Administration of local anesthetics (LAs) in the cerebrospinal fluid (CSF) during spinal anesthesia can be associated with postoperative neurological complications, such as transient neurological symptoms and persistent lumbosacral neuropathy. Transient neurological symptoms manifest as radiating pains uni- or bilaterally in the buttocks and legs of variable severity, with the symptoms usually resolving within 1–2 wk(s). It is quite common and can occur as often as one in every three patients receiving lidocaine.1 The risk of persistent lumbosacral neuropathy can be as high as 1 in 200 after continuous lidocaine spinal anesthesia or 1 in 1300 after a single lidocaine injection.2 The most severe complication, cauda equina syndrome, which is characterized by bowel and bladder sphincter incontinence, sexual dysfunction, paresthesias, and leg muscle weakness, can occur with an incidence from 1 in 1000 to 1 in 10,000 depending on the LA used.2 LAs are used clinically at very high molar concentrations (Table 1). Although they mix with the CSF, there could be a short period of time in which nerve tissues may be surrounded by very high LA concentration. In addition, with sacral pooling or migration of an epidural catheter, LA concentrations can remain high in one location and cause severe nerve injury.
In animal models, neurotoxicity of different LAs at clinical concentrations has also been demonstrated. In in vitro neuronal models (isolated nerves, spinal cord, or neuronal cells), LAs have been shown to cause conduction block, loss of action and/or resting potential, increase of intracellular Ca2+, growth cone inhibition, and cell death.3–7 LAs inhibit voltage-dependent Na+ channels, as well as K+ and Ca2+ channels,5,7–14 but at concentrations that are much lower (100- to 1000-fold, in μM-range) than clinical concentrations during regional anesthesia (in mM-range). A number of studies have shown that LAs may cause mitochondrial dysfunction by collapsing the mitochondrial membrane potential, uncoupling oxygen consumption, and adenosine diphosphate-phosphorylation, releasing cytochrome c, and activating caspases, which may all lead to apoptotic cell death.6,15–28 Although LA toxicity observed with in vitro studies does not necessarily correlate with toxicity in the clinical setting, information about the mechanism of LA toxicity might be obtained.
Although neurological complications have been observed with the use of different LAs, lidocaine appears to be more neurotoxic than other LAs.4,29–31 Neurological complications are associated with lidocaine at an incidence of one order of magnitude greater than other LAs.2 In comparative studies, higher cytotoxicity has also been observed with lidocaine than with other LAs, including bupivacaine, mepivacaine, ropivacaine, and procaine.32–34 However, many of the toxicity studies have been conducted using different conditions and exposure times ranging from 3 min to 24 h of LA treatment. In this study, we determined and compared the cytotoxicity of commonly used LAs in a human neuronal cell model. Cell viability after a 10-min LA treatment was determined as the primary indicator for LA toxicity. Caspase activation after bupivicaine and lidocaine treatment, as a measure of apoptosis, was examined in cells treated for 10-min with or without a 3-h recovery period. In addition, the LA effect on intracellular Ca2+ responses was evaluated.
Undifferentiated human SH-SY5Y neuroblastoma cells (originally provided by Dr. J. Biedler, Sloan-Kettering Institute for Cancer Research, Rye, NY) were seeded on 24-well tissue cultures plates with a plating density of approximately 5 × 105 cells/well, then maintained in RPMI 1640 medium containing 12% fetal bovine serum and penicillin (100 U/mL) plus streptomycin (100 μg/mL). The cell culture was kept at 37°C in a tissue culture incubator (Heraeus BBD 6220) with 5% CO2 supply and an average humidity of 95%. Measurements of cell viability were performed on cells 3 days after seeding.
Ca2+ measurements were performed on cells 3–5 days after the confluence was reached.
All cell culture components were GIBCO products and purchased from Invitrogen (Carlsbad, CA).
Measurement of Cell Viability Using MTT-Colorimetry Assay
Cell viability was evaluated based on the method of Mosmann35 with some modifications. In brief, 3 or 4 days after seeding, phenol red-containing culture medium was removed and cells in multiwell plates were gently rinsed twice with Dulbecco’s phosphate buffered saline (DPBS) with 10 mM glucose-addition (pH 7.4). Colorless RPMI 1640 medium 300 μL was then added to each well, and cells were temperature-equilibrated for 5 min at 37°C in a thermo-stated heat block. Subsequently, LAs were added to each well and incubated with cells for 10 min at 37°C. At the end of the LA treatment, medium (containing detached cells) was collected in microcentrifuge tubes from each well and centrifuged for 1 min at 5000 rpm in an Eppendorf Bench-top centrifuge (type 5415C). Each pellet was resuspended in 300 μL fresh colorless RPMI 1640 medium. In addition, 300 μL fresh colorless RPMI 1640 medium were added to each well containing still-attached cells. Both cells remaining in wells and collected in microcentrifuge tubes were then mixed with 30 μL/well (or sample) of yellow-colored MTT-reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) and incubated at 37°C for 4 h in a tissue culture incubator. During this incubation, MTT was converted by cellular dehydrogenases to a purple-colored formazan product. To dissolve the formazan precipitant, 300 μL isopropanol was added to each sample either in culture wells or in microcentrifuge tubes, containing attached and detached cells, respectively. The optic density (OD) of the homogenous purple solutions was then measured using a spectrophotometer (Beckman DU 7400) at 570 nm (OD570) and at 630 nm (the reference wavelength) (OD630). The difference between OD570 and OD630 (i.e., OD570 − OD630) was used for quantifying the specific MTT conversion. Because the MTT conversion can only be completed by living cells, the degree of the MTT conversion (i.e., the amount of purple formazan product) was proportional to the number of live cells.35 Total MTT conversion by cells in a given well was the sum of the MTT conversion by both attached and detached cells. MTT conversion found in cells without LA treatment (the control sample) was set as 100% and the MTT conversion found in cells treated with LA was normalized to the corresponding control values and used as an indicator for cell viability (in % versus the untreated control groups).
