Abstract
The distribution of the α4-subunit of the neuronal nicotinic acetylcholine receptor (nAChR) in the rat brain was examined at light and electron microscopy levels using immunohistochemical staining. In the present study we demonstrate the specificity, in both tissue homogenates and brain sections, of a polyclonal antibody raised against the rat nAChR α4-subunit. The characterization of this antibody involved: (1) Western blot analysis of rat brain homogenates and membrane extracts from cells previously transfected with diverse combinations of neuronal nAChR subunits, and (2) immunohistochemistry using transfected cells and rat brain tissue.
At the light microscope level, the α4-subunit-like-immunoreactivity (LI) was widely distributed in the rat brain and matched the distribution of the α4-subunit transcripts observed previously by in situhybridization. Strong immunohistochemical labeling was detected in the mesencephalic dopaminergic nuclei. The nAChRs in this region are thought to be responsible for the modulation of dopaminergic transmission. The neurotransmitter identity of α4-immunolabeled neurons in the substantia nigra pars compacta (SNpc) and the ventral tegmental area was thus assessed by investigating the possible colocalization of the nAChR α4-subunit with tyrosine hydroxylase using confocal microscopy. The double labeling experiments unambiguously indicated that the α4-subunit-LI is present in dopaminergic neurons.
At the electron microscope level, the neurons in the SNpc exhibited α4-subunit-LI in association with a minority of postsynaptic densities, suggesting that the α4-subunit may be a component of functional nAChRs mediating synaptic transmission between midbrain cholinergic neurons and mesencephalic dopaminergic neurons.
- neuronal nAChR
- immunohistochemistry
- immunogold
- postsynaptic localization
- substantia nigra
- dopaminergic nuclei
- rat brain
The nicotinic acetylcholine receptor (nAChR) is a pentameric protein that forms a cation-selective channel at the neuromuscular junction (Changeux, 1991) and in the peripheral nervous system and CNS (Sargent, 1993). In rat brain, the neuronal nAChRs comprise at least six α subunits, named α2–α7, and three β subunits named β2–β4 (Le Novère and Changeux, 1995). Both types of subunits can contribute to the various pharmacological profiles of neuronal nAChRs and may form a number of functionally different hetero-oligomers (Boulter et al., 1987; Luetje and Patrick, 1991; Conroy et al., 1992; Vernallis et al., 1993;Bertrand and Changeux, 1995; Conroy and Berg, 1995; Ramirez-Latorre et al., 1996; Role and Berg, 1996; Wang et al., 1996; Ragozzino et al., 1997).
In situ hybridization (ISH) studies revealed that the neuronal nAChR subunits mRNAs display diverse, yet overlapping expression patterns in the CNS (Wada et al., 1989, 1990; Duvoisin et al., 1989; Dineley-Miller and Patrick, 1992; Seguela et al., 1993). Immunohistochemistry provides further information on the topological distribution of neuronal nAChR protein subunits at the cellular and subcellular levels. Neuronal nAChRs have been mapped using a variety of antibodies on tissue sections of rat brain (Mason, 1985; Swanson et al., 1987; Schroder et al., 1989; Bravo and Karten, 1992; Okuda et al., 1993; Dominguez del Toro et al., 1994; Nakayama et al., 1995; Goldner et al., 1997; Rogers et al., 1998; Sorenson et al., 1998). For instance, the staining patterns obtained using antibodies against the β2-subunit (Deutch et al., 1987; Hill et al., 1993) parallel those observed with [3H]nicotine and [3H]acetylcholine (ACh) binding (Clarke et al., 1985b).
Several lines of evidence indicate that activation of the mesotelencephalic dopaminergic systems is involved in the reinforcing properties of nicotine (Imperato et al., 1986; Corrigall et al., 1992;Pontieri et al., 1996) as well as of several other drugs of abuse, such as opiates, cocaine, amphetamine, and ethanol (Koob, 1992, 1996). In the case of nicotine, activation of dopaminergic systems is thought to be principally mediated by nAChRs located in the mesencephalon (Nisell et al., 1994). Indeed, electrophysiological experiments have shown that nicotine can activate dopaminergic (DA) neurons in preparations of rodent mesencephalon (Clarke et al., 1985a; Pidoplichko et al., 1997;Picciotto et al., 1998; Sorenson et al., 1998).
Little information is available, however, on the exact cellular localization and subunit composition of the nAChRs responsible for the effects of nicotine. Neurochemical-selective lesions in rats indicated that[3H]nicotine binding sites are specifically associated with DA neurons (Clarke and Pert, 1985; Clarke et al., 1985b). Also [3H]cytisine labeling, another high-affinity agonist at nicotinic receptors, was found on DA neurons (Happe et al., 1994). Antibodies against the α4-subunit were able to immunoprecipitate receptors labeled either by [3H]nicotine (Whiting and Lindstrom, 1987, 1988) or [3H]cytisine (Flores et al., 1992) binding. The substantia nigra pars compacta (SNpc) and ventral tegmental area (VTA) contain moderate to high levels of the α3, α4, α5, α6, β2, and β3 nAChR-subunit mRNAs (Wada et al., 1989, 1990; Deneris et al., 1989; Dineley-Miller and Patrick, 1992; Le Novère et al., 1996).
Electrophysiological experiments have shown that nAChRs with putative α4–β2 and α7 compositions are present on DA cell bodies of the rodent mesencephalon (Pidoplichko et al., 1997; Picciotto et al., 1998;Sorenson et al., 1998). Neurochemical studies further suggest the existence of α3- or α6-containing nAChRs on DA nerve terminals (for discussion, see Le Novère et al., 1996).
