Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Colchicine is a plant-derived alkaloid that binds to tubulin and depolymerizes microtubules (Osborn and Weber, 1976; Walker and Whitfield, 1985), disrupts axonal transport (Karlsson and Sjostrand, 1969; Fink et al., 1973; Wooten et al., 1975) and inhibits mitosis (Wilson and Friedkin, 1966, 1967; Wang et al., 1975). This compound has been used clinically in the treatment of gout (Hastie, 1991) and has also been used extensively as an experimental tool to characterize cellular processes that involve microtubule structure and axoplasmic transport (Chiaia et al., 1996; Schmalz et al., 1996; Tandon et al., 1996; Saib et al., 1997). Numerous studies have also used colchicine as a neurotoxin to cause lesions in discrete brain regions, such as the dentate gyrus (Brady et al., 1992; Newell et al., 1993; Tandon et al., 1994; Gobbi et al., 1996) and the septum (Peterson and McGinty, 1988; Gilbert and Peterson, 1991), and several recent studies have suggested that colchicine neurotoxicity may model some of the neuropathological aspects of Alzheimer's dementia (Nakagawa et al., 1987; Mattson, 1992). In all of these cases, the mechanism of colchicine action has been attributed to its inhibition of tubulin binding and subsequent disruption of processes that depend on the integrity of cytoskeletal architecture. However, the mechanism(s) underlying these diverse actions of colchicine have not been determined experimentally.

Several studies have also demonstrated that colchicine can significantly inhibit the function of several ion channels, including voltage-gated sodium (Matsumoto et al., 1984) and calcium (Johnson and Byerly, 1993) channels and ligand-gated nicotinic (Hardwick and Parsons, 1995) and GABA_A receptor-gated ion channels (Whatley et al., 1994; Whatley and Harris, 1996). The mechanism(s) through which colchicine exerts these effects are unclear, but have been hypothesized to involve the depolymerizing actions of this drug.

We have characterized the interaction of colchicine and other microtubule disrupting agents with recombinant GABA_A receptors in a stable transfection system to further delineate the mechanism(s) through which microtubule depolymerization might inhibit GABA_A receptor function. Here we report that colchicine appears to have an additional direct inhibitory effect on GABA_A receptor function, one that is independent of the microtubule disrupting actions of this drug. Unlike the indirect inhibition of GABA_A receptor function that appears to be related to effects on microtubules (Whatley et al., 1994; Whatley and Harris, 1996), this direct effect is rapid and competitive, and it is not mimicked or occluded by other microtubule disrupting agents.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture conditions. Stable transfection of mouse L (tk) cells with human GABA_A receptor subunits was carried out as described previously (Hadingham et al., 1992). Expression was controlled by a dexamethasone-sensitive promoter. Cells were grown to confluence in 75-ml flasks and then plated onto 12-mm-diameter glass coverslips coated with poly-L-lysine with Dulbecco's Modified Eagle's Medium (DME/high glucose, Hyclone, Logan, UT) supplemented with 10% fetal bovine serum (Gemini Bio-Product, Calabasas, CA), 100 U/ml penicillin, 0.1 mg/ml streptomycin (Sigma Chemicals, St. Louis, MO) and 2 mM L-glutamine (Dexter CO/Gibco Labs Division, Grand Island, NY). Cells were grown on coverslips for 24 to 48 hr at 37°C/5% CO₂, treated with 1 µM dexamethasone (Sigma Chemical, St. Louis, MO) and then grown for an additional 3 to 6 days.

Electrophysiology. Coverslips were transferred to a recording chamber perfused with a HEPES-buffered external solution (in mM): NaCl, 130; KCl, 5; CaCl₂, 2; MgCl₂, 1; HEPES, 10; D-glucose, 11; pH adjusted to 7.4 with NaOH; 300-310 mOsm. Whole-cell patch recordings were made from individual cells with borosilicate glass electrodes (1.5 mm outer diameter, 0.86 mm internal diameter, Sutter Instruments, Novato, CA). The intracellular recording solution contained (in mM): KCl, 130; CaCl₂, 0.1; EGTA, 1.0; Mg-ATP, 2; Tris-GTP, 0.2; HEPES, 10; pH adjusted with KOH; 275-285 mOsm. Currents were recorded with an Axopatch 200 amplifier (Axon Instruments, Foster City, CA), low pass filtered with a 4 pole Bessel Filter (2 KHz) and analyzed on- and off-line with software developed in this laboratory. All external solutions were gravity fed from glass 12-cc syringe reservoirs through a two-barrel flow tube array (450 µm internal diameter, Polymicro Tech. Inc., Phoenix, AZ). The flow tube array was affixed to a piezoelectric double ceramic plate (Bimorph) (Morgan Matroc, Inc., Bedford, OH). By applying a voltage across the Bimorph, the solution interface across the surface of the cell being recorded could be rapidly exchanged (approximately 100 msec half-time). Cells were voltage-clamped at -40 to -60 mV and inward currents were evoked by 2- or 3-sec applications of GABA at an interval of 30 or 60 sec. Unless otherwise indicated, all drugs were applied only during the GABA pulses.