Tetrodotoxin/Tetraethyl-Ammonium Effect on Cell Viability
Because LAs inhibit voltage-dependent Na+ channels, as well as K+ and Ca2+ channels,8–12,36 we tested whether inhibition of channel function contributed to cell death within the same timeframe as LA treatment. MTT conversion by SH-SY5Y cells was examined in the presence of Na+ channel blocker tetrodotoxin (TTX) or K+ channel blocker tetraethyl-ammonium (TEA) or both TTX plus TEA. Cells were incubated with either 1 μM TTX or 10 mM TEA, or 1 μM TTX plus 10 mM TEA for 10 min at 37°C. After the treatment, cells were incubated with MTT reagent as above and the MTT conversion was evaluated.
Microscopic Observation of Cell Morphology During an LA Treatment
Cells in multiwell plates were placed on the stage of a microscope (Zeiss Axiovert 200M, Jena, Germany), lidocaine (10 mM) or bupivacaine (1 mM) was slowly added to the wells to be examined. Images in Normaski mode were taken before LA addition and during the entire incubation with LA, using a 40× LD Plan Neofluar® objective (Zeiss, Jena, Germany). This observation was made at room temperature (20°C–22°C).
Fluorescence Imaging of Cell Viability and LA Cytotoxicity
SH-SY5Y Cells (approximately 7 × 104 cells/well) were seeded on glass coverslips (φ 12 mm) precoated with poly-l-lysine (12.5 μg/mL) and held in 12-well culture plates. Cells were then maintained under the same conditions described in the section “Cell Culture” for 5–6 days before use.
Adherent cells (on coverslips) were gently washed twice with warm (37°C) DPBS. Cells were then exposed to different concentrations of LAs (in colorless RPMI 1640 medium) for 10 min at 37°C. Control coverslips were incubated in LA-free colorless RPMI 1640 medium under the same conditions. Subsequently, the buffer containing detached cells was collected from each well into microcentrifuge tubes, the cells were centrifuged at 5000 rpm (Eppendorf Centrifuge 5415C, room temperature), and the pelleted cells were washed twice with DPBS. Cells still attached to coverslips were also washed twice with DPBS in the wells. Both adherent and detached cells were then stained with the LIVE/DEAD® assay reagents (calcein/AM and ethidium homodimer-1) (Invitrogen, Carlsbad, CA) following the protocols provided by Invitrogen. After washing twice with DPBS, some cells were killed with 70% methanol for 30 min at room temperature. These samples served as the death-controls. All slides were stored at 4°C overnight to allow the coverslip sealing to be tried.
All slides were imaged using a Zeiss microscope (Axiovert 200M, Jena, Germany). Images in fluorescence mode were taken using 20/40× EC Plan Neofluar objectives (Zeiss, Jena, Germany). Because live cells possess ubiquitous intracellular esterase activity and can convert the nonfluorescent cell-permeant calcein/AM to the intensely green-fluorescent cell-impermeant calcein, the live cells were labeled with evenly distributed intracellular green fluorescence, visualized using the filter set BP 450–490 (Ex)/BP 515–565 (Em) with a beam splitter FT 510. In contrast, ethidium homodimer-1 only enters cells with damaged membranes (of the dead cells) and undergoes a 40-fold enhancement of a red fluorescence (Ex/Em = 528/617 nm) upon binding to nucleic acids, but can be excluded by intact membrane of live cells. Therefore, the dead cells were labeled by red fluorescence of ethidium homodimer-1-DNA, visualized using the filter set BP 546/12 (Ex)/LP 590 (Em) with a beam splitter FT 580. The fluorescence illumination was provided by an X-Cite 120 illuminator with a 120 W metal halide lamp (EXFO Life Sciences & Industrial Division, Mississauga, Canada).
Detection of Early and Delayed Caspase Activation upon LA Treatment Using the Image-iT™ LIVE Red Caspase-3/-7 Detection Kit
Cells were seeded and cultured on coverslips as described above. Culture medium was replaced with fresh medium containing different concentrations of bupivacaine (0.5/1/5 mM) or of lidocaine (10/20/40 mM) and cells were incubated for 10 min at 37°C.
The LA-containing medium was removed and cells were further treated through two different paths: 1) for examining immediate LA effect, cells were washed twice with fresh culture medium and then incubated with a fluorescent inhibitor of caspases (FLICA™, Invitrogen Kit # I-35102 with SR-DEVD-FMK) for 60 min. After rinsing away the FLICA, cells were incubated with 1 μM Hoechst 33342 for 5 min to stain the nuclei and washed with DPBS. All coverslips were then mounted on microscope slides using the apoptosis fixative solution provided by the detection kit and sealed with nail polish. 2) To examine delayed LA effect, cells were first given a 3-h recovery after LA exposure in the regular culture medium and under the normal culturing conditions. All coverslips were then incubated with FLICA and Hoechst 33342, washed, fixed, and mounted. As an apoptotic positive control, some cells were treated with DNAse I for 10 min instead of LA before incubation with FLICA. Cells exposed only to regular medium served as the negative controls.
Evaluation of all slides was performed using a Zeiss microscope (Axiovert 200M, Jena, Germany). Images in fluorescence mode (illuminated by X-CITE 120, see above) were taken using 20/40× EC Plan Neofluar objectives (Zeiss, Jena, Germany). A red fluorescence (Ex/Em = 550/595 nm) indicating cells with activated caspases were visualized using the filter set BP 546/12 (Ex)/LP 590 (Em) with a beam splitter FT 580. The blue fluorescence (Ex/Em = 350/461 nm) representing the Hoechst-stained nuclei was visualized using the filter set G 365 (Ex)/LP 420 (Em) with a beam splitter FT 395.
