The mesocorticolimbic dopamine (DA) system plays important roles in reward, motivation, learning, memory, and movement. This system arises from the A10 region, comprising the ventral tegmental area and three adjacent midline nuclei (caudal linear nucleus, interfascicular nucleus, and rostral linear nucleus of the raphe). DAergic and GABAergic neurons are intermingled in this region with recently discovered glutamatergic neurons expressing the vesicular glutamate transporter 2 (VGluT2). Here, we show by in situ hybridization and immunohistochemistry that there are two subpopulations of neurons expressing VGluT2 mRNA in the A10 region: (1) a major subpopulation that expresses VGluT2 but lacks tyrosine hydroxylase (TH; VGluT2-only neurons), present in each nucleus of the A10 region, and (2) a smaller subpopulation that coexpresses VGluT2 and TH (VGluT2-TH neurons). By quantitative real-time PCR, we determined the mRNA copy numbers encoding VGluT2 or TH in samples of individual microdissected TH immunoreactive (IR) neurons. Data from both in situ hybridization and from mRNA quantification showed that VGluT2 mRNA is not present in every TH-IR neuron, but restricted to a subset of TH-IR neurons located in the medial portion of the A10 region. By integration of tract tracing, in situ hybridization, and immunohistochemistry, we found that VGluT2-only neurons and VGluT2-TH neurons each innervate both the prefrontal cortex and the nucleus accumbens. These findings establish that in addition to the well-recognized mesocorticolimbic DA-only and GABA-only pathways, there exist parallel mesocorticolimbic glutamate-only and glutamate-DA pathways.
The mesocorticolimbic dopamine (DA) system plays important roles in reward, motivation, learning, memory, and movement. This system arises from the A10 region that comprises two major nuclei of the ventral tegmental area (VTA), parabrachial pigmented (PBP) and paranigral (PN) nuclei, and three midline nuclei, the caudal linear nucleus (CLi), interfascicular nucleus (IF), and rostral linear nucleus of the raphe (RLi), that are not, strictly speaking, part of the VTA (Swanson, 1982). Within the VTA proper, DA neurons are interspersed with GABA neurons that establish local connections (Johnson and North, 1992; Omelchenko and Sesack, 2009) and innervate the prefrontal cortex (PFC) and the nucleus accumbens (nAcc) (Van Bockstaele and Pickel, 1995; Carr and Sesack, 2000) or both. Recent electrophysiological and anatomical findings indicate that glutamate signaling neurons are also present in the A10 region and that they too project to the nAcc (Chuhma et al., 2004; Hur and Zaborszky, 2005; Lavin et al., 2005; Kawano et al., 2006; Yamaguchi et al., 2007; Nair-Roberts et al., 2008; Dobi et al., 2010; Stuber et al., 2010; Tecuapetla et al., 2010).
Glutamate neurons can be identified through the detection of mRNA encoding vesicular glutamate transporters (VGluT1, VGluT2, and VGluT3), which transport glutamate into synaptic vesicles for release at presynaptic terminals. Detection of mRNA encoding these transporters is needed for the accurate labeling of cell bodies of glutamate neurons, as levels of the transporters are often undetectable with standard immunolabeling methods. Two subpopulations of neurons expressing VGluT2 mRNA are present in the A10 region (Kawano et al., 2006; Yamaguchi et al., 2007): (1) VGluT2-only neurons that express VGluT2 mRNA but lack tyrosine hydroxylase (TH) and GABA markers (Yamaguchi et al., 2007), and (2) VGluT2-TH neurons that coexpress VGluT2 mRNA and TH (Kawano et al., 2006). Whereas the subpopulation of VGluT2-only neurons is the prevalent cell population in the two nuclei of the VTA (PBP and PN) (Yamaguchi et al., 2007), a small subpopulation of VGluT2-TH neurons is found in the midline nuclei of the A10 region (Kawano et al., 2006).
A role for VGluT2-only neurons in VTA neurotransmission has been suggested from electrophysiological and anatomical findings showing that some of the VTA VGluT2 neurons establish local glutamatergic synapses on DA and non-DA neurons (Dobi et al., 2010). These studies indicate that VTA VGluT2 neurons provide excitatory local neurotransmission, contrary to the notion that all glutamatergic regulation to the A10 region is from extrinsic neurons.
To further explore the cellular heterogeneity of the A10 region, we used radioactive in situ hybridization together with immunohistochemistry to map the distribution of VGluT2-only, VGluT2-TH, and TH-only neurons in each subdivision of the A10 region. To quantify the levels of TH mRNA or VGluT2 mRNA present in individual TH cells in vivo, we used a single-cell quantitative real-time PCR (qRT-PCR) method involving microdissected cells. To investigate whether subpopulations of VGluT2-only or VGluT2-TH neurons target the nAcc or PFC, we implemented a combination of tract tracing, immunohistochemistry and in situ hybridization.
