Restoration of Glutamatergic Transmission by Dopamine D4 Receptors in Stressed Animals*

Background: The dopamine D4 receptors in prefrontal cortex (PFC) play a key role in mental health and disorders. Results: D4 activation caused a bi-directional, homeostatic regulation of glutamatergic responses in rats exposed to acute or chronic stress. Conclusion: The altered synaptic excitation in stress conditions is restored by D4 signaling in PFC. Significance: It provides a potential mechanism for the role of D4 in stress-related neuropsychiatric disorders. The prefrontal cortex (PFC), a key brain region for cognitive and emotional processes, is highly regulated by dopaminergic inputs. The dopamine D4 receptor, which is enriched in PFC, has been implicated in mental disorders, such as attention deficit-hyperactivity disorder and schizophrenia. Recently we have found homeostatic regulation of AMPA receptor-mediated synaptic transmission in PFC pyramidal neurons by the D4 receptor, providing a potential mechanism for D4 in stabilizing cortical excitability. Because stress is tightly linked to adaptive and maladaptive changes associated with mental health and disorders, we examined the synaptic actions of D4 in stressed rats. We found that neural excitability was elevated by acute stress and dampened by repeated stress. D4 activation produced a potent reduction of excitatory transmission in acutely stressed animals and a marked increase of excitatory transmission in repeatedly stressed animals. These effects of D4 targeted GluA2-lacking AMPA receptors and relied on the bi-directional regulation of calcium/calmodulin kinase II activity. The restoration of PFC glutamatergic transmission in stress conditions may enable D4 receptors to serve as a synaptic stabilizer in normal and pathological conditions.

The dopaminergic system in prefrontal cortex (PFC) 2 plays a key role in regulating high level cognitive functions, such as working memory (1,2). Dopamine D 4 receptors, which are largely restricted to PFC neurons (3,4), are critically involved in PFC functions both under normal condition and in many neuropsychiatric disorders (5,6). For example, attention deficit hyperactivity disorder and the responsiveness to its treatment have been associated with a D 4 gene polymorphism that weak-ens D 4 receptor function (7)(8)(9)(10). The highly effective antipsychotic drug used in schizophrenia treatment, clozapine, has a high affinity to D 4 receptors (11,12). Preclinical studies also indicate that D 4 receptor antagonists alleviate stress-induced working memory problems in monkeys (13) and ameliorate cognitive deficits caused by psychotomimetic drugs (14). Mice lacking D 4 receptors exhibit supersensitivity to psychomotor stimulants, reduced exploration of novel stimuli, cortical hyperexcitability, and juvenile hyperactivity and impulsive behaviors associated with attention deficit hyperactivity disorder (15)(16)(17)(18).
To understand the role of D 4 in mental health and disorders, it is important to elucidate the molecular and cellular mechanisms underlying the impact of D 4 on cortical excitability and working memory. It has been proposed that glutamate receptor-mediated synaptic transmission that controls PFC neuronal activity is crucial for working memory (19). Dysfunction of glutamatergic transmission is considered as the core feature and fundamental pathology of mental disorders (20 -22). Recently we have demonstrated that D 4 stimulation causes a profound depression of AMPA receptor (AMPAR) responses in PFC pyramidal neurons when their activity is elevated in vitro and causes a marked potentiation of AMPAR responses when their activity is dampened in vitro (23,24). It suggests that D 4 receptors may use the homeostatic control of glutamatergic transmission to stabilize the activity of PFC circuits.
The PFC is one of the primary targets of stress hormones (25). Stress has profound and divergent effects on the body and the brain. Acute stress is essential for adaptation and maintenance of homeostasis, whereas chronic stress often exacerbates deficiencies in emotional and cognitive processes associated with psychiatric disorders, such as depression, anxiety, and posttraumatic stress disorder (25,26). In this study, we sought to determine whether animals exposed to acute or repeated behavioral stress have altered neuronal activity in vivo and whether D 4 exerts an activity-dependent regulation of glutamatergic transmission in PFC pyramidal neurons from stressed animals. Knowledge gained from this study should help us understand the role of D 4 in the adaptive and maladaptive changes linked to stress.

EXPERIMENTAL PROCEDURES
Stress Paradigm-All experiments were conducted in accordance to the Institutional Animal Care and Use Committee (IACUC) of the State University of New York at Buffalo. Juvenile (3-4 weeks old) Sprague-Dawley male rats were used in this study. For acute stress, rats were restrained in air-accessible cylinders for 2 h (10:00 a.m. to 12:00 p.m.) as described previously (27,28). For repeated stress, rats were restrained daily (2 h) for 5-7 days.