Measurements of Cytosolic Ca2+
Cells in confluent culture were loaded with 5 μM Fura-2/AM (Molecular Probes, Eugene, OR) for 30 min at 37°C in the original culture medium. After removal of the dye-containing medium, the cells (still attached) were rinsed twice with DPBS without CaCl2 and MgCl2 but 10 mM glucose-addition (pH 7.4), then gently resuspended in an incubation buffer (IB, pH 7.4) containing 140 mM NaCl, 5 mM KCl, NaHCO3, 1 mM MgCl2, 10 mM HEPES, and 10 mM glucose. Cells were incubated in buffer containing 1.5 mM CaCl2 for 1 h at room temperature to allow cell recovery. After a brief centrifugation (1 min at 3000 rpm) before each measurement, the cell pellet was resuspended in the same incubation buffer and thermo-equilibrated at 37°C for 5 min.
Ca2+ measurements were performed using a FluoroMax-3 fluorescence spectrofluorometer (Jobin Yvon, Edison, NJ). After a short (2–3 min) baseline recording, 1.5 mM CaCl2 was added to the suspension and cells usually showed an elevated [Ca2+]i baseline [Ca2+]i increase. Three minutes later, cells were challenged by either 100 mM KCl or 1 mM CAB. Fluorescence ratios were monitored at 510 nm after alternated excitation at 340 and 380 nm. Maximal and minimal fluorescence ratios were acquired by using 0.01% sodium dodecyl sulfate and 60 mM ethylene glycol-bis-(β-aminoethyl ether)N,N,N′,N′-tetraacetic acid, and [Ca2+]i was then determined from fluorescence ratios based on the method of Grynkiewicz et al.37 and a Kd of fura-2 for Ca2+ of 224 nM, as described previously.38
In the drug-treated groups, cells were preincubated with either procaine or chloroprocaine (0.1–10 mM) for 5 min at 37°C before CaCl2-addition, before 1.5 mM CaCl2 was added.
In medical practice, concentrations for LAs are usually expressed in percentile (%) instead of in molarity (mM). Table 1 shows the comparison between these two modes of expression of concentration for the six LAs used in this study.
All chemicals used for this study were obtained from Sigma Chemicals Co. (St. Louis, MO), unless otherwise indicated. Stock LAs used in this study are from the following sources: bupivacaine 0.5% (15 mM, pH 4.0–6.5), and 0.75% (23 mM, pH 4.0–6.5) Sensorcaine®-MPF from AstraZeneca LP (Wilmington, DE), ropivacaine 0.5% (16 mM, pH 4.0–6.0) Naropin® from AstraZeneca LP, mepivacaine 1.5% (53 mM, pH 4.5–6.8) and 2% (71 mM, pH 4.5–6.8) Polocaine®-MPF from AstraZeneca LP, 3% (98 mM, pH 2.7–4.0) Chloroprocaine from Bedford Laboratories (Bedford, OH). LA solutions of all other concentrations than the stock solutions (i.e., the clinical solutions) used in this study were prepared by mixing the stocks with a physiological buffer (IB, pH 7.4). All mixtures were then checked to ensure a physiological pH value (7.35–7.40). Stock solutions for lidocaine (500 mM) and procaine (100 mM) were made from the powder chemicals acquired from Sigma in a physiological buffer (IB, pH 7.4). Dilutions were made from the stocks with the same physiological buffer.
All Six LAs Decreased Cell Viability
Figure 1 shows the effect of four amide (bupivacaine, ropivacaine, mepivacaine, and lidocaine) and two ester (procaine and chloroprocaine) LAs on cell viability after a 10-min drug exposure. All six LAs decreased cell viability in a concentration-dependent fashion. Based on their LD50 (concentration at which 50% of the cells were dead), the potency of these six LAs showed the following order: bupivacaine > ropivacaine > chloroprocaine > lidocaine > mepivacaine ≥ procaine (Table 2). In our assay system and in the concentration range examined (up to 100 mM), only bupivacaine and lidocaine, but none of the other LAs, was able to kill all cells with increasing concentration. At 10 mM and lower, procaine appeared similar to lidocaine in toxicity, whereas mepivacaine, chloroprocaine, and ropivacaine were all approximately equivalent and intermediate in toxicity between lidocaine and bupivacaine.
An alternative way to view the data is to compare equivalent concentrations of LAs. For example 1.2 mM (0.037%) bupivacaine kills 50% of the cells after a 10 min exposure, whereas 1.2 mM (0.030%) lidocaine and 1.2 mM procaine have essentially no effect. At 1.2 mM (0.031%) mepivacaine, 1.2 mM (0.034%) chloroprocaine, and 1.2 mM (0.034%) ropivacaine kill approximately 15%–25% of the cells.
TTX and TEA Did Not Affect the LA Effect on Cell Viability
A 10-min treatment with Na+ channel blocker TTX at 1 μM (a concentration shown to completely block voltage-dependent Na+ currents in SH-SY5Y cells39–41) had no effect on cell viability (Fig. 2). A 10-min treatment with the K+ channel blocker TEA (10 mM) did not cause cell death (Fig. 2). Co-application of 1 μM TTX and 10 mM TEA did not affect cell viability (Fig. 2).
Morphological Changes of SH-SY5Y Cells Caused by Lidocaine or Bupivacaine
Figure 3 shows representative morphological changes of SH-SY5Y cells in the absence and presence of 10 mM lidocaine (approximately 0.27%) or 1 mM bupivacaine (approximately 0.032%) over a period of 20–30 min. Although the morphology of the control cells (without LA) remained unchanged (Figs. 3A–C), a 20-min treatment with 10 mM lidocaine (causing approximately 34% cell death within 10 min, Fig. 1) resulted in retraction of the cell extensions (Fig. 3F). The cell shape gradually became rounded. Twenty minutes after adding drugs, very few cell extensions were visible and almost all cells looked rounded (Fig. 3F). In comparison with lidocaine, 1 mM bupivacaine (causing approximately 42% cell death within 10 min, Fig. 1) showed little effect on cell morphology over a period of 30 min (Fig. 3I).
