date and are used extensively in clinical practices. To ation, and synapse formation has not yet been performed. render such choices feasible, comparable data for the Methods: In this study, using cell culture, electrophysiologic long-term effects on synaptic transmission, regeneration, and imaging techniques on unequivocally identified presynap- and synapse formation are needed; however, with a few tic and postsynaptic neurons from the mollusc Lymnaea, the notable exceptions1–3 for both general and local anes- authors provided a comparative account of the effects of both general and local anesthetics on synaptic transmission, nerve thetics agents, no such data are currently available. regeneration, and synapse formation between cultured neurons. Notwithstanding the fact that various inhalation (such Results: The data show that clinically used concentrations of as sevoflurane, halothane, and isoflurane), intravenous both inhalation and local anesthetics affect synaptic transmis- (propofol, thiopental, and ketamine), and local anes- sion in a concentration-dependent and reversal manner. The thetic agents (such as lidocaine and bupivacaine) per- authors provided the first direct evidence that long-term over- turb nervous system functions by disrupting either syn- night treatment of cultured neurons with sevoflurane and isoflurane does not affect neurite regeneration, whereas both aptic transmission or nerve conductions, their precise lidocaine and bupivacaine suppress neurite outgrowth com- modes of actions vary considerably from preparation to pletely. The soma–soma synapse model was then used to com- preparation.4 Inhalation anesthetics, for example, affect pare the effects of both types of agents on synapse formation. synaptic transmission by either blocking presynaptic The authors found that local but not inhalation anesthetics transmitter release5 or suppressing postsynaptic recep- drastically reduced the incidence of synapse formation. The 6–9 10–12 local anesthetic–induced prevention of synapse formation most tor function at both excitatory and inhibitory 13,14 likely involved the failure of presynaptic machinery, which , and these responses may invoke a variety otherwise developed normally in the presence of both sevoflu- of ion channels and second messengers.13,15–18 Similarly, rane and isoflurane. intravenous anesthetics such as propofol bring about Conclusion: This study thus provides the first comparative, synaptic depression by enhancing either the function of albeit preclinical, account of the effects of both general and ␥ 8,12,19–21 local anesthetics on synaptic transmission, nerve regeneration, -aminobutyric acid receptors or perhaps by sup- 5 and synapse formation and demonstrates that clinically used pressing presynaptic glutamate release. Despite our cur- lidocaine and bupivacaine have drastic long-term effects on rent understanding of the cellular basis of anesthesia, the neurite regeneration and synapse formation as compared with precise mechanisms by which both general and intrave- sevoflurane and isoflurane. nous agents perturb nervous system function remain largely unknown. This limited understanding of how anes- thetics affect neuronal communications in the nervous sys- tem stems from anatomical challenges that are often met in * Research Associate, Calgary Brain Institute, Faculty of Medicine, University of most mammalian preparations where direct cell–cell inter- Calgary, and Department of Anesthesiology, Miyazaki Medical College, University of Miyazaki. † Professor, Department of Anesthesiology, Miyazaki Medical Col- actions are difficult to study unequivocally. Moreover, as lege, University of Miyazaki. ‡ Professor, Calgary Brain Institute, Faculty of Medicine, University of Calgary. compared with their intravenous and inhalation counter- Received from the † Calgary Brain Institute, Faculty of Medicine, University of parts, even less is known about the mechanisms by which Calgary, Calgary, Alberta, Canada, and the * Department of Anesthesiology, local agents affect synaptic transmission during pain, sur- Miyazaki Medical College, University of Miyazaki, Miyazaki, Japan. Submitted for publication July 20, 2004. Accepted for publication October 15, 2004. Supported gery, and functional recovery. by the Canadian Institutes for Health Research, Ottawa, Ontario, Canada, and In addition to some undesired side effects of all acutely Alberta Heritage Foundation for Medical Research, Edmonton, Alberta, Canada. applied anesthetics,22 long-term treatments of neuronal Address reprint requests to Dr. Syed: Department of Cell Biology and Anatomy, 23 24,25 Faculty of Medicine, University of Calgary, 3330 Hospital Drive Northwest, Calgary, tissue with both inhalation and local agents cause Alberta T2N 4N1, Canada. Address electronic mail to: [email protected]. Indi- widespread learning defects and degeneration. Whether vidual article reprints may be purchased through the Journal Web site, www.anesthesiology.org. such long-term (hours to days) exposure of the neuronal

Anesthesiology, V 102, No 2, Feb 2005 353 354 ONIZUKA ET AL. tissue to these anesthetics also affects nerve regenera- at the University of Calgary Animal Resource Centre, tion and synapse formation has not yet been determined. Calgary, Alberta, Canada.) In this study, we took advantage of an ideal model preparation in which synaptic transmission between Cell Culture uniquely identified neurons can be investigated at the Cells were isolated individually and cultured as de- level of single presynaptic and postsynaptic neurons. scribed previously.34 Briefly, snails were anesthetized Individually isolated neurons from the mollusc Lymnaea with 10% Listerine (21.9% ethanol, 0.042% menthol; not only regenerate their neurites in cell culture but also Pfizer Inc., New York, NY) solution in normal Lymnaea recapitulate their specific patterns of synapses, which saline (containing 51.3 mM NaCl, 1.7 mM KCl, 4.1 mM are similar to those seen in vivo. This in vitro approach CaCl2, and 1.5 mM MgCl2) buffered to pH 7.9 with

