Long-term Exposure to Local but Not Inhalation Anesthetics Affects Neurite Regeneration and Synapse Formation between Identified Lymnaea  Neurons

Laboratory-raised stocks of the fresh water pond snail Lymnaea stagnalis  were maintained at room temperature (18–20°C) in an aquarium containing well-aerated and dechlorinated tap water and were fed lettuce. Animals aged 1–2 months (shell length, 10–15 mm) were used for cell isolation while the brain-conditioned medium (CM) was prepared from 3- to 4-month-old animals (shell length, 15–25 mm). (Animal Care Certification is not required for invertebrate species such as L. stagnalis  at the University of Calgary Animal Resource Centre, Calgary, Alberta, Canada.)

Cells were isolated individually and cultured as described previously.34Briefly, snails were anesthetized with 10% Listerine (21.9% ethanol, 0.042% menthol; Pfizer Inc., New York, NY) solution in normal Lymnaea  saline (containing 51.3 mm NaCl, 1.7 mm KCl, 4.1 mm CaCl², and 1.5 mm MgCl²) buffered to pH 7.9 with HEPES. The central ring ganglia were removed under sterile conditions and washed with antibiotic saline (40 μg/ml gentamycin; two washes, 10 min each). The ganglia were then treated with 0.2% trypsin (Sigma type III; Sigma Chemical Company, St. Louis, MO) for 22 min followed by 0.2% soybean trypsin inhibitor (Sigma type 1-S; Sigma Chemical Company) for 10 min, both in defined medium (DM). DM consisted of serum-free 50% L-15 medium with added inorganic salts (40 mm NaCl, 1.7 mm KCl, 4.1 mm CaCl², 1.5 mm MgCl², and 10 mm HEPES). The pH was adjusted to 7.9 with 1 N NaOH, and 20 μm gentamycin was added. The enzymatically treated ganglia were pinned to the bottom of a dissection dish that contained 8 ml high-osmolarity DM (DM with 37.5 mm glucose).

The cells were isolated by applying gentle suction to a fire-polished and Sigmacote (Sigma Chemical Company)–treated pipette. The isolated cells were plated on poly-l-lysine pretreated coverslips glued to the bottom of a 35-mm Falcon dish34in the presence of CM. The CM was prepared by incubating central ganglia in DM (12 ganglia/6 ml DM) for 4 or 5 days and frozen until used. Isolated somata of identified neurons were either allowed to extend neurites or were juxtaposed in a soma–soma configuration and left undisturbed overnight.35 

To assess neuronal regeneration in either the absence or the presence of each anesthetic, identified neurons were isolated in cell culture and plated on poly-l-lysine–coated dishes containing CM plus the anesthetics. Presynaptic (visceral dorsal 4 [VD4]) and postsynaptic (left pedal dorsal 1 [LPeD1]) neurons were selected for neurite outgrowth and synapse formation assays. To test for the effects of each anesthetic agent on neurite outgrowth and synapse formation, the anesthetic was added to the dish 1 h after cell plating, and the neurons were maintained overnight in the dark. Neuronal sprouting was assessed as described previously.36Specifically, only those neurons exhibiting outgrowth (multiple branches, active growth cones, and so forth) equivalent to five somata diameter were considered as sprouted. The extent of outgrowth was calculated as a function of maximum neurite length. Neurons cultured overnight in CM alone served as controls.

To prepare neurite–neurite synapses, the isolated cells were plated in CM a few soma diameters apart and were allowed to extend neurites. After 24–48 h, the cells exhibited extensive outgrowth with multiple overlapping neurites. The cells were cultured overnight either in CM alone or CM plus the anesthetic, which was subsequently washed out during electrophysiologic recordings. To demonstrate the reversibility of anesthetic-induced effects, the compound was washed away with normal saline for more than an hour before the intracellular recordings.

Conventional intracellular recording techniques were used.34Specifically, glass microelectrodes (1.5 mm ID, with filament; World Precision Instruments, Sarasota, FL) were pulled on a vertical electrode puller (Kopf, 700C; David Kopf Instruments, Tujunga, CA) and filled with a saturated solution of K²SO⁴(resistance, 30–60 MΩ). Isolated cells were viewed under a Zeiss (Telaval 31; Carl Zeiss Canada Ltd., North York, Ontario, Canada) inverted microscope and impaled using Narishigi micromanipulators (model MO-103; Narishigi Instruments, Tokyo, Japan). The intracellular signals were amplified via  a preamplifier (Neurodata model IR-283; Cygnus Technology Inc., Delaware, PA), displayed on a storage oscilloscope (Tektronix R5103N; Tektronix, Montreal, Quebec, Canada), and recorded and stored on a computer using Digidata software (Axon Instruments, Union City, CA). All experiments were performed at room temperature (18°–22°C).

