Evaluation of Substituted N-Phenylpiperazine Analogs as D3 vs. D2 Dopamine Receptor Subtype Selective Ligands

N-phenylpiperazine analogs can bind selectively to the D3 versus the D2 dopamine receptor subtype despite the fact that these two D2-like dopamine receptor subtypes exhibit substantial amino acid sequence homology. The binding for a number of these receptor subtype selective compounds was found to be consistent with their ability to bind at the D3 dopamine receptor subtype in a bitopic manner. In this study, a series of the 3-thiophenephenyl and 4-thiazolylphenyl fluoride substituted N-phenylpiperazine analogs were evaluated. Compound 6a was found to bind at the human D3 receptor with nanomolar affinity with substantial D3 vs. D2 binding selectivity (approximately 500-fold). Compound 6a was also tested for activity in two in-vivo assays: (1) a hallucinogenic-dependent head twitch response inhibition assay using DBA/2J mice and (2) an L-dopa-dependent abnormal involuntary movement (AIM) inhibition assay using unilateral 6-hydroxydopamine lesioned (hemiparkinsonian) rats. Compound 6a was found to be active in both assays. This compound could lead to a better understanding of how a bitopic D3 dopamine receptor selective ligand might lead to the development of pharmacotherapeutics for the treatment of levodopa-induced dyskinesia (LID) in patients with Parkinson’s disease.


Introduction
Dopamine is a neurotransmitter that controls a variety of physiological functions in both the periphery and in the central nervous system (CNS). In the periphery, dopamine receptors have been implicated in the modulation of renal sodium excretion, urinary output, inhibition of norepinephrine release, vasodilatory activity of blood vessels and the regulation of insulin secretion in pancreatic β-cells [1,2]. Dopamine receptor expression has also been reported to be expressed on regulatory T cells and antigen-presenting dendritic cells associated with the immune system [3,4]. Dopaminergic pathways in the brain have been implicated in the reward system, psychostimulant seeking behaviors, movement, memory and learning, sleep regulation, feeding, attention, olfaction, vision, hormonal regulation, and emotional tone [5].
There are five human dopamine receptor subtypes classified into two major categories: D1-like (D1 and D5) and D2-like (D2, D3 and D4) dopamine receptor subtypes. The D1-like and D2-like dopamine receptor classes have been categorized based upon similarities in genomic organization, amino acid sequence homology and pharmacological properties [6]. All five of the dopamine receptor subtypes are members of the G protein-coupled receptor (GPCR) protein superfamily. The two genes which encode for the D1 and D5 are both  [19,20,26,27,30].
The synthesis strategy for the compounds described in this communication is outlined in Scheme 1. Briefly, the orthosteric N-phenylpiperazine moieties 2a-f were afforded by a one-pot Pd-catalyzed Buchwald-Hartwig amination reaction of six different aryl halides with piperazine [34]. The N-alkylation of aryl piperazines 2a-f with 1-bromobutane in acetone using potassium carbonate as a base provided fragments of D3 dopamine receptor selective compounds, 3a-f. Butyl amine compounds 5a-f, which were used for conjugation with substituted benzoic acids, were synthesized using a Gabriel amine synthesis reaction. N-arylpiperazine compounds 2a-f were reacted with N-(4-bromobutyl)phthalimide in acetonitrile to give butyl phthalimide compounds 4a-f and the hydrazinolysis of these compounds gave the corresponding amines 5a-f in high yields (70-95%). The free amines 5a-f were coupled with 4-(thien-3-yl)benzoic acid or 4-(1,3-thiazol-4-yl)benzoic acid using 1ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and 1-hydroxybenzotriazole (HOBt) hydrate in dichloromethane to give the targeted benzamide compounds 6a-f (52-88% yields) and 7a-f (64-84% yields). The final products were converted to the corresponding hydrochloride salts. Detailed synthesis procedure and chemical analysis of synthesized compounds ( 1 H, 13 C NMR, and Mass) were reported in Supplementary Material.

5-HT1A Receptor Binding
Because the 5-HT1A receptor is a commonly observed off-site binding site for benzamide N-phenylpiperazine ligands, compounds were simultaneously screened for binding of our compounds at the human 5-HT1A receptor expressed in CHO-K1 cells, using the radioligand [ 3 H]-8-hydroxy-DPAT (8-OH-DPAT) ( Table 2). Based upon the analysis of the displacement assay shown in Table 2, the orthosteric compounds 3c-f appeared to bind with low affinity at human 5-HT1A receptors. However, at a concentration of 91 nM compounds 3a and 3b appeared to be able to partially displace the 5-HT1A selective radioligand. Compounds 6b, 7a, 7b and 7f displaced tritiated 8-OH-DPAT binding in a manner that suggested substantial affinity for the 5-HT1A receptor.

