Discovery of Novel Dual Adenosine A2A and A1 Receptor Antagonists with 1H-Pyrazolo[3,4-d]pyrimidin-6-amine Core Scaffold as Anti-Parkinson’s Disease Agents

New compounds with 1H-pyrazolo [3,4-d]pyrimidin-6-amine core scaffolds were synthesized and characterized in vitro to determine their affinity for human A2A and A1 receptors. Among the tested compounds, a few compounds displayed nanomolar binding affinities for both receptors. One particular compound, 11o, showed high binding activities (hA2A Ki = 13.3 nM; hA1 Ki = 55 nM) and full antagonism (hA2A IC50 = 136 nM; hA1 IC50 = 98.8 nM) toward both receptors. Further tests showed that 11o has low hepatic clearance and good pharmacokinetic properties in mice, along with high bioavailability and a high brain plasma ratio. In addition, 11o was associated with very low cardiovascular risk and mutagenic potential, and was well-tolerated in rats and dogs. When tested in an MPTP-induced mouse model of Parkinson’s disease, 11o tended to improve behavior. Moreover, 11o dose-dependently reversed haloperidol-induced catalepsy in female rats, with graded ED50 of between 3 and 10 mg/kg. Taken together, these results suggest that this potent dual A2A/A1 receptor antagonist, 11o, is a good candidate for the treatment of Parkinson’s disease with an excellent metabolic and safety profile.


Introduction
Parkinson's disease (PD) is the second most prevalent neurodegenerative disease worldwide, but still lacks a curative therapeutic treatment. Loss of dopaminergic neurons in the substantia nigra pars compacta, leading to striatal dopamine depletion, is the core pathophysiological finding in patients with PD. Currently, levodopa (L-DOPA) is the standard of care (SOC) to treat the motor symptoms of PD [1]. Levodopa alleviates the motor symptoms of patients with PD by replacing the depleted dopamine in the brain. However, PD symptoms become increasingly refractory to SOC as the disease progresses, resulting in motor complications, including motor-response oscillations (on-off periods) and drug-induced dyskinesias, probably due to discontinuous drug delivery and its short half-life [2]. Several FDA-approved therapies exist to treat off periods in patients with PD, including levodopa inhalation powder, entacapone tablets, carbidopa/levodopa enteral suspension, carbidopa/levodopa/entacapone tablets, apomorphine hydrochloride injection, safinamide tablets and istradefylline tablets. Despite these options, patients with PD still need therapies that are more efficacious, longer lasting and safer to use.
Adenosinergic modulation of the indirect striatal output pathway remains one of the most well-investigated nondopaminergic approaches to control PD symptoms. Adenosine affects the dopaminergic system via receptor interactions and intracellular signaling in pre-and postsynaptic neurons. Four known subtypes of adenosine receptors have been Istradefylline is highly selective for the A 2A receptor and much less selective for the other adenosine receptors A 1 , A 2B and A 3 [5]. In the brain, presynaptic A 1 receptors on dopaminergic neurons may function synergistically with A 2A receptors in the presence of antagonists: inhibition of the A 1 receptor will facilitate dopamine release, while inhibition of the A 2A receptor will enhance postsynaptic responses to dopamine [6]. Thus, a highly plausible hypothesis is that a dual antagonist of the A 2A /A 1 receptor may exert even more beneficial effects on PD than a selective A 2A antagonist [7,8].
In this study, based on the aforementioned hypothesis, we designed and synthesized novel dual A 2A /A 1 antagonists with a 1H-pyrazolo [3,4-d]pyrimidin-6-amine core scaffold (11o) that has high brain permeability and metabolic stability, along with outstanding safety and pharmacokinetic (PK) profiles. Preladenant and vipadenant exhibited better in vitro activities than istradefylline but were less metabolically stable with relatively shorter PK profile due to the exposed furan on the right wing. We introduced a phenyl or pyridyl ring in place of the furan ring to compensate for this shortcoming (Figure 2). The comparatively low activity observed after replacing the furan ring with the phenyl ring was overcome by introducing a nitrile group on the ring. Similar result was obtained after replacing this group with pyridyl ring, and similar activity compared to the phenyl ring was detected when nitrile group was introduced. Indeed, the activity was maintained only when the nitrile group was placed in the meta position. For core structure design, we initially investigated the representative compounds with core bicyclic ring compounds in which triazole, pyrazole and imidazole were fused to pyrimidine (Figure 2). Among them, the pyrazolopyrimidine ring yielded the best physicochemical property and required the simplest synthetic route. Hence, in the present study, a series of derivatives with pyrazolopyrimidine core structures were investigated as A 2A and A 1 receptor antagonists (4a-b, 8, 11a-x) and drug candidate 11o was finally discovered as a potential treatment for Parkinson's disease.

