Precise Localization of α7 Nicotinic Acetylcholine Receptors on Glutamatergic Axon Terminals in the Rat Ventral Tegmental Area

Confocal microscopy

Triple-labeling confocal microscopy was used to investigate the cellular distribution of α7 nAChRs within the VTA with respect to dopaminergic neurons, GFAP-immunopositive astrocytes, and glutamatergic and cholinergic axon terminals. At a low magnification (Fig. 1A,B), α7 nAChRs (identified by fluorescent αBgt binding sites) were not discernable. However, both VGluT1 (Fig. 1A) and VGluT2 (Fig. 1B) immunoreactivities were observed in the VTA, with differential labeling of the neighboring interpeduncular nucleus. At a higher magnification, αBgt binding sites were visualized in the VTA (Fig. 1C), and subpopulations were associated with VGluT-immunopositive structures likely to be axon terminals (yellow arrows), tyrosine hydroxylase-positive dopaminergic neurons (blue arrows), and nondopaminergic structures (green arrows). Notably, not all of the glutamatergic axonal boutons expressed αBgt binding sites (red arrows). Preincubation of the sections with an excess of unconjugated αBgt (Fig. 1D) or α-cobratoxin (data not shown) eliminated fluorescent αBgt labeling, demonstrating that the fluorescent αBgt binds specifically to α7 nAChRs. Sections labeled for both VGluT1 and VGluT2 isoforms, using different fluorochromes (Fig. 1E), showed complementary labeling of structures interpreted to be axon terminals, with no evidence of colocalization. This indicates that VGluT1 and VGluT2 are likely to be localized to separate axon terminals in the VTA.

The colocalization of α7 nAChRs with each VGluT isoform was demonstrated (Fig. 2A-J) and confirmed using fluorescence profiling (Fig. 2E,J). This analytical tool (see Materials and Methods) permits individual fluorescence intensities to be plotted along a user-defined line. In Figure 2, D and I, the line of analysis is represented by white arrows, and Figure 2, E and J, illustrates the respective fluorescence intensities along these paths in the direction of the arrows. In both of the representative examples shown, the αBgt signals (green peaks) overlap with the VGluT1 (Fig. 2D) and VGluT2 (Fig. 2I) signals (red peaks), indicating colocalization of α7 nAChRs with glutamatergic axon terminals. Profile analysis of 100 VGluT1- and 100 VGluT2-labeled axonal boutons in the VTA revealed that 25% of VGluT1- and 31% of VGluT2-immunoreactive boutons expressed detectable αBgt binding sites. In comparison, similar analyses within the shell of the NAcc from the same rats found that 30% of VGluT1- and 72% of VGluT2-immunoreactive boutons expressed detectable αBgt binding sites.

To address the identity of VGluT-immunoreactive structures in the VTA, we determined whether VGluT is expressed by glial cells, as well as axonal boutons. Aldehyde-fixed sections were triple-labeled with α7 nAChRs, VGluT1 and VGluT2, and GFAP for confocal analysis (Fig. 3A). VGluT immunoreactivity was detected within a subpopulation of astrocytic processes within the VTA, and this association was confirmed by fluorescence profile analysis (Fig. 3B). Triple labeling, however, failed to demonstrate any expression of α7 nAChRs within VTA astrocytes; in all of the cases examined, αBgt binding sites were restricted to non-GFAP-immunopositive structures, including glutamatergic axon terminals (Fig. 3C).

Similar experiments comparing the localization of α7 nAChRs and cholinergic axon terminals (identified by VAChT immunoreactivity) in the VTA failed to find any incidences of colocalization in coronal VTA sections from five adult rats (Fig. 4A). Serial optical sections through the VTA (z-stacks) were analyzed to check for fluorescent αBgt binding sites around the entire perimeter of the cholinergic axon terminals, again failing to find any instances of colocalization (data not shown). The absence of detectable αBgt binding sites associated with cholinergic axon terminals in the VTA is in contrast to immunolabeled sections of the NAcc shell from the same rats, in which a proportion of αBgt binding sites colocalized with cholinergic axon terminals (Fig. 4B, yellow arrows).

