Abstract
Withdrawal anxiety after chronic alcohol is likely to contribute to drug seeking and relapse in alcoholics. The brain regions regulating fear/anxiety behaviors, especially neurotransmitter systems with acute ethanol sensitivity, are potential targets for chronic ethanol-induced adaptations. We have therefore examined N-methyl-d-aspartate (NMDA) receptors after chronic ethanol ingestion in rat lateral/basolateral amygdala. Whole cell patch-clamp measurements indicate that chronic ethanol ingestion significantly increased NMDA receptor current density. This enhanced NMDA receptor function was also associated with an increase in ifenprodil inhibition and a decrease in apparent calcium-dependent current inactivation. These findings suggest that NR2B-containing receptors may be specifically enhanced and suggest that processes dependent upon calcium influx through amygdala NMDA receptors may potentially be enhanced by chronic ethanol ingestion. We measured subunit mRNA expression to investigate possible molecular mechanisms that control functional receptor adaptations to chronic ethanol. Quantitative real-time reverse transcription-polymerase chain reaction (RT-PCR) demonstrated that NR1 subunit mRNA expression, but not NR2 or NR3 expression, was enhanced in samples from chronic ethanol-exposed animals. Single-cell RT-PCR was then used to confirm that NR2 mRNA expression was unaltered by chronic ethanol. Most GAD–, presumed projection neurons expressed both NR2A and NR2B mRNAs, and this profile did not change during chronic ethanol exposure. Our results suggest that both transcriptional and nontranscriptional adaptations to chronic ethanol ultimately contribute to alterations in NMDA receptor function. Because amygdala NMDA receptors play a significant role in many learned fear behaviors, chronic ethanol-induced adaptations in these receptors may influence the expression of withdrawal anxiety.
Recent advances in understanding the neuroanatomy of anxiety behaviors have identified the amygdala as a central regulatory brain area. The lateral and basolateral subdivisions are the amygdala's primary input areas where sensory information is extensively processed and forwarded to both the central amygdala and other “emotional” forebrain regions. Because these subdivisions are functionally related and are closely juxtaposed anatomically, they will be referred as the lateral/basolateral amygdala (BLA) throughout this work. Inactivation or chemical lesions of the BLA disrupt learned fear measured in several different behavioral paradigms. For example, direct muscimol injections before training, but not after training, block the acquisition of learned fear (Wilensky et al., 1999). Lesions of the BLA also disrupt the ability of drug-associated cues to reinstate drug-seeking behavior (Meil and See, 1997). Analysis of neurotransmitter systems within the lateral/basolateral amygdala has shown that NMDA receptors play a unique role in fear-learning. NMDA receptor antagonists injected directly into the lateral/basolateral amygdala block the association of sensory cues with footshock in many fear-learning paradigms (Fanselow and Kim, 1994; Hatfield and Gallagher, 1995). Like these antagonists, ethanol can inhibit NMDA receptor function and can disrupt fear learning as well (Sonner et al., 1998), suggesting that the lateral/basolateral amygdala is an important site of ethanol action.
NMDA receptors are multimeric complexes containing distinct families of protein subunits. The NR1 subunit family is required for the formation of functional channels and is represented by a single gene theoretically encoding eight unique splice variants (Laurie and Seeburg, 1994). NR2 subunits produce channels similar to those found in native systems when coexpressed with NR1 (Williams et al., 1994) and are encoded by four separate genes (NR2A–D). Expression of the different NR2 subunits can confer unique pharmacological and biophysical properties to native receptors. Because the expression of NR2 subunits is both temporally and spatially regulated (Monyer et al., 1994), the characteristics of NMDA receptors vary both during development and from brain region to brain region. NR3 subunits share sequence similarity with other NMDA subunits and seem to function in a dominant negative manner when coexpressed in a receptor complex (Sucher et al., 1995). These subunits are highly expressed in embryonic central nervous system (Sucher et al., 1995) and seem to regulate the development of NMDA receptor-containing synapses (Das et al., 1998). NR3 subunit expression declines in adults to very low levels, although substantial expression is still evident in the temporal lobe, including the amygdala (Ciabarra et al., 1995). Thus, a diversity of NMDA receptor subunits may help to determine the functional characteristics of these important channels.
Acute inhibition of NMDA receptor-gated currents by intoxicating concentrations of ethanol (Lovinger et al., 1989) has been demonstrated for numerous isolated neuronal preparations as well as for NMDA-mediated synaptic responses in several brain regions. Subunit composition seems to influence this interaction. For example, NR2A- or NR2B-containing receptors are relatively more sensitive to acute ethanol inhibition compared with those channels containing NR2C or NR2D when expressed in heterologous systems (Masood et al., 1994). Consistent with this, both ethanol inhibition and inhibition by the NR2B-selective antagonist decreases in cortical neurons as their time in culture increases (Lovinger, 1995). However, subunit-specific post-translational modifications and protein-protein interactions can also dramatically influence the acute ethanol sensitivity of NMDA receptors (Yaka et al., 2003). Thus, many factors may influence the interaction between ethanol and the NMDA receptor. This diversity is manifest frequently as variations in ethanol potency or efficacy across brain regions.