Electrophysiological Recording in Slices-PFC slice preparation procedures were similar to what was described previously (29). In brief, animals were anesthetized by inhaling 2-bromo-2-chloro-1,1,1-trifluoroethane (1 ml/100 g; Sigma) and decapitated. Brains were quickly removed and sliced (300 m) with a Leica VP1000S Vibrotome while bathed in a HEPES-buffered salt solution. Slices were then incubated for 1-5 h at room temperature (22-24°C)  Standard voltage clamp recording techniques were used to measure AMPAR-EPSC in layer V PFC pyramidal neurons as described before (23,28). Neurons were visualized with a ϫ40 water-immersion lens and illuminated with near infrared (IR) light, and the image was captured with an IR-sensitive CCD camera. Recordings were obtained using a Multiclamp 700A amplifier (Axon Instruments). Tight seals (2-10 gigohms) were obtained by applying negative pressure. The membrane was disrupted with additional suction, and the whole-cell configuration was obtained. Neurons were held at Ϫ70 mV, and EPSCs were evoked by stimulating the neighboring neurons with a bipolar tungsten electrode (FHC, Inc.). The internal solution contained 130 mM cesium methanesulfonate, 10 mM CsCl, 4 mM NaCl, 1 mM MgCl 2 , 10 mM HEPES, 5 mM EGTA, 2.2 mM QX-314, 12 mM phosphocreatine, 5 mM MgATP, 0.5 mM Na 2 GTP, 0.1 mM leupeptin, pH 7.2-7.3, 265-270 mosM. To record miniature EPSCs (mEPSCs) in the slice, 1 mM MgCl 2containing ACSF (tetrodotoxin added) was used.
Whole-cell current clamp techniques were used to measure action potential firing (30). During the recording of spontaneous firing, PFC slices were bathed in a modified ACSF with reduced Mg 2ϩ (0.5 mM) and slightly elevated K ϩ (3.5 mM) at room temperature. This modified ACSF is more similar to in vivo rodent CSF than standard ACSF (31). The internal solution contained 100 mM potassium gluconate, 20 mM KCl, 10 mM HEPES, 4 mM MgATP, 0.3 mM NaGTP, 10 mM phosphocreatine, and 0.1 mM leupeptin, pH 7.2-7.3, 265-270 mosM. A small depolarizing current was applied to adjust the interspike potential to Ϫ60 mV.
Dopamine receptor ligands PD168077 maleate (Tocris) and the second messenger reagents KN-93, KN-92, purified active CaMKII␣ protein (74 kDa; Abcam ab60899) and CaM proteins (Calbiochem) were made up as concentrated stocks in water or dimethyl sulfoxide and stored at Ϫ20°C or Ϫ80°C. Stocks were thawed and diluted immediately before use. Reagents were dialyzed into neurons through the patch electrode for 20 min before electrophysiological recordings were started. Data anal-yses were performed with the Clampfit software (Axon Instruments). Experiments with two groups were analyzed statistically using Student's t tests. Experiments with more than two groups were subjected to one-way ANOVA followed by post hoc Tukey tests. Miniature synaptic currents were analyzed with Mini Analysis Program (Synaptosoft, Leonia, NJ). Statistical comparisons of mEPSCs were made using the Kolmogorov-Smirnov test.
Western Blotting-After treatment, slices were homogenized in boiling 1% SDS, followed by centrifugation (13,000 ϫ g, 20 min). The supernatant fractions were subjected to 7.5% SDSpolyacrylamide gels and transferred to nitrocellulose membranes. The blots were blocked with 5% nonfat dry milk for 1 h at room temperature followed by incubation with various primary antibodies of CaMKII (1:1000; Santa Cruz Biotechnology sc-9035) and Thr(P) 286 -CaMKII (1:1000; Santa Cruz Biotechnology sc-12886). After incubation with horseradish peroxidase-conjugated secondary antibodies (Sigma-Aldrich), the blots were exposed to the enhanced chemiluminescence substrate (Amersham Biosciences). Quantification was obtained from densitometric measurements of immunoreactive bands on film using ImageJ software.