Electron microscopy

Both unfixed and aldehyde-fixed sections through the VTA were used to examine the ultrastructural localization of α7 nAChRs in the VTA using αBgt conjugates. nAChRs do not tolerate conventional glutaraldehyde fixation that provides optimum preservation of tissue morphology at the electron microscope level (Jones et al., 2002). Indeed, αBgt binding and anti-nAChR subunit antibody labeling are significantly reduced even at low concentrations (0.1%) of glutaraldehyde (Jones et al., 2002), so other strategies were necessary to circumvent this problem. Unfixed sections (Fig. 5) are advantageous, because the receptors are in their native state for labeling and can be postfixed by immersion fixation in glutaraldehyde. However, unfixed sections are incompatible with permeabilization steps, so αBgt penetration is poor and limited to within a few micrometers of the section surface (our unpublished observations). In addition, because cell membrane integrity is maintained, αBgt can bind only to receptors expressed on the cell surface. In this study, optimal tissue preservation was observed within a few micrometers of the section surface, coinciding with the maximal αBgt penetration and binding. In these superficial areas, tissue morphology was poorer than that seen with in vivo fixative perfusion, reflecting a combination of postmortem degeneration in ACSF and the slow rate of slice penetration by glutaraldehyde during postlabeling fixation.

Tissue morphology was improved in fixed tissue from brains perfused with PLP fixative in situ before sectioning (Fig. 6). This fixative was chosen because it has been shown previously to be a suitable alternative to glutaraldehyde-based fixatives for immunoelectron microscopy studies (McLean and Nakane, 1974). Although morphology is inferior to that seen with glutaraldehyde perfusion in vivo, retention of antigenicity is greatly improved. In contrast to unfixed sections, fixed sections are also compatible with detergent permeabilization, revealing both surface and cytoplasmic α7 nAChR populations.

In both unfixed and fixed samples, tissue morphology preservation was sufficient to identify critical cellular structures such as axonal boutons and synapses. At the electron microscope level, assembled α7 nAChRs were precisely localized using two αBgt conjugates, biotinylated and gold-tagged αBgt. Biotinylated αBgt has been used previously to localize α7 nAChRs in the stratum radiatum of the CA1 region of the hippocampus (Fabian-Fine et al., 2001). This conjugate has been shown to be specific for α7 nAChRs, confirmed in this study by competition with excess αBgt or α-cobratoxin, which obliterated labeling (data not shown) in a manner similar to that of the confocal control experiments (Fig. 1D). However, the requirement for a secondary labeling step to detect biotinylated αBgt may be affected adversely by endogenous tissue biotin. For this reason, we have characterized recently a 1.4 nm gold conjugate of αBgt for directly visualizing α7 nAChRs at the light and electron microscope levels (Jones at al., 2004). We showed previously that this conjugate is specific for α7 nAChRs: it displaced [³H]methyllycaconitine binding to rat brain membranes with affinity comparable with that of native αBgt, and PLP-fixed tissue labeling was abolished by preincubation with excess αBgt or α-cobratoxin (Jones et al., 2004).

In unfixed tissue slices through the VTA, α7 nAChRs were localized both on the surface of axon terminals (Fig. 5) and on dendrites themselves (data not shown). No labeling of glial cells, based on morphological criteria, was noted within the VTA. In some instances, the silver enhancement of the gold particle spanned the synaptic cleft, so it is not possible to determine whether the α7 nAChRs are presynaptic or postsynaptic (Fig. 5D). There was no apparent difference in labeling with either biotinylated (Fig. 5A,B) or gold (Fig. 5C,D) αBgt. Notably, the use of gold-conjugated anti-biotin, in postembedding labeling of epoxy resin-embedded sections to detect biotinylated αBgt, represents a novel technique for identifying receptors; it was mainly used previously to detect amino acids (Ingham, 1992). Interestingly, α7 nAChRs on the surface of axon terminals were rarely seen at active zones of neurotransmitter release and were instead found within the synaptic cleft at the periphery of active zones (perisynaptic loci) (Fig. 5D).

