A2A receptor antagonists

Bioorganic Chemistry 77 (2018) 136–143

Contents lists available at ScienceDirect

Bioorganic Chemistry journal homepage: www.elsevier.com/locate/bioorg

Benzopyrone represents a privilege scaffold to identify novel adenosine A1/A2A receptor antagonists Mietha M. van der Walt a,b,⇑, Gisella Terre’Blanche a,c a

Centre of Excellence for Pharmaceutical Sciences, School of Pharmacy, North-West University, Private Bag X6001, Potchefstroom 2520, South Africa Human Metabolomics, North-West University, Private Bag X6001, Box 269, Potchefstroom 2520, South Africa c Pharmaceutical Chemistry, School of Pharmacy, North-West University, Private Bag X6001, Potchefstroom 2520, South Africa b

a r t i c l e

i n f o

Article history: Received 7 July 2017 Revised 7 December 2017 Accepted 2 January 2018 Available online 4 January 2018 Keywords: Benzopyrone Adenosine A2A receptor Adenosine A1 receptor Antagonist GTP shift Radioligand binding assay Cancer Parkinson’s disease

a b s t r a c t Adenosine receptor antagonists are under investigation as potential drug candidates for the treatment of certain cancers, neurological disorders, depression and potentially improve tumour immunotherapy. The benzo-c-pyrone scaffold is well-known in medicinal chemistry with diverse pharmacological activities attributed to them, however, their therapeutic potential as adenosine receptor antagonists have not been investigated in detail. To expand on the structure–activity relationships, the present study explored the adenosine A1 and A2A receptor binding affinities of a selected series of benzo-c-pyrone analogues. In vitro evaluation led to the identification of 5-hydroxy-2-(3-hydroxyphenyl)-4H-1-benzopyran-4-one with the best adenosine A2A receptor affinity among the test compounds and was found to be non-selective (A1Ki = 0.956 µM; A2AKi = 1.44 µM). Hydroxy substitution on ring A and/or B play a key role in modulating the binding affinity at adenosine A1 and A2A receptors. Adenosine A1 receptor affinity was increased to the nanomolar range with hydroxy substitution on C6 (ring A), while meta-hydroxy substitution on ring B governed adenosine A2A receptor affinity. The double bond between C2 and C3 of ring C as well as C2 phenyl substitution was shown to be imperative for both adenosine A1 and A2A receptor affinity. Selected benzo-c-pyrone derivatives behaved as adenosine A1 receptor antagonists in the performed GTP shift assays. It may be concluded that benzo-c-pyrone based derivatives are suitable leads for designing and identifying adenosine receptor antagonists as treatment of various disorders. Ó 2018 Elsevier Inc. All rights reserved.

1. Introduction There’s a world-wide interest in identifying potential adenosine receptor (AR) drug candidates, since a variety of pharmacological applications are attributed to them. Currently four AR subtypes are known, namely A1, A2A, A2B and A3 [1]. The A1 and/or A2A AR subtypes are potential drug targets for Parkinson’s disease (A1 and A2A AR) [2], Alzheimer’s disease (A1 AR) [3], depression (A2A AR) [4], certain cancer therapies (A1 and A2A AR) [5] and improving tumour immunotherapy (A2A AR) [5]. Initially, the A1 and A2A AR subtypes have been recognized as promising drug targets for the treatment of neurodegenerative disAbbreviations: AR, adenosine receptor; GTP, guanosine triphosphate; [3H] DPCPX, [3H]-dipropyl-8-cyclopentylxanthine; [3H]NECA, N-[3H]-ethyladenosin-50 uronamide; CPA, N6-cyclopentyladenosine. ⇑ Corresponding author at: Human Metabolomics, North-West University, Private Bag X6001, Box 269, Potchefstroom 2520, South Africa. E-mail addresses: [email protected], [email protected] (M.M. van der Walt). https://doi.org/10.1016/j.bioorg.2018.01.004 0045-2068/Ó 2018 Elsevier Inc. All rights reserved.

