Consumption of alcoholic beverages produces alterations in motor coordination and equilibrium that are responsible for millions of accidental deaths. Studies indicate that ethanol produces these alterations by affecting the cerebellum, a brain region involved in the control of motor systems. Purkinje neurons of the cerebellar cortex have been shown to be particularly important targets of ethanol. However, its mechanism of action at these neurons is poorly understood. We hypothesized that ethanol could modulate Purkinje neuron function by altering the excitatory input provided by the climbing fiber from the inferior olive, which evokes a powerful all-or-none response denoted as the complex spike. To test this hypothesis, we performed whole-cell patch-clamp electrophysiological and Ca2+ imaging experiments in acute slices from rat cerebella. We found that ethanol potently inhibits the late phase of the complex spike and that this effect is the result of inhibition of type-1 metabotropic glutamate receptor-dependent responses at the postsynaptic level. Moreover, ethanol inhibited climbing fiber long-term depression, a form of synaptic plasticity that also depends on activation of these metabotropic receptors. Our findings identify the climbing fiber→Purkinje neuron synapse as an important target of ethanol in the cerebellar cortex and indicate that ethanol significantly affects cerebellar circuits even at concentrations as low as 10 mm (legal blood alcohol level in the United States is below 0.08 g/dl = 17 mm).
Acute and chronic ethanol (EtOH) exposure produces profound impairments in motor functioning. Acutely, EtOH causes motor control alterations, ataxia, and vertigo, increasing the risk of accidental injuries, which are responsible for millions of deaths around the world (Dinh-Zarr et al., 2004). Individuals with lower sensitivity to the acute motor impairing effects of EtOH have an increased risk of developing alcohol dependence and abuse (Schuckit, 1994), and infants with subtle delays in motor coordination development may have an increased risk for subsequently developing alcoholism (Manzardo et al., 2005). Chronically, EtOH produces alterations in the mature nervous system leading to progressive impairments in stance and gait (Diamond and Messing, 1994), and developmental exposure induces long-lasting motor deficits that cause significant disability and require intensive rehabilitation (Klintsova et al., 1998, 2002; Autti-Ramo et al., 2002; Coffin et al., 2005). Thus, understanding the mechanism by which EtOH disrupts motor functioning is paramount.
Studies suggest that the motor impairing effects of EtOH are, in part, a consequence of cerebellar function alterations. The Purkinje neuron (PN) is an important target of EtOH (Siggins and French, 1979; Rogers et al., 1980; Sorensen et al., 1980; Basile et al., 1983), and evidence indicates that its effects are in part mediated by an increase in GABAergic transmission (Lin et al., 1994; Lee et al., 1995; Freund and Palmer, 1997; Yang et al., 2000). Recent studies indicate that EtOH increases GABAergic tone at molecular layer interneuron→PN (Criswell and Breese, 2005; Mameli et al., 2005a) and Golgi→granule cell synapses (Carta et al., 2004; Hanchar et al., 2005; Valenzuela et al., 2005). Thus, alcohol affects mossy fiber–parallel fiber excitatory input to PNs via an increase in inhibitory neurotransmission at different relay points along this pathway.
A mature PN also receives excitatory input from the inferior olive via a single climbing fiber (CF). This pathway provides powerful input via ∼1500 synapses that release glutamate with high release probability at the proximal portion of the dendritic tree (Schmolesky et al., 2002). CF activation elicits a unique response known as the complex spike (CS) (see Fig. 1). This response is triggered by activation of AMPA receptors (AMPARs), which evoke an all-or-none EPSP that activates dendritic voltage-gated Ca2+ channels (VGCCs). Depolarization then spreads to the soma and axon, activating voltage-gated Na+ channels and generating the first spike of the CS. Evidence suggests that the depolarization plateau and spikelets that follow the first spike are mediated by Ca2+ and resurgent Na+ currents (Schmolesky et al., 2002). However, the precise identity of the slow CS components remains to be determined.
