Dopaminergic Modulation of Striatal Inhibitory Transmission and Long-Term Plasticity

Dopamine (DA) modulates glutamatergic synaptic transmission and its plasticity in the striatum; however it is not well known how DA modulates long-term plasticity of striatal GABAergic inhibitory synapses. This work focused on the analysis of both dopaminergic modulation of inhibitory synapses and the synaptic plasticity established between GABAergic afferents to medium spiny neurons (MSNs). Our results showed that low and high DA concentrations mainly reduced the amplitude of inhibitory synaptic response; however detailed analysis of the D1 and D2 participation in this modulation displayed a wide variability in synaptic response. Analyzing DA participation in striatal GABAergic plasticity we observed that high frequency stimulation (HFS) of GABAergic interneurons in the presence of DA at a low concentration (200 nM) favored the expression of inhibitory striatal LTD, whereas higher concentration of DA (20 μM) primarily induced LTP. Interestingly, the plasticity induced in an animal model of striatal degeneration mimicked that induced in the presence of DA at a high concentration, which was not abolished with D2 antagonist but was prevented by PKA blocker.


Introduction
Dopamine (DA) is involved in different functions of the nervous system like cognition, reward mechanisms, motor functions, learning, and memory. In the striatum, DA modulates synaptic transmission and synaptic plasticity through the activation of its DA receptors [1][2][3][4]. The modulation of DA depends on receptor subtype stimulated in a particular synapse; for example, activation of D1-class dopamine receptors (D1) increases glutamatergic responses mediated by Nmethyl-D-aspartate (NMDA) and non-NMDA receptors in the corticostriatal pathway, and the stimulation of D2-class dopamine receptors (D2) attenuates them [5,6]. Additionally, D1 and D2 activation is involved in the generation of longterm depression (LTD) and D1 contributes to long-term potentiating (LTP) of excitatory synapses of corticostriatal pathway [7,8].
Although the striatum receives a massive amount of glutamatergic influences from the cortex and thalamus, MSNs, the projection cells, and most of local circuits of interneurons use GABA as a neurotransmitter [9]. GABAergic synapses on MSNs are also modulated by DA; for instance, activation of postsynaptic D1 receptors reduces GABA currents in MSNs [10], whereas the presynaptic activation of D1 receptors of axon collaterals of other MSNs increases the IPSC amplitude on MSNs and presynaptic D2 receptors decrease it [11].
Stimulation of MSNs or GABAergic interneurons with low and high frequency stimulation protocol induces shortterm depression (STD) in striatal GABAergic synapses as the main form of synaptic plasticity [12,13]. This short-term synaptic plasticity is also modulated by DA; activation of D1 receptors increases STD, while D2 agonist decreases it [14]. Long-term plasticity can be produced at inhibitory synapses on MSNs [15]; however, it remains unclear how DA modulates this type of plasticity.
Changes in DA content in the striatum are related to motor impairments, procedural learning, and cognitive deficits in animal models of neurodegeneration; low content of DA is associated with the development of Parkinson's Disease (PD), while high concentration of DA is related to symptoms of early stages of Huntington's Disease (HD) [16][17][18]; perhaps changes in synaptic plasticity underlie some of the deficits observed in such pathologies. Altered corticostriatal synaptic plasticity has been described after modifying DA receptors or DA content [8,19], but it is not known how DA modulates inhibitory long-term synaptic plasticity on MSNs in striatal degeneration, nor if DA concentration is determinant to produce an specific type of plasticity. In this study, DA effects on GABAergic inhibitory synaptic connections on principal neurons as well as in its synaptic long-term plasticity were analyzed. In addition, inhibitory long-term synaptic plasticity of an animal model of striatal degeneration was compared with the plasticity obtained using different concentrations of DA in control conditions.