We demonstrate here the colocalization of tyrosine hydroxylase (TH) and α4-subunit-like-immunoreactivity (LI) in mesencephalic DA neurons and provide new ultrastructural data using both immunoperoxidase and immunogold techniques on the subcellular localization of α4-containing nAChRs in the SNpc.
MATERIALS AND METHODS
Antibodies. We used a polyclonal antibody (catalog #1772, lot #D256; Santa Cruz Biotechnology, Santa Cruz, CA) raised in goat against a synthetic peptide corresponding to a portion of the intracellular domain of the rat (Rattus norvegicus) nAChR α4-subunit. The immunogenic peptide is a 20 amino acid sequence comprising residues 573–592 of the cDNA deduced sequence α4-subunit (RAVEGVQYIADHLKAEDTDF), which corresponds to the highly variable cytoplasmic loop between the membrane-spanning domains M3 and M4. This antibody should recognize both identified splice variants, α4–1 and α4–2, because the sequence used for immunization is localized upstream from the splicing region of the cDNA sequence according to Goldman et al. (1987).
The specificity of this antibody was assessed by several methods in both tissue homogenates and sections.
Western blot analysis in rat homogenates. Whole extracts from three rat brains and lungs were analyzed separately by six Western blots. Tissues were homogenized in five volumes of boiling lysis buffer (1% SDS, 10 mm Tris-HCl, pH 7.4) and centrifuged at 550 × g for 10 min. Supernatant was collected, aliquoted, and frozen at −80°C until use.
Fifty microgram aliquots of either sample were separated by SDS-PAGE (10% gels). Proteins were transferred to nitrocellulose membranes, blocked overnight with 5% nonfat dry milk in TBST buffer (10 mm Tris-HCl, pH 7.5, 100 mmNaCl, and 0.1% Tween 20) at 4°C and then incubated at room temperature (RT) with the anti-α4 antibody diluted 1:5000 (0.04 μg/ml) in the same blocking buffer for 1 hr, washed with TBST, and incubated with a peroxidase-conjugated rabbit anti-sheep antibody (Cappel, West Chester, PA) diluted 1:5000 (9.9 μg/ml) for 1 hr, after which membranes were washed again. Bound peroxidase was detected using enhanced chemiluminescence (ECL; Amersham). As a control, an identical gel was run in parallel with the anti-α4 antibody preadsorbed at 37°C with a 10-fold molar excess (2 μg/ml) of the synthetic peptide (Santa Cruz Biotechnology) for 3–4 hr. No immunostaining was observed under these circumstances.
Western blot and immunohistochemistry in human embryonic kidney cells transfected with nicotinic receptor subunit cDNAs. Human embryonic kidney (HEK-293) cells were transfected with cDNAs corresponding either to the rat α3, α4, and β4, or the human α4 and β2, or (for control purposes) the chimeric α7-V201–5HT3 receptor (Eiselé et al., 1993) by the calcium phosphate DNA precipitation method as previously described (Chen and Okayama, 1987).
The level of expression of nicotinic receptors in this transient system was assessed by an equilibrium binding assay. Briefly, cells were harvested as previously described (Corringer et al., 1995) and incubated for 30 min at RT with [3H]epibatidine (0.4 nm), in the absence or presence of 0.5 mmunlabeled nicotine to measure nonspecific binding. The level of expression was typically 0.2 pmol of binding sites per confluent 10 cm dish. The cell membrane fraction was obtained for subsequent use in Western blot analysis.
Cells were recovered and homogenized using an antiprotease buffer [0.1m PBS, pH 7.4, 50 mm EDTA, and 0.5 mm phenylmethylsulfonyl fluoride (PMSF)]. Pellets were resuspended after centrifugation at 600 × g for 10 min in cold buffer (50 mm Tris–acetate, pH 7.4, 50 mm EDTA, 0.5 mm PMSF, and 0.2 U/μl aprotinine). DNA was eliminated by centrifugation at 400 ×g for 10 min. Supernatants were centrifuged at 12,000 × g for 1 hr to pellet the membranes. Forty microgram aliquots were run in SDS-PAGE gels and developed as indicated above.
The immunohistochemical reaction was performed after fixation of cell cultures with 3% paraformaldehyde for 30 min at RT. After rinsing, free aldehyde radicals were neutralized in 50 mmNH4Cl, 1 mm glycine, and 1 m lysine in 0.1 m PBS buffer, pH 7.4, for 20 min at RT. Cells were then blocked in PBS buffer containing 5% normal horse serum for 30 min and then incubated at 4°C with the anti-α4-subunit nAChR antibody (Santa Cruz Biotechnology) diluted 1:1000 (0.2 μg/ml) in PBS containing 1% normal horse serum overnight. After three 10 min washes in PBS, cells were incubated at RT with Cy3-conjugated anti-goat IgG (Amersham, Pittsburgh, PA) diluted 1:500 (2 μg/ml). Controls for method specificity consisted in omission of the primary antibody. The use of nontransfected cells and of α7-V201–5HT3(Eiselé et al., 1993) transfected cells in the same conditions yielded no immunostaining. For the peptide competition control, the primary antibody was preincubated at 37°C with 2 μg/ml of the synthetic peptide used for immunization for 3–4 hr. No immunostaining was observed under these circumstances.