CF activation of PNs plays a central role in motor adaptation and learning by communicating an error signal that depresses concomitantly active parallel fiber inputs, allowing appropriate motor activity to occur (Hansel et al., 2001). CF input is thought to encode the unconditioned stimulus in pavlovian eyeblink conditioning, and EtOH acutely modulates this paradigm (Hernandez et al., 1986). We hypothesized that CF input to PNs could be affected by acute EtOH exposure. To test this, we used patch-clamp electrophysiological and Ca2+ imaging techniques in rat cerebellar slices.
Materials and Methods
Unless indicated, all chemicals were from Sigma (St. Louis, MO). Experiments were performed in parasagittal vermis cerebellar slices, which were prepared from 20- to 30-d-old male Sprague Dawley rats (Harlan, Indianapolis, IN) as described previously (Carta et al., 2004). Artificial CSF (ACSF) contained the following (in mm): 126 NaCl, 3 KCl, 1.25 NaH2PO4, 1 MgSO4, 26 NaHCO3, 2 CaCl2, and 10 glucose equilibrated with 95% O2 plus 5% CO2. After a recovery time of ≥80 min, slices were transferred to a chamber perfused with ACSF containing 20 μm bicuculline methiodide at a rate of 2–3 ml/min. When indicated, EtOH (AAPER Chemical, Shelbyville, KY), 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo[f]quinoxaline (NBQX), (+)-2-methyl-4-carboxyphenylglycine (LY367385), and/or dl-threo-β-benzyloxyaspartate (TBOA) (all from Tocris Cookson, Ellisville, MO) were added to the ACSF.
Whole-cell patch-clamp electrophysiological recordings were performed under infrared-differential interference contrast microscopy at 31–32°C with Axopatch 200B or Multiclamp 700B amplifiers (Molecular Devices, Sunnyvale, CA). Data were acquired and analyzed with pClamp-9 (Molecular Devices). The effect of ethanol on membrane proteins is sensitive to changes in intracellular signaling pathways, which are affected during whole-cell patch-clamp recordings. This uncertainty must be kept in mind when interpreting our results. Stimulation was performed with a concentric bipolar electrode (Frederick Haer Company, Bowdoinham, ME) placed in the granule cell layer near the patched PN. Constant-current steps (50–100 μA per 100 μs) were delivered every 20 s. The location of the stimulating electrode was gently adjusted until all-or-none CF responses were observed. Patch pipettes had resistances of 3–5 MΩ. Access resistances were between 20 and 35 MΩ; if access resistance changed >20%, the recording was discarded.
CSs were recorded in the whole-cell current-clamp configuration using an internal solution containing the following (in mm): 135 K-gluconate, 10 MgCl2, 0.1 CaCl2, 1 EGTA, 10 HEPES, pH 7.3, and 2 Na2-ATP. Whole-cell voltage-clamp recordings of type-1 metabotropic glutamate receptor (mGluR1)-dependent EPSCs were recorded using the same internal solution at −65 mV. Whole-cell voltage-clamp recordings of AMPAR-mediated EPSCs were performed at a holding potential between −40 and −10 mV using an internal solution containing the following (in mm): 60 CsCl, 30 Cs d-gluconate, 20 tetraethylammonium-Cl, 20 EGTA, 4 MgCl2, 4 ATP, and 30 HEPES, pH 7.3, adjusted with CsOH.
For the long-term depression (LTD) studies, the internal solution contained the following (in mm): 128 CsOH, 111 gluconic acid, 4 NaOH, 10 CsCl, 2 MgCl2, 10 HEPES, 4 Na2ATP, 30 sucrose, and 4 Lidocaine N-ethyl bromide quaternary salt, pH 7.25. LTD was induced by stimulating the CF at 5 Hz for 30 s, as described previously (Weber et al., 2003).