orders [6,7]. It is estimated that between 8.7 and 9.3 million individuals will be diagnosed with Parkinson’s disease by 2030, making this disease the 2nd most common neurological condition world-wide [8]. Selective A1 AR antagonists have shown potential to treat Parkinson’s disease associated cognitive impairment [9]. In turn, selective A2A AR antagonists may improve motor dysfunction of Parkinson’s disease [10], exhibit neuroprotective properties [11] and lower the risk of developing dyskinesia [12]. Thus dual antagonism of the A1 and A2A ARs may find therapeutic value as Parkinson’s disease therapy. The safety profile of AR antagonists has been documented in clinical trials intended to explore the validity of these drugs as Parkinson’s disease treatment. For example, istradefylline (a selective A2A AR antagonist) was studied in phase III clinical trials and is currently approved as Parkinson’s disease therapy in Japan [13]. More recently, ARs—especially the A2A AR subtype—have been implicated as a potential drug target for the development of anticancer treatment and improving tumour immunotherapy [5]. Hepatocellular carcinoma (HCC) is a common malignancy in

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

developing countries with hepatitis C and B as risk factors [14]. It’s the 5th most common cancer world-wide [15] with the majority of cases occurring in Asia and Africa [14]. Approximately 600,000 deaths are attributed to HCC per year [14]. In vivo studies with rodents under normoxic and hypoxic conditions together with HCC (Hep3B) cells in culture have shown that adenosine via A2A ARs enhance production of erythropoietin (EPO) [16,17]. This suggests that A2A AR antagonists are a promising novel therapy to inhibit increased levels of EPO. It is therefore plausible that A2A AR antagonists may find therapeutic relevance in the treatment of HCC and other EPO-secreting tumours (e.g. renal cell carcinoma and cerebellar hemangioblastoma) [5]. Activation of the ARs by specific ligands, agonists or antagonists, modulates tumour growth via various signalling pathways. The A1 AR play a role in preventing the development of glioblastomas, where this antitumor effect is mediated via tumour-associated microglial cells. Activation of the A2A AR results in inhibition of the immune response to tumours via suppression of T regulatory cells and by blocking A2A AR, anti-tumour immunity is enhanced [18]. Benzopyrones are a class of compounds with significantly diverse biological activities and for this reason is considered a privileged scaffold in medicinal chemistry [19]. Chemically, the benzopyrones represent a class of ketone containing benzopyrone derivatives which constitute the basic framework of various natural occurring and/or synthetic compounds such as flavonoids (benzo-c-pyrones) [20,21] and the structurally related coumarins [19] and isocoumarines (benzo-a-pyrones) [22] (Fig. 1). Based on their chemical structure, flavonoids are primarily divided into six subgroups namely flavones, isoflavones, flavanones, flavanols and anthocyanidins [23]. Flavones and isoflavones possess a basic benzo-c-pyrone skeleton (ring A and C is fused), with the substitution position of the phenyl side chain (ring B), at the pyrone core (ring C), dividing these flavonoids into flavones (C2 position) and isoflavones (C3 position) (Fig. 1; Table 1). In turn, flavanones are structurally related to flavones, however, they lack the double bond between the C2 and C3 position of the pyrone core (ring C) (Fig. 1; Table 1). Flavonoids are found naturally in plants (including fruits and vegetables) and food prepared from flavonoid containing plants [21]. Consumption of flavonoids through the diet is essential, since animals and humans cannot synthesize flavonoids [24]. The beneficial effects of flavonoids, however, extend beyond dietary necessity. Various pharmacological activities have been attributed to flavonoids in the past, for instance antibacterial activity [25], anti-fungal properties [25,26], anti-inflammatory effects [26,27], antioxidant activity [20,26,28], antiviral activity [25,29], cancer chemopreventative agents [30] and hepatoprotective activity [26]. To further emphasize the therapeutic potential of flavonoids, these compounds have been under investigation for the treatment of Parkinson’s disease as AR antagonists [31,32,33,34,35] and for their ability to exhibit neuroprotective properties [23]. The neuroprotective mechanism of flavonoids has been reviewed previously and may be ascribed to their ability to suppress lipid peroxidation, to inhibit inflammatory mediators, to activate endogenous antiox-