using Lymnaea neurons has been used extensively to Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 HEPES. The central ring ganglia were removed under decipher both cellular and synaptic mechanisms by sterile conditions and washed with antibiotic saline which various anesthetics affect neuronal function and (40 ␮g/ml gentamycin; two washes, 10 min each). The synaptic transmission.13,26–34 ganglia were then treated with 0.2% trypsin (Sigma type Here, we sought to determine and compare how clin- III; Sigma Chemical Company, St. Louis, MO) for 22 min ically relevant concentrations of inhalation (sevoflurane followed by 0.2% soybean trypsin inhibitor (Sigma type and isoflurane) and local anesthetics (lidocaine and bu- 1-S; Sigma Chemical Company) for 10 min, both in de- pivacaine) affect synaptic transmission and whether fined medium (DM). DM consisted of serum-free 50% these actions involve presynaptic mechanisms, postsyn- L-15 medium with added inorganic salts (40 mM NaCl, aptic mechanisms, or both. Moreover, we provide the 1.7 mM KCl, 4.1 mM CaCl , 1.5 mM MgCl , and 10 mM first direct and comparative account of how inhalation 2 2 HEPES). The pH was adjusted to 7.9 with 1 N NaOH, and and local anesthetics affect neurite regeneration and 20 ␮M gentamycin was added. The enzymatically treated synapse formation. Specifically, our data show that both ganglia were pinned to the bottom of a dissection dish sevoflurane/isoflurane and lidocaine/bupivacaine block that contained 8 ml high-osmolarity DM (DM with cholinergic synaptic transmission between the paired 37.5 mM glucose). presynaptic and postsynaptic neurons. Long-term (over- The cells were isolated by applying gentle suction to a night) sevoflurane/isoflurane treatment of the cultured fire-polished and Sigmacote (Sigma Chemical Company)– neurons did not affect neurite outgrowth, whereas the treated pipette. The isolated cells were plated on poly- cultured cells did not exhibit regeneration in the pres- L-lysine pretreated coverslips glued to the bottom of a ence of both lidocaine and bupivacaine. Because cells 35-mm Falcon dish34 in the presence of CM. The CM was did not extend neurites in the presence of local anes- prepared by incubating central ganglia in DM (12 gan- thetics, we thus adopted the soma–soma model (syn- glia/6 ml DM) for 4 or 5 days and frozen until used. apses develop between the cell bodies in the absence of Isolated somata of identified neurons were either al- neurites) to ask the question whether lidocaine and lowed to extend neurites or were juxtaposed in a soma– bupivacaine also affect synapse formation between the soma configuration and left undisturbed overnight.35 paired cells. Although both anesthetics reduced the in- cidence of synapse formation between the paired cells, this synaptogenesis was severely compromised by lido- Neurite Outgrowth caine and bupivacaine. Taken together, our data provide To assess neuronal regeneration in either the absence the first direct evidence that long-term treatment of or the presence of each anesthetic, identified neurons neurons with local but not inhalation anesthetics is se- were isolated in cell culture and plated on poly-L-lysine– verely detrimental for neurite outgrowth and synapse coated dishes containing CM plus the anesthetics. Pre- formation. synaptic (visceral dorsal 4 [VD4]) and postsynaptic (left pedal dorsal 1 [LPeD1]) neurons were selected for neu- rite outgrowth and synapse formation assays. To test for Materials and Methods the effects of each anesthetic agent on neurite out- Animals growth and synapse formation, the anesthetic was added Laboratory-raised stocks of the fresh water pond snail to the dish 1 h after cell plating, and the neurons were Lymnaea stagnalis were maintained at room tempera- maintained overnight in the dark. Neuronal sprouting ture (18–20°C) in an aquarium containing well-aerated was assessed as described previously.36 Specifically, only and dechlorinated tap water and were fed lettuce. Ani- those neurons exhibiting outgrowth (multiple branches, mals aged 1–2 months (shell length, 10–15 mm) were active growth cones, and so forth) equivalent to five used for cell isolation while the brain-conditioned me- somata diameter were considered as sprouted. The ex- dium (CM) was prepared from 3- to 4-month-old animals tent of outgrowth was calculated as a function of maxi- (shell length, 15–25 mm). (Animal Care Certification is mum neurite length. Neurons cultured overnight in CM not required for invertebrate species such as L. stagnalis alone served as controls.

Anesthesiology, V 102, No 2, Feb 2005 ANESTHETICS AFFECT REGENERATION, SYNAPSE FORMATION 355

Neurite–Neurite Synapses trophysiologic techniques to facilitate the uptake of To prepare neurite–neurite synapses, the isolated cells FM1-43 in cells that were paired overnight in either the were plated in CM a few soma diameters apart and were presence or the absence of each anesthetic. The styryl allowed to extend neurites. After 24–48 h, the cells dye and anesthetics were then replaced with cold saline exhibited extensive outgrowth with multiple overlap- to prevent neuronal firing during the washout and to ping neurites. The cells were cultured overnight either remove background fluorescence. Fluorescent images of in CM alone or CM plus the anesthetic, which was the FM1-43–labeled cells were acquired using a Zeiss subsequently washed out during electrophysiologic re- (Carl Zeiss Canada Ltd.) Axiovert 200M inverted micro- cordings. To demonstrate the reversibility of anesthetic- scope. Excitation light was from a 100-W Hg lamp, ex- induced effects, the compound was washed away with citation filters (490/30 nm), a dichroic mirror (505 nm),

normal saline for more than an hour before the intracel- and emission filters (570 low pass nm or 610 nm). Phase Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 lular recordings. and fluorescent images were captured with a Photomet- rics (Tuscon, AZ) Sensys 1400 camera (1- to 100-ms Electrophysiologic Recording exposure), connected to a computer running Axiovision Conventional intracellular recording techniques were 3.0 for Windows (Carl Zeiss Canada Ltd.). The pixels in used.34 Specifically, glass microelectrodes (1.5 mm ID, cross-section of the fixed area (200 ϫ 150 ␮m2) were with filament; World Precision Instruments, Sarasota, FL) integrated and measured with NIH image software (ver- were pulled on a vertical electrode puller (Kopf, 700C; sion 1.62; National Institutes of Health, Bethesda, MD). David Kopf Instruments, Tujunga, CA) and filled with a ⍀ saturated solution of K2SO4 (resistance, 30–60 M ). Statistical Analysis Isolated cells were viewed under a Zeiss (Telaval 31; Carl Parametric data are expressed as mean Ϯ SE and were Zeiss Canada Ltd., North York, Ontario, Canada) in- analyzed for significance using one-way analyses of vari- verted microscope and impaled using Narishigi micro- ance with repeated measures (anesthetic concentration manipulators (model MO-103; Narishigi Instruments, To- was the between-subjects factor) and a Tukey post hoc kyo, Japan). The intracellular signals were amplified via test. Nonparametric data are expressed as percent and a preamplifier (Neurodata model IR-283; Cygnus Tech- were analyzed for significance using the chi-square test. nology Inc., Delaware, PA), displayed on a storage oscil- Significance was assumed if P was less than 0.05. loscope (Tektronix R5103N; Tektronix, Montreal, Que- bec, Canada), and recorded and stored on a computer using Digidata software ( Instruments, Union City, Results CA). All experiments were performed at room tempera- ture (18°–22°C). Specific Synapse Formation between Identified Neurons in Cell Culture Drugs A well-established model of identified Lymnaea neu- Sevoflurane (Sevofrane; Maruishi, Osaka, Japan) and rons37 was used to reconstruct synapses in either a neurite– isoflurane (Forane; Abbott, Abbott Park, IL) were se- neurite33,37 or a soma–soma configuration.35,37–40 Spe- lected as inhalation anesthetics, whereas lidocaine (Xy- cifically, identified presynaptic VD4 and its locaine; AstraZeneca, London, United Kingdom) and bu- postsynaptic partner LPeD1 were isolated from the adult pivacaine (Anapin; AstraZeneca) were diluted into CM animals and cultured overnight in the presence of CM. and served as local agents. VD4 and LPeD1 synapse pairs Within 12–24 h, neurons (n ϭ 20) exhibited extensive were exposed to sevoflurane and isoflurane (bubbled outgrowth (fig. 1A). Simultaneous intracellular record- overnight) at concentrations of 0.2, 2, 10, 20, and ings revealed that current injection in VD4 (at arrow, 100 mM, and lidocaine and bupivacaine were used at fig. 1B) generated 1:1 excitatory postsynaptic potentials 38 concentrations of 0.001, 0.005, 0.01, 0.1, and 1 mM. (EPSPs) in a manner similar to those observed in vivo. Exogenous application of acetylcholine (Sigma Chemical Similarly, when paired in a soma–soma configuration Company), was performed (80-ms pulses, 1–2 psi) in (n ϭ 32; fig. 1C) in CM, synapses (90–100%) similar to some experiments using 1 ␮M acetylcholine applied di- those observed in a neurite–neurite mode also devel- rectly to the synaptic site via a Pneumatic PicoPump oped (80–100%), albeit in the absence of neurite out- (PV800; World Precision Instruments) pressure injector. growth.41 Specifically, action potentials in VD4 (at ar- row, fig. 1D) resulted in 1:1 EPSPs in LPeD1. This FM1-43 Imaging synaptic transmission has previously been shown to be 38,39 Cells were incubated in 20 ␮M FM1-43 (Molecular cholinergic. It is important to note that the efficacy Probes, Eugene, OR) for 10 min before the addition of of synaptic transmission (consistency of EPSP amplitude) 0.01 mM lidocaine or 5 mM sevoflurane to the bath. The between VD4 and LPeD1 was more consistent in our presynaptic cell (VD4) was stimulated to generate 100 soma–soma model as compared with the neurite–neurite action potentials (10 spikes/burst) by conventional elec- configuration. Therefore, we opted to use the soma–

