Differential distribution of voltage-gated calcium channel alpha-2 delta (α₂δ) subunit mRNA-containing cells in the rat central nervous system and the dorsal root ganglia

Animals

The protocols performed in the present study were approved by the Institutional Animal Care and Use Committee of Merck Research Laboratories, in accordance with the guidelines of the National Institutes of Health and the US Department of Agriculture. Adult male Sprague-Dawley albino rats (250–300 g) were obtained from Harlan (Indianapolis, IN) and housed in individual cages under controlled temperature and humidity on a 12:12 light/dark cycle (lights on at 07:00). Food and water were available ad libitum.

Perfusion and tissue processing

Animals were deeply anesthetized with pentobarbital anesthesia (65 mg/kg, i.p.). Each rat was perfused transcardially with normal saline (50–100 ml), followed by ice-cold 4% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA) in 0.1 M borate buffer at pH 9.5 for 20 minutes. The brains were quickly removed, postfixed for 2 hours in the same fixative, then placed overnight in 20% sucrose in 0.02 M potassium PBS (KPBS) for cryoprotection. Five series of 30-μm thick frozen sections through the brain and the lumbar enlargement of the spinal cord were cut using a sliding microtome. Sections were collected at 150-μm intervals into an antifreeze solution and stored at –20°C until use. The L4 or L5 dorsal root ganglia (DRG) were cut at 20 μm on a cryostat, thaw-mounted onto poly-L-lysine-coated slides, and stored at –80°C until use.

In situ hybridization

Sections were mounted onto gelatin-subbed, poly-L-lysine-coated microscopic slides, postfixed with 10% formalin for 30 minutes, and processed for in situ hybridization as described previously (Simmons et al.,1989). To minimize procedural artifacts, all tissue hybridized with each probe was processed together in a single in situ hybridization histochemistry experiment. After a 30-minute proteinase K digestion (10 μg/ml at 37°C; Boehringer Mannheim, Indianapolis, IN) and acetylation (0.0025% acetic anhydride at room temperature), the sections were dehydrated in ascending alcohols and dried under vacuum overnight. T7 polymerase (Promega, Madison, WI) was used to transcribe ³⁵S-labeled antisense cRNA probes. To prepare probes, first-strand cDNA was synthesized from rat DRG total RNA using the RETROscript kit (Ambion, Austin, TX). α₂δ-specific DNA fragments were generated by PCR using PfuTurbo DNA polymerase (Stratagene, La Jolla, CA). For the α₂δ-1 specific probe, a α₂δ-1-specific sense strand 25-mer, 5′-GCACCAAGGGAATACTGCAATGACC-3′, and an antisense strand 27-mer, 5′-CCACCATCATCTAGAATGACACAGTCC-3′ were used to generate a 587-bp fragment (nucleotides 1948–2534 of rat coding sequence, Access. no. AF286488). For the α₂δ-2-specific probe, a α₂δ-2-specific sense strand 30-mer, 5′-GTCTGATGATGACTATGTGAATGTGGCCTC-3′, and an antisense strand 24-mer, 5′-CCAGGGCCATCCTGTGTCAGGTTG-3′ were used to generate a 579-bp fragment (nts 972–1550 of rat coding sequence, Access. no. AF486277). For the α₂δ-3-specific probe, a α₂δ-3 specific sense strand 35-mer, 5′-CCATCGCTGCCAAGTACTCGGGCTCCCAGCTTCTG-3′, and an antisense strand 28-mer, 5′-GCACTTCCAAAGTACTGCCATATGACAG-3′ were used to generate a 487-bp fragment (nts 164–650 of rat coding sequence, Access. no. AF486278). All PCR was performed under the following conditions: 5 minutes at 95°C followed by 35 cycles of 20 seconds at 95°C, 20 seconds at 50°C, and 1 minute at 72°C. The PCR products were cloned into the pCR-Blunt II-TOPO vector (Invitrogen, La Jolla, CA) and linearized at the 5′ or 3′ end of the cloned fragment for use as an in vitro transcription template. The identity of each probe was confirmed by sequence analysis, and each yielded hybridization patterns in accord with available evidence. The specificity of the hybridization was examined by hybridizing the sections with ³⁵S-labeled sense probes with similar specific activities. The radiolabeled cRNA probe was purified by passing the transcription reaction solution over a Sephadex G-50 Nick column (Pharmacia, Piscataway, NJ), and four 100-μl fractions were collected and counted by using a scintillation counter (Packard, Meridian, CT). The leading fraction was heated at 65°C for 5 minutes with 500 μg/ml yeast tRNA (Sigma, St. Louis, MO) and 50 μM dithiothreitol (DTT) (Stratagene) in DEPC (Sigma) water and then diluted to an activity of ∼10⁷ with hybridization buffer containing 50% formamide (Boehringer Mannheim), 0.25 M sodium chloride, 1× Denhardt's solution (Sigma), and 10% dextran sulfate (Pharmacia). This hybridization solution was pipetted onto the sections (80 μl/slide), which were covered with a glass coverslip and sealed with DPX (Electron Microscopy Sciences) before incubation for 20 hours at 58°C. After hybridization, the slides were washed four times (5 minutes each) in 4× SSC before RNase digestion (20 μg/ml for 30 minutes at 37°C; Sigma) and rinsed at room temperature in decreasing concentrations of SSC that contained 1 mM DTT (2×, 1×, 0.5× for 10 minutes each) to final stringency of 0.1× SSC at 75°C for 30 minutes. After dehydration in ascending alcohols, the sections were exposed to X-ray films (BioMax MS; Eastman Kodak, Rochester, NY) for 2–3 days, together with autoradiographic ¹⁴C microscales (Amersham, Arlington Heights, IL), before being dipped in NBT-2 liquid emulsion (Eastman Kodak). The dipped autoradiograms were developed 21 days later with Kodak D-19 developer and the sections were counterstained with thionin.