[³H]SR 95531 binding. For analysis of [³H]SR 95531 binding, cells stably transfected with recombinant subunits were harvested and prepared as described previously (Klein et al., 1994). Cells were homogenized in assay buffer (in mM): NaCl, 145; KCl, 5; MgCl₂, 1; CaCl₂, 1; D-glucose, 10; HEPES, 10; pH adjusted to 7.5 with Tris base, pelleted twice at 100,000 × g for 15 min and resuspended in assay buffer. [³H]SR 95531 binding was performed essentially as described by Buck and Harris (1990). Aliquots of cell membrane (150-250 µg protein) were incubated with 3 nM [³H]SR 95531, 3 to 300 nM nonradioactive SR 95531 plus/minus 100 µM colchicine in a final volume of 0.3 ml for 30 min at 34°C. Nonspecific binding was determined in the presence of 100 µM GABA.

Drugs. Unless otherwise stated, all reagents used in the electrophysiological experiments were purchased from Sigma Chemicals (St. Louis, MO). The reagents used in the intracellular recording solution were purchased from Fluka (Buchs, Switzerland), except Mg-ATP and Tris-GTP. Taxol and vinblastine were purchased from Calbiochem (La Jolla, CA), and flumazenil from Hoffman-LaRoche (Nutley, NJ). Flumazenil and all microtubule disrupting agents were made up as concentrated stock solutions in 100% DMSO, stored at -20°C and diluted to their final concentrations immediately before each experiment. The final concentration of DMSO never exceeded 0.1%, a concentration that had no effect on GABA currents in this study.

Statistics. Drug effects were quantified as the percent change in GABA-evoked current amplitude relative to the mean of control and washout values. Statistical analyses of drug effects were performed by two-tailed Student's paired and unpaired t tests as indicated, with a minimal level of significance of P < .05. pA₂ values for colchicine inhibition were estimated from the electrophysiological and binding assays with the dose ratio of EC₅₀ concentrations of GABA in the presence and absence of 100 µM colchicine.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Colchicine rapidly antagonizes GABA-evoked currents. All recordings were carried out on murine L(tk) cells stably transfected with human recombinant 122L GABA_A receptor subunits. Three to six days after dexamethasone induction of GABA_A receptor expression, rapid application of GABA reliably generated dose-dependent inward currents in cells voltage-clamped at -40 to -60 mV. These responses reversed at the chloride equilibrium potential (0 mV under our recording conditions), were completely antagonized by 20 µM bicuculline (data not shown) and were therefore mediated by the activation of GABA_A receptors.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Acute application of colchicine inhibits GABAA receptor-gated chloride currents. The time course of 100 µM colchicine inhibition of currents evoked by 2 µM GABA with () and without colchicine preincubation is illustrated at the bottom. Traces at the top are averages of three to seven sweeps recorded at the times indicated by corresponding letters on the graph. The solid bars above the averages represent the time during which GABA was applied via the flow tubes, and the unfilled bars indicate the duration of colchicine application. Note that filled circles indicate application of 2 µM GABA and unfilled circles indicate concurrent application of 100 µM colchicine and 2 µM GABA. The apparent slow onset of the effects of colchicine on the GABA response reflects the time required to change the solutions in the flow tubes; once this has occurred, colchicine can antagonize the GABA response within 100 to 200 msec of its application (e.g., in D; see also fig. 2).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Rapid time course of colchicine and bicuculline inhibition of GABA-evoked currents. The time course of the reduction in inward current elicited by a 2-sec application of 100 µM colchicine (A) or 1 µM bicuculline (B) during continuous application of 500 nM GABA is shown. Note that 500 nM GABA was present throughout the entire period and that both the onset and offset of the inhibition were identical for both drugs.

