The Multimodal Serotonergic Agent Vilazodone Inhibits L-DOPA-Induced Gene Regulation in Striatal Projection Neurons and Associated Dyskinesia in an Animal Model of Parkinson’s Disease

Levodopa (L-DOPA) treatment in Parkinson’s disease is limited by the emergence of L-DOPA-induced dyskinesia. Such dyskinesia is associated with aberrant gene regulation in neurons of the striatum, which is caused by abnormal dopamine release from serotonin terminals. Previous work showed that modulating the striatal serotonin innervation with selective serotonin reuptake inhibitors (SSRIs) or 5-HT1A receptor agonists could attenuate L-DOPA-induced dyskinesia. We investigated the effects of a novel serotonergic agent, vilazodone, which combines SSRI and 5-HT1A partial agonist properties, on L-DOPA-induced behavior and gene regulation in the striatum in an animal model of Parkinson’s disease. After unilateral dopamine depletion by 6-hydroxydopamine (6-OHDA), rats received repeated L-DOPA treatment (5 mg/kg) alone or in combination with vilazodone (10 mg/kg) for 3 weeks. Gene regulation was then mapped throughout the striatum using in situ hybridization histochemistry. Vilazodone suppressed the development of L-DOPA-induced dyskinesia and turning behavior but did not interfere with the prokinetic effects of L-DOPA (forelimb stepping). L-DOPA treatment drastically increased the expression of dynorphin (direct pathway), 5-HT1B, and zif268 mRNA in the striatum ipsilateral to the lesion. These effects were inhibited by vilazodone. In contrast, vilazodone had no effect on enkephalin expression (indirect pathway) or on gene expression in the intact striatum. Thus, vilazodone inhibited L-DOPA-induced gene regulation selectively in the direct pathway of the dopamine-depleted striatum, molecular changes that are considered critical for L-DOPA-induced dyskinesia. These findings position vilazodone, an approved antidepressant, as a potential adjunct medication for the treatment of L-DOPA-induced motor side effects.


Introduction
Levodopa (L-DOPA) remains the gold standard treatment for Parkinson's disease (PD) since its introduction in the 1960s [1,2]. As a dopamine precursor, L-DOPA helps restore the disrupted dopamine tone in the basal ganglia in PD patients. However, after 4 to 10 years of L-DOPA treatment, 50-90% of patients experience abnormal hyperkinetic motor complications termed L-DOPA-induced dyskinesias ( [3]; reviewed in [4,5]), which severely limit the usefulness of this treatment.
A large literature has related the development of these motor complications to various presynaptic and postsynaptic neuronal changes in the striatum produced by the loss of dopamine terminals and the L-DOPA treatment (e.g., [6][7][8][9][10]). For example, presynaptic changes include large pulsatile fluctuations in striatal dopamine levels following intermittent L-DOPA treatment due to disrupted control of dopamine release and clearance [11][12][13][14]. Studies have revealed that these non-physiological extracellular dopamine levels seen after L-DOPA administration originate from serotonin terminals in the striatum, which, following severe dopamine denervation, become the primary source for dopamine release [14][15][16][17][18][19].
One drug of recent interest is vilazodone, an antidepressant approved by the U.S. Food and Drug Administration (FDA). Vilazodone combines SSRI properties with 5-HT1A receptor partial agonist activity [49][50][51]. A recent study found that vilazodone co-administered with L-DOPA was sufficient to significantly attenuate dyskinesia scores, without interfering with the prokinetic benefits of L-DOPA [52]. At least some of these antidyskinetic effects were mediated by the 5-HT1A partial agonist property of vilazodone [52].
In the present study, we used the unilateral 6-hydroxydopamine (6-OHDA) lesion model in the rat to further investigate the behavioral and gene regulation effects of vilazodone co-administered with L-DOPA. For one, we focused on the expression of two neuropeptide markers-dynorphin and enkephalin-which are selectively expressed by dMSNs and indirect pathway (striatopallidal) MSNs (iMSNs), respectively [53], as a means to determine the projection neuron subtypes impacted by this treatment. We also assessed the effects of these drugs on the expression of a transcription factor, zif268, and the 5-HT1B serotonin receptor [33], both of which are regulated by dopamine agonists. Changes in gene regulation were mapped throughout the striatum in order to identify the functional domains affected. Our findings demonstrate that a dose of vilazodone that suppresses all major subtypes of L-DOPA-induced dyskinesia in this PD model inhibited L-DOPA-induced increases in dynorphin, zif268, and 5-HT1B expression, but did not affect enkephalin expression, in the dopamine-depleted striatum, or either gene in the intact striatum.