Neuronal Excitability, Which Is Elevated by Acute Stress and Dampened by Repeated Stress, Is Restored by D 4 Activation-To
compare overall excitability of PFC circuits in control versus stressed animals, we measured the spontaneous action potential firing, which helps to reveal the circuit excitability changes induced by altered synaptic drive onto pyramidal neurons (32). As shown in Fig. 1A, compared with neurons from nonstressed (NS) control rats, the firing rate was significantly increased in neurons from acutely stressed (AS) animals (NS: 1.14 Ϯ 0.25 Hz, n ϭ 9; AS: 2.04 Ϯ 0.31 Hz, n ϭ 10, p Ͻ 0.01, t test). On the other hand, a significant reduction in the firing rate was observed in repeatedly stressed (RS) animals ( Fig. 1B, NS: 1.25 Ϯ 0.41 Hz, n ϭ 8; RS: 0.37 Ϯ 0.10 Hz, n ϭ 11, p Ͻ 0.01, t test). These data indicate that the excitability of PFC pyramidal neurons is elevated by acute stress and dampened by chronic stress. Thus, stress provides a physiological condition that has bi-directional changes in neuronal activity in vivo.
We then examined the impact of D 4 receptor activation on the excitability of PFC pyramidal neurons from stressed animals. As shown in Fig. 1, C and D, application of the D 4 receptor agonist PD168077 (40 M) produced a significant reducing effect on the firing rate in acutely stressed animals (control: 2.14 Ϯ 0.04 Hz, PD: 1.33 Ϯ 0.03 Hz, n ϭ 7). On the other hand, PD168077 significantly enhanced the firing rate in repeatedly stressed animals (control: 0.39 Ϯ 0.02 Hz, PD: 0.97 Ϯ 0.04 Hz, n ϭ 5). Thus, the PFC neuronal excitability has been restored by D 4 in stress conditions. D 4 Restores the Excitatory Synaptic Transmission in Stressed Animals-Because D 4 stimulation induces an activity-dependent regulation of glutamatergic transmission in vitro (23), we next examined the effect of D 4 in PFC neurons from animals whose neural activity has been perturbed by in vivo stressors. As shown in Fig. 2, A-C, the basal evoked AMPAR-EPSC amplitude was markedly increased by acute stress (NS: 109.   PD168077 significantly enhanced AMPAR-EPSC in repeatedly stressed animals (53.6 Ϯ 6.5% enhancement, n ϭ 14, p Ͻ 0.001, ANOVA). Thus, the excitatory synaptic strength has been brought back to the control level by D 4 in stress conditions (ASϩPD: 120.8 Ϯ 8.3 pA, n ϭ 16; RSϩPD: 123.6 Ϯ 7.3 pA, n ϭ 14).
We next measured AMPAR-mediated mEPSC, a response from quantal release of single glutamate vesicles. Acute stress caused a significant increase of the mEPSC amplitude (NS: 12.1 Ϯ 0.6 pA; AS: 17.2 Ϯ 2.0 pA, p Ͻ 0.001, ANOVA), which was recovered to the nonstressed level by PD168077 (10.9 Ϯ 0.3 pA, 2.1 Ϯ 0.2 Hz, n ϭ 8, Fig. 2, D and E). The mEPSC frequency was not significantly changed by acute stress, consistent with our previous report (27). On the other hand, the mEPSC amplitude and frequency were significantly reduced in animals exposed to repeated stress (9.5 Ϯ 0.3 pA, 1.5 Ϯ 0.3 Hz, n ϭ 7), and PD168077 restored mEPSC similar to that observed in nonstressed animals (14 Ϯ 0.5 pA, 2.8 Ϯ 0.3 Hz, n ϭ 7, Fig. 2, F and G). Taken together, these results suggest that D 4 activation provides a homeostatic regulation of excitatory synaptic transmission in stressed animals.
The Homeostatic Effect of D 4 Targets Mainly GluA2-lacking AMPA Receptors-Because GluA2-containing and GluA2lacking AMPAR channels have distinct channel conductance, open probability, Ca 2ϩ permeability and rectification, we examined whether D 4 receptors differentially affect AMPAR subunits in stress conditions. It has been found that GluA2lacking AMPARs have prominent rectification (33), and changes in GluA2-lacking AMPARs alter the inward rectification due to voltage-dependent blockade of intracellular polyamine (34). Thus, rectification index (RI) of AMPAR responses (ratio of the current amplitude at Ϫ60 mV to that at ϩ40 mV) was measured with a spermine (100 M)-containing intracellular solution (35). As shown in Fig. 3A, the RI in acutely stressed animals was much bigger than that in nonstressed control animals (NS: 2.3 Ϯ 0.34, n ϭ 7; AS: 4.0 Ϯ 0.39, n ϭ 11; p Ͻ 0.001, ANOVA), suggesting that acute stress may predominantly increase GluA2-lacking AMPARs at the synapse. Application of PD168077 (40 M) reduced the RI back to the control level (2.4 Ϯ 0.2, n ϭ 11), indicating that D 4 activation removes the synaptic GluA2-lacking AMPARs that are previously delivered by acute stress. On the other hand, in repeatedly stressed animals, the RI was unchanged (NS: 2.1 Ϯ 0.2, n ϭ 9; RS: 2.2 Ϯ 0.2, n ϭ 9, p Ͼ 0.05, ANOVA) but was markedly increased by PD168077 (4.1 Ϯ 0.6, n ϭ 8), suggesting that D 4 increases the synaptic recruitment of GluA2-lacking AMPARs.