137

idant enzymes, to modulate gene expression in neuronal cells and as a mitochondrial target therapy [23]. The AR antagonistic properties of selected compounds bearing a benzo-c-pyrone backbone were demonstrated by Ji and colleagues [35], where the flavone derivative 2-phenyl-4H-benzopyran-4-one (1a) was reported as a non-selective AR antagonist exhibiting dissociation constant (Ki) values of 3.28 lM and 3.45 lM for the A1 and A2A ARs, respectively [35]. In another study by Jacobson and co-workers [34], only the A1 AR affinity of selected isoflavone derivatives were explored and it was found that 7-hydroxy-3-(4-hydroxyphenyl)-4H-benzopyran4-one (2c daidzein) possess moderate A1 AR binding [34]. Based on the above, the benzo-c-pyrone backbone is considered a promising lead scaffold for identifying novel A1 and/or A2A AR antagonists, however, their therapeutic potential as AR antagonists have not been investigated in detail. Since, effective antagonism of the A1 and/or A2A ARs are deemed beneficial for a variety of disorders, including Parkinson’s disease and anti-cancer therapy, the current study aim to further explore the structure-activity relationships for benzopyrone-based compounds and to discover high affinity AR antagonists. The present study investigated the A1 and A2A AR binding affinities of a selected series of known benzo-cpyrone analogues (flavone, flavanone, isoflavone and thioflavanone). In addition, the necessity of a benzo-c-pyrone scaffold to govern AR affinity was further explored by comparing the Ki values of structurally related benzo-c-pyrones to selected benzo-apyrone moieties. Table 1 summarize the experimentally determined A1 and A2A AR binding affinities (via radioligand binding assays) of the selected benzo-c-pyrone (1a–g; 2a–c and 3a–c) and benzo-a-pyrone (4a–c) derivatives, respectively. 2. Results and discussion 2.1. Chemistry All the test and reference compounds were obtained from standard commercial sources (Sigma Aldrich). These compounds were of analytical grade and used without further purification. 2.2. Radioligand binding assays The affinities of the selected benzo-c-pyrone (1a–g; 2a–c and 3a–c) and benzo-a-pyrone (4a–c) derivatives at rat A1 and A2A AR subtypes were determined with radioligand competition experiments as described previously [36,37]. The test compounds (1a–g, 2a–c, 3a–c and 4a–c) displayed varying degrees of affinity and selectivity towards the A1 and A2A AR subtypes (Table 1). The flavone derivatives (1a–g) were identified as the most promising benzo-c-pyrone based compounds, since this class exhibited the highest A1 and A2A AR affinities among the test compounds. The Ki values for the A1 AR activity of the flavones (1a–g) ranged between 0.59 and 2.25 µM. Furthermore, five of the seven flavones (1a, 1c–f) displayed A2A AR affinity, with Ki values ranging between 1.44 and 5.14 µM. On the other hand, the benzo-apyrone derivatives (4a–c) presented one isochromenone, compound

Fig. 1. The general chemical structures of benzo-c-pyrone (flavonoids) and benzo-a-pyrone (isocoumarins and coumarins) derivatives.

138

Table 1 The Ki values for the binding of the investigated benzopyrones at rat A1 and A2A ARs. Ki ± SEM (µM)a/(% displacement)b c