LA Effect on Cell Viability Demonstrated by Fluorescence Labeling
Figure 4 shows representative images of cell viability after a 10-min treatment with different LAs at a concentration causing approximately 50% cell death. Methanol was used to kill cells to validate the staining for dead cells. Figures 4A and B present untreated and methanol-treated cells as the live (green-colored) and dead (red-colored) controls, respectively. In some LA-treated cells, superposition of green and red images (representing the intracellular distribution of calcein- and ethidium-homodimer-DNA, respectively) resulted in orange-colored images, indicating that the integrity of the plasma membrane and the nuclei was impaired. In general, LA treatment showed a similar dose-dependent effect on cell viability in the imaging data in agreement with the results obtained from the assays using the MTT-colorimetry in attached and detached cells. Depending on the concentration applied, different degrees of cell death were reflected on images by different intensity of the orange/red staining. In comparison with the images shown in Figure 4, images obtained from cells treated by lower LA concentrations contained mostly green-colored vital cells (data not shown). Chloroprocaine at 5 mM caused extensive cell detachment. The fluorescence staining of the detached cells showed strongly impaired membrane integrity (Fig. 4E). Procaine at 10 mM caused less cell-detachment but a significant number of cells were also damaged (Fig. 4F).
Caspase Activation by LA Treatments
To determine whether apoptosis is involved in cell death caused by bupivacaine and lidocaine, we examined caspase activation. Figures 5A and B represent the negative and positive controls of caspase activation. In Figure 5A (untreated cells), only blue nuclear staining was visible; the red fluorescence in the background of the blue staining in Figure 5B (cells treated with DNAse I) shows positive caspase activation recognized by FLICA. The lack of red fluorescence in Figures 5C–E indicates that a 10-min treatment with bupivacaine (0.5/1/5 mM) did not cause immediate caspase activation. In cells with a 3-h recovery after 10-min bupivacaine-treatment, however, visible red fluorescence underneath the blue nuclear staining (Figs. 5F–H) suggests delayed caspase-activation. In comparison with bupivacaine, lidocaine at 1 and 5 mM did not cause significant caspase-acitivation, without or with the 3-h recovery (data not shown). However, lidocaine at 10 mM caused caspase-activation in cells both without (Fig. 6A) and with (Fig. 6D) a 3-h recovery. At a concentration >10 mM, lidocaine-caused caspase inactivation was visible immediately after the 10-min treatment (Figs. 6B and C).
LAs Inhibited Carbachol- and KCl-Evoked [Ca2+]i-Transients
In the absence of LA, the basal [Ca2+]i measured was 84 ± 5 (mean ± sd in nM, n = 133) in SH-SY5Y cells incubated in nominally Ca2+-free buffer. The addition of 1.5 mM CaCl2 resulted in a Ca2+-induced [Ca2+]i-increase of about 55 nM. The [Ca2+]i reached a new steady state-level of 139 ± 47 (mean ± sd in nM, n = 35).
Upon adding procaine or chloroprocaine (0.1–10 mM), no increase in basal [Ca2+]i was observed. However, after 5-min exposure to procaine or chloroprocaine, the Ca2+-induced [Ca2+]i-increase appeared to be slightly larger, but no clear LA concentration dependency was apparent in the Ca2+-induced [Ca2+]i-increase. The [Ca2+]i increase (versus untreated control group) was not statistically significant at most LA concentrations, including 10 mM, the highest LA concentration applied in this study. Only at 2.5 and 4 mM procaine or at 1.5, 2, and 4 mM chloroprocaine the increase was significant (P < 0.05, tested by one-way analysis of variance plus Dunnett’s multiple comparison posttests) (Fig. 7).
Both procaine and chloroprocaine (0.1–10 mM) inhibited CAB-evoked [Ca2+]i-transients [resulting from intracellular Ca2+ release from the IP3-sensitive stores42 in a concentration-dependent manner (Fig. 8A)]. Chloroprocaine was 10-fold more potent than procaine (IC50: 0.028 mM for chloroprocaine versus 0.28 mM for procaine). Both procaine and chloroprocaine also inhibited the KCl-evoked [Ca2+]i-transients (Fig. 8B) mediated mainly through Ca2+ entry and Ca2+-induced Ca2+-release from the caffeine-sensitive store.42 Although chloroprocaine showed a concentration-dependent inhibition in the range of 0.1–10 mM and an IC50 of 0.81 mM, procaine inhibited the KCl-evoked [Ca2+]i-transients only at concentrations >1 mM with an IC50 of 7.8 mM, nearly 10-fold less potent than chloroprocaine.
LA Effect on Cell Viability
The current study demonstrates that all six tested LAs are toxic to neuronal SH-SY5Ycells. Cell death occurred at LA concentrations far below the peak clinical concentration in the CSF and quickly after a 10-min LA treatment. Based on the curves in Figure 1 depicting cell viability after LA treatment, bupivacaine is the most and lidocaine and procaine are the least potent LAs (Fig. 1). All six LAs killed cells at subclinical concentrations, but only bupivacaine and lidocaine were able to kill all cells at the concentrations examined. Treatment with ropivacaine, mepivacaine, procaine, or chloroprocaine only caused approximately 50%–70% cell death at the highest concentrations used in this study and the concentration-dependency curves were shallower than those of bupivacaine and lidocaine (Fig. 1). Because the starting LA solutions of mepivacaine, ropivacaine, and chloroprocaine were unbuffered and had to be mixed with buffer solution to ensure the proper pH value for the conducted experiments, the highest concentrations for LAs applied to SH-SY5Y cells were lower than concentrations administered in the clinical setting. This together with the shallower behavior in their concentration-dependency did not allow us to determine whether these LAs at higher concentrations would kill all cells. The difference in steepness of the concentration-dependency curves suggests that there are more paths involving cell death caused by these three LAs than by bupivacaine and lidocaine.
