Electrophysiological Effects of SKF83959 on Hippocampal CA1 Pyramidal Neurons: Potential Mechanisms for the Drug's Neuroprotective Effects

Although the potent anti-parkinsonian action of the atypical D1-like receptor agonist SKF83959 has been attributed to the selective activation of phosphoinositol(PI)-linked D1 receptor, whereas the mechanism underlying its potent neuroprotective effect is not fully understood. In the present study, the actions of SKF83959 on neuronal membrane potential and neuronal excitability were investigated in CA1 pyramidal neurons of rat hippocampal slices. SKF83959 (10–100 µM) caused a concentration-dependent depolarization, associated with a reduction of input resistance in CA1 pyramidal neurons. The depolarization was blocked neither by antagonists for D1, D2, 5-HT2A/2C receptors and α1-adrenoceptor, nor by intracellular dialysis of GDP-β-S. However, the specific HCN channel blocker ZD7288 (10 µM) antagonized both the depolarization and reduction of input resistance caused by SKF83959. In voltage-clamp experiments, SKF83959 (10–100 µM) caused a concentration-dependent increase of Ih current in CA1 pyramidal neurons, which was independent of D1 receptor activation. Moreover, SKF83959 (50 µM) caused a 6 mV positive shift in the activation curve of Ih and significantly accelerated the activation of Ih current. In addition, SKF83959 also reduced the neuronal excitability of CA1 pyramidal neurons, which was manifested by the decrease in the number and amplitude of action potentials evoked by depolarizing currents, and by the increase of firing threshold and rhoebase current. The above results suggest that SKF83959 increased Ih current through a D1 receptor-independent mechanism, which led to the depolarization of hippocampal CA1 pyramidal neurons. These findings provide a novel mechanism for the drug's neuroprotective effects, which may contributes to its therapeutic benefits in Parkinson's disease.

Our previous work demonstrated that SKF83959 exerted a potent neuroprotective action in rat pheochromocytoma cells (PC12 cells) treated with H 2 O 2 [12]. This action, however, was only partially attributed to inhibition of glycogen synthase kinase-3b (GSK3b) by SKF83959 via activation of D 1 -like receptor. Therefore, other mechanisms independent of D 1 -like receptor may be involved in the neuroprotection by SKF83959. Accumulating evidence shows that enhanced delayed rectifier K + channel induces neuronal death, while blocking K + outflow through the K + channel promotes the survival of neurons [13][14][15]. We have shown recently that SKF83959 is a potent blocker of the delayed rectifier K + channels in rat hippocampal pyramidal neurons [16], which may contribute to the non-receptor mechanisms of the neuroprotection of the drug. The membrane properties and excitability play critical roles in physiological and pathological activity of brain neurons. It has been demonstrated that an increased neuronal excitability in the pathological state of 6-OH-DOPA-lesioned rat model for Parkinson's disease [17][18][19]. Alongside our previous study [16], the present study was designed to further explore the effects of SKF83959 on the neuronal membrane properties and excitability in rat hippocampal pyramidal neurons. The present results demonstrate that SKF83959 not only induces membrane depolarization of CA1 pyramidal neurons via an enhancement of Ih current in a DAreceptor-independent manner, but also reduces the neuronal excitability. These findings may provide a novel mechanism for the drug's neuroprotective effects and its anti-Parkinsonian efficacy.

Results
Effects of SKF83959 on passive membrane properties of hippocampal CA1 pyramidal neurons Bath application of SKF83959 caused a reversible depolarizing response of CA1 pyramidal neurons in rat hippocampal slices (Fig. 1A). The maximal responses caused by SKF83959 at the concentrations of 10, 50 and 100 mM were 2.560.4 mV (n = 5), 6.961.4 mV (n = 6) and 9.761.0 mV (n = 7), respectively. The depolarization never led to spontaneous firing of the recorded neuron even when the resting membrane potential was set to a level close to the threshold of action potential firing by injecting steady depolarizing current. The depolarization caused by SKF83959 (50 mM) persisted, when TTX (0.5 mM) was included in the perfusion medium (6.860.7 mV, n = 5, unpaired t test, P.0.05 vs. SKF83959 alone, Fig.1B). Depolarizing responses were also observed in acutely dissociated CA1 pyramidal neurons (PND5-6). Superfusion of SKF83959 (50 mM) caused a depolarization of 4.960.5 mV (n = 6) in dissociated single CA1 pyramidal neurons. The results suggest that the depolarization effect of SKF83959 was independent of synaptic connections.
In a representative experiment shown in Fig. 1A, the membrane potential was manually clamped to the control level during SKF83959 application to monitor the change of input resistance of the recorded neurons. The pooled data from the 10 neurons shows that SKF83959 (50 mM) caused a small but statistically significant reduction of the input resistance (from 143.5611.2 MV to 135.769.5 MV, paired t test, P,0.05, Fig. 1C). The results suggest that SKF83959-induced membrane depolarization was accompanied by a reduction of the input resistance.

Effects of SKF83959 on subthreshold responses of hippocampal CA1 pyramidal neurons
The hyperpolarization-activated non-selective cation current (Ih), which is mediated by the hyperpolarization-activated, cyclic- Figure 1. SKF83959 induced depolarizing response in CA1 pyramidal neuron in hippocampal slices. A. Resting membrane potential recorded in a representative neuron. The upper trace shows the membrane potential of the neuron. The resting potential was 260 mV, whereas the input resistance change was monitored in the upper trace through injecting hyperpolarizing current pulses (400 ms, 250 pA, lower trace) every 10 sec (the downward deflections). The black bar denotes the perfusion with SKF83959 (50 mM). To exclude the change of input resistance caused indirectly by depolarizing response, the membrane potential during SKF83959 application was manually clamped to the pre-drug level. Bicuculline (5 mM) was added in ASCF to suppress the spontaneous IPSPs. B. Bar graphs showing the maximal depolarization caused by SKF83959 (50 mM) in the presence of TTX (0.5 mM, n = 5), SCH (D1 receptor antagonist SCH23390, 10 mM, n = 5), Rac (D2 receptor antagonist raclopride, 10 mM, n = 5), Mes (5-HT2A/2C receptor antagonist mesulergine, 10 mM, n = 5), Pra (Alpha1-adrenoceptor antagonist prazosin, 10 mM, n = 5), or following intracellular dialysis of GDP-b-S (0.5 mM, n = 6). C. Bar graphs showing the input resistance in the control (Ctrl) and during perfusion with SKF83959 (SKF, 50 mM, n = 10, *P,0.05). doi:10.1371/journal.pone.0013118.g001 neucleotide gated (HCN) channels, plays a crucial role in setting the resting membrane potential of neurons [20]. To explore the mechanism underlying SKF83959-induced depolarization, we examined the effect of SKF83959 on the voltage sag caused by prolonged hyperpolarizing current pulse, which is a hallmark of Ih activation [21]. In the presence of SKF83959 (50 mM), injecting the hyperpolarizing current pulses produced more pronounced voltage sags ( Fig. 2A). In a group of neurons tested, the voltage sag ratio was significantly increased from 1.460.03 to 1.560.03 (n = 10, paired t test, P,0.01, Fig. 2B).
In order to confirm the involvement of Ih in SKF83959-induced depolarization, we pretreated the slice with ZD7288, a specific HCN channel blocker, and found that ZD7288 (10 mM) completely abolished the voltage sag either in the absence or in the presence of SKF83959 (Fig. 2C). Furthermore, inhibition of Ih by ZD7288 (10 mM) significantly antagonized both the depolarizing response and input resistance reduction caused by SKF83959. As shown in Figs. 2D, 2E, the maximal depolarizing response caused by SKF83959 (50 mM) alone was 6.961.4 mV (n = 6), whereas SKF83959 induced depolarization response was significantly reduced to 2.660.5 mV (n = 8, unpaired t test, P,0.01) in the presence of ZD7288. Moreover, pretreatment with ZD7288 (10 mM) drastically increased the input resistance of the neurons tested to 276.3616.5 MV (n = 8). However, subsequent perfusion with SKF83959 failed to change the neuronal input resistance in the presence of ZD7288 (n = 8, 282.4619.2 MV, paired t test, P.0.1). The above results suggest that activation of Ih is responsible for SKF83959-induced depolarization.

SKF83959 enhanced Ih current in hippocampal CA1 pyramidal neurons
The above data indicated that SKF83959 enhances the activity of HCN channels. Consistently, perfusion with SKF83959 (50 mM) markedly increase the amplitude of Ih of CA1 pyramidal neurons in hippocampal slice at all the potentials tested (Fig. 3A). Plot of the averaged current/voltage (I/V) relationship of Ih in the presence or absence of SKF83959 reveals a down-shift of the I/V curve, and an enhancement of the maximal steady-state current amplitude (from 270 to 2120 mV, Fig. 3B). The amplitude of Ih at 2120 mV was increased from 336623 pA to 454663 pA (n = 6, paired t test, P,0.05). Moreover, SKF83959-induced enhancement of Ih was reversible (Fig. 3C) and in a dose-dependent manner (Fig. 3D). At the concentrations of 10, 50, and 100 mM, SKF83959 increased the amplitude of Ih by 18.962.0% (n = 5, ANOVA, P,0.05), 54.765.0% (n = 11, ANOVA, P,0.001) and 85.9610.6% (n = 6, ANOVA, P,0.001), respectively.
In order to elucidate the mechanisms for the drug-mediated enhancement of Ih, we first examined whether SKF83959 modulates the gating mechanisms of Ih channels. Perfusion with SKF83959 (50 mM) caused a right shift of the activation curve of Ih (Fig. 4A). In the control period, the half-activation potential for Ih (V 1/2 ) was 290.861.8 mV (n = 6), and the slop factor was 11.561.0 (n = 6). In the presence of SKF83959, the value of V 1/2 changed to 284.261.7 mV (n = 6, paired t test, P,0.001), and the slop factor to 14.161.1 (n = 6, paired t test, P,0.01). Furthermore, SKF83959 accelerated the activation of Ih in steps to large hyperpolarizing voltage steps (from 2120 to 2100 mV) (Fig. 4B). The Ih current traces could be fitted with bi-exponential functions with a relatively stable fast component, which accounted for the majority of Ih, followed by a variable slow component. The fast activation time constant (t f ) of Ih in steps to 2120, 2110 and 2100 mV in the control period were 29.561.8 ms, 35.262.1 ms and 39.962.6 ms, respectively; whereas subsequent perfusion with SKF83959 (50 mM), the values of t f were significantly reduced to 24.561.7 ms (n = 6, paired t test, P,0.01), 30.461.7 ms (n = 6, paired t test, P,0.01) and 33.261.7 ms (n = 6, paired t test, P,0.05), respectively.
We then examined whether the enhancement of Ih by SKF83959 was mediated through activation of D 1 -like receptors. As shown in Fig. 5A, in the presence of D 1 -like receptor antagonist SCH23390 (10 mM), perfusion with SKF83959 (50 mM) increased the amplitude of Ih by 62.9612.4% (n = 6, paired t test, P,0.05), which was close to that obtained in the absence of SCH23390 (54.765.0%, unpaired t test, n = 11). Intracellular dialysis of GDPb-S did not prevent the enhancement of Ih by SKF83959 either (Figs. 5B). With GDP-b-S (0.5 mM) present in the recording pipettes, perfusion with SKF83959 (50 mM) increased the amplitude of Ih by 81.9610.1% (n = 9, paired t test, P,0.001). Similar result was obtained with intracellular dialysis of GppNHp, a hydrolysis-resistant GTP analog, which uncoupled G-protein (Figs. 5C). With GppNHp (0.5 mM) present in the recording pipettes, perfusion with SKF83959 (50 mM) increased the amplitude of Ih by 79.3610.8% (n = 6, paired t test, P,0.001).
It has been shown that the activation of Ih is facilitated by cAMP in a direct, PKA-independent manner, and there is a cyclic nucleotide-binding domain (CNDB) on the C-terminal of each subunit of the channel [20]. However, intracellular dialysis of cAMP (100 mM) or Rp-cAPM (100 mM), a hydrolysis-resistant cAMP analog, did not occlude the enhancement of Ih by SKF83959 (Figs. 5D, 5E), indicating that SKF83959-induced enhancement of Ih current is independent of intracellular cAMPrelated mechanisms. With cAMP (100 mM) or Rp-cAPM (100 mM) present in the recording pipettes, perfusion with SKF83959 (50 mM) increased the amplitude of Ih by 70.7619.6% (n = 6, paired t test, P,0.01) or by 55.168.3% (n = 6, paired t test, P,0.001), respectively. Effects of SKF83959 on somatic excitability of hippocampal CA1 pyramidal neurons Repetitive discharge of CA1 pyramidal neurons was elicited by injecting prolonged depolarizing current pulses. Comparing the records prior to and after bath application of SKF83959 (50 mM) reveals that the drug markedly reduced the number and the amplitude of action potentials evoked by the depolarizing current pulses (Figs. 6A, 6B), and increased the latency of the first spike firing in the train (Fig. 6C). Furthermore, the rheobase current (the minimum current to evoke a single action potential) was also increased from 12267 pA to 13267 pA (n = 5, paired t test, P,0.05) (Fig. 6D).  To further characterize the influence of SKF83959 on action potential, single spike was elicited by injecting depolarizing current pulse. Perfusion of SKF83959 (50 mM) slowed down both the upstroke and repolarizing phases of the spike. As a result, the action potential was broadened (Fig. 7A, 7C). Pooled data from 8 neurons showed that SKF83959 significantly reduced the amplitudes of action potential (from 96.362.1 to 81.862.7 mV, paired t test, P,0.001) and increased the half-width of the spikes from 1.260.1 to 1.760.1 ms, paired t test, P,0.01) (Fig. 7C). In addition, the threshold of action potential was also significantly raised from 239.260.5 to 234.761.2 mV (P,0.01, paired t test, Fig. 7D). All the above data indicated that the somatic excitability of hippocampal CA1 pyramidal neurons was dramatically reduced by SKF83959.

Discussion
In the present study we characterized the electrophysiological effects of SKF83959, an atypical D 1 -like receptor agonist, on the passive membrane properties and excitability of hippocampal CA1 pyramidal neurons. The major findings are summarized as following: (1) SKF83959 caused depolarizing response associated with a reduction of input resistance; (2) SKF83959-induced depolarization was mediated mainly by an enhancement of Ih via a D 1 -like receptor-independent mechanism. (3) SKF83959 reduced the neuronal excitability.
In the present study we demonstrate that SKF83959 causes a concentration-dependent depolarizing response in rat hippocampal CA1 pyramidal neurons. Furthermore, we found that the effect was sensitive neither to bath application of TTX, SCH23390, raclopride, mesulergine, prazosin, nor to intracellular dialysis of GDP-b-S. Therefore, it is clear that the response of SKF83959 is not mediated by interaction of SKF83959 with G-protein coupled receptors. This implicates that the drug's effect was mediated by an action on ion channels responsible for setting the resting membrane potential, such as TASK-1 channels, HCN channels, etc. [20,22]. Indeed, in the current-clamp experiments we showed that SKF83959 enhanced the voltage sag caused by prolonged hyperpolarizing current pulse, which is a hallmark of HCN channel activation [20]. In voltage-clamp experiments we demonstrated that SKF83959 enhanced the Ih current. Furthermore, the enhancement of Ih by SKF83959 is also insensitive to bath application of SCH2339 or intracellular dialysis of GDP-b-S or GppNHp (Figs. 5A, 5B, 5C), suggesting that the effect was independent of activation of D 1 -like receptors. Therefore, both the results form current-and voltage-clamp experiments suggest that SKF83959-induced depolarization mainly due to the enhancement of Ih current via DA receptor-independent mechanisms.
There is a cyclic nucleotide-binding domain (CNDB) on the Cterminal of each HCN channel subunit, and the binding of cAMP directly facilitates activation of HCN channels [20]. In fact, the effects of SKF83959 on Ih resemble those caused by cAMP: (1) causing a right shift of the activation curve of Ih (Fig. 4A), and (2) speeding up the activation kinetics (Fig. 4B). However, intracellular dialysis of high concentrations of cAMP or Rp-cAPM did not occlude the enhancement of Ih by SKF83959 (Figs. 5D, 5E), suggesting that the compound's effect was not mediated by intracellular cAMP. At the present, several specific blockers of HCN channels including ZD7288 are commercially available, which have been developed as ''heart rate-lowering'' agents to block a pacemaker current I f in cardiomyocyte [23]. On the other hand, few small molecule compounds other than cyclic nucleotides were found thus far to enhance Ih current. The anticonvulsant lamotriqine was reported to preferentially alter the dendritic excitability of hippocampal CA1 pyramidal neurons through increase of Ih current (Poolos et al. 2002). To our knowledge, our study provides the first evidence that SKF83959 represents another small molecule activator of Ih current. The molecular target of SKF83959 on HCN channel, however, remains to be identified.
