Dopamine (DA) exerts a strong influence on inhibition in prefrontal cortex. The main cortical interneuron subtype targeted by DA are fast-spiking γ-aminobutyric acidergic (GABAergic) cells that express the calcium-binding protein parvalbumin. D1 stimulation depolarizes these interneurons and increases excitability evoked by current injection. The present study examined whether this direct DA-dependent modulation of fast-spiking interneurons involves DARPP-32. Whole-cell patch-clamp recordings were made from fast-spiking interneurons in brain slices from DARPP-32 knockout (KO) mice, wild-type mice, and rats. Low concentrations of DA (100 nM) increased interneuron excitability via D1 receptors, protein kinase A, and cyclic adenosine 3′,5′-monophosphate in slices from both normal and DARPP-32 KO mice. Immunohistochemical staining of slices from normal animals revealed a lack of colocalization of DARPP-32 with calcium-binding proteins selective for fast-spiking interneurons, indicating that these interneurons do not express DARPP-32. Therefore, although DARPP-32 impacts cortical inhibition through a previously demonstrated D2-dependent regulation of GABAergic currents in pyramidal cells, it is not involved in the direct D1-mediated regulation of fast-spiking interneurons.
Cortical dopamine (DA) modulates glutamate and γ-aminobutyric acid (GABA) transmission, and this interaction is critical for optimal cognitive function (Sawaguchi and Goldman-Rakic 1994; Zahrt et al. 1997; Wang 1999; Wang et al. 2002; Seamans and Yang 2004). Moreover, GABA and DA modulation of prefrontal cortex (PFC) functions are altered in schizophrenia (Benes 1997; Egan and Weinberger 1997; Beasley et al. 2002; Guidotti et al. 2005). DA activates diverse signaling cascades that target numerous intracellular substrates to regulate cellular excitability (Greengard et al. 1999). Recently, we showed that the concentration of DA plays an important role in determining DA receptor subtype activation. D1 antagonists and protein kinase A (PKA) inhibitors blocked the enhancement of GABA currents that occurs with low concentrations of DA, suggesting this was due to activation of the D1–cyclic adenosine 3′,5′-monophosphate (cAMP)–PKA signaling pathway (Trantham-Davidson et al. 2004).
Traditionally, the D1–PKA cascade is associated with phosphorylation of DARPP-32 (Snyder et al. 1998; Nishi et al. 2000), which modulates the activity of a large number of receptors and ion channels through inhibition of protein phosphatase 1 (Greengard et al. 1999). However, a number of recent studies suggest that DA can affect membrane excitability via additional signaling methods. For example, DA can modulate an inwardly rectifying hyperpolarization-activated current (Ih), which regulates resting membrane potential (Wu and Hablitz 2005), an effect that appears to be independent of PKA activation (Rosenkranz and Johnston 2006). Modulation of other K+ channels in pyramidal cells involves cAMP and PKA activation but appears to be independent of downstream effector proteins (Dong and White 2003; Dong et al. 2004). Similarly, we have previously shown that low [DA] and D1 agonists can increase GABAA responses in pyramidal neurons both in wild-type (WT) and DARPP-32 knockout (KO) animals, suggesting that this phosphoprotein is not required for low [DA]/D1-mediated effects on cortical inhibition (Trantham-Davidson et al. 2004). Previous studies showed that D1 stimulation increases GABAA currents in pyramidal cells by increasing excitability of fast-spiking interneurons and thereby increasing GABA release (Gorelova et al. 2002; Kröner et al. 2007). However, it is not clear whether DARPP-32 was simply not involved in this effect or whether fast-spiking interneurons lack DARPP-32 expression altogether. Given the general importance of DARPP-32 to DA signaling and the central role of DA regulation of inhibition to large-scale PFC networks and behavior (Constantinidis and Goldman-Rakic 2002; Constantinidis et al. 2002), we aimed to test both possibilities.
The present study used in vitro patch-clamp recordings from DARPP-32 KO mice, control mice, and rats to investigate DA's effects on fast-spiking interneurons. DA increased membrane excitability of these cells in both control and KO animals, suggesting that DA affects fast-spiking interneurons independently of DARPP-32 activation. Additionally, immunohistochemical staining in the PFC of normal animals support this finding as they failed to show colocalization of DARPP-32 with the calcium-binding proteins parvalbumin (PV) and calbindin, which serve as markers of fast-spiking interneurons.
