Xanthine–Dopamine Hybrid Molecules as Multitarget Drugs with Potential for the Treatment of Neurodegenerative Diseases

Multitarget drugs based on a hybrid dopamine–xanthine core were designed as potential drug candidates for the treatment of neurodegenerative diseases. Monoamine oxidase B (MAO-B) inhibitors with significant ancillary A2A adenosine receptor (A2AAR) antagonistic properties were further developed to exhibit additional phosphodiesterase-4 and -10 (PDE4/10) inhibition and/or dopamine D2 receptor (D2R) agonistic activity. While all of the designed compounds showed MAO-B inhibition in the nanomolar range mostly combined with submicromolar A2AAR affinity, significant enhancement of PDE-inhibitory and D2R-agonistic activity was additionally reached for some compounds through various structural modifications. The final multitarget drugs also showed promising antioxidant properties in vitro. In order to evaluate their potential neuroprotective effect, representative ligands were tested in a cellular model of toxin-induced neurotoxicity. As a result, protective effects against oxidative stress in neuroblastoma cells were observed, confirming the utility of the applied strategy. Further evaluation of the newly developed multitarget ligands in preclinical models of Alzheimer’s and Parkinson’s diseases is warranted.


Introduction
Neurodegeneration is described as the progressive loss of structure and function of neurons leading to central nervous system dysfunction [1,2]. Although neurodegenerative diseases (NDs) are a clinically and pathologically heterogeneous group of disorders, including Parkinson's disease (PD), Huntington's disease (HD), and Alzheimer's disease (AD), they share some general molecular mechanisms leading to a pathogenic cascade, which includes protein misfolding and aggregation, mitochondrial dysfunction, oxidative stress and free radical formation, and inflammation and deregulation of microRNAs [3][4][5][6][7][8].
The primary causes of neurodegeneration are still unclear, and a wide spectrum of potential biological targets is considered for the prevention of neurodegeneration and the treatment of neurodegenerative diseases. Therefore, a multitarget-drug approach, also described as polypharmacology, presents a strategy adopted by medicinal chemists to provide new molecules with increased therapeutic efficacy [9,10]. However, the prevalent management in various NDs is still based on a mono-target approach, and the primary drug targets are related to the modulation of neurotransmitter levels to compensate disturbances caused by the loss of neurons. Dopamine replacement therapy in PD, and acetylcholine esterase (AChE) inhibitors in AD, were milestones in the symptomatic therapy of these diseases; nevertheless, they are insufficient to prevent or stop the chronic progression of neurodegeneration [11,12].
From a medicinal chemistry perspective, multitarget drugs can interact with several targets, thereby improving therapeutic efficacy, attenuating side effects, and providing a more predictable pharmacokinetic profile than combination therapies [13,14]. Thus, the strategy to develop multitarget ligands is now intensively explored for ND, generating new molecules that simultaneously address various biological targets [10,15].
Monoamine oxidases (MAO) are enzymes involved in the metabolism of various amines, including neurotransmitters. There are two isoenzymes, MAO-A and MAO-B, that differ in substrate preference and tissue distribution. Elevated levels of MAO-B have been detected in the aging brain, resulting in increased MAO-B enzymatic activity around amyloid β-plaques in aged AD mice, reducing levels of neurotransmitters, in particular, dopamine [16,17]. Even worse, MAO-B may lead to an acceleration of the neurodegenerative process by generating potentially neurotoxic substances, such as dopaldehyde and hydrogen peroxide [18]. On the other hand, the neuroprotective effect of MAO-B inhibition may be also be associated with the regulation of the mitochondrial apoptosis system, maintenance of mitochondrial function, suppression of α-synuclein oligomerization and aggregation, increased expression of genes encoding antioxidant enzymes, and anti-apoptotic Bcl-2 and pro-survival neurotropic factors [19][20][21]. Those potential disease-modifying properties of MAO-B inhibitors were widely investigated to create new anti-ND agents designed as multitarget drugs, e.g., via combination of MAO-B inhibition and a blockade or activation of G-protein-coupled receptors or inhibition of AChE [22][23][24].