Theoretically, a depolarizing response would increase the spontaneous firing and the number of action potentials evoked by depolarizing pulses. In the present study, spontaneous firing was never observed during SKF83959-induced depolarization. In contrast, we found that SKF83959 significantly reduced the number (Figs. 6A, 6B)and amplitude (Figs. 7A, 7B) of action potentials evoked by depolarizing current pulses, prolonged the latency of the first spike (Fig. 6C) and increased the rheobase current ( Fig. 6D) as well as the threshold of action potential firing (Fig. 7D). The somatic recordings demonstrate that SKF83959 suppress the excitability of postsynaptic CA1 pyramidal neurons. However, the effects do not seem to be due to the increased Ih current. A non-uniform gradient of HCN channel distribution has been demonstrated in hippocampal CA1 pyramidal neuron with the distal dendrites containing a much higher density of HCN channels than that of the soma [24]. As a result, lamotriqine that increased Ih current preferentially reduced dendritic excitability, while minimally affecting the somatic excitability of CA1 pyramidal neuron [25]. We recently demonstrated that SKF83959 exerted potent inhibition on voltage-activated Na + current in acutely dissociated hippocampal pyramidal neurons (data not shown), which may explain the reduction of overall excitability of CA1 pyramidal neurons reported here. The broadening of action potential (Fig. 7A, 7C) could be attributed to the blockade of the delayed rectifier K + current by SKF83959 [16], one of the outward currents responsible for the repolarization of action potentials [26].
The inhibition of neuronal excitability by SKF83959 may contribute to its therapeutic benefits in Parkinson's disease (PD). It was found that the spontaneous activity of striatal neurons in 6-OHDA-lesioned PD rats was several folds higher than in control animals [17][18][19]. Intracellular recording conducted in striatal slices also demonstrated that dopamine-denervation increased neuronal excitability [27]. The hyperactivity of striatal neurons in PD would augment the GABAergic control over the output nucleus of basal ganglia, which may associate with some motor symptoms observed in the disease [28]. SKF83959 may reduce the hyperexcitability of striatal neurons in PD, which, in turn, contributes to its therapeutic effects, including the attenuation of the development of dyskinesia. In addition, inhibition of neuronal excitability by SKF83959 may also improve neuronal survival and contribute to its neuroprotective effect. In the context of PD, neuroprotective effect is important for slowing down the progressive loss of dopamine neurons. It is noted that the present data were obtained from hippocampal pyramidal neurons, whereas Parkinson's disease is more relevant to striatal MSN neurons or DA neurons in SNC. However, hippocampal pyramidal neurons are the most widely studied cell in the central nervous system and serve as excellent models for the general neuronal activity seen in other parts of the nervous system. We believe that the effects found in the present study should occur in other brain regions, such as in striatum where Ih channel also widely expressed [29]. Regardless, these receptor-independent mechanisms provide a novel insight for the drug's potent neuroprotective action.

Electrophysiological recordings from hippocampal pyramidal neurons
Male Sprague-Dawley rats (2-3 weeks of age) were anesthetized with 10% chloral hydrate (400 mg/kg, i.p.) and decapitated. The brain was rapidly removed and placed in an ice-cold ACSF containing the following (in mmol/L): NaCl 119, KCl 2.5, CaCl 2 2.5, MgSO 4 1.3, NaH 2 PO 4 1, NaHCO 3 26.2, and glucose 11, bubbled with a gas mixture (95%O 2 and 5%CO 2 ). Transverse hippocampal slices (350 mm) were cut using a M752 vibroslice (Campden Instruments Ltd., UK), and incubated in the ACSF at room temperature. After equilibration for at least 1 hour, one piece of the slices was transferred to recording chamber and perfused with oxygenated ACSF at a rate of 2-3 ml/min.

Data acquisition and analysis
Data are presented as mean6S.E.M. Statistical significance was assessed using paired or unpaired Student's t test or ANOVA, and P,0.05 was considered to be significant. Data analyses were performed using the software Excel 2003 (Microsoft) and Origin 8.0.
For preparing stock solutions, SKF83959 and prazosin were dissolved in dimethylsulfoxide (DMSO), other drugs in distilled water. The solutions were stored at 220uC, and diluted in ACSF to desired concentrations before use. DMSO with a final concentration less than 0.1% had no detectable effect on the membrane properties and Ih current of the recorded neurons, nor did the receptor antagonists at desired concentrations. Most drugs were delivered to the slice through perfusion, expect GDP-b-S, GppNHp, cAMP and Rp-cAMP, which were added in the pipette solution, and dialyzed into the neurons recorded.