Similar results were obtained with PLP-fixed sections, incubated with either biotinylated or gold αBgt (Fig. 6). In these experiments, however, the sections were permeabilized, permitting the visualization of intracellular αBgt binding sites. α7 nAChRs were observed in the cytoplasm of a subpopulation of VTA neurons (Fig. 6A,B), often associated with the endoplasmic reticulum (Fig. 6B). α7 nAChRs were also observed in presynaptic boutons (Fig. 6C,D). Semiquantitative analysis of gold αBgt labeling in 100 random fields revealed that 57% of α7 nAChRs were somatodendritic, whereas 43% were on axonal boutons within the VTA (n = 1647 gold particles). Most α7 nAChRs were cytoplasmic (82% of somatodendritic and 65% of axonal receptors), and the remainder were associated with the plasma membrane (defined as either touching or being within one gold particle diameter distance from the membrane).

In agreement with the unfixed tissue experiments, most gold particles on the surface of axon terminals in contact with dendrites were found at nonsynaptic loci (Fig. 6D). Of 140 membrane-associated gold particles, 61% were within the synaptic cleft adjacent to the active zone (perisynaptic), 27% were outside of the synaptic cleft (extrasynaptic), and the remainder (12%) were within the synaptic cleft at the active zone (synaptic).

Many of the axon terminals labeled with biotinylated or gold-conjugated αBgt appear to form asymmetric synapses with dendrites, characteristic of glutamatergic synapses, although the morphological preservation makes this distinction less clear (Figs. 5C,D, 6C,D). To determine definitively whether α7 nAChRs are present on glutamatergic axon terminals, as inferred by the confocal results (Fig. 2) and single-labeling studies (Figs. 5B,D,6D), double-labeling experiments were undertaken at the electron microscope level (Fig. 7). α7 nAChRs were identified using either biotinylated or gold-conjugated αBgt, whereas glutamatergic axon terminals were labeled using anti-VGluT1 or -VGluT2 antibodies and revealed with the DAB reaction product, which forms a dark precipitate visible in electron micrographs. Because gold conjugates of αBgt or anti-biotin IgG do not penetrate sections as efficiently as unconjugated antibodies, analysis was restricted to zones containing gold particles, typically within 10 μm of the section surface. α7 nAChRs were identified in both VGluT1-immunopositive (Fig. 7A,C) and VGluT2-immunopositive (Fig. 7B,D) axon terminals in the VTA. In addition to a cytoplasmic location, gold particles were occasionally observed to be associated with the membrane at perisynaptic sites (Fig. 7A). Only a subpopulation of glutamatergic axon terminals was found to contain α7 nAChRs (9% of VGluT1 and 11% of VGluT2 terminals; n = 100 immunolabeled terminals each). Additionally, when sections were incubated with antibodies to both VGluT1 and VGluT2, a proportion of αBgt binding sites was observed in nonglutamatergic axon terminals (75% associated with either VGluT1 or VGluT2 and 25% associated with unlabeled terminals; n = 100 αBgt-labeled axon terminals).

To investigate whether a subpopulation of dopaminergic neurons within the VTA express α7 nAChRs, sections were colabeled with antibodies against tyrosine hydroxylase. Approximately one-half (45%; n = 30 neurons) of tyrosine hydroxylase-positive neurons contained gold-conjugated αBgt particles in the VTA. Labeling of α7 nAChRs was observed both in the cytoplasm (Fig. 8A,B) and on the surface membrane at synaptic sites (Fig. 8C) of dopaminergic neurons.

Similar experiments to determine whether α7 nAChRs are present on, or apposed to, cholinergic axonal boutons failed to find any association (n = 43 VAChT-immunopositive axonal boutons from 5 rats) (Fig. 8D), in agreement with the confocal microscopy results (Fig. 4). As in the latter experiments, serial sections in the z-axis were analyzed to check for αBgt binding sites around the whole perimeter of the cholinergic axonal boutons. To ensure that this lack of association between α7 nAChRs and cholinergic axonal boutons was not attributable to variability in tissue penetration between gold conjugates and antibodies (see above), the presence of gold particles was confirmed at the same level in the section (Fig. 8D, arrow).