Combined immunohistochemistry and hybridization histochemistry

Colocalization studies were performed to examine the neurochemical identity of the scattered α₂δ-1 and α₂δ-2 mRNA-containing cells in the hippocampus, the cerebral cortex, the caudoputamen, and some other selected regions. In the hippocampus and cortex the scattered distribution pattern of α₂δ-2 mRNA appeared to be reminiscent of GABAergic interneurons and it was hypothesized that α₂δ-2, but not α₂δ-1, might be selectively expressed in GABAergic interneurons. To address this hypothesis, colocalization of α₂δ-2 or α₂δ-1 with parvalbumin (PV) was performed. PV is a calcium-binding protein that is known to be expressed in a subpopulation of GABAergic interneurons in the hippocampus and cortex, and therefore has been widely used as a surrogate marker for GABAergic interneurons (Miettinen et al.,1996; Sloviter et al.,2001). Furthermore, because α₂δ-2 mRNA was expressed in scattered cells in the caudoputamen and appeared to be reminiscent of the pattern of cholinergic interneurons, colocalization of α₂δ-2 or α₂δ-1 with choline acetyltransferase (ChAT), a marker for cholinergic neurons, was performed. Both ChAT and PV antibodies were obtained from Chemicon (Temecula, CA, nos. AB5042 and MAB1572, respectively). The ChAT polyclonal antibody was raised against a synthetic peptide (H-GLFSSYRLPGHTQDTLVAQKSS-NH₂) from porcine ChAT, coupled to KLH. The PV monoclonal antibody was raised against PV protein purified from frog muscle, and reacts with PV from both brain and muscle. The antibody specifically stains the Ca⁺⁺ bound form of PV. The staining of PV and ChAT immunoreactive cells in this study using these two antibodies was identical to previous reports (Armstrong et al.,1983; Celio,1990) and no extraneous objects were stained. Immunolocalization of PV (1:2,000), or ChAT (1:1,000) was combined with hybridization histochemical detection of α₂δ-1 or α₂δ-2 subunit mRNAs. The protocol for combining immunoperoxidase labeling with isotopic in situ hybridization involves minor modifications (Chan et al., 1993) of the procedure described by Watts and Swanson (1989). Immunohistochemistry was carried out first, using a 48-hour at 4°C primary antibody incubation. A nonimmune (blocking) serum, a potential source of RNAse contamination, was replaced with 2% heparin sulfate and 2% BSA. A conventional avidin-biotin immunoperoxidase protocol (Sawchenko et al., 1990), Vectastain Elite reagents (Vector Laboratories, Burlingame, CA), and a diaminobenzidine reaction were used to visualize PV or ChAT immunoreactivities.

Data collection and analysis

The optical density of the autoradiographic images of α₂δ-1, α₂δ-2, or α₂δ-3 mRNA in different brain regions was measured from x-ray films by using a microcomputer-assisted video imaging densitometer (MCID system, Imaging Research, St. Catherines, Ontario, Canada). The boundaries of different brain regions were determined from observation of the corresponding Nissl-stained sections. The mean optical density over a large irregularly shaped region of the lateral and the third ventricles was also measured and used to calculate the mean background density, which was subtracted from the optical density measurement of signals over different brain regions. Although these mean optical density measurements do not correspond to absolute optical density units, they reflect relative mRNA levels. As mentioned above, commercially available ¹⁴C autoradiographic standards were exposed to each x-ray film along with the experimental material. The mean optical density of an interactively defined region over each standard was measured. The mean optical densities of the autoradiographic images recorded over different brain regions all fell within the linear range of the standard values.