5-HT1A Receptor Binding
Because the 5-HT1A receptor is a commonly observed off-site binding site for benzamide N-phenylpiperazine ligands, compounds were simultaneously screened for binding of our compounds at the human 5-HT1A receptor expressed in CHO-K1 cells, using the radioligand [ 3 H]-8-hydroxy-DPAT (8-OH-DPAT) ( Table 2). Based upon the analysis of the displacement assay shown in Table 2, the orthosteric compounds 3c-f appeared to bind with low affinity at human 5-HT1A receptors. However, at a concentration of 91 nM compounds 3a and 3b appeared to be able to partially displace the 5-HT1A selective radioligand. Compounds 6b, 7a, 7b and 7f displaced tritiated 8-OH-DPAT binding in a manner that suggested substantial affinity for the 5-HT1A receptor.

5-HT1A Receptor Binding
Because the 5-HT1A receptor is a commonly observed off-site binding site for benzamide N-phenylpiperazine ligands, compounds were simultaneously screened for binding of our compounds at the human 5-HT1A receptor expressed in CHO-K1 cells, using the radioligand [ 3 H]-8-hydroxy-DPAT (8-OH-DPAT) ( Table 2). Based upon the analysis of the displacement assay shown in Table 2, the orthosteric compounds 3c-f appeared to bind with low affinity at human 5-HT1A receptors. However, at a concentration of 91 nM compounds 3a and 3b appeared to be able to partially displace the 5-HT1A selective radioligand. Compounds 6b, 7a, 7b and 7f displaced tritiated 8-OH-DPAT binding in a manner that suggested substantial affinity for the 5-HT1A receptor. A comparison of the binding affinity and binding selectivity of the N-phenylpiperazine orthosteric ligands (compounds 3a-f) with the longer thiophene-containing (compounds 6a-f) and thiazole-containing (compounds 7a-f) analogs further suggest that the benzamide N-phenylpiperazine class of compound likely engage the D3 dopamine receptor, but not the D2 dopamine receptor, in a bitopic binding mode, leading to enhanced affinity at the D3 receptor subtype and consequently increased D3 vs. D2 binding selectivity. In general, the thiophene analogs (compounds 6a-f) bound with slightly higher affinity at D3 dopamine receptors and/or with slightly greater D3 vs D2 binding selectivity than the corresponding thiazole analogs (compounds 7a-f).

5-HT1A Receptor Binding
Because the 5-HT1A receptor is a commonly observed off-site binding site for benzamide N-phenylpiperazine ligands, compounds were simultaneously screened for binding of our compounds at the human 5-HT1A receptor expressed in CHO-K1 cells, using the radioligand [ 3 H]-8-hydroxy-DPAT (8-OH-DPAT) ( Table 2). Based upon the analysis of the displacement assay shown in Table 2, the orthosteric compounds 3c-f appeared to bind with low affinity at human 5-HT1A receptors. However, at a concentration of 91 nM compounds 3a and 3b appeared to be able to partially displace the 5-HT1A selective radioligand. Compounds 6b, 7a, 7b and 7f displaced tritiated 8-OH-DPAT binding in a manner that suggested substantial affinity for the 5-HT1A receptor.   Phenylpiperazine ligands 3a-f, 6a-f, and 7a-f were screened for the binding activity of [ 3 H]-8-hydroxy-DPAT at human 5-HT1A receptors expressed in CHO-K1 cells using a single point radioligand displacement assay. The radioligand displacement percentages were determined by at least three experiments using 91 nM concentration of test compound. Nonspecific binding was defined using 10 μM metergoline. Compounds 3a and 3b exhibited moderate displacement of the radioligand at the 5-HT1A receptor. Thiophenyl compound 6b and thiazolyl compounds 7a, 7b, and 7f exhibited a greater percentage of radioligand displacement. Data is presented as percent displacement of radioligand (n ≥ 3) ± S.E.M.
A comparison of the effect of compounds 3a, 6a and 7a on [ 3 H]-8-hydroxy-DPAT binding was conducted. Based upon the radioligand displacement assessment the orthosteric ligand (compound 3a) exhibited moderate affinity at 5-HT1A receptors. The thiazole analog of compound 3a (compound 7a) exhibited increased affinity at 5-HT1A, while the thiophene analog (compound 6a) appeared to exhibit decreased binding affinity at the 5-HT1A receptor compared to compound 3a.
Compound 6a bound at the D3 dopamine receptor with high affinity (Ki value = 1.4 ± 0.21 nM) and exhibited >400-fold D3 vs. D2 receptor binding selectivity, while exhibiting a low level of radioligand displacement at 5-HT1A receptors, suggesting that it bound with low affinity at the 5-HT1A receptor binding site. To verify the low affinity binding of compound 6a at the 5-HT1A receptor, a dose dependent competitive radioligand binding curve was performed. Compound 6a was found to bind to the 5-HT1A receptor with  Phenylpiperazine ligands 3a-f, 6a-f, and 7a-f were screened for the binding activity of [ 3 H]-8-hydroxy-DPAT at human 5-HT1A receptors expressed in CHO-K1 cells using a single point radioligand displacement assay. The radioligand displacement percentages were determined by at least three experiments using 91 nM concentration of test compound. Nonspecific binding was defined using 10 μM metergoline. Compounds 3a and 3b exhibited moderate displacement of the radioligand at the 5-HT1A receptor. Thiophenyl compound 6b and thiazolyl compounds 7a, 7b, and 7f exhibited a greater percentage of radioligand displacement. Data is presented as percent displacement of radioligand (n ≥ 3) ± S.E.M.
A comparison of the effect of compounds 3a, 6a and 7a on [ 3 H]-8-hydroxy-DPAT binding was conducted. Based upon the radioligand displacement assessment the orthosteric ligand (compound 3a) exhibited moderate affinity at 5-HT1A receptors. The thiazole analog of compound 3a (compound 7a) exhibited increased affinity at 5-HT1A, while the thiophene analog (compound 6a) appeared to exhibit decreased binding affinity at the 5-HT1A receptor compared to compound 3a.
Compound 6a bound at the D3 dopamine receptor with high affinity (Ki value = 1.4 ± 0.21 nM) and exhibited >400-fold D3 vs. D2 receptor binding selectivity, while exhibiting a low level of radioligand displacement at 5-HT1A receptors, suggesting that it bound with low affinity at the 5-HT1A receptor binding site. To verify the low affinity binding of compound 6a at the 5-HT1A receptor, a dose dependent competitive radioligand binding curve was performed. Compound 6a was found to bind to the 5-HT1A receptor with Phenylpiperazine ligands 3a-f, 6a-f, and 7a-f were screened for the binding activity of [ 3 H]-8-hydroxy-DPAT at human 5-HT1A receptors expressed in CHO-K1 cells using a single point radioligand displacement assay. The radioligand displacement percentages were determined by at least three experiments using 91 nM concentration of test compound. Nonspecific binding was defined using 10 μM metergoline. Compounds 3a and 3b exhibited moderate displacement of the radioligand at the 5-HT1A receptor. Thiophenyl compound 6b and thiazolyl compounds 7a, 7b, and 7f exhibited a greater percentage of radioligand displacement. Data is presented as percent displacement of radioligand (n ≥ 3) ± S.E.M.
A comparison of the effect of compounds 3a, 6a and 7a on [ 3 H]-8-hydroxy-DPAT binding was conducted. Based upon the radioligand displacement assessment the orthosteric ligand (compound 3a) exhibited moderate affinity at 5-HT1A receptors. The thiazole analog of compound 3a (compound 7a) exhibited increased affinity at 5-HT1A, while the thiophene analog (compound 6a) appeared to exhibit decreased binding affinity at the 5-HT1A receptor compared to compound 3a.
Compound 6a bound at the D3 dopamine receptor with high affinity (Ki value = 1.4 ± 0.21 nM) and exhibited >400-fold D3 vs. D2 receptor binding selectivity, while exhibiting a low level of radioligand displacement at 5-HT1A receptors, suggesting that it bound with low affinity at the 5-HT1A receptor binding site. To verify the low affinity binding of compound 6a at the 5-HT1A receptor, a dose dependent competitive radioligand binding curve was performed. Compound 6a was found to bind to the 5-HT1A receptor with Phenylpiperazine ligands 3a-f, 6a-f, and 7a-f were screened for the binding activity of [ 3 H]-8-hydroxy-DPAT at human 5-HT1A receptors expressed in CHO-K1 cells using a single point radioligand displacement assay. The radioligand displacement percentages were determined by at least three experiments using 91 nM concentration of test compound. Nonspecific binding was defined using 10 µM metergoline. Compounds 3a and 3b exhibited moderate displacement of the radioligand at the 5-HT1A receptor. Thiophenyl compound 6b and thiazolyl compounds 7a, 7b, and 7f exhibited a greater percentage of radioligand displacement. Data is presented as percent displacement of radioligand (n ≥ 3) ± S.E.M.
A comparison of the effect of compounds 3a, 6a and 7a on [ 3 H]-8-hydroxy-DPAT binding was conducted. Based upon the radioligand displacement assessment the orthosteric ligand (compound 3a) exhibited moderate affinity at 5-HT1A receptors. The thiazole analog of compound 3a (compound 7a) exhibited increased affinity at 5-HT1A, while the thiophene analog (compound 6a) appeared to exhibit decreased binding affinity at the 5-HT1A receptor compared to compound 3a.
Compound 6a bound at the D3 dopamine receptor with high affinity (K i value = 1.4 ± 0.21 nM) and exhibited >400-fold D3 vs. D2 receptor binding selectivity, while exhibiting a low level of radioligand displacement at 5-HT1A receptors, suggesting that it bound with low affinity at the 5-HT1A receptor binding site. To verify the low affinity binding of compound 6a at the 5-HT1A receptor, a dose dependent competitive radioligand binding curve was performed. Compound 6a was found to bind to the 5-HT1A receptor with a Ki value of 199 nM (Figure 2), which is approximately 140-fold lower than its affinity at the D3 dopamine receptor. The rank order of affinity of these analogs at the 5-HT1A receptor was found to be 7a > 3a > 6a (Figure 2). a Ki value of 199 nM (Figure 2), which is approximately 140-fold lower than its affinity at the D3 dopamine receptor. The rank order of affinity of these analogs at the 5-HT1A receptor was found to be 7a > 3a > 6a (Figure 2).  Table 2.

Ligand Efficacy Analysis
Compounds exhibiting high affinity binding at the D3 dopamine receptor were subsequently evaluated for efficacy at the D3 dopamine receptor subtype using both a forskolin-dependent inhibition of adenylyl cyclase assay and a β-arrestin binding assay (Figure 3). Quinpirole was used as a reference prototypic full agonist and haloperidol was used as a reference prototypic antagonist. Compounds 7e and 6d were found to be antagonists in both assays. Compound 6c was found to be functionally selective because it was  Table 2.

Ligand Efficacy Analysis
Compounds exhibiting high affinity binding at the D3 dopamine receptor were subsequently evaluated for efficacy at the D3 dopamine receptor subtype using both a forskolindependent inhibition of adenylyl cyclase assay and a β-arrestin binding assay (Figure 3). Quinpirole was used as a reference prototypic full agonist and haloperidol was used as a reference prototypic antagonist. Compounds 7e and 6d were found to be antagonists in both assays. Compound 6c was found to be functionally selective because it was a weak partial agonist for the adenylyl cyclase assay while being a very weak partial agonist/antagonist in the β-arrestin binding assay. Compounds 6a and 6b were found to be weak partial agonists in both the adenylyl cyclase inhibition assay and the β-arrestin binding assay. a weak partial agonist for the adenylyl cyclase assay while being a very weak partial agonist/antagonist in the β-arrestin binding assay. Compounds 6a and 6b were found to be weak partial agonists in both the adenylyl cyclase inhibition assay and the β-arrestin binding assay. Values were normalized to reference full agonist quinpirole (10,000 nM). The concentration of the test compound was equal to 10× the Ki value as determined from competitive radioligand binding studies, therefore 90% of the binding sites were occupied.

In-Vivo Analysis of Compound 6a
Based upon the data indicating that compound 6a bound with high affinity to the human D3 dopamine receptor and with low affinity at both the human D2 dopamine receptor subtype (Table 1) and the 5-HT1A receptor ( Table 2 and Figure 2), two in-vivo assays were performed to determine if compound 6a was capable of in-vivo engagement of the D3 receptor in the brain. While we do appreciate the potential that possible in-vivo differences in pharmacological responses for the same drug in experimental animals and humans of sex differences, these initial in-vivo studies were primarily conduced to assess the ability of compound 6a to cross the blood brain barrier (BBB) and engage the pharmacologically relevant target. Further sex differences in the pharmacological response of animals to compound 6a will be conduction as further in-vivo studies are necessitated.
First, we have previously published that D3 receptor selective compounds could attenuate a hallucinogen-dependent head twitch response (HTR) in male DBA/2J mice [20,33]. Figure 4 indicates that at a dose of 10 mg/kg compound 6a attenuated the number of head twitches induced by 5 mg/kg of the hallucinogen 2,5-dimethoxy-4-iodoamphetamine (DOI) over a 30 min time period. A two-way ANOVA with repeated measures revealed a main effect of Group (F2,18 = 6.49; p = 0.008), a main effect of Time (F5,90 = 348.16; p < 0.001) and an interaction between Time and Group (F10,90 = 14.92; p < 0.001). Second, we previously reported that D3 vs. D2 dopamine receptor subtype selective ligands could attenuate L-dopa dependent abnormal involuntary movement (AIM) scores in unilaterally lesioned (hemiparkinsonian) male rats, which is a model for L-dopa-Induced Dyskinesia (LID) [35,36]. Figure 5 indicates that at a dose of 10 mg/kg compound 6a could attenuate the AIM scores induced by 8 mg/kg of L-dopa and benserazide (each) in unilaterally lesioned rats. A two-way ANOVA with repeated measures revealed a main effect of Values were normalized to reference full agonist quinpirole (10,000 nM). The concentration of the test compound was equal to 10× the K i value as determined from competitive radioligand binding studies, therefore 90% of the binding sites were occupied.

In-Vivo Analysis of Compound 6a
Based upon the data indicating that compound 6a bound with high affinity to the human D3 dopamine receptor and with low affinity at both the human D2 dopamine receptor subtype (Table 1) and the 5-HT1A receptor ( Table 2 and Figure 2), two in-vivo assays were performed to determine if compound 6a was capable of in-vivo engagement of the D3 receptor in the brain. While we do appreciate the potential that possible invivo differences in pharmacological responses for the same drug in experimental animals and humans of sex differences, these initial in-vivo studies were primarily conduced to assess the ability of compound 6a to cross the blood brain barrier (BBB) and engage the pharmacologically relevant target. Further sex differences in the pharmacological response of animals to compound 6a will be conduction as further in-vivo studies are necessitated.
First, we have previously published that D3 receptor selective compounds could attenuate a hallucinogen-dependent head twitch response (HTR) in male DBA/2J mice [20,33]. Figure 4 indicates that at a dose of 10 mg/kg compound 6a attenuated the number of head twitches induced by 5 mg/kg of the hallucinogen 2,5-dimethoxy-4-iodoamphetamine (DOI) over a 30 min time period. A two-way ANOVA with repeated measures revealed a main effect of Group (F 2,18 = 6.49; p = 0.008), a main effect of Time (F 5,90 = 348.16; p < 0.001) and an interaction between Time and Group (F 10,90 = 14.92; p < 0.001). Second, we previously reported that D3 vs. D2 dopamine receptor subtype selective ligands could attenuate L-dopa dependent abnormal involuntary movement (AIM) scores in unilaterally lesioned (hemiparkinsonian) male rats, which is a model for L-dopa-Induced Dyskinesia (LID) [35,36].

Discussion
This communication describes part of our ongoing studies on the development of dopamine receptor selective ligands in vivo to dissect the role of the D2 and D3 dopamine receptor subtypes in the rewarding and motivational aspects of psychostimulant abuse [35,[37][38][39], the treatment of neurological and neurodegenerative disorders [36,40] and for the development of selective imaging agents [41].
We previously reported that structural variations in the aryl portion of 4-(thiophen-3-yl)benzamides play an important role in D3 dopamine receptor affinity, D3 vs. D2 receptor binding selectivity and D3 dopamine receptor functional selectivity (biased agonism) [20][21][22]30,35,42]. The structural templates for these studies included LS-3-134 and WW-III-55 (Figure 1). Both compounds possess a common 4-(thiophen-3-yl)benzamide with a saturated four-carbon chain adjacent to a substituted phenylpiperazine moiety. LS-3-134 was found to bind at the human D3 dopamine receptor with high affinity (K i = 0.17 nM) and demonstrated >150-fold D3 vs. D2 dopamine receptor binding selectivity [30]. LS-3-134 was found to be a partial agonist at the D3 dopamine receptor (35% of maximum efficacy) using an adenylyl cyclase inhibition assay. While WW-III-55 exhibits only moderate binding affinity (approximate K i of 20 nM) at D3 dopamine receptors, this compound binds with high (>800-fold) D3 vs. D2 dopamine receptor binding selectivity. Using an adenylyl cyclase assay, WW-III-55 was found to be a strong partial agonist (67.6 ± 12.5% of maximum efficacy) while LS-3-134 was found to be a weak partial agonist (34.4 ± 1.7%). Therefore, we projected that structurally related analogs of WW-III-55 and LS-3-134 might be identified that could exhibit high affinity binding at the D3 dopamine receptor, substantial D3 vs. D2 receptor binding selectivity and varying efficacy for G-protein and β-arrestin mediated signaling.
This expectation was verified when we examine the binding properties of a panel of structural analogs of WW-III-55 and LS-3-134, where the 4-(thiophen-3-yl)benzamide structure, which we project interacts with the SBS, was invariant while substituents on the phenyl moiety of the phenylpiperazine (the structure binding to the orthosteric binding site) were varied. This synthetic strategy led to the identification of several compounds that exhibited subnanomolar affinity (Ki values < 1.0 nM) at the human D3 dopamine receptor subtype with high D3 vs. D2 dopamine receptor binding selectivity (≥1000-fold) [22]. In the context of a dynamic ensemble theory of GPCR structure and activation [43][44][45], this observation suggests that each of the 4-(thiophen-3-yl)benzamide analogs used in that study interacts with the D3 dopamine receptor protein in a slightly different manner (depending on the position and chemical properties of the substituent), stabilizing a unique ensemble of receptor structures resulting in a collection of ligand-dependent structural variations in the D3 dopamine receptor energy landscape [42][43][44][45].
In this communication our initial goal was first to characterize a number of synthetic analogs in which the 4-(thiophen-3-yl)benzamide group and the adjacent saturated four carbon chain was invariant, while the substitution pattern on the phenyl ring of the phenylpiperazine moiety substituted at the 2nd, 3rd and 4th position with a fluorine. Once we established that the 2nd position was optimal for a fluoride substitution (compounds 3a and 6a), our second goal was to attempt to further optimize ligand binding affinity by exploring additional substitutions on the phenyl ring of the N-phenylpiperazine moiety (compounds 6d-f). Compound 6a remained our best candidate, with the highest affinity at the D3 receptor subtype (Ki value = 1.4 nM), while it exhibited substantial D3 vs. D2 receptor binding selectivity (>450-fold). Our third synthetic strategy was to substitute the thiophene ring (6a-f) with a thiazole ring (7a-f). In general, this structural alteration led to a decrease in D3 vs. D2 receptor binding selectivity.
Using the aforementioned strategy, we were able to identify a number of ligands which exhibited high affinity binding at the D3 dopamine receptor or substantial D3 vs. D2 dopamine receptor binding selectivity. However, compound 6a remained of particular interest because it bound at the D3 dopamine receptor with high affinity (K i value = 1.4 nM) and exhibited substantial D3 vs. D2 binding selectivity (>400-fold).
A direct comparison of the binding affinities for compounds 3a-f and 6a-f suggests that compound 6a-f likely bound at the D3 dopamine receptor subtype, but not the D2 dopamine receptor, in a bitopic mode. This observation was similar to what we previously reported [23]. While the identification of a ligand that can interact with a dualistic ligand binding mode may not be applicable for the development of receptor subtype selective ligands for every GPCR subfamily, it appears not to be unique to the dopamine receptor subtypes. Bitopic ligands have also been reported for other GPCRs, including the muscarinic and angiotensin II receptors [46][47][48]. Furthermore, binding data provided by NIMH Psychoactive Drug Screening Program (PSDP) indicated that compound 6a had reduced affinity at the 5-HT1A receptor, which is a common off-site binding component that has limited the development of in-vivo D3 dopamine receptor selective imaging agents. However, this binding data indicated that several alpha-adrenergic receptor subtypes bound compound 6a.
The studies in this communication also demonstrated that after intraperitoneal injection compound 6a was capable of engaging CNS D3 dopamine receptors in vivo leading to a significant decrease in the DOI-dependent head twitch response in mice and AIM scores in dyskinetic hemiparkinsonian rats. These observations suggest that compound 6a was capable of traversing the blood-brain barrier in a manner such that a therapeutic concentration could be achieved. We anticipate that studies such as those reported in this communication could lead to a better understanding of how the development of bitopic ligands might be designed and used to selectively target structurally similar GPCR receptor subtypes in vivo.

Chemistry
All reagents and solvents for synthesis were purchased and used without further purification. Structures of synthesized chemicals were identified using 1 H and 13 C nuclear magnetic resonance (NMR) spectra, and mass spectroscopy. 1 H and 13 C NMR were recorded in δ units relative to deuterated solvent (CDCl 3 ) as an internal reference by Bruker DMX 500 MHz NMR instrument (Bruker, Billerica, MA, USA). 1 H chemical shifts are reported in parts per million (ppm) and measured relative to tetramethylsilane (TMS). Mass spectra were acquired using a 2695 Alliance LC/MS (Milford, MA, USA). Purification of synthesized chemicals was conducted on Biotage Isolera One (Biotage, Salem, NH, USA) with a dual-wavelength UV-VIS detector. A detailed description of the synthesis and characterization of the compounds described in this paper can be found in the Supporting Information.

D2 and D3 Dopamine Receptor Competitive Radioligand Binding Assays
For these competitive binding studies, transfected HEK293 cell homogenates were suspended in homogenization buffer and incubated with radioligand [ 125 I]IABN, in the presence or absence of inhibitor at 37 • C for 60 min (total volume = 150 µL), as previously described [49]. The final radioligand concentration was approximately equal to the K d value for the binding of the radioligand. Nonspecific binding was defined as the binding of the radioligand in the presence of 25 µM (+)-butaclamol. For each competition curve, triplicates were performed using two concentrations of inhibitor per decade over five orders of magnitude. Binding was terminated by the addition of cold wash buffer (10 mM Tris-HCl/150 mM NaCl, pH = 7.5) and filtration over a glass-fiber filter (Pall A/B filters, #66198).

Cyclase Inhibition Assay
The forskolin-dependent adenylyl cyclase inhibition assay was performed using a cAMP-Glo Assay kit (Promega). For this assay cAMP stimulates protein kinase A (PKA) activity, which leads to a decrease in ATP levels and a decrease in the luciferase-coupled production of light. This is a 96 well plate assay using HEK cells stably transfected to express the human D3 dopamine receptor (10,000 cells/well/0.1 mL complete media). After removing media and rinsing with phosphate buffered saline, cells were incubated with 20 µL of forskolin (6000 nM) and test compound or vehicle. The dose of test compound was equal to 10x the K i value obtained from our competitive radioligand binding studies, such that 90% of the binding sites are occupied. After cells, test compound (or vehicle) and forskolin (or vehicle) were incubated for 15 min, cells were lysed using 20 µL lysis buffer (provided with the kit) and mixed for 15 min at room temperature. A cAMP Detection Solution (40 µL) was added to each well, gently mixed for 60 s and incubated at room temperature for 20 min. The Kinase-Glo Solution was added (40 µL) to terminate the PKA reaction. The reaction mixture was gently mixed for 60 s and then incubated at room temperature for 10 min. The remaining ATP was detected via a luciferase reaction. The plates were read using an Enspire Alpha 2390 Multilabel Reader/Luminometer. Haloperidol (200 nM) and quinpirole (10,000 nM) was used as a prototypic reference antagonist and full agonist, respectively.

β-Arrestin Binding Assay
The β-arrestin binding assay was performed using a DiscoverX Pathhunter ® kit according to the manufacturer's instructions. PathHunter ® eXpress β-Arrestin GPCR cells are engineered to co-express the ProLink™ (PK) tagged GPCR and the Enzyme Acceptor (EA) tagged β-Arrestin. Activation of the GPCR-PK induces β-Arrestin-EA recruitment, forcing complementation of the two β-galactosidase enzyme fragments (EA and PK). The resulting functional enzyme hydrolyzes substrate to generate a chemiluminescent signal. DRD3-U2OS cells were seeded according to kit instructions into white plastic 96-well plates (provided with the kit) 48 h prior to the assay in a volume of 100 µL/well using the Cell Plating Reagent (CPR) provided. Vehicle or test drug, diluted in CPR, was added to the designated wells and incubated for 90 min at 37 • C. Working detection solution (a component provided with the kit) was added to the wells and incubated for 1 h in the dark at room temperature. As discussed for the cyclase assay, an Enspire Alpha 2390 Multilabel Reader/Luminometer was used to detect the luminescent signal with haloperidol and quinpirole used as reference antagonist and full agonist, respectively.

In-Vivo Studies
The treatment of the animals and the experimental procedures were approved by the UNTHSC Institutional Animal Care and Use Committee (IACUC). Animal care and housing were in adherence with the conditions set forth in the "Guide for the Care and Use of Laboratory Animals" (Institute of Laboratory Animal Resources on Life Sciences, National Research Council, 1996).