Chemistry
The synthesis of A 2A and A 1 receptor antagonists with 1H-pyrazolo [3,4-d]pyrimidin-6-amine as a core scaffold (compounds 4a-b, 8, 11a-x) was accomplished using three methods with a short three-step route. The general procedures for the three methods are described in Schemes 1-3. For core structure design, we initially investigated the representative compounds with core bicyclic ring compounds in which triazole, pyrazole and imidazole were fused to pyrimidine (Figure 2). Among them, the pyrazolopyrimidine ring yielded the best physicochemical property and required the simplest synthetic route. Hence, in the present study, a series of derivatives with pyrazolopyrimidine core structures were investigated as A2A and A1 receptor antagonists (4a-b, 8, 11a-x) and drug candidate 11o was finally discovered as a potential treatment for Parkinson's disease.

Chemistry
The synthesis of A2A and A1 receptor antagonists with 1H-pyrazolo [3,4-d]pyrimidin-6-amine as a core scaffold (compounds 4a-b, 8, 11a-x) was accomplished using three methods with a short three-step route. The general procedures for the three methods are described in Schemes 1-3.
For the synthesis of compounds 4a and 4b, the core bicyclic ring was synthesized directly using hydrazine (Scheme 1). The core scaffold, 1H-pyrazolo [3,4-d]pyrimidin-6amine, was generated through ring condensation by reacting hydrazine with 2-amino-4,6dichloropyrimidine-5-carboxaldehyde. Since the pyrimidine reagent has two chloro atoms symmetrical on both sides, the resulting hydrazine obtained after condensing with aldehyde reacts with any chloro group on both sides of the pyrimidine to obtain the same intermediate 2. The chlorine of compound 2 was substituted with 3-benzonitrile groups by Suzuki coupling, and then nitro reduction with iron dust provided the final compounds.
For the synthesis of compound 8, commercially available 4-chloro-1H-pyrazolo [3,4d]pyrimidin-6-amine (5) was reacted with 3-cyanophenylboronic acid by Suzuki coupling to provide compound 6. Alkylation with benzyl halide followed by nitro reduction provided the final product (Scheme 2). For the synthesis of compounds 4a and 4b, the core bicyclic ring was synthesized directly using hydrazine (Scheme 1). The core scaffold, 1H-pyrazolo [3,4-d]pyrimidin-6-amine, was generated through ring condensation by reacting hydrazine with 2-amino-4,6-dichloropyrimidine-5-carboxaldehyde. Since the pyrimidine reagent has two chloro atoms symmetrical on both sides, the resulting hydrazine obtained after condensing with aldehyde reacts with any chloro group on both sides of the pyrimidine to obtain the same intermediate 2. The chlorine of compound 2 was substituted with 3-benzonitrile groups by Suzuki coupling, and then nitro reduction with iron dust provided the final compounds. For the synthesis of compound 8, commercially available 4-chloro-1H-pyrazolo [3,4-d]pyrimidin-6-amine (5) was reacted with 3-cyanophenylboronic acid by Suzuki coupling to provide compound 6. Alkylation with benzyl halide followed by nitro reduction provided the final product (Scheme 2).
For the synthesis of compounds 11a-11x, intermediate 5 was first alkylated with the corresponding benzyl halides, and then substituted phenyl/pyridyl was incorporated by Pd-catalyzed cross-coupling in a different order from Scheme 3. For ortho-substituted cyanopyridyls (A1 and A2, Table 1), classic Suzuki, Stille and Negishi coupling did not work well due to strong electron deficiency. This deficit was resolved by applying Pdcatalyzed desulfination coupling [9,10]. Finally, nitro reduction with iron dust provided the final products. Notably, during nitro reduction, the cyano group of pyridyl was also reduced in Fe/NH 4 Cl and then readily hydrolyzed. A similar observation was recorded under Pd/C hydrogenolysis conditions. However, the problem was solved by shortening the reduction time in Fe/NH 4 Cl, and the reaction was generally completed within 1 h. All final compounds were efficiently purified to >95% purity using normal phase prep-HPLC. well due to strong electron deficiency. This deficit was resolved by a desulfination coupling [9,10]. Finally, nitro reduction with iron du products. Notably, during nitro reduction, the cyano group of pyri in Fe/NH4Cl and then readily hydrolyzed. A similar observation Pd/C hydrogenolysis conditions. However, the problem was solved duction time in Fe/NH4Cl, and the reaction was generally complete compounds were efficiently purified to >95% purity using normal p well due to strong electron deficiency. This deficit was resolved by applying Pd-catalyzed desulfination coupling [9,10]. Finally, nitro reduction with iron dust provided the final products. Notably, during nitro reduction, the cyano group of pyridyl was also reduced in Fe/NH4Cl and then readily hydrolyzed. A similar observation was recorded under Pd/C hydrogenolysis