In addition to acute ethanol action, chronic ethanol exposure can also alter NMDA receptor function and subunit composition. Functional NMDA receptor binding is increased after chronic ethanol in several brain areas, including the hippocampus and cortex (Gulya et al., 1991). Chronic ethanol also enhances NMDA-mediated currents in medial septum neurons (Grover et al., 1998), although direct measures of chronic ethanol's influence on NMDA channel function in native cells are not abundant in the literature. Due to the prominent role of NMDA receptors in amygdala-dependent behaviors, we have therefore focused chronic ethanol's effects on native NMDA receptor pharmacology, function, and expression in rat lateral/basolateral amygdala. Portions of this work have published in abstract form.
Materials and Methods
Chronic Ethanol Administration. All animal procedures were in accordance with the guidelines set forth by National Institutes of Health animal care and use policy. Adult male Sprague-Dawley rats (Harlan, Indianapolis, IN) were housed individually in an American Association for Accreditation of Laboratory Animal Care-accredited facility and were kept on a 12-h light/dark cycle. Animals were randomly placed into two treatment groups, those receiving a non-ethanol-containing liquid diet (control group) and those chronically exposed via an ethanol-containing liquid diet. These diets were commercially obtained (Bio-Serv Inc., Frenchtown, NJ) and are similar to that reported by Lieber and DeCarli (1989). Chronic ethanol rats received the ethanol diet [4–6% (v/v)] for a total of 10 to 12 days. Control rats were “yoked” to the ethanol rats and received a volume of the control diet equivalent to the consumption by the ethanol rats on the previous day. Both control and chronic ethanol rats gained a similar amount of weight during the liquid diet treatments (P > 0.6, two-tailed t test); start weights were 132 ± 10 g for controls (n = 16) and 133 ± 8 g for chronic ethanol rats (n = 17), and finishing weights were 195 ± 10 and 203 ± 11 g for control and chronic ethanol rats, respectively. Water was given ad libitum and diet intake monitored daily. Chronic ethanol rats consumed 11.2 ± 0.4g/kg/day ethanol and were sacrificed while intoxicated. Where it was measured, blood ethanol concentrations at the time of sacrifice were 153 ± 16 mg/dl (n = 17).
Brain Slice Preparation and Neuronal Isolation. To prepare brain slices, rats were anesthetized with isoflurane and decapitated. Tissue slices containing the BLA were prepared as reported previously (McCool et al., 2003). After preparation, slices were stored in oxygenated Ringer's solution for up to 5 h. Individual neurons were isolated from tissue by incubation of a single slice for 20 min at 37°C in oxygenated Ringer's solution containing 0.5 to 0.75 mg/ml Pronase protease (Calbiochem, San Diego, CA). Digested tissue was passed through a series of fire polished Pasteur pipettes, and neurons were allowed to settle on tissue culture-treated plastic dishes or coverslips. Ethanol (150 mg/dl) was added to brain slices from chronic ethanol-exposed rats during storage to prevent withdrawal in vitro. In a separate set of experiments, we determined that an ∼2.5-h in vitro ethanol exposure of control brain slices did not affect maximal NMDA responses. This time was chosen because it represents the average time slices were stored during a typical experiment. Using 1 mM NMDA, current densities were 27 ± 4 pA/pF for slices incubated without ethanol (n = 18) and 24 ± 7 pA/pF for slices incubated with 150 mg/dl ethanol during storage (n = 6; P > 0.1, t test). Ethanol was absent from chronic ethanol cells during the enzymatic treatment, neuronal isolation, and recording periods.
Electrophysiology. The whole cell patch-clamp technique was performed on acutely dissociated neurons that were continuously perfused with a HEPES-buffered saline solution consisting of 150 mM NaCl, 10 mM HEPES, 2.5 mM KCl, 2.5 mM CaCl2, 1.0 mM MgCl2, and 10 mM d-glucose, pH 7.4 with NaOH, osmolality 320 mmol/kg adjusted with sucrose. Drugs were diluted from concentrated stocks into HEPES-buffered saline lacking MgCl2 and containing only 0.2 mM CaCl2 along with 0.5 to 1 μM tetrodotoxin. This solution was applied within 100 μm of the cell using a linear array of fused silica tubes (150 mm i.d.; Hewlett Packard, Palo Alto, CA) mounted on a manipulator. A Cs+-based internal solution (120 mM CsCl, 10 mM HEPES, 11 mM EGTA, 1 mM CaCl2, and 4 mM Mg-ATP, pH 7.2 with CsOH, osmolality 305 mmol/kg adjusted with sucrose) was used in the patch electrode.