Sevoflurane (Sevofrane; Maruishi, Osaka, Japan) and isoflurane (Forane; Abbott, Abbott Park, IL) were selected as inhalation anesthetics, whereas lidocaine (Xylocaine; AstraZeneca, London, United Kingdom) and bupivacaine (Anapin; AstraZeneca) were diluted into CM and served as local agents. VD4 and LPeD1 synapse pairs were exposed to sevoflurane and isoflurane (bubbled overnight) at concentrations of 0.2, 2, 10, 20, and 100 mm, and lidocaine and bupivacaine were used at concentrations of 0.001, 0.005, 0.01, 0.1, and 1 mm. Exogenous application of acetylcholine (Sigma Chemical Company), was performed (80-ms pulses, 1–2 psi) in some experiments using 1 μm acetylcholine applied directly to the synaptic site via  a Pneumatic PicoPump (PV800; World Precision Instruments) pressure injector.

Cells were incubated in 20 μm FM1-43 (Molecular Probes, Eugene, OR) for 10 min before the addition of 0.01 mm lidocaine or 5 mm sevoflurane to the bath. The presynaptic cell (VD4) was stimulated to generate 100 action potentials (10 spikes/burst) by conventional electrophysiologic techniques to facilitate the uptake of FM1-43 in cells that were paired overnight in either the presence or the absence of each anesthetic. The styryl dye and anesthetics were then replaced with cold saline to prevent neuronal firing during the washout and to remove background fluorescence. Fluorescent images of the FM1-43–labeled cells were acquired using a Zeiss (Carl Zeiss Canada Ltd.) Axiovert 200M inverted microscope. Excitation light was from a 100-W Hg lamp, excitation filters (490/30 nm), a dichroic mirror (505 nm), and emission filters (570 low pass nm or 610 nm). Phase and fluorescent images were captured with a Photometrics (Tuscon, AZ) Sensys 1400 camera (1- to 100-ms exposure), connected to a computer running Axiovision 3.0 for Windows (Carl Zeiss Canada Ltd.). The pixels in cross-section of the fixed area (200 × 150 μm²) were integrated and measured with NIH image software (version 1.62; National Institutes of Health, Bethesda, MD).

Parametric data are expressed as mean ± SE and were analyzed for significance using one-way analyses of variance with repeated measures (anesthetic concentration was the between-subjects factor) and a Tukey post hoc  test. Nonparametric data are expressed as percent and were analyzed for significance using the chi-square test. Significance was assumed if P  was less than 0.05.

A well-established model of identified Lymnaea  neurons37was used to reconstruct synapses in either a neurite–neurite33,37or a soma–soma configuration.35,37–40Specifically, identified presynaptic neuron VD4 and its postsynaptic partner LPeD1 were isolated from the adult animals and cultured overnight in the presence of CM. Within 12–24 h, neurons (n = 20) exhibited extensive outgrowth (fig. 1A). Simultaneous intracellular recordings revealed that current injection in VD4 (at arrow, fig. 1B) generated 1:1 excitatory postsynaptic potentials (EPSPs) in a manner similar to those observed in vivo .38Similarly, when paired in a soma–soma configuration (n = 32; fig. 1C) in CM, synapses (90–100%) similar to those observed in a neurite–neurite mode also developed (80–100%), albeit in the absence of neurite outgrowth.41Specifically, action potentials in VD4 (at arrow, fig. 1D) resulted in 1:1 EPSPs in LPeD1. This synaptic transmission has previously been shown to be cholinergic.38,39It is important to note that the efficacy of synaptic transmission (consistency of EPSP amplitude) between VD4 and LPeD1 was more consistent in our soma–soma model as compared with the neurite–neurite configuration. Therefore, we opted to use the soma–soma preparation to test for the effects of anesthetics on synaptic transmission. Similar results, however, were obtained from the neurite–neurite preparation.