Anesthesiology, V 102, No 2, Feb 2005 356 ONIZUKA ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021

Fig. 1. Specific synapses reform between Lymnaea neurons in a neurite–neurite and soma–soma configuration. Identified pre- synaptic (visceral dorsal 4 [VD4]) and postsynaptic (left pedal dorsal 1 [LPeD1]) neurons were isolated from visceral and left pedal ganglia, respectively. (A) Neurons were plated in close proximity and allowed to extend neurites in conditioned me- Fig. 2. Inhalation anesthetics suppress cholinergic synaptic dium. Extensive outgrowth occurred within 24–48 h, and neu- transmission between the soma–soma paired cells. (A) Intracel- rites overlapped. (B) Simultaneous intracellular recordings lular recordings between visceral dorsal 4 (VD4) and left pedal were made from both cells. Induced action potentials in VD4 (at dorsal 1 (LPeD1) revealed an excitatory synapse in conditioned arrows) generated 1:1 excitatory postsynaptic potentials (EP- medium, and action potentials in the presynaptic cell generated C) Neurons were isolated individually 1:1 excitatory postsynaptic potentials (EPSPs). These excitatory) .(20 ؍ SPs) in LPeD1 (n and paired in conditioned medium in a soma–soma configura- responses were mimicked by exogenously applied acetylcho- tion. Although neurons did not exhibit neurite outgrowth, the line (ACh) (at arrow), which generated a compound postsyn- induced actions potentials in VD4 (D) generated 1:1 EPSPs in aptic potential. Both synaptic (VD4) and nonsynaptic (ACh) which were similar to those seen between responses were significantly reduced by 5 mM sevoflurane ,(32 ؍ LPeD1 (n The sevoflurane- and isoflurane (iso)-induced .(11 ؍ neurite–neurite pairs. (sevo; n depression was concentration dependent, and the synaptic transmission returned to its baseline within minutes of wash- soma preparation to test for the effects of anesthetics on out (Wash) with normal saline (B). * Data were statistically dif- synaptic transmission. Similar results, however, were ferent from the control value. obtained from the neurite–neurite preparation. isoflurane (control, 13.9 Ϯ 3.6 mV; 0.3 nM, 7.1 Ϯ 2.6 mV; Both Inhalation and Local Anesthetics Block 5mM, 2.3 Ϯ 0.9 mV [n ϭ 10 for each]) (fig. 2B). Excitatory Synaptic Transmission between VD4 The above data thus show that both sevoflurane and and LPeD1 isoflurane depress synaptic transmission between VD4 To test and compare the effects of inhalation (sevoflu- and LPeD1 and also the postsynaptic cholinergic respon- rane and isoflurane) and local anesthetics (lidocaine and siveness in LPeD1. These effects were concentration- bupivacaine) on cholinergic synaptic transmission be- dependent, and the synaptic transmission recovered tween VD4 and LPeD1, the soma–soma paired cells were within minutes of washout with normal saline (fig. 2B). recorded in either the absence or the presence of vari- Next, we sought to determine the effects of local ous anesthetics. Synapses were tested electrophysiologi- anesthetics lidocaine and bupivacaine on cholinergic cally. Action potentials in VD4 under control saline con- synaptic transmission between VD4 and LPeD1. Syn- ditions (fig. 2A) generated 1:1 EPSPs in LPeD1, and these apses were reconstructed as described above, and both excitatory responses were mimicked by exogenous ace- the synaptic transmission and the cholinergic responses tylcholine (at arrow), which was pressure-applied di- were tested in either the absence or the presence of rectly at the synapse under a fast perfusion system.42 The lidocaine (fig. 3A). Action potentials in VD4 induced 1:1 perfusion solution was then switched to the saline con- EPSPs in LPeD1, and these responses were mimicked by taining 5 mM sevoflurane, and the synaptic transmission exogenous acetylcholine (applied at arrow). Although was tested again. We found that both the synaptic trans- lidocaine significantly reduced the amplitude of VD4- mission and the cholinergic responses were reduced induced EPSPs in LPeD1 (EPSP amplitude: control, significantly (control, 11.9 Ϯ 0.7 mV; 0.3 mM, 6.4 Ϯ 11.4 Ϯ 2.6 mV [n ϭ 10]; 0.1 mM lidocaine, 5.6 Ϯ 1.5 mV 1.1 mV; 5 mM, 2.2 Ϯ 0.3 mV [n ϭ 11 for each case]) in [n ϭ 8];1mM lidocaine, 1.9 Ϯ 0.8 mV [n ϭ 9]) (fig. 3A), a concentration-dependent and reversible manner (fig. the cholinergic response remained unperturbed as com- 2B). Similar results were obtained in the presence of pared with sevoflurane (acetylcholine amplitude: con-

Anesthesiology, V 102, No 2, Feb 2005 ANESTHETICS AFFECT REGENERATION, SYNAPSE FORMATION 357 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021