Materials and Methods
Tissue preparation for anatomical studies.
Nine adult male Sprague Dawley rats (300–350 g body weight) were anesthetized with chloral hydrate (35 mg/100 g) and perfused transcardially with 4% (w/v) paraformaldehyde (PFA) in 0.1 m phosphate buffer (PB), pH 7.3. Brains were left in 4% PFA for 2 h at 4°C, rinsed with PB, and transferred sequentially to 12, 14, and 18% sucrose solutions in PB. Coronal serial sections of 5 μm (four rats) or 12 μm (five rats) in thickness were prepared. All animal procedures were approved by the NIDA Animal Care and Use Committee.
Fluoro-Gold injections into the PFC or the nAcc.
Eight 300–330 g Sprague Dawley male rats (4 for PFC and 4 for nAcc) were anesthetized with chloral hydrate (3 ml/kg, i.p.) in a physiological saline solution. Deeply anesthetized rats were fixed in a stereotaxic apparatus, and the retrograde tracer Fluoro-Gold (FG; 1% in cacodylate buffer, pH 7.5) was delivered bilaterally into the PFC [3.0 mm anteroposterior (AP); 0.7 mm mediolateral (ML); and −4.6, −3.8, and −3.0 mm dorsoventral (DV)] or the nAcc (1.9 mm AP; 2.5 mm ML; −6.8 and −7.7 mm DV) with pipettes lowered at a 10° angle in the coronal plane. The FG was delivered iontophoretically through a stereotaxically positioned glass micropipette (inner tip diameter between 60–70 μm for PFC injections and 40 μm for nAcc injections) by applying 5 μA current in 5 s pulses at 10 s intervals for 20 min. The micropipette was left in place for an additional 10 min to prevent backflow of tracer up the injection track after each injection. One week after FG injections, the rats were perfused as indicated above.
Combination of in situ hybridization and TH immunolabeling.
Coronal free-floating sections (12 μm thick) or sections collected on glass slides (5 μm thick) were processed as described previously for free-floating sections (Wang and Morales, 2008) and for sections on glass slides (Morales and Wang, 2002). Sections were incubated for 10 min in PB containing 0.5% Triton X-100, rinsed two times for 5 min each with PB, treated with 0.2N HCl for 10 min, rinsed two times for 5 min each with PB, and then acetylated in 0.25% acetic anhydride in 0.1 m triethanolamine, pH 8.0, for 10 min. Sections were rinsed two times for 5 min each with PB and postfixed with 4% PFA for 10 min. Before hybridization and after a final rinse with PB, the free-floating sections were incubated in hybridization buffer (50% formamide, 10% dextran sulfate, 5× Denhardt's solution, 0.62 m NaCl, 50 mm DTT, 10 mm EDTA, 20 mm PIPES, pH 6.8, 0.2% SDS, 250 μg/ml salmon sperm DNA, 250 μg/ml tRNA) for 2 h at 55°C. Sections collected on glass slides were dehydrated through a series of graded ethanol (50, 70, and 95%, 5 min for each concentration). Sections were hybridized for 16 h at 55°C in hybridization buffer containing [35S]- and [33P]-labeled single-stranded antisense or sense of rat VGluT1 (nucleotides 53-2077; GenBank accession number NM-053859.1), VGluT2 (nucleotides 317-2357; GenBank accession number NM-053427), or VGluT3 (nucleotides 1-1729; GenBank accession number BC117229.1) probes at 107 cpm/ml. Plasmids that contained the VGluT1 and VGluT2 were generously provided by Dr. Robert H. Edwards (University of California, San Francisco). Sections were treated with 4 μg/ml RNase A at 37°C for 1 h, washed with 1× SSC, 50% formamide at 55°C for 1 h, and with 0.1× SSC at 68°C for 1 h. After the last SSC wash, sections were rinsed with PB and incubated for 1 h in PB supplemented with 4% bovine serum albumin and 0.3% Triton X-100. This was followed by the overnight incubation at 4°C with an anti-TH mouse monoclonal antibody (1:500; MAB 318; Millipore) for which specificity has been documented (Tagliaferro and Morales, 2008). After being rinsed three times for 10 min each in PB, sections were processed with an ABC kit (Vector Laboratories). The material was incubated for 1 h at room temperature in a 1:200 dilution of the biotinylated secondary antibody, rinsed with PB, and incubated with avidin-biotinylated horseradish peroxidase for 1 h. Sections were rinsed and the peroxidase reaction was then developed with 0.05% 3, 3-diaminobenzidine-4 HCl (DAB) and 0.03% hydrogen peroxide (H2O2). Free-floating sections were mounted on coated slides. Slides were dipped in Ilford K.5 nuclear tract emulsion (Polysciences; 1:1 dilution in double distilled water) and exposed in the dark at 4°C for 4 weeks before development.