Light microscope immunohistochemistry. Fifteen adult (200–250 gm) Sprague Dawley rats (R. norvegicus) of either sex were anesthetized with 35% chloral hydrate (0.1 ml/100 gm) and perfused through the ascending aorta with 200 ml of a 4% paraformaldehyde (freshly depolymerized) in 0.1 m PBS, pH 7.2, at RT. Brains were removed and post-fixed in the same fixative overnight at 4°C. After rinsing in abundant PBS, 50- to 60-μm-thick vibratome sections were collected in PBS and processed as free-floating sections at 4°C. Endogenous peroxidase activity was inhibited by pretreatment of the free-floating sections with 3% H2O2 in PBS for 15 min at RT. Nonspecific sites were blocked for at least 1 hr in PBS containing 4% bovine serum albumin (BSA) and 0.25% Triton X-100 at RT, rinsed three times for 10 min in PBS, and incubated overnight with the primary antibody anti-nAChR α4-subunit (Santa Cruz Biotechnology) diluted 1:1000 (0.2 μg/ml) in blocking buffer, at 4°C. After three 10 min rinses in PBS, sections were incubated in biotinylated rabbit anti-goat antibody (Amersham) diluted 1:200 (6.5 μg/ml) in blocking buffer at RT for 1 hr, rinsed again, and incubated in the avidin–biotin–peroxidase complex (ABC kit; Vectastain; Vector Laboratories, Burlingame, CA), all of them at RT. Bound peroxidase was revealed using 0.05% diaminobenzidine tetrahydrochloride (DAB; Sigma, St Louis, MO) and 0.01% H2O2 in PBS, at RT. Sections were mounted on gelatinized slides, dehydrated in a graded series of ethanols, and coverslipped with Eukitt mounting medium.
Method specificity was controlled either by omitting the primary antibody or by its preadsorption with the immunogenic peptide. For the peptide competition control, the primary antibody was preincubated at 37°C with 2 μg/ml of the peptide for 3–4 hr. No immunostaining could be observed in either case.
Double immunostaining for confocal microscopy. Free-floating vibratome sections (50 μm) from six rats [bregma levels, −5.20 to −5.80 mm, according to Paxinos and Watson (1986)] were pretreated in blocking buffer (4% BSA and 0.25% Triton X-100 in PBS) at RT for 1 hr and then incubated with two different primary antibodies: the anti-α4-subunit nAChR (Santa Cruz Biotechnology) used at 1:1000 dilution (0.2 μg/ml) and a monoclonal antibody to TH (MAB318; Chemicon, Temecula, CA) at a 1:500 dilution (4 μg/ml). Two different protocols were used with identical results.
In the first one, both antibodies were diluted together in the blocking buffer, the incubation proceeded at 4°C overnight. After washing, all of the following steps were performed in the dark and at RT. Sections were incubated with a mixture of biotinylated rabbit anti-goat IgG (Jackson ImmunoResearch, West Grove, PA) diluted 1:200 (6.5 μg/ml) and Cy5-conjugated anti-mouse IgG (Amersham, Arlington Heights, IL) diluted 1:600 (1.6 μg/ml) in blocking buffer for 1 hr. Sections were then washed in PBS and incubated with Cy2-conjugated streptavidin (Amersham) diluted 1:600 (1.6 μg/ml) in blocking buffer for 1 hr.
In the second protocol, immunostaining for the two primary antibodies was developed sequentially. The sections were incubated first with the goat polyclonal antibody to α4-subunit followed by incubations in biotinylated anti-goat IgG (Jackson Immunoresearch) and then in Cy2-conjugated streptavidin (Amersham). After three 10 min rinses in PBS the sections were incubated in the monoclonal anti-TH antibody followed by Cy5-conjugated anti-mouse IgG (Amersham, Arlington Heights, IL). After rinsing in PBS, the sections were mounted, coverslipped with a 1:1 solution of PBS and glycerol, and examined in a Leica TCS NT confocal laser-scanning microscope equipped with an argon/krypton-mixed gas laser with excitation peaks at 488 nm (for Cy2) and 647 nm (for Cy5). Confocal image series were recorded through separate channels for the fluorescence of Cy2 and Cy5.
Immunofluorescently stained neurons were quantified on three rats (Table 1) by collecting images from a total of 26 distinct fields, each representing 329 × 329 × 35 μm of SNpc tissue.
Triplicate quantification of immunopositive (+) and unlabeled (−) neurons after double-labeling of the α4-subunit and TH by confocal microscopy (first column) in the rat SNpc
Electron microscope pre-embedding immunoperoxidase.Vibratome sections from nine rats were used for electron microscopy. They were cryoprotected in sucrose–glycerol–PBS mixtures before being submitted to several freeze–thaw cycles in liquid nitrogen. This membrane permeabilization treatment was preferred to exposure to Triton X-100 because it was less detrimental to ultrastructural morphology.Vibratome sections were then rinsed in large volumes of PBS, blocked and incubated overnight at RT in primary antiserum, and then processed as above for peroxidase. Addition of 0.1% glutaraldehyde did not modify the pattern of immunolocalization. Once the peroxidase activity was detected, sections were post-fixed in OsO4 dehydrated in a graded series of ethanol, propylene oxide, and flat embedded in epoxy resin (Durcupan ACM). Coronal sections of the mesencephalon were examined with the light microscope and drawn using a camera lucida. The region of SNpc corresponding to bregma −5.80 mm, ventral 8.4 mm, and 1.3 mm from the midsaggital plane (Paxinos and Watson, 1986) was one of the richest in α4-subunit-LI positive neurons. This medial part of the SNpc was selected because it is also densely populated with DA neurons and axonal branches containing ACh (Fallon and Loughlin, 1995) and choline acetyltransferase (ChAT; Henderson and Sherriff, 1991) immunoreactivity. It is also the zone in which Futami el al. (1995) found synaptic nicotinic transmission. Samples were cut out from the slides and glued to blank blocks of resin. Ultrathin sections were cut and examined, without any further contrasting, with a Philips CM10 electron microscope (EM) at 60 kV.