Paired-pulse experiments were performed at an interpulse interval of 50 ms; the paired-pulse ratio (PPR) was calculated as the ratio between the second and first EPSCs. For some PPR determinations, the concentrations of CaCl2 and MgSO4 in the ACSF were changed to 0.5 and 2.5 mm, respectively, to lower glutamate release probability.
Calcium imaging experiments were performed with patch pipettes back-filled with an internal solution containing the following (in mm): 9 KCl, 10 KOH, 120 K-gluconate, 3.48 MgCl2, 10 HEPES, pH 7.25, 4 NaCl, 4 Na2ATP, 17.5 sucrose, and fura-2 AM (100–150 μm; Invitrogen, Carlsbad, CA). Twenty to 30 min were allowed for the dye to diffuse distally and reach a steady-state level. Ca2+ measurements were made by ratio imaging of fura-2 AM using 350/380 nm excitation from a monochrometer (TILL Photonics, Gräfelfig, Germany). Emission was collected using a cooled CCD camera system using DCLP 410 dichroic and LP510/80 emission filters (Chroma Technology, Rockingham, VT). Images were acquired every minute. Each acquisition consisted of ∼20 frames (individual exposure time was 20 ms) at an acquisition frequency of 10 Hz. Data acquisition and analysis were performed with TILLvisION imaging software (TILL Photonics). Fluorescence changes were normalized to resting levels and expressed as ΔF/F, where ΔF = F350/F380, and F is the fluorescence obtained during baseline before CF activation (average of 8–12 frames).
Each slice was exposed once to a single EtOH concentration, and the duration of EtOH exposure was limited to 5 min in all cases to avoid the development of rapid tolerance (Miyakawa et al., 1997). In all cases, effects of EtOH were calculated with respect to the average of control and washout responses. Data are presented as mean ± SEM.
Stimulation of CFs elicited CSs with duration of 140 ± 10 ms and amplitude of the fast Na+ spike of 72 ± 0.8 mV (n = 30). The average number of secondary spikelets was 2.6 ± 0.2 per CS (n = 30). Acute exposure to 50 mm EtOH decreased the area of the CS plateau without affecting the amplitude of the fast Na+ spike (Fig. 1A1,B1). Traces at a more compressed time further illustrate the EtOH-induced decrease in CS area (Fig. 1A2,B2) and also show that the afterhyperpolarization that typically follows the CS (amplitude, −3.2 ± 0.2 mV; n = 22) was not significantly affected by 50 mm EtOH (−8.02 ± 6.8% of control; n = 9). The onset of the effect of EtOH was evident as early as 1 min after the start of application, reaching a maximum 5 min later (Fig. 1C). CS area returned to baseline levels ∼10 min after termination of the application. The effect of EtOH was dose dependent, producing a significant effect even at 10 mm (Fig. 1D). One-way ANOVA followed by Dunnett’s post hoc test revealed statistically significant differences between 10 and 75 mm EtOH (p < 0.05). EtOH (10–75 mm) did not significantly affect the resting membrane potential or the amplitude of the fast Na+ spike (data not shown).
To assess whether EtOH produced changes in the Ca2+ transients associated with the CS, simultaneous Ca2+ imaging and electrophysiological recordings were performed in a subset of neurons used to collect the data shown in Figure 1C. We measured Ca2+ transients in soma, proximal dendrites, and distal dendrites (Fig. 2A) and found that application of 50 mm EtOH did not significantly affect the amplitude or kinetics of these transients (Fig. 2B,C).
It was shown recently that CF glutamate release activates mGluR1, which mediates an EPSC (Dzubay and Otis, 2002). We hypothesized that this EPSC could contribute to the late phase of the CS. Consequently, we examined the effect of the mGluR1 antagonist LY367385 (50 μm) on the CS. Like EtOH, this agent reduced the CS area and duration without affecting the fast Na+ spike (Fig. 3A–C). The effect of this antagonist on the CS area was greater than that of 50 mm EtOH (compare Figs. 1C,D, 3B,C). Importantly, LY367385 occluded the effect of EtOH (Fig. 3B,C).