conditions. However, the problem was solved by shortening the reduction time in Fe/NH4Cl, and the reaction was generally completed within 1 h. All final compounds were efficiently purified to >95% purity using normal phase prep-HPLC. well due to strong electron deficiency. This deficit was resolved by applying Pd-catalyzed desulfination coupling [9,10]. Finally, nitro reduction with iron dust provided the final products. Notably, during nitro reduction, the cyano group of pyridyl was also reduced in Fe/NH4Cl and then readily hydrolyzed. A similar observation was recorded under Pd/C hydrogenolysis conditions. However, the problem was solved by shortening the reduction time in Fe/NH4Cl, and the reaction was generally completed within 1 h. All final compounds were efficiently purified to >95% purity using normal phase prep-HPLC. desulfination coupling [9,10]. Finally, nitro reduction with iron dust provided the final products. Notably, during nitro reduction, the cyano group of pyridyl was also reduced in Fe/NH4Cl and then readily hydrolyzed. A similar observation was recorded under Pd/C hydrogenolysis conditions. However, the problem was solved by shortening the reduction time in Fe/NH4Cl, and the reaction was generally completed within 1 h. All final compounds were efficiently purified to >95% purity using normal phase prep-HPLC. desulfination coupling [9,10]. Finally, nitro reduction with iron dust provided the final products. Notably, during nitro reduction, the cyano group of pyridyl was also reduced in Fe/NH4Cl and then readily hydrolyzed. A similar observation was recorded under Pd/C hydrogenolysis conditions. However, the problem was solved by shortening the reduction time in Fe/NH4Cl, and the reaction was generally completed within 1 h. All final compounds were efficiently purified to >95% purity using normal phase prep-HPLC.  [11]. As shown in Table 1, we investigated the structure-activity relationship (SAR) for the A-group (cyanopyridyl or cyanophenyl) and the R 1 , R 2 and R 3 of benzylamine substituted to the core scaffold toward human A 2A receptors (hA 2A R).
In the SAR of R 1-3 on the benzylamine, compounds with unsubstituted benzylamine rings (11a-d and 4a-b) showed a low binding affinity to the hA 2A receptor, whereas CF 3 on R 2 showed the best result. The order of change in activity was CF 3 > di-F > Me > F > H. A similar pattern was obtained for both pyridyl and phenyl groups, irrespective of the A group.
Next, we examined the SAR of A group. In the pyridyl A group, the activity was substantially affected by the position of the 'N' on the pyridyl ring. Among the series of compounds A1-6, 6-picolinonitrile derivatives (A2) showed the best activity, followed by 2-isonicotinonitrile-substituted compounds (A1) with moderate activity. 5-Nicotinotrile (A3) and 4-picolinonitrile derivatives (A4) displayed very low activity. In the phenyl A-region, they all exhibited promising activity. Little significant change in activity was observed between 3-benzonitrile (A5) and 2-fluoro-3-benzonitrile (A6). Overall, the cyanophenyl group showed slightly better activity than the cyanopyridyl group in the SAR of the A group.
Among the compounds tested, all compounds with A2, A5 and A6 groups displayed good activities. However, cyanopyridyl groups A1-A4 showed less inhibitory activity toward hA 1 receptor than compounds with cyanophenyl groups (A5 and A6, unpublished data). Recent findings reported using PD animal models suggested that affinity for hA 1 receptors might positively contribute to PD [6,8]. Therefore, we hypothesized that the use of cyanophenyl groups (A5 and A6) may exert a higher synergistic effect than cyanopyridyl (A2) on PD.
From the list of compounds with K i value of less than 30 nM for A 2A receptor, we have further selected for compounds with K i value below 100 nM for A 1 receptor. In addition, it was deemed necessary to select a highly soluble compound because only intranasal delivery was considered to maximize compound's brain exposure. Compounds with CF 3 , a strong electron withdrawing group (EWG), next to aniline, such as 11v, 11w and 11x, have greatly reduced solubility due to increased difficulties in forming salt. In the case of 11t and 11u where fluoro group was substituted in R 1 and R 2 , salt formation was also not feasible due to strong EWG. Among the selected compounds capable of salt formation, 11o displayed the best binding affinity for both A 2A and A 1 receptors. Thus, we have selected 11o as a potential drug candidate and conducted further studies.
Next, the functional activity of 11o against A 1 and A 2A adenosine receptors was evaluated using c-AMP assay in CHO-K1 (A 1 ) and HEK-293 (A 2A ) cells. As shown in Figure 3, 11o displayed a full antagonism with IC 50 values of 98.8 nM and 136 nM in A 1 and A 2A adenosine receptors, respectively, indicating that 11o is a dual A 1 /A 2A antagonist.