To test and compare the effects of inhalation (sevoflurane and isoflurane) and local anesthetics (lidocaine and bupivacaine) on cholinergic synaptic transmission between VD4 and LPeD1, the soma–soma paired cells were recorded in either the absence or the presence of various anesthetics. Synapses were tested electrophysiologically. Action potentials in VD4 under control saline conditions (fig. 2A) generated 1:1 EPSPs in LPeD1, and these excitatory responses were mimicked by exogenous acetylcholine (at arrow), which was pressure-applied directly at the synapse under a fast perfusion system.42The perfusion solution was then switched to the saline containing 5 mm sevoflurane, and the synaptic transmission was tested again. We found that both the synaptic transmission and the cholinergic responses were reduced significantly (control, 11.9 ± 0.7 mV; 0.3 mm, 6.4 ± 1.1 mV; 5 mm, 2.2 ± 0.3 mV [n = 11 for each case]) in a concentration-dependent and reversible manner (fig. 2B). Similar results were obtained in the presence of isoflurane (control, 13.9 ± 3.6 mV; 0.3 nm, 7.1 ± 2.6 mV; 5 mm, 2.3 ± 0.9 mV [n = 10 for each]) (fig. 2B).

The above data thus show that both sevoflurane and isoflurane depress synaptic transmission between VD4 and LPeD1 and also the postsynaptic cholinergic responsiveness in LPeD1. These effects were concentration-dependent, and the synaptic transmission recovered within minutes of washout with normal saline (fig. 2B).

Next, we sought to determine the effects of local anesthetics lidocaine and bupivacaine on cholinergic synaptic transmission between VD4 and LPeD1. Synapses were reconstructed as described above, and both the synaptic transmission and the cholinergic responses were tested in either the absence or the presence of lidocaine (fig. 3A). Action potentials in VD4 induced 1:1 EPSPs in LPeD1, and these responses were mimicked by exogenous acetylcholine (applied at arrow). Although lidocaine significantly reduced the amplitude of VD4-induced EPSPs in LPeD1 (EPSP amplitude: control, 11.4 ± 2.6 mV [n = 10]; 0.1 mm lidocaine, 5.6 ± 1.5 mV [n = 8]; 1 mm lidocaine, 1.9 ± 0.8 mV [n = 9]) (fig. 3A), the cholinergic response remained unperturbed as compared with sevoflurane (acetylcholine amplitude: control, 19.7 ± 2.8 mV [n = 10]; 0.1 mm lidocaine, 15.4 ± 5.4 mV [n = 8]; 1 mm lidocaine, 11.8 ± 4.5 mV [n = 9]) (fig. 3B). Similar results were obtained with bupivacaine, and these effects were both concentration-dependent and reversible (EPSP amplitude: control, 12.1 ± 2.6 mV; 0.1 mm bupivacaine, 5.0 ± 1.1 mV; 1 mm bupivacaine, 1.6 ± 0.5 mV; acetylcholine amplitude: control, 19.7 ± 2.8 mV; 0.1 mm bupivacaine, 14.5 ± 5.0 mV; 1 mm bupivacaine, 11.8 ± 4.5 mV [n = 10 for each]) (fig. 3B). These data thus demonstrate that the local anesthetics depress cholinergic synaptic transmission between VD4 and LPeD1 and that these effects most likely involve the presynaptic mechanisms.

To test the hypothesis that both inhalation and local anesthetics affect neurite regeneration and synapse formation from isolated Lymnaea  neurons, cells were cultured in CM in either the absence or the presence of any given anesthetic agent and were allowed to extend neurites. Specifically, cells were paired in close proximity and were allowed to extend neurites. We reasoned that if neurons regenerated their processes, the synapses would develop between the neurites, which could then be detected electrophysiologically. Under normal culture conditions, both cells extended neurites (fig. 4), and when tested electrophysiologically, synapses were detected as described previously (fig. 1; not shown here). Similarly, cells cultured in 5 mm sevoflurane (bubbled and superfused overnight) exhibited extensive outgrowth, which was indistinguishable from control (fig. 4). To demonstrate that the inhalation anesthetics had not evaporated or catalyzed overnight, the synaptic transmission was first tested electrophysiologically in the presence of the incubating medium (CM plus sevoflurane) before washout with normal saline. Under these experimental conditions, action potentials in VD4 did not generate any response in the postsynaptic cell (n = 11; data not shown). However, within a few minutes after washout with normal saline, synapses similar to those observed under control conditions were detected (data not shown). In contrast to the general anesthetics (control, 1.4 ± 0.4 mm [n = 12]; 2 mm sevoflurane, 1.4 ± 0.5 mm [n = 10]; 5 mm sevoflurane, 1.2 ± 0.4 mm [n = 11]), the cultured neurons did not extend neurites in the presence of lidocaine or bupivacaine (0.01 mm lidocaine, 0.4 ± 0.3 mm [n = 10]; 0.1 mm lidocaine, 0.2 ± 0.3 mm [n = 10]; 0.01 mm bupivacaine, 0.3 ± 0.3 mm [n = 10]; 0.1 mm bupivacaine, 0.1 ± 0.2 mm [n = 11]) (fig. 5).