Animals and Slice Preparation.
Male C57/BL6 mice (Harlan Laboratories Inc., Mexico) ∼40 days old and kept at room temperature (RT, 25 ∘ C) under a 12 : 12 h light : dark cycle with free access to food and water were used. To evaluate plasticity of inhibitory synapses on MSN in an animal model of striatal degeneration, some mice were treated with the mitochondrial toxin 3-nitropropionic acid (3-NP) for 5 days (15 mg/kg, once a day) as previously described [20,21]. To identify MSNs that exhibited plasticity BACD1-GFP mice were used in some experiments. Experimental procedures were performed in accordance with international guidelines of animal care and the National Committee on Animal Research Ethics for the Care of Laboratory Animals (NOM-062-ZOO-1999).
Inhibitory synaptic currents were evoked via a tungsten stimulating electrode (12 m tip; FHC, 0.1 Hz) positioned inside the striatum (100 m from the recording pipette) and connected to an isolation unit (DS2AK, Digimiter Ltd., Hertfordshire, UK). All recordings were conducted in MSNs (Figure 1(a), = −70 mV) in the presence of glutamatergic antagonists (CNQX 10 m and APV 50 m) to isolate the GABAergic component, which was abolished in the presence of Bicuculline (10 M; Figure 1(b)). To induce striatal plasticity, a high frequency stimulation (HSF) protocol was used (3 trains, 3 seconds, of 100 Hz with an interval of 10 seconds between each train), and the cells were recorded for at least 40 min after the HFS. . Fluorescent images were obtained  with a Hamamatsu (Orca C4742-95) camera coupled to the  Olympus image acquisition system Cell M (excitation light  450-490 nm through a dichroic filter; emission light 502-538). Some slices were fixed and covered with Vectashield (Vector Laboratories, Burlingame, CA) for reconstruction. Visualization was made through a confocal fluorescence microscopy (Olympus Fv-1000) and acquired with the OLYM-PUS FLUO VIEW 3.1 software.

Statistical
Analysis. Data were analyzed and plotted offline using Microcal Origin 7 (Microcal Origin Lab Corporation, Northampton, MA, USA) and the statistical software Sigma Plot (Systat Software, Inc., San José, CA) with a parametric test or a nonparametric test if the data did not display a normal distribution. Data are expressed as mean ± SEM and significance was set at < 0.05. The final figures were edited using Adobe Illustrator 10 or Adobe Creative Suite 5 (Adobe Systems, Inc., San José, CA).

Dopaminergic Modulation of GABAergic Synaptic Transmission.
MSNs of the striatum have two types of GABAergic synapses, those from axon collaterals of other MSNs and those from GABAergic interneurons. We performed intrastriatal stimulation; then, most of the GABAergic synaptic response was due to the stimulation of GABAergic interneurons [11,14,22,23]. To determine the role of DA on these synapses, the effects of DA, DA agonists and antagonists on the GABAergic Inhibitory Postsynaptic Currents (IPSCs) were analyzed.