Brain Slice Preparation and Whole-Cell Patch-Clamp Recordings
DARPP-32 KO mice (Nally et al. 2003) and C57BL/6 WT mice (Rockefeller University; [Fienberg et al. 1998]; 23 KO animals and 32 WT animals; 14–120 days; animals kindly provided by Drs Paul Greengard and H.C. Hemmings), or Sprague Dawley rats (Harlan; 14–28 days of age) were deeply anesthetized with chloral hydrate (15 mg/kg). Following decapitation, brains were rapidly dissected and immersed for 1 min in cold (4 °C), oxygenated (95% O2–5% CO2) artificial cerebrospinal fluid (ACSF), consisting of (in mM) 200 sucrose, 1.9 KCl, 1.2 NaH2PO4, 33 NaHCO3, 0.5 CaCl2, 6 MgCl2, 25 glucose, and 0.4 ascorbic acid. Coronal sections (300 μm thickness) containing the prelimbic and infralimbic region of the PFC were transferred to an incubation solution consisting of (in mM) 125 NaCl, 2.5 KCl, 25 NaHCO3, 4 MgCl2, 1 CaCl2, 1.25 NaH2PO4, 10 glucose, and 0.4 ascorbic acid until use. Slices were transferred to a recording chamber and superfused with oxygenated ACSF (in mM): 125 NaCl, 2.5 KCl, 25 NaHCO3, 1.3 MgCl2, 2.0 CaCl2, 10 glucose, and 0.4 ascorbic acid. Submerged slices were perfused at a rate of 1–3 ml/min and viewed using differential interference contrast optics. Recordings were made at 33–36 °C. Thick-walled borosilicate pipettes (3–7 MΩ tip resistance) were used for whole-cell patch-clamp recordings and were filled with 125 mM K+-gluconate, 3 mM KCl, 2 MgCl2, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, and 0.1 mM ethylene glycol-bis(2-aminoethylether)-N,N,N-tetraacetic acid 2.0 mM Na-ATP, 0.5 mM Tris-GTP, 10 Na2-phosphocreatinine. In addition, 0.3% biocytin or 40 μM Alexa 594 was included in the internal solution to aid morphological recovery using either standard 3,3 diaminobenzidine histochemistry for biocytin (see below) or confocal laser-scanning microscopy, respectively. Pipettes were connected to the headstage of a Heka EPC 9/3 or axopatch 200b amplifier via an Ag/AgCl wire. Data were acquired to a PC running TIDA software (Heka, Lambrecht, Germany) or custom Labview software. An Ag/AgCl reference pellet was placed in the bath and voltage shifts were corrected using offset. Fast and slow capacitance and series resistance were corrected automatically and were optimized manually.
Categorization of Fast-Spiking Interneurons
Cells were targeted for recording based on their apparent lack of apical dendrites, and interneuron morphology was confirmed by staining for biocytin or confocal imaging of Alexa 594 (Fig. 1A). Intrinsic membrane properties and the evoked firing pattern were used to distinguish subtypes of GABAergic interneurons (Kawaguchi 1993) as well as to determine changes in neuronal excitability following application of DA or other drugs. Therefore, series of current steps (1000 ms duration, 10–300 pA at 1 Hz, or 500ms, 10-300 pA at 0.5 Hz in some experiments) were injected to evoke spike firing at various steady-state membrane potentials. Fast-spiking interneurons had short-duration action potentials and fast monophasic afterhyperpolarizations (AHPs) and showed little or no spike-frequency adaptation during large depolarizing pulses (Fig. 1B). Cells that did not fit this physiological profile or the morphological characteristics of interneurons were omitted from analysis.