Fig. 3. Local anesthetics depress synaptic transmission between visceral dorsal 4 (VD4) and left pedal dorsal 1 (LPeD1). (A) Soma–soma synapses were recorded in either the absence or Fig. 4. Long-term local but not inhalation anesthetic treatment the presence of lidocaine (lido; 3% mM). In the presence of suppresses neurite outgrowth from cultured neurons. To test lidocaine, synaptic (VD4-induced) but not cholinergic (at ar- for the effects of both general and local anesthetics on neurite row) responses were significantly depressed. Similar results regeneration, cells were isolated and cultured in conditioned were obtained in the presence of bupivacaine (bupi). The local medium in either the absence (control) or the presence of anesthetic–induced effects were concentration dependent and various agents. Neurons cultured overnight in sevoflurane reversible (B). # Comparative data sets; * data were statistically (5 mM) and isoflurane (5 mM) extended neurites, which were indistinguishable from that of control. However, both low ؍ acetylcholine; EPSP ؍ different from the control value. ACh ؍ excitatory postsynaptic potential; Wash washout. (0.01 mM) and higher concentrations (0.1 mM) of lidocaine and bupivacaine suppressed neurite outgrowth completely and ir- trol, 19.7 Ϯ 2.8 mV [n ϭ 10]; 0.1 mM lidocaine, 15.4 Ϯ reversibly. Specifically, even though neuronal somata seemed healthy, they did not exhibit neurite outgrowth. 5.4 mV [n ϭ 8];1mM lidocaine, 11.8 Ϯ 4.5 mV [n ϭ 9]) (fig. 3B). Similar results were obtained with bupivacaine, and these effects were both concentration-dependent be detected electrophysiologically. Under normal cul- and reversible (EPSP amplitude: control, 12.1 Ϯ 2.6 mV; ture conditions, both cells extended neurites (fig. 4), and 0.1 mM bupivacaine, 5.0 Ϯ 1.1 mV; 1 mM bupivacaine, when tested electrophysiologically, synapses were de- 1.6 Ϯ 0.5 mV; acetylcholine amplitude: control, 19.7 Ϯ tected as described previously (fig. 1; not shown here). 2.8 mV; 0.1 mM bupivacaine, 14.5 Ϯ 5.0 mV; 1 mM Similarly, cells cultured in 5 mM sevoflurane (bubbled bupivacaine, 11.8 Ϯ 4.5 mV [n ϭ 10 for each]) (fig. 3B). and superfused overnight) exhibited extensive out- These data thus demonstrate that the local anesthetics growth, which was indistinguishable from control (fig. depress cholinergic synaptic transmission between VD4 4). To demonstrate that the inhalation anesthetics had and LPeD1 and that these effects most likely involve the not evaporated or catalyzed overnight, the synaptic presynaptic mechanisms. transmission was first tested electrophysiologically in the presence of the incubating medium (CM plus sevoflu- Local but Not Inhalation Anesthetics Suppress rane) before washout with normal saline. Under these Neurite Outgrowth from Cultured Neurons experimental conditions, action potentials in VD4 did To test the hypothesis that both inhalation and local not generate any response in the postsynaptic cell (n ϭ anesthetics affect neurite regeneration and synapse for- 11; data not shown). However, within a few minutes mation from isolated Lymnaea neurons, cells were cul- after washout with normal saline, synapses similar to tured in CM in either the absence or the presence of any those observed under control conditions were detected given anesthetic agent and were allowed to extend neu- (data not shown). In contrast to the general anesthetics rites. Specifically, cells were paired in close proximity (control, 1.4 Ϯ 0.4 mm [n ϭ 12]; 2 mM sevoflurane, and were allowed to extend neurites. We reasoned that 1.4 Ϯ 0.5 mm [n ϭ 10]; 5 mM sevoflurane, 1.2 Ϯ 0.4 mm if neurons regenerated their processes, the synapses [n ϭ 11]), the cultured neurons did not extend neurites would develop between the neurites, which could then in the presence of lidocaine or bupivacaine (0.01 mM

Anesthesiology, V 102, No 2, Feb 2005 358 ONIZUKA ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021

Fig. 5. Summary data showing the extent of neurite outgrowth in either the absence or the presence of anesthetics. Local anesthetics retarded neurite outgrowth extensively, whereas the growth exhibited in the presence of sevoflurane (sevo) and isoflurane (iso) was indistinguishable from that of control. * Data -bupiv ؍ were statistically different from the control value. bupi .lidocaine ؍ acaine; lido lidocaine, 0.4 Ϯ 0.3 mm [n ϭ 10]; 0.1 mM lidocaine, 0.2 Ϯ 0.3 mm [n ϭ 10]; 0.01 mM bupivacaine, 0.3 Ϯ 0.3 mm [n ϭ 10]; 0.1 mM bupivacaine, 0.1 Ϯ 0.2 mm ϭ Fig. 6. Incidence of synapse formation between neurons paired [n 11]) (fig. 5). under various clinically used anesthetic concentrations. Sum- These data thus demonstrate that both lidocaine and mary data showing the incidence of synapse formation between bupivacaine significantly retard neurite outgrowth from cells paired either in conditioned medium (CM) alone or con- ditioned medium plus the anesthetics. In sevoflurane and cultured neurons, whereas in the presence of their in- isoflurane, the incidence of pairs exhibiting normal synapses halation counterparts, both neurite processes and syn- was similar to that of conditioned medium, whereas both lido- apses develop normally (fig. 5). Because cells paired in caine and bupivacaine used within a clinical range completely prevented synapse formation between visceral dorsal 4 and left local anesthetics did not extend neurites (and thus the pedal dorsal 1. synapse formation), their effects on synapse formation could not be tested and compared with sevoflurane and isoflurane, 95% [n ϭ 15]). Cells paired overnight in the isoflurane. To test their effects on synaptogenesis di- presence of both local anesthetics, however, did not rectly, we therefore resorted to our soma–soma model develop synapses within the clinical range, even though where synapses could develop in the absence of neurite morphologically they seemed healthy and exhibited nor- outgrowth. mal intrinsic membrane properties (resting membrane parameters and others) (1 mM lidocaine, Local but Not Inhalation Anesthetics Retard 0% [n ϭ 10]; 0.1 mM lidocaine, 0% [n ϭ 10]; 0.01 mM Synapse Formation between Soma–Soma Pairs lidocaine, 6% [n ϭ 11]; 0.005 mM lidocaine, 15% [n ϭ To test and compare the effects of both general and 11]; 0.001 mM lidocaine, 0% [n ϭ 15]; 1 mM bupivacaine, local anesthetics on synapse formation, cells were soma– 0% [n ϭ 10]; 0.1 mM bupivacaine, 3% [n ϭ 9]; 0.01 mM soma paired in CM overnight in either the absence or the bupivacaine, 11% [n ϭ 9]; 0.005 mM bupivacaine, 21% presence of various agents. The incidence of synapse [n ϭ 9]; 0.001 mM bupivacaine, 50% [n ϭ 12]) (fig. 6). It formation between the paired cells was tested electro- is interesting to note that in instances in which cells did physiologically. Under normal CM conditions, the inci- form synapses in lower concentrations of both local dence of synapse formation ranged between 90 and agents, the synaptic transmission between these pairs 100% (fig. 6). Similarly, neurons paired within the clini- failed rapidly and after 5–10 min of stimulation; no syn- cal concentration range of sevoflurane/isoflurane also aptic transmission was detectable. After several minutes formed normal synapses, although the percent of cells of rest, however, the synaptic transmission resumed, forming excitatory synapses decreased with increasing although the amplitude of postsynaptic potential was anesthetic concentrations (10 mM sevoflurane, 70% [n ϭ significantly reduced (data not shown). 30; P Ͻ 0.05]; 2 mM sevoflurane, 90% [n ϭ 18]; 0.3 mM Figure 7 shows examples of representative traces from sevoflurane, 100% [n ϭ 12]; 10 mM isoflurane, 65% [n ϭ cells cultured under control conditions and in the pres- 20]; 2 mM isoflurane, 80% [n ϭ 20; P Ͻ 0.05]; 0.3 mM ence of sevoflurane, isoflurane, lidocaine, and bupiva-

Anesthesiology, V 102, No 2, Feb 2005 ANESTHETICS AFFECT REGENERATION, SYNAPSE FORMATION 359 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021