Colchicine inhibition of GABA-evoked currents is dose-dependent and competitive. A separate series of experiments were carried out to characterize the pharmacological characteristics of the colchicine antagonism of GABA_A receptor-mediated responses. With use of the concurrent application protocol, the colchicine inhibition of currents evoked by 2 µM GABA was dose-dependent, with an EC₅₀ of 56 µM and a Hill coefficient of 1.2 (fig. 3). The lowest concentration of colchicine that significantly inhibited currents evoked by 2 µM GABA was 10 µM (14 ± 2%, P < .01, paired t test, n = 7), and a concentration of 1 mM colchicine almost completely antagonized these currents.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Dose dependence of colchicine inhibition of GABA-evoked currents. The mean inhibition of currents evoked by 2 µM GABA by 1 to 1000 µM colchicine is shown; each point is the mean ± S.E.M. obtained from between 4 and 13 cells. Traces at the top of the figure are averages of three to six sweeps recorded from a single representative cell during acute application of 10 µM, 50 µM and 1 mM colchicine (COL). CTL = control.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Colchicine antagonizes GABAA receptor function in a competitive manner. (A) Representative traces illustrate the effects of 100 µM colchicine (COL) on currents evoked by 500 nM, 2 µM and 10 µM GABA. Note that the magnitude of the colchicine inhibition decreases with increasing GABA concentrations. (B) Effect of colchicine on the GABA dose-response curve. Each point represents the mean ± S.E.M. from 3 to 13 cells tested at each concentration. In some cases, error bars fall within the symbols.

View this table: [in this window] [in a new window]   TABLE 1 Effect of colchicine on [3H]SR 95531 binding in L(tk) cells expressing 122L GABAA receptor subunitsa

The colchicine inhibition does not involve the benzodiazepine binding site. -Lumicolchicine is a structural analog of colchicine that has no effect on microtubules (Wilson and Friedkin, 1967; Walker and Whitfield, 1985). A previous report demonstrated that this analog potentiated GABA_A receptor function by a direct interaction with the benzodiazepine site on GABA_A receptors (Mihic et al., 1994). We therefore determined whether the antagonism of GABA_A receptor function observed with colchicine was mediated by an interaction with this same site. Under the recording conditions used in this study, acute application of 100 µM -lumicolchicine significantly potentiated currents evoked by 1 µM GABA in cells expressing 122L GABA_A receptor subunits (fig. 5). The mean potentiation was 147 ± 4% (P < .001, paired t test, n = 5) and this enhancement was completely blocked by flumazenil. Colchicine (100 µM) significantly inhibited GABA currents, as observed in the initial experiments, however flumazenil had no effect on the colchicine inhibition (fig. 5).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Colchicine inhibition does not involve the benzodiazepine binding site. (A) Representative traces illustrating the effect of 1 µM flumazenil on the enhancement of GABA responses induced by 100 µM -lumicolchicine and on the antagonism of GABA responses induced by 100 µM colchicine. (B) Bar graph summarizing the effect of 100 µM -lumicolchicine and 100 µM colchicine on GABA-evoked currents in the presence and absence of 1 µM flumazenil. n = 3-7 cells/group. Note that flumazenil completely antagonized the -lumicolchicine-induced potentiation of GABA-evoked currents but had no effect on the colchicine-mediated inhibition.