3

A1 vs [ H]DPCPX

c

3

A2A vs [ H]NECA

Ki ± SEM (µM)a/(% displacement)b SI

d

Benzo-c-pyrones Flavones

A1c

vs [3H]DPCPX

A2Ac vs [3H]NECA

SId

Benzo-c-pyrones Flavanones 5.14 ± 0.69a 3.45e

2.3

18.2 ± 1.8a

>100 (52%)b

–

1.94 ± 0.17a 2.17e

>100 (57%)b 6.20e

–

3.55 ± 0.44a

>100 (38%)b

–

>100 (44%)b

>100 (62%)b

–

Thioflavanone

0.592 ± 0.011a

3.93 ± 0.08a

6.6

1.69 ± 0.01a 3.03e

2.69 ± 0.51a 2.68e

1.6

Benzo-a-pyrones Isocoumarins 0.956 ± 0.094a

1.44 ± 0.08a

1.5

>100 (63%)b

>100 (42%)b

–

1.27 ± 0.15a

1.84 ± 0.04a

1.4

7.41 ± 0.76a

3.35 ± 0.80a

0.5

>100 (51%)b

>100 (77%)b

–

Coumarin

0.744 ± 0.055a

>100 (35%)b

–

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

2.25 ± 0.27a 3.28e

0.163 ± 0.001a 0.331h

0.545 ± 0.204a 0.530h

24d 22h

DPCPX (A1 antagonist)

0.0004 ± 0.0002a 0.0005h

0.0068 ± 0.0001a 0.0079g

h

f

g

e

c

d

>100 (45%)b 12.6 ± 1.1a 23.4f

All Ki values determined in triplicate and expressed as mean ± SEM. Percentage displacement of the radioligand at the indicated concentration. Rat receptors were used (A1: rat whole brain membranes; A2A: rat striatal membranes). Selectivity index (SI) for the A1 AR isoform calculated as the ratio of Ki (A2A)/Ki (A1). Literature value obtained from reference [35]. Literature value obtained from reference [34]. Literature value obtained from reference [38]. Literature value obtained from reference [36]. a

b

>100 (65%)b 23.1 ± 2.8a

–

CPA (A1 agonist)

Reference compounds >100 (71%)b >100 (54%)b

– Isoflavones

–

SId A2Ac vs [3H]NECA

Ki ± SEM (µM)a/(% displacement)b

A1c vs [3H]DPCPX SId A2Ac vs [3H]NECA A1c vs [3H]DPCPX

Ki ± SEM (µM)a/(% displacement)b Table 1 (continued)