The method specificity was controlled either by omitting the primary antibody or by preabsorption with the antigenic peptide. No immunoperoxidase labeling was observed with either of these controls.
Electron microscope postembedding immunogold labeling. For immunogold labeling, several combinations of fixative were tested: 4% formaldehyde (freshly depolymerized from paraformaldehyde) applied according to a “pH shift protocol” (Nagelhus et al., 1998), 4% formaldehyde plus 0.1% glutaraldehyde, or 4% formaldehyde plus 0.5% glutaraldehyde. The best immunogold labeling was obtained with the first two of these fixatives. Blocks of substantia nigra from two “pH-shift” rats, and from one “0.1% glutaraldehyde” rat were dissected out from vibratome sections, 400 μm in thickness, at the same stereotaxic levels as indicated above.
Embedding of tissue in Lowicryl HM20 resin was performed as described in Matsubara et al. (1996). Ultrathin sections, 90 nm, were etched for 1–3 sec in sodium ethanolate. Remnant aldehyde residues were neutralized by a 10 min treatment in 0.1% sodium borohydride plus 50 mm glycine in Tris-buffered saline plus 0.1% Triton X-100 (TBST) for 10 min. Grids were then incubated for 10 min in a blocking solution of TBST and 2% human serum albumin to block the nonspecific binding sites.
Different concentrations of NaCl ranging from 0.3 to 0.9% were tested. The best immunolabeling was obtained with 0.5–0.6% NaCl in the vehicle buffer, and when the grids were incubated at RT successively in the blocking solution for 120 min, first antibody (goat polyclonal IgG anti-α4 from Santa Cruz Biotechnology) 1:1000 (0.2 μg/ml), followed by washes in TBST, and 60 min incubation in an immunogold reagent 1:40 (IgG rabbit anti-goat 10 nm gold-conjugate; Aurion) in the same blocking solution plus 0.5% polyethylene glycol (20,000 molecular weight).
For controls, as in the immunoperoxidase protocol, the method specificity was examined either by omitting the primary antibody or by preabsorption with the antigenic peptide. No immunogold labeling was observed when the first antibody was omitted. When using preadsorbed primary antibodies, some gold particles were present mainly in the somatodendritic cytoplasm, and very few were found on plasma membranes of extrasynaptic and postsynaptic domains of the neuronal cell surface.
After washes, the ultrathin sections were dried, stained with 1% uranyl acetate and 0.3% lead citrate, dried, and observed in a Philips CM10 EM at 80 kV. Magnification was calibrated with a cross grating replica.
RESULTS
Specificity of the immunohistochemistry
The anti-α4-subunit antibody used in this study is a peptide-purified polyclonal antibody that recognizes an intracellular epitope in the α4-subunit protein. The specificity of this polyclonal antibody was tested using several methods.
First, in Western blots, the anti-α4-subunit antibody reacted with a single band present in rat brain homogenates and absent in lung homogenates. This specific band had an apparent molecular weight of ∼70 kDa, which corresponds to the approximate range where immunopurified α4-subunit protein is found (Whiting et al., 1987). Preadsorption with the corresponding synthetic peptide used for immunization led to a dramatic diminution in the labeling intensity of this band (Fig. 1).
Analysis of the α4-subunit antibody specificity by immunoblot of total homogenates of rat brain and lung. Equivalent amounts of protein from each fraction (50 μg) were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the polyclonal antibody against the α4-subunit of the nAChR. A, After incubation with the anti-α4-subunit antibody, one specific band was positive in brain homogenates and was not present in lung homogenates.B, Incubation with the α4-subunit antibody pre-adsorbed with the corresponding synthetic peptide used for immunization has led to a dramatic diminution in the labeling intensity of this band in brain homogenates.
Second, SDS solubilized membranes from HEK-293 cells transfected with rat cDNAs corresponding to the α4β4 and α3β4 subunits, α7–5HT3 construct, the α4β2 human subunit cDNAs, and nontransfected cells were used in immunoblots studies. We detected one band in the expected molecular weight range, 70 kDa, in the membranes from cells that had been transfected with either α4β4 rat cDNAs or α4β2 human cDNAs, but this band was not present in either α3β4, α7–5HT3-transfected, or in nontransfected cell extracts (Fig. 2).
Immunoblot analysis of the α4-subunit antibody specificity in membranes extracted from HEK-293 cells transfected with different cDNAs of nAChRs. We detected one band of the expected molecular weight range only in the cells that were transfected with either α4β4 rat cDNAs (left) or α4β2 human cDNAs (right). This band was not present in either α3β4 rat cDNAs, in α7–5HT3-transfected, or in nontransfected cell extracts (NT).
When the immunofluorescence technique was used directly in transfected HEK-293 cells, an immunopositive reaction was present in association with membranes of cells that had been transfected with either α4β4 rat cDNA or α4β2 human cDNA (Fig. 3) but not in nontransfected cells.
Staining of transfected HEK-293 cells with the α4-subunit antibody. Immunofluorescence positive signal was present in HEK-293 cells transfected with either (A) α4β2 human cDNAs or (B) α4β4 rat cDNAs. Scale bar, 10 μm.
Additional evidence for specificity of the antibody used in the present study comes also from Western blot analysis using homogenates of brains of mice lacking the α4-subunit (Marubio et al., 1999), which shows the disappearance of the 70 kDa band corresponding to the α4-subunit.