Fig. 8. Summary data depicting the amplitude of excitatory postsynaptic potentials (EPSPs) and cholinergic responses re- corded from cells paired in various anesthetics. Visceral dorsal 4 paired with left pedal dorsal 1 in lidocaine and bupivacaine did not generate EPSPs in its postsynaptic partner, whereas cholinergic responses were similar to those of control. Normal synapses developed in sevoflurane and isoflurane, and the am- plitude of both synaptic and cholinergic responses were iden- tical to those of control. # Data sets compared; * data were -acetylcho ؍ statistically different from the control value. ACh ;lidocaine ؍ isoflurane; lido ؍ bupivacaine; iso ؍ line; bupi .sevoflurane ؍ sevo Fig. 7. Local anesthetics block synapse formation between vis- 0.8 Ϯ 0.3 mV [n ϭ 9]) (fig. 8), their cholinergic re- ceral dorsal 4 (VD4) and left pedal dorsal 1 (LPeD1). Cells were paired overnight in either the absence or the presence of vari- sponses in LPeD1 were similar to those of control and ous anesthetics. Normal synapses were detected in conditioned the inhalation anesthetics (acetylcholine amplitude: medium (control), and the postsynaptic cell exhibited a depo- 0.01 mM lidocaine, 16.5 Ϯ 4.8 mV [n ϭ 11]; 0.01 mM larizing response to exogenously applied acetylcholine (ACh) Ϯ ϭ (at arrow). Similarly, both synaptic and cholinergic responses bupivacaine, 19.6 2mV[n 9]) (fig. 8). These data recorded from cells paired in sevoflurane and isoflurane were lend further support to the idea that neuronal failure to identical to those of control. Cells paired in lidocaine and bu- develop synapses in the presence of local anesthetic is pivacaine (0.01 mM) exhibited significantly reduced incidence of electrophysiologically detectable synapses, whereas their not due to its toxic effects on neuronal excitability or the cholinergic responses were normal. desensitization of the postsynaptic receptors. We next sought to determine whether the absence of caine. Synapses similar to those observed under control synapses between VD4 and LPeD1 cultured in local conditions developed in sevoflurane and isoflurane. anesthetics could be due to the fact that the presynaptic Moreover, the cells treated with general anesthetic ex- machinery may not have developed in their presence. hibited responses similar to those of cells with exog- Specifically, cells were paired overnight in either the enously applied acetylcholine, as was seen under control absence or the presence of anesthetics (fig. 9). On day 2, conditions (fig. 7). Conversely, cells paired in both lido- the anesthetic solution was replaced with normal saline caine (0.01 mM) and bupivacaine (0.01 mM) did not containing the dye FM1-43, and the presynaptic cell was develop synapses, whereas their responses to exog- stimulated electrically. We reasoned that if cells pos- enously applied acetylcholine were similar to those of sessed normal, functional presynaptic machinery (exo- control and both sevoflurane and isoflurane. These data cytosis and endocytosis), we should expect labeling of are summarized in figure 8 (EPSP amplitude: control, the presynaptic cell at its contact with the postsynaptic 10.3 Ϯ 3.4 mV [n ϭ 38]; 5 mM sevoflurane, 8.8 Ϯ 1.7 mV neuron. To test this possibility, VD4 was stimulated to [n ϭ 18]; 5 mM isoflurane, 8.2 Ϯ 1.7 mV [n ϭ 20]). fire action potentials via current injections (100 actions Similarly, the amplitude of acetylcholine-induced depo- potentials) in the presence of the dye FM1-43. The dye larizing responses in LPeD1 observed in sevoflurane and was then washed away with normal saline, and images isoflurane was similar to that observed under control were acquired. We found that under both normal (figs. conditions (acetylcholine amplitude: control, 19.8 Ϯ 9A–C) and sevoflurane (figs. 9D–F) conditions, exclusive 3.4 mV [n ϭ 38]; 5 mM sevoflurane, 14.7 Ϯ 5.1 mV [n ϭ labeling of the presynaptic cells was discernible at its 18]; 5 mM isoflurane, 13.7 Ϯ 4.9 mV [n ϭ 20]) (fig. 8). contact site with LPeD1 and also in its processes sur- Interestingly, although the incidence of synapse forma- rounding the postsynaptic somata (fig. 9B, control, and tion between cells paired in local anesthetics was signif- fig. 9E, sevoflurane). In contrast, faint or no staining of icantly reduced (P Ͻ 0.05) (EPSP amplitude: 0.01 mM the presynaptic cell was observed in VD4 paired in the lidocaine, 0.5 Ϯ 0.4 mV [n ϭ 11]; 0.01 mM bupivacaine, presence of lidocaine (figs. 9G and H). The pixel values

Anesthesiology, V 102, No 2, Feb 2005 360 ONIZUKA ET AL.

and reversible manner, long-term lidocaine and bupiva- caine but not sevoflurane and isoflurane treatment re- tards neurite outgrowth and synapse formation. The neuronal failure to exhibit outgrowth and synapse for- mation in the presence of local agents, however, cannot be attributed to their toxic effects on intrinsic membrane properties or postsynaptic cholinergic receptors; rather, the presynaptic secretory machinery seems to have been perturbed. The state of anesthesia almost exclusively involves

either the suppression of excitatory or an enhancement Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 of inhibitory synaptic transmission between any given pair or sets of functionally related neurons. Therefore, an imbalance between excitatory and inhibitory synaptic transmission is invoked by various agents, used in a variety of combinations, to create desired results for pain management. However, despite their extensive use in surgical procedures and pain management, our under- standing is limited vis-à-vis the modes and sites of action of various anesthetic agents. Notwithstanding our recent Fig. 9. Long-term lidocaine treatment perturbs the development progress toward elucidating how inhalation and intrave- of presynaptic machinery. To test whether the transmitter se- nous anesthetics affect neuronal communications4 in the cretory machinery of the presynaptic cell paired overnight in lidocaine was perturbed, the exocytotic and endocytotic pro- brain and achieve desired effects, much less is known files were analyzed through FM1-43 imaging. Visceral dorsal 4 about local agents (oral, injected, local) such as lido- (VD4) cells paired overnight under control (A and B), sevoflu- caine, which are multimodel and serve as antiarrhyth- rane (D and E), or lidocaine (G and H) but recorded under 43 normal saline conditions were loaded with the dye FM1-43. mics, antiepileptics, and anticonvulsants agents. Be- Specifically, VD4 was stimulated (100 action potentials) in the cause the core of all nervous system functions resides at presence of FM1-43, and the dye was washed away with normal the synaptic transmission between any given presynap- saline. Images were acquired under both control (B) and sevoflurane (E) conditions. VD4 was extensively loaded at ei- tic and postsynaptic neuron or a subset of such neurons, ther its contact site with left pedal dorsal 1 (LPeD1) or in its it is therefore pivotal to examine how anesthetics affect processes surrounding the postsynaptic somata (at arrow). The neuronal communication at the level of individual neu- corresponding panels under each image (C, F, I) represent the pixel values, and these values are averaged in J. * Data were rons. Such an approach may not be relevant for our statistically different from the control value. understanding of the state of unconsciousness that is experienced during anesthesia or how pain perception for each category are depicted in the corresponding is managed by the nervous system; nevertheless, it will panels (figs. 9C, F, and I), and the summary data for pixel provide basic understanding of how anesthetics function values are presented in figure 9J. at a fundamental level. With this caveat in mind, we took Taken together, our data demonstrate that although advantage of a simple model system approach and used both local and general anesthetics are effective in block- in vitro reconstructed synapses between functionally ing synaptic transmission between the paired cells, the well-defined neurons. VD4 and LPeD1 are members of 37 former is detrimental to nerve regeneration and synapse the cardiorespiratory network that comprises the re- formation and that these effects are likely due to the spiratory central pattern generator. The synaptic con- failure of the presynaptic assembly in the presence of nections between various central pattern–generating lidocaine. neurons are well characterized, and the entire network has been reconstituted in cell culture, where it generates the patterned rhythmic activity in a manner similar to 37 Discussion that observed in vivo. Therefore, adult Lymnaea neu- rons not only regenerate but also recapitulate their spe- This study provides the first direct comparative ac- cific connections in cell culture.44 Moreover, specific count of the effects of general and local anesthetics on synapses between the central pattern generator neurons synaptic transmission (short term), nerve regeneration, can also be reconstructed in a soma–soma configura- and synapse formation (long term). We have demon- tion.35,45 The soma–soma synapses are also target cell strated that although both general (sevoflurane and contact specific and require new protein synthesis and isoflurane) and local anesthetics used at clinically rele- gene transcription.35 Therefore, both morphologically vant concentrations affect synaptic transmission be- and electrophysiologically, the soma–soma synapses are tween identified neurons in a concentration-dependent similar to those seen in vivo.35,42 The soma–soma prep-