We next measured the effect of EtOH on EPSCs mediated by mGluR1. In agreement with a previous report (Dzubay and Otis, 2002), these EPSCs could be evoked by stimulating the CF with a train of four pulses at a frequency of 20 Hz in the presence of 10 μm NBQX and 20 μm bicuculline (Fig. 4). These EPSCs were abolished by LY367385, confirming that they depend on mGluR1 activation (Fig. 4A,B,D). Application of 50 mm EtOH decreased the amplitude of these events by ∼70% (Fig. 4A,B,D). As reported previously (Dzubay and Otis, 2002), the glutamate transporter inhibitor TBOA (100 μm) increased the amplitude of the mGluR1-dependent EPSCs (Fig. 4C). EtOH decreased the amplitude of the mGluR1-dependent EPSCs to a similar extent in the absence and presence of TBOA (Fig. 4D). In contrast to a recent report (Zhu et al., 2005), we did not observe enhancement of the EPSC under conditions of mGluR1 antagonism in any of the cells tested (Fig. 4D).
We next investigated whether EtOH modulates mGluR1 function via a presynaptic mechanism by measuring its effect on paired-pulse plasticity. Pharmacologically isolated AMPAR-mediated fast EPSCs evoked by CF stimulation displayed paired-pulse depression, consistent with the high release probability of these fibers (Fig. 5). The EPSCs were abolished by 10 μm NBQX (n = 4; data not shown). Application of 50 mm EtOH did not significantly affect either the amplitude of the EPSCs or the PPR (Fig. 5A,B). An effect of 50 mm EtOH on the PPR or EPSC amplitude could not be observed, even under conditions of low probability of glutamate release (i.e., low extracellular Ca2+ levels) (Fig. 5C–F).
Finally, we assessed the effect of EtOH on CF LTD, a form of synaptic plasticity that requires activation of mGluR1 (Hansel and Linden, 2000). Shown in Figure 6 are recordings of CF-evoked AMPAR-mediated fast EPSCs obtained before and after CF tetanization for 30 s at 5 Hz. Tetanization induced a long-lasting decrease in EPSC amplitude, and this effect could not be observed in the presence of 50 mm EtOH (Fig. 6A,B).
We report here that EtOH potently modulates CF→PN synapses even at a concentration of 10 mm, placing them among a select group that is significantly affected by levels near the legal intoxication limit in the United States (17 mm). Examples of synapses belonging to this group include those in the cerebellar granule cell layer (Carta et al., 2004; Hanchar et al., 2005), CA1 and CA3 hippocampal regions (Carta et al., 2003; Galindo et al., 2005), and amygdala (Roberto et al., 2003).
EtOH does not modulate the initial phase of the CS
The fast voltage-gated Na+ channel-mediated spike was unaffected by EtOH, in agreement with the literature indicating that these channels are insensitive to subanesthetic EtOH concentrations (Shiraishi and Harris, 2004). This finding also indicates that EtOH does not affect AMPARs, which are responsible for initiating the CS (Schmolesky et al., 2002). This was confirmed by the finding that EtOH does not decrease the amplitude of CF AMPAR-mediated EPSCs. These results are in agreement with those from other studies indicating that AMPARs are relatively insensitive to EtOH in the hippocampus (Lovinger et al., 1989, 1990; Weiner et al., 1999; Carta et al., 2003; Hendricson et al., 2003) and cerebellum (Carta et al., 2004). However, hippocampal AMPARs can be very sensitive to EtOH under some circumstances, such as early in development (Lu and Yeh, 1999; Moykkynen et al., 2003; Mameli et al., 2005b), and this could also be the case at developing CF→PN synapses.