Having demonstrated its specificity by Western blot analysis, we then used the anti-α4 antibody in an immunohistochemical study of the expression of α4-subunit protein in the adult rat brain. Method specificity was studied by omitting the primary antibody as well as by incubation in the presence of an excess of the synthetic peptide. No cellular labeling was apparent under either of these conditions (data not shown).
Altogether these data strongly support the specificity of this antibody in tissue sections. The α4-subunit-LI pattern thus suggests the localization of the α4-subunit.
Immunohistochemical localization in rat brain
Incubation of tissue sections with the anti-α4 antiserum generated reproducible patterns of staining within discrete populations of neurons and neuronal processes. The sensitivity of the immunohistochemical procedure was sufficient for strong labeling of cells of the SNpc complex (Fig.4A) and reliable staining in other brain areas shown by ISH to express α4 mRNA, such as the cerebral cortex (Fig.5A), the globus pallidus (Fig.5B), and some nuclei of the thalamus. In general, the pattern of immunolabeling was in good agreement with the pattern detected earlier by ISH (Wada et al., 1989).
Distribution of nAChR α4-subunit-LI neurons in the adult rat substantia nigra. A, Coronal sections containing rat SNpc and SNpr were stained with the α4-subunit antibody detected using immunoperoxidase method. Scale bar, 200 μm.B, The α4-subunit-LI was present in neuronal perikarya and dendrites, but absent from cell nuclei as shown in this detail of SNpc-immunopositive cells. Scale bar, 50 μm.
Distribution of α4-subunit-LI in cerebral cortex, globus pallidus (GP), and striatum (ST) after immunoperoxidase labeling.A, Coronal sections of rat somatosensory cortex show a laminar distribution of α4-subunit-LI, mainly in neurons of the infragranular layers. Cell bodies and apical dendrites of pyramidal neurons show a prominent labeling. Scale bar, 100 μm.B, Cell bodies of globus pallidus were clearly stained with α4-subunit antibody but few neurons were positive in the striatum. Scale bar, 100 μm.
The immunohistochemical reaction was restricted to neurons and was present in cell perikarya and dendrites but absent from cell nuclei (Fig. 4B). In the mesencephalon, immunoreactivity for the α4-subunit was more intense in nuclei containing dopaminergic cells than in other groups of neurons. α4-subunit-LI neurons were present mainly in the SNpc and in the VTA, but some of them were also detected in the predominantly nondopaminergic reticular portion of the substantia nigra (SNpr) (Fig. 4A). No evidence for staining of axon terminals in the caudate putamen and the nucleus accumbens, the terminal projection area of the dopaminergic neurons in the midbrain, was observed (Fig. 5B). A few somata in the caudate putamen, however, were found immunoreactive for the α4-subunit.
Colocalization studies
Given the high level of expression of α4-subunit mRNA and -LI in the dopaminergic cell groups and the importance of these nuclei in the central actions of nicotine, we directly tested whether the α4-subunit expressing neurons in the midbrain were dopaminergic. The transmitter identity of the α4-subunit-LI positive neurons was examined by double-labeling sections of the rat midbrain for both the α4-subunit-LI and TH, the rate-limiting enzyme responsible for the synthesis of DA. A quantitative estimation of double- or single-labeled neurons in SNpc revealed that 92% of neurons were double-labeled for the α4-subunit-LI and TH (Table 1). Thus, the majority of cells immunoreactive for TH were found to be labeled by the antibody against α4 (Fig. 6); conversely, only 5.8% of neurons immunoreactive for the α4-subunit did not coexpress TH. The identity of the neurotransmitter of these cells remains to be determined. Moreover, the intensity of its α4-subunit immunostaining was weak as compared with that of double-stained neurons. Thus, a strong correlation exists between the dopaminergic status of the neuron and the expression of the α4-subunit-LI. These results are in agreement with recent colocalization studies done in SNpc using a different an-tibody against nAChRs α4-subunit (Sorenson et al., 1998). Double labeling of cells with antibodies against α4-subunit and glial fibrillary acidic protein showed no localization of the α4-subunit-LI in glial cells (data not shown).
Confocal microscopy shows colocalization of α4-subunit-LI and TH in dopaminergic neurons in the SNpc (A–C) and VTA (D–F). A,D, Cy2-conjugated streptavidin decorates neurons immunolabeled with the antibody raised against the α4-subunit.B, E, Cy5-conjugated anti-mouse IgG-stained neurons containing TH. Superimposition of the images (C, F) show that most of the neurons were double-labeled. Scale bar, 20 μm.
Ultrastructural localization of α4-subunit-LI in the SNpc
The subcellular localization of the α4-subunit-LI was initially examined by a pre-embedding immunoperoxidase method. Electron microscopy observations confirmed that in the SNpc, the α4-subunit-LI reaction was found associated with neurons. For illustration (Figs.7, 8), we selected neurons in which the reaction product was not so massive that cytoplasmic compartments were obscured. In these cells, the α4-subunit-LI reaction product was localized in perikarya and dendrites. It appeared as clumps of dense material associated with the rough endoplasmic reticulum (RER) and the cytoplasmic matrix (Fig.7A). No immunostaining was observed with preadsorbed antibody or in the absence of the first antibody in the pre-embedding immunoperoxidase procedure (Fig. 8D), as well as in the postembedding immunogold procedure (Fig.9H). As in previous work with the β2-subunit (Hill et al., 1993), we never observed α4-subunit-LI reaction in the Golgi apparatus (Fig. 7B). In the perikarya, the α4-subunit-LI reaction was distributed in a more patchy manner (Fig. 7A,B) than the β2-subunit-LI reaction product (Hill et al., 1993). The α4-subunit-LI DAB reaction product was found associated with microtubules in dendrites (Fig.8B,C). Dense accumulations of α4-subunit-LI reaction product were also observed associated with a few asymmetrical synaptic profiles located on both perikarya (Fig.8A) and dendritic shafts (Fig. 8B), but never on dendritic spines. These observations suggest that α4-subunit-LI positive synaptic profiles are preferentially located on the perikarya and proximal dendrites of the neurons in the SNpc. Large deposits of DAB immunoreaction product could not be detected (Fig. 8A–C) in the perisynaptic, i.e., around the edge of postsynaptic density (unlike as in Luján et al., 1996, their Fig. 6) domains of the neuronal cell surface. The frequency of the DAB-decorated postsynaptic densities was low.