Anesthesiology, V 102, No 2, Feb 2005 ANESTHETICS AFFECT REGENERATION, SYNAPSE FORMATION 361 aration has also been used in a number of other species be attributed to “use-dependent” or “frequency-depen- to define the cellular mechanisms underlying synapse dent block” of Naϩ channel function.43 Additional ex- formation, synaptic transmission, and plasticity at a res- periments are required to test for the direct versus indi- olution not achievable elsewhere (Aplysia,46 Heli- rect actions of local anesthetics on various intrinsic soma,47 leech48,49; Nicholls50). neuronal properties. In addition, to deduce whether The usefulness of identified Lymnaea neurons for an- local anesthetics may also affect presynaptic Ca2ϩ chan- esthetic research has previously been validated by a nels, the soma–soma model offers a wonderful opportu- number of laboratories (halothane,28–30 sevoflurane,13 nity for such research. It is also noteworthy that synaptic and propofol33,51). In this study, we have compared the recovery from lidocaine and bupivacaine took much effects of both inhalation and local anesthetics on cho- longer (several tens of minutes) upon drug washout as

linergic synaptic transmission, which was suppressed compared with sevoflurane and isoflurane, and these Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 equally by these agents. The sevoflurane- and isoflurane- long-term effects may also be due to a slower removal of induced suppression most likely involved their actions these compounds from the target sites and/or the hydro- on postsynaptic receptors because cholinergic re- lysis resistance nature of these ester-type compounds. sponses were also significantly reduced by these anes- In addition to their short-term effects on synaptic trans- thetics. However, their effects on presynaptic Ca2ϩ mission, long-term exposure of the neuronal cells to a channels cannot be ruled out. Consistent with this no- variety of general anesthetics has been shown to cause tion are our earlier data where sevoflurane was shown to widespread degeneration and learning defects in devel- depress inhibitory synaptic transmission between soma– oping rats.23 Similarly, long-term treatment of the rabbit soma pairs, and these effects involved both presynaptic facial nerve with lidocaine and bupivacaine perturbed its and postsynaptic ion channels.13 Interestingly, although degeneration and regeneration,25 whereas in other stud- both lidocaine and bupivacaine depressed synaptic re- ies, these agents were reported to induce apoptosis in sponses, they did not significantly reduce the amplitude various cell types.24,53 Lidocaine has also been reported of the compound postsynaptic potentials in LPeD1 that previously to cause axonal degeneration in the posterior were elicited by acetylcholine. These results therefore root of rats.54 To add to the list of these toxic effects, the suggest that as compared with the inhalation anesthet- current study demonstrated that lidocaine and bupiva- ics, the synaptic depression induced by the local agents caine also completely suppress neurite outgrowth from may involve presynaptic mechanisms. cultured neurons and blocked synapse formation be- In contrast with their inhalation counterparts, which tween the soma–soma paired cells. Regarding the effects exert myriad effects on a variety of presynaptic and of local anesthetic on synapse formation, these data are postsynaptic ion channels,18 the local anesthetics pri- consistent with our previous study in which long-term marily affect Naϩ channels43 and block action potential propofol treatment of the soma–soma paired cells also propagation across the nerve. Their direct effects on blocked synapse formation between VD4 and LPeD1. either cell bodies or the synaptic sites, however, have However, it is interesting to note that the cells paired in not been fully investigated. Local anesthetics, such as propofol did reestablish their specific synapses after sev- procaine, suppressed nerve propagation in squid giant eral hours of anesthetic washout,33 whereas in the cur- axon by blocking action potentials,52 whereas in this rent study, synapses did not develop even after a day of study, we could successfully generate spikes in Lym- the anesthetic removal (5 ␮M propofol, 70% [n ϭ 10]; naea neurons, even in the presence of higher concen- 10 ␮M, 60% [n ϭ 9]; 25 ␮M, 25%, [n ϭ 8] 50 ␮M,0%[nϭ trations of lidocaine and bupivacaine. Interestingly, 6]).33 One may argue that a long-term treatment of the within a few seconds after anesthetic perfusion, both paired cells with local anesthetics may have rendered cells were depolarized (5–10 mV), and this observation the synaptic transmission undetectable. To rule out this is consistent with an earlier study in which lidocaine possibility, we performed the experiments in which increased intracellular Naϩ concentration in Lymnaea cells were first allowed to develop synapses in CM and neurons.51 It is also important to note that even though then the pair was exposed overnight to the local anes- we could trigger single spikes in VD4, compound action thetics. On day 3, the anesthetic was washed, and within potentials often did not occur (not shown) in response an hour of medium removal, normal synaptic transmis- to a sustained depolarizing pulse. A similar voltage-in- sion was detected (not shown). These results therefore duced inactivation of spikes has also been reported for argue in favor of the idea that local anesthetics block Lymnaea neurons.51 It is therefore plausible that local synapse formation between VD4 and LPeD1. Another anesthetics may not immediately affect Naϩ channels interesting difference between propofol33 and local an- required for a single action potential at rest (tonic esthetic was that in the presence of the former, the cells block43), but their depolarizing effects may shift the that normally form chemical synapses exhibited electri- steady state inactivation curve, thus reducing the avail- cal coupling, which is not observed in vivo between the ability of various Naϩ channels for spike generation. The cells.33 However, this electrical coupling was not ob- neuronal inability to generate repetitive spikes can also served in either lidocaine or bupivacaine, further under-