Treatment with TTX and TEA did not result in cell death, suggesting that the block of Na+ and K+ channels had little involvement in the LA-induced cell death observed in this study in agreement with the work of Sakura et al.43
The viability results are consistent with the previous findings from in vitro investigations studying different LAs in different cell types. In SH-SY5Y cells, Friederich and Schmitz6 found that lidocaine (3–30 mM for 3 h) dose-dependently decreased cell viability and killed almost all cells at 30 mM, with marked loss of neurite extensions, which is very similar to our results with lidocaine (Figs. 1 and 3D–F). In another study, Nishimura et al.44 found that the threshold concentration for lidocaine to cause rat thymocyte death ranged between 10 and 30 mM, which is close to our observation (Fig. 1). Gold et al.7 observed lidocaine-induced cell death after just a 4-min treatment of dorsal root ganglion neurons with an EC50 of 14 mM, which is close to the LD50 for lidocaine-induced cell death (approximately 14 mM, Table 2) in our SH-SY5Y cells. Bupivacaine and tetracaine at clinical concentrations were also found to cause death in human renal cells.26,45 Using the collapse of chick embryonic growth cones and neurites in cultured neurons as a measure of toxicity, Kasaba et al.33 compared toxicity of different LAs (in a 30-min treatment) and found the toxicity order procaine ≤ mepivacaine < ropivacaine < bupivacaine < lidocaine < tetracaine < dibucaine. We found the following order of the cell-killing potency (based on LD50, at which 50% of the cells were killed, Table 2): procaine ≤ mepivacaine < lidocaine < chloroprocaine < ropivacaine < bupivacaine. Similar to Kasaba et al., we also found procaine and mepivacaine were least potent in killing cells but, unlike their findings, we found lidocaine was much less potent in killing cells than bupivacaine. Morphologically, we also observed that lidocaine caused more severe retraction of the neurite-like cell extensions in addition to the cell rounding-up effect (Figs. 3D–F) than bupivacaine, both at a concentration around the LD50 (Figs. 3D–F versus Figs. 3G–I). The examination of detached cells that we collected at the end of the 10-min LA treatment demonstrated many live cells after incubation with 1 or 5 mM lidocaine, but almost only dead cells in samples taken after incubation with 10 mM lidocaine (data not shown). As for bupivacaine, it caused less cell shape change and detachment, but the majority of collected detached cells after incubation with bupivacaine (0.5, 1 or 5 mM) appeared to be dead (data not shown). Thus, lidocaine and bupivacaine may exert their toxicity in different ways: lidocaine being more “effective” in causing cell retraction and detachment (possibly due to a disruptive effect on the cytoskeleton46,47) and bupivacaine more potent in killing cells. This might explain the different cytotoxic potency of lidocaine or bupivacaine found in studies using morphological indicators (collapse of chick embryonic growth cones and neurites in cultured neurons33) versus using cell viability as a measure as we have. Park et al.48 also used the MTT assay to assess the effect of LAs on the viability of a Schwann cell line. They studied a panel of LAs similar to ours over a concentration range of 0–1 mM but with an exposure of up to 18 h. Bupivacaine at 1 mM killed approximately 90% of the cells, whereas no cell death was observed with lidocaine, mepivicaine, and procaine.
Necrosis or Apoptosis?
In numerous in vitro studies, investigators have tried to determine the mechanisms underlying LA toxicity. LAs were found to cause increased permeability of the mitochondrial membrane, collapse of the mitochondrial membrane potential, and decrease in adenosine triphosphate production by either uncoupling of oxidative phosphorylation or inhibition of complex I of the mitochondrial respiratory chain.15–18,20–25 In recent years, some studies have investigated whether LAs can trigger apoptosis. In addition to their effect on mitochondrial bioenergetics, lidocaine, bupivacaine, and tetracaine were shown to have apoptotic effects, such as causing DNA-fragmentation and plasma membrane blebbing, activating caspases, and releasing cytochrome c from mitochondria.16,23,26–28 In the current study, we examined bupivacaine and lidocaine for their capacity to activate the caspase-3/-7 (two caspases responsible for proteolytic cleavage that lead to cell disassembly). We found that bupivacaine and lidocaine each activate caspases-3/-7, but in different manners (Figs. 5 and 6). The caspase activation by bupivacaine occurred with a time delay, suggesting that the bupivacaine-induced cell death determined after 10-min treatment was necrotic. Park et al.48 found that bupivacaine (0.5 mM) induced the generation of reactive oxygen species in a Schwann cell line after a treatment of 2–4 h followed by apoptosis triggering processes, including the activation of caspase-3. It is possible that the delayed activation of caspase-3/-7 by bupivacaine observed in the current study might also be attributed to a similar cause.
Lidocaine-induced cell death was mainly due to necrosis when the concentration was low. At a higher lidocaine concentration, although the caspases were activated immediately upon addition, as shown in Figure 6, apoptosis was unlikely the direct cause of cell death after the 10-min treatment because the morphological observation showed no visible cell disassembly even after 20-min (at 10 mM, Fig. 3F) treatment. No DNA-laddering (a common indicator for apoptosis) was found in cells treated with up to 40 mM for up to 120 min (data not shown). With a lidocaine treatment longer than 10 min and higher than 10 mM, apoptosis may have had no chance to manifest before cells died necrotically despite activation of caspases. A study done by Nishimura et al.44 also suggested that in the presence of mM-lidocaine (up to 30 mM in their study), the process of apoptosis cannot be completed. In another study, in the human histiocytic lymphoma cell line U937 Kamiya et al.28 found that lidocaine below 12 mM induced apoptosis but caused necrosis above 12 mM. These results seem to be contradictory to our finding that showed lidocaine below 10 mM did not activate caspase-3/-7. However, in the study of Kamiya et al., cells were treated with lidocaine for 24 h, whereas our cells were only treated for 10 min. Our cells treated with 10 mM lidocaine showed a delayed caspase activation 3 h after a 10-min treatment, which is then consistent with the observation made by Kamiya et al. The same group also found that collapse of mitochondrial membrane potential (Δψm) was independent of caspase activation, which suggests multiple pathways leading to acute or delayed cell death. Because of their capacity to activate caspases, bupivacaine and lidocaine at clinical concentrations may also be able to exert delayed damaging effects on the cells through apoptotic pathway(s).