Phenotypic characterization of retrograde labeled cells by combination of TH immunofluorescence, FG immunolabeling, and in situ hybridization.
Midbrain coronal free-floating sections (18 μm thick) were incubated for 2 h at 30°C with a mixture of rabbit anti-FG antibody (1:500; AB153; Millipore) and the mouse monoclonal anti-TH antibody (1:500) in antibody buffer (DEPC-treated PB with 0.5% Triton X-100) supplemented with RNasin (40 U/μl stock; 5 μl/ml of buffer; Promega). Sections were rinsed three times for 5 min each with DEPC-treated PB and incubated in biotinylated goat anti-rabbit antibody (1:200; Vector Laboratories) and fluorescein-conjugated donkey anti-mouse antibody (1:50; Jackson ImmunoResearch) in DEPC-treated PB supplemented with RNasin for 1 h at 30°C. Sections were rinsed three times for 5 min each with DEPC-treated PB, transferred to 4% PFA, and visualized by epifluorescence with a Leica DM LB microscope to identify FG- or TH-labeled neurons. Sections were rinsed three times for 5 min each with DEPC-treated PB, incubated for 10 min in PB containing 0.5% Triton X-100, rinsed two times for 5 min each with PB, treated with 0.2N HCl for 10 min, rinsed two times for 5 min each with PB, and then acetylated in 0.25% acetic anhydride in 0.1 m triethanolamine, pH 8.0, for 10 min. Sections were rinsed two times for 5 min each with PB, and postfixed with 4% PFA for 10 min. Before hybridization and after a final rinse with PB, the sections were incubated in hybridization buffer for 2 h at 55°C. Sections were hybridized for 16 h at 55°C in hybridization buffer containing [35S]- and [33P]-labeled single-stranded antisense or sense of rat VGluT2. Sections were treated with 4 μg/ml RNase A at 37°C for 1 h and washed with 1× SSC, 50% formamide at 55°C for 1 h and with 0.1× SSC at 68°C for 1 h. After the last SSC wash, sections were rinsed with PB and incubated for 1 h at room temperature in avidin-biotinylated horseradish peroxidase (ABC kit; Vector Laboratories). Sections were rinsed, and the peroxidase reaction was then developed with 0.05% DAB and 0.03% H2O2. Sections were mounted on coated slides. Slides were dipped in Ilford K.5 nuclear tract emulsion (Polysciences; 1:1 dilution in double distilled water) and exposed in the dark at 4°C for 4 weeks before development.
Tissue preparation for single-cell quantitative RT-PCR.
Four adult Sprague Dawley male rats (300–350 g body weight) were anesthetized with chloral hydrate (35 mg/100 g) and perfused transcardially with 10% RNAlater (Ambion) in 0.1 m PB, pH 7.4, as described previously (Sanna et al., 2005). Brains were immediately removed and frozen in isopentane at −40°C. Coronal cryosections (10 μm thick) through the midbrain (−5.04 mm to −5.40 mm from bregma) were collected onto polyethylene-naphthalene-membrane-coated microscope slides (Leica Microsystems) and stored at −80°C.
TH immunofluorescent labeling for laser microdissection and RNA extraction.
Coronal cryosections stored at −80°C were moved to −20°C and 30 min later were transferred to room temperature, air dried, and fixed with ice-cold acetone for 2 min. Fixed sections were rinsed twice in PB and incubated for 4 min with the mouse anti-TH antibody (1:40 dilution) in PB. Samples were rinsed twice with PB and incubated for 4 min with an Alexa Fluor 488-conjugated goat anti-mouse antibody (1:30 dilution; Invitrogen). Antibody solutions were supplemented with 0.1% Triton X-100 and 400 U/ml of an RNase inhibitor (RNasin; Promega). Sections were rinsed in PB, dehydrated in graded ethanol solution (70, 95, and 100%; 30 s each step), and air dried. Individual TH-positive cellular profiles were microdissected from the PBP, RLi, and IF. TH-positive cellular profiles were microdissected under a 40× objective lens and collected by gravity directly into a cap of a 0.5 ml PCR tube (Eppendorf) containing 20 μl of RNA extraction buffer. Total RNA was immediately extracted using the PicoPure RNA isolation kit (Molecular Devices). Genomic DNA was removed by DNase digestion using the DNA-free kit (Qiagen). Extracted RNA was stored at −80°C.
qRT-PCR with TaqMan primer/probes.