The blockade of adenosine A 2A receptors (A 2A AR) enhances dopamine D 2 receptor (D 2 R)-dependent signaling in the basal ganglia of the central nervous system. Therefore, the selective A 2A AR antagonist istradefylline, which is based on an 8-styrylxanthine core, was approved in Japan (in 2013) and later on in the USA (in 2019) as an adjuvant therapeutic in combination with levodopa for the treatment of PD [25]. Besides the improvement in motor-related dysfunction in PD patients, A 2A AR antagonists may exhibit neuroprotective effects by reducing oxidative stress, glutamate excitotoxity, and microglial reactivity [26][27][28]. Several in vivo models of neurodegenerative diseases indicated an improvement in cognitive impairment via a blockade of A 2A AR. This effect was further enhanced by A 1 AR antagonism in dual A 1 /A 2A AR blocking agents [29][30][31]. Also, in models of AD, A 2A AR antagonists or a knockdown of A 2A ARs showed neuroprotective effects [32,33]. Recently, the chronic administration of caffeine (1,3,7-trimethylxanthine), a non-selective AR antagonist, was found to increase information encoding and processing in the hippocampus of mice, indicating beneficial effects on memory and learning [34].
Phosphodiesterases (PDE) are a family of enzymes that degrade cyclic nucleotides via the hydrolysis of phosphodiester bonds. Eleven PDE family members (PDE1-PDE11) were identified, which show various substrate selectivities and biological profiles [35]. In the context of NDs, PDE4 and PDE10 seem to be particularly involved. PDE4 is a cAMP-specific enzyme and its selective inhibitors have been found to improve memory and cognition deficits in rodent AD models, associated with misfolded amyloid β protein. Moreover, PDE4 inhibitors may increase tyrosine hydroxylase phosphorylation, leading to an enhancement of dopamine biosynthesis and a protection of dopaminergic neurons [36][37][38]. PDE10 hydrolyzes both cAMP and cGMP, and high expression was detected in striatum. The dysregulation of this brain region may lead to motor dysfunction and cognition impairment. PDE10 inhibitors are intensively explored as drug candidates for HD, PD, and schizophrenia [39,40].
Dopamine D 2 receptor agonists are applied as monotherapy or combined with L-DOPA, the precursor of dopamine, for the treatment of PD [41]. The enhancement of D 2 neurotransmission in the nigrostratial pathway improves motor dysfunction. Moreover, recent studies showed potential neuroprotective and anti-inflammatory effects for D 2 R agonists [42,43]. The activation of D 2 R expressed on CD4 + T cells was suggested to protect against neuroinflammation and neurodegeneration in animal studies [44].
The object of the present study was the optimization of novel hybrid compounds targeting MAO-B and adenosine receptors as anti-neurodegenerative agents. The hybrid chemical scaffold shares two pharmacophore units: a xanthine part (including pyrimidineor diazepine-containing tricyclic analogues) and a dopamine part. Chemical modifications focus on the N7-position of the xanthine core via the introduction of various substituents, such as alkyl or arylalkyl, or annelation leading to tricyclic structures. Moreover, a preliminary structure-activity relationship (SAR) analysis of xanthine derivatives with extended alkyl substituents at the N1and the N3-position was initiated. The final compounds were tested in vitro for their inhibitory potency on MAO-B, PDE4, and PDE10, and for their interaction with adenosine and dopamine receptor subtypes.
Xanthine and dopamine fragments, which are present in designed compounds, were reported in the literature to act as potent radical scavengers [45,46]. Since oxidative stress is postulated as one of the factors contributing to neurodegeneration, the selected compounds of the current series were additionally tested in vitro for their potential to exhibit antioxidant properties. Furthermore, a cellular model was employed to investigate whether the newly designed compounds are able to provide neuronal protection against oxidative damage.