Ultrastructural localization of the α4-subunit of the nAChR in the rat SNpc, using immunoperoxidase labeling.A, Dense deposits of DAB are observed in spots of the somatodendritic compartments (asterisks).B, Immunoreactivity for the α4-subunit is distributed heterogeneously. In the perikarya, immunolabeling is frequently associated with endoplasmic reticulum (asterisk), but not with the Golgi apparatus (Go). Scale bar, 0.4 μm.
Postsynaptic localization of the α4-subunit-LI in the rat SNpc (A–C). Dense deposits of immunoreaction product are associated with postsynaptic densities (arrows) and not with perisynaptic plasma membrane.A, Densely immunolabeled axosomatic synaptic contact (arrow) is close to a lightly labeled endoplasmic reticulum (asterisks). B, Immunolabeled synaptic profile (arrow) on a dendritic shaft. Dendroplasm is also lightly immunolabeled. C, In the same field as in B and C unlabeled synaptic contacts (arrowhead) were present beside labeled synaptic profiles (arrow). D, No labeling was apparent in controls (arrow) in which the first antibody was either omitted or preadsorbed with the peptide used as antigen. Scale bar, 0.4 μm.
Postsynaptic localization of the α4-subunit-LI in the rat substantia nigra (A–G). Consistent with our observations using immunoperoxidase, the immunogold labeling is observed at some synaptic profiles (arrowheads). Most of the postsynaptic densities were decorated with two to four gold particles. To estimate the nonspecific labeling, the primary antibody was either omitted or preadsorbed with the antigen peptide before incubations. In that case, most of the synaptic profiles (arrows) were unlabeled (H). Scale bar, 0.25 μm.
Because DAB precipitates may be deposited at some distance from the epitopes, we attempted to confirm a possible synaptic localization of the α4-subunit by postembedding immunogold experiments, which provides a better resolution. As shown in Figure 9A–G, gold particles were observed associated with some postsynaptic densities. The scarcity of the observations of gold-decorated postsynaptic densities paralleled that of DAB-positive synaptic profiles. Although the resolution of the postembedding immunogold labeling of single sections is high, it will necessarily give a lower proportion of immunopositive synapses than pre-embedding peroxidase labeling because only the epitopes exposed at the surface of the ultrathin sections will be accessible to the immunoreagents (Griffiths, 1993). Only a small proportion of the receptor-bearing synapses will be recognized as such. Accordingly, we found fewer immunolabeled synapses with the immunogold labeling. From an estimated 60,000 synapses observed directly on the EM screen, we found 102 synaptic profiles decorated with at least one gold particle (Fig. 10A). The distribution of gold particles per profile did not follow a Poisson distribution (Fig. 10B) for synaptic profiles with two, and more, gold particles per profile, suggesting that a threshold of two gold particles per profile should be used to consider it as labeled. Thus, only 63 of the synaptic profiles could be considered as labeled. Of these, only 44 were decorated with gold particles located within 30 nm distance from the postsynaptic plasma membrane (Fig.11), which is the criterion defined byMatsubara et al. (1996). Finally, only 31 of the narrowed group of synaptic profiles showed distinct appositions of presynaptic and postsynaptic plasma membranes, a synaptic cleft 10–20 nm wide, and the presence of synaptic vesicles in the nerve terminals (Peters et al., 1991). This selected population contained 114 gold particles with a mean of 3.6 ± 3.1 (mean ± SD; n = 31) gold particles per synaptic profile. The mean of the density of gold particles calculated for each of these 31 synaptic profiles is 10.7 ± 11.6 (n = 31) gold particles per micrometer of decorated postsynaptic plasma membrane (total measured length of 14.2 μm). The estimated mean length of postsynaptic plasma membranes was 0.49 ± 0.27 μm (n = 114), giving 38.8 × 10−4 gold particles per micrometer of total postsynaptic plasma membrane (total length of 60,000 × 0.49 μm). The background was estimated after immunolabeling with antibody preadsorbed with the immunogenic peptide. We scanned the same number of synaptic profiles as in the experimental sample (n = 60,000). We found seven profiles decorated with at least one gold particle (Fig. 10A). After the same criteria used for the experimental sample, only one profile decorated with two gold particles was observed, giving 0.68 × 10−4gold particles per micrometer of total postsynaptic plasma membrane (total length of 60,000 × 0.49 μm).
Histogram distribution of synaptic profiles decorated with one or more gold particles per profile (before applying the morphological and gold particle distribution criteria), in both experimental (dark gray) and control (light gray) conditions (A). In experimental conditions (B) this distribution (open diamonds) rapidly diverges from a Poisson distribution (black diamonds) for two and more gold particles per synaptic profile.
Histogram showing the radial distribution of gold particles over synaptic profiles sampled in the SNpc. The distances between the center of each gold particles and the postsynaptic plasma membrane were grouped into bins 10 nm wide. The values along thex-axis indicate bin centers, with bin “0” centered on the postsynaptic plasmic membrane. Minus signsindicate direction of the presynaptic domain. These data were pooled from 44 synaptic profiles.