cDNA from extracted RNA was obtained by reverse transcription with the SuperScript III First-Strand Synthesis Super Mix (Invitrogen) and was brought to a final volume of 20 μl. For each assay, 5 μl was used. Quantification of VGluT2 or TH mRNA copy numbers from the cDNA was done using the Gene Expression Master Mix (Applied Biosystems) containing either VGluT2 or TH primers. Beta-actin (Applied Biosystems; catalog #4352931E) was amplified and used as an on-sample normalizing control. The primers were obtained from Applied Biosystems to span exons 1 to 2 of the rat VGluT2 (catalog #Rn00584780_m1) or exons 2 to 3 of the rat TH (catalog # Rn01451452_m1). The qRT-PCR was performed with the iCycler iQ Real-Time PCR Detection System (Bio-Rad Laboratories) with the following conditions: 50°C for 2 min, 95°C for 10 min, 50 cycles at 95°C for 15 s, and 60°C for 60 s. For TH and VGluT2 qRT-PCR assays, an external calibration curve was used based on a rat TH-plasmid or rat VGluT2-plasmid cDNA. For preparation of the calibration curve, samples were run in quadruplicates using serial dilutions of cDNA (5 to 105 copies). qRT-PCR results were analyzed using the iCycler iQ Real-Time PCR Detection System software and Excel software. The number of mRNA copies in each sample was interpolated from its detection threshold value using the TH- or rat VGluT2-plasmid cDNA.
Data analysis of cellular subpopulations.
Sections were viewed, analyzed, and photographed with bright-field or epiluminescence microscopy using a Nikon Eclipse E 800 microscope fitted with 4× and 20× objective lenses. Single- and double-labeled neurons were observed within each traced region at high power (20× objective lens) and marked electronically. Subdivisions of the midbrain dopamine system were traced according to Swanson (1982), Phillipson (1979a,b), Halliday and Törk (1986), German and Manaye (1993), and Paxinos and Watson (2007). TH/VGluT2 double-labeled material was analyzed using epiluminescence to increase the contrast of silver grains (neither dark-field nor bright-field optics allow clear visualization of silver grains when colocalized with high concentration of immunoproducts). A cell was considered to express VGluT2 mRNA when its soma contained concentric aggregates of silver particles above background level. A neuron was considered to express TH immunoreactivity when its soma was clearly labeled as brown. A TH-immunolabeled neuron was included in the calculation of total population of TH cells when the stained cell was at least 5 μm in diameter. The cells expressing VGluT2 mRNA, TH immunoreactivity, or both markers were counted separately. To determine cellular coexistence of VGluT2 mRNA and TH immunolabel, (1) silver grains corresponding to VGluT2 expression were focused under epiluminescence microscopy, (2) the path of epiluminescence light was blocked without changing the focus, and (3) bright-field light was used to determine whether a brown neuron, expressing TH in focus, contained the aggregates of silver grains seen under epiluminescence. Labeled cells were counted three times, each time by a different observer. The background was evaluated from slides hybridized with sense probes. For FG/TH/VGluT2 triple-labeled material, FG fluorescent cells containing or lacking TH fluorescent signal were photographed before processing for in situ hybridization (for detection of VGluT2 transcripts) and immunohistochemistry (for detection of FG in retrograde labeled cells, seen as brown because of the reaction of DAB). This FG/VGluT2 double-labeled material was analyzed following the procedure described above for the analysis of TH/VGluT2 double-labeled material. Pictures were adjusted to match contrast and brightness by using Adobe Photoshop (Adobe Systems).
Within the A10 region, many neurons express VGluT2 mRNA, but none express VGluT1 or VGluT3 mRNAs
To determine whether glutamatergic neurons are present within the different subdivisions of the A10 region, we used radioactive in situ hybridization to identify cellular expression of transcripts encoding VGluT1, VGluT2, or VGluT3 (Fig. 1B′,F′, H′). We detected cells expressing VGluT2 mRNA (but not VGluT1 or VGluT3) intermingled with TH immunoreactive (TH-IR) cells in all subdivisions of the A10 region: PBP, PN, CLi, RLi, and IF (Figs. 2A–B′, 4, 11A). The specificity of the detection of VGluT2 mRNA was confirmed by the lack of signal when sections were hybridized with the VGluT2 radioactive sense riboprobe (Figs. 1C–D′). We next determined the distribution of neurons expressing VGluT2 mRNA within the different subdivisions of the A10. There is a general agreement that the PN and the PBP are part of the VTA (Phillipson, 1979a,b; Swanson, 1982; Halliday and Törk, 1986; German and Manaye, 1993; Paxinos and Watson, 2007). However, some investigators include the RLi (Phillipson, 1979a,b; German and Manaye, 1993), the CLi (Phillipson, 1979a,b), or the IF (Phillipson, 1979a,b) as part of the VTA. Here, we followed the nomenclature established by Swanson (1982) and considered the PN and the PBP as regions of the VTA that, together with the RLi, the CLi, and the IF, comprise the A10 region.