Head Twitch Response
Upon arrival DBA/2J mice (Jackson Labs) were allowed to acclimate for at least one week prior to testing. Mice were housed at ≤4 animals per cage with unlimited access to food and fresh water. The protocol for the HTR assessment has been previously reported [32]. Briefly, on the day of testing mice were weighed and placed individually in an open-ended Plexiglas cylinder (21 × 34 cm) with a clean paper towel on the floor in a dimly lighted room. Animals were allowed to habituate to the cylinder for 10 min prior to the intraperitoneal (i.p.) injection of vehicle (90% sterile deionized water and 10% dimethylsulfoxide (DMSO: Sigma) in sterile deionized water) or test drug (10 mg/kg). Five minutes later DOI (5 mg/kg) was administered, and the mouse was immediately returned to the cylinder.
The mice exhibited various responses to DOI administration including a vertical jerking movement. However, a head twitch was defined as a rapid movement of the head, without the involvement of the front paws during grooming. We estimate that the HTR lasted approximately 0.6 s. At least two observers counted the number of head twitches by visual examination and the number of head twitches was recorded in 5-min cumulative intervals. The data points are presented as the mean values obtained by two observers.
The effect of Group was assessed using two-way analysis of variance (ANOVA) with Time as the repeated measure. The data was further analyzed at each time point with a one-way ANOVA with Group as the factor followed by planned individual comparisons between different groups using a single degree-of-freedom F test involving the error term from the overall ANOVA when the Time and Group interaction was significant. The alpha level was set at 0.05 and Systat 13 statistical package was used.

AIMs Studies
The Abnormal Involuntary Movement (AIMs) studies were performed using male Sprague-Dawley rats that were unilaterally lesioned, using injections of 6-hydroxydopamine (6-OHDA) in the medial forebrain bundle (MFB) by the commercial vendor (Charles River Laboratories). Upon arrival to our vivarium, the lesioned rats were housed under a 12-h light: 12-h dark cycle. The animals were housed singly with unlimited access to food and water. 8 mg/kg L-dopa with 8 mg/kg benserazide was administered to each rat as a daily i.p. injection for 21 consecutive days to induce the development of dyskinesia-like movements. Benserazide was included because it is a peripherally-acting aromatic L-amino acid decarboxylase inhibitor which prevents the peripheral metabolism of L-dopa. The L-dopa/benserazide solutions were dissolved in water. Test drug (compound 6a) was prepared using 90% sterile water with 10% DMSO.
8 mg/kg of L-dopa combined with 8 mg/kg benserazide was administered via intraperitoneal injection followed by either the vehicle control or test drug (compound 6a, 10 mg/kg). Experiments were designed such that animals received test drug only once per week. Throughout the course of these studies each animal received a minimum of one dose of L-dopa/benserazide (8 mg/kg each) per week to maintain the involuntary movements.
AIMs ratings were performed as previously described [31,38,52]. The severity of the AIMs was quantified using lesioned rats that were observed individually in their home cages at 15 min intervals, starting 15 min after the injection of L-dopa for approximately 2 h.
The AIMs were scored in four categories: (1) axial AIMs, including dystonic or choreiform torsion of the trunk and neck towards the side contralateral to the lesion; (2) limb AIMs, including jerky and/or dystonic movements of the forelimb contralateral to the lesion; (3) orolingual AIMs, including twitching of orofacial muscles with empty masticatory movements and protrusion of the tongue towards the side contralateral to the lesion and (4) locomotive AIMs, including increased locomotion with contralateral side bias. Each of the four categories were scored on a severity scale from 0 to 4, where 0 = absent, 1 = present for less than half of the observation time, 2 = present for more than half of the observation time, 3 = present all the time but suppressible by external stimuli and 4 = present all the time and not suppressible by external stimuli. For these experiments the external stimuli that was used was gentle tapping on the cage with a pencil.
The change in crossovers from controls was calculated for each test compound dosage group by subtracting the average number of crossovers of the respective vehicle group from the number of crossovers of individual subjects in the test compound dosage groups. The latter measure is graphed to illustrate the effects of the compounds on locomotion.
The effect of Group was assessed using two-way analysis of variance (ANOVA) with time as the repeated measure. The data was further analyzed at each time point with a one-way ANOVA with Group as the factor. The alpha level was set at 0.05 and Systat 13 statistical package was used.

Informed Consent Statement: Not applicable.
Data Availability Statement: The article contains complete data used to support the findings of this study.