These data thus demonstrate that both lidocaine and bupivacaine significantly retard neurite outgrowth from cultured neurons, whereas in the presence of their inhalation counterparts, both neurite processes and synapses develop normally (fig. 5). Because cells paired in local anesthetics did not extend neurites (and thus the synapse formation), their effects on synapse formation could not be tested and compared with sevoflurane and isoflurane. To test their effects on synaptogenesis directly, we therefore resorted to our soma–soma model where synapses could develop in the absence of neurite outgrowth.

To test and compare the effects of both general and local anesthetics on synapse formation, cells were soma–soma paired in CM overnight in either the absence or the presence of various agents. The incidence of synapse formation between the paired cells was tested electrophysiologically. Under normal CM conditions, the incidence of synapse formation ranged between 90 and 100% (fig. 6). Similarly, neurons paired within the clinical concentration range of sevoflurane/isoflurane also formed normal synapses, although the percent of cells forming excitatory synapses decreased with increasing anesthetic concentrations (10 mm sevoflurane, 70% [n = 30; P < 0.05]; 2 mm sevoflurane, 90% [n = 18]; 0.3 mm sevoflurane, 100% [n = 12]; 10 mm isoflurane, 65% [n = 20]; 2 mm isoflurane, 80% [n = 20; P < 0.05]; 0.3 mm isoflurane, 95% [n = 15]). Cells paired overnight in the presence of both local anesthetics, however, did not develop synapses within the clinical range, even though morphologically they seemed healthy and exhibited normal intrinsic membrane properties (resting membrane action potential parameters and others) (1 mm lidocaine, 0% [n = 10]; 0.1 mm lidocaine, 0% [n = 10]; 0.01 mm lidocaine, 6% [n = 11]; 0.005 mm lidocaine, 15% [n = 11]; 0.001 mm lidocaine, 0% [n = 15]; 1 mm bupivacaine, 0% [n = 10]; 0.1 mm bupivacaine, 3% [n = 9]; 0.01 mm bupivacaine, 11% [n = 9]; 0.005 mm bupivacaine, 21% [n = 9]; 0.001 mm bupivacaine, 50% [n = 12]) (fig. 6). It is interesting to note that in instances in which cells did form synapses in lower concentrations of both local agents, the synaptic transmission between these pairs failed rapidly and after 5–10 min of stimulation; no synaptic transmission was detectable. After several minutes of rest, however, the synaptic transmission resumed, although the amplitude of postsynaptic potential was significantly reduced (data not shown).

Figure 7shows examples of representative traces from cells cultured under control conditions and in the presence of sevoflurane, isoflurane, lidocaine, and bupivacaine. Synapses similar to those observed under control conditions developed in sevoflurane and isoflurane. Moreover, the cells treated with general anesthetic exhibited responses similar to those of cells with exogenously applied acetylcholine, as was seen under control conditions (fig. 7). Conversely, cells paired in both lidocaine (0.01 mm) and bupivacaine (0.01 mm) did not develop synapses, whereas their responses to exogenously applied acetylcholine were similar to those of control and both sevoflurane and isoflurane. These data are summarized in figure 8(EPSP amplitude: control, 10.3 ± 3.4 mV [n = 38]; 5 mm sevoflurane, 8.8 ± 1.7 mV [n = 18]; 5 mm isoflurane, 8.2 ± 1.7 mV [n = 20]). Similarly, the amplitude of acetylcholine-induced depolarizing responses in LPeD1 observed in sevoflurane and isoflurane was similar to that observed under control conditions (acetylcholine amplitude: control, 19.8 ± 3.4 mV [n = 38]; 5 mm sevoflurane, 14.7 ± 5.1 mV [n = 18]; 5 mm isoflurane, 13.7 ± 4.9 mV [n = 20]) (fig. 8).