To test further whether the change in AMPAR trafficking is involved in D 4 regulation of synaptic transmission at stress conditions, we carried out surface biotinylation assays. As shown in Fig. 3, G and H, PD168077 caused a marked reduction of surface GluA1 expression in PFC slices from acutely stressed animals, but significantly increased the level of surface GluA1 in PFC slices from repeatedly stressed animals (NS: Ϫ29.6 Ϯ 2.6%, n ϭ 11; AS: Ϫ55.9 Ϯ 7.0%, n ϭ 7; RS: 46.9 Ϯ 11.2%, n ϭ 10; p Ͻ 0.01, ANOVA). The biochemical evidence is consistent with what was observed in electrophysiological recordings.
CaMKII Is Involved in the D 4 Regulation of AMPAR Transmission in Stressed Animals-Our previous studies have shown that D 4 exerts an activity-dependent bi-directional regulation of CaMKII (38), which is critical for the D 4 -induced homeostatic regulation of AMPARs in vitro (23,24). Thus, we tested the role of CaMKII in the D 4 actions in stressed animals. As shown in Fig. 4, A and B, in PFC slices from nonstressed animals, application of PD168077 (40 M, 10 min) reduced the level of activated (Thr 286 -phosphorylated) CaMKII (32.1 Ϯ 7.1% reduction, n ϭ 5). In acutely stressed animals, the basal level of activated CaMKII was elevated (1.59 Ϯ 0.19-fold of control, n ϭ 5), and the reducing effect of PD168077 was significantly augmented (71.4 Ϯ 7.3% reduction, n ϭ 5). Conversely, in repeatedly stressed animals, the basal level of activated CaMKII was unchanged (1.12 Ϯ 0.12-fold of control, n ϭ 6), whereas PD168077 significantly increased the level of activated CaMKII (61.4 Ϯ 10.5% increase, n ϭ 6). This bi-directional regulation of CaMKII activity by D 4 receptors provides a basis for the dual effects of D 4 on AMPAR transmission in PFC pyramidal neurons from acutely versus chronically stressed animals.
To examine whether the D 4 reduction of AMPAR-EPSC in acutely stressed animals is through suppression of CaMKII, we dialyzed neurons with an EGTA-free internal solution containing purified active CaMKII␣ protein (0.6 g/ml), calmodulin (30 g/ml) and CaCl 2 (0.3 mM). The active CaMKII protein itself did not significantly alter the basal EPSC (Ϫ7.8 Ϯ 1.7%, n ϭ 6), which is likely due to the high level of CaMKII in these cells. As shown in Fig. 5A, the reducing effect of PD168077 on AMPAR-EPSC was largely blocked by intracellular infusion of the active CaMKII protein (6.2 Ϯ 2.1% reduction, n ϭ 9, Fig.  5C), but not the heat-inactivated CaMKII protein (46.4 Ϯ 3.9% reduction, n ϭ 11; p Ͻ 0.001, t test, Fig. 5C).
To examine whether the D 4 enhancement of AMPAR-EPSC in repeatedly stressed animals is through stimulation of CaMKII, we dialyzed the CaMKII inhibitor, KN-93 (20 M). KN-93 itself did not significantly alter the basal EPSC (Ϫ7 Ϯ 1.4%, n ϭ 7), which is likely due to the low level of CaMKII in these cells. SEPTEMBER 6, 2013 • VOLUME 288 • NUMBER 36
Finally, we examined the role of CaMKII in D 4 regulation of neuronal excitability in stressed animals. As shown in Fig. 6, in animals exposed to acute stress, the reducing effect of PD168077 on spontaneous action potential firing was largely Taken together, these results suggest that D 4 restores glutamatergic transmission in stress conditions through bi-directional regulation of CaMKII activity, which contributes the dual regulation of neuronal excitability by D 4 receptors.