Recordings were done at room temperature using an Axopatch ID amplifier (Axon Instruments, Union City, CA) in voltage-clamp mode as described previously (Hamill et al., 1981). Whole-cell capacitance and series resistance were determined by fits of the capacitive transients during square-wave voltage steps using standard software procedures contained within pClamp 7.0 software (Axon Instruments) and monitored throughout the recordings. Series resistance and capacitance were compensated manually.
Current amplitudes were measured from the apparent peak of the response and are reported as values standardized to the whole cell capacitance (in picofarads) derived from fits to a square-wave depolarization (pClamp 7.0; Axon Instruments). Concentration-response relationships were derived using commercially available software (Prism; GraphPad Software Inc., San Diego, CA) where data were fit to the standard logistic equation Y = Ymax/(1 + 10((log EC50–X)·Hill slope) Summarized data are reported as the mean ± S.E.M.).
Real-time RT-PCR. Total RNA was isolated from the lateral/basolateral amygdala of individual control and chronic ethanol rats or from whole forebrain using affinity chromatography (RNeasy mini kit; QIAGEN, Valencia, CA). Contaminating genomic DNA was removed by digestion with DNase I according to the manufacturer's instructions. RNA concentrations were determined using fluorescent detection (RiboGreen; Molecular Probes). The reverse transcription reaction was performed on 4 to 8 ng/μl total RNA using random hexanucleotides as described previously (McCool and Farroni, 2001). Real-time PCR was performed on cDNA products using the TaqMan detection method (for review, see Giulietti et al., 2001) and a Prism 7000 sequence detection system (Applied Biosystems, Foster City, CA). Taqman Universal PCR Mix (Applied Biosystems) containing TaqDNA polymerase, dNTPs (+dUTP), and buffers was used according to manufacturer's directions. Pilot experiments demonstrated that 0.9 μM primers and 0.25 μM probe produced the largest change in fluorescence during the course of a PCR reaction for all gene products examined. Primer and probe combinations were designed using PrimerExpress software (version 3.0; Applied Biosystems) and rat sequences available on public databases (GenBank). For the rat NR1 subunit, regions around the various splice sites were excluded from consideration. Similarly, cDNA sequences for the NR2 or NR3 family were compared and regions greater than 85 to 90% similar were not included for primer/probe design. Probes were labeled with 5′5-carboxyfluorescein and 3′5-carboxytetramethylrhodamine. Primers and probes for each NMDA subunit mRNA and for the ubiquitous gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) are shown in Table 1.
The PCR reaction consisted of initial incubations at 50°C for 2 min followed by 95°C for 10 min; steps were 40 cycles of 95°C for 15 s to melt DNA duplexes followed by 60°C for 1 min for annealing and product formation. Fluorescence was measured at the end of the annealing/extension step. Background fluorescence, defined as the fluorescence during early portions of the PCR reaction before any substantial accumulation of PCR product, was subtracted from each reaction. CT values were defined as an arbitrarily “cut-off” change in fluorescence above background, typically 0.1 to 0.3 log units, depending upon the absolute level of fluorescent intensity, during the log-linear phase of the PCR reaction.
The “relative standard curve” method (Johnson et al., 2000) was used to compare expression levels of mRNAs between the control and chronic ethanol samples. For this, serial dilutions of cDNAs prepared from total forebrain RNA were subjected to real-time PCR with gene-specific primer/probe combinations to establish a standard curve. Real-time PCR reactions on cDNAs prepared from control or chronic ethanol total RNA samples were performed at the same time; and, cycle threshold (CT) values of these samples were related to “nanograms of forebrain RNA” equivalents using the linear relationship between the log(ng forebrain RNA) in a particular standard curve reaction and its corresponding CT value (Fig. 5). Each gene's nanograms of forebrain RNA equivalent in a given sample was normalized to the relative expression level GAPDH in that same sample.
Single-Cell RT-PCR. Single-cell RT-PCR on dissociated basolateral amygdala neurons was performed as described previously (McCool and Farroni, 2001). For some experiments, individual neurons were harvested by simple aspiration into a borosilicate glass pipette containing ∼5 μl of ribonuclease-free water and ribonuclease inhibitor (0.1 U/μl). In other experiments, neurons were harvested after electrophysiology; in these cases, ribonuclease inhibitor (0.1 U/μl) was added to the recording internal solution described above. There were no apparent differences in the NR2 expression profile between neurons harvested in water versus those harvested after electrophysiology.