Interestingly, although the incidence of synapse formation between cells paired in local anesthetics was significantly reduced (P < 0.05) (EPSP amplitude: 0.01 mm lidocaine, 0.5 ± 0.4 mV [n = 11]; 0.01 mm bupivacaine, 0.8 ± 0.3 mV [n = 9]) (fig. 8), their cholinergic responses in LPeD1 were similar to those of control and the inhalation anesthetics (acetylcholine amplitude: 0.01 mm lidocaine, 16.5 ± 4.8 mV [n = 11]; 0.01 mm bupivacaine, 19.6 ± 2 mV [n = 9]) (fig. 8). These data lend further support to the idea that neuronal failure to develop synapses in the presence of local anesthetic is not due to its toxic effects on neuronal excitability or the desensitization of the postsynaptic receptors.

We next sought to determine whether the absence of synapses between VD4 and LPeD1 cultured in local anesthetics could be due to the fact that the presynaptic machinery may not have developed in their presence. Specifically, cells were paired overnight in either the absence or the presence of anesthetics (fig. 9). On day 2, the anesthetic solution was replaced with normal saline containing the dye FM1-43, and the presynaptic cell was stimulated electrically. We reasoned that if cells possessed normal, functional presynaptic machinery (exocytosis and endocytosis), we should expect labeling of the presynaptic cell at its contact with the postsynaptic neuron. To test this possibility, VD4 was stimulated to fire action potentials via  current injections (100 actions potentials) in the presence of the dye FM1-43. The dye was then washed away with normal saline, and images were acquired. We found that under both normal (figs. 9A–C) and sevoflurane (figs. 9D–F) conditions, exclusive labeling of the presynaptic cells was discernible at its contact site with LPeD1 and also in its processes surrounding the postsynaptic somata (fig. 9B, control, and fig. 9E, sevoflurane). In contrast, faint or no staining of the presynaptic cell was observed in VD4 paired in the presence of lidocaine (figs. 9G and H). The pixel values for each category are depicted in the corresponding panels (figs. 9C, F, and I), and the summary data for pixel values are presented in figure 9J.

Taken together, our data demonstrate that although both local and general anesthetics are effective in blocking synaptic transmission between the paired cells, the former is detrimental to nerve regeneration and synapse formation and that these effects are likely due to the failure of the presynaptic assembly in the presence of lidocaine.

This study provides the first direct comparative account of the effects of general and local anesthetics on synaptic transmission (short term), nerve regeneration, and synapse formation (long term). We have demonstrated that although both general (sevoflurane and isoflurane) and local anesthetics used at clinically relevant concentrations affect synaptic transmission between identified neurons in a concentration-dependent and reversible manner, long-term lidocaine and bupivacaine but not sevoflurane and isoflurane treatment retards neurite outgrowth and synapse formation. The neuronal failure to exhibit outgrowth and synapse formation 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 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 understanding is limited vis-à-vis the modes and sites of action of various anesthetic agents. Notwithstanding our recent progress toward elucidating how inhalation and intravenous anesthetics affect neuronal communications4in the brain and achieve desired effects, much less is known about local agents (oral, injected, local) such as lidocaine, which are multimodel and serve as antiarrhythmics, antiepileptics, and anticonvulsants agents.43Because the core of all nervous system functions resides at the synaptic transmission between any given presynaptic and postsynaptic neuron or a subset of such neurons, it is therefore pivotal to examine how anesthetics affect neuronal communication at the level of individual neurons. Such an approach may not be relevant for our understanding of the state of unconsciousness that is experienced during anesthesia or how pain perception is managed by the nervous system; nevertheless, it will provide basic understanding of how anesthetics function at a fundamental level. With this caveat in mind, we took advantage of a simple model system approach and used in vitro  reconstructed synapses between functionally well-defined neurons. VD4 and LPeD1 are members of the cardiorespiratory network37that comprises the respiratory central pattern generator. The synaptic connections between various central pattern–generating 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 that observed in vivo .37Therefore, adult Lymnaea  neurons not only regenerate but also recapitulate their specific connections in cell culture.44Moreover, specific synapses between the central pattern generator neurons can also be reconstructed in a soma–soma configuration.35,45The soma–soma synapses are also target cell contact specific and require new protein synthesis and gene transcription.35Therefore, both morphologically and electrophysiologically, the soma–soma synapses are similar to those seen in vivo .35,42The soma–soma preparation has also been used in a number of other species to define the cellular mechanisms underlying synapse formation, synaptic transmission, and plasticity at a resolution not achievable elsewhere (Aplysia ,46,Helisoma ,47leech48,49; Nicholls50).