Selectivity for Adenosine Receptor Subtypes
We examined and compared the binding activities of 11o toward four adenosine receptors, A 1 , A 2A , A 2B and A 3 , to investigate its selectivity for adenosine receptors.
Whereas 11o showed relatively high selectivity toward both receptors A 1 and A 2A , with K i values of 55 nM and 13.3 nM, respectively, it exerted low binding affinities on A 2B and A 3 receptors with K i values 0.4 µM and 1.05 µM, respectively, indicating that compound 11o is a dual specific A 2A /A 1 receptor antagonist (Table 2).

In Vivo Activity in Animal Models of Parkinson's Disease
In a subsequent study, we investigated the therapeutic efficacy of 11o in PD using an MPTP-lesioned mouse model [12]. This MPTP-lesion mouse model is one of the most well-established PD models and causes an approximately 90% loss of striatal dopamine, as analyzed using HPLC 7 days after the administration of the last dose of MPTP [13].
When 11o was administered 30 min after the last MPTP treatment and once daily for 7 days thereafter, it was unable to rescue dopamine depletion caused by MPTP (data not shown) at doses up to 10 mg/kg. However, when mice were tested for nest-building activity after the intranasal administration of 1 and 10 mg/kg doses, compound 11o partially restored the activity level to near sham level, although a significant difference was not observed (Figure 4). Taken together, compound 11o positively affected the behavioral outcome of MPTP-lesioned mice without affecting the dopamine level. An interesting approach would be to determine the synergistic effect of L-DOPA and 11o in a 6-OHDA rat model, which is currently being investigated by our group. Nest-building score of sham or MPTP-lesioned mice treated either with vehicle, rasagiline or compound 11o (intranasal). The graph presents the mean nesting scores ± SEM (n = 12 rats per group). Data were analyzed using the Kruskal-Wallis test and Dunn's multiple comparison test (all mean values were compared to MPTP, vehicle). * p < 0.05 and ** p < 0.01.
In another experiment, compound 11o resulted in a dose-dependent reversal of catalepsy in female rats induced by haloperidol ( Figure 5). Vehicle or varying amount of 11o (0.3, 1, 3, 10 mg/kg) was administered first, followed by IP injection of 1 mg/kg haloperidol 60 min later. When catalepsy was assessed 60 min after haloperidol injection by measuring latency to step down from the rod, 11o dose-dependently reversed haloperidol-induced catalepsy, with maximum effect observed at 10 mg/kg, which was comparable to the effect observed with positive control istradefylline (1 mg/kg). . Latency to step-down in female rats injected with haloperidol (HP). Rats were administered with 11o (PO) 60 min before haloperidol injection at increasing dose levels. Istradefylline (1 mg/kg, PO) was used as positive control. The graph presents the mean latency (seconds) ± SEM (n = 10 rats per group). Data were analyzed using the one-way ANOVA with Dunnett's multiple comparisons test (all mean values were compared to vehicle). * p < 0.01.

In Vitro Metabolic Stability
In vitro intrinsic clearance (CL int ) was investigated using cryopreserved hepatocytes from five different species-CD-1 mice, Sprague Dawley rats, beagle dogs, cynomolgus monkeys and humans-to understand the metabolic stability of 11o (Table 3).  The pharmacokinetic parameters of 11o in mouse plasma are described in Table 4. T max was 2 h for all doses following PO dosing. T 1/2 values were 4.38 h, 3.66 h and 5.69 h for doses of 1, 3 and 10 mg/kg, respectively. C max values were 36.8 ng/mL, 81.4 ng/mL and 367 ng/mL for doses 1, 3 and 10 mg/kg, respectively, and AUC last values were 176 h·ng/mL, 577 h·ng/mL and 2248 h·ng/mL for the three doses. Following IV dosing, T max was 0.083 h, and the T 1/2 value was 3.95 h. C max and AUC last values were 459 ng/mL and 332 h·ng/mL, respectively. Intranasal (IN) dosing resulted in a shorter T max of 0.5 h for all groups, indicating that the compound reached the system faster via this route compared to PO. The T 1/2 value was 7.4 h for 10 mg/kg dose group. The T 1/2 values for the other 2 dose groups were not calculated due to insufficient data points in the elimination phase or poor goodness-of-fit (R 2 < 0.8) for the elimination phase. The C max values were 23.3 ng/mL, 53.9 ng/m and 88.5 ng/mL for doses of 1, 3 and 10 mg/kg, respectively, with AUC last values of 74.9, 183 and 444 h·ng/mL, respectively.
Oral bioavailability was assessed relative to the AUC last for the IV dose group, and the data are presented in Table 5. Bioavailability of 11o after oral dosing was 53%, 57.9% and 67.7% for the 1, 3 and 10 mg/kg dose groups, respectively, indicating good bioavailability. Bioavailability after IN administration was relatively poorer, ranging from 13.4-22.6%, depending on the dose. The pharmacokinetic parameters of 11o in the mouse brain are described in Table 5