DA Modulation of Striatal GABAergic Transmission.
Several studies have shown that DA modulatory effects on GABAergic transmission depend on the activation of different DA receptor subtypes [24,25]. Then, to evaluate modulatory effects of DA on striatal GABAergic transmission, we studied the effect of DA treatment on MSNs using two different concentrations of DA (200 nM and 20 M) previously reported to have differential effects on modulation of GABAergic transmission (Li et al., 2012). Low DA concentration (200 nM) decreased IPSC amplitude compared with the control in 42.9% ( = 3) of the recorded cells ( 2 = 5.968, = 0.027; two-tailed paired -test; Figures 2(a)-2(c)). In these cells, the PPR did not exhibit a significant change (Figure 2(d)), which would suggest a postsynaptic modulatory effect; however the kinetics current did not change in the presence of low DA concentration (Figures 2(e) and 2(f)) which suggests a presynaptic modulatory effect. The remaining recorded cells ( = 4, 57.1%) did not exhibit any amplitude change in the presence of low DA concentration (Figure 2(c)), suggesting that DA in low concentration did not modulate all GABAergic transmission on recorded MSNs.
Thirteen MSNs were evaluated with higher concentration of DA (20 M), and 53.8% ( = 7) of the recorded cells did not exhibit any modulation (Figure 2(i)), but 38.5% ( = 5) of the recorded cells did decrease the IPSC amplitude in its presence. Figures 2(g) and 2(h) illustrate that the IPSC amplitude was reduced by 39.9% from control amplitude ( 4 = 3.150, = 0.0345; two-tailed paired -test; Figure 2(g)). The PPR did not exhibit significant changes (PPR: 4 = −1.171, = 0.307; Figure 2(j)); in addition, the rise time significantly increased by 28% compared with control (rise time: 4 = 3.393, = 0.0275; two-tailed paired -test; Figure 2(k)), and the decay time constant remained the same (decay time: 2 = −1.928, = 0.12; Figure 2(l)). This data suggests that 20 M of DA modulates GABAergic transmission through a postsynaptic mechanism. Only one cell (7.7%) exhibited an amplitude increase after the application of DA at the high concentration (data not shown).

D1 and D2 Receptor Activation Modulates Striatal GAB-Aergic Transmission.
In order to identify the DA receptor responsible for specific effects of GABAergic transmission modulation, it was evaluated in the presence of D1 or D2 agonist and antagonist.
Paired-pulse analysis and measurement of the rise time and the decay time of the IPSC were analyzed to determine the pre-or postsynaptic nature of the modulatory effect. For the recordings in which the amplitude of the IPSC increased, the paired-pulse ratio (PPR) did not change (PPR: 2 = 0.301, = 0.778; two-tailed paired -test; Figure 3(b)), nor did the time constants (rise time: 2 = −0246, = 0.818; two-tailed paired -test; Figure 3(c); decay time: 2 = 1677, = 0.169; two-tailed paired -test; Figure 3(d)). For those recordings in which the IPSC exhibited a decrease in the presence of the D1 agonist, the decrease in the PPR was not significantly different ( = 4, PPR: 3 = 2324, = 0.059), nor were the changes in the kinetics (rise time: 3 = −0287, = 0.784; two-tailed paired -test; decay time: 3 = −0340, = 0.746; two-tailed paired -test). To construct the kinetics bars presented in Figures 3(c) and 3(d), all of the data from the cells were combined because they did not exhibit statistical changes in their kinetics. PPR exhibited no change in the presence of the D1 agonist, which may indicate a postsynaptic modulatory mechanism of the SKF81297; however the kinetics of the currents did not change in the presence of the D1 agonist which indicates a presynaptic mechanism for this modulation.
We next went to evaluate if endogenous DA affected IPSC amplitude by acting on D1 receptors; then D1 receptors were blocked. In the presence of the D1 antagonist, SCH23390 (1 M), the IPSC amplitude increased by 54% compared with  Neural Plasticity To evaluate if DA effects on IPSC amplitude were mediated by the activation of D2 receptors, 11 MSNs were evaluated in the presence of the D2 agonist (Quinelorane, 10 M). In this situation, the IPSC amplitude decreased in 54.5% ( = 6), increased in 27.3% ( = 3), and produced no effect in 18.2% ( = 2) of the recorded cells (Figure 3(i)). In the cells with an IPSC amplitude decrease, the amplitude was reduced by 41% compared with control and the modulation was statistically significant (control 100.11 ± 0.26 versus Quinelorane 58.69 ± 3.046, 5 = 13.246, = 0.00004; two-tailed paired -test; Figure 3(i)). These results indicated that DA through the activation of D2 receptors reduced GABAergic transmission. The PPR did not change suggesting a postsynaptic mechanism (Figure 3(j)), nor however the IPSC kinetics did, suggesting a presynaptic mediated mechanism (Figures 3(k) and 3(l)).
To evaluate if endogenous DA affected IPSC amplitude through D2 activated mechanisms, the D2 antagonist sulpiride (1 M) was studied. Sulpiride produced no change in the IPSC amplitude, PPR, or current kinetics ( = 5, Figures 3(m) and 3(n)). These experiments illustrated that endogenous DA do not modulate IPSC amplitude by D2 activation.