Two types of VGluT2 mRNA-expressing neurons are present in the VTA
Using 5-μm-thick sections together with single in situ hybridization (for detection of VGluT2 mRNA) and TH immunolabeling, we found VGluT2 mRNA-expressing neurons throughout the rostrocaudal levels of the VTA intermingled with TH-IR neurons (Figs. 2⇓–4). In the lateral portions of PBP and PN, the VGluT2 cells rarely coexpressed TH (Figs. 2⇓–4). However, we did identify some neurons coexpressing VGluT2 mRNA and TH immunoreactivity (VGluT2-TH neurons) (Table 1) within the medial portion of the VTA at the level of −5.28 mm from bregma. These VGluT2-TH neurons were observed in a zone initiated at the border between the PBP and the RLi and expanding 400 μm laterally (medial PBP) (Figs. 2E,E′, 4A′, 11A), and in a zone initiated at the border between the PN and the IF and expanding 200 μm laterally (medial PN) (Figs. 4A′, 11A). In these medial zones, the VGluT2-TH neurons were a fraction of the total population of VGluT2-expressing neurons, in which their majority was VGluT2-only neurons (72–79%). However, these VGluT2-TH neurons represented about half of the total population of TH-IR neurons detected in these medial zones (Table 1). In summary, the cellular phenotyping of VTA neurons reveals an uneven cellular compartmentalization within the classically defined VTA. Whereas the lateral aspects of the VTA were comprised mostly of TH-only neurons, the medial aspects of the VTA were comprised mostly of VGluT2-only neurons with a few VGluT2-TH neurons.
Neurons expressing VGluT2 mRNA are present in each medial nuclei of the A10 region and most of them lack TH immunoreactivity
Neurons expressing VGluT2 mRNA were found in each medial nucleus of the A10 region (RLi, IF, and CLi). The VGluT2-only neurons constituted the major subpopulation of neurons expressing VGluT2 mRNA in each medial nucleus of A10 region (63–89% depending on the rostromedial level of the nucleus) (Tables 2⇓–4; Fig. 4A–D). Conversely, the VGluT2-TH neurons were a minor neuronal subpopulation in the total population of VGluT2 mRNA-expressing neurons (12–37%) (Tables 2⇓–4; Fig. 4A′–D′).
The VGluT2-TH neurons were also a minor neuronal subpopulation within the total population of TH-IR neurons (Tables 2⇑–4) detected in the middle and caudal IF (10–14% VGluT2-TH; 86–90% TH only) and in the CLi (8–17% VGluT2-TH; 83–92% TH only). However, the VGluT2-TH neurons were a major neuronal subpopulation within the total population of TH-IR neurons found in the rostromedial aspects of the RLi (about 60% VGluT2-TH; 40% TH only) and at the rostral level of the IF (44.5% VGluT2-TH; 55.5% TH only). The differential distribution of VGluT2-only, VGluT2-TH, and TH-only neurons within each of the subdivisions of the A10 region (Figs. 4, 11A) underscores the cellular difference between the VTA proper (PBP and PN) and medial subdivisions of the A10 region.
Quantification of TH mRNA or VGluT2 mRNA present in individual microdissected TH-IR cells (qRT-PCR analysis)
To address whether lack of sensitivity by in situ hybridization could account for the lack of detection of VGluT2 mRNA in TH-IR cells within some subdivisions of the A10 region, we collected individual TH-IR cell profiles by UV-laser microdissection from coronal cryosections (Fig. 5A–B′). The copy numbers of either TH mRNA or VGluT2 mRNA were evaluated from every individual microdissected TH-IR cell profile using qRT-PCR amplification curves (Fig. 5C,D). Fifty TH-IR cell profiles were analyzed from the lateral portion of the PBP. All of them were confirmed to contain TH mRNA (142 ± 13 TH mRNA copy numbers) (Table 5), but coexpression of VGluT2 mRNA was detected in only three of these cells (14 ± 9 VGluT2 mRNA copy numbers). These findings confirmed the results obtained by in situ hybridization showing that the majority of TH cells within the lateral aspects of the PBP and PN do not coexpress VGluT2 mRNA.
The qRT-PCR analysis further established that a population of TH-IR cells restricted to the medial subdivisions of the A10 region does coexpress VGluT2 mRNA. We analyzed 32 TH-IR cell profiles from the medial portion of the PBP (−5.28 mm from bregma); all of them contained TH mRNA (222 ± 33 TH mRNA copy numbers) (Table 5), and about half (43.8 ± 2.5%) coexpressed VGluT2 mRNA (13 ± 1 VGluT2 mRNA copy numbers). Thirty-eight TH-IR cell profiles were analyzed from the IF. All of them showed to contain TH mRNA (96 ± 12 TH mRNA copy numbers), but only a quarter coexpressed VGluT2 mRNA (10 ± 2 VGluT2 mRNA copy numbers). In contrast, about half of the TH-IR microdissected neurons from the RLi (54.5 ± 6.0%) coexpressed TH mRNA (76 ± 12 TH mRNA copy numbers) and VGluT2 mRNA (8 ± 1 VGluT2 mRNA copy numbers).
Therefore, within the total population of TH-IR neurons in the VTA and adjacent midline nuclei, the proportionality of VGluT2-TH neurons detected by qRT-PCR analysis of individual TH-IR microdissected neurons is similar to the proportionality of VGluT2-TH neurons found by in situ hybridization and immunohistochemistry.
The PFC and the nAcc receive inputs from VGluT2-only neurons and from VGluT2-TH neurons
As previous electrophysiological studies have shown that electrical stimulation of the VTA induces EPSCs in medium spiny neurons of the nAcc (Chuhma et al., 2009) or neurons in the PFC (Lavin et al., 2005), we next investigated whether either of the two subpopulations of VGluT2 neurons (VGluT2-only or VGluT2-TH neurons) within the A10 region innervate the PFC or the nAcc. Toward this end, we injected the retrograde tract tracer Fluoro-Gold into the PFC (Fig. 6A–E) or the nAcc (Fig. 6F–O) and used a combination of immunolabeling and in situ hybridization to phenotype the retrogradely labeled FG neurons.
FG-labeled cells from PFC injections were heterogeneously distributed within the A10 region (Table 6); most of them were located in the medial portions of the PBP (30.8%) and PN (14.5%) and in the IF (18.2%), with fewer cases in the RLi (8.9%) and the CLi (8.9%). We identified four different phenotypes of mesocortical neurons within the A10 region (Figs. 7, 8): mesocortical VGluT2-only neurons (FG-labeled cells expressing VGluT2 mRNA without TH immunoreactivity), mesocortical VGluT2-TH neurons (FG-labeled cells coexpressing VGluT2 mRNA and TH immunoreactivity), mesocortical TH-only neurons (FG-labeled cells lacking VGluT2 mRNA but expressing TH immunoreactivity), and mesocortical VGluT2-TH-negative neurons (cells lacking both VGluT2 and TH). The mesocortical VGluT2-only neurons were concentrated in the medial aspects of both the PBP and the PN, and in the IF (Fig. 8B); these areas had fewer mesocortical VGluT2-TH neurons (Fig. 8B′). The highest concentration of mesocortical VGluT2-TH neurons was in the caudal RLi (Fig. 8C). Within the total population of neurons from the A10 region targeting the PFC, 39.1% were VGluT2-only neurons, 27.1% were VGluT2-TH neurons, 24.1% were TH-only, and 9.7% were VGluT2-TH-negative neurons (Table 6; Fig. 11B). These results reveal a higher mesocortical input fromVGluT2-only neurons than from TH-only neurons.
FG-labeled neurons from nAcc injections were distributed in different subdivisions of the A10 region (Table 6). The highest concentration was in the PBP (35.4%), followed by the IF (20.1%) and the PN (14.4%). The lowest concentrations of FG-labeled cells were found in the RLi (7.4%) and the CLi (4.5%). In common with cells projecting to the PFC, we identified four different phenotypes of mesoaccumbal neurons (Figs. 9, 10): mesoaccumbal VGluT2-only neurons, mesoaccumbal VGluT2-TH neurons, mesoaccumbal TH-only neurons, and mesoaccumbal VGluT2-TH-negative neurons. The mesoaccumbal neurons expressing VGluT2 mRNA (with or without TH immunoreactivity) were clustered in the medial aspects of the A10 region (Fig. 10). In contrast, both the mesoaccumbal TH-only neurons and the mesoaccumbal VGluT2-TH-negative neurons had a mediolateral heterogeneous distribution in the A10 region (Fig. 10A″–D″). Within the total population of A10 region neurons targeting the nAcc (Table 6; Fig. 11B), half of them were mesoaccumbal TH-only cells (53.6%), about one-third were mesoaccumbal VGluT2-TH neurons (27.4%), less than one-tenth were mesoaccumbal VGluT2-only neurons (8.3%), and one-tenth were mesoaccumbal VGluT2-TH-negative neurons (10.6%). These results showed that most of the mesoaccumbal innervations were from TH neurons; these TH neurons had two different phenotypes.
This study identifies a glutamate mesocorticolimbic pathway—a pathway arising in the A10 region and projecting to the PFC and nAcc—that parallels the established DAergic and GABAergic pathways from the same region of origin to the same two major target regions. This glutamate pathway comprises two subpopulations of neurons: a subpopulation that expresses VGluT2 mRNA without TH (VGluT2-only neurons) and a subpopulation that coexpresses VGluT2 mRNA and TH (VGluT2-TH neurons). The PFC and the nAcc each receive inputs from VGluT2-only neurons and VGluT2-TH neurons from the VTA and midline A10 region.
We found a high prevalence of neurons expressing VGluT2 mRNA in each nucleus of the A10 region in mature rats. Neurons expressing VGluT2 mRNA are differentially distributed in the A10 region and fall into two neuronal subpopulations; a major subpopulation of neurons expressing VGluT2 mRNA but lacking TH (VGluT2-only neurons) with a lateromedial increasing gradient of distribution within the A10 region, and a smaller subpopulation of neurons coexpressing VGluT2 mRNA and TH (VGluT2-TH neurons) clustered in the medial aspects of the A10 region. Early findings indicated that with exception of the RLi, there is a low prevalence of VGluT2-TH neurons within the total population of TH neurons (Kawano et al., 2006). Unlike these early findings, we found by in situ hybridization and also by single-cell qRT-PCR that almost half of the TH neurons coexpress VGluT2 mRNA in the medial PBP, medial PN, RLi, and the rostral level of the IF. The differential distribution of VGluT2-only, VGluT2-TH, and TH-only neurons within the A10 region suggest (1) a differential compartmentalization of neurons with different signaling phenotype within the A10 region, (2) a major subpopulation of glutamate signaling neurons in the A10 region, and (3) a restricted high prevalence of TH neurons expressing VGluT2 in the medial aspects of the A10 region.
The roles of VGluT2-only neurons and VGluT2-TH neurons in brain function remain to be determined. A role for VGluT2 in vesicular DA filling in TH neurons has been proposed based on in vitro studies (Hnasko et al., 2010), and it has been suggested that lower levels of vesicular DA in knock-out mice depleted of VGluT2 in TH neurons might explain the attenuated locomotion induced by acute injections of cocaine or methamphetamine in these knock-out mice (Birgner et al., 2010; Hnasko et al., 2010). However, support for the coexistence of DA and VGLUT2 within the same synaptic terminal or within the same synaptic vesicle in vivo requires further investigation.
A mesocorticortical glutamate pathway from glutamate-only and glutamate-TH neurons
To successfully phenotype and quantify the glutamate-only and glutamate-TH neurons innervating the PFC or the nAcc, we implemented a novel multistep procedure that combined tracer deposit by iontophoretic application (to diminish tracer uptake by axons of passages or damage axons), use of radioactive riboprobes (to detect VGluT2 mRNA), and double immunohistochemistry (to detect TH neurons and be able to retrieve FG-containing cells after the loss of FG during the in situ hybridization procedure). By applying this experimental approach, we identified four classes of A10 neurons innervating the PFC and innervating the nAcc and found that the glutamatergic-only and the glutamatergic-DAergic neurons innervating the PFC or the nAcc are confined to the medial aspects of the A10 region. These results further support the view of neurochemical and functional differences between the lateral and medial aspects of the A10 region and strengthen the proposition that the PN and the PBP are regions of the VTA that, together with the RLi, the CLi, and the IF, comprise the A10 region (Swanson, 1982). An early study indicated the presence of VGluT2 neurons projecting to the PFC in the rat (Hur and Zaborszky, 2005). Here we show that two types of VGluT2 neurons (glutamate-only and glutamate-TH neurons) within the A10 region innervate the PFC. Thus, there exists a substantial mesocortical glutamate pathway in parallel to the well-known mesocortical DAergic and GABAergic pathways. Here, we show an unexpectedly higher mesocortical input from VGluT2-only neurons than from TH-only neurons. These anatomical findings together with previous findings showing EPSCs in the PFC after stimulation of the VTA (Lavin et al., 2005) indicate that in addition to the regulation of PFC activity by the DA mesocorticolimbic pathway, PFC activity is regulated by a glutamate mesocorticolimbic pathway from VGluT2 neurons distributed mostly in the medial aspects of the A10 region.
A mesolimbic glutamate pathway from glutamate-only and from glutamate-TH neurons
In common with the mesocortical pathway, we found that both glutamate-only and glutamate-TH neurons from the A10 region innervate the nAcc. However, most of the mesoaccumbal projections are from TH neurons, and as many as 34% of these TH neurons express VGluT2 mRNA. These findings, together with electrophysiological results showing EPSCs in nAcc spiny neurons after VTA electrical stimulation (Chuhma et al., 2009), provide evidence for a mesoaccumbal glutamatergic pathway that might be originated from VGluT2-only or VGluT2-TH neurons.
The glutamatergic signaling by TH neurons has been demonstrated recently by optogenetics methodology using transgenic mice (Stuber et al., 2010; Tecuapetla et al., 2010). In these optogenetic studies, the in vitro selective stimulation of afferents from TH neurons elicit EPSCs in all tested spiny neurons in nAcc slice preparations (Stuber et al., 2010; Tecuapetla et al., 2010); responses are no longer detected in slice preparations from conditional knock-out mice lacking VGluT2 in TH neurons (Stuber et al., 2010). The mechanism that mediates EPCS in all tested neurons in the nAcc slice preparations is unclear. Our findings from in vivo studies in the rat indicate that only a fraction of the TH mesoaccumbal neurons have VGluT2 mRNA. These findings, together with those from slice preparations, could be integrated by assuming that a single VGluT2-TH mesoaccumbal neuron affects multiple neurons in the nAcc. However, it remains to be determined whether (1) the VGluT2-TH mesoaccumbal pathway of the rat is similar to the one of the mouse; (2) a single VGluT2-TH mesoaccumbal neuron interacts with multiple neurons in the nAcc through highly arborized axons, as documented for nigrostriatal TH neurons (Matsuda et al., 2009); or (3) a single VGluT2-TH mesoaccumbal neuron interacts with a selective population of neurons in the nAcc that may affect the activation of other local neurons.
The corelease of glutamate and DA from TH neurons has been suggested from in vitro and in vivo findings (Sulzer et al., 1998; Bourque and Trudeau, 2000; Joyce and Rayport, 2000; Sulzer and Rayport, 2000; Chuhma et al., 2004, 2009; Stuber et al., 2010; Tecuapetla et al., 2010). However, anatomical evidence for the coexistence of glutamate neurotransmitter and DA at the synaptic level requires further investigation. For instance, ultrastructural studies have shown that VGLUT2-TH dual-labeled axon terminals, although present in the nAcc of immature rats, appear to be absent in the nAcc of adult rats (Bérube-Carriére et al., 2009). It remains uncertain whether this lack of TH/VGLUT2 terminals results from undetectable levels of TH in VGLUT2-positive terminals or undetectable levels of VGLUT2 in TH-positive terminals. In this regard, we found by qRT-PCR that TH-only neurons and VGluT2-TH neurons have similar amounts of TH mRNA. Thus, it seems unlikely that lack of TH detection in VGLUT2 terminals making asymmetric synapses, but not in TH terminals making symmetric synapses, results from lower levels of TH transcripts. Asymmetric synapses have been traditionally associated with excitatory synapses, and indeed almost all VGLUT2 terminals make asymmetric synapses in different brain areas. So, even though lower levels of VGLUT2 protein may exist in TH terminals in the nAcc, this does not necessary explain the well-documented infrequent occurrence of TH-asymmetric synapses in the rat nAcc (Arluison et al., 1984; Bouyer et al., 1984; Voorn et al., 1986). As an alternative, lack of coexistence of VGLUT2 and TH within the same axon terminal may be explained by a segregation of VGLUT2 terminals and TH terminals from a common VGluT2-TH neuron. A segregation of glutamatergic and DAergic terminals appears to occur in cultured DA neurons (Sulzer et al., 1998; Joyce and Rayport, 2000). We speculate that if VGluT2-TH neurons have the capability to segregate TH and VGluT2 into different processes, this may result in the dendritic production and release of DA within the VTA, but vesicular glutamate accumulation and synaptic release in target areas, such as the nAcc.
VGluT2 neurons are present in each subdivision of the A10 region and may provide fast non-DA excitatory mesocorticolimbic signaling. The A10 region contains two classes of VGluT2 neurons: VGluT2 neurons lacking TH (glutamate-only neurons), which are present in all subdivisions of the A10 region, and VGluT2 neurons coexpressing TH (glutamate-TH neurons), which are restricted to the medial portions of the A10 region. Both classes of A10-glutamate neurons innervate the PFC and the nAcc; thus, we propose that in addition to the well-recognized mesocorticolimbic DA-only and GABA-only pathways, there exist parallel mesocorticolimbic glutamate-only and glutamate-DA pathways. Our documentation of mesocorticolimbic glutamate pathways in the adult rat may provide the theoretical background to experimentally advance the suggestion that glutamatergic signaling from the VTA may play a role in a fast neurotransmission of salient stimuli (Lapish et al., 2006, 2007).
This work was supported by the Intramural Research Program of the National Institute on Drug Abuse. We thank Dr. Roy Wise for discussions.
- Correspondence should be addressed to Marisela Morales, Intramural Research Program, Neuronal Networks Section, National Institute on Drug Abuse, 251 Bayview Boulevard, Baltimore, MD 21224.