DA stock solutions were made daily and the carbogen line was removed 1–2 min before introducing DA to reduce DA oxidation. Overhead and microscope lights were extinguished, and DA was bath applied for 2–3 min. In experiments that used specific antagonists to test the effects of D1-receptor blockade (SCH-23390, 5–10 μM), the inhibition of cAMP (Rp-cAMP, 100 μM), or PKA activity (H-89, 10 μM; KT5720, 10 μM), the antagonists were bath applied for 10 min before DA was introduced. In other experiments, the membrane-impermeable PKA inhibitor PKI 5-24 (20 μM) was added to the recording pipette solution. Results using PKA inhibitors were similar regardless of application method or type of inhibitor; therefore, these data were pooled into a single group (“PKA inhibitors”) for analysis (Fig. 4B).
Although sampling occurred continuously at 30-s intervals, for statistical comparisons measures of the intrinsic membrane excitability (number of evoked spikes, input resistance, action potential threshold and amplitude, and amplitude of the fast AHP) were compared at 2 time intervals (0–5 min before and 10–15 min after drug application) and the measurements were averaged for each period. For analysis, a current step was chosen for each neuron that was the minimal current necessary to produce stable repetitive firing under control conditions. Averages for each interval were compared using 2-tailed, paired t-tests (differences of alpha ≤ 0.05 were considered significant). All data are presented as means ± standard error of the mean.
Following recording, slices were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB) for 2 h at room temperature. Slices were then transferred to 0.05 M Tris buffer containing 1% Triton X-100 and stored overnight at 4 °C. Activity of endogenous peroxidases was neutralized by incubation with hydrogen peroxide in methanol (1:200). After washing, slices were incubated for 2 h at room temperature in the avidin–biotin complex (Vector Laboratories, Burlingame, CA) and biocytin-filled neurons were visualized by standard 3,3′-diaminobenzidine histochemistry, resulting in a dark brown staining product. Only recordings from cells that fit both the electrophysiological criteria outlined above and that showed basic morphological criteria of interneurons (lack of an apical dendrite, smooth aspiny dendrites, and dense localized axonal projections) were used for analysis of DA effects. For double-labeling experiments of either PV or calbindin with DARPP-32, 4 adult rats (>250 g) were perfused with 4% PFA and 50 μm thick slices were prepared on a freezing microtome. Sections were preincubated in 1% (w/v) bovine serum albumin (Sigma, St. Louis, MO) and 20% normal goat serum in PB. For double labeling of DARPP-32 and PV, free-floating sections were incubated overnight at 4 °C in mouse anti-DARPP-32 (C24-6a; kindly provided by H.C. Hemmings Jr [Hemmings and Greengard 1986]; working dilution 1:1000) and rabbit anti-PV (Swant, Bellizona, Switzerland [Kagi et al. 1987]; working dilution 1:1000) in PB (pH 7.4) and 0.3% (v/v) Triton X-100 (Sigma). The following steps were carried out at room temperature, separated by 3 washes in PB of 10 min each. Slices were incubated in a mixture of fluorescein isothiocyanate (FITC)-conjugated donkey anti-mouse and tetramethyl rhodamine iso-thiocyanate donkey anti-rabbit (both Jackson; working dilutions 1:500) in 0.3% Triton X PB for 1 h. Treatment of the sections was identical for double labeling of DARPP-32 and calbindin, except that we used the following combinations of antibodies: monoclonal mouse anti-calbindin (Swant, [Celio et al. 1990]; 1:1,000) and rabbit anti-DARPP-32 (Chemicon, Temecula, CA; 1:1,000), followed by rhodamine red-X anti-mouse and FITC donkey anti-rabbit (both Jackson; both 1:500). Slices were mounted and coverslipped using Vectashield (Vector Labs, Burlingame, CA). Fluorescence images were acquired using a 60× oil immersion, NA 0.9 objective, or 40× water immersion, NA 0.8 objective on a laser-scanning confocal microscope (Zeiss, Peabody, MA, LSM 510). Fluorescent dyes were excited at 488 and 543 nm, respectively, and emission signals were band-pass filtered between 500–550 nm for FITC and long-pass filtered at 560 nm for rhodamine conjugates.