EtOH does not affect the Ca2+ transients associated with the CS
EtOH significantly reduced the late phase of the CS, and based on the literature (Llinas and Sugimori, 1980a,b; Schmolesky et al., 2002), we hypothesized that it might act by inhibiting VGCCs. This hypothesis was not supported by our finding that EtOH did not alter Ca2+ transients associated with the CS. Although the identity of the VGCCs involved in CS generation is currently unknown, dendritic P/Q-type and T-type channels have been implicated in this process (Schmolesky et al., 2002). Thus, our findings suggest that these channels are insensitive to EtOH in PNs. In agreement with this suggestion, P-type channels were shown to be insensitive to EtOH concentrations as high as 200 mm in acutely dissociated rat cerebellar PNs (Hall et al., 1994; Walter and Messing, 1999). However, the effect of EtOH on T-type channel function in PNs has not been tested, and we cannot eliminate the possibility that EtOH modulates these channels in light of the recent demonstration that it potently regulates them in thalamic neurons (Mu et al., 2003). Voltage-clamp electrophysiological studies, which have faster temporal resolution than our Ca2+ imaging studies, may reveal effects of ethanol on VGCCs in PNs.
mGluR1-dependent EPSCs contribute to CS generation and are targets of EtOH
Our finding that LY367385 decreases the late phase of the CS suggests that mGluR1 is involved in the generation of this response. This is consistent with the slow timing of the mGluR1-mediated EPSC (Dzubay and Otis, 2002). It was generally believed that voltage-gated Na+ and Ca2+ channel-mediated currents were the principal components of the slow phase of this response (Schmolesky et al., 2002). Our findings indicate that the late portion of this slow phase is mediated in part by mGluR1, which has been shown to be physically coupled to the type-1 transient receptor potential cation (TRPC1) channel that ultimately mediates the EPSC in PNs (Kim et al., 2003). Future experiments will be required to clarify the precise mechanism by which TRPC1 channels mediate the slow components of the CS.
Three pieces of evidence suggest that EtOH inhibits the slow phase of the CS in an mGluR1-dependent manner. First, the effect of EtOH is mimicked by the mGluR1 antagonist LY367385. Second, this antagonist occludes the effect of EtOH. Third, EtOH inhibits the mGluR1-dependent EPSC. The EtOH-induced inhibition of the mGluR1-dependent EPSC is not significantly different in the presence or absence of TBOA, indicating that it is not the result of an indirect action of EtOH at glutamate transporters. The lack of an effect of EtOH on the PPR suggests that its mechanism of action does not involve a decrease in glutamate release probability. Therefore, we hypothesize that EtOH acts at the level of the mGluR1 receptor itself or the intracellular signaling pathways that it activates. Netzeband and Gruol (1995) found that, in cerebellar PNs cultured for 21–37 d, EtOH (33 mm) had little effect on changes in firing rates evoked by the nonselective mGluR agonist (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid (ACPD). At 66 mm, it increased the duration of the ACPD-induced response. Subsequently, Gruol et al. (1997) reported that acute EtOH exposure (33–66 mm) had little or no effect on Ca2+ signals to ACPD and the mGluR1/5 agonist (R,S)-3,5-dihydroxyphenylglycine (DHPG). Interestingly, lower concentrations of EtOH (10 mm) significantly enhanced the Ca2+ signals to DHPG. Based on the results of these studies and those of our study, it cannot be discerned whether EtOH modulates PN function via direct modulation of mGluRs or a downstream indirect mechanism. However, a downstream effect of EtOH is more likely, given that mGluR1-driven Ca2+-dependent Cl− currents were found to be insensitive to 10–100 mm EtOH in Xenopus oocytes (Minami et al., 1998). Recent reports indicate that members of the TRP family, such as TRPV1 and TRPM8, are modulated by EtOH and that this may be in part responsible for EtOH-induced sensory and inflammatory responses in the skin, digestive, and respiratory tracts (Trevisani et al., 2002, 2004; Lyall et al., 2005a,b). Thus, EtOH might affect the mGluR1-dependent EPSC by altering TRPC1 channel function or coupling of this channel with the mGluR1 receptor–Gq–homer complex (Yuan et al., 2003).