Chemistry
The synthetic strategy to obtain xanthine derivatives fused with a dopamine fragment included: first, the preparation of the xanthine part of the molecules, and then, the conjugation with a dopamine moiety whose hydroxy functions were protected by methyl groups (Scheme 1).
The deprotection of the methoxy groups in the 3,4-dimethoxyphenylethylaminosubstituted compounds was carried out via reflux with 48% or 40% HBr and alkalization with 20% Na 2 CO 3 to pH 8-9, leading to the final products.
The structure of all compounds was confirmed by elemental analyses and UV, IR, and NMR spectra. The UV spectra showed a bathochromic shift of λ max from 275 nm to about 290 nm (xanthine compounds) or from 275 nm to about 300 nm (tricyclic compounds), typical for 8-aminoxanthines [49]. Compounds showed IR absorption bands typical for xanthine derivatives [50], and in proton and carbon NMR spectra, adequate shifts could be observed. The purity of all tested compounds was at least 95%, determined by UPLC/MS. Spectroscopic data are presented in Supplementary Data.

Monoamine Oxidase Inhibition
The final compounds were evaluated in vitro for their human MAO-B (hMAO-B) inhibitory potencies, and selected structures were also investigated for human MAO-A (hMAO-A) inhibition. The results are presented in Table 1.

Monoamine Oxidase Inhibition
The final compounds were evaluated in vitro for their human MAO-B (hMAO-B) inhibitory potencies, and selected structures were also investigated for human MAO-A (hMAO-A) inhibition. The results are presented in Table 1.
All structures displayed high MAO-B inhibitory potency, with IC 50 values in the nanomolar concentration range. The results strongly support the utility of a hybrid dopamine-xanthine scaffold for activity towards MAO-B. Compound 11a, a caffeine combined with a dopamine moiety, presented high MAO-B inhibitory potency (IC 50 = 50.7 nM). The elongation of the alkyl chain in the N7-position (11b, 11e, 11h, 11i) had no or only a slight impact on biological activity, maintaining potencies in the same range. The exception was compound 11c, containing a propyl residue, which displayed three-fold lower inhibition potency. A similar effect was also observed for structures with branched alkyl substituents (11d, 11f, 11g). In contrast, compounds with aromatic residues at the N7position (11j, 11k, 11l) showed comparable MAO-B inhibitory potencies to structures with linear alkyl chains. The modification of compound 11c via the introduction of longer propyl (12) or butyl (13) substituents at the N1and N3-positions slightly improved the inhibitory activity compared to the 1,3-dimethyl analogue.   Tricyclic derivatives (20a, 20b, 21, 22) generally showed somewhat lower MAO-B inhibitory potency; however, the IC 50 values were still in the nanomolar concentration range. The most active tricyclic compound, 20b (IC 50 = 101 nM) contained a 1,3-diazepino ring and a short methyl chain at the N1and N3-positions.
Within the obtained library of compounds, only two compounds (11i, 11k) showed significant MAO-A inhibitory activity (IC 50 < 500 nM), indicating that the compounds were selective for MAO-B, as desired.