α7 nAChRs are present on glutamatergic terminals

At the EM level, α7 nAChRs, identified by either biotin- or gold-conjugated αBgt, were found on axon terminals forming synapses onto dendrites (Fig. 5). Morphologically, these synapses often appeared to be asymmetric, consistent with multiple labeling strategies that identified α7 nAChRs on a subpopulation of VGluT-immunoreactive (glutamatergic) terminals. At the cellular level, two populations of glutamatergic synapses in the VTA could be distinguished by their immunoreactivity for either VGluT1 or VGluT2, with VGluT2 predominating (Fig. 1E). The lack of colocalization of VGluT1 and VGluT2 immunoreactivities (Fig. 1E) indicates that these markers define two populations of glutamatergic synapses. α7 nAChRs were associated with a proportion of both VGluT1- and VGluT2-positive structures. At the confocal microscope level, αBgt binding was associated with 25% of VGluT1- and 31% of VGluT2-immunoreactive boutons. The association was lower at the electron microscope level (9 and 11%, respectively), which may reflect the increased sensitivity of confocal microscopy compared with immunoelectron microscopy. Complementary, nonoverlapping expression of VGluT1 and VGluT2 has been generally observed (Fremeau et al., 2001, 2004; Herzog et al., 2001; Kaneko and Fujiyama, 2002). The significance is unclear, although it has been suggested that VGluT2 is expressed at synapses with a higher probability of glutamate release (Fremeau et al., 2001). VGluT1 mRNA expression predominates in cortical areas, whereas VGluT2 is the predominant transcript in brainstem (Fremeau et al., 2001, 2004; Herzog et al., 2001; Kaneko and Fujiyama, 2002). Tracing studies will be necessary to discover whether the two populations of glutamate terminals recognized in the present study reflect different pathways projecting to the VTA. Although α7 nAChRs were identified in a similar proportion of both populations of glutamatergic terminals, the greater abundance of VGluT2-positive boutons in the VTA means that, numerically, many more α7 nAChRs are associated with these terminals.

Of the presynaptic α7 nAChRs in the VTA, 35% were associated with the plasma membrane, and the remainder were cytoplasmic. Thus, a higher proportion is membrane-associated than observed previously for the distribution of the nAChR β2 subunit immunoreactivity in the dorsal striatum, in which 15% of β2 subunit immunolabeling was membrane-bound in nigrostriatal dopaminergic axon terminals compared with only 3% for the cytoplasmic enzyme tyrosine hydroxylase (Jones et al., 2001). Interestingly, 61% of presynaptic α7 nAChRs were situated within the synaptic cleft at perisynaptic loci, 27% were extrasynaptic (outside of the synaptic cleft), and only 11% were seen at the active zone, opposed to the postsynaptic density. It is unlikely that this is a result of partial accessibility of the synaptic cleft to the αBgt conjugates because both biotinylated αBgt and gold αBgt have been shown previously to bind to synaptic sites in the hippocampus (Fabian-Fine et al., 2001; Jones et al., 2004). The perisynaptic location of α7 nAChRs at the edges of glutamatergic synapses in the VTA places them in an ideal location for modulating glutamatergic transmission via local depolarization and/or increases in intracellular calcium through α7 nAChRs. Perisynaptic loci for α7 nAChRs have also been described in the cerebellum (Caruncho et al., 1997) and hippocampus (Fabian-Fine et al., 2001), using antibodies against the α7 nAChR subunit and biotinylated αBgt.

α7 nAChRs are also associated with nonglutamatergic structures

Not all synaptic boutons in the VTA that were positive for αBgt were double-labeled with VGluT1 or VGluT2 antibodies. Thus, a proportion (25%) of presynaptic αBgt binding sites in the VTA reside on nonglutamatergic axon terminals. These are unlikely to be cholinergic because double labeling with VAChT failed to find any incidences of α7 nAChRs on cholinergic axonal boutons (Figs. 4A,8D). It is possible that αBgt labels a subset of GABAergic axon terminals because 50% of GABAergic neurons in the VTA express α7 nAChRs, as judged by single-cell PCR (Klink et al., 2001), and some of these will be local interneurons with axon terminals in the VTA.