Properties of Fast-Spiking Interneurons from Mouse Medial PFC
Figure 1 shows the morphological and physiological characteristics of typical fast-spiking interneurons recorded in the mouse medial PFC. Fast-spiking interneurons from both wild-type (n = 43) and DARPP-32 KO (n = 28) animals were identified by their lack of significant adaptation in firing during prolonged depolarizing current pulses (adaptation ratios were 1.2 ± 0.09 for WT and 1.1 ± 0.1 for DARPP-32 KO animals, respectively). Fast-spiking cells generated spikes of short duration (WT, 0.6 ± 0.08 ms; KO, 0.53 ± 0.18 ms), which were followed by fast, monophasic AHPs of large amplitude (WT, 13.9 ± 1.5 mV; KO, 14.8 ± 1.8 mV). Similarly, other basic electrophysiological properties such as input resistance (WT, 180.1 ± 11.5 MΩ; KO 197.0 ± 13.6 MΩ), action potential threshold (WT, −38.5 ± 2.6 mV; KO, −41.0 ± 1.6 mV), or action potential amplitude (WT, 46.3 ± 3.9 mV; KO, 43.7 ± 4.5 mV) did not significantly differ between WT and DARPP-32 KO mice.
DA and D1 Receptor Modulation of Evoked Firing in Fast-Spiking Interneurons
In both WT and DARPP-32 KO mice, we assessed the effects of bath application of DA and/or D1 receptor-specific compounds on intrinsic membrane excitability and evoked spike firing in response to a series of somatic current injections. In control recordings under identical conditions (i.e., repeated somatic current injections every 5 min) but without application of DAergic drugs, neither the number of evoked spikes (WT, 24.3 ± 3.3 baseline, 24.0 ± 1.9, 10–15 min after mock drug application, n = 6; KO, 15.4 ± 3.8 baseline, 15.8 ± 3.5, 10–15 min after mock drug application, n = 3) nor the input resistance (WT, 139 ± 14.7 MΩ baseline, 143.1 ± 19.1 MΩ 10–15 min after mock drug application; KO 140.0 ± 6.6 MΩ baseline, 146.4 ± 11.1 MΩ 10–15 min after mock drug application, n = 3; not shown) changed as a measure of time. During these mock drug applications, the oxygen line was removed from the recording solution to mimic conditions during DA experiments (see Methods). As these and previously published (Trantham-Davidson et al. 2004) results suggest, the brief oxygen deprivation and potential minor changes in pH do not result in detectable changes in cell excitability.
Experiments focused on the effects of a low dose of DA (100 nM) because we have previously shown that this dose increases inhibitory postsynaptic current (IPSC) amplitude in pyramidal cells via a D1 receptor–mediated mechanism, whereas higher doses decrease IPSC amplitude through activation of a D2-associated signaling pathway (Trantham-Davidson et al. 2004). Consistent with these previous findings, bath application of 100 nM DA increased the firing rate of fast-spiking interneurons. This increase was observed in both WT and DARPP-32 KO animals, suggesting that low DA activates a signaling pathway that is independent of DARPP-32 (Figs 2 and 3).
This change in evoked firing was long lasting and rarely reversed during the duration of the experiments (Fig. 3A). In the majority of cells, increases in spike firing were accompanied by a reduction in apparent outward rectification at membrane potentials just subthreshold to spike initiation (Fig. 3B), which indicates DA modulation of active conductances at depolarized potentials, as previously shown in fast-spiking cells and pyramidal neurons in the PFC (Yang and Seamans 1996; Gorelova et al. 2002; Kröner et al. 2007). Additionally, in cells from both WT and KO animals, a small membrane depolarization (2–5 mV) was observed upon DA application that was compensated with direct current injection in current clamp to maintain the membrane potential at −70 mV before changes in excitability were measured (not shown). This is in accord with previous studies, which showed that D1 stimulation depolarizes target cells (Shi et al. 1997; Gorelova and Yang 2000; Dong and White 2003; Dong et al. 2004; Tseng and O'Donnell 2004; Kröner et al. 2007). Consistent with these findings, in both WT and KO animals the effects of DA on evoked firing and on membrane depolarization were blocked by preapplication of the D1 receptor antagonist SCH-23390 (Fig. 2B), and they were mimicked by application of the D1 receptor agonist SKF-38393 (Fig. 2C). Similarly, the full D1 receptor agonist SKF-81297 (5 μM; n = 4) also significantly increased the number of evoked spikes (baseline, 9.6 ± 2.6; SKF-81297, 17.9 ± 1.6; P < 0.05, n = 4) in WT animals (not shown).