For the polymerase chain reaction, “hot-start” was used to initiate the reaction by addition of Taq polymerase after a 2-min heat denaturation. Samples were subjected to 40 cycles of 94°C for 1 min, 50°C for 1 min, and 72°C for 1.5 min in a PTC-100 thermal cycler (MJ Research, Waltham, MA). Amplicons were analyzed by agarose gel electrophoresis using ethidium bromide fluorescence. Oligonucleotide primers were derived from unique sequences to specifically amplify individual NR2 subunits (Porter et al., 1998), the 67-kDa form of glutamic acid decarboxylase (Ceranik et al., 1997), or GAPDH (Salazar et al., 1999). All oligonucleotides were obtained from commercial sources (Sigma Genosys, The Woodlands, TX).
To statistically examine the largely qualitative scRT-PCR data, a binomial nomenclature (0, no expression; 1, expression detected) was used to describe results for each gene in any given cell. Arrays were generated for the expression of a given mRNA derived from each treatment and were compared (control versus chronic ethanol) using resampling methods contained within the resampling stats add-in for Excel (Resampling Stats Inc., Arlington, VA). Briefly, if treatment had an effect on the expression of a given gene, the difference between the fraction of cells expressing the gene (number of 1s divided by total number of neurons tested) in each treatment group would be relatively large, and the magnitude of this difference would occur only rarely in artificially generated arrays where the total number of 1s and 0s as the original data set were randomly assigned between the treatment groups. The distribution of the differences between treatments for these random arrays was used to discern the “probability” of obtaining the difference derived from the real data. In our case, we used the original data to generate 10,000 random arrays. A difference value in the real data that was represented by equal or larger magnitudes in less than 500 of 10,000 simulated arrays would be assumed to predict a P < 0.05 (see http://www.resample.com/content/software/excel/userguide/index.shtml for methodological details).
Results
Chronic Ethanol Enhances NMDA Efficacy but Not Potency. To assess NMDA receptor function in control (n = 7 individuals) and chronic ethanol (CE) rats (n = 7 individuals), we established NMDA concentration-response relationships in isolated lateral/basolateral amygdala neurons (Fig. 1A). We normalized currents (in picoamperes) to apparent cell capacitance (in picofarads) to reduce cell-to-cell variability. Neither whole-cell capacitance (18 ± 1 versus 19 ± 1pF; P > 0.5 with two-tailed t test) nor access resistance (24 ± 1 versus 24 ± 2 MΩ; P > 0.9 with two-tailed t test) was significantly different between control neurons (n = 39) and CE neurons (n = 43), respectively. In the presence of 3 μM glycine, current densities from 30 μM to 1 mM NMDA were significantly larger in neurons isolated from CE rats compared with controls (Fig. 1B; P < 0.05 at each concentration, two-tailed t test). For example, at the maximal concentration tested (1 mM), chronic ethanol NMDA currents were 56 ± 8 pA/pF (n = 21 neurons), whereas control currents were 35 ± 6 pA/pF (n = 16 neurons). A similar trend was noted for comparisons between current amplitudes without normalization, with current amplitudes with 0.3 mM from control neurons (532 ± 94 pA) being significantly smaller (P < 0.05) than amplitudes from chronic ethanol neurons (868 ± 125 pA). In one experiment, neurons from a chow-fed rat exhibited a 1 mM NMDA/3 μM glycine current density (33 ± 9 pA/pF, n = 3) that was almost identical to control rats, suggesting that the differences between control and CE current density was due to the ethanol exposure in the latter group and that the liquid diet had little impact on “basal” NMDA receptor function. Despite the elevated NMDA receptor function in CE neurons, the apparent affinity and Hill slope defined by the concentration-response relationship were not significantly different between treatment groups (Fig. 1B). In cells where complete concentration-response relationships were defined, the log [EC50] values were –4.1 ± 0.1 for control neurons (n = 14, equivalent to 75 μM) and –4.1 ± 0.1 (n = 18, 68 μM) for CE neurons (P ≫ 0.05, t test with Welch's correction). For these same neurons, the Hill slopes, 1.3 ± 0.1 for controls and 1.5 ± 0.1 for chronic ethanol cells, were also not significantly different.
More detailed examination of the current densities for 300 μM NMDA from individual neurons in each treatment group (Fig. 1C) revealed that chronic ethanol may have enhanced the average current density by increasing the number of neurons with densities >50 pA/pF. Although there were neurons with comparable densities in the control group (n = 3 of 21 or 14%), there were substantially more of these cells (n = 9 of 19 or 47%) in the chronic ethanol group. Thus, although the range of current densities was similar between treatment groups (5–97pA/pF for control neurons versus 12 to 126pA/pF for chronic ethanol neurons), the median current density for chronic ethanol neurons (50 pA/pF) was over twice that for control neurons (23 pA/pF).