Clustering analysis of the numerical optical density measurements of autoradiographic images of each channel subunit recorded over different brain regions was performed with SciTegic Pipeline Pilot software using a maximal dissimilarity-based partitioning method (Abercrombie,1946; Kaufman and Rousseeuw,1990). Untransformed α₂δ abundance measurements for each subunit were clustered to assign brain regions to qualitative abundance bins. A five-scale rating method was adopted, in which “+” represents close to background, “++” represents low, “+++” represents moderate, “++++” represents high, and “+++++” represents very high optical densities. Ratings are relative within each probe, and because of differences in size, sequence, and isotope-labeling efficiency among probes, the exact quantitative comparisons of subunit expression across probes are treacherous. Thus, comparisons of the expression pattern of different subunits were made based on the relative ratings.

Analysis of cells double-labeled for PV and α₂δ-1 or -2 in the hippocampus was made by standard counting methods. This involved counting the number of cells in the CA1, CA3, and DG of the hippocampus that contained clusters of silver grains whose density was >5 times background. Data was collected from the hippocampal regions as the number of cells that were PV-positive but did not express an α₂δ subunit, those which expressed both PV and an α₂δ subunit, as well as those that were PV-negative, but expressed an α₂δ subunit. Average counts per section were generated from five sections through the hippocampal region of interest in each animal, then averaged again to obtain group means. Abercrombie's (1946) method was used to correct for double-counting errors. Double-labeling in other brain regions were estimated by counting single and double-labeled cells under 20× or 40× objectives at representative levels in a complete series.

Images from representative levels of the brain were scanned from the x-ray film using a desktop scanner and photomicrographs taken using an Olympus microscope. All images were processed with Adobe PhotoShop software (San Jose, CA) for optimal sharpness, contrast, and brightness, and plotted using Freehand software (Macromedia, San Francisco, CA).

Expression of α₂δ subunit mRNA in the rat CNS and DRG

The α₂δ-1, α₂δ-2, and α₂δ-3 mRNA-containing cells were found to be widely distributed throughout the rat brain, the spinal cord, and the DRG (Figs. 1-3, 8, 9), including areas involved in somatosensory, motor, autonomic, and neuroendocrine functions. The three subunits showed unique distribution patterns within each major subdivision of the brain as well as the spinal cord (Table 1), and appeared to be expressed in the majority of, if not all, neurons because each cell examined at high power showed typical neuronal morphology in Nissl counterstaining, and no significant labeling was found over fiber tracts. Similarly, in the DRG all labeling was localized to neurons and no satellite cells appeared to express α₂δ mRNA. Specific signals were obtained in all experiments, as silver grains were not detected over cells when sections were hybridized with isotope-labeled sense strand probes (data not shown). Moreover, the specificity of the labeling in the present study is also supported by our results in mouse brain using the same probes (data not shown), which is in good agreement with data previously reported in the literature (Hobom et al.,2000).

Distribution of cells containing α₂δ-1 mRNA

Cerebrum (telencephalon).

The most extensive labeling of α₂δ-1 mRNA was observed in the cortical regions, with the heaviest labeling over neurons in areas involved in olfaction, learning and memory, such as the piriform and the entorhinal cortices, postpiriform transition area, and the hippocampal formation (Figs. 1A–G, 4A,B; Table 1). A low (layers IV–VI) to moderate (layers II–IV) density of labeled neurons was found in each field of the isocortex, but a higher density of labeling appeared to be over neurons in layer II of the piriform cortex. The entorhinal cortex, one of the major regions that have strong connections with the hippocampal formation, contained a high to very high density of labeled neurons in layers II and III. In the hippocampus the pyramidal layer of each subdivision of Ammon's horn (CA1–CA3) was highly labeled, but only a low level of labeling was seen in the dorsal and ventral subiculum. Moreover, the granular layer of the dentate gyrus contained heavily labeled neurons. A low level of signals was also found in the polymorph layer of the dentate gyrus (Fig. 4B). The main olfactory bulb and the anterior olfactory nucleus contained a high density of α₂δ-1 mRNA-containing cells (Fig. 1A), although only a low level of the signal was detected in the endopiriform nucleus. Likewise, a high density of α₂δ-1 mRNA-containing cells was found in the dorsal part of the taenia tecta (Fig. 1B), a structure possibly differentiated from the adjacent anterior olfactory nucleus (Swanson,1998).

Overall, the amygdala contained many labeled neurons, with the heaviest labeling observed in regions that receive cortical or hippocampal inputs, such as the amygdalohippocampal area, the cortical nucleus, and the basolateral and basomedial nuclei (Fig. 1E–G; Table 1). We also observed a moderate density of labeled neurons in the intercalated nucleus that receives intra-amygdalar inputs. Only a low level of α₂δ-1 mRNA signal was seen in the medial and the central nuclei of the amygdala.

Except for low to moderate levels of α₂δ-1 mRNA expression in the bed nucleus of the anterior commissure and the triangular nucleus of the septum, the septal region, including the lateral and medial septal nuclei, and the bed nuclei of the stria terminalis expressed very low levels of α₂δ-1 mRNA (Fig. 1C–E; Table 1).