In spinal or epidural anesthesia, the parts of the nerve tissue bearing the brunt of the high initial LA concentrations are the axons in the nerve roots of the cauda equina. Under certain circumstances, such as poor mixing and sacral pooling, the potential for high perineural concentrations of drug may increase as well as the risk of nerve injury or functional disturbance. In addition, although nerves in vivo are usually surrounded by myelin and other types of tissues, which may buffer some damaging effects exerted by LAs, prolonged or continuous anesthesia may eventually “use up” such buffering capacity and result in nerve or neuronal damage. The results of this study demonstrate that 1) LAs can cause severe damage to neuronal cells and their extensions and 2) the duration of an LA treatment seems to matter in terms of whether this drug at a given concentration may cause neurological consequences.
In addition, different LAs may exert their toxic effect in different ways. Lidocaine has been reported to have a degenerative effect causing physical reduction of the myelin,15 whereas bupivacaine has been found to kill the Schwann cells more effectively than lidocaine.48 It is possible that lidocaine has a greater degenerative effect on myelin layers, which increases the chance for the nerves being exposed to LA and resulting in certain disturbance (not necessarily injury) to the nerves. In vivo, bupivacaine’s toxicity might be largely damped in killing the nerve-surrounding Schwann cells, so that less nerve disturbance is registered during a timely short application. Prolonged or continuous bupivacaine treatment may still cause nerve damage after the “buffering capacity” of the myelin has been exhausted.
Role of Cytosolic Ca2+ ([Ca2+]i) in LA-Induced Cell Death
Various studies found that LAs can cause an increase in the cytosolic Ca2+ level ([Ca2+]i),5,7,14,23,49 and suggested that the [Ca2+]i-increase was involved in LA cytotoxicity and might trigger LA-induced apoptosis. In our previous study and the current study, we found four amide LAs (bupivacaine, ropivacaine, mepivacaine, and lidocaine, up to 2.3 mM38) and two ester LAs (procaine and chloroprocaine, up to 10 mM, Fig. 8) inhibited both depolarization-evoked (by high K+) and muscarinic receptor activation-evoked (by carbachol) [Ca2+]i-increase in a concentration-dependent manner. Because the depolarization-evoked [Ca2+]i-increase is primarily attributed to Ca2+-induced Ca2+ release from the caffeine-sensitive intracellular store after Ca2+-influx through voltage-dependent Ca2+ channels and the carbachol-evoked [Ca2+]i-increase is primarily mediated through Ca2+ released from the IP3-sensitive store,42,50 the LA-induced inhibition of evoked-[Ca2+]i-increase suggests that the LAs prevented intracellular Ca2+ mobilization over the timeframe studied (5 min pretreatment with LA followed by 20–30 min for the reaction).
The Ca2+-evoked [Ca2+]i-increase was not affected by any of the amide LAs (bupivacaine, ropivacaine, mepivacaine, or lidocaine).38 Procaine and chloroprocaine in the current study produced a significant increase in Ca2+-evoked [Ca2+]i only within 1.5–4 mM but not at lower or higher concentrations (Fig. 8). Based on our previous and current studies, cytosolic Ca2+ does not appear to have a direct and immediate involvement in LA-induced cell death within the concentration range studied (up to 10 mM) and during short (≤10 min) exposure times. Because none of the LAs used in our studies directly caused concentration-dependent Ca2+ mobilization, the triggering role of an [Ca2+]i-increase in LA-induced apoptosis appears unlikely. However, our previous data38 and this study suggest that LAs can change the Ca2+ homeostasis through their effects on Ca2+ entry and on intracellular Ca2+ mobilization, which in turn may indirectly affect cell viability through different mechanism(s). For example, Sattler et al.51 reported that Ca2+ neurotoxicity is a function of the Ca2+ influx pathway, not the intracellular Ca2+ load itself.
These six LAs possess distinct chemical structures and properties. The octanol/buffer partition coefficients (at pH 7.4°C and 36°C) for bupivacaine (560) are much higher than for the other five LAs studied (versus 3.1 for procaine, 17.4 for chloroprocaine, 42 for mepivacaine, 110 for lidocaine at pH 7.4°C and 36°C, and 115 for ropivacaine at pH 7.4°C and 25°C).12 This suggests that bupivacaine accumulates in a much larger amount in membranes and may distribute and/or remain longer within the plasma membrane to exert more interference with the membrane-bound ion channels, receptors, or other functional proteins, as well as to cause more lipid membrane permeability changes.52 This might contribute to the higher toxicity of bupivacaine that we have observed after a short exposure time. It is also noteworthy that the least toxic LA, procaine, among the six LAs studied has the smallest octanol/buffer partition coefficients (3.1 at pH 7.4°C and 36°C).