1363d 958h

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

139

4b, to possess affinity for the A1 and A2A ARs with Ki values of 7.41 µM and 3.55 µM, respectively. Three of the flavones (1d–f) was found to be non-selective for both AR subtypes, while compounds 1b, 1c and 1 g displayed selectivity for the A1 AR over the A2A AR. Compound 1c possessed a selectivity index (SI) of 6.6. In contrast, the isochromenone 4b exhibited selectivity for the A2A AR over the A1 AR subtype (SI = 0.5). The above mentioned observations prompted further exploration to the effect of simple structural modifications to the benzo-c-pyrone scaffold, e.g. OH-substitution on ring A and B, saturation of ring C, substitution on position 2 or 3 of ring C and replacing the oxygen with sulphur in ring C. Of note, it was found that our reported Ki values were generally in agreement with literature results reported by other research groups, with the only exception being compound 1b. The latter compound was previously reported with a Ki value of 6.20 µM at A2A ARs and was obtained by measuring the displacement of specific [3H]CGS-21,680 binding in rat striatal membranes [35]. With our A2A AR radioligand binding assay using [3H]NECA, we were unable to obtain a sigmoidal dose–response curve. The A2A AR radioligand assay of 1b was performed in triplicate with at least nine concentrations (0.5–500 µM) spanning three orders of magnitude. The percentage displacement of specific [3H]NECA binding at the latter mentioned concentrations ranged between 55% and 99%. In addition, the A2A AR radioligand binding assay for 1b was repeated three times in separate competition experiments. Furthermore, 1H and 13C NMR results confirmed the purity and structure of 1b. Flavones, isoflavones and flavanones – OH-substitution on ring A and B: First the influence of hydroxy substitution to govern A1 and/or A2A AR affinity was evaluated. In the case of the flavone derivatives (1a–g), mono-substitution of a OH-group on position C5, C6 or C7 of ring A (1b–d) or di-substitution with OH-groups on both rings A and B (1e–g) were explored. The position of the hydroxy group on ring A was found to be important to modulate both the A1 and A2A AR affinity. Upon comparison of the unsubstituted flavone (1a) to its mono-substituted homologues, bearing a OH-group at position C5 (1b), C6 (1c) or C7 (1d) of ring A, it was found that C6-substitution (1c) is the most optimal for increasing A1 AR affinity. Noteworthy, compound 1c exhibited the highest increase (fourfold) in A1 AR affinity (A1Ki = 0.592 µM) compared to the unsubstituted compound 1a (A1Ki = 2.25 µM), followed by C7-substitution (A1Ki = 1.69 µM) and C5-substitution (A1Ki = 1.94 µM). At A2A AR binding, however, C6 (1c A2AKi = 3.93 lM) and C7 (1d A2AKi = 2.69 lM) OH-substituted homologues was only slightly more potent than 1a (A2AKi = 5.14 lM). It appears that C5 OHsubstitution is not tolerated (1b A2AKi = 57%) for A2A AR affinity. OH-substitution on both rings A and B (1e–g) also showed an improvement in A1 and A2A AR affinity when compared to the unsubstituted flavone (1a). Compound 1e, possessing a C5-OH (ring A) and a meta-OH on ring B, presented the highest A2A AR affinity (Ki = 1.44 lM) and compound 1f with a C7-OH (ring A) and meta-OH the second highest A2A AR affinity (Ki = 1.84 lM). Comparing the di OH-substituted flavones to their monosubstituted counterparts (1e vs 1b and 1f vs 1d) it was observed that OH-substitution on ring A together with meta-OH substitution on ring B increased A2A AR affinity. Similarly, OH-substitution (ring A) together with ring B meta-OH substitution slightly improved the A1 AR affinity, e.g. 1e (A1Ki = 0.956 lM) vs 1b (A1Ki = 1.94 lM) and 1f (A1Ki = 1.27 lM) vs 1d (A1Ki = 1.69 lM). Conversely, the C6-OH on ring A together with a para OH-substitution on ring B resulted in 1 g (A1Ki = 0.744 lM and A2AKi = 35%) exhibiting a minor decrease in A1 AR binding and a loss of A2A AR activity compared to its homologue 1c (A1Ki = 0.592 lM and A2AKi = 3.93 lM). This suggests that A2A AR affinity is diminished when the OH-group on ring B is para-substituted, however, further investigation is necessary.

140

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

may provide a planarity and electron coupling to the molecule so that conjugation between ring C and ring A/B could be obtained. This is in agreement with observations of other research groups that explored flavonoids in various other pharmacological fields, including antibacterial, anti-viral, anti-cancer and antiinflammatory activity. They also observed that the C4-carbonyl group of ring C may induce electron shifts via resonance effects and that the C2@C3 double bond might attribute to molecular planarity [39]. Considering the above, the C2@C3 double bond in ring C may govern increased A1 and A2A AR affinity. A similar trend was observed with the isocoumarins (4a vs 4b) representing the benzoa-pyrone based derivatives. The absence of a double bond between C3 and C4 in ring C of compound 4b (3-phenyl-isochromenone) resulted in compound 4a (3-phenyl-isochromanone), which showed to be detrimental to both the A1 and A2A AR affinity. Flavone vs isoflavone – 2-phenyl vs 3-phenyl substitution of ring C: The benzo-c-pyrone scaffold was further explored by moving the phenyl-substituent (ring B) from the C2 position (flavones) of ring C to the C3 position (isoflavones), retaining the aforementioned double bond (C2@C3) in ring C deemed necessary for AR affinity. Unfortunately, C3 phenyl-substitution (isoflavones) resulted in a significant decrease of A1 AR affinity and a diminished A2A AR affinity compared to their corresponding flavone counterparts. The isoflavone derivative 2a (A1Ki = 54%; A2AKi = 71%) was devoid of A1 and A2A AR affinity compared to its flavone homologue 1a (A1Ki = 2.25 µM; A2AKi = 5.14 µM) and the A1 AR affinity of the isoflavone 2b (A1Ki = 23.1 µM; A2AKi = 65%) was approximately 14 times lower than its corresponding flavone analogue 1d (A1Ki = 1.69 µM ; A2AKi = 2.69 µM) and devoid of A2A AR affinity. The A2A AR affinity of isoflavone 2c was also diminished (A1Ki = 12.6 µM; A2AKi = 45%). Thus, the optimal substitution pattern for the benzo-c-pyrone based derivatives to favour A1 and A2A AR affinity seems to be C2 phenyl-substitution of ring C.