The most prominent labeling in the basal ganglia was found in the subthalamic nucleus (Fig. 1H; Table 1). The caudoputamen and the fundus of the striatum expressed a low level of α₂δ-1 mRNA (Figs. 1B–G, 5A; Table 1), while the nucleus accumbens appeared to have a moderate level of the labeled cells (Fig. 1C). Except for the low level of labeled neurons in the magnocellular preoptic nucleus that highly expressed acetylcholine in the basal forebrain, the other parts of the globus pallidus including the entopeduncular nucleus, the internal segment, and the substantia innominata contained only a few sporadic, if any, α₂δ-1 mRNA-containing neurons. Similar results were also observed in the pars reticularis of the substantia nigra (Fig. 1E–G; Table 1).

Hypothalamus.

A high to very high density of labeled neurons were found in three nuclei in the periventricular zone of the hypothalamus; these were the median preoptic nucleus, the supraoptic nucleus, and the paraventricular nucleus, including its autonomic, parvicellular, and magnocellular parts (Figs. 1D–H, 6A; Table 1). Low to moderate levels of labeling were observed in the periventricular zone including the anteroventral periventricular nucleus and the arcuate nucleus (Fig. 6B; Table 1).

The greatest density of labeling in the hypothalamus was found in the ventromedial nucleus in the medial zone (Figs. 1G, 6B; Table 1). Strongly labeled neurons were also found in the dorsomedial, tuberomammillary, and pre- and supramammillary nuclei. The medial preoptic nucleus, the anterior hypothalamic nucleus, and the lateral and medial mammillary nuclei appeared to express low to moderate levels of α₂δ-1 mRNA (Fig. 1E–H; Table 1).

Only a few labeled neurons were found in the lateral hypothalamic and posterior hypothalamic areas in the lateral zone. Sporadically labeled cells were found in the lateral preoptic area.

Cerebellum.

A high density of labeled neurons was found over regions of the granular cell layer. Sporadically labeled cells were observed in the fastigial nucleus and the parvicelluar part of the interposed nucleus. Signals over regions of the molecular layer and Purkinje layer, as well as the rest of the interposed nucleus and the dentate nucleus, were not detectable (Fig. 7A; Table 1).

Spinal cord and DRG.

In the spinal cord a moderate level of labeling was found in the lamina I and the lamina II of the dorsal horn. However, the emulsion-coated autoradiograms also showed labeled cells scattered in the remaining laminae of the spinal gray, while only a few motor neurons in the ventral horn appeared to be labeled (Fig. 8A,B; Table 1).

Nearly all neurons with different diameters in the DRG were found to express α₂δ-1 mRNA, but the highest levels of expression appeared to be in the large-diameter neurons (Fig. 9A).

Distribution of cells containing α₂δ-2 mRNA

Distribution of cells containing α₂δ-3 mRNA

Characterization of the chemical identity of α₂δ-1 and α₂δ-2 mRNA-containing cells

Colocalization of α₂δ-1 and α₂δ-2 subunits with parvalbumin (PV), a marker for GABAergic neurons, and choline acetyltransferase (ChAT), a marker for cholinergic neurons, were used to assess the neurochemical identity of these cells.

Approximately 75% of PV-positive neurons coexpressed α₂δ-2 mRNA in the pyramidal layer of the Ammon's horn in the hippocampus. In contrast, this population of cells did not express α₂δ-1 mRNA. Likewise, α₂δ-2 mRNA was present in the majority of PV-positive neurons within the cortex, while very little α₂δ-1 expression was detected in this cell population (Fig. 10). In addition, all PV-positive Purkinje cells in the cerebellum coexpressed α₂δ-2 mRNA. In other brain regions, a high percentage of α₂δ-2 and PV coexpression, ranging from 55–95%, was found in the lateral and basolateral nuclei of the amygdala, the lateral and medial septal nuclei, the nucleus of the diagonal band, the magnocellular preoptic nucleus, the reticular nucleus of the thalamus, the substantia innominata, and the intermediate subnucleus of the interpeduncular nucleus. Approximately 40% of the PV-positive cells in the caudoputamen coexpressed α₂δ-2 mRNA. In all of these regions, very few, if any, PV-positive cells appeared to coexpress α₂δ-1 mRNA.