On the extrasynaptic plasma membrane of the dendrites we found a density of 0.10 ± 0.15 (n = 27) gold particles per micrometer (total length sampled, 316 μm), and the background was 0.01 ± 0.04 (n = 31) gold particles per micrometer (total length sampled, 185 μm). Evidence thus exists that the α4-subunit is localized at authentic, albeit very scarce, postsynaptic membranes in DA neurons.
DISCUSSION
Together with the β2-subunit, the α4-subunit of the nAChR is one of the most widely expressed nAChR subunits in the CNS. Using an antibody specific for the α4-subunit we have shown a widespread distribution of this subunit in the rat brain. The expression of the α4-subunit was particularly intense in DA neurons of the SNpc and VTA where α4-subunit-LI colocalizes with TH. Furthermore, α4-subunit-LI was found in association with some postsynaptic densities in the DA neurons.
Specificity of α4-subunit immunostaining
The α4-subunit immunolabeling was found to be highly specific both in tissue homogenates and sections of rat brain. Such specificity was further confirmed using mice lacking the α4-subunit (Marubio et al., 1999). Western blot analysis using the anti-α4-subunit antibody showed the disappearance of a band corresponding to 70 kDa in brain extracts from these mice lacking the α4-subunit, although it is still present in control mice (Marubio et al., 1999). There was a good correspondence between areas where neurons exhibited a dense cytoplasmic α4-subunit-LI staining and regions where a positive ISH signal was observed using probes for the α4-subunit mRNA (SNpc–VTA, thalamus, and cortex) (Wada et al., 1989).
The regional distribution of α4-subunit-LI differs unambiguously from that of other nAChR subunits. For instance, the α2-subunit that has the highest peptide sequence similarity to the α4-subunit of the known nAChR is present in high mRNA levels in the interpeduncular nucleus (Wada et al., 1989). However, α4-subunit mRNA levels are lower, and α4-subunit-LI cannot be detected in this area.
The β2-subunit of the nAChR is present in most neurons of the rat brain, as shown by both ISH (Wada et al., 1989) and immunohistochemistry techniques (Swanson et al., 1987; Hill et al., 1993). The α4-subunit-LI pattern differs from that of β2-subunit-LI especially in the striatum and hippocampus, where only a few positive α4-subunit-LI cells were detected. Moreover, the α4-subunit-LI continues to be expressed in unaltered levels (data not shown) in mice lacking the β2-subunit of the nAChR (Picciotto et al., 1995) indicating that this antibody does not cross-react with the β2-subunit.
The pattern of the α4-subunit-LI in general parallels the pattern of high-affinity [3H]nicotine binding sites in the rat brain (Clarke et al., 1985b). There is moderate to strong labeling in the medial habenula, thalamus, SNpc, and VTA, and weaker labeling in the hippocampus. Furthermore, the pattern of the α4-subunit-LI does not match the pattern of125I-α-bungarotoxin (α-BTX) binding sites, which reflects the distribution of the α7-subunit (Seguela et al., 1993;Orr-Urtreger et al., 1997). The distribution of125I-α-BTX labeling is strong in the hippocampus and weak in the superficial cortical layers, whereas α4-subunit-LI is strong in the cortex and weak in hippocampus. At the level of the hypothalamus α4-subunit-LI is present but the densities of high-affinity [3H]nicotine binding sites or α4 mRNA are rather low (Clarke et al., 1985b; Wada et al., 1989). In the striatum, we found some neurons with α4-subunit-LI, although mRNA have not been reported in this structure (Wada et al., 1989). Differences in the distribution of nAChRs and their mRNAs at the cellular level may result either from higher sensitivity of immunocytochemical techniques as compared with ISH, or from higher translational efficiencies.
The α4-subunit in dopaminergic neurons of the SNpc and VTA
The composition of the nAChRs that mediate the effect of nicotine on DA neurons is only partially known. Recent results have shown that the response of SNpc and VTA neurons to nicotine is absent in mice lacking the β2-containing receptors (Picciotto et al., 1998). This suggests that the β2-subunit is an integral part of the nAChRs that mediates nicotine-elicited responses on DA neurons.
It remains to be determined which other nicotinic subunits assemble with the β2-subunit in DA neurons. Moderate to high levels of the α3, α4, α5, α6, β2, β3 nAChR-subunit mRNAs are present in these neurons (Wada et al., 1989, 1990; Deneris et al., 1989;Dineley-Miller and Patrick, 1992; Le Novère et al., 1996). The α6-subunit immunoreactivity was found to be present in DA neurons of the midbrain (Goldner et al., 1997). In addition, an α7-containing nAChR has been demonstrated with electrophysiological techniques (Pidoplichko et al., 1997).
Our present results, along with those from Sorenson et al. (1998), suggest that the α4-subunit protein is highly concentrated in DA cell bodies and proximal dendrites. This is in line with previous work showing that a nAChR with a putative composition of α4β2 is responsible for nicotinic responses in mesencephalic DA cell bodies and disappears in β2 knock-out mice (Picciotto et al., 1998).
Indeed, different kinds of experiments have shown that high-affinity binding sites, which are enriched in the mesencephalic DA cell bodies, are most likely composed of α4β2-containing nAChRs (Whiting and Lindstrom, 1987; Flores et al., 1992; Happe et al., 1994; Zoli et al., 1998). It cannot be excluded that nicotinic transmission is mediated via α4/nonβ2 receptors in other brain regions (Zoli et al., 1998).
Taken together, the present and previous results suggest, but do not yet demonstrate, that the α4-subunit might be part of the nAChRs that contribute to the nicotinic activation of DA neurons in the mesotelencephalic pathways.