Adenosine Receptors
The final compounds were tested in radioligand binding assays to evaluate their affinity for all four ARs subtypes (see Table 2) [51,52]. Rat brain cortical membrane preparations were used as a source of rat A 1 (rA 1 ) ARs, and rat brain striatal membrane preparations served as a source of rat A 2A (rA 2A ) ARs. For A 2B -and A 3 AR assays, membrane preparations of Chinese hamster ovary (CHO) cells recombinantly expressing either human A 2B (hA 2B ) or human A 3 (hA 3 ) ARs were employed. 1 8-(3,4-Dihydroxyphenylethylamino)caffeine (11a) was found to be a moderately potent A 2A AR ligand with additional micromolar affinity for the A 1 AR. The elongation of the alkyl chain at the N7-position (in 11b, 11c, 11e, 11h) resulted in increased affinity and selectivity for the A 1 AR subtype. Compound 11b (8-(3,4-dihydroxyphenylethylamino)-7-ethyltheophylline) presented balanced, dual A 1 /A 2A AR affinity (A 1 AR, K i = 479 nM; A 2A AR, K i = 672 nM). A significant decrease in affinity at both A 1 -and A 2A ARs was observed for compound 11i containing an N7-hexyl substituent. Furthermore, the N7propyl derivative 11c presented significantly lower affinity than the N7-ethyl (11b) and N7-butyl (11e) analogues at both A 1 -and A 2A ARs. Interestingly, compounds with N7propyl substituent and longer alkyl chains, such as di-propyl (12; A 1 AR, K i = 130 nM; A 2A AR, K i = 833 nM) or di-butyl (13) fragments introduced at the N1,N3-positions, displayed higher affinity towards both A 1 -and A 2A ARs than the N1,N3-dimethylsubstituted analogue.
The replacement of a linear alkyl chain at the N7-position by a branched one (11d, 11f, 11g) resulted in low, micromolar affinity at both A 2A -and A 1 ARs, with the exception of compound 11d, displaying a sub-micromolar K i value at the A 1 AR. A comparable effect was observed in the group containing an aromatic phenyl ring connected by a short linear linker to the xanthine core at the N7-position (11j, 11k, 11l). Here, only compound 11l with phenoxyethyl residue at the N7-position presented high affinity for the A 1 AR (K i = 124 nM).
In the group of tricyclic derivatives, compound 20a (N9-(3,4-dihydroxyphenylehyl)-1,3-dimethyltetrahydroprymido[2,1-f ]purinedione) was found to be a moderately potent A 2A AR ligand (K i = 507 nM). The enlargement of the third heterocyclic ring fused to the f -bond (20b) resulted in an increase in affinity for the A 2A AR (K i = 234 nM). A loss of selectivity and a significant enhancement of AR affinity was observed for compounds 21 and 22, with a longer alkyl chain at the N1,N3-postions. Compound 21 showed triple A 1 /A 2A /A 2B AR affinity compared to compound 22 binding to all four AR subtypes with nanomolar affinities.

Affinity at Dopamine Receptors
The final compounds were tested in radioligand displacement assays to evaluate their affinity to dopamine D 1 and D 2 receptors (see Table 2). Rat brain cortical membranes were used as source of native rat D 1 receptor protein, and human D 2 receptor was recombinantly expressed in HEK293 cells.
None of the tested compounds was found to interact with dopamine D 1 receptor, at least at the highest tested concentration of 1 µM. Therefore, the presented compounds may be considered inactive or very weak D 1 receptors ligands.
Compound 13, as the most potent D 2 R ligand in this series, was evaluated for its intrinsic activity in a functional cAMP assay, and it showed agonistic properties with an EC 50 value of 11.3 µM (Figure 1). This feature can be considered beneficial for the treatment of PD to elevate decreasing D 2 transmission in the nigro-striatal pathway related to the neurodegenerative process. EC50 value of 11.3 µM (Figure 1). This feature can be considered beneficial for the treatment of PD to elevate decreasing D2 transmission in the nigro-striatal pathway related to the neurodegenerative process.