Within the caudoputamen, the vast majority of the ChAT-positive cells, if not all, coexpressed α₂δ-2 mRNA, but did not coexpress α₂δ-1 mRNA (Fig. 11A,C). However, only a few ChAT-positive cells along the lateral border of the magnocellular preoptic nucleus appeared to coexpress α₂δ-2 mRNA, but none of the ChAT-positive cells in this nucleus coexpressed α₂δ-1 mRNA (Fig. 11B,D). Similarly, only ∼20–30% of ChAT-positive cells in the medial septal nucleus and the nucleus of the diagonal band coexpressed α₂δ-2 mRNA, although α₂δ-2 mRNA was generally highly expressed in these regions. However, a high percentage, ranging from 50–90%, of the ChAT-positive cells in the islands of Calleja, the substantia innominata, the medial segment of the globus pallidus, the medial habenula, the laterodorsal tegmental nucleus, and the pedunculopontine nucleus appeared to coexpress α₂δ-2 mRNA. Moreover, the majority of the ChAT-positive motor neurons in the third, fifth, seventh, tenth, and twelfth cranial nuclei, as well as those in the nucleus ambiguus and the spinal ventral horn, coexpressed α₂δ-2 mRNA. Notably, only a small portion of ChAT-positive cells in the lamina X and the lateral horn of the spinal cord coexpressed α₂δ-2 mRNA. In addition, α₂δ-2 mRNA did not appear to be coexpressed in the ChAT-positive cells in the intermediate subnucleus of the interpeduncular nucleus.

Distribution of α₂δ-1, α₂δ-2, and α₂δ-3 mRNA-containing cells

High-voltage-activated calcium channels are extensively distributed in the CNS. Consistent with this fact, α₂δ-1, α₂δ-2, and α₂δ-3 mRNA-containing cells are widely found throughout the rat brain and the spinal cord. The distribution pattern of the mRNAs encoding the α₂δ-1 and α₂δ-2 subunits are often observed to be complementary in rat brain regions, suggesting differential regulatory mechanisms underlying the modification of the biophysical and pharmacological properties of the pore-forming α₁ subunit in different brain regions. In areas where α₂δ-1 expression is high, α₂δ-2 expression tends to be low, and vice versa. For example, in regions such as the cortex and in the pyramidal cell layers of the hippocampus, α₂δ-1 expression is high and α₂δ-2 expression is low. In contrast, regions including the septum, the reticular nucleus of the thalamus, the medial habenula, and the arcuate nucleus show high levels of α₂δ-2 mRNA expression and low levels of α₂δ-1 mRNA expression. A similar phenomenon is also observed within the amygdala and the cerebellum. However, there are a few brain regions which express high levels of both α₂δ-1 and α₂δ-2 including the median preoptic nucleus, the subthalamic nucleus and the premammillary nucleus. Expression of α₂δ-3 mRNA is observed more ubiquitously, although it can be found in distinct patterns that overlap with either α₂δ-1 or α₂δ-2 mRNA expression in many brain areas. In some regions all three subunits can be found in the same nucleus. This raises the possibility that different heterooligomeric combinations of the auxiliary subunits exist. Alone, the α₂δ-1, α₂δ-2, or α₂δ-3 subunits have been shown to shift the voltage dependence of channel activation and inactivation in a hyperpolarizing direction and accelerate the kinetics of current inactivation when coexpressed with an α₁ subunit such as α₁C (Klugbauer et al.,1999; Hobom et al.,2000). However, the function of different combinations of these auxiliary units, especially with regard to binding drugs such as gabapentin (Marais et al.,2001), is yet to be determined. In instances when two or three subunits are found in the same brain region, it is also unknown if the different α₂δ subunits exist in the same or different cell populations.

Extensive labeling of α₂δ-1 and α₂δ-3 mRNA is found in the cortex, the hippocampus, and the amygdala. The α₂δ-2 mRNA is considerably less abundant in these regions and appears to be expressed in a specific cell group. Regions involved in somatosensory, visual, auditory, visceral, and motor functions in the cortex, as well as the association cortices all contain a significant amount of α₂δ-1 mRNA-containing cells, suggesting that the α₂δ-1 subunits may be essential in multiple cortical neuronal functions. Similarly, extensive and heavy labeling of α₂δ-1 mRNA is found in the hippocampal formation, including each subdivision of Ammon's horn, the dentate gyrus, and in the perientorhinal region that receives extensive cortical inputs and provides major reciprocal connections with the hippocampus. Because α₂δ-1 mRNA does not appear to coexpress with PV in the hippocampus or the cortex, it is likely that most of the above α₂δ-1 mRNA-containing neurons are glutamatergic. Although neuronal activity in the hippocampus, mediated by VGCCs containing the α₂δ-1 subunit, may play an important role in learning and memory in the normal physiological state, abnormal firing, due to brain trauma, infection, or tumor, of neurons in the hippocampus, cortex, or amygdala has been found to be attributable to the onset of seizures (Nestler,2001; McIntyre et al.,2002). Thus, consistent with our results, gabapentin and its analogs, which specifically bind to α₂δ-1 (Gee et al.,1996) and are thought to reduce calcium currents (Stefani et al.,1998; Calabresi et al.,1999; Fink et al.,2000), are efficacious against focal epilepsy in adult humans and in animal models of focal seizures (Dalby and Nielsen,1997; Maj et al.,1999; Lado et al.,2001). Also consistent with our results, radiolabeled gabapentin binds heavily to the rat cortex, the hippocampus, and the amygdala in rats (Gu et al., unpubl. obs.; Hill et al.,1993). The presence of α₂δ-2 mRNA in PV-containing cells suggests that VGCCs containing the α₂δ-2 subunit may play a role in the activation of GABAergic interneurons in the cortex, the hippocampus, and the amygdala. Although gabapentin and its analogs also bind to the α₂δ-2 subunit, the antiepileptic effect of gabapentin does not seem to be mediated through the binding to the α₂δ-2 subunit, because gabapentin, especially its analog pregabalin, has much higher affinity for the α₂δ-1 than the α₂δ-2 subunit (Bryans et al.,1998; Marais et al.,2001). Moreover, reduction of calcium currents through VGCCs containing the α₂δ-2 subunit, based on our findings, would be expected to suppress the activities of GABAergic neurons and therefore compromise the antiepileptic effect. Because no specific ligand has been found for the α₂δ-3 subunit and gabapentin and its analogs do not seem to bind to this subunit with high affinity, the outcome of suppressing the α₂δ-3 subunit in the cortex, the hippocampus, and the amygdala remains unclear. However, the α₂δ-3 subunit may serve as a better target for epilepsy because it is only expressed in the brain and, like α₂δ-2, the α₂δ-3 subunit is involved in the enhancement and modulation of the currents through several different high-voltage-activated (HVA) and low-voltage-activated (LVA) Ca⁺⁺ channels (Klugbauer et al.,1999; Gao et al.,2000; Hobom et al.,2000; Marais et al.,2001). The low level of expression of the α₂δ-1 subunit in the thalamus suggests that the thalamus may not be the primary target for the antiepileptic effect of gabapentin and its analogs, but the high expression of both α₂δ-2 and α₂δ-3 mRNA in the GABAergic reticular nucleus suggests that VGCCs containing these two subunits may be involved in the control of information outflow from the thalamus.