In most instances, the effects of 100 nM DA or the selective D1 receptor agonists on evoked firing were independent of changes in input resistance (Fig. 2A3–C3), action potential amplitude, spike threshold, or amplitude of the fast AHP (data not shown). An exception to this occurred in KO animals, where changes in the evoked number of spikes following DA application were accompanied by a significant increase in input resistance (Fig. 2A3). However, the functional significance of this effect is unclear, particularly because it was not evident following D1 receptor activation with SKF-38393, which otherwise mimicked the effects of DA in both WT and KO animals (Fig. 2C).
In an effort to test the contribution of the D1-adenyl cyclase–cAMP–PKA pathway to the effects of low DA on fast-spiking interneurons, subsequent experiments used either a cAMP inhibitor (Rp-cAMPS, 10 μM) or a variety of PKA inhibitors (H-89, 10 μM; KT5720, 20 μM, both bath applied; or PKI 5-24, 10 μM in the pipette solution). As shown in Figure 4, inhibition of either cAMP or PKA prevented the effects of 100 nM DA on interneuron excitability in both WT and DARPP-32 KO mice. Results for all PKA inhibitors used were similar and so the data were pooled (Fig. 4B). Qualitatively identical results were also obtained in recordings from rat PFC interneurons using the PKA inhibitor H-89 (n = 4; data not shown).
Taken together, these data suggest that in fast-spiking interneurons, low concentrations of DA (100 nM) stimulate D1 receptors to increase excitability via cAMP and PKA activity but not phosphorylation of DARPP-32.
Lack of Colocalization of DARPP-32 with Markers of Fast-Spiking Interneurons
Given that the physiological effect of D1 receptor stimulation on fast-spiking interneurons appeared to be independent of DARPP-32, we asked whether fast-spiking interneurons in WT animals actually possess this protein. To this end, we used confocal microscopy and double immunofluorescent labeling of DARPP-32 with one of 2 different calcium-binding proteins that serve as molecular markers of fast-spiking interneurons. Specifically, fast-spiking interneurons express the calcium-binding protein PV (Kawaguchi 1993; Zaitsev et al. 2005) and possibly also calbindin (Kawaguchi 1993; Cauli et al. 1997; Zaitsev et al. 2005). For these experiments, we used tissue from rat PFC because we had found qualitatively similar physiological effects in interneuron recordings from young adult rats and mice (see above); this in turn enabled us to use well characterized primary antibodies against PV and DARPP-32 (see Methods), which were the main focus of these experiments. Figure 5 shows double immunolabeling for PV and DARPP-32 (A,B) and calbindin and DARPP-32 (C), respectively, in rat PFC slices. Although many cells show robust labeling for one or the other marker at the level of the soma and proximal dendrites, there is no overlap of the 2 signals. We examined 615 neurons immunopositive for PV and more than 2000 neurons labeled for DARPP-32 throughout the entire rostral–caudal extent of the prelimbic and infralimbic cortex in slices from 4 animals. We found no evidence of colocalization, suggesting that PV-positive, fast-spiking interneurons in the PFC do not contain DARPP-32. Similarly, no calbindin-positive cells were found that showed double labeling for DARPP-32.
We show that a low concentration of DA increases excitability of fast-spiking interneurons in PFC via activation of D1 receptors, cAMP, and PKA, but independently of DARPP-32 activation. Immunohistochemical labeling revealed no evidence of double labeling of PV, a marker for fast-spiking interneurons, and DARPP-32. Because these neurons are primary targets of cortical DA innervation and D1 modulation of GABA currents occurs presynaptically (Gorelova et al. 2002), our results suggest that D1 modulation of inhibition in PFC does not require phosphorylation of DARPP-32.