In summary, all six LAs exert cytotoxicity in a concentration-dependent manner, but their capacity to kill cells differs. Bupivacaine seems to be the most toxic LA. It not only has the highest potency for killing cells, but was also able to kill all available cells with an increasing concentration. Lidocaine also can kill all cells with an increasing concentration. In contrast, none of the other four LAs, ropivacaine, mepivacaine, procaine, and chloroprocaine, killed all cells even at the highest test concentrations studied, and their concentration-dependency curves were shallower than those of bupivacaine and lidocaine. Bupivacaine and lidocaine can activate caspase-3/-7 leading to apoptosis, but their patterns of activation are different. Although bupivacaine-induced caspase activation occurred slowly, lidocaine-induced caspase activation occurred more rapidly at concentrations larger than 10 mM. At high concentrations, lidocaine causes an immediate caspase activation, but cells cannot survive long enough to allow apoptosis to be complete.
1.Pollock JE. Transient neurologic symptoms: etiology, risk factors, and management. Reg Anesth Pain Med 2002;27:581–6
2.Johnson ME. Potential neurotoxicity of spinal anesthesia with lidocaine. Mayo Clin Proc 2000;75:921–32
3.Selander D. Neurotoxicity of local anesthetics: animal data. Reg Anesth 1993;18:461–8
4.Pollock JE. Neurotoxicity of intrathecal local anaesthetics and transient neurological symptoms. Best Pract Res Clin Anaesthesiol 2003;17:471–84
5.Johnson ME, Saenz JA, DaSilva AD, Uhl CB, Gores GJ. Effect of local anesthetic on neuronal cytoplasmic calcium and plasma membrane lysis (necrosis) in a cell culture model. Anesthesiology 2002;97:1466–76
6.Friederich P, Schmitz TP. Lidocaine-induced cell death in a human model of neuronal apoptosis. Eur J Anaesthesiol 2002;19:564–70
7.Gold MS, Reichling DB, Hampl KF, Drasner K, Levine JD. Lidocaine toxicity in primary afferent neurons from the rat. J Pharmacol Exp Ther 1998;285:413–21
8.Strichartz GR, Ritchie, JM. The action of local anesthetics on ion channels of excitable tissues. In: Strichartz GR, ed. Local anesthetics. Berlin: Springer-Verlag, 1987:21–53
9.Moorman JR. Sodium channels. In: Yaksh TL, Lynch C III, Zapol WM, Maze M, Biebuyck JF, Saidman LJ, ed. Anesthesia biologic foundations, Philadelphia, New York: Lippincott-Raven Publishers, 1998:145–62
10.Rossner KL, Freese KJ. Bupivacaine inhibition of L-type calcium current in ventricular cardiomyocytes of hamster. Anesthesiology 1997;87:926–34
11.Liu BG, Zhuang XL, Li ST, Xu GH, Brull SJ, Zhang JM. Effects of bupivacaine and ropivacaine on high-voltage-activated calcium currents of the dorsal horn neurons in newborn rats. Anesthesiology 2001;95:139–43
12.Strichartz GR, Sanchez V, Arthur GR, Chafetz R, Martin D. Fundamental properties of local anesthetics. II. Measured octanol:buffer partition coefficients and pKa values of clinically used drugs. Anesth Analg 1990;71:158–70
13.Butterworth JFt, Strichartz GR. Molecular mechanisms of local anesthesia: a review. Anesthesiology 1990;72:711–34
14.Johnson ME, Gores GJ, Uhl CB, Sill JC. Cytosolic free calcium and cell death during metabolic inhibition in a neuronal cell line. J Neurosci 1994;14:4040–9
15.Lirk P, Haller I, Myers RR, Klimaschewski L, Kau YC, Hung YC, Gerner P. Mitigation of direct neurotoxic effects of lidocaine and amitriptyline by inhibition of p38 mitogen-activated protein kinase in vitro and in vivo. Anesthesiology 2006;104:1266–73
16.Unami A, Shinohara Y, Ichikawa T, Baba Y. Biochemical and microarray analyses of bupivacaine-induced apoptosis. J Toxicol Sci 2003;28:77–94
17.Schonfeld P, Sztark F, Slimani M, Dabadie P, Mazat JP. Is bupivacaine a decoupler, a protonophore or a proton-leak-inducer? FEBS Lett 1992;304:273–6
18.Grouselle M, Tueux O, Dabadie P, Georgescaud D, Mazat JP. Effect of local anaesthetics on mitochondrial membrane potential in living cells. Biochem J 1990;271:269–72
19.Garlid KD, Nakashima RA. Studies on the mechanism of uncoupling by amine local anesthetics. Evidence for mitochondrial proton transport mediated by lipophilic ion pairs. J Biol Chem 1983;258:7974–80
20.Grant RL, Acosta D Jr. A digitized fluorescence imaging study on the effects of local anesthetics on cytosolic calcium and mitochondrial membrane potential in cultured rabbit corneal epithelial cells. Toxicol Appl Pharmacol 1994;129:23–35
21.Terada H, Shima O, Yoshida K, Shinohara Y. Effects of the local anesthetic bupivacaine on oxidative phosphorylation in mitochondria. Change from decoupling to uncoupling by formation of a leakage type ion pathway specific for H+ in cooperation with hydrophobic anions. J Biol Chem 1990;265:7837–42
22.Sztark F, Malgat M, Dabadie P, Mazat JP. Comparison of the effects of bupivacaine and ropivacaine on heart cell mitochondrial bioenergetics. Anesthesiology 1998;88:1340–9
23.Irwin W, Fontaine E, Agnolucci L, Penzo D, Betto R, Bortolotto S, Reggiani C, Salviati G, Bernardi P. Bupivacaine myotoxicity is mediated by mitochondria. J Biol Chem 2002;277:12221–7