Toxicity Studies
hERG (human-ether-a-go-go) liability was investigated in an hERG-expressing HEK cell line using the manual patch clamp method to assess the potential cardiotoxicity of 11o [15]. The IC 50 was 116 µM, suggesting a low hERG liability of the compound at physiologically relevant dose (Table 6). The mutagenic potential of 11o was also tested by assessing its ability to reverse mutations in a bacterial mutagenesis study [16,17]. 11o did not cause any positive mutagenic response when applied at up to 1867 µg/plate, suggesting that the compound is not mutagenic.
Next, we investigated the maximum tolerated dose (MTD) of 11o in rats and dogs. In male and female Sprague Dawley rats, MTD following single oral administrations of 11o was >1000 mg/kg, with no body weight loss or mortality observed. When MTD was assessed in male and female beagle dogs, the value obtained following single oral administrations of the compound was >400 mg/kg, with slight decreases in body weight in both males (−5.7%) and females (−6.5%) that were not considered adverse and test article-related. These results indicate that a single dose of 11o was well-tolerated up to 1000 mg/kg and 400 mg/kg in rats and dogs, respectively.
To a solution of 2-amino-4,6-dichloropyrimidine-5-carboxaldehyde (1.0 eq) in DMF TEA (5.0 eq) and 4-nitrophenylhydrazine 1 (1.0 eq, 2 HCl) were added at 0 • C, and then the mixture was stirred at 25 • C for 1 h. The reaction mixture was diluted with Ethyl acetate and washed with brine. The organic layer was concentrated to give a brown solid.

In Vitro Human Adenosine Receptor-Binding and Functional Assays
The radioligand-binding assays for adenosine receptors were performed at Eurofins Discovery Services (Panlabs, Taiwan). Briefly, stock solutions of the compounds were prepared in DMSO and further diluted with the incubation buffer to the desired concentration. Final DMSO concentrations in the assay were 1.0%. IC 50 values were determined by a nonlinear, least-squares regression analysis using MathIQ TM (ID Business Solutions Ltd., Guildford, UK). Where inhibition constants (K i ) were presented, the K i values were calculated using Cheng and Prusoff equation [18] using the observed IC 50 of the compound, the concentration of radioligand employed in the assay, and the historical values for the K D of the ligand (obtained experimentally at Eurofins Panlabs, Inc.). The Hill coefficient (n H ), which defines the slope of the competitive binding curve, was calculated using MathIQ TM . Hill coefficients significantly different than 1.0, may suggest that the binding displacement does not follow the laws of mass action with a single binding site.
For functional studies, A 1 and A 2A receptor cyclic AMP HTRF (homogeneous time resolved fluorescence) assay was performed at EuroscreenFast (Charleroi, Belgium) using CHO-K1 (A 1 ) and HEK293 (A 2A ) stable cell line with CPA (A 1 ) and NECA (A 2A ) as reference agonist and DPCPX (A 1 ) and ZM241385 (A 2A ) as reference antagonist compound.