More evidence has emerged that the binding of gabapentin to the α₂δ-1 subunit may be responsible for its efficacy in the treatment of neuropathic pain (Gee et al.,1996; Bramwell,2004), although the mechanism of action has yet to be determined. While some pathological changes have been found in the cortex and the thalamus in neuropathic pain (Apkarian et al.,2004; Weigel and Krauss,2004), it is surprising to find that α₂δ-1 and α₂δ-2 mRNA expression is low in the somatosensory relay nuclei in the thalamus. Nevertheless, a heavy density of α₂δ-1 and α₂δ-2 mRNA-containing cells is found in the lateral parabrachial nucleus, an area believed to convey spinal noxious information to the hypothalamus and the amygdala (Craig,2003), and also in the superficial layers of the spinal dorsal horn. The spinal cord results are consistent with the observations in multiple gabapentin binding studies (Gu et al., unpubl. obs.; Thurlow et al.,1996), biochemical and behavioral investigations showing that the α₂δ-1 subunit in the spinal cord is significantly altered in the pain states (Luo et al.,2002; Li et al.,2004), and the fact that intrathecal injection of gabapentin is efficacious in suppressing nociception in different animal pain models (Shimoyama et al.,1997; Patel et al.,2001; Kumar,2004). Although α₂δ-3 mRNA labeling is low in the thalamic somatosensory relay nuclei and appears to be scant in the lateral parabrachial nucleus, it is highly expressed in the superficial layers of the spinal dorsal horn. In addition to the lateral parabrachial nucleus and the spinal dorsal horn, the dorsal root ganglia may be another important target for gabapentin because all three α₂δ subunits are heavily expressed in neurons of all diameters and a high level of gabapentin binding is found in this region (Gu et al., unpubl. obs.).

Consistent with the high expression of α₂δ-1 and α₂δ-2 mRNA in the visceral sensory areas, heavily labeled α₂δ-1 and α₂δ-2 mRNA-expressing cells are also found in the visceral motor structure—the dorsal motor nucleus of the vagus nerve. The somatic motor neurons of the spinal cord express a high level of α₂δ-3 mRNA. A moderate level of α₂δ-2 and low level of α₂δ-1 mRNA are also seen in these cells. Nevertheless, the high level of α₂δ-2 mRNA in ChAT-positive motor neurons indicates that VGCCs containing the α₂δ-2 subunit may play an important role in somatic as well as visceral motor functions. The inferior olive, a structure in the brainstem that sends strong afferents (climbing fibers) to Purkinje cells in the cerebellar cortex, expresses a moderate level of α₂δ-1 mRNA. Purkinje cells, however, do not appear to express α₂δ-1 mRNA, but rather express high levels of α₂δ-2 mRNA. The expression of the three α₂δ subunits in the cerebellum appears to be differential. This pattern suggests that different VGCCs that are associated with different α₂δ subunits may play roles in motor control and other cerebellar functions at multiple levels (Manto and Padolfo,2002).