Anesthesiology, V 102, No 2, Feb 2005 362 ONIZUKA ET AL. scoring their effects on both chemical and electrical site or a disruption of other candidate molecules in the synapses. secretory pathway remains to be investigated. When comparing the effects of sevoflurane, isoflurane, In summary, notwithstanding the need to develop bet- and propofol with lidocaine/bupivacaine on neurite out- ter therapeutic drugs for long-term pain management, growth, both inhalation (sevoflurane/isoflurane, this analgesia, and anesthesia, the current preclinical study study) and intravenous (propofol33) anesthetics did not underscores the importance of a rigorous screening of affect neuronal sprouting, and cells extended processes such agents regarding their long-term effects on nerve that were indistinguishable from those of cells cultured regeneration and synapse formation. The ageing popula- in CM. This negative result could not be attributed to the tion and advanced surgical procedures necessitate the volatile nature of this compound because no synaptic use and development of long-lasting local anesthetics for transmission was observed between the paired cells be- both short- and long-term pain management. However, Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 fore the anesthetic washout with saline. However, neu- these strategies should not be at the expense of their rite outgrowth was completely retarded by both local deleterious affects on nerve regeneration and synapse anesthetics (lidocaine and bupivacaine). Consistent with formation, which are pivotal for functional recovery af- these data are earlier studies in which lidocaine was ter the loss of all nervous system function after surgery. found to induce the collapse of growth cones from On one hand, the data presented here provide a com- dorsal root ganglia neurons,2 and these effects may in- parison between various anesthetics for their effects on volve an intracellular enhancement of growth cone Ca2ϩ synaptic transmission, regeneration, and synapse forma- concentration in a manner similar to that observed after tion, and on the other hand, this model system provides 3 us with an unparalleled opportunity to identify various tetracaine treatment. On one hand, these studies dem- 58 onstrate a Naϩ channel–independent action of local an- genes and their products that are differentially regu- esthetics on neuronal tissues, and together with our lated in single cells after long-term anesthetic treatment. data, they provide unequivocal evidence that long-term The authors thank Wali Zaidi (Technician, Calgary Brain Institute, University of exposure of neurons to local but not general anesthetics Calgary, Calgary, Alberta, Canada) for excellent technical support. is detrimental to nerve regeneration. However, these effects could not be attributed to the local anesthetic– induced toxic influences on neuronal viability or a per- References turbation of their intrinsic membrane properties because 1. Anderson PL, Bamburg JR: Effects of local anesthetics on nerve growth in all neurons seemed healthy and electrophysiologically culture. Dev Neurosci 1981; 4:273–90 2. Radwan IA, Saito S, Goto F: Growth cone collapsing effect of lidocaine on viable (firing action potential and response to acetylcho- DRG neurons is partially reversed by several neurotrophic factors. ANESTHESIOLOGY line pulses). Therefore, as demonstrated previously, one 2002; 97:630–5 3. Saito S, Radwan IA, Nishikawa K, Obata H, Okamoto T, Kanno T, Goto F: of the likely mechanisms by which local agents may Intracellular calcium increases in growth cones exposed to tetracaine. Anesth prevent neurite outgrowth is to increase intracellular Analg 2004; 98:841–5 4. Antognini JF, Raines DE, Carstens EE: The future of anesthetic mechanisms concentration in the growth cones from a growth per- research, Neural Mechanisms of Anesthesia. Edited by Antognini JF, Raines DE, missive to a growth suppressive range.55–57 Whether the Carstens EE. Totowa, New Jersey, Humana Press, 2003, pp 449–53 5. Perouansky M, Hemmings HC Jr: Presynaptic actions of general anesthetics, local anesthetic–induced suppression observed here also Neural Mechanisms of Anesthesia. Edited by Antognini JF, Raines EE, Carstens EE. 2ϩ Totowa, New Jersey, Humana Press, 2003, pp 345–70 involved intracellular Ca remains to be determined. 6. Daniels S: General anesthetic effects on glycine receptors, Neural Mecha- We have previously demonstrated that voltage-induced nisms of Anesthesia. Edited by Antognini JF, Raines EE, Carstens EE. Totowa, New 2ϩ Jersey, Humana Press, 2003, pp 333–44 Ca hot spots develop between soma–soma paired cells 7. Forman SA, Flood P, Raines D: Nicotinic acetylcholine receptors and anes- and that these are both target cell and contact site thetics, Neural Mechanisms of Anesthesia. Edited by Antognini JF, Raines EE, specific.42 Consistent with these data are the studies that Carstens EE. Totowa, New Jersey, Humana Press, 2003, pp 283–98 8. Pearce RA: General anesthetic effects on GABAA receptors, Neural Mecha- showed that the presynaptic secretory machinery is spe- nisms of Anesthesia. Edited by Antognini JF, Raines EE, Carstens EE. Totowa, New Jersey, Humana Press, 2003, pp 265–82 cialized at the contact site between the cells. Specifi- 9. Perouansky M, Antognini JF: Glutamate receptors: Physiology and anes- cally, the dye FM1-43 (which labels vesicles during exo- thetic pharmacology, Neural Mechanisms of Anesthesia. Edited by Antognini JF, Raines EE, Carstens EE. Totowa, New Jersey, Humana Press, 2003, pp 319–32 cytosis and is subsequently internalized during 10. El-Beheiry H, Puil E: Anesthetic depression of excitatory synaptic trans- endocytosis) did not label presynaptic neurons that had mission in neocortex. Exp Brain Res 1989; 77:87–93 11. MacIver MB, Tauck DL, Kendig JJ: General anesthetic modification of not developed synapses after long-term propofol treat- synaptic facilitation and long-term potentiation in hippocampus. Br J Anesth ment.33 In the current study, we have demonstrated that 1989; 62:301–10 12. Wakasugi M, Hirota K, Roth SH, Ito Y: The effects of general anesthetics on cells treated chronically with local anesthetics also did excitatory and inhibitory synaptic transmission in area CA1 of the rat hippocam- not label with the dye, suggesting that in the presence of pus in vitro. Anesth Analg 1999; 88:676–80 13. Hamakawa T, Feng Z-P, Grigoriev N, Inoue T, Takasaki M, Roth S, Lukow- these anesthetics, the transmitter secretory machinery iak K, Hasan SU, Syed NI: Sevoflurane induced suppression of inhibitory synaptic responsible for exocytosis and endocytosis of cholin- transmission between soma-soma paired Lymnaea neurons. J Neurophysiol 1999; 82:2812–9 ergic vesicles does not assemble at the contact site be- 14. Pocock G, Richards CD: Cellular mechanisms in general anesthesia. Br J tween the paired cells Whether this failure is due to Anaesth 1991; 66:116–28 15. Bazil CW, Minneman KP: Anesthetic concentrations of enflurane and 2ϩ neuronal inability to target Ca channels at the synaptic methoxyflurane in a rat brain in vitro and in vivo. J Pharmacol 1989; 41:835–9

Anesthesiology, V 102, No 2, Feb 2005 ANESTHETICS AFFECT REGENERATION, SYNAPSE FORMATION 363