Previous results demonstrated that fast-spiking interneurons are the main target of DA modulation in the rat prelimbic cortex and primate dorsolateral PFC. DA modulates 3 distinct K+ currents in fast-spiking interneurons; a “leak” current, an inward rectifier, and a slowly inactivating K+ current (ID). DA-mediated attenuation of the K+ leak current slightly depolarizes resting membrane potential (Gorelova et al. 2002; Kröner et al. 2007). In our data, we similarly found evidence for a slow depolarization following DA application, which was evident as an increase in the negative holding current required to hold the cells at −70 mV in current clamp. The most apparent effect of DA receptor activation by 100 nM DA was the significant increase in the firing rate of fast-spiking interneurons. This effect was previously shown to result from suppression of the slowly inactivating K+ current, which activates at more depolarized potentials and regulates repeated firing. Our present observation that DA apparently reduced outward rectification in a membrane range just below action potential threshold (albeit without significantly altering spike threshold itself) similarly indicates a modulation of ID. Furthermore, consistent with data presented here, these effects on outward rectification and evoked spike firing were previously shown to require D1 receptor stimulation (Yang and Seamans 1996; Gorelova et al. 2002; Kröner et al. 2005, 2007).
In dissociated pyramidal neurons from the PFC, basal PKA activity and elevations of cAMP also decreased K+ currents, resulting in increased excitability (Dong and White 2003; Dong et al. 2004). In these studies, the PKA-dependent modulation was also observed in outside-out patches, indicating a direct modulation that did not require downstream signaling molecules such as DARPP-32. Members of the inwardly rectifying potassium family of K+ channels contain nucleotide-binding sites (Nichols and Lopatin 1997) and phosphorylation sites for PKA (Rudy and McBain 2001; Lien et al. 2002; Tanemoto et al. 2002; Vogalis et al. 2003). Stimulation of the D1–cAMP–PKA pathway could therefore directly phosphorylate these channels. If a similar situation holds for fast-spiking interneurons, this would explain how low concentrations of DA via D1 receptor activation are able to increase the excitability of fast-spiking interneurons in the absence of DARPP-32.
Recent reports also suggest other novel mechanisms by which DA and D1 receptor activation can modulate neuronal excitability without involvement of PKA and/or DARPP-32: In the rat, DA affects the excitability of a variety of neurons, including layer I cortical interneurons from young (<11d) animals (Wu and Hablitz 2005), via modulation of an inwardly rectifying hyperpolarization-activated current (Ih) (Jiang et al. 1993; Vargas and Lucero 1999; Rosenkranz and Johnston 2006). This modulation has been shown to be independent of PKA activation (Rosenkranz and Johnston 2006) and in the cortex may require cooperative activation of D1 and D2 receptors (Wu and Hablitz 2005). In all these cells, the presence of hyperpolarization-activated, nonselective cation channels is indicated by a prominent time-dependent “sag” in the cells' response to hyperpolarizing current injection (Wu and Hablitz 2005; but see Aponte et al. 2006; Rosenkranz and Johnston 2006). However, in fast-spiking cortical interneurons in layers 2–5 from rat and monkey PFC, these sags are rarely seen, and instead, cells often show prominent fast inward and outward rectification in both the hyperpolarized and depolarized voltage range (c.f. Figs 1 and 3; Kawaguchi 1995; Gorelova et al. 2002; Zaitsev et al. 2005; Krimer et al. 2005; Kröner et al. 2007). Thus, whether DA modulation of Ih could contribute to the effects observed in our study appears questionable.
Another noteworthy observation is the fact that the effects of DA and D1 receptor stimulation on the excitability of interneurons in WT and KO animals were slow to develop and persistent in time. This is in line with many previous reports that used bath application of DA to study changes in membrane excitability of principal cells and interneurons (Yang and Seamans 1996; Henze et al. 2000; Gulledge and Jaffe 2001; Gorelova et al. 2002; Kröner et al. 2005, 2007) or IPSCs as an indicator of changes in action potential-dependent GABAergic transmission (Seamans et al. 2001; Trantham-Davidson et al. 2004; Kröner et al. 2007). On the other hand, when using fast pressure application of DA in rats, depolarization of fast-spiking interneurons can occur for less than 5 s after which the membrane potential returns to the baseline (Lapish et al. 2006; Kröner and Seamans, unpublished data). Similarly, in vivo the synaptic release of DA in the PFC following ventral tegmental area stimulation has been shown to last approximately 4 s, but nevertheless was able to influence the firing of intracellularly recorded PFC neurons for tens of minutes (Lavin et al. 2005). Taken together, these observations indicate that in the present experiments the rate of bath exchange was not a limiting factor and that although some effects of DA on fast-spiking interneurons can be fast and transient, others are very protracted and difficult to reverse both in vitro and under certain conditions in vivo.