24.Floridi A, Barbieri R, Pulselli R, Fanciulli M, Arcuri E. Effect of the local anesthetic bupivacaine on the energy metabolism of Ehrlich ascites tumor cells. Oncol Res 1994;6:593–601
25.Floridi A, Di Padova M, Barbieri R, Arcuri E. Effect of local anesthetic ropivacaine on isolated rat liver mitochondria. Biochem Pharmacol 1999;58:1009–16
26.Lee HT, Xu H, Siegel CD, Krichevsky IE. Local anesthetics induce human renal cell apoptosis. Am J Nephrol 2003;23:129–39
27.Johnson ME, Uhl CB, Spittler KH, Wang H, Gores GJ. Mitochondrial injury and caspase activation by the local anesthetic lidocaine. Anesthesiology 2004;101:1184–94
28.Kamiya Y, Ohta K, Kaneko Y. Lidocaine-induced apoptosis and necrosis in U937 cells depending on its dosage. Biomed Res 2005;26:231–9
29.Hodgson PS, Liu SS, Batra MS, Gras TW, Pollock JE, Neal JM. Procaine compared with lidocaine for incidence of transient neurologic symptoms. Reg Anesth Pain Med 2000;25:218–22
30.Hodgson PS, Neal JM, Pollock JE, Liu SS. The neurotoxicity of drugs given intrathecally (spinal). Anesth Analg 1999;88:797–809
31.YaDeau JT, Liguori GA, Zayas VM. The incidence of transient neurologic symptoms after spinal anesthesia with mepivacaine. Anesth Analg 2005;101:661–5, table of contents
32.Radwan IA, Saito S, Goto F. Growth cone collapsing effect of lidocaine on DRG neurons is partially reversed by several neurotrophic factors. Anesthesiology 2002;97:630–5
33.Kasaba T, Onizuka S, Takasaki M. Procaine and mepivacaine have less toxicity in vitro than other clinically used local anesthetics. Anesth Analg 2003;97:85–90, table of contents
34.Boselli E, Debon R, Allaouchiche B, Chassard D, Portokalian J. Induction of apoptosis by local anesthetics. Anesthesiology 2005;95:A984
35.Mosmann T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods 1983;65:55–63
36.Hollmann MW, Durieux ME. Local anesthetics and the inflammatory response: a new therapeutic indication? Anesthesiology 2000;93:858–75
37.Grynkiewicz G, Poenie M, Tsien RY. A new generation of Ca2+
indicators with greatly improved fluorescence properties. J Biol Chem 1985;260:3440–50
38.Xu F, Garavito-Aguilar Z, Recio-Pinto E, Zhang J, Blanck TJJ. Local anesthetics modulate neuronal calcium signaling through multiple sites of action. Anesthesiology 2003;98:1139–46
39.Forsythe ID, Lambert DG, Nahorski SR, Lindsdell P. Elevation of cytosolic calcium by cholinoceptor agonists in SH-SY5Y human neuroblastoma cells: estimation of the contribution of voltage-dependent currents. Br J Pharmacol 1992;107:207–14
40.Toselli M, Tosetti P, Taglietti V. Functional changes in sodium conductances in the human neuroblastoma cell line SH-SY5Y during in vitro differentiation. J Neurophysiol 1996;76:3920–7
41.Tosetti P, Taglietti V, Toselli M. Functional changes in potassium conductances of the human neuroblastoma cell line SH-SY5Y during in vitro differentiation. J Neurophysiol 1998;79:648–58
42.Xu F, Zhang J, Recio-Pinto E, Blanck TJ. Halothane and isoflurane augment depolarization-induced cytosolic CA2+
transients and attenuate carbachol-stimulated CA2+
transients. Anesthesiology 2000;92:1746–56
43.Sakura S, Bollen AW, Ciriales R, Drasner K. Local anesthetic neurotoxicity does not result from blockade of voltage-gated sodium channels. Anesth Analg 1995;81:338–46
44.Nishimura Y, Kanada A, Yamaguchi JY, Horimoto K, Kobayashi M, Tatsuishi T, Kanemaru K, Ueno SY, Oyama Y. Cytometric analysis of lidocaine-induced cytotoxicity: a model experiment using rat thymocytes. Toxicology 2006;218:48–57
45.Lee HT, Krichevsky IE, Xu H, Ota-Setlik A, D’Agati VD, Emala CW. Local anesthetics worsen renal function after ischemia-reperfusion injury in rats. Am J Physiol Renal Physiol 2004;286:F111–9
46.Haschke RH, Byers MR, Fink BR. Effects of lidocaine on rabbit brain microtubular protein. J Neurochem 1974;22:837–43
47.Nicolson GL, Smith JR, Poste G. Effects of local anesthetics on cell morphology and membrane-associated cytoskeletal organization in BALB/3T3 cells. J Cell Biol 1976;68:395–402
48.Park CJ, Park SA, Yoon TG, Lee SI, Yum KW, Kim HJ. Bupivacaine induces apoptosis via ROS in the Schwann cell line. J Dent Res 2005;84:852–7
49.Arai Y, Kondo T, Tanabe K, Zhao QL, Li FJ, Ogawa R, Li M, Kasuya M. Enhancement of hyperthermia-induced apoptosis by local anesthetics on human histiocytic lymphoma U937 cells. J Biol Chem 2002;277:18986–93
50.Garavito-Aguilar ZV, Recio-Pinto E, Corrales AV, Zhang J, Blanck TJ, Xu F. Differential thapsigargin-sensitivities and interaction of Ca2+
stores in human SH-SY5Y neuroblastoma cells. Brain Res 2004;1011:177–86
51.Sattler R, Tymianski M, Feyaz I, Hafner M, Tator CH. Voltage-sensitive calcium channels mediate calcium entry into cultured mammalian sympathetic neurons following neurite transection. Brain Res 1996;719:239–46
52.Pardo L, Blanck TJ, Recio-Pinto E. The neuronal lipid membrane permeability was markedly increased by bupivacaine and mildly affected by lidocaine and ropivacaine. Eur J Pharmacol 2002;455:81–90
© 2009 International Anesthesia Research Society
53.Weiss JN. The Hill equation revisited: uses and misuses. Faseb J 1997;11:835–41