Other microtubule disrupting agents do not mimic or occlude the acute actions of colchicine. The acute effects of other microtubule disrupting agents were also tested by the same protocol used for colchicine. We tested the acute actions of 10 µM vinblastine, 1 µM taxol and 10 µg/ml nocodazole on currents evoked by 2 µM GABA. None of these compounds had any significant effect on GABA-evoked currents (fig. 6). We also tested another inactive analog of colchicine, -lumicolchicine, which was also without effect on GABA-evoked currents.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Other microtubule disrupting compounds have no acute effect on GABA-evoked currents. The bar graph summarizes the acute effect of 100 µM colchicine (COL), 100 µM -lumicolchicine (-Lumi), 10 µM vinblastine (VIN), 1 µM taxol and 10 µg/ml nocodazole (NOCOD) on currents evoked by 2 µM GABA. n = 5-12 cells/group.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Disruption of microtubules does not occlude the acute effect of colchicine. The bar graph summarizes the acute effect of colchicine under control conditions (white bar), and after incubation of cells for 1 to 4 hr with 100 µM colchicine (COL), 10 µM vinblastine (VIN), 1 µM taxol or 10 µg/ml nocodazole (NOCOD). n = 4-12 cells/group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The results of this study demonstrate that colchicine can act as a competitive antagonist at GABA_A receptors through a mechanism that appears to be distinct from the effects of this drug on microtubule depolymerization. Two studies have previously demonstrated that colchicine and other microtubule depolymerizing agents can inhibit GABA_A receptor function in cerebral cortical microsacs, in Xenopus oocytes transiently transfected with recombinant GABA_A receptor subunits (Whatley et al., 1994) and in cells stably transfected with GABA_A receptor subunits (Whatley et al., submitted). In both of these studies, this inhibition was attributed to the depolymerizing actions of these compounds on microtubule structure. However the time course and characteristics of the inhibition were significantly different. In microsacs and oocytes, 100 µM colchicine inhibited GABA_A receptor function in as little as 3 to 20 sec and the inhibition appeared to be competitive. Other microtubule depolymerizing drugs also inhibited GABA_A receptor-mediated chloride flux, although their actions were only characterized at 30 to 60 min. In contrast, in chloride flux studies of the stably transfected cell line characterized in the present study, 100 µM colchicine required 60 min to significantly inhibit GABA_A receptor function, and this inhibition was noncompetitive, having no effect on GABA_A receptor affinity. Moreover, using immunolabeling of tubulin, it was shown that colchicine did not appreciably depolymerize microtubule structure for at least 30 min. Based on the results of these studies, it is clear that colchicine (but not the other depolymerizing agents) has a direct, competitive inhibitory interaction with the GABA_A receptor that is apparent within 100 msec of colchicine application, and does not involve the microtubule depolymerizing actions of this drug. In contrast, the noncompetitive inhibitory effects of colchicine and other similar agents that are observed at longer treatment intervals (60 minutes) are likely to involve a mechanism related to the depolymerizing actions of these drugs.

Several lines of evidence indicate that the inhibition of GABA_A receptor function observed in this study did not involve the microtubule depolymerizing actions of colchicine. First, the time course of the colchicine inhibition of GABA-evoked currents is inconsistent with the microtubule depolymerizing actions of this drug (Whatley and Harris, 1996; Whatley et al., submitted). With concurrent application of colchicine and GABA, the inhibition developed with a half-time of approximately 140 msec, whereas colchicine-mediated microtubule depolymerization in these cells required at least 30 min (Whatley et al., submitted). In addition, within the temporal resolution of the drug delivery system used in this study, no difference was observed between colchicine antagonism of GABA-evoked currents and that produced by equieffective concentrations of bicuculline methiodide, a competitive GABA_A receptor antagonist. Second, the effects of colchicine in the present study were characterized by a competitive interaction with GABA, whereas the inhibition attributed to depolymerization produces a noncompetitive decrease in GABA efficacy with no change in receptor affinity for GABA (Whatley et al., submitted). Third, the rapid inhibitory effect of colchicine on GABA-evoked currents was not observed with other microtubule disrupting agents like nocodazole, vinblastine and taxol. In contrast, all three of these compounds have been shown to significantly inhibit GABA_A receptor function when applied for durations that are sufficient to depolymerize microtubule assembly (Whatley et al., 1994; Whatley and Harris, 1996). Finally, prolonged incubation of cells in colchicine or other microtubule depolymerizing drugs for durations that were long enough to fully depolymerize microtubules in these cells (Whatley and Harris, 1996; Whatley et al., submitted) had no effect on the rapid colchicine inhibition of GABA_A receptor function, which again suggests the independence of these two actions.

The results of this study suggest that colchicine is a competitive antagonist at GABA_A receptors, an action that is independent from the noncompetitive colchicine inhibition associated with the microtubule depolymerizing actions of this drug. Presently, the mechanism by which colchicine exerts this effect is not known. Colchicine may inhibit GABA_A receptor function by direct interaction with the GABA binding site or alternatively it may modulate GABA affinity via an allosteric interaction with another site on the GABA_A receptor or even with some protein closely associated with it. This latter possibility, that colchicine is acting as an allosteric, apparent competitive antagonist, is supported by the observation that colchicine had no effect on GABA_A receptors reconstituted into proteoliposomes (Whatley et al., 1994) where any proteins that might normally associate with these receptors in neurons would not be present. There is evidence that a structural analog of colchicine, -lumicolchicine, can interact directly with the benzodiazepine binding site on GABA_A receptors (Mihic et al., 1994). This analog does not bind tubulin and is therefore often used as a control for the microtubule disrupting actions of colchicine. However, -lumicolchicine was shown to compete with flunitrazepam binding in cerebral cortical microsacs and potentiate muscimol-stimulated chloride flux in microsacs and GABA-evoked currents in Xenopus oocytes stably transfected with recombinant GABA_A receptor subunits (Mihic et al., 1994). In addition, the potentiation of GABA-evoked currents was completely antagonized by flumazenil, a competitive benzodiazepine antagonist. In the present experiments, similar effects of -lumicolchicine on GABA-evoked currents were observed, but the acute inhibitory effect of colchicine was not antagonized by flumazenil. This suggests that colchicine and -lumicolchicine, despite being structural analogs, do not interact with the same site on the GABA_A receptor complex.