Given that PN TRPC1 channels are Ca2+ permeable (Dzubay and Otis, 2002; Canepari et al., 2004), it might have been expected for EtOH to reduce the CS-associated Ca2+ transient. However, evidence indicates that VGCCs, as well as Ca2+ internal stores, are the predominant sources for these transients (for review, see Schmolesky et al., 2002). The large contribution of these sources likely masks that of TRPC channels, which is expected to be relatively small.
Acute EtOH exposure inhibits CF LTD
Hansel and Linden (2000) demonstrated that antagonism of group I mGluRs blocks CF LTD. Therefore, the effect of EtOH on CF LTD is likely to be a consequence of its inhibitory effect on mGluR1-dependent signaling. Our findings are in general agreement with those of Overstreet et al. (1997), who reported that 10 mm EtOH inhibits LTD induced by the nonselective mGluR agonist ACPD in CA1 pyramidal hippocampal neurons from immature rats.
The EtOH-induced inhibition of CF LTD could have an impact at different levels. First, this form of LTD controls the direction of plasticity at parallel fiber→PN synapses. CF-dependent Ca2+ elevations induce parallel fiber LTD, which has a higher Ca2+ threshold than parallel fiber long-term potentiation (LTP) (Coesmans et al., 2004). CF LTD is associated with decreased dendritic Ca2+ transients in PNs, and this increases the probability of parallel fiber LTP induction (Weber et al., 2003; Coesmans et al., 2004). By inhibiting CF LTD, EtOH could increase the probability of parallel fiber LTD induction. Second, CF LTD might be a neuroprotective mechanism that prevents excessive Ca2+ elevations in PNs (Hansel and Linden, 2000). The inhibitory effects of EtOH on this mechanism could trigger Purkinje cell loss in alcoholics (Sullivan et al., 2003). Third, the associative interaction between parallel and climbing fibers enhances endocannabinoid-mediated retrograde inhibition of parallel fiber→PN synapses that lasts for several seconds (Brenowitz and Regehr, 2005). Previous CF LTD induction is likely to modulate this short-term form of synaptic plasticity, and EtOH might affect this regulatory process, leading to alterations in ongoing motor activity and fine motor control.
CF input to Purkinje cells controls simple spike activity driven by mossy fiber–parallel fiber input or an intrinsic spike generator (Cerminara and Rawson, 2004). CF input produces a pause in simple spike activity, which could be impacted both by the EtOH-induced decrease in CS area that is shown here and by the increase in CF bursts that was demonstrated previously by Rogers et al. (1980, 1986). Moreover, not all individual spikelets in the CS are propagated down the PN axon (Khaliq and Raman, 2005; Monsivais et al., 2005). On average, only two spikes, corresponding to the fast Na+ spike and a late spikelet occurring at an interspike interval >4 ms, are propagated to deep cerebellar nuclei (Monsivais et al., 2005). This axonal doublet might relay important information to and/or control the firing properties of deep cerebellar nuclear neurons. Given that EtOH alters the number of secondary spikelets in some cases (Fig. 1B1), it is possible that CF axonal propagation is altered by this agent.
Although the physiological relevance of changes in axonal firing or synaptic strength triggered by the CS is presently unknown, it has been shown that CF input to PNs participates both in movement generation and refinement (Kitazawa et al., 1998). The data presented here indicate that the powerful CF input is diminished by EtOH and this could contribute to the motor impairing actions of this widely abused substance.
This work was supported by National Institutes of Health Grant AA014973. We thank L. D. Partridge for critically reading this manuscript and assistance with calcium imaging studies. We are also grateful to W. Muller and J. A. Connor for invaluable comments.
↵*M.C. and M.M. contributed equally to this work.
M. Carta’s present address: Department of Experimental Biology, Section of Neuroscience, University of Cagliari, 09123 Cagliari, Italy.
- Correspondence should be addressed to Dr. C. Fernando Valenzuela, Department of Neurosciences, MSC08 4740, 1 University of New Mexico, Albuquerque, NM 87131-0001. Email: