Mesocortical dopamine (DA) is a key neurotransmitter in cognitive processes and is involved in schizophrenia and antipsychotic drug action. DA exerts a highly complex modulation of network activity in prefrontal cortex (PFC), possibly due to the recruitment of multiple signaling pathways and to specialized cellular localizations of DA receptors in cortical microcircuits. Using double in situ hybridization, we quantitatively assessed the expression of D1 and D2 receptor messenger RNAs (mRNAs) in pyramidal and γ-aminobutyric acidergic (GABAergic) neurons of rat PFC. The proportion of pyramidal and GABA cells expressing these transcripts shows great regional variability in PFC, with little overlap (layer V). More pyramidal and GABA cells express D1 than D2 receptors. D1 receptors are expressed by a greater proportion of GABA than pyramidal neurons, yet the number of D1-positive pyramidal cells outnumbers D1-positive interneurons due to the greater abundance of pyramidal neurons. Occasional PFC cells show high levels of mRNA, similar to those in striatal neurons. Finally, pyramidal and GABAergic cells expressing the same transcript were almost never found in close apposition, yet D2-containing pyramidal neurons were often found close to non-D2 GABA neurons. Thus, cellular and network DA actions in PFC are region and layer specific and may depend on precise cellular interactions.
Brain dopamine (DA) is involved in a wide variety of functions including motivation, attention, reward, cognition, and movement (Graybiel et al. 1994; Williams and Goldman-Rakic 1995; Schultz 1998; Koob and Le Moal 2001; Dalley et al. 2004; Grace et al. 2007; Iversen and Iversen 2007). Derangements of the DA pathways arising in the ventral tegmental area are suspected in severe psychiatric conditions such as schizophrenia, depression, attention deficit hyperactivity disorder, and in drug addiction (Grace 1991; Willner et al. 1992; Everitt and Robbins 2000; Koob and Le Moal 2001; Castellanos and Tannock 2002). In particular, the mesocortical DA system plays a major role in cognitive processes by modulating the memory fields of pyramidal neurons in the dorsolateral prefrontal cortex (PFC). Partial blockade of D1 receptors in monkey PFC magnified the persistent neuronal activity during the delay period in working memory tasks (Williams and Goldman-Rakic 1995), yet a complete D1 receptor blockade canceled this effect, suggesting a bell-shaped relationship between DA D1 receptor occupancy and cognitive performance (Goldman-Rakic et al. 2000; Vijayraghavan et al. 2007). On the other hand, blockade of DA D2/3 receptors attenuated the activity of cortical neurons during the delay period in this type of tasks (Williams and Goldman-Rakic 1995). Moreover, the cognitive deficits induced by chronic D2/3 receptor blockade by antipsychotics are reversed by D1 agonist administration (Castner et al. 2000).
The cellular basis for the DA actions in PFC have been extensively investigated, and a plethora of studies have examined the effect of endogenously or exogenously applied DA or of selective D1 and D2 receptor agonists in the PFC of rodents and nonhuman primates (Thierry et al. 1990; Grace 1991; O'Donnell 2003). The vast majority of studies have assessed the effects of DA receptor stimulation on projection (pyramidal) neurons, whereas some examined DA effects on γ-aminobutyric acidergic (GABAergic) interneurons. However, despite an extensive research, there is not a unified view of DA actions in PFC, and many of these studies have yielded nonconvergent and often contradictory conclusions (for review, see Seamans and Yang 2004).
This situation possibly results from the use of different methodological approaches (e.g., in vitro vs. in vivo) and the multiplicity of short-term (milliseconds) and long-term (minutes, hour) signaling mechanisms coupled to D1 and D2 receptor families (Svenningsson et al. 2004; Tseng and O'Donnell 2004; Beaulieu et al. 2007; Bolan et al. 2007; Lapish et al. 2007). Also, the ability of DA to modulate pyramidal activity directly and indirectly (via DA receptor activation in GABAergic interneurons) (Zhou and Hablitz 1999; Gorelova et al. 2002; Tseng and O'Donnell 2007a) adds a further element of complexity. Moreover, recent studies indicate that the D1 and D2 receptor–mediated effect of DA on PFC interneurons is age and cell type dependent (Tseng and O'Donnell 2007b). Thus, the net effect of DA or dopaminergic agents on output cells of the PFC may depend on the relative abundance of each DA receptor in recorded cell and in local network interneurons and on the different signaling mechanism involved.
In order to provide a more detailed insight on the cellular basis of the DA actions in PFC, we performed a detailed quantitative estimation of the proportion of pyramidal and GABAergic neurons expressing DA D1 and D2 receptors in the various areas of the rat PFC, paying also a special attention to their layer distribution and relative localization.
Materials and Methods
Male albino Wistar rats (250–300 g) were from Iffa Credo, Lyon, France. Animals were kept in a controlled environment (12-h light–dark cycle and 22 ± 2 °C room temperature) with food and water provided ad libitum. Animal care followed the European Union regulations (O.J. of E.C. L358/1 18/12/1986) and was approved by the local Institutional Animal Care and Use Committee. The rats were killed by decapitation, and the brains were rapidly removed, frozen on dry ice, and stored at −20 °C. Tissue sections, 14-μm thick, were cut using a microtome–cryostat (HM500 OM; Microm, Walldorf, Germany), thaw mounted onto 3-aminopropyltriethoxysilane (Sigma, St Louis, MO) coated slides, and kept at −20 °C until use.
The oligodeoxyribonucleotide probes used were complementary to the following bases: 752–796, 1805–1852, and 797–841 of the rat D1 receptor messenger RNA (mRNA) (GenBank Accession No. NM_012546.1); 347–388, 1211–1258, and 1211–1258 of the D2 receptor mRNA (GenBank Accession No. NM_012547.1); 514–558 of the GAD65 isoform of the enzyme glutamate decarboxylase mRNA (GenBank Accession No. NM_012563); 191–235 of the GAD67 isoform mRNA (GenBank Accession No. NM_017007); and 127–172 and 1756–1800 of the vGluT1 mRNA (vesicular glutamate transporter) (GenBank Accession No. U07609). The probes for D1 and D2 receptors were synthesized on a 380 Applied Biosystem DNA synthesizer (Foster City Biosystem, Foster City, CA) and purified on a 20% polyacrylamide/8 M urea preparative sequencing gel. The rest of the probes were synthesized and high-performance liquid chromatography purified by Isogen Bioscience BV (De Meern, The Netherlands). Each D1 and D2 receptor oligonucleotide was individually labeled (2 pmol) at the 3′-end with [33P]-dATP (>2500 Ci/mmol; DuPont-NEN, Boston, MA) using terminal deoxynucleotidyltransferase (TdT, Calbiochem, La Jolla, CA). GAD and vGluT oligonucleotides (100 pmol) were nonradioactively labeled with TdT (Roche Diagnostics GmbH, Mannheim, Germany) and digoxigenin-11-deoxyuridine triphosphate (Boehringer Mannheim). Oligonucleotides were purified by centrifugation using QIAquick Nucleotide Removal Kit (QIAGEN GmbH, Hilden, Germany).
In Situ Hybridization Histochemistry Procedure
The protocols for single and double label in situ hybridization was used as previously described (Serrats et al. 2003). Frozen tissue sections were first brought to room temperature, fixed for 20 min at 4 °C in 4% paraformaldehyde in phosphate-buffered saline (1× PBS: 8 mM Na2HPO4, 1.4 mM KH2PO4, 136 mM NaCl, and 2.6 mM KCl), washed for 5 min in 3× PBS at room temperature, twice for 5 min each in 1× PBS, and incubated for 2 min at 21 °C in a solution of predigested pronase (Calbiochem, San Diego, CA) at a final concentration of 24 U/mL in 50 mM Tris–HCl, pH 7.5, and 5 mM ethylenediaminetetraacetic acid (EDTA). The enzymatic activity was stopped by immersion for 30 s in 2 mg/mL glycine in 1× PBS. Tissues were finally rinsed in 1× PBS and dehydrated through a graded series of ethanol. For hybridization, the radioactively labeled and the nonradioactively labeled probes were diluted in a solution containing 50% formamide, 4× standard saline citrate (SSC) (1× SSC: 150 mM NaCl and 15 mM sodium citrate), 1× Denhardt's solution (0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin), 10% dextran sulfate, 1% sarkosyl, 20 mM phosphate buffer, pH 7.0, 250 μg/mL yeast tRNA, and 500 μg/mL salmon sperm DNA. The final concentrations of radioactive and Dig-labeled probes in the hybridization buffer were in the same range (∼1.5 nM). Tissue sections were covered with hybridization solution containing the labeled probes, overlaid with Nescofilm coverslips (Bando Chemical Ind., Kobe, Japan), and incubated overnight at 42 °C in humid boxes. Sections were then washed 4 times (45 min each) in a buffer containing 0.6 M NaCl and 10 mM Tris–HCl (pH 7.5) at 60 °C.
Development of Radioactive and Nonradioactive Hybridization Signal
Hybridized sections were treated as described previously (Serrats et al. 2003). Briefly, after washing, the slides were immersed for 30 min in a buffer containing 0.1 M Tris–HCl, pH 7.5, 1 M NaCl, 2 mM MgCl2, and 0.5% bovine serum albumin (Sigma) and incubated overnight at 4 °C in the same solution with alkaline phosphate–conjugated anti-digoxigenin-F(ab) fragments (1:5000; Boehringer Mannheim). Afterward, they were washed 3 times (10 min each) in the same buffer (without antibody) and twice in an alkaline buffer containing 0.1 M Tris–HCl, pH 9.5, 0.1 M NaCl, and 5 mM MgCl2. Alkaline phosphatase activity was developed by incubating the sections with 3.3 mg nitroblue tetrazolium and 3.3 mg bromochloroindolyl phosphate (Gibco BRL, Gaithersburg, MD) diluted in 10 ml of alkaline buffer. The enzymatic reaction was blocked by extensive rinsing in the alkaline buffer containing 1 mM EDTA. The sections were then briefly dipped in 70% and 100% ethanol, air-dried, and dipped into Ilford K5 nuclear emulsion (Ilford, Mobberly, Cheshire, UK) diluted 1:1 with distilled water. They were exposed in the dark at 4 °C for 5 weeks and finally developed in Kodak D19 (Kodak, Rochester, NY) for 5 min and fixed in Ilford Hypam fixer (Ilford).
Specificity of the Probes
The specificity of the hybridization signals has been previously established and published (Pompeiano et al. 1992; Serrats et al. 2003). These controls included the following procedures. 1) The thermal stability of the hybrids obtained was checked for every probe. 2) For a given oligonucleotide probe, the hybridization signal was completely blocked by competition of the labeled probe in the presence of 50-fold excess of the same unlabeled oligonucleotide. 3) Because we synthesized more than one probe for each mRNA analyzed, the hybridization signal obtained with each oligonucleotide for the same mRNA was identical at both regional and cellular levels when used independently. 4) To assure the specificity of the nonradioactive hybridization signal, we compared the results obtained with the same probe radioactively labeled.
Analysis of the Results
Tissue sections were examined in bright and dark field in a Wild 420 macroscope (Leica, Heerbrugg, Germany) and in a Nikon Eclipse E1000 microscope (Nikon, Tokyo, Japan) equipped with bright and dark field condensers for transmitting light and with epi-illumination. Micrography was performed using a digital camera (DXM1200 3.0; Nikon) and analySIS software (Soft Imaging System GmbH, Münster, Germany). Bright field images were captured with transmitted light. Dark field images were captured with Darklite illuminator (Micro Video Instruments, Avon, MA). The figures were prepared for publication using Adobe Photoshop software (Adobe Software, Mountain View, CA). The cellular counting was made in an Olympus AX70 Stereo Microscope using CAST software for stereological analysis. Glutamatergic and GABAergic cells (Dig-labeled probes) were identified as cellular profiles exhibiting a dark precipitate (alkaline phosphatase reaction product) clearly distinguishable from background. D1 and D2 mRNA hybridization signal (33P-labeled probes) was considered positive when accumulation of silver grains over the stained cellular profiles was 4 times greater than the background signal. Due to the presence of silver grains at different focal levels, a careful examination of each putatively positive cell was carried out at the microscope to validate positive signals. Cells were counted in 4 adjacent sections of each rat and averaged to obtain individual values. Results are given as mean ± standard error of the mean of 3 different rats. Data have been analyzed using 1- or 2-way analysis of variance (ANOVA) followed by post hoc Tukey's test. P <0.05 was considered statistically significant.
Distribution of D1 and D2 Receptor mRNAs in Rat PFC
Figure 1 shows dark field images of the localization of DA D1 and D2 receptor mRNAs at different antero-posterior (AP) levels of the PFC, corresponding approximately to bregma +4.20, +3.24, and +3.00 mm (Paxinos and Watson 2005). In agreement with the denser dopaminergic innervation of deep layers in rat PFC, both receptor transcripts were more densely expressed in layers V–VI than in superficial or intermediate layers in the medial PFC. They were relatively abundant in the cingulate, prelimbic, and infralimbic cortices, with an overlapped expression in layer V and some remarkable differences in layers II–III and VI (see below). Both receptors, specially the D2 receptor mRNA, were also expressed by cells in the agranular insular cortex, the claustrum, and the endopiriform nucleus.
Overall, the expression of DA D1 and D2 receptors in PFC was much lower than in striatal areas. For comparison, note the high expression of both transcripts in the more rostral part of “nucleus accumbens” and olfactory tubercle in Figure 1 (panels A3 and B3; see also Fig. 2).
Expression of DA D1 Receptors in Pyramidal (vGluT1 Positive) and GABAergic (GAD65/67 Positive) Neurons of the Rat PFC
In a previous report examining the cellular localization of 5-HT1A and 5-HT2A receptors in PFC, we examined the distribution of pyramidal (vGluT1 positive) and GABAergic neurons (GAD positive) in rat PFC using the same methodology, and rats of the same age and weight, than those used in the present study (Santana et al. 2004). High densities of pyramidal cells, as labeled by vGluT1 mRNA, were found at various cortical levels. Dense clusters of these cells were observed in the tenia tecta and piriform cortex which also showed a greater density of label compared with that in other cortical areas, such as the prelimbic area or the anterior cingulate. In contrast, vGluT1-expressing cells were absent in layer I, as expected, whereas GAD-expressing cells were scattered throughout all PFC layers, including layer I close to midline. We calculated the percentage of vGluT1- and GAD-positive cells by reference to the total number of Nissl-stained cells in adjacent sections. These values were 75 ± 5% for vGluT1- and 16 ± 1% for GAD-positive cells (prelimbic area, data from 3 rats; each individual value is the average of 3 adjacent sections except for the Nissl-stained section, which were duplicate sections). These values were used for the calculation of the absolute proportions of pyramidal and GABAergic interneurons expressing D1 or D2 receptors in the different layers shown in Figure 3 (see below).
D1 receptor mRNA was found in a significant proportion of pyramidal and GABAergic cells in various subdivisions of the PFC. Figure 2 shows examples of pyramidal (vGluT1 positive) and GABAergic (GAD positive) neurons expressing the transcript for D1 (panels A1 and B1, respectively) and D2 receptors (A2 and B2, respectively). Panels A3 and B3 show the presence of abundant GABAergic cells expressing D1 receptors (A3) and D2 receptors (B3) in the more rostral part of the nucleus accumbens, approximately at bregma +2.70 mm (Paxinos and Watson 2005). Contrary to the pyramidal and GABAergic cells in PFC, which expressed these DA receptors in a limited proportion, nearly all GABAergic cells in the nucleus accumbens expressed a high density of D1 and D2 receptor mRNAs.
A careful inspection of tissue sections revealed the existence of local relationships between GABAergic cells not expressing D2 receptors with pyramidal neurons expressing D2 receptors. These cells were adjacent to each other and were found mainly in layer V of the mPFC. Several examples are given in lower panels of Figure 2 (C1–C6). Although we did examine neither the anatomical relationships nor the Ca2+-binding protein of these interneurons, the proximity of the cell bodies and the size of the digoxigenin label (GAD+) suggest that they may be large perisomatic GABA cells (e.g., basket or chandelier cells). Unlike the examples shown in panels C1–C6 of Figure 2, pyramidal and GABAergic neurons expressing the same receptor mRNA (either D1 or D2) were very rarely found in close proximity.
Table 1 shows the percentage of pyramidal and GABAergic neurons in the various layers of the mPFC (prelimbic area) expressing D1 and D2 receptor mRNAs. Overall, the proportion of cells expressing D1 receptors was greater than that expressing D2 receptors. Two-way ANOVA analysis of the data in pyramidal neurons revealed a significant effect of the receptor type (F1,4 = 24.5, P < 0.01), layer (F2,8 = 45.0, P < 0.0001), and receptor × layer interaction (F2,8 = 48.6, P < 0.0005). The proportion of pyramidal neurons expressing D1 receptors was similar in layers II–III and V (19–21%) and was significantly higher in layer VI (38%). Pyramidal cells expressing D2 receptors were more abundant in layer V (25%) than in layers II–III (5%) or VI (13%).
|Layers II–III||Layer V||Layer VI|
|Pyramidal neurons||19 ± 3a||21 ± 2a||38 ± 3bc|
|GABAergic neurons||28 ± 1||30 ± 2||38 ± 4|
|Pyramidal neurons||5 ± 1ac||25 ± 2ab||13 ± 1bc|
|GABAergic neurons||5 ± 2||8 ± 2||17 ± 2b|
|Layers II–III||Layer V||Layer VI|
|Pyramidal neurons||19 ± 3a||21 ± 2a||38 ± 3bc|
|GABAergic neurons||28 ± 1||30 ± 2||38 ± 4|
|Pyramidal neurons||5 ± 1ac||25 ± 2ab||13 ± 1bc|
|GABAergic neurons||5 ± 2||8 ± 2||17 ± 2b|
P < 0.05 versus layer VI, Tukey test post-ANOVA.
P < 0.05 versus layers II–III.
P < 0.05 versus layer V.
Data are means ± standard errors of the mean of 3 rats (4 adjacent sections per rat) and show the percentage of neurons of each type (pyramidal or GABAergic) expressing D1 or D2 receptors in superficial, intermediate, or deep layers of the prelimbic area in the mPFC.
A similar analysis for the data in GABAergic cells revealed a significantly greater abundance of D1- than D2-positive neurons (F1,4 = 159.6, P < 0.0005) and significant differences between layers (F2,8 = 13.3, P < 0.003) with a nonsignificant receptor × layer interaction. Contrary to the remarkable layer differences in the expression of DA D1 and D2 receptors in pyramidal neurons, the distribution of both transcripts in GABAergic neurons was quite homogenous across layers, with the only exception of D2-positive GABAergic cells in layer VI (17% vs. 5 and 8% in layers II–III and V, respectively). Figure 3 shows the percentages of PFC neurons (either pyramidal or GABAergic) expressing D1 and D2 receptors in layers II–III, V, and VI.
Differential Regional and Cellular Expression of D1 and D2 Receptor mRNAs
Despite the overall greater density of D1 and D2 receptor mRNAs in deep layers (V, VIa, and VIb) of the mPFC, there was little overlap (mainly in layer V) and there were some remarkable differences in the regional and cellular expression of both receptor transcripts. On the one hand, the cellular level of expression of both transcripts by PFC neurons was much lower and heterogeneous (particularly D1 mRNA) than that by GABA cells in the rostral part of nucleus accumbens (Fig. 2; A3–B3), as judged from the density of silver grains, which is proportional to the mRNA present in the labeled cells (Jonker et al. 1997). Occasional PFC cells were found that expressed very high levels of the transcripts, particularly of D1 receptors. The D2 receptor transcript appeared to be more homogeneously expressed and with an average cellular density greater than that of D1 receptors. Figure 4 shows some examples of pyramidal and GABAergic cells expressing high levels of D1 receptor mRNA. A high expression of D2 receptor mRNA was only observed in pyramidal neurons. These cells had a less conspicuous label than those expressing high levels of D1 receptors and were mainly found in layer V of the mPFC.
In addition to individual differences in the cellular expression of D1 and D2 receptors, there were marked differences in their regional distribution. Hence, DA D2 receptor mRNA had a more restricted distribution in mPFC than D1 receptor mRNA. Figure 5 shows a detailed comparison of the expression of D1 and D2 receptor mRNAs in the prelimbic and cingulate subdivisions of the PFC. Cells expressing DA D2 receptor mRNA were primarily concentrated in layer V, whereas DA D1 receptor mRNA showed a more widespread distribution, with a larger number of positive cells in layer VI (particularly VIb; see Fig. 1A–C and Fig. 5) together with a thin layer of positive cells in layer II–III, putatively in layer II. Although in rat mPFC it is difficult to distinguish between layers II and III (Swanson 1998), these D1-positive cells were consistently found very close to layer I, which would likely correspond to layer II. These superficial D1-positive neurons were restricted to the infralimbic, prelimbicand—to a lesser extent—cingulate subdivisions of the PFC; dorsal and lateral aspects of the PFC were devoid of these cells. Also, D1-positive neurons were found in layer V, overlapping with D2-positive cells.
In these 2 locations where D2 receptors are not expressed (layer II and VI), the D1 receptor mRNA was present in pyramidal and GABAergic neurons. Panels B1 and C1 in Figure 5 show, respectively, pyramidal and GABA cells expressing D1 receptor mRNA in layer II. The absence of D2 receptor mRNA in this cortical layer is shown in panel A2 (dark field), whereas panels B2 and C2 show the presence of pyramidal (B2) and GABAergic (C2) neurons not expressing D2 receptor mRNA (the presence of D1 receptor mRNA in layer VI and of D2 receptor mRNA in layer V is shown in Fig. 2).
Regional differences in the expression of D1 and D2 receptor mRNAs were also found in ventral areas at this AP level. D1 receptor mRNA was much more abundant than D2 receptor mRNA in the tenia tecta (dorsal and ventral) and the piriform cortex (Fig. 1). These differences are shown in more detail in Figs 6 and 7, respectively. In both locations, there was a substantial proportion of pyramidal cells expressing D1 (but not D2) receptor mRNA, as shown in panels B1 and B2, respectively, of the corresponding figures. GABAergic cells also expressed D1 receptors (not shown).
Comparison with Other Monoamine Receptors Present in PFC
In order to provide a general view of the possible influence of monoamines in PFC cell excitability, Figure 8 shows a comparison of the laminar distribution of DA D1, DA D2, and serotonin 5-HT1A, 5-HT2A, and 5-HT3 receptors in mPFC (panels A1–A5) and their expression in glutamatergic and GABAergic cells (panels B1–C5) of prelimbic cortex. Overall, the population of cells expressing 5-HT1A and 5-HT2A receptors is greater than that expressing DA D1 and D2 receptors.
Numerous histological studies using in situ hybridization, immunohistochemistry, and autoradiography have established the presence of the members of D1 and D2 receptor families and their encoding mRNAs in rodent and primate PFC (Mansour et al. 1990, 1991; Bouthenet et al. 1991; Fremeau et al. 1991; Weiner et al. 1991; Mengod et al. 1992; Vincent et al. 1993; Smiley et al. 1994; Gaspar et al. 1995; Khan et al. 1998; Le Moine and Gaspar 1998; Muly et al. 1998). However, to our knowledge, there are not detailed quantitative studies on the expression of D1 and D2 receptors by pyramidal and GABA cells in PFC. This information may be relevant for the interpretation and design of functional studies examining the role of DA in the modulation of PFC networks and the involvement of DA receptors in antipsychotic drug action.
Immunohistochemical techniques have been often used to identify DA receptors in pyramidal and GABA cells (see above studies). However, the use of antibodies is not free from artifactual problems, particularly with proteins expressed in low levels, such as monoamine receptors. Cross-reactivity has led to strict recommendations for the correct use of antibodies (Saper and Sawchenko 2003; Saper 2005). Some of these recommendations may have not been used in some previous studies on DA receptors. Thus, we performed the present study using double in situ hybridization, which provides an unequivocal identification of the receptor transcripts and of the cellular phenotypes of the neurons expressing them. Glutamatergic and GABAergic neurons have been labeled with the selective markers vGluT1 and GAD65/67, respectively, which yield a clear signal even using nonradioactive, Dig-labeled probes. Despite the evidence that the 2 GAD genes are coexpressed in most GABA-containing neurones of the central nervous system (Feldblum et al. 1993; Esclapez et al. 1994), all experiments were performed using probes for both GAD isoforms (see Materials and methods) to allow for the identification of cortical GABA interneurons expressing one or other form of GAD. This also provided additional sensitivity for their identification using nonradioactive probes. This procedure was used in previous studies to examine the expression of 5-HT receptors in PFC (Puig et al. 2004; Santana et al. 2004) (a more extensive discussion of the validity of these markers can be found in Santana et al. 2004). We are also examining the expression of α-adrenoceptors (Santana N, López-Giménez JF, Milligan G, Mengod G, Artigas F, in preparation), which together with present and past studies will provide a more complete picture of the quantitative distribution of 5-HT and catecholamine receptors in pyramidal and GABAergic neurons of the PFC. Yet, a clear limitation of in situ hybridization is that it does not allow to establish the subcellular localization of the protein. Moreover, in situ hybridization cannot detect the presence of axonal receptors whose mRNA is located in cell bodies of afferent neuronal groups to PFC.
The sum of the percentages of vGluT1- and GAD-positive cells (75 ± 5% and 16 ± 1%, respectively) obtained in our results does not reach the 100% of PFC cells obtained in adjacent Nissl-stained sections. This difference does not appear to be caused by the presence of vGluT2-positive pyramidal neurons in PFC (Santana N, Mengod G, Artigas F, unpublished data). More likely, it is due to methodological reasons. The percentage of GAD-positive neurons in PFC (16 ± 1%) agrees well with that reported previously in other cortical areas (15%, Beaulieau 1993). One possibility to explain the difference is that we were very strict in defining positive neurons, so that some false negatives may exist. Also, the difference may be lower than apparent due to the deviation of the measures.
Cellular and Regional Heterogeneity of D1 and D2 Receptor Expression in PFC
The main findings of the present study are as follows: 1) D1 and D2 receptor mRNAs are segregated in different neuronal populations, with the exception of layer V, which contains a similar percentage of cells expressing one or other receptor transcript. 2) The proportion of pyramidal and GABA cells expressing one or other transcript varies considerably among PFC areas and layers. 3) More pyramidal and GABA cells express D1 than D2 receptors. 4) A relatively greater proportion of GABA than pyramidal neurons express D1 receptors (although the absolute number of pyramidal neurons expressing this receptor is greater, due to the 4.5:1 ratio in the abundance of pyramidal vs. GABAergic neurons; see Fig. 3). 5) The cellular level of expression of DA D1 and D2 receptor mRNAs by PFC cells is much lower than that in striatal GABAergic neurons, yet occasional cells show a high level of expression, particularly of D1 receptor mRNA and 6) pyramidal and GABAergic cells expressing the same receptor subtype were almost never found in close proximity, yet pyramidal neurons expressing D2 receptors were often found in apposition with large, possibly perisomatic, GABA cells not expressing D2 receptors.
In accordance with previous literature, D1 receptor mRNA was mainly found in deep layers (V–VI) of mPFC as well as in perirhinal areas and piriform cortex (Weiner et al. 1991). D1 receptor mRNA was also found in layer II of the mPFC (Mansour et al. 1991; Gaspar et al. 1995). In contrast, D2 receptor mRNA was essentially localized in layer V of the mPFC (Mansour et al. 1990; Bouthenet et al. 1991; Weiner et al. 1991; Gaspar et al. 1995; Le Moine and Gaspar 1998). Ligand-binding studies showed also the presence of both receptors in deep layers of mPFC (Mansour et al. 1990), in good match with the mesocortical DA innervation (Lindvall et al. 1978; Thierry et al. 1978; van Eden et al. 1987).
Interestingly, there is an striking similarity between the distribution of D1 receptor mRNA in PFC and that of DA- and adenosine 3′:5′ monophosphate–regulated phosphoprotein (DARPP-32). DARPP-32 is major target for DA-activated adenylyl cyclase in striatum (Walaas et al. 1983; Greengard 2001) and contributes to acute D1 receptor–mediated responses at the cellular and behavioral levels (Bibb et al. 1999; Hemmings et al. 1984, Greengard et al. 1999). Our results show a prominent D1 receptor expression in layers II and V–VI of PFC, which is in good agreement with previous inmunocytochemical and in situ hybridization experiments, where DARPP-32 was found in layers II–III and VI throughout the neocortex and particularly in PFC. DARPP-32 was found to be enriched in D1-containing neurons (Ouimet et al. 1984; Walaas and Greengard 1984; Schalling et al. 1990).
Despite the discordant effects of DA on PFC cells reported in functional studies, some common features have been proposed (Seamans and Yang 2004; Lapish et al. 2007). Briefly, DA exerts biphasic effects on PFC cells, both at the temporal and concentration levels. Second, DA can activate multiple signaling mechanisms through the same receptor and can exert synapse- and cell-specific effects, often through a direct interaction with N-methyl-D-aspartic acid and GABA receptors. Likewise, DA effects depend on the membrane potential of the recorded cell and on the activity level of local networks, and DA can have opposite effects at low and high activity levels. Finally, DA can exert long-lasting actions that extend beyond the application period. The present results can provide a further insight into the complexity of these actions by identifying the cellular elements involved. First, the presence of D1 and D2 receptor transcripts in pyramidal and GABAergic neurons is consistent with the existence of direct and GABA-mediated DA actions on pyramidal cells through both receptors (Thierry et al. 1990; Grace 1991; O'Donnell 2003). The marked differences in the proportion of pyramidal and GABAergic cells expressing each receptor subtype in different cortical layers and PFC subdivisions is also consistent with the large variability of DA effects in mPFC. Indeed, early in vivo studies already indicated a differential action of DA on superficial and deep layers (Sesack and Bunney 1989). On the other hand, the presence of high levels of D1 receptors in some PFC cells and the observed associations between non-D2 GABAergic and D2-positive pyramidal cells suggest the existence of microenvironments with differential physiological characteristics from the surrounding tissue.
The high proportion of GABA cells expressing D1 receptors (but not of D2 receptors) may perhaps be related to the inverted-U relationship between D1 receptor activation and cognitive performance in nonhuman primates (Goldman-Rakic et al. 2000), although it remains to be established whether the present results can be extrapolated to primate brain.
DA has been suggested to increase the signal/noise ratio in mPFC through the concurrent activation of D1 and D2 receptors, which results in an increased time in “up” (e.g., more depolarized) states (D1 effect) and a concurrent reduction of firing probability (D2 effect) (O'Donnell 2003). The present data indicate that this effect can take place only in layer V of the mPFC, where a substantial proportion (20–25%) of pyramidal cells may express one or other receptor (although we do not examine whether both receptors colocalize in the same neurons). Given the connectivity of layer V pyramidal neurons, one corollary of the present observations is that DA may predominantly filter cortical information reaching subcortical targets.
Marked differences in the expression of D1 and D2 receptors were also noted in ventral areas. Pyramidal and GABA cells in the piriform cortex and tenia tecta expressed D1 but not D2 receptor mRNA. Few studies have examined the role of monoamines, and particularly DA, in piriform cortex (Sheldon and Aghajanian 1990; Gellman and Aghajanian 1993, 1994; Marek and Aghajanian 1994, 1996a, 1996b). Bath application of 5-HT, noradrenaline, and DA increased GABAA-mediated inhibitory postsynaptic potentials recorded in pyramidal neurons in vitro (Gellman and Aghajanian 1993). Additionally, direct excitatory actions of noradrenaline and 5-HT have been reported in piriform pyramidal neurons, mediated, respectively, by α1B-adrenoceptors and 5-HT2A receptors (Marek and Aghajanian 1996a, 1996b). These observations are consistent with the presence of 5-HT2A (Santana et al. 2004), DA D1 receptors (this study), and α1-adrenoceptors (Santana et al., unpublished data) in pyramidal and GABA interneurons of the piriform cortex. However, to our knowledge, no direct excitatory actions of DA have been reported in piriform pyramidal neurons, which, according to the present observations, could be directly excited through D1 receptor activation.
Comparison with 5-HT Receptors
DA and 5-HT appear to play different roles in the modulation of PFC-based functions (Robbins 2005), perhaps due to the control of different neuronal populations in PFC by both transmitters as well as to the activation of different signaling mechanisms. To examine the first possibility, we compared the distribution of cells expressing D1 and/or D2 receptors with those expressing 5-HT receptors (5-HT1A, 5-HT2A, and 5-HT3 receptors), obtained using the same methodology than used herein (Amargos-Bosch et al. 2004; Puig et al. 2004; Santana et al. 2004). 5-HT1A and 5-HT2A receptors were rather homogenously expressed by ∼50–60% of pyramidal neurons in the mPFC, except 5-HT2A receptors in layer VI, which were expressed only by ∼26% of pyramidal neurons (Santana et al. 2004). This regional and cellular expression pattern differs markedly from that of 5-HT3 receptors, selectively localized in GABAergic neurons of all mPFC layers yet enriched in layer I (40%, 18%, 6%, and 8% of all GABA cells in layers I, II–III, V, and VI, respectively [Puig et al. 2004]). This suggests that 5-HT and DA differentially modulate the activity of cells and neuronal networks in mPFC, with some overlap among 5-HT1A, 5-HT2A, D1, and D2 receptors in layer V and between 5-HT1A and D1 receptors in layer VI.
The lower proportion of PFC cells expressing DA receptors (compared with 5-HT receptors) cannot be attributed to methodological factors (e.g., poor labeling of DA receptor mRNAs) because virtually all cells in the nucleus accumbens and olfactory tubercle in the same tissue sections showed a very high abundance of both transcripts (Figs 1 and 2).
A second difference between DA and 5-HT receptors is the relative proportion of GABA cells expressing one or other transcript. Hence, the proportion of pyramidal neurons expressing 5-HT1A and/or 5-HT2A receptors in layers III–VI is 2–3 times greater than that of GABAergic neurons. In contrast, this ratio is lower than 1 for D1 receptor mRNA (see Table 1). Likewise, the proportion of pyramidal and GABA cells in rat PFC expressing α1-adrenoceptors is very similar (Santana et al., in preparation). This suggests that excitatory actions of catecholamines on GABAergic neurons are more relevant than those of 5-HT in intermediate–deep layers, but not in superficial layers, where up to 40% of GABAergic interneurons contain excitatory 5-HT3 receptors whose physiological activation results in large GABAergic excitations (Puig et al. 2004).
In summary, the expression of DA D1 and D2 receptors by pyramidal and GABAergic neurons of the PFC shows remarkable differences in terms of their layer distribution, proportion of each neuronal type expressing one of other receptor, and cellular level of expression. These results may be useful to better understand the complex actions of DA and antipsychotic drugs in PFC. Further work is required to examine similar aspects of other members of the D1 and D2 receptor families.
Ministry of Education and Science (SAF-2007.62378) and to N.S. SENY Fundació.
We thank Judith Ballart for skillful technical assistance. Conflict of Interest: None declared.