In preliminary experiments, continuous application of microtubule disrupting drugs such as nocodozole and taxol for up to 30 min did not significantly inhibit GABA-evoked currents. Similar treatments with colchicine did inhibit GABA_A receptor function; however, this inhibition could all be attributed to the acute, competitive inhibition that appears to be independent of colchicine's microtubule depolymerizing effects. In contrast, a 20-min continuous application of vinblastine did significantly inhibit GABA responses (Whatley et al., submitted). These disparate findings may reflect differences in the depolymerization kinetics of these compounds under our experimental conditions; however, incubating cells for up to 4 hr with these compounds, an interval sufficient to ensure maximal microtubule depolymerization (Whatley and Harris, 1996; Whatley et al., submitted), had no effect on the acute inhibitory action of colchicine.

Colchicine has been used as an experimental tool in numerous studies to characterize the role of microtubule assembly in various aspects of cellular function (Chiaia et al., 1996; Schmalz et al., 1996; Tandon et al., 1996; Saib et al., 1997) and as a selective neurotoxin in studies of epilepsy (Lee and Hong, 1990; Barnes and Mitchell, 1993; Gobbi et al., 1996) and Alzheimer's disease (Nakagawa et al., 1987; Mattson, 1992). The mechanism underlying the neurotoxic effects of colchicine is not known, but has usually been attributed to its ability to disrupt microtubule assembly, because structural analogs of colchicine, such as - and -lumicolchicine, that do not bind tubulin, do not mimic the neurotoxic effects of colchicine.

Our findings suggest that at least some of the neurotoxic actions of colchicine may be a result of direct inhibition of GABA_A receptor function. This is particularly likely in those studies in which colchicine was injected directly into specific brain regions to induce seizures (Lee and Hong, 1990; Barnes and Mitchell, 1993; Gobbi et al., 1996). Other competitive GABA_A receptor antagonists can elicit seizures in vivo and frequently have been used in models of epileptogenesis (Uemura and Kimura, 1988; Cataltepe et al., 1995; Reigel and Bourn, 1995). In addition, some of the neuropathological actions of colchicine on the phosphorylation state of the microtubule-associated protein, tau, are also observed with glutamate administration (Nakagawa et al., 1987; Mattson, 1992) and are not mimicked by other microtubule disrupting agents (Sygowski et al., 1993). These effects may stem from a generalized increase in neuronal excitability resulting, in the case of glutamate, from direct activation of (±)--amino-3-hydroxy-5-methylisoxazole-4-proprionic acid and/or N-methyl-D-aspartate receptors and, in the case of colchicine, from acute inhibition of GABA_A receptor function. Preliminary experiments from our laboratory have shown that colchicine, at concentrations used in this study, can significantly depress GABA_A receptor-mediated synaptic currents in rat hippocampal CA1 neurons. These findings further support the hypothesis that direct antagonism of neuronal GABA_A receptor function may underlie some of colchicine's neurotoxic actions.

Our results also suggest that the use of structural analogs of colchicine that do not bind tubulin may not be sufficient to demonstrate that effects of colchicine are mediated by microtubule depolymerization. Because these analogs either potentiate (-lumicolchicine) or have no acute effect (-lumicolchicine) on GABA_A receptor function, they would not mimic the effects of colchicine that reflect its acute inhibitory action on GABA_A receptor function. Because other microtubule disrupting agents, such as nocodazole, taxol and vinblastine, do not share this acute inhibitory effect on GABA_A receptor function, these agents may prove more useful in identifying the actions of colchicine that are mediated by depolymerization of microtubules. A detailed understanding of the cellular substrates that underlie colchicine-mediated neurotoxicity is needed to evaluate and interpret the results of studies that use this compound in animal models of complex disease states such as epilepsy and Alzheimer's disease. Further studies will be needed to determine the extent to which competitive antagonism of GABA_A receptors contributes to the neuropathological actions of colchicine.