Phosphodiesterase Inhibition
Further biological evaluation of the presented compound library was performed determining inhibitory potencies against selected human PDE isoforms, namely, PDE4 and PDE10 (for results see Table 3).
Regarding PDE10A1, the most active compound was 11d (IC50 = 2.30 µM), which was also the most potent dual PDE4B1/PDE10A inhibitor. The introduction of a short, straight alkyl chain with 2-4 carbons in the N7-position (11b, 11c, 11e) maintained an IC50 value in the range of 3-5 µM. A similar inhibitory potency was also observed for compound 11k with an N7-phenylethyl substituent. The modification of the linker between the xanthine core and the aromatic ring in the N7-position (11j, 11l) resulted in a decrease in inhibitory potency in PDE10A. A significant reduction in inhibitory activity was also observed for the rest of the tested compounds, including the series of tricyclic structures, except for N1,N3-dipropyl derivative (12), which maintained the activity in a low micromolar range.
Regarding PDE10A1, the most active compound was 11d (IC 50 = 2.30 µM), which was also the most potent dual PDE4B1/PDE10A inhibitor. The introduction of a short, straight alkyl chain with 2-4 carbons in the N7-position (11b, 11c, 11e) maintained an IC 50 value in the range of 3-5 µM. A similar inhibitory potency was also observed for compound 11k with an N7-phenylethyl substituent. The modification of the linker between the xanthine core and the aromatic ring in the N7-position (11j, 11l) resulted in a decrease in inhibitory potency in PDE10A. A significant reduction in inhibitory activity was also observed for the rest of the tested compounds, including the series of tricyclic structures, except for N1,N3-dipropyl derivative (12), which maintained the activity in a low micromolar range.

Determination of Antioxidant Activity
The 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay was used to determine the antioxidant properties of selected compounds using ascorbic acid, trolox, and quercetin as references. The study demonstrated that most of the tested xanthine derivatives have high capacity for DPPH reduction and that their free radical scavenging activity exceeds 50% of the antioxidant effect of reference quercetin tested at the same concentration ( Figure 2 and Table 4). Among the evaluated structures, four compounds, 11i, 11l, 21, and 22, represented significantly lower antioxidative activity than quercetin. From a structural point of view, 21 and 22 share a common pyrimido[2,1-f ]purinedione core with propyl or butyl substituents at the N1and N3-position. Compounds 11i and 11l, containing hexyl or phenoxyethyl as N7-substituents, were the only structures in the groups of bicyclic xanthine derivatives that lacked antioxidant activity in the DPPH assay.

Neuroprotective Effects of Hybrid Compounds in a Cellular Model
Based on their promising antioxidant potency, selected xanthine derivatives were further tested for neuroprotective effects in an oxidative-stress-induced cell death model. Oxidative stress plays an important role in the degeneration of dopaminergic neurons in Parkinson's disease (PD), but also in other neurodegenerative diseases [55]. Compounds with capacity to scavenge reactive oxygen species could potentially represent beneficial neuroprotective properties in neurodegeneration therapies [56].
For neuroprotection studies, a cellular model of neuroblastoma SH-SY5Y cells exposed to hydrogen peroxide was applied. Lactate dehydrogenase (LDH) release was used as a measure of cell death. First, the compounds (11a, 11c, 12, 13, 20a, 21, and 22) were tested for their possible intrinsic neurotoxicity to the SH-SY5Y cell line (see Supplementary Figure S1). One of the tested compounds (13) showed significant cytotoxicity at 50 µM concentration and was excluded from further evaluation. All other compounds were tested for cytoprotective effects against H 2 O 2 -induced toxicity in SH-SY5Y cells.
The compounds presented protective effects, significantly decreasing cytotoxicity induced by the reactive oxygen species (Figure 3). The most potent compounds were 12, 21, and 22. Dose-dependent responses were observed for 11c, 12, and 21. Interestingly, the considered structures represented cytoprotective activity in SH-SY5Y cells exposed to oxidative stress, independent on the degree of their antioxidant potency determined in the DPPH assay. Although the antioxidant activity could be one of the functionalities that drive the protection against the accumulation of reactive oxygen species in neural cells, other mechanisms of action were also proposed for neuroprotective agents. For example, it has been widely reported that MAO-B inhibitors possess the ability to attenuate the toxic effects in progression of neurodegenerative diseases [21,57]. Since all of the compounds tested in neuroprotection experiments represented high inhibitory activity at MAO-B (IC 50 values: 50.7-170 nM), the retaining of neuroprotective activity by compounds 21 and 22, which lacked antioxidant properties, could potentially be attributed to their activity towards MAO-B.
ample, it has been widely reported that MAO-B inhibitors possess the ability to attenuate the toxic effects in progression of neurodegenerative diseases [21,57]. Since all of the compounds tested in neuroprotection experiments represented high inhibitory activity at MAO-B (IC50 values: 50.7-170 nM), the retaining of neuroprotective activity by compounds 21 and 22, which lacked antioxidant properties, could potentially be attributed to their activity towards MAO-B.

Conclusions
A library of 19 novel dopaminyl-substituted derivatives of purinediones was synthesized. Systematic modification of the substituent in the N7-position of the xanthine core was performed, and SARs were analyzed regarding MAO-B, PDE4, and PDE10, adenosine and dopamine receptors. The hybrid dopaminyl-xanthine scaffold showed high MAO-B inhibitory potency with IC50 values in the range of 44 to 205 nM. On the other side, SARs for PDE inhibition and AR affinity were steeper, and more rigid structural requirements were observed. For PDE inhibitory activity, the presence of a short 2-4 carbon chain, or a benzyl or phenylethyl residue, in position N7 was required. AR affinity was increased by the introduction of a methyl substituent in position N7 or by annelation of a third ring to the f-bond. Although a dopaminyl fragment was present, all compounds showed low or only moderate affinity to dopamine receptors. For the most

Conclusions
A library of 19 novel dopaminyl-substituted derivatives of purinediones was synthesized. Systematic modification of the substituent in the N7-position of the xanthine core was performed, and SARs were analyzed regarding MAO-B, PDE4, and PDE10, adenosine and dopamine receptors. The hybrid dopaminyl-xanthine scaffold showed high MAO-B inhibitory potency with IC 50 values in the range of 44 to 205 nM. On the other side, SARs for PDE inhibition and AR affinity were steeper, and more rigid structural requirements were observed. For PDE inhibitory activity, the presence of a short 2-4 carbon chain, or a benzyl or phenylethyl residue, in position N7 was required. AR affinity was increased by the introduction of a methyl substituent in position N7 or by annelation of a third ring to the f -bond. Although a dopaminyl fragment was present, all compounds showed low or only moderate affinity to dopamine receptors. For the most active compound 13, D 2 R agonistic properties were confirmed, which is desired for the intended application, namely, the treatment of neurodegenerative diseases by multitarget drugs. In the light of the above, novel potent MAO-B inhibitors/adenosine receptor antagonists were identified in the current series, with ancillary activity at PDE isoforms. Although more balanced activity at all considered biological targets would be expected for optimized multitarget drugs, our early-stage study provided an insight into the key structural elements that may introduce additional beneficial properties to MAO-B/A 2A AR dual compounds that may constitute for leading structures in further studies. An additional advantage of the discussed hybrid compounds is their antioxidant activity, determined in the DPPH assay, and neuroprotective properties observed in an oxidative-stress-induced neuroblastoma cell death model.
Those above outcomes justify the conclusion that the presented hybrid chemical scaffold may be considered a novel multitarget drug entity in the search for potent antineurodegenerative agents. The preference of the dopaminyl-xanthine core for MOA-B inhibition indicates its suitability for further structural optimization towards dual targeting MAO-B/PDE and MAO-B/AR agents.

Materials and Methods
All commercially available reagents and solvents were used without further purification. Melting points (mp.) were determined on a MEL-TEMP II (LD Inc., Kansas City, KS, USA) melting point apparatus and were uncorrected. IR spectra were taken as KBr discs on a Jasco FT/IR-410 spectrometer. UV spectra were recorded on Jasco UV-Vis V530 in concentration of 10 −5 mol/L in methanol. 1 H-NMR spectra were determined with a Varian Mercury 300 MHz apparatus in CDCl 3 (3,4-dimethoxyphenylethyl derivatives) or in DMSO d 6 (dopaminyl derivatives) using tetramethylsilane as an internal standard. 13 C NMR data were recorded on a 75 MHz on Varian-Mercury-VX 300 MHz PFG spectrometer.
The J values are reported in Hertz (Hz), and the splitting patterns are designated as follows: s (singlet), d (doublet), t (triplet), dd (doublet of doublets), dt (doublet of triplets), quin (quintet), and m (multiplet). Elemental analyses (CHN) were performed on an Elemental Vario-El III (Hanau, Germany) apparatus and were in accordance with theoretical values within ± 0.5%. The purity of the tested compounds was determined (%) on Waters TQD mass spectrometer coupled with a Waters ACQUITY UPLC system. Retention times (t R ) are presented in minutes. The reactions were monitored via thin-layer chromatography (TLC) using aluminum sheets coated with silica gel 60F254 (Merck, Darmstadt, Germany) as developing system cyclohexane/dioxane 1:1. Spots were detected under UV light.

Phosphodiesterase Inhibition
Tested and reference compounds were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mM and further diluted in assay buffer (10 mM Tris-HCl, 10 mM magnesium chloride and 0.05% Tween-20; pH 7.4). Inhibition of human PDE10A and 4B1 was measured using PDElight HTS cAMP phosphodiesterase assay kit (Lonza, Basel, Switzerland) according to manufacturer's recommendations. Then, 10 ng of PDE10A and 5 ng of PDE4B1 (BPS Biosciences, San Diego, CA, USA) in appropriate buffer was incubated with reference and tested compound for 20 min. After incubation, the cAMP substrate (final concentration 1.25 µM for PDE10A and 5 µM for PDE4B1) was added and incubated for 1 h. Then PDELight AMP Detection Reagent was added and incubated 10 min. All reactions were carried out at 37 • C in 96 white-walled, half-area-well plates (PerkinElmer). Luminescence was measured in a multifunction plate reader (POLARstar Omega, BMG Labtech, Ortenberg, Germany). The percentage of inhibition and IC 50 s were computed using GraphPad Prism Version 6.0 software.

Radioligand Binding Assays at Adenosine Receptors
Adenosine receptor radioligand binding assays were performed as previously described [61] using rat brain cortical membrane preparations for A 1 and rat brain striatal membrane preparations for A 2A AR assays. Frozen rat brains (unstripped) were obtained from Pel Freez, Rogers, AR, USA. For assays at human AR subtypes, cell membranes of CHO cells recombinantly expressing the respective receptor were used. The following compounds were employed as radioligands:  [65]. Initially, a single high concentration of compound was tested (1 µM). For potent compounds, full concentration-inhibition curves were determined using six to ten different concentrations of test compound spanning three orders of magnitude. At least three independent experiments were performed. Data were analyzed using GraphPad PRISM program version 4.0 or higher (Graph Pad, San Diego, CA, USA).

Dopamine D 1 receptors
Rat striatum tissue was thawed in 50 volumes of ice-cold 50 mM potassium phosphate buffer, pH 7.4, homogenized and centrifuged at 20,000× g for 20 min. The resulting pellet was resuspended in the same quantity of the buffer and centrifuged again for 20 min. Binding experiments were conducted in 96-well microplates in a total volume of 250 µL of appropriate buffers. Reaction mix included 50 µL solution of test compound, 50 µL of radioligand, and 150 µL of tissue suspension (3 mg/mL). [ 3 H]SCH23390 (spec. act. 81.9 Ci/mmol, PerkinElmer, NET 930) was used for labeling the D 1 -receptor. For measuring unspecific binding, cis(Z)flupentixol-5 µM was applied. Samples were incubated at 30 • C for 60 min. The incubation was terminated by rapid filtration over glass fiber filters GF/B using a Harvester (PerkinElmer, USA). The radioactivity was measured in MicroBeta TriLux 1450-liquid scintillation counter (PerkinElmer). Each compound was tested in a screening assay in duplicate at final concentrations of 1 µM. Results were expressed as a percent of inhibition of specific binding.