The usefulness of identified Lymnaea  neurons for anesthetic research has previously been validated by a number of laboratories (halothane,28–30sevoflurane,13and propofol33,51). In this study, we have compared the effects of both inhalation and local anesthetics on cholinergic synaptic transmission, which was suppressed equally by these agents. The sevoflurane- and isoflurane-induced suppression most likely involved their actions on postsynaptic receptors because cholinergic responses were also significantly reduced by these anesthetics. However, their effects on presynaptic Ca²⁺channels cannot be ruled out. Consistent with this notion are our earlier data where sevoflurane was shown to depress inhibitory synaptic transmission between soma–soma pairs, and these effects involved both presynaptic and postsynaptic ion channels.13Interestingly, although both lidocaine and bupivacaine depressed synaptic responses, they did not significantly reduce the amplitude of the compound postsynaptic potentials in LPeD1 that were elicited by acetylcholine. These results therefore suggest that as compared with the inhalation anesthetics, the synaptic depression induced by the local agents may involve presynaptic mechanisms.

In contrast with their inhalation counterparts, which exert myriad effects on a variety of presynaptic and postsynaptic ion channels,18the local anesthetics primarily affect Na⁺channels43and block action potential propagation across the nerve. Their direct effects on either cell bodies or the synaptic sites, however, have not been fully investigated. Local anesthetics, such as procaine, suppressed nerve propagation in squid giant axon by blocking action potentials,52whereas in this study, we could successfully generate spikes in Lymnaea  neurons, even in the presence of higher concentrations of lidocaine and bupivacaine. Interestingly, within a few seconds after anesthetic perfusion, both cells were depolarized (5–10 mV), and this observation is consistent with an earlier study in which lidocaine increased intracellular Na⁺concentration in Lymnaea  neurons.51It is also important to note that even though we could trigger single spikes in VD4, compound action potentials often did not occur (not shown) in response to a sustained depolarizing pulse. A similar voltage-induced inactivation of spikes has also been reported for Lymnaea  neurons.51It is therefore plausible that local anesthetics may not immediately affect Na⁺channels required for a single action potential at rest (tonic block43), but their depolarizing effects may shift the steady state inactivation curve, thus reducing the availability of various Na⁺channels for spike generation. The neuronal inability to generate repetitive spikes can also be attributed to “use-dependent” or “frequency-dependent block” of Na⁺channel function.43Additional experiments are required to test for the direct versus  indirect actions of local anesthetics on various intrinsic neuronal properties. In addition, to deduce whether local anesthetics may also affect presynaptic Ca²⁺channels, the soma–soma model offers a wonderful opportunity for such research. It is also noteworthy that synaptic recovery from lidocaine and bupivacaine took much longer (several tens of minutes) upon drug washout as compared with sevoflurane and isoflurane, and these long-term effects may also be due to a slower removal of these compounds from the target sites and/or the hydrolysis resistance nature of these ester-type compounds.