GABAergic Synaptic Plasticity.
To analyze the synaptic plasticity of the IPSCs in MSNs, HFS was given to 14 MSNs cells (see Section 2). Figure 4(a) presents the IPSC amplitude before (top) and after HFS (middle). HFS significantly decreased the IPSC amplitude by 49% compared with the amplitude before the train ( 6 = 5919, < 0.001), in 50% ( = 7) of the recorded cells; this reduction persisted for more than 30 minutes; then, the current amplitude decrease was considered to be long-term depression (LTD; Figure 4(b)). The PPR before and after HFS did not change ( = 7; Figure 4(d)) nor did the time constants (Figures 4(e) and 4(f)). 42.9% ( = 6) of the recorded cells showed no change in the current amplitude after HFS, and only 7.14% ( = 1) of recorded cells exhibited an amplitude increase (Figure 4(c)).
To evaluate if all types of MSN developed inhibitory longterm plasticity, some experiments were performed on slices coming from BACD1-GFP mice. 13 cells were recorded, 4 cells did not exhibit any plasticity, and both BACD1-GFP-positive ( = 3) and BACD1-GFP-negative ( = 3) MSNs developed LTD after HFS; no one of the evaluated BACD1-GFP-positive cells developed LTP; however 3 BACD1-GFP-negative cells developed LTP. Figure 4(g) illustrates the reconstruction of a D1-GFP-negative expressing neuron that displayed GABAergic LTD; this neuron was considered as D2 expressing MSN.
From these results we conclude that MSNs expressing D1 and D2 developed LTD as the prevalent form of inhibitory long-term plasticity, and only neurons expressing D2 receptors displayed LTP.

GABAergic Synaptic Plasticity in Presence of DA.
Once we showed that GABAergic synapses exhibit LTD as the main form of synaptic plasticity, we investigated whether DA modulate this form of plasticity. The role of DA in modulating GABAergic plasticity was also evaluated using two concentrations (200 nM and 20 M). In the presence of a low concentration of DA (200 nM), HFS induced LTD (Figure 5(a)) in 60% ( = 3) of the recorded cells, LTP was produced in only 20% ( = 1) of the recorded cells ( Figure 5(c)), and 20% ( = 1) did not displayed any plasticity. In the cells in which LTD was observed, the amplitude was reduced by 73.11% compared with the current amplitude before HFS

D2 Receptor Modulation of Striatal GABAergic Synaptic
Plasticity. We next went to evaluate D2 receptors role in striatal plasticity of inhibitory synapses on MSN. In the presence of the D2 agonist Quinelorane (10 M), HFS generated LTD in 50% ( = 7) of the recorded cells and LTP in 28.6% ( = 4) of them. In those cells where LTD was generated, the amplitude of IPSC was reduced by 45.06% compared with the amplitude before HFS (Pre-HFS = 100.895 ± 0.687 versus Post-HFS = 55.833 ± 7.316; 6 = 5.838, = 0.001; twotailed paired -test, Figures 7(a) and 7(b)). The LTD did not produce changes in the PPR or in the current kinetics ( Figures  7(d)-7(f)). In those cells that exhibited LTP in the presence of Quinelorane (Figure 7(c)), the increase in the IPSC amplitude was statistically significant ( 3 = 5685, = 0.002). The PPR significantly decreased ( 3 = 2.398, = 0.0373) without any change in the current kinetics, indicative of presynaptic mechanisms in D2 modulation of LTP (data not shown). The rest of the recorded cells (21.4%, = 3) demonstrated no change after HFS (Figure 7(c)). Data obtained in this section suggested that D2 receptor participates in the induction of LTD but also in LTP of GABAergic plasticity because the percentage of cells that exhibited LTP increased from 7.14 in control conditions to 28.6%.  Figures 8(a) and 8(b)), without changes in the PPR (Figure 8(c)). Conversely, in the presence of     was prevented (Figures 9(g) and 9(h)) and LTD development was favored (Pre-HFS = 98.16 ± 2.95 versus Post-HFS = 55.22 ± 9.54; 2 = 6.52; = 0.02; two-tailed -test) without changes in the PPR (Figure 9(i)). This result supports the idea that LTP of inhibitory synapses on MSNs is triggered by PKA signaling pathway in striatal degeneration and its postsynaptic inhibition promotes the development of LTD.