An extensive and unique distribution of α₂δ-1, α₂δ-2, and α₂δ-3 mRNA-containing neurons was observed in the periventricular, the medial, and the lateral zones of the hypothalamus, suggesting their broad involvement in neuroendocrine, homeostatic, motivated behavior, and autonomic functions. For example, regions in the hypothalamus known to mediate defensive behavior, such as the dorsomedial part of the ventromedial nucleus and the dorsal part of the premammillary nucleus of the hypothalamus (Canteras,2002), contained a high density of α₂δ-1 mRNA-containing neurons. Similarly, the parvicellular part of the paraventricular nucleus that contains corticotropin-releasing hormone (CRH), the autonomic part of the paraventricular nucleus that activates autonomic motor neurons (Swanson and Kuypers,1980), and the dorsomedial nucleus that is associated with behavioral responses to stressors (Inglefield et al.,1994; Thompson and Swanson,1998) also contained a high density of α₂δ-1 mRNA-containing neurons. Structures governing fluid homeostasis, such as the median preoptic nucleus, as well as the circumventricular organs—the subfornical organ and the area postrema—express high levels of both α₂δ-1 and α₂δ-2 mRNA, with a relatively higher density of α₂δ-2 mRNA-containing cells. Furthermore, heavily labeled α₂δ-2 mRNA-containing neurons were found in the suprachiasmatic nucleus, the arcuate nucleus, the lateral hypothalamic area, the anterior hypothalamic nucleus, and the dorsal part of the premammillary nucleus, suggesting that VGCCs containing the α₂δ-2 subunit may play an important role in the regulation of circadian rhythm, ingestion (Sawchenko,1998), and defensive behaviors. Likewise, the high level of expression of α₂δ-3 mRNA in the suprachiasmatic nucleus strongly suggests the important role of α₂δ-3 in the regulation of circadian rhythm. The expression of α₂δ-1, α₂δ-2, and α₂δ-3 mRNAs in the anteroventral periventricular nucleus, the ventral part of the premammillary nucleus, the medial preoptic nucleus, and the ventrolateral part of the ventromedial nucleus of the hypothalamus suggests the involvement of VGCCs containing these subunits in the control of gonadotropin release and the expression of sexual behavior (Pfaff and Schwartz-Giblin,1988; Simerly and Swanson,1988; Canteras et al.,1992,1994; Gu and Simerly,1997). The supramammillary nucleus relays information to the hippocampus and was recently found to be involved in the genesis and spread of limbic seizures in the hippocampus (Saji et al.,2000). Heavily labeled α₂δ-1 mRNA-containing cells were found in this nucleus. This structure could be another possible target for gabapentin in the prevention of seizures.

Both α₂δ-1 and α₂δ-2 mRNAs are highly expressed in the locus coeruleus and the dorsal raphe in the reticular core of the brainstem. These two nuclei send massive ascending outputs containing norepinephrine and serotonin, respectively, to the cortex and the intralaminar nuclei of the thalamus and are involved in arousal and selective attention. Moderate levels of α₂δ-1 and α₂δ-2 mRNA are also expressed in the pedunculopontine nucleus, an interface for the basal ganglia to influence sleep and waking (Mena-Segovia et al.,2004). In addition, heavily labeled α₂δ-2 mRNA-containing neurons are found in the laterodorsal tegmental nucleus that, like the pedunculopontine nucleus, contains cholinergic neurons and is also involved in the sleep and waking (Kayama and Koyama,2003). Moreover, high expression of α₂δ-2 mRNA in the nucleus incertus also supports the involvement of VGCCs containing this subunit in arousal mechanisms (Goto et al.,2001; Olucha-Bordonau et al.,2003). High density of α₂δ-2 mRNA-containing neurons are also found in the interpeduncular nucleus. This nucleus is thought to play an important role in the regulation of normal sleep patterns (Haun et al.,1992). The high level of expression of α₂δ-2 mRNA in this nucleus and in other sleep/waking structures may be associated with the severe somnolence caused by gabapentin administration.

Functional implications

It is thought that a single α₂δ subtype may complex with different α₁ subunits (Hobom et al.,2000) since the expression patterns of α₂δ subunits do not suggest a selective interaction with known α₁ subunits. However, as summarized by Hobom et al. (2000) based on the individual mapping of α₂δ subunit and α₁ subunit, at least the α₂δ-2 subunit could complex with the α₁A, α₁E, and/or α₁G channel in cerebellar Purkinje cells, and the α₁E and/or α₁I channel in the habenulae. Moreover, α₂δ-2 may complex with α₁E in the subthalamic nucleus (Trimmer and Rhodes,2004), and the α₂δ-2 or α₂δ-3 subunit may complex with α₁B in the hippocampus (Chung et al.,2000), α₁C, D, and E in the motor neurons in the spinal cord (Westenbroek et al.,1998). Indeed, detailed colocalization studies are needed to clarify this issue. Nevertheless, current understanding of the functional role of α₂δ subunits is largely based on in vitro coexpression studies with members of the L-type, non-L-type, and T-type calcium channels. Overall, α₂δ subunits modulate channel activity by increasing calcium current density, shifting the voltage dependence of activation to more negative potentials, or increasing steady-state inactivation (see Klugbauer et al.,2003, for review).