Dopamine D 2 receptors
HEK293 cells with stable expression of human D 2L receptor were maintained at 37 • C in a humidified atmosphere with 5% CO 2 and grown in Dulbecco's Modifier Eagle Medium containing 10% dialyzed fetal bovine serum and 500 µg/mL G418 sulfate. For membrane preparation, cells were subcultured in 150 cm 2 flasks, grown to 90% confluence, washed twice with prewarmed to 37 • C phosphate buffered saline (PBS), and pelleted by centrifugation (200× g) in PBS containing 0.1 mM EDTA and 1 mM dithiothreitol. Prior to membrane preparation, pellets were stored at −80 • C.
Cell pellets were thawed and homogenized in 10 volumes of assay buffer using an Ultra Turrax tissue homogenizer and centrifuged twice at 35,000× g for 15 min at 4 • C. The incubation buffer consisted of 50 mM Tris-HCl, 1 mM EDTA, 4 mM MgCl 2 , 120 mM NaCl, 5 mM KCl, 1.5 mM CaCl 2 , and 0.1% ascorbate. Assay was incubated in a total volume of 200 µL in 96-well microtiter plate for 1 h at 37 • C. The process of equilibration was terminated by rapid filtration through Unifilter plate with a 96-well cell harvester and radioactivity retained on the filter was quantified on a Microbeta plate reader (PerkinElmer). For competition studies, the assay samples contained radioligand 2.5 nM [ 3 H]-raclopride (PerkinElmer). Non-specific binding was defined with 10 µM of haloperidol. Each compound was tested in triplicate at 7 concentrations (10 −10 -10 −4 M). The inhibition constants (K i ) were calculated from the Cheng-Prusoff equation [66]. Results were expressed as means of at least two separate experiments.

cAMP Accumulation Assay in Cells Expressing hD 2 R
Intrinsic activity of compound 13 and natural agonist-dopamine-was determined in cAMP accumulation assay. Briefly, HEK293 cells stably expressing human D 2L R were incubated (30 min, room temperature) with forskolin (1 µM) and evaluated compounds in triplicates (7 appropriate concentrations in a range spanning over 6 log units) in presence of phosphodiesterase inhibitor (RO-201724, 100 µM). Subsequently, intracellular cAMP accumulation was measured with homogenous TR-FRET immunoassay, using LANCE Ultra cAMP kit (PerkinElmer) and following manufacturer's instructions. TR-FRET signal, inversely proportional to cAMP concentration in sample, was measured using Infinite M1000 Pro microplate reader (Tecan, Männedorf, Switzerland). Sigmoidal dose-response curve fitting was performed, using GraphPad Prism™ software (version 5.01, San Diego, CA, USA).

Determination of the Antioxidant Activity
Antioxidant properties of considered structures were evaluated in DPPH test. The assay utilizes a stable free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH). DPPH has a deep violet color in solution and strong absorption band at 520 nm. The antioxidant activity of the tested compounds was defined by their ability to induce the formation of the reduced form of DPPH. Upon reduction, DPPH color turned pale yellow, which manifested by decreased sample absorption at 520 nm and was measured in conducted experiments using EnSpire microplate reader (PerkinElmer). Antioxidant activity of xanthine derivatives and reference compounds was tested in triplicate at 50 µM concentration and normalized to the activity of reference radical scavenger-quercetin.

Neuroprotection Studies
Neuroblastoma SH-SY5Y cells were seeded at the density of 20,000 cells per well to 96-well plates one day before the experiment. Twelve hours prior to start of the assay, cell culture medium was replaced with the fresh portion of reduced serum (1%) medium. Tested compounds were added to the cells at the final concentration of 10 or 50 µM in 1% DMSO. In the neuroprotection studies, hydrogen peroxide, at a final concertation of 300 µM, was pipetted to the assay plate 1 h after tested compounds addition. Cell death, represented by the release of LDH through the perforated cell membrane, was measured following 14 h incubation of cells with the compounds and hydrogen peroxide. LDH levels in cell culture medium were determined using fluorometric CytoTox-ONE Homogeneous Membrane Integrity Assay (Promega Corporation, Madison, WI, USA), according to manufacturer's instructions.