Our physiological data from DARPP-32 KO animals that suggest that the D1 modulation of interneurons can be independent of DARPP-32 was further corroborated by our histochemical analysis in normal rats. We show here that cells, which express the Ca2+-binding proteins PV or calbindin do not coexpress DARPP-32. In several previous studies, PV was shown to label the vast majority of fast-spiking interneurons (Cauli et al. 1997; Kawaguchi and Kondo 2002; Zaitsev et al. 2005). Calbindin labels a distinct, partially overlapping group of interneurons, some of which also have been shown to display fast-spiking firing characteristics (Cauli et al. 1997; Kawaguchi and Kubota 1997). Thus, the absence of DARPP-32 from fast-spiking neurons in the PFC indicates that DARPP-32 in the PFC shows cell-specific expression and is not ubiquitous in all neurons possessing DA receptors. Similarly in the striatum, DARPP-32 is expressed in medium spiny neurons but not GABAergic interneurons or large aspiny cholinergic cells (Ouimet and Greengard 1990; Anderson and Reiner 1991).
DA, GABA, and DARPP-32 all have been implicated in the pathophysiology of schizophrenia: In the brains of schizophrenic individuals, both the synthesis and reuptake of GABA are reduced in a subset of PFC interneurons, and the subpopulation of PV+ neurons appears to be particularly vulnerable (Lewis et al. 2005). PV+ cells show differences in the expression of several specific genes compared with normal subjects (i.e., GAD-67, GAT1, PV, and DNMT-1) (Benes 1997; Veldic et al. 2005). Likewise, alterations in DA neurotransmission also have long been discussed in the etiology of the disease (Winterer and Weinberger 2004). Remarkably, in postmortem tissue from the brains of schizophrenic patients, the DA innervation of interneurons may be up to 3-times higher than for pyramidal neurons (Benes 1997). Thus, alterations in both DA and its main inhibitory target, PV+ interneurons, seem to be critically involved in the pathology of schizophrenia.
Changes in DARPP-32–mediated functions may also be involved in schizophrenia because levels of this protein are decreased in postmortem brain tissue from these patients (Albert et al. 2002). Additionally, psychotomimetic drugs such as D-amphetamine and phencyclidine, which exacerbate psychosis in schizophrenics and induce stereotypical movements in normal rats, do not have these effects when given to animals with point mutations at specific phosphorylation sites of the DARPP-32 protein (Svenningsson et al. 2003).
In spite of the individual importance of DA, PV+ interneurons, and DARPP-32 to normal and abnormal PFC function, the present data suggest that these factors are not necessarily correlated at a cellular level. Although it is clear that DA modulates both pyramidal and interneurons in the PFC, the involvement of DARPP-32 in DA's effects appears to be restricted to the modulation of pyramidal neurons. DARPP-32 modulates N-methyl-D-aspartic acid (NMDA) currents in striatum via D1 receptors (Flores-Hernandez et al. 2002) and GABA currents via D2 receptors in PFC (Trantham-Davidson et al. 2004). These currents are of paramount importance to our theorized global function of DA in the PFC (Yang et al. 1999; Durstewitz et al. 2000; Durstewitz and Seamans 2002; Seamans and Yang 2004). Taken together, reduced NMDA currents and abnormalities in the activity of the GABAergic network are expected to prevent the initiation, maintenance, and tuning of active network states thought to encode working memory information (Goldman-Rakic 1995; Seamans and Yang 2004; Winterer and Weinberger 2004). Therefore, although independent at the cellular level, alterations in PV+ interneurons and DARPP-32 would together have deleterious effects on working memory processes and cognition.
National Alliance for Research on Schizophrenia and Depression (C06 RR015455); the National Institutes of Health (R01MH064569-02, RO1MH065924-02); the Extramural Research Facilities Program of the National Center for Research Resources.
The authors are grateful to Dr Paul Greengard for his generous gift of wild-type and DARPP-32 knockout animals and to Dr Hugh C. Hemmings for his donation of the DARPP-32 antibody. Conflict of Interest: None declared.