hERG, AMES and Metabolic Stability Study
hERG ion channel-blocking profile of acute exposure to 11o was performed in stably expressing HEK cells at Nova Research Laboratories (New Orleans, LA, USA). Briefly, cloned equivalent of the human IKr, hERG, was used. Cells were obtained from Cytogentrics Bioscience GmbH (Jachim-Jungius-Strafe 9, 18,059 Rostock, Germany). Internal and external recording solutions as well as data acquisition and analysis were as previously described [15].
AMES bacterial reverse-mutation test was performed at Frontage laboratories inc. (Frontage laboratories, Concord, OH, USA) according to testing facility's standard operating procedures (SOPs) following OECD guideline 471, "Bacterial Reverse Mutation Test", and ICH S2(R1), "Guidance on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use" [19]. Briefly, test system was exposed to test article via plateincorporation methodology originally described by Ames et al. [16] and updated by Maron and Ames [17]. This test system has been shown to detect a wide range of classes of chemical mutagens [16,20]. For mutagenicity assay, TA98 and TA100 (tester strains source: Molecular Toxicology Inc. (Boone, NC, USA). TA98 lot number 5464D, part number 71-098L; TA100 lot number 5439D, part number 71-100L) was exposed to vehicle alone and at least 8 concentrations of 11o, with duplicate plates/condition, in both the presence and absence of S9 using the plate incorporation method. Unless limited by solubility, 11o was to be evaluated at a maximum concentration of 5000 µg/plate (the dose levels were 5000, 1581, 500, 158, 50, 15.8, 5.0 and 1.5 µg/plate). However, 11o showed limited solubility and the highest tested concentration permissible as a workable preparation was 1867 µg/plate. Following incubation, plates were examined for the condition of the bacterial background lawn and precipitate and recorded. Colonies were enumerated either by visually or by an automatic colony counter (Scan 1200 Automated colony counter and scan software).
Metabolic stability of 11o was determined in cryopreserved hepatocytes from mice, rats, dogs, monkeys and human according to the SOPs of Frontage Laboratories Inc. Briefly, cryopreserved hepatocytes for pre-clinical animals and for human (BioIVT, Westbury, NY, USA) were thawed in a 37 • C water bath with gentle shaking until ice had almost melted. The suspensions were immediately transferred to a centrifuge tube (50 mL) containing pre-warmed thawing media at 37 • C with gentle handshaking to prevent the cells from settling. Cell suspensions were centrifuged at 50× g and 4 • C for 5 min. The supernatant was discarded and the pellets were resuspended in prewarmed incubation media (5 mL) at 37 • C. The percentage of viable cells in the suspension was determined by Trypan blue staining method to ensure viability of ≥70 %. 11o was incubated in a 24-well plate containing hepatocyte suspensions (1million cells/mL) at 37 • C for 4 h under 5% CO 2 and 95% air. Aliquots were removed in duplicate at 0, 15, 30, 60, 120, 180 and 240 min and the samples were treated with ACN (3 volumes containing an internal standard). The samples were centrifuged and aliquots of the supernatant were analyzed for parent compounds by LC/MS/MS using suitable multiple reaction monitoring (MRM) transitions. Midazolam (0.2 µM) was included as positive control to confirm the viability of the hepatocytes from each species. The disappearance of 11o and midazolam was monitored by LC/MS/MS (MRM) analysis of the hepatocyte extracts at each time point. The peak area ratio of the analyte to internal standard was obtained for each sample and the rate of disappearance of 11o was assessed by comparing the peak area ratios at various time points to the peak area ratio obtained at t = 0. The t 1/2 is obtained (using the linear portion of semi-log plot of the area ratio vs. time) and the CL int values determined using well-stirred model. As soon as the animals arrived, they were brought to the assigned animal room, unpacked and checked for their health status. Animals were habituated at least for one week prior to study start. Animals were single-housed in individual ventilated cages on standardized rodent bedding supplied by Rettenmaier. The room temperature was maintained at 20~24 • C and the relative humidity was maintained between 45 to 65%. Animals were housed under a constant light cycle (12 h light/dark). Dried, pelleted standard rodent chow (Altromin) as well as normal tap water was available to the animals ad libitum.
Animals with apparently good health conditions were included to the study. All animals were randomly assigned to starting groups (cohort). The number of animals in a starting group was limited to ensure same age and uniform handling. A total of 144 male C57BL/6J mice were allocated to 12 groups (n = 12/group). Eleven groups (groups 2~12) were intraperitoneally (IP) treated four times with 20 mg/kg MPTP with 2 h interval on day 1. Probenecid or vehicle was administered via IP injection (250 mg/kg) with the first MPTP or vehicle treatment (e.g., MPTP left abdomen, Probenecid right abdomen). One group received vehicle only (group 1) and served as control. Group 4 received a positive control/reference item (rasagiline) via IP injection 30 min before the first MPTP injection. Thirty minutes after the last MPTP treatment the first compound treatment was performed. Groups received either vehicle (group 2 PO and 3 IN), four different doses of 11o via oral gavage (groups 5~8) or intranasal (groups 9~12). Animals were treated with the compound or vehicle for 7 consecutive days. For PO, each mouse received 10 µL/g body weight of compound. For IN, the body weight of animals was averaged to 30 g. This was basis for concentration calculation for the whole treatment period. Each animal received a volume of 5 µL per naris (10 µL per animal).

Nest Building
Behavior assessment was performed on day 3, two hours after the scheduled treatment. All animals were tested in nest-building and open-field testing in a randomized order.
To test individual nest-building behavior, mice were housed individually in cages containing wood-chip bedding and one square of pressed cotton ('nestlet'). No other nesting material (e.g., wood wool) was present. The nestlet was introduced on the day before the evaluation of the nest status in about 2 to 3 h before the dark phase was initiated, and the nest-building behavior was evaluated on the following day of the experiment within 2~3 h after the light-phase start. Time span between introduction of the cotton square and evaluation of the nest status was same for all examinations. Manipulation of the nestlet and the constitution of the built nest were assessed according to five-point scale [21] as follows: 1. Nestlet not noticeably touched (>90%) intact); 2. Nestlet partially torn up (50%~90% remaining intact); 3. Mostly shredded but often not identifiable nest site: <50% of the nestlet remains intact but <90% is within a quarter of the cage floor area, i.e., the cotton is not gathered into a nest, but spread around the cage; 4. An identifiable, but flat nest: >90% of the nestlet is torn up, the material is gathered into a nest within a quarter of the cage floor area, but the nest is flat, with walls higher than mouse body height (curled up on its side) on less than 50% of its circumference; 5. A (near) perfect nest: >90% of the nestlet is torn up, the nest is a crater, with walls higher than mouse body height on more than 50% of its circumference.

Haloperidol-Induced Catalepsy in Female Rats
A total of 120 female Sprague Dawley rats (275~300 g) were purchased from Charles River Laboratories (Charles River, Senneville, QC, Canada) and maintained at Atuka, Inc. All animal studies were conducted at Atuka, Inc. according to CCAC guidelines and under IACUC-approved Animal Use Protocols (AUPs), with IACUC number 6606.0.3 for this study.
A one-week period of acclimatization was allowed between delivery of rats and commencement of treatments. Animals were weighed on the first day of acclimatization and then on a weekly basis thereafter. Rats were housed 2/cage at standard temperature (21 ± 2 • C) in a light-controlled environment (lights on 6:00 am to 6:00 pm) with access to food (Teklad 7912, Harlan, Madison, WI, USA) and water ad libitum. The studies were conducted according to CCAC guidelines and under IACUC-approved Animal Use Protocols (AUPs).
The catalepsy study incorporated 12 groups of rats. Ten rats were incorporated into each treatment group (total N = 120 rats). Treatments were divided over 3 consecutive treatment blocks and the order of treatments were randomized within each block and all treatments completed within that block before proceeding to the next. Vehicle (0.5% methyl cellulose, Sigma-Aldrich, St-Louis, MI, USA), 11o or istradefylline was administered PO at t = 0 min and followed 60 min later (t = 60 min) by haloperidol (1 mg/kg, SC). Catalepsy was assessed a further 60 min later (t = 120).
Behavior was evaluated in each animal by assessing performance on the bar test. Catalepsy was assessed 60 min following haloperidol administration (120 min following administration of vehicle, 11o or istradefylline). To assess catalepsy, animals were positioned to place both front paws upon a dowel rod (approximately 0.7 cm diameter) suspended 6 cm above a stable surface. The time taken for the animal to remove both paws (decent latency) from the rod and place them back on the surface, in seconds was determined. The test was conducted in triplicate and the average of the three tests were calculated.

In Vivo Pharmacokinetics, Bioavailability and Brain Plasma Ratio Study in Mouse
Male Crl:CD-1 mice of 9~10 weeks of age were purchased from Charles River Laboratories, USA and maintained at standard food and water ad libitum in temperature-controlled room with 12 h on-off light cycle before the experiment. All procedures described were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Frontage Laboratories Inc.
For PO dosing, mice were given 11o dissolved in DMSO: PEG400 (1:3, v:v) via gavage. For IV dosing, 11o solubilized in DMSO: PEG400: saline (1:3:6, v:v) were injected into tail veins at dose volume of 1 mL/kg. For IN dosing, 11o solubilized in DMSO and 25 µL of the appropriate dosing solution was placed near the nares to allow animal to inhale the dose. Three mice were sampled from each group for each blood and brain collection time point. Time points are 0.5, 1, 2, 4, 8, 24 h for PO and IN dosing and 0.083, 1, 2, 4, 8 and 24 h for IV dosing. Mice were euthanized by terminal retro-orbital blood collection under isoflurane anesthesia, followed by opening of the thoracic cavity per facility SOPs. A target volume of 1 mL was collected into tubes containing K 3 EDTA anti-coagulant, inverted several times and held on wet ice until centrifuged under refrigeration (set at 5 • C for 10 min at 2000× g). Resulting plasma were stored in a freezer set to maintain −70 • C until analyzed for 11o concentration. Brain was removed from each animal and weighed following termination, snap frozen and stored until later use.
Plasma and brain PK analysis was conducted using WinNolin Version 6.2.1 (Pharsight, Mountain View, CA, USA), operating as a validated software system. Noncompartmental analysis was conducted using the extravascular administration model for oral dosing and the IV bolus model for IV dosing. The peak plasma and brain concentration, time to achieve peak plasma concentration, half-life, and area under the plasma concentration-time curve (C max , T max , T 1/2 and AUC) were calculated from mean plasma concentrations for each sampling time/dose group for 11o. Nominal blood collection times were used for calculation of PK parameters. Dose proportionality was assessed, and bioavailability was calculated for PO dosing.

Maximum Tolerated Dose Study in Rat and Dog
Rat and dog MTD studies were performed at Frontage Laboratories Inc following testing facility's SOPs. All procedures described were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Frontage Laboratories Inc.
For rat MTD study, male and female Sprague Dawley rats (Rattus norvegicus) of 8.1~9.6 weeks of age were purchased from Charles River Laboratories, USA and maintained at standard food and water ad libitum in temperature-controlled room with 12 h on-off light cycle before the experiment. Animals were acclimated to the testing facility for at least 6 to 16 days prior to initiation of dosing. Animals received a single dose of 11o in 0.5% methyl cellulose via oral gavage starting from 150 mg/kg dose level (this initial dose level was selected based on available data and the estimated toxicity using EPA's toxicity estimation software tool (TEST) version 4.2.1) in 10 mL/kg dose volume. Subsequent doses occurred 3 or 4 days following the previous dose and were increased based on findings from the previous dosing(s). The procedure continued until four dosing events occurred. Animals were observed once in the morning and once in the afternoon throughout the study for clinical observations. Body weights were measured on the day prior to dose administration. Data were collected and reported electronically using Provantis TM , (Instem LSS Ltd. Staffordshire, UK).
For dog MTD study, male and female beagle dogs (Canis familiaris) of 1.3 to 1.4 years of age were from Frontage Laboratories' test facility stock colony, originally procured from Marshall BioResources (North Rose, NY, USA). The health status of the animals was reviewed by a staff veterinarian. Data collected on animals prior to release to study were maintained as testing facility records. Same dogs were used for each dose event. Animals received a single dose of 11o in 0.5% methyl cellulose via oral gavage starting from 50 mg/kg dose level in 5 mL/kg dose volume. Subsequent doses occurred 3 or 4 days following the previous dose and were increased based on findings from the previous dosing(s). The procedure continued until four dosing events occurred. Animals were fasted overnight prior to each dose administration with food provided approximately 4 h postdose. Animals were observed once in the morning and once in the afternoon throughout the study for clinical observations. Body weights were measured on the day prior to dose administration. Data were collected and reported electronically using Provantis TM (Instem LSS Ltd., Staffordshire, UK).

Statistical Analysis
Statistical analysis was performed using GraphPad Prism 8.3.0. Data were tested for normality using the Kolmogorov-Smirnov test. If normal distribution was confirmed, differences between groups were tested with the one-way ANOVA followed by either Bonferroni or Dunnett's post hoc analysis. If data were not normally distributed, differences between groups were tested with Kruskal-Wallis test, followed by Dunn's post hoc test for multiple comparisons. Data were presented as mean ± SEM.

Conclusions
In this study, a series of novel adenosine A 2A receptor antagonists with 1H-pyrazolo [3,4-d]pyrimidin-6-amine core scaffolds was designed and synthesized as potential anti-Parkinson's disease agents using efficient synthetic methods with three-step short routes. Among them, compound 11o with a 2-fluoro-3-cyanophenyl A group exhibited high selectivity for A 2A and A 1 receptors with full antagonism for both A 1 and A 2A receptors. Animal studies indicated that this compound displayed a positive tendency toward improving nesting behavior in MPTP-induced mouse model of Parkinson's disease, and a dose-dependent reversal of catalepsy induced by haloperidol in female rats. In vitro metabolism study in hepatocytes from mice, rats, dogs, monkeys and humans showed that the compound exhibited relatively low clearance in these species. When administered orally and intranasally to mice, 11o exhibited excellent bioavailability, good brain/plasma ratio and PK profile. Compound 11o had low to no cardiotoxicity risk and mutagenic potential, with MTD in rats and dogs of >1000 mg/kg and >400 mg/kg, respectively. Taken together, these results indicate that 11o is a dual A 2A /A 1 receptor antagonist, and a good potential drug candidate for treating Parkinson's disease that merits further exploration.