Discussion
The present study describes that DA modulation of GABAergic synapses on MSNs and the type of plasticity developed depend on the DA concentration and the activation of D1-or D2-class receptors. Also DA effect on striatal plasticity was altered in striatal degeneration.

DA Modulation of GABAergic Synapses on
MSNs. DA at a low and high concentration decreased the IPSC amplitude compared with the control conditions in most of the recorded cells. However when specific D1 and D2 agents were evaluated modulation on GABAergic synapses depended on DA receptors activation or inhibition. The D1 agonist produced decrease and increase in the IPSC amplitude of the MSNs recorded. The D1 family receptors include the D 1 and D 5 receptors, which are differentially located in striatal neurons, and exhibited different affinities for the agonist [26] and different dependencies on the G protein/adenylyl cyclase signaling pathway [27]. Additionally, there are also diverse responses due to the activation of pre-or postsynaptic mechanisms. In recordings from dissociated MSNs, SKF81297 (1-10 M) reduces the postsynaptic GABA ligandgated-currents [10,28]. On the contrary, the presynaptic D 1 receptor activation of axonal collaterals from MSNs stimulates GABAergic synapses on other MSNs [29]. Furthermore, D 5 receptors are expressed in the terminals of GABAergic interneurons [30], because most of the cells recruited with intrastriatal stimulation are the GABAergic interneurons [23]; changes in the IPSC amplitude in the presence of the D1 agonist may be due to the activation of the presynaptic D 5 receptors on GABAergic interneurons.
Blocking the D1 receptors mainly produced an increase in the GABAergic currents, suggesting that when D1 receptors are blocked, the D2 activation by endogenous DA is unmasked. The agonist D2, Quinelorane 10 M, also produces decrease and increase of the IPSC amplitude. D2 receptors belong to a family that has several subtypes, such as D 2 , D 3 , and D 4 ; the striatum mostly expresses D 2 and D 3 [26] and Quinelorane affects both of them and may produce different responses depending on receptor sensitivity and selectivity [31]. D2 receptors are located postsynaptically in the MSNs that express enkephalins, but they are also autoreceptors in the striatal dopaminergic terminals; low concentrations of DA activate them, decreasing the endogenous release of DA as a response [26]. Subsequently, in those cells in which the D2 agonist produced an IPSC amplitude increase, the effect may be postsynaptically mediated, whereas a decrease in the IPSC may be related to a presynaptic effect. Earlier study showed that the stimulation of presynaptic D2 receptors decreases GABAergic synaptic amplitude [29].