Animals
Adult male Sprague-Dawley rats (225-249 g at the beginning of the study; Harlan, Indianapolis, IN, USA) were housed 2-3 per cage under standard laboratory conditions (12:12 h light/dark cycle, lights on at 07:00 h; with food and water available ad libitum). Experiments were performed between 13:00 and 17:00 h. All procedures met the NIH guidelines for the care and use of laboratory animals and were approved by the Rosalind Franklin University Animal Care and Use Committee (protocol # 17-05; approved 19 April 2017).

Forelimb Stepping Test
The effectiveness of the 6-OHDA lesion was evaluated 4 weeks post-surgery by assessing deficits in forelimb movements using a stepping test [55]. In this test, the rat held by an experimenter is moved sideways, with its forelimb on the side opposite to the movement direction touching the bench surface; the rat will perform adjusting steps during this lateral movement, in our settings, typically 10-14 steps. After dopamine depletion, the number of adjusting steps with the forelimb contralateral to the lesion drops to 3 steps or fewer in animals with a >90% dopamine cell loss in the substantia nigra [56,57]. The present study included only 6-OHDA-infused rats that exhibited 3 or fewer adjusting steps. The lesion was further characterized by measuring tyrosine hydroxylase (TH) immunoreactivity in the striatum.

Behavioral Analysis
In order to assess the prokinetic effects of L-DOPA treatment, the stepping test was performed on the second day (Tue) of weeks 1 and 2, 60 min after L-DOPA administration. L-DOPA-induced dyskinesias (measured as abnormal involuntary movements, AIMs [58]) were evaluated once daily on three days per week (Wed-Fri) during the first 2 weeks [59]. Rats were individually placed in a clear plastic cylinder, and the behavior was videotaped. AIMs were rated by a rater unaware of the treatment using a time-sampling procedure, i.e., during 1-min periods at 30-min intervals 30-180 min after the L-DOPA injection. Three AIM subtypes were assessed, "axial", "limb" (forelimb), and "orolingual" ( [58], for exact definitions). The frequency of these AIM subtypes was rated using a standard scale (0 = absent; 1 = occasional; 2 = frequent with many interruptions; 3 = continuous but interrupted by external sensory stimuli; 4 = continuous, not interrupted by strong sensory stimuli) [31,58]. In addition, the amplitude of AIMs was scored as follows [58]: axial AIMs (1 = lateral deviation of head and neck at approximately 30 • angle; 2 = lateral deviation of head and neck at 30 • < angle ≤ 60 • ; 3 = lateral deviation of the head, neck, and upper trunk at 60 • < angle ≤ 90 • ; 4 = torsion of the head, neck, and trunk at > 90 • angle, often causing the rat to lose balance), forelimb AIMs (1 = small involuntary movements of the distal forelimb; 2 = movements of low amplitude, causing translocation of both distal and proximal forelimb; 3 = involuntary movements of the whole limb, including shoulder muscles; 4 = strong limb and shoulder movements, often similar to ballism), and orolingual AIMs (1 = small involuntary movements of the orofacial muscles without tongue protrusions; 2 = involuntary orofacial movements of high amplitude with tongue protrusions). Partial scores, such as 0.5, 1.5, 2.5, and 3.5, were given to increase the sensitivity of the rating. Frequency and amplitude scores were multiplied for each monitoring period (i.e., 30, 60, 90, 120, 150, and 180 min), and the values were added up for a total AIM score for each subtype and an overall total for each testing day.
Turning behavior in the open-field test on the last day (week 3) was assessed during 4 sampling periods, at 5-10, 15-20, 25-30, 35-40 min after L-DOPA injection. Tight circling contraversive to the lesion (i.e., towards the left side) and ipsiversive to the lesion was measured (counted as the number of half-turns).

Analysis of Autoradiograms
Gene expression in the striatum was assessed in sections from 3 rostrocaudal levels ("rostral", approximately at +1.6 mm relative to bregma [54]; "middle", +0.4; "caudal", −0.8), in a total of 23 sectors. These sectors are mostly defined by their predominant cortical inputs and thus reflect different functional domains ( [61,62]). Eighteen of these sectors represent the caudate-putamen and 5 the nucleus accumbens.
Hybridization signals on film autoradiograms were measured by densitometry (NIH Image; Wayne Rasband, NIMH, Bethesda, MD, USA), as described [63]. Mean densities were corrected for background by subtracting mean density values measured over white matter (corpus callosum). The illustrations of film autoradiograms depicted in the figures are computer-generated images and are contrast-enhanced where necessary.

Tyrosine Hydroxylase Immunohistochemistry
Striatal sections were processed for TH immunohistochemistry to determine the extent of the dopamine depletion, following previously published procedures [33,60,64]. In short, fresh-frozen, thaw-mounted 12 µm sections were first fixed in 4% paraformaldehyde/saline for 10 min. TH immunolabeling was then revealed with a rabbit peroxidase-antiperoxidase (1:500) reaction, followed by a standard 3,3 -diaminobenzidine protocol with nickel intensification. The signal was measured by densitometry.

Statistics
Treatment effects on gene expression, TH, and stepping were determined by two-factor ANOVA (Prism 8.0, GraphPad, San Diego, CA, USA). Newman-Keuls post hoc tests were used to describe differences between individual groups. Effects on AIMs and turning were assessed with Mann-Whitney tests with alpha adjustments.
During the open-field test in week 3, treatment with L-DOPA alone induced high rates of tight contraversive turning (6-OHDA/Veh/LD, half-turns, mean ± SEM, total in the four 5-min sampling periods: 466.3±47.8). This L-DOPA-induced turning behavior emerged between 8 and 20 min after L-DOPA injection and was always accompanied (and often preceded) by the various types of AIMs (not measured). Vilazodone co-administration suppressed contraversive turning ( Figure 1D, left). During the 40-min open-field test, rats of the 6-OHDA/VIL/LD group displayed no or very low counts of contraversive turning (total half-turns: 6.1±3.5 vs. 466.3±47.8 in the 6-OHDA/Veh/LD group; p < 0.001; data not shown). However, towards the end of their 60-min survival time (in their holding cage and especially during the final CO 2 exposure), some rats did show tendencies for contraversive turning and, most noticeable, contraversive forelimb AIMs. This turning seemed to have been triggered or enhanced by the handling when rats were transferred back to their holding cage and during the CO 2 exposure (arousal/stress).

Effects of Dopamine Depletion and L-DOPA Treatment on Dynorphin and Enkephalin Expression
The increase was minor in medial and ventral striatal sectors and was maximal in dorsal and lateral ("sensorimotor") sectors on all three rostrocaudal levels (Figures 2 and 5). This effect was also minimal or absent in sectors of the nucleus accumbens (pooled sectors, "limbic"; Figure 5). Therefore, the most pronounced L-DOPA-induced increase in dynorphin expression in the sensorimotor striatum markedly contrasted with the fairly uniform distribution of dopamine depletion (loss of TH signal) and dopamine depletion only-induced changes in dynorphin expression (Figures 1 and 2), confirming our earlier findings [33].
The 6-OHDA lesion alone (6-OHDA/Veh/Veh vs. Sham/Veh/Veh) produced significantly increased enkephalin expression in 18 of the 23 striatal sectors on all three rostrocaudal levels, with lesser effects in ventral sectors (Figures 3 and 5). Repeated L-DOPA treatment (6-OHDA/Veh/LD vs. 6-OHDA/Veh/Veh) caused a modest, but statistically significant, further increase in enkephalin expression in nine of these sectors, an effect that was most robust in pooled "associative" sectors ( Figure 5).

Effects of Vilazodone on Dynorphin and Enkephalin Expression after Dopamine Depletion and L-DOPA Treatment
Vilazodone strongly attenuated the L-DOPA-induced increase in dynorphin expression throughout the striatum. Thus, the vilazodone + L-DOPA-treated group displayed significantly lower levels of dynorphin mRNA compared to the L-DOPA only-treated group (6-OHDA/VIL/LD vs. 6-OHDA/Veh/LD) in 16 of the 23 sectors (Figures 2 and 5). However, vilazodone did not completely eliminate the L-DOPA-induced increase in dynorphin expression, as the vilazodone + L-DOPA-treated group showed significantly higher levels of dynorphin expression than the 6-OHDA lesion-only group (6-OHDA/VIL/LD vs. 6-OHDA/Veh/Veh) in 4 of the 23 sectors (Figures 2 and 5). This was observed in the dorsolateral sector on the rostral level (not shown) and the dorsal, dorsolateral, and ventrolateral sectors on the middle level (Figure 2), that is, in the sectors with the greatest L-DOPA-induced increase in dynorphin expression (sensorimotor sectors; Figure 5). Indeed, vilazodone normalized dynorphin expression in that dynorphin mRNA levels were similar (p > 0.05) to those in the sham lesion controls (6-OHDA/VIL/LD vs. Sham/Veh/Veh) in 22 of the 23 striatal sectors.
In contrast to the strong impact of vilazodone on L-DOPA-induced dynorphin expression (see above), vilazodone alone had no significant effect on dynorphin expression, either on the lesion side (6-OHDA/VIL/Veh vs. 6-OHDA/Veh/Veh) or on the intact side (Figures 2 and 5).
In contrast to dynorphin expression, enkephalin expression was minimally or not affected by vilazodone, either on the lesion side or on the intact side (Figures 3 and 5). For example, on the lesion side, enkephalin expression did not differ significantly (p > 0.05) between the vilazodone + L-DOPA-treated group and the L-DOPA only-treated group (6-OHDA/VIL/LD vs. 6-OHDA/Veh/LD) in any but one of the 23 sectors (Figures 3 and 5). Also in contrast to dynorphin expression, there was a tendency for increased enkephalin expression after vilazodone-only treatment (6-OHDA/VIL/Veh vs. 6-OHDA/Veh/Veh; Figure 3) on the lesion side (borderline significant in six sectors and thus more robust after pooling sectors; Figure 5). Again, no effects on enkephalin expression were seen on the intact side (Figures 3 and 5).

Effects of Vilazodone on 5-HT1B Receptor Expression after Dopamine Depletion and L-DOPA Treatment
Consistent with our previous findings [33], dopamine depletion and repeated L-DOPA treatment produced differential changes in 5-HT1B mRNA expression in the striatum ipsilateral to the lesion. The lesion effects were widespread, present on all three rostrocaudal levels (for effects on the middle level, Figure 4; rostral and caudal effects are not shown). After dopamine depletion alone (6-OHDA/Veh/Veh vs. Sham/Veh/Veh), there was a statistically significant increase in 5-HT1B expression in 10 of the 23 sectors, with lesser or no effects in ventral sectors, including the nucleus accumbens (Figures 4 and 5).
Repeated L-DOPA treatment produced a pronounced further increase in 5-HT1B expression (Figure 4). This increase was seen in comparison to lesion-only animals (6-OHDA/Veh/LD vs. 6-OHDA/Veh/Veh) in 15 of the 23 sectors and to sham lesion controls (6-OHDA/Veh/LD vs. Sham/Veh/Veh) in 18 sectors, on all three rostrocaudal levels. Regionally, these effects were again strongest in the dorsal and lateral (sensorimotor) striatum and weakest in the medial and ventral striatum, including the nucleus accumbens (Figures 4 and 5) and, thus, matching the distribution of the L-DOPA-induced increase in dynorphin expression (Figures 3 and 5).
Again similar to dynorphin and in contrast to the pronounced impact of vilazodone on L-DOPA-induced 5-HT1B expression (see above), vilazodone alone had no significant effect on 5-HT1B expression, neither on the lesion side (6-OHDA/VIL/Veh vs. 6-OHDA/Veh/Veh) (with the exception of one sector; not shown) nor on the intact side (Figures 4 and 5).

Effects of Vilazodone on Zif268 Expression after Dopamine Depletion and L-DOPA Treatment
The treatment effects on zif268 expression in the middle striatum are shown in Figure 6 (data from rostral and caudal levels were similar and are not shown). L-DOPA produced a pronounced increase in zif268 expression, an effect that was fairly uniform throughout the striatum ( Figure 6). Vilazodone attenuated but did not eliminate the L-DOPA-induced increase in zif268 expression ( Figure 6). Thus, the vilazodone effect on L-DOPA-induced zif268 expression mimicked that on dynorphin and 5-HT1B expression.

Discussion
L-DOPA-induced dyskinesia is associated with aberrant neuronal activity and altered gene regulation in striatal projection neurons produced by the L-DOPA treatment. These changes in gene regulation are thought to be driven by supraphysiological dopamine release from serotonin terminals. The present study used a rat model of L-DOPA-induced dyskinesia to investigate the effects of vilazodone, an SSRI/5-HT1A receptor partial agonist, on L-DOPA-induced changes in gene regulation in striatal projection neurons and behavior. Our main findings showed that a dose of vilazodone that suppressed the development of L-DOPA-induced dyskinesia also attenuated L-DOPA-induced gene regulation in direct pathway neurons (dMSNs), as demonstrated by the inhibition of the L-DOPA-induced increase in expression of the cell type marker dynorphin. Vilazodone also inhibited L-DOPA-induced expression of 5-HT1B and zif268, which also occurs in dMSNs. In contrast, no effects of vilazodone on the expression of enkephalin (a marker for iMSNs), nor on gene regulation in the intact striatum, were found. Overall, these results indicate that vilazodone co-administration selectively attenuated L-DOPA-induced gene regulation in dMSNs, and suggest that the antidyskinetic effects of vilazodone are mediated via preventing such L-DOPA-induced pathophysiological changes in the direct pathway neurons.

Vilazodone Attenuates L-DOPA-Induced Dyskinesia Without Inhibiting Its Prokinetic Effects
A number of previous studies have found that treatments with SSRIs and 5-HT1A agonists attenuate L-DOPA-induced dyskinesia in animal models (e.g., [18,[36][37][38][39][40]). As these agents temper serotonin release, it is taken that at least part of this behavioral effect is due to attenuation of abnormal dopamine release from serotonin terminals [14,19,65]. However, some of these agents, especially at higher doses, also impair therapeutic (prokinetic) effects of L-DOPA or have other side effects [41,42,47], questioning the usefulness of this pharmacological approach. In contrast, a recent study showed that vilazodone, which combines SSRI and 5-HT1A partial agonist properties [49][50][51], given in conjunction with L-DOPA inhibited the development and expression of L-DOPA-induced dyskinesia while leaving L-DOPA's prokinetic effects intact, as assessed by the forepaw stepping test [52].
Our present findings confirm and extend these previous results. We demonstrated that the dose of 10 mg/kg of vilazodone, which displayed a 100% occupancy of serotonin transporter sites in several brain regions [49], almost completely prevented the development of dyskinesia (measured as AIMs) during the course of a three-week, 13-injection L-DOPA treatment. Moreover, our results also showed that vilazodone suppressed all major subtypes of L-DOPA-induced dyskinesia in this rat model (axial, forelimb, orolingual AIMs), as well as L-DOPA-induced contraversive turning behavior. Importantly, also consistent with previous results [52], this vilazodone treatment did not impair the improved forepaw stepping, a simple motor response that is sensitive to dopamine loss and replacement by L-DOPA [55,66]. It has been shown that, while eliminating the dopamine spikes that produce L-DOPA-induced dyskinesia, this vilazodone + L-DOPA treatment elevates dopamine levels in the denervated striatum to some degree [52], thus presumably facilitating the execution of this simple behavior.
However, our findings also showed that more complex sensorimotor behavior (orienting to a novel environment [67]) was not normalized. The vilazodone + L-DOPA-treated animals still showed marked turning asymmetries in direction ipsiversive to the lesion in the open-field test, which reflects a dopamine denervation-induced sensorimotor neglect on the body surface contralateral to the dopamine lesion (e.g., [60,67,68]). Thus, while simple motor responses, such as stepping, were improved by this treatment, more complex sensorimotor integration might remain impaired.
Consistent with our previous findings [33] and those of others (e.g., [31,32]), our results showed decreased expression of dynorphin after the 6-OHDA lesion alone and markedly increased expression after the subsequent repeated L-DOPA treatment. As described before [33], the L-DOPA-induced increase in dynorphin expression was widespread throughout the striatum but displayed distinct regional variations. The increase was most pronounced in the sensorimotor striatum and was lower or minimal in medial (associative) and ventral (limbic) sectors. Previous work showed that this distinct pattern of sensorimotor-dominant gene regulation after L-DOPA treatment is not specific for dynorphin, but extends to other genes as well, for example, the transcription factors dFosB and JunB [32,73], the serotonin receptor 5-HT1B ( [33]; present study), and other genes [74,75]. These findings thus highlight the close association between abnormal signaling in dMSNs of the sensorimotor striatum and dyskinesia, in general [76][77][78][79].
Our findings demonstrate that vilazodone inhibited the L-DOPA-induced increase in striatal dynorphin expression. This result is in agreement with a recent study using RT-PCR to measure the effects of vilazodone on striatal dynorphin expression (in total striatal tissue [52]), as well as studies investigating other antidyskinetic agents [44,80]. Our present mapping study showed that this effect occurred throughout the striatum. It was most striking in the sensorimotor striatum but was also present in the associative striatum, which, although less affected by L-DOPA, nevertheless appears to contribute to L-DOPA-induced dyskinesia (for discussion [33]). The associative striatum has also been linked to non-motor side effects of L-DOPA treatment, such as impulsive/compulsive behaviors and visual hallucinations [81]. Attenuated gene regulation by vilazodone in both the sensorimotor and associative striatum may thus be essential to effectively reduce various L-DOPA-induced adverse effects.
Interestingly, after vilazodone + L-DOPA treatment, dynorphin mRNA levels were not entirely reduced to lesion-only levels, but were similar to normal levels (sham lesion controls) and, in this sense, were "normalized" (most clearly seen in the sensorimotor striatum). It is unclear whether a higher dose of vilazodone would have prevented increased dynorphin expression entirely. It is interesting to speculate that this normalization of gene regulation reflected the (partially) increased dopamine steady-state levels [52] and normalized corticostriatal synaptic activation (Altwal et al., in preparation) produced by this vilazodone dose, thus enabling some normalization of behavior in addition to preventing L-DOPA-induced dyskinesia.

Vilazodone Does Not Inhibit Enkephalin Expression (iMSNs)
Another well-established gene regulation effect of dopamine depletion is increased enkephalin expression in iMSNs (e.g., [69,71,[82][83][84]; for review, [53]). L-DOPA tends to further increase enkephalin expression, but our direct comparison showed that L-DOPA had a more modest effect on enkephalin expression than on dynorphin expression (e.g., [33]; present results). In our previous study, we found a minor, but statistically significant, further increase in enkephalin expression in several striatal sectors, after repeated L-DOPA treatment [33]. Here, we obtained a similar but less robust effect, perhaps due to the shorter L-DOPA treatment in the present study (13 injections over 3 weeks) compared to the previous study (18 injections over 4 weeks [33]).
Our results showed that, in marked contrast to dynorphin, vilazodone did not affect enkephalin expression after dopamine depletion + L-DOPA treatment. Again in contrast to dynorphin expression, after dopamine depletion alone, vilazodone tended to increase enkephalin mRNA levels, at least in the limbic and associative striatum, which displayed a lesser increase in enkephalin expression by the dopamine depletion.
The reasons for these differential effects remain unclear. Gene regulation of enkephalin is more complex than that of dynorphin, with a significant impact by cholinergic activity [85] and an antagonistic regulation by cortical glutamate vs. striatal dopamine ( [86]), which each may be under distinct serotonin receptor control. Besides, it has been shown that continuous (D2) dopamine receptor stimulation is required to reverse dopamine lesion-induced enhanced enkephalin expression (e.g., [69,83]), while intermittent (D1) dopamine receptor stimulation is sufficient to reverse changes in dynorphin expression (e.g., [69]).
Irrespective of the underlying mechanisms, our findings demonstrate that vilazodone had a markedly differential impact on gene regulation in dMSNs vs. iMSNs and that attenuation of the former was associated with the suppression of L-DOPA-induced dyskinesia. Thus, these results support the notion that vilazodone's antidyskinetic properties are mediated by direct pathway-dependent mechanisms (reducing aberrant dMSN signaling), in the dopamine-depleted striatum. However, given that enkephalin expression (iMSN signaling) was not affected (not normalized), it is not surprising that vilazodone did not appear to completely normalize more complex behavior (sensorimotor integration).

Vilazodone Inhibits L-DOPA-Induced Increases in 5-HT1B and Zif268 Expression
The effects of vilazodone on 5-HT1B and zif268 expression support the above conclusion of a selective impact of this serotonergic agent on gene regulation in dMSNs. The 5-HT1B serotonin receptor subtype is expressed in both types of striatal projection neurons-dMSNs and iMSNs [29]. This receptor is dynamically regulated by a number of pharmacological treatments and experimental manipulations, for example, dopamine depletion and treatments with dopamine agonists, including psychostimulants and L-DOPA (cf. [33]). There is evidence that dopamine depletion and L-DOPA treatment differentially upregulate 5-HT1B in dMSNs vs. iMSNs [29]. Similar to enkephalin, dopamine depletion seems to preferentially increase 5-HT1B expression in iMSNs [29], with a rather uniform regional distribution, matching that of increased enkephalin expression ( [33]; present results). On the other hand, similar to dynorphin, repeated L-DOPA treatment after dopamine depletion produces increased 5-HT1B expression selectively in dMSNs [29], with a regional distribution (sensorimotor dominant) that matches that of increased dynorphin expression ( [33]; present results). Our previous findings showed that L-DOPA-induced increases in 5-HT1B expression correlated with the magnitude of L-DOPA-induced dyskinesia [33]. Underscoring this association between striatal gene regulation and behavior, our present results demonstrate that a dose of vilazodone that suppressed the development of L-DOPA-induced dyskinesia inhibited this L-DOPA-induced expression of 5-HT1B, presumably in dMSNs [29], while not inhibiting the lesion-induced increase in 5-HT1B, presumably in iMSNs.
Our findings on zif268 expression support a preferential effect of this serotonergic agent on dMSNs. Zif268 is a transcription factor (an immediate-early gene) that is induced by various dopamine agonists, including L-DOPA, and plays a role in several forms of neuroplasticity (cf. [86]). After dopamine depletion, dopamine receptor agonists dramatically induce zif268, selectively in dMSNs throughout the striatum [87], a reflection of supersensitive D1 receptor signaling. Our results demonstrate that vilazodone attenuated this abnormal zif268 response in the dopamine-depleted striatum. These findings thus also indicate a preferential impact of vilazodone on gene regulation in the D1 receptor-regulated direct pathway.

Possible Mechanisms Underlying the Effects of Vilazodone
The mechanisms by which vilazodone elicits its antidyskinetic action seem to involve a reduction of aberrant dMSN signaling following L-DOPA treatment. The attenuated increase in dynorphin, 5-HT1B, and zif268 expression is consistent with a vilazodone-induced inhibition of L-DOPA-derived dopamine release from striatal serotonin terminals.
Studies have shown that the serotonergic innervation of the striatum is essential for the development of L-DOPA-induced dyskinesia [13,16,18,88]. As monoaminergic cells, serotonin neurons contain the necessary cellular mechanisms to take up L-DOPA, convert it to dopamine, package it into vesicles, and then release it [15,89]. In advanced PD with a severe loss of striatal dopamine terminals, it is thought that serotonin neurons become the main source for dopamine release in the striatum. However, serotonin neurons lack the feedback control mechanism that is vital to regulate the release of dopamine (i.e., D2 autoreceptors), resulting in extreme fluctuations in extracellular dopamine levels following administration of L-DOPA (e.g., [12]).
Similarly, vilazodone seems to convey its antidyskinetic effects, at least in part, via a 5-HT1A receptor-dependent mechanism, as blocking 5-HT1A receptors using a selective 5-HT1A receptor antagonist (WAY100635) inhibits the ability of vilazodone to reduce AIMs [52]. Together, these findings suggest that the antidyskinetic effects of vilazodone may be mediated through its direct and indirect activation of 5-HT1A autoreceptors, which would reduce the activity of serotonin neurons innervating the striatum and temper their dopamine release, thus reducing stimulation of supersensitive D1 receptors and concomitant abnormal molecular signaling.
However, vilazodone was administered systemically, which precludes conclusions regarding the sites of action; other mechanisms could thus also be involved. Serotonin receptors have a wide distribution throughout the brain [92] and modulate activity in various nodes of cortico-basal ganglia-thalamocortical circuits. For example, the cortex receives a prominent serotonin innervation, and cortical neurons express several 5-HT receptor subtypes (e.g., 5-HT1A, 5-HT2A, and others) (reviewed in [93]). Increased extracellular serotonin levels via the SSRI action of vilazodone may thus also modulate the excitability of cortical neurons by activating these 5-HT heteroreceptors. Given that aberrant facilitation of corticostriatal synaptic transmission is considered a major contributor to L-DOPA-induced dyskinesia (reviewed in [94]), moderating cortical neuronal activity, or glutamate release from corticostriatal terminals via 5-HT1A receptors located on these terminals [44], could be additional mechanisms underlying the vilazodone effects on dyskinesia.
Indeed, increased corticostriatal input (Altwal et al., in preparation) likely contributes to the increased striatal gene regulation associated with dyskinesia. We have hypothesized before [33] that the striking increase in dynorphin (and 5-HT1B) expression predominantly in the sensorimotor striatum (which markedly contrasts with the uniform loss of dopamine terminals/tyrosine hydroxylase throughout the striatum [33], present results) may at least in part reflect enhanced re-entrant corticostriatal input, as a consequence of abnormally enhanced D1 receptor-regulated striatal output in the direct pathway [95] and enhanced thalamocortical activity ( [33], for discussion). Thus, 5-HT receptor-mediated dampening of cortical input may have contributed to vilazodone-induced inhibition of striatal gene regulation.
However, striatal gene regulation is under considerable influence of cortical glutamate also in the intact brain [86], yet vilazodone alone had no clear effect on striatal gene regulation, either in the dopamine-depleted or intact striatum. Therefore, a major contribution of dampened glutamate release per se to vilazodone-induced attenuation of dMSN gene regulation and dyskinesia is questionable, and interaction with abnormal dopamine release seems more likely. Further studies will have to investigate the specific mechanisms targeted by vilazodone.

Conclusions
Supraphysiological dopamine release from serotonin terminals in the striatum seems to drive L-DOPA-induced dyskinesia via abnormal gene regulation in striatal direct pathway neurons. In animal models, SSRIs and 5-HT1A receptor agonists independently have had beneficial antidyskinetic effects when given in conjunction with L-DOPA, presumably via tempering of such dopamine release; however, such agents can interfere with L-DOPA-induced prokinetic effects or may have other unwanted side effects (see above). Vilazodone, an FDA-approved antidepressant, has a unique pharmacological profile, combining SSRI and 5-HT1A receptor partial agonist properties. Our results in a rat model of PD show that vilazodone attenuates the development of L-DOPA-induced dyskinesia and selectively inhibits L-DOPA-induced aberrant gene regulation in striatal direct pathway neurons. These findings situate this multimodal serotonergic drug as a promising candidate for improving L-DOPA treatment and alleviate motor dysfunction in patients with PD.