Chronic Ethanol Does Not Alter the Acute Ethanol Sensitivity of NMDA Receptors. As in many other preparations, NMDA-gated currents in acutely isolated lateral/basolateral amygdala neurons were inhibited by acute application of ethanol (Fig. 2A). However, chronic ethanol ingestion did not seem to alter this acute sensitivity. Using 30 mM ethanol, NMDA currents were inhibited by 18 ± 7% in control neurons (n = 4; Fig. 2B) and by 15 ± 2% (n = 9) in chronic ethanol neurons (P ≫ 0.1, t test). Likewise, 100 mM ethanol inhibited NMDA currents by 51 ± 7% (n = 10) and 42 ± 5% (n = 11) in control and chronic ethanol neurons, respectively. Although there was an apparent trend for reduced inhibition at this higher ethanol concentration, the differences between treatment groups were not significantly different. The distribution of the individual percentage of inhibition values at 100 mM ethanol was also similar between groups (20–81%, median = 47% for control neurons versus 21–71%, median = 41% for chronic ethanol neurons).
Chronic Ethanol Increases Inhibition by Ifenprodil. The noncompetitive antagonist ifenprodil (10 μM) substantially inhibited NMDA-gated currents elicited by 100 μM NMDA and 3 μM glycine in a majority of neurons tested. Ifenprodil (10 μM) represents a concentration that maintains some selectivity for the inhibition of NR2B-containing NMDA receptors relative to receptors composed of other subunits (Williams, 1993). In many neurons, ifenprodil inhibition was slow to reach its maximal extent during application of an NMDA/ifenprodil admixture, as has been described for the action of this antagonist (Williams, 1993). There were also a large number of neurons where ifenprodil inhibition did not have this slow-onset but was none-the-less substantial (for examples, see Figs. 3 and 7). To ensure that we examined ifenprodil inhibition at its fullest extent, we compared the percentage of inhibition at the apparent steady-state current level (“ISS”), approximately 4 s after the peak of the current response (Fig. 3). Ifenprodil inhibited steady-state current by 29 ± 4% in control neurons (n = 16) and by 46 ± 5% (n = 18) in chronic ethanol neurons (P < 0.05, t test). In addition of steady-state current amplitudes, peak-to-steady-state current ratios are often used to describe ifenprodil inhibition and can provide a better appreciation for its slow onset. In control neurons, IPeak/ISS ratios were 1.5 ± 0.1 in the presence of 10 μM ifenprodil and significantly increased to 2.0 ± 0.4 in CE neurons (P ≪ 0.01, two-tailed t test). Both results strongly suggest that chronic ethanol ingestion increased the contribution to functional receptors by the NR2B subunit.
Influence of Extracellular Calcium on Current Kinetics: Interactions with Chronic Ethanol. Because chronic ethanol seemed to influence the NR2 subunit composition of NMDA receptors, we next examined the impact of chronic ethanol exposure on the effects of extracellular calcium because NR2 subunits can also influence calcium-dependent current inactivation (Krupp et al., 1996). With low extracellular calcium (0.2 mM), currents evoked using an EC50 concentration of NMDA plus 3 μM glycine did not appreciably inactivate during the agonist exposure (Figs. 1, 2, 3). However, increasing the extracellular Ca2+ concentration (Cao2+) from 0.2 to 2 mM substantially increased the apparent rate of current inactivation during the agonist exposure (Fig. 4A). Because high Cao2+ attenuated the steadystate portion of the current (ISS) much more than the peak (IPeak) current amplitude, we specifically examined the effects of 2 mM Cao2+ on ISS. The inhibition of ISS in control neurons (58 ± 5%, n = 6; Fig. 4B) was significantly larger than in chronic ethanol neurons (35 ± 2%, n = 12; P < 0.01 in two-tailed t test). Current inactivation described by the IPeak/ISS ratio during exposure to 2 mM Cao2+ was also increased in control neurons relative to CE neurons, as demonstrated by the average current traces from control and chronic ethanol neurons in Fig. 4C. The IPeak/ISS ratio in presence of 2 mM Cao2+ was 2.15 ± 0.18 in control neurons (n = 6; Fig. 4D), whereas the ratio in CE neurons was 1.71 ± 0.09 (n = 12; P < 0.05, two-way t test). In contrast, the ratios in 0.2 mM Cao2+ for control neurons (1.39 ± 0.09) was not different from chronic ethanol neurons (1.35 ± 0.05), indicating the differences between the IPeak/ISS ratios of control and chronic ethanol neurons in 2 mM Cao2+ was likely specific to permeation of calcium through the NMDA channels.