While the exact in vivo functional significance of the differential distribution of the α₂δ subtypes in the CNS remains to be explored, some insight has been gained through the observation of spontaneous mutation mouse models. Mutations in the Cacna2 (α₂δ-2) gene have been found to underlie the phenotype in the spontaneously mutant ducky mouse (du and du^2j strain) (Barclay et al.,2001; Brodbeck et al.,2002) and the newly identified entla mutant (Brill et al.,2004). These animals exhibit ataxia, seizures, and infertility, suggesting that VGCCs containing α₂δ-2 subunit play a pivotal role in neural development and function. The R217A mutation in the α₂δ-1 subunit results in the loss of efficacy of gabapentin and pregabalin in a model of neuropathic pain (Bramwell et al.,2004), indicating the involvement of VGCCs containing the α₂δ-1 subunit in the mediation of neuropathic pain.

The present study reveals that α₂δ-2 is expressed in many neuronal structures that are known to be GABAergic, and thus, VGCCs containing the α₂δ-2 subunit may play a specific role at central inhibitory synapses. These structures include the reticular nucleus of the thalamus, the zona incerta, the septum, cerebellar Purkinje cells, as well as hippocampal and cortical PV-positive interneurons. Interestingly, while α₂δ-2 expression is highly prominent in GABAergic neurons in the above regions, it is not highly present in the medium-sized spiny GABAergic neurons of the striatum. Only less than half of the PV-positive cells in the striatum coexpresses α₂δ-2 mRNA. In this structure, α₂δ-2 is highly localized to the large cholinergic interneurons. Because the cholinergic interneurons send their projections to and excite the GABAergic neurons of the striatum (Nestler,2001), activation of VGCCs containing the α₂δ-2 subunit on cholinergic interneurons augments the inhibitory output from the striatum.

These results showing selective expression of α₂δ-2 in GABAergic structures, may contribute to understanding the neurochemical basis of the ducky (du) mouse phenotype mentioned above. Ducky is a spontaneous autosomal recessive mouse mutation in the α₂δ-2 VGCC subunit gene, which results in the expression of a truncated, nonfunctional α₂δ-2 protein (Barclay et al.,2001; Brodbeck et al.,2002). Mice carrying homozygous ducky (du) alleles have a phenotype that includes epileptic seizures with features similar to those occurring in human idiopathic generalized epilepsy (Puranam and McNamara,1999), as well as ataxia and paroxysmal dyskinesia (Snell,1955). The loss of α₂δ-2 subunit protein on hippocampal and cortical GABAergic terminals could underlie the seizure phenotype. As α₂δ-2 is selectively localized on GABAergic interneurons in the hippocampus and cortex, as well as the reticular nucleus of the thalamus, loss of this protein could result in a decrease in GABAergic inhibition, and thus an overly excitable state. The ataxia and paroxysmal dyskinesia associated with the phenotype in ducky (Snell,1955) may result from loss of inhibitory/modulatory influences in motor circuitry, especially from Purkinje cells of the cerebellum. It is well documented that degenerative changes of the cerebellar Purkinje cells caused by CAG repeats lead to spinocerebellar ataxia type-1, and the disturbance of calcium homeostasis has been linked to this autosomal dominant disorder (Vig et al.,2001). In agreement with this observation, the altered morphology of Purkinje cell dendrites has been found to be one of the anatomical correlates of the du phenotype (Brodbeck et al.,2002), and the peak P-type calcium current shows a 35% decrease in Purkinje cells (Barclay et al.,2001).

α₂δ-2 mRNA is also highly expressed in several other brain nuclei known to be cholinergic, including the nucleus of the diagonal band, the magnocellular preoptic nucleus, the medial septal nucleus, the nucleus of the diagonal band, the substantia innominata, the pedunculopontine nucleus, the laterodorsal tegmental nucleus, and the motor neurons in the cranial nuclei and spinal cord. However, colocalization studies with ChAT show that α₂δ-2 is rarely found in the cholinergic neurons in the magnocellular preoptic nucleus, the medial septal nucleus, and the nucleus of the diagonal band, indicating its expression in other cell groups in these nuclei. Thus, the α₂δ-2 subunit is highly expressed by cholinergic interneurons in the striatum, but not in the principle cholinergic neurons in the basal forebrain that provides cholinergic inputs to the cortex. Because cholinergic interneurons in the striatum play an important role in the control of movement, it is not surprising that mutating the α₂δ-2 subunit causes dyskinesia. The high expression of α₂δ-2 mRNA in almost all motor neurons suggests that the α₂δ-2 subunit may be essential in somatic and visceral movements, which may also help explain why dyskinesia occurs when the α₂δ-2 subunit is mutated.