The cholinergic circuits in the mesencephalic dopaminergic nuclei
Anatomical observations show that the mesencephalic DA nuclei receive a cholinergic innervation from the pedunculopontine tegmental nuclei (PPN) and the laterodorsal tegmental nucleus (Woolf and Butcher, 1986; Beninato and Spencer, 1987, 1988; Gould et al., 1989;Tokuno et al., 1988, but see Lee et al., 1988). High densities of muscarinic (Nastuk and Graybiel, 1991) and nicotinic receptors have been observed in the SNpc (see above), with a pattern resembling that of acetylcholinesterase (AChE) (Paxinos et al., 1980).
There is also functional evidence for cholinergic innervation (Clarke et al., 1987). For example, quinolinic acid lesions of the PPN attenuate the stimulatory effects of intranigral neostigmine, an AChE inhibitor, on DA efflux in the striatum, and intranigral nicotine enhances striatal DA efflux (Blaha and Winn, 1993; Blaha et al., 1996). Electrophysiological studies with slice preparations reveal nicotinic postsynaptic responses on DA neurons in the SNpc after stimulation of the PPN (Futami et al., 1995, but see Scarnati et al., 1986). These observations suggested that nicotinic receptors are involved in the pontonigral cholinergic transmission.
Ultrastructural studies demonstrated that ChAT-positive terminals form asymmetrical synapses with the cell bodies and dendrites of SNpc DA cells (Beninato and Spencer, 1987, 1988; Henderson and Greenfield, 1987; Martinez-Murillo et al., 1989a,b; Bolam et al., 1991; Henderson and Sherriff, 1991).
To date, few ultrastructural studies of nAChR localization have been performed in rat brain. Using two different monoclonal antibodies, mAb WF6 (raised against purified Torpedo marmorata electric organ) and mAb 299 (against chicken α4-subunit), α4-subunit-LI was found to be associated with some postsynaptic membranes in the cerebral cortex (Whiting and Lindstrom, 1988; Schroder et al., 1989; Okuda et al., 1993; Nakayama et al., 1995) and in dendrites of neurons from the SNpc of the rat (Sorenson et al., 1998). However, mAb WF6 competes for the α-BTX-binding site on Torpedo nAChRs and may recognize an α-BTX nAChR in brain.
We found α4-subunit-LI in perikarya and dendritic shafts of neurons in the substantia nigra and in association with some postsynaptic membranes. We also observed a strong staining of neuronal RER but failed to detect it over Golgi membranes. It is proposed that the α4-subunit-containing receptors accumulate in the RER in the course of biosynthesis and assemble, before being targeted to the neuronal cell surface.
Nonsynaptic transmission in this system cannot be excluded (Descarries et al., 1997). The α4-subunit might be localized extrasynaptically as was the case for other nAChRs subunits. For instance, in chick ciliary ganglion, immunohistochemical staining at the electron microscopy level indicated that α-BTX-sensitive receptors are localized in regions adjacent to postsynaptic membrane thickenings and are concentrated in the vicinity of short dendrites emanating from the cell bodies (Jacob and Berg, 1983). nAChRs that are recognized by mAb 35 (Tzartos et al., 1981) but not α-BTX can be found, in part, in the specialized postsynaptic membrane (Jacob et al., 1984; Loring and Zigmond, 1987). Similar conclusions were reached with confocal immunofluorescence that revealed clusters of α-BTX-binding nAChRs in perisynaptic locations (Horch and Sargent, 1995).
The synaptic profiles decorated with α4-subunit LI are quite rare, but their topological distribution, restricted to the perikarya and proximal dendrites of the DA neurons in the SNpc, might have a significant functional role that remains to be fully elucidated.
We suggest that the α4-subunit is part of functional nAChRs present on DA neurons. It seems likely to be involved in mediating the increase of dopamine release caused by nicotine. This mechanism may be relevant to brain reward circuits that plausibly could contribute to cognitive learning (Dehaene and Changeux, 1989, 1997; Levin et al., 1992, 1995) and nicotine addiction (Koob, 1996; Pontieri et al., 1996).
Footnotes
This research was supported by grants from The Collège de France, the Association Française contre la Myopathie, the Council for Tobacco Research, Reynolds Pharmaceutics, the European Union (Biomed BMH1-CT94–1060 and Biotech 960236), the French Embassy in Spain, the Ministerio de Educacion y Cultura (PB94–0219-CO2-01), the Comunidad de Madrid (AE00268/95), and the National Alliance for Research on Schizophrenia and Depression. We are grateful to Dr. Y. Pass for his kind and qualified help in several experiments. We are indebted to Dr. P. J. Corringer, S. Bohler for scientific discussions and materials provided, and Drs. R. Miles, M. Zoli, C. Léna, and N. Le Novère for critical reading of this manuscript. We thank P. Gounon, director of the Station Centrale de Microscopie Electronique at Institut Pasteur, and K. M. Gujord at Oslo University for their constant and expert advice during all this work. We would also to thank R. Hellio, B. Martı́n-Clemente, and Dr. J. R Martı́nez-Galán for their help in confocal microscopy.
Correspondence should be addressed to Jean-Pierre Changeux, Neurobiologie Moléculaire, Département des Biotechnologies, Institut Pasteur, 25, rue du Dr. Roux, 75724 Paris Cedex 15, France.
Dr. M. del Mar Arroyo-Jiménez’s present address: Facultad de Medicina, Universidad de Castilla–La Mancha, Edificio Benjamín Palencia, Campus Universitario, 02071 Albacete, Spain.