Chronic Ethanol Alters the Expression of Some NMDA Receptor Subunit mRNAs. Because our functional data indicated that chronic ethanol could influence NMDA receptor levels and alter NR2 subunit contributions, we examined regulation of mRNA expression levels as a potential molecular mechanism. Given that the lateral/basolateral amygdala is a relatively small brain region, comparisons between control and chronic ethanol animals were achieved using real-time RT-PCR with “TaqMan” detection (Giulietti et al., 2001). Total forebrain RNA was used with the relative standard curve method (Johnson et al., 2000) to quantify gene expression within individual samples. An example of the real-time method is shown in Fig. 5A. Importantly, the relationship between the log[nanograms of total forebrain RNA] put into the PCR reaction and the relative expression level, represented by the CT (see Materials and Methods), of each gene product was linear over a wide range concentrations (Fig. 5B). This ensured that CT could be directly related to the amount of the mRNA in the original sample.
Using this real-time RT-PCR method, we compared relative expression levels of each NMDA receptor subunit mRNA in control and chronic ethanol rats. Expression levels of NMDA subunits were normalized to levels of the ubiquitous gene product for GAPDH in each sample. Importantly, GAPDH mRNA expression was not significantly different between control and chronic ethanol samples (Fig. 6A). Although all NMDA subunit mRNAs were detected in the BLA, neither NR2 nor NR3 subunit mRNA levels were significantly different between the treatment groups (Fig. 6C; Table 2). In contrast, NR1 subunit mRNA levels were significantly increased by chronic ethanol exposure (Fig. 6B; P < 0.05, t test). Relative expression levels of NR1 mRNA were 22.0 ± 1.4 for control samples and 28.0 ± 1.5 in BLA exposed to chronic ethanol, a ∼25% increase. The relative expression level of NR1 as well as the remaining subunits was not significantly correlated with absolute amount of ethanol consumed by an individual rat (not shown).
Chronic Ethanol Does Not Alter the NR2 Expression Profile of Individual Amygdala Neurons. Because the expression of NR2C and NR2D was detected in total lateral/basolateral amygdala RNA and was not anticipated, we subsequently examined NR2 expression in single isolated neurons to identify the cellular source for these various mRNAs. As shown in Fig. 7, A and B, many neurons expressed multiple NR2 subunit mRNAs, with NR2A and NR2B mRNAs being the predominant species detected with this method (Fig. 7, A and B). Although we attempted to examine isolated neurons with morphological characteristics consistent with glutamatergic projection neurons (McDonald, 1982), we confirmed the cell phenotype in these molecular studies by using glutamic acid decarboxylase (GAD)65 mRNA expression as a marker for GABAergic interneurons. In 12 control neurons, 10 cells did not seem to express GAD (Fig. 7C). Three of 10 of these GAD– control neurons (30%) expressed NR2A subunit mRNA alone. Five of the remaining seven neurons (50%) expressed both NR2A and NR2B. Interestingly, one control neuron expressed NR2A, NR2B, and NR2D (10%), whereas the final neuron (10%) expressed all known NR2 subunits. Chronic ethanol exposure did not seem to qualitatively alter this pattern of expression. Fourteen of 17 chronic ethanol neurons yielded signal for NR2 expression. 12 of these neurons seemed to be GAD–. Five of the 12 NR2+/GAD– chronic ethanol neurons (42%) expressed only NR2A (Fig. 6D), whereas one of 12 (8%) expressed only NR2B. Of the remaining six neurons, five chronic ethanol neurons (42%) expressed both NR2A and NR2B, whereas a single NR2A/B+ neuron (8%) also expressed NR2D. To provide some quantitative comparisons of these largely qualitative data, we used resampling techniques (see Materials and Methods) but did not detect any significant statistical differences between the NR2 expression profiles of the control and chronic ethanol groups when the NR2 mRNAs were considered individually.
Our morphological selection process limited the number of GAD+ neurons examined, preventing us from making any specific measures of chronic ethanol on NR2 expression in this cell population. However, our preliminary results indicate that the NR2 expression profile of GAD+ neurons may be distinct from that found in the GAD– population. Although two of six GAD+ neurons from all experimental groups expressed only NR2A and NR2B, the expression profile in the remaining four GAD+ neurons was heterogeneous and consisted of NR2A+D, NR2A+B+C, NR2A+C+D, and NR2A+B+D. Thus, the GAD+ population of neurons is the most significant source for NR2C and 2D expression in the lateral/basolateral amygdala.
Discussion
One of the major findings of our study was that chronic ethanol exposure resulted in a pronounced functional increase in NMDA receptors expressed by lateral/basolateral amygdala neurons. Because NMDA receptors in this brain region are important both for fear responses to unconditioned stimuli (Adamec et al., 1999) and for conditioned “fear learning” (Miserendino et al., 1990), chronic ethanol-induced increases in NMDA receptor function may help “identify” the withdrawal state as an aversive event and ultimately help accentuate drug-seeking behaviors. It will be particularly interesting to correlate the time course of withdrawal anxiety and the subsequent development of drug seeking with the functional properties of lateral/basolateral amygdala NMDA receptors.