αBgt labeling was also associated with cell bodies and dendrites of both dopaminergic (Figs. 1C, 8A,B) and nondopaminergic (Fig. 1C) neurons within the VTA. α7 nAChRs were localized on somatodendritic membranes, as well as in the cytoplasm of these neurons (Fig. 6A,B); analysis revealed that 82% of αBgt-associated gold particles were cytoplasmic. Within the cytoplasm, αBgt tended to be associated with the endoplasmic reticulum (Fig. 6B), possibly reflecting sites of receptor assembly before translocation to the plasma membrane. The presence of α7 nAChRs in dopaminergic neurons is consistent with single-cell PCR analysis (Klink et al., 2001) and with the demonstration of α7 nicotinic responses in midbrain dopaminergic neurons (Pidoplichko et al., 1997; Tsuneki et al., 2000).

αBgt binding was not observed in glial cell bodies or processes (Fig. 3). Originally, VGluT1 and VGluT2 immunoreactivity was not detected in glial cells (Fremeau et al., 2001, 2004), although a recent study has reported low levels of immunoreactivity for both transporters in a proportion of astrocytic processes in the rat hippocampus (Bezzi et al., 2004). Here, we confirm that this is also the case for the VTA because we also detected low levels of VGluT immunoreactivity in a proportion of GFAP-immunopositive astrocytes within this brain region (Fig. 3B), but αBgt labeling was never associated with GFAP immunoreactivity. Although astrocytes have been reported to express functional α7 nAChRs (Sharma and Vijayaraghavan, 2001), these studies have been performed on primary cultures and may not reflect the situation in the mature brain.

Cholinergic innervation in relation to α7 nAChRs

The second major objective of this study was to ascertain the physical relationship between α7 nAChRs and sites of ACh release. Confocal microscopy provided no evidence for the coincidence of fluorescent αBgt and cholinergic axon terminals identified by VAChT immunoreactivity in VTA sections from five rats (Fig. 4A). This was verified at the EM level, at which the juxtaposition of VAChT-positive boutons and axon terminals labeled by αBgt was never observed, despite being present in the same field (Fig. 8D). Thus, α7 nAChRs are in the vicinity of, but not opposed to, cholinergic axonal boutons. Indeed, VAChT-immunoreactive boutons were only observed in synaptic contact with dendrites in the VTA. Although these results do not preclude the rare occurrence of axoaxonic cholinergic synapses onto glutamatergic terminals in the VTA, they are compatible with other studies indicating that cholinergic afferents from the pedunculopontine tegmental nucleus synapse directly onto the somatodendritic membranes of local GABAergic interneurons and, to a lesser extent, onto dopaminergic neurons in the VTA and do not make axoaxonic synapses (Garzon et al., 1999). The lack of axoaxonic synapses is consistent with a paracrine mode of cholinergic signaling (Descarries et al., 1997; Zoli et al., 1999; Zoli, 2000) to activate presynaptic α7 nAChRs in the VTA. Indeed, the ability of α7 nAChRs to respond to choline, as well as to ACh (Alkondon et al., 1997), adds additional significance to diffuse or volume transmission for the physiological functioning of these receptors and may be pertinent to the involvement of cholinergic nicotinic mechanisms in brain reward systems (Ikemoto and Wise, 2002; Schoffelmeer et al., 2002; Fagen et al., 2003). The EC₅₀ values of ACh and choline estimated to activate α7 nAChRs (for example, 130 μm and 1.6 mm, respectively, in rat hippocampal neurons) (Alkondon et al., 1997) greatly exceed ambient levels likely to be in the low micromolar range (Descarries et al., 1997). This raises the question of whether α7 nAChRs would be desensitized rather than activated under such conditions. Functional measurements in vivo will be necessary to evaluate this relationship.

In summary, this study provides direct neuroanatomical evidence for presynaptic α7 nAChRs on a subset of glutamatergic axon terminals in the VTA. These receptors are strategically placed to enhance the release of glutamate onto dopaminergic neurons. The absence of direct cholinergic synaptic input to presynaptic α7 nAChRs indicates that these receptors are likely to be activated by endogenous ACh or choline diffusing from local cholinergic terminals. This insight into the neuroanatomical relationships in the VTA is relevant for the emerging role of the cholinergic input in reward in general, as well as for the specific mediation of the reinforcing properties of nicotine.