DISCUSSION
Stress serves as a key controller for neuronal responses that underlie behavioral adaptation, as well as maladaptive changes that lead to cognitive and emotional disturbances in stress-related mental disorders (26). Corticosterone, the major stress hormone, has been found to exert a complex effect on PFC excitatory synapses and PFC-mediated behaviors (25,39). Acute stress induces synaptic potentiation by increasing surface delivery of AMPARs and NMDARs via glucocorticoid/ SGK/Rab4 signaling, resulting in enhanced working memory performance (27,28,40). Conversely, chronic stress induces dendritic shortening, spine loss, and impairment in cognitive flexibility and perceptual attention (41)(42)(43). Our recent study has found that the detrimental effect of repeated stress on cognition is causally linked to the ubiquitin/proteasome-mediated degradation of AMPAR and NMDAR subunits and the suppression of glutamatergic transmission in PFC (44). In addition, the alteration of RNA editing of AMPAR subunits (45,46) is potentially another mechanism underlying the effects of stress.
In this study, we have found that the spontaneous action potential firing is bi-directionally changed by acute versus repeated stress, suggesting that the synaptic drive onto PFC pyramidal neurons is altered by stress, leading to the change in PFC circuit excitability. Activation of D 4 reduces the potentiated AMPAR responses in acutely stressed animals and enhances the depressed AMPAR responses in repeatedly stressed animals. The restoration of synaptic strength to the control level by D 4 in stress conditions supports the notion that D 4 serves as a homeostatic synaptic factor to stabilize cortical excitability (23,24).
AMPA receptors are tetramers assembled by GluA1-4 subunits. Different populations of AMPARs have distinct properties and regulatory mechanisms. GluA2-containing AMPARs are dominant in glutamatergic synapses of neocortex and hippocampus (47, 48), whereasGluA2-lacking AMPARs can be  synaptically recruited under certain conditions, such as neuronal activity blockade (49), sensory deprivation (50), or ischemia (51). GluA2-lacking AMPARs have prominent inward rectifying property (34), Ca 2ϩ permeability (52) and larger single channel conductance (53), compared with GluA2-containing AMPARs. The unique Ca 2ϩ permeability of GluA2-lacking AMPARs plays an important role in homeostatic plasticity (49) and LTP induction in hippocampus (35). Here, we show that GluA2-lacking (GluA1 homomeric) AMPARs is the major target of D 4 . D 4 reduces the GluA2-lacking AMPARs that are delivered by acute stress, and D 4 delivers more GluA2-lacking AMPARs to the synapse in repeatedly stressed animals. Thus, by redistributing GluA2-lacking AMPARs, D 4 restores glutamatergic transmission in stress conditions. It is expected that GluA2 knock-out animals will show differential effects of D 4 R activation when acutely stressed and repeatedly stressed animals are compared.
One of the most important regulators of AMPARs and synaptic plasticity is CaMKII (54). CaMKII activation has been found to increase channel conductance (55) and promote synaptic insertion of AMPARs (56). In this study, we demonstrate that D 4 markedly decreases CaMKII activity in acutely stressed animals, which is important for the D 4 reduction of AMPAR responses. On the other hand, D 4 potently increases CaMKII activity in repeatedly stressed animals, which is important for the D 4 enhancement of AMPAR responses. It further confirms that the D 4 -induced homeostatic regulation of AMPARs depends on the activity-dependent bi-directional regulation of CaMKII (23,24,38).
Our previous study has shown that at the high activity state, D 4 suppresses AMPAR responses by disrupting the kinesin motor-based transport of GluA2 along microtubules via a mechanism dependent on CaMKII inhibition. On the other hand, at the low activity state, D 4 potentiates AMPAR responses by facilitating synaptic targeting of GluA1 through the scaffold protein SAP97 via a mechanism dependent on CaMKII stimulation (24). Thus, the major mechanism underlying the actions of D 4 in stress conditions is likely to be CaMKII-mediated regulation of AMPAR trafficking, not phosphorylation-dependent changes on channel conductance.
In summary, we conclude that D 4 is a synaptic gatekeeper that stabilizes glutamatergic transmission in prefrontal cortex. When neuronal activity changes in situations such as acute or repeated stress, D 4 redistributes GluA2-lacking AMPARs at synapses through the bi-directional control of CaMKII. This homeostatic action of D 4 provides a potential mechanism for the unique role of D 4 in many stress-related neuropsychiatric disorders.