A second important finding of our studies was that chronic ethanol exposure altered the pharmacological properties of amygdala NMDA receptors. However, acute ethanol sensitivity was not changed, suggesting little or no tolerance to ethanol by amygdala NMDA receptors in this system. The impact of chronic ethanol exposure on acute ethanol sensitivity on NMDA receptors has varied in the literature. For example, the acute ethanol sensitivity of NMDA-induced neurotoxicity seems diminished after chronic ethanol exposure of cultured cerebellar granule cells (Cebere et al., 1999), whereas chronic ethanol did not attenuate the acute ethanol inhibition of NMDA receptor-dependent intracellular calcium increases in these same neurons (Iorio et al., 1992). Similarly, acute ethanol inhibition in acutely dissociated medial septum/diagonal band neurons is attenuated after chronic ethanol (Grover et al., 1998). However, acute ethanol inhibition of amygdala NMDA currents (this study) and NMDA currents in hippocampal slices or cultured neurons (White et al., 1990) were not altered by chronic ethanol exposure. Although different endpoint measures might explain the disparate findings in cultured cerebellar neurons, the divergent findings in the latter electrophysiology studies may indicate brain region-specific tolerance of NMDA receptor acute ethanol inhibition after chronic ethanol exposure. Furthermore, it is unclear how absolute amounts of ethanol exposure or the total length of exposure might have influenced these studies.
In contrast to acute ethanol, chronic ethanol ingestion did seem to influence the functional contribution by specific NMDA receptor subunits. The inhibition of amygdala NMDA currents by the noncompetitive, NR2B-specific antagonist ifenprodil was enhanced in neurons from ethanol-exposed animals. This increase in NR2B function is also reflected by the apparent increase in IPeak/ISS ratio in the presence of ifenprodil. Although we used an ifenprodil concentration that has maximal effects on NR2B (Williams, 1993), we cannot exclude the possibility that combinations of NR2B with other NR2 subunits might reduce the apparent efficacy of ifenprodil and mask chronic ethanol-induced increases in these subunits (Blevins et al., 1997). Our approach of combining NR2 expression profiling using single-cell RTPCR with electrophysiology did in fact demonstrate the presence of multiple NR2 subunit mRNAs in many neurons. Regardless, these ifenprodil findings are supported by the alterations in calcium-dependent inactivation after chronic ethanol exposure. Alterations in the peak-to-steady-state ratio may further suggest that calcium-dependent inactivation may be decreased after chronic ethanol exposure. The influence of chronic ethanol on absolute current amplitude in different extracellular calcium environments could also be interpreted as a relative increase in Ca2+ permeability (Plant et al., 1997). Regardless of the interpretation, NR2 subunits can dramatically influence Ca2+ entry through NMDA receptors (Blevins et al., 1997), with the rank order of permeability for NMDA-gated channels containing different NR2 subunits being NR2B > NR2A+NR2B > NR2A. These results together suggest that the larger current amplitudes in the presence of “high” extracellular Ca2+ for chronic ethanol neurons are consistent with an increased functional contribution by ifenprodil-sensitive NR2B-containing channels. Although the ultimate consequences of enhanced Ca2+ entry or decreased calcium-dependent inactivation after chronic ethanol ingestion are yet to be directly addressed, NR2B-containing amygdala NMDA receptors participate in fear-learning behaviors (Rodrigues et al., 2001). Furthermore, calcium entry through NMDA receptors is essential for long-term potentiation-like increases in synaptic efficacy (Regehr and Tank, 1990) that can ultimately result in increased fear learning during or after chronic ethanol ingestion. Our results therefore suggest important consequences are likely to be related to functional adaptations of lateral/basolateral amygdala NR2B subunit.
Alterations in functional contributions by different NR2 subunits after chronic ethanol exposure led us to examine the influence of gene expression as a potential contributor to receptor adaptation. However, real-time RT-PCR measures of NR2 subunit mRNA expression failed to detect any significant alterations in these subunits during chronic ethanol exposure. This suggests that the altered functional contributions by the NR2B subunit during chronic ethanol are not mediated by changes in NR2B mRNA transcription or stability. Recently, acute ethanol has been shown to influence the phosphorylation status of the NR2B subunit (Yaka et al., 2003). Because phosphorylation of this subunit can also have a profound impact on receptor function (Yaka et al., 2002), it is entirely possible that NR2B adaptation in lateral/basolateral amygdala is governed by nontranscriptional mechanisms. Regardless, the up-regulation of NR1 subunit mRNA levels during chronic ethanol may help accentuate chronic ethanol-dependent up-regulation of receptor function. Our molecular findings further suggest that the molecular mechanisms related to NMDA receptor adaptation in lateral/basolateral amygdala may be distinct from those in the cerebral cortex where chronic ethanol up-regulates functional receptors (Gulya et al., 1991) without altering NR1 subunit mRNA expression (Morrow et al., 1994). This contrast also highlights the diversity of region-specific transcriptional and post-transcriptional mechanisms responsible for chronic ethanol-induced alterations in NMDA receptor expression and function.