16. Dilger JP, Vidal AM, Mody HI, Liu Y: Evidence for direct actions of general 36. Ridgway RL, Syed NI, Lukowiak K, Bulloch AGM: Nerve growth factor anesthetics on an ion channel protein: A new look at a unified mechanisms of (NGF) induces sprouting of specific neurons of the snail, Lymnaea stagnalis. action. ANESTHESIOLOGY 1994; 81:431–43 J Neurobiol 1991; 22:377–90 17. Hirota K, and Roth SH: The effects of sevoflurane on population spikes in 37. Syed NI, Bulloch AGM, Lukowiak K: In vitro reconstruction of the respi- CA1 and dentate gyrus of the rat hippocampus in vitro. Anesth Analg 1997; ratory central pattern generator (CPG) of the mollusk Lymnaea. Science 1990; 85:426–30 250:282–5 18. Topf N, Recio-Pinto, E, Blanck TJJ, Hemmings HC Jr: Actions of general 38. Woodin MA, Munno DW, Syed NI: Trophic factor-induced excitatory anesthetic on voltage-gated ion channels, Neural Mechanisms of Anesthesia. synaptogenesis involves postsynaptic modulation of nicotinic acetylcholine re- Edited by Antognini JF, Raines DE, Carstens EE. Totowa, New Jersey, Humana ceptors. J Neurosci 2002; 22:505–14 Press, 2003, pp 299–318 39. Munno DW, Prince DJ, Syed NI: Synapse number and synaptic efficacy are 19. Bai D, Zhu G, Pennefather P, Jackson MF, MacDonald JF, Orser BA: Distinct regulated by presynaptic cAMP and protein kinase A. J Neurosci 2003; 23: functional and pharmacological properties of tonic and quantal inhibitory 4146–55 postsynaptic currents mediated by gamma-aminobutyric acid (A) receptors in 40. Smit AB, Syed NI, Schapp D, Klumperman J, Kits KS, Lodder H, van der hippocampal neurons. Mol Pharmacol 2001; 59:814–24 Schors RC, van Elk R, Sorgedrager B, Brejc K, Sixma T, Geraerts WPM: AChBP, a 20. Hirota K, Roth SH, Fujimura J Masuda A, Ito Y: GABAergic mechanisms in glia-derived acetylcholine-receptor-like modulation of cholinergic synaptic trans- the action of general anesthetics. Toxicol Lett 1998; 100–101:203–7 mission. Nature 2001; 411:261–8 21. Nelson LE, Guo TZ, Lu J, Saper CB, Franks NP, Maze M: The sedative 41. Feng Z-P, Hasan S, Lukowiak K, Syed NI: Target cell contact suppresses Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/102/2/353/357598/0000542-200502000-00018.pdf by guest on 28 September 2021 component of anesthesia is mediated by GABA(A) receptors in an endogenous neurite outgrowth from soma-soma paired Lymnaea neurons. J Neurobiol 2000; pathway. Nat Neurosci 2002; 5:979–84 42:357–69 22. Horlocker TT, Neal JM: One hundred years later, I can still make your heart 42. Feng Z-P, Grigoriev N, Lukowiak K, Goldberg JI, Syed NI: Development of stop and your legs weak: The relationship between regional anesthesia and local Ca2ϩ hotspots during synaptogenesis between Lymnaea neurons is target cell anesthetic toxicity. Reg Anesth Pain Med 2002; 27:543–4 contact specific. J Physiol (London) 2002; 539:53–65 23. Zorumski CE, Olney JW, Wozniak DF: Early exposure to common anes- 43. Wang GK: Local anesthetics, Neural Mechanisms of Anesthesia. Edited by thetic agent causes widespread neurodegeneration in the developing rat brain Antognini JF, Raines DE, Carstens EE. Totowa, New Jersey, Humana Press, 2003, and persistent learning defects. J Neurosci 2003; 23:876–82 pp 427–40 24. Boselli E, Duflo F, Debon R, Allaouchiche B, Chassard D, Thomas L, 44. Bulloch AGM, Syed NI: In vitro reconstruction of neural circuits. Trends Portoukalian J: The induction of apoptosis by local anesthetics: A comparison Neurosci 1992; 15:442–7 between lidocaine and ropivacaine. Anesth Analg 2002; 96:755–6 45. Munno DW, Syed NI: Synaptogenesis in the CNS: an odyssey from wiring 25. Gozil R, Kurt I, Erdogan D, Calguner E, Keski S, Kurt MN, Elmas C: together to firing together. J Physiol (London) 2003; 552:1–11 Long-term degeneration and regeneration of the rabbit facial nerve blocked with 46. Klein M: Synaptic augmentation by 5-HT at rested Aplysia sensorimotor conventional lidocaine and bupivacaine solutions. Anat Histol Embryol 2002; synapses: Independence of action potential prolongation. Neuron 1994; 13: 31:293–9 159–66 26. Girdlestone D, Cruicshank SG, Winlow W: A system for the application of 47. Haydon PG: The formation of chemical synapses between cell-cultured general anesthetics and other volatile agents to superfused, isolated tissue prep- neuronal somata. J Neurosci 1988; 8:1032–8 arations. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1989; 92C: 48. Fuchs PA, Henderson LP, Nicholls JG: Chemical transmission between 35–7 individual Retzius cells and sensory neurons of leech in culture. J Physiol (Lon- 27. Girdlestone D, McCrohan CR, Winlow W: The actions of three volatile don) 1992; 323:195–210 general anesthetics on withdrawal responses of the pond snail Lymnaea stag- 49. Fuchs PA, Nicholls JG, Ready DF: Membrane properties and selective nalis. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1989; 92C:39–43 connections of identified leech neurons in culture. J Physiol (London) 1981; 28. Spencer GE, Syed NI, Lukowiak K, Winlow W: Halothane-induced synaptic 316:203–23 depression at both in vivo and in vitro reconstructed synapses between identi- 50. Nicholls JG, Liu Y, Payton BW, Kuffler DP: The specificity of synapse fied Lymnaea neurons. J Neurophysiol 1995; 74:2604–13 formation by identified leech neurons in culture. J Exp Biol 1990; 153:141–54 29. Spencer GE, Syed NI, Lukowiak K, Winlow W: Halothane affects both 51. Onizuka S, Kasaba T, Hamakawa T, Ibusuki S, Takasaki, M: Lidocaine inhibitory and excitatory synaptic transmission at a single identified molluscan increase intracellular sodium concentration through voltage-dependent sodium synapse, in vivo and in vitro. Brain Res 1996; 714:38–48 channels in an identified Lymnaea neuron. ANESTHESIOLOGY 2004; 101:110–9 30. Franks NP, Lieb WR: Volatile general anesthetics activate a novel neuronal 52. Taylor RE: Effect of procaine on electrical properties of squid axon mem- Kϩ current. Nature 1988; 333:662–4 brane. Am J Physiol 1959; 196:1071–8 31. Franks NP, Lieb WR: Selective effects of volatile general anesthetics on 53. Lee HT, Xu H, Siegel CD, Krichevsky IE: Local anesthetics induce human identified neurons. Ann N Y Acad Sci 1991; 625:54–70 renal cell apoptosis. Am J Nephrol 2003; 23:129–39 32. Franks NP, Lieb WR: Stereospecific effects of inhalational general anes- 54. Takenami T, Yagishita S, Asato F, Arai M, Hoka S: Intrathecal lidocaine thetic optical isomers on nerve ion channels. Science 1991; 254:427–30 causes posterior root axonal degeneration near entry into the spinal cord in rats. 33. Woodall AJ, Naruo H, Prince D, Feng Z-P, Winlow W, Takasaki M, Syed NI: Reg Anesth Pain Med 2002; 27:58–67 Chronic anesthetic treatment blocks synaptogenesis but not neuronal regenera- 55. Rehder V, Kater SB: Regulation of neuronal growth cone filopodia by tion in cultured Lymnaea neurons. J Neurophysiol 2003; 90:2232–9 intracellular calcium. J Neurosci 1992; 12:3175–86 34. Syed NI, Hasan Z, Lovell P: In vitro reconstruction of neuronal circuits: A 56. Kater SB, Mattson MP, Cohan C, Connor J: Calcium regulation of the simple model system approach, Modern Techniques in Neuroscience Research. neuronal growth cone. Trends Neurosci 1988; 11:315–21 Edited by Windhorst U, Johnsson H. Berlin, Springer, 1999, pp 361–77 57. Kater SB, Mills LR: Regulation of growth cone behaviour by calcium. 35. Feng Z-P, Klumperman J, Lukowiak K, Syed NI: In vitro synaptogenesis J Neurosci 1991; 11:891–9 between the somata of identified Lymnaea neurons requires protein synthesis 58. van Kesteren RE, Syed NI, Munno DW, Bosewan J, Feng Z-P, Geraerts but not extrinsic growth factors or substrate adhesion molecules. J Neurosci WPM, Smith AB: Synapse formation between central neurons requires postsyn- 1997; 17:7839–49 aptic expression of the MEN1 tumor suppressor gene. J Neurosci 2001; 21:1–5

Anesthesiology, V 102, No 2, Feb 2005