GABAergic Synaptic Plasticity and Its DA Modulation.
HFS decreased the IPSC amplitude in 50% of the experiments; these results are consistent with those obtained by others using different stimulation protocols [32,33].
DA concentration was crucial for determining the type of inhibitory plasticity triggered at the striatum. In the presence of DA, HFS generated both LTD and LTP; however, a low DA concentration (200 nM) produced mostly LTD, whereas higher concentration of DA (20 M) favored LTP in half of the experiments. DA at low concentrations activates high affinity receptors, whereas higher concentrations also activate low affinity receptors. D2 receptors possess higher affinity for DA and their stimulation facilitates LTD. Moreover, the administration of sulpiride 10 M prevented the LTD induced with HFS in the presence of DA at 200 nM, whereas the administration of sulpiride in the presence of DA at 20 M did not prevent the induction of LTP, but only H89 did, supporting the notion that D2 receptors play an important role in triggering striatal LTD in inhibitory synapses at the striatum, whereas D1 activation coupled to PKA signaling pathway is mainly involved in generating LTP in the GABAergic synapses. However, we cannot rule out that postsynaptic D2 receptors may mediate increases in the IPSCs after HFS in the presence of 20 M of DA because postsynaptic D2 in MSNs of the indirect pathway can stimulate intracellular calcium through the phospholipase C pathway [34]. This pathway enhances GABA A currents by mobilizing intracellular Ca 2+ [35].

D1 and D2 Modulation of Striatal GABAergic Synaptic
Plasticity. To better understand DA receptors role in plasticity, specific DA agonist and antagonist were evaluated. In the presence of the D1 agonist, HFS produced LTD in all recorded cells. Despite this finding, HFS in the presence of the D1 antagonist still generated LTD in a large percentage of the cells (77.8%), demonstrating that the inhibition of D1 receptors did not prevent the development of striatal LTD in the majority of the inhibitory synapses.
HFS in the presence of the D2 agonist generated LTD in 50% of the cells and LTP in 28% of the cells. Nevertheless, in the presence of the D2 antagonist, all Inhibitory plasticity in MSNs was prevented. In fact, in D2 knockout mice LTD of corticostriatal excitatory plasticity is abolished, generating LTP instead [36]. Our results indicate that D2 receptors play an important role in triggering LTD of inhibitory striatal plasticity as well.

GABAergic Plasticity in Striatal Degeneration.
Synaptic abnormalities in corticostriatal pathway have been described after acute 3-NP treatment [21,37,38], but there is no study of changes on inhibitory synapses on MSNs. Evaluation of synaptic plasticity of this connection on damaged MSNs indicated that 75% of the recorded cells did not exhibit any plasticity, while the other 25% of the cells exhibited only LTP. The LTP produced in slices of 3-NP-treated mice resembled the LTP triggered under high levels of DA in normal tissue (Figures 9(b) and 5(b), resp.). This LTP was affected by the D1 antagonist and the PKA inhibitor (Figures 9(g) and 9(h)) which in its presence generates LTD, suggesting that D1 receptors are involved in the plasticity observed in damaged striatal tissue. Striatal LTP in IPSCs may result from a dysregulation in DA release [39] produced by 3-NP, as observed in other animal models of striatal degeneration [19]. It appears that a sustained elevation of DA or an imbalance in its concentration causes selective degeneration of striatal GABAergic neurons and motor dysfunction [17,18]. A reduction in D1 receptors has been documented in acute 3-NP administration [37]; if this were the case within in vivo subchronic treatment, an alternative mechanism to explain LTP in striatal IPSCs is that feed forward inhibition mechanism, mediated mainly by fast spiking (FS) interneurons [40] on MSNs expressing D1 receptors (direct pathway), was overactivated in damaged striatum; in fact we have shown a reduction in number of spines, and dendrites tick in our 3-NP model of degeneration [21] which may indicate that interneurons projecting on the perisomatic area would have more synaptic impact on MSNs than those arriving in distal areas; FS are the ones heavily innervating perisomatic area in MSNs [9]; then LTP exhibited in MSN expressing D1 receptors may be produced when FS are activated with field stimulation in control slices as well as in slices from damaged striatum; further experiments should test this hypothesis.

Conclusion
DA effect on striatal plasticity was different from its modulatory action in inhibitory synapses on MSNs. The variability on synaptic responses may be due in part to the set of GABAergic afferents to MSNs (mainly interneurons) stimulated that exhibit different DA receptors as well as DA receptors that are present in recorded MSNs.