To our knowledge, this is the first report of NR2C, NR2D, NR3A, and NR3B mRNA expression in lateral/basolateral amygdala. Although the absolute levels of these subunits cannot be defined using the methods used here, we can consider their expression level relative to forebrain without normalizing to GAPDH. Combining data from control and chronic ethanol samples (n = 17), NR2C expression per nanograms of total lateral/basolateral RNA was equivalent to 5.6 ± 1.2 ng of forebrain RNA, whereas NR2D expression was equivalent to 0.7 ± 0.1 ng of forebrain RNA per nanogram of lateral/basolateral RNA. Similarly, NR3A expression per nanogram of total lateral/basolateral RNA was similar to that found in 2.5 ± 0.4 ng of forebrain RNA, whereas NR3B expression in 1 ng of total lateral/basolateral amygdala RNA is equivalent to 1.4 ± 0.3 ng forebrain RNA. These results indicate that NR2C and NR3A transcripts are relatively more abundant in the lateral/basolateral amygdala compared with the forebrain, whereas NR2D and NR3B expression is approximately equivalent.
Like the real-time RT-PCR experiments, our single-cell RT-PCR data were unable to identify any significant shift in NR2 expression profile after chronic ethanol. These results support our more quantitative findings with real-time measures. In addition, several general findings are of some interest. First, we found that the majority of neurons examined by scRT-PCR were GAD-negative, suggesting that our physiological examinations were performed primarily upon GAD–, presumably projection neurons. Second, although our neurons were taken from adult rats (ca. 200 g), most GAD– neurons expressed both NR2A and NR2B mRNAs, indicating the NR2B-to-NR2A transition evident in other brain regions during the juvenile-to-adult phase of development (Monyer et al., 1994) may not take place in the BLA or be incomplete in the animals examined here. We were also surprised to find that a small population of GAD-negative neurons (approximately 20% across all treatment groups) seemed to express NR2D subunit mRNA. Although previous work in other laboratories (Monyer et al., 1994) did not indicate any significant expression of this subunit in this brain region, it is possible that the modest-to-low levels of expression noted in our real-time experiments and the relatively small number of neurons identified by scRT-PCR might have prevented detection by less sensitive methods. Finally, despite our attempts to select neurons that morphologically resemble projection neurons, GAD expression was none-the-less detected in eight of 42 cells (19%) examine be single-cell RT-PCR. Importantly, a subset of the GAD+ presumed interneurons seemed to be the primary source of NR2C mRNA expression in the lateral/basolateral amygdala. Regardless, the GAD+ cells sampled by our molecular studies may represent the distinct class of relatively “large, pyramidal” interneurons identified in previous anatomical studies (McDonald, 1982). However, the number of GAD+ interneurons with this phenotype should be quite small relative to the large number of GAD-negative projection neurons. This in turn suggests that the procedure for acutely separating individual neurons from tissue may somehow bias for the isolation of GAD+ cells.
In conclusion, chronic ethanol ingestion enhances lateral/basolateral amygdala NMDA receptor function. Although this enhancement seems to involve alterations in NR2 subunit content of the functional receptor complex, we find no indication that pronounced alterations in the NR2 subunit mRNA expression plays any role in these adaptations. Because lateral/basolateral amygdala NMDA receptors can regulate a number of important behaviors, from cue-specific fear learning to drug seeking, it is likely that functional adaptations of these receptors may play a significant role for many of the behaviors manifest during and/or after chronic ethanol exposure.
Acknowledgments
We thank Drs. David Lovinger and Jeff Weiner for helpful editorial comments on this manuscript.
Footnotes
-
This work was supported by (AA13120) and a pilot grant from the Texas A&M University Center for Environmental and Rural Health.
-
DOI: 10.1124/jpet.103.057505.
-
ABBREVIATIONS: BLA, lateral/basolateral amygdala; NMDA, N-methyl-d-aspartate; RT-PCR, reverse transcription-polymerase chain reaction; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; scRT-PCR, single-cell RT-PCR; CE, chronic ethanol; Cao2+, extracellular calcium concentration; GAD, glutamic acid decarboxylase.
-
↵1 Current address: Soonchunhyang University, College of Natural Science, Division of Genetic Engineering, A-san city, Choong-Nam 336-745, South Korea.
- Received July 23, 2003.
- Accepted August 21, 2003.
- The American Society for Pharmacology and Experimental Therapeutics