In addition to their short-term effects on synaptic transmission, long-term exposure of the neuronal cells to a variety of general anesthetics has been shown to cause widespread degeneration and learning defects in developing rats.23Similarly, long-term treatment of the rabbit facial nerve with lidocaine and bupivacaine perturbed its degeneration and regeneration,25whereas in other studies, these agents were reported to induce apoptosis in various cell types.24,53Lidocaine has also been reported previously to cause axonal degeneration in the posterior root of rats.54To add to the list of these toxic effects, the current study demonstrated that lidocaine and bupivacaine also completely suppress neurite outgrowth from cultured neurons and blocked synapse formation between the soma–soma paired cells. Regarding the effects of local anesthetic on synapse formation, these data are consistent with our previous study in which long-term propofol treatment of the soma–soma paired cells also blocked synapse formation between VD4 and LPeD1. However, it is interesting to note that the cells paired in propofol did reestablish their specific synapses after several hours of anesthetic washout,33whereas in the current study, synapses did not develop even after a day of the anesthetic removal (5 μm propofol, 70% [n = 10]; 10 μm, 60% [n = 9]; 25 μm, 25%, [n = 8] 50 μm, 0% [n = 6]).33One may argue that a long-term treatment of the paired cells with local anesthetics may have rendered the synaptic transmission undetectable. To rule out this possibility, we performed the experiments in which cells were first allowed to develop synapses in CM and then the pair was exposed overnight to the local anesthetics. On day 3, the anesthetic was washed, and within an hour of medium removal, normal synaptic transmission was detected (not shown). These results therefore argue in favor of the idea that local anesthetics block synapse formation between VD4 and LPeD1. Another interesting difference between propofol33and local anesthetic was that in the presence of the former, the cells that normally form chemical synapses exhibited electrical coupling, which is not observed in vivo  between the cells.33However, this electrical coupling was not observed in either lidocaine or bupivacaine, further underscoring their effects on both chemical and electrical synapses.

When comparing the effects of sevoflurane, isoflurane, and propofol with lidocaine/bupivacaine on neurite outgrowth, both inhalation (sevoflurane/isoflurane, this study) and intravenous (propofol33) anesthetics did not affect neuronal sprouting, and cells extended processes that were indistinguishable from those of cells cultured in CM. This negative result could not be attributed to the volatile nature of this compound because no synaptic transmission was observed between the paired cells before the anesthetic washout with saline. However, neurite outgrowth was completely retarded by both local anesthetics (lidocaine and bupivacaine). Consistent with these data are earlier studies in which lidocaine was found to induce the collapse of growth cones from dorsal root ganglia neurons,2and these effects may involve an intracellular enhancement of growth cone Ca²⁺concentration in a manner similar to that observed after tetracaine treatment.3On one hand, these studies demonstrate a Na⁺channel–independent action of local anesthetics on neuronal tissues, and together with our data, they provide unequivocal evidence that long-term exposure of neurons to local but not general anesthetics is detrimental to nerve regeneration. However, these effects could not be attributed to the local anesthetic–induced toxic influences on neuronal viability or a perturbation of their intrinsic membrane properties because all neurons seemed healthy and electrophysiologically viable (firing action potential and response to acetylcholine pulses). Therefore, as demonstrated previously, one of the likely mechanisms by which local agents may prevent neurite outgrowth is to increase intracellular concentration in the growth cones from a growth permissive to a growth suppressive range.55–57Whether the local anesthetic–induced suppression observed here also involved intracellular Ca²⁺remains to be determined.

We have previously demonstrated that voltage-induced Ca²⁺hot spots develop between soma–soma paired cells and that these are both target cell and contact site specific.42Consistent with these data are the studies that showed that the presynaptic secretory machinery is specialized at the contact site between the cells. Specifically, the dye FM1-43 (which labels vesicles during exocytosis and is subsequently internalized during endocytosis) did not label presynaptic neurons that had not developed synapses after long-term propofol treatment.33In the current study, we have demonstrated that cells treated chronically with local anesthetics also did not label with the dye, suggesting that in the presence of these anesthetics, the transmitter secretory machinery responsible for exocytosis and endocytosis of cholinergic vesicles does not assemble at the contact site between the paired cells Whether this failure is due to neuronal inability to target Ca²⁺channels at the synaptic site or a disruption of other candidate molecules in the secretory pathway remains to be investigated.

In summary, notwithstanding the need to develop better therapeutic drugs for long-term pain management, analgesia, and anesthesia, the current preclinical study underscores the importance of a rigorous screening of such agents regarding their long-term effects on nerve regeneration and synapse formation. The ageing population and advanced surgical procedures necessitate the use and development of long-lasting local anesthetics for both short- and long-term pain management. However, these strategies should not be at the expense of their deleterious affects on nerve regeneration and synapse formation, which are pivotal for functional recovery after the loss of all nervous system function after surgery. On one hand, the data presented here provide a comparison between various anesthetics for their effects on synaptic transmission, regeneration, and synapse formation, and on the other hand, this model system provides us with an unparalleled opportunity to identify various genes58and their products that are differentially regulated in single cells after long-term anesthetic treatment.

The authors thank Wali Zaidi (Technician, Calgary Brain Institute, University of Calgary, Calgary, Alberta, Canada) for excellent technical support.