In the case of the isoflavones (2a–c), mono-substitution with a OH-group at C7 of ring A led to gained A1 AR binding (2b A1Ki = 23. 1 lM vs 2a), but the A2A AR affinity remained elusive. Furthermore, the isoflavone derivative 2c (A1Ki = 12.6 lM) was documented with an approximate twofold increase in A1 AR affinity after the introduction of a second OH-substitution (para-position) on ring B, however, A2A AR affinity was not gained. This trend is similar to the flavones above where di-substitution with OH-groups favours A1 AR affinity, but para OH-substitution on ring B disfavours A2A AR affinity (2b vs 2c). A pronounced (fivefold) improvement of the A1 AR affinity was documented for the flavanone derivative 3b (A1Ki = 3.55 lM), bearing mono-substitution with a OH-group at position C6 of ring A, when compared to the unsubstituted flavanone 3a (A1Ki = 18.2 l M). Unfortunately, this modification did not afford 3b with A2A AR affinity. Based on the above, the results suggest that OHsubstitution and the substitution pattern may modulate the affinity and selectivity of benzo-c-pyrone based derivatives (flavones, isoflavones and flavanones) at both AR subtypes. The introduction of sulphur (3c) to ring C of compound 3a afforded the resulting thioflavanone with diminished A1 AR affinity and the A2A AR binding remained elusive. For this reason the oxygen heteroatom in ring C is preferred in the case of the flavanone class of compounds. Saturation of ring C: The flavones (1a–g) and flavanones (3a–b) are both 2-phenyl substituted benzo-c-pyrone derivatives, with the exception that the flavanones do not possess a double bond between C2 and C3 of ring C (Fig. 1; Table 1). The necessity of this double bond to govern AR activity was examined by comparing selected flavone and flavanone derivatives. Saturation of ring C (absence of the double bond) was found to be detrimental to A2A AR affinity (1a vs 3a and 1c vs 3b), whereas A1 AR affinity decreased six- to eightfold. The unsaturated C2@C3 double bond

Table 2 The A1 AR affinities (in the absence and presence of GTP) and calculated GTP shifts of selected benzo-c-pyrone (1c, 1e and 1g) and benzo-a-pyrone (4b) analogues. Ki ± SEM (µM)a A1b vs [3H] DPCPX

A1b + GTPc vs [3H] DPCPX

Ki ± SEM (µM)a GTP shift

Benzo-c-pyrones Flavone derivatives

d

0.592 ± 0.011a

0.806 ± 0.121a

1.4

0.956 ± 0.094a

1.11 ± 0.13a

1.2

0.744 ± 0.055a

0.578 ± 0.021a

0.8

CPA (A1 agonist)

DPCPX (A1 antagonist) a

c d e f

A1b + GTPc vs [3H] DPCPX

GTP shift

7.41 ± 0.76a

6.49 ± 0.82a

0.9

0.0068 ± 0.0001a 0.0079e

0.099 ± 0.015a 0.099f

15 14f

0.0004 ± 0.0002a 0.0005f

0.0004 ± 0.0002a 0.0004f

1.0 1.0f

Benzo-a-pyrones Isochromenone derivative

Reference compounds

b

A1b vs [3H] DPCPX

All Ki values determined in triplicate and expressed as mean ± SEM. Rat A1 AR were used expressed by rat whole brain membranes. GTP shift assay, where the 100 µM GTP was added to the A1 AR radioligand binding assay. GTP shifts calculated by dividing the Ki in the presence of GTP by the Ki in the absence of GTP. Literature value obtained from reference [38]. Literature value obtained from reference [36].

d

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

141

Fig. 2. The binding curves of compounds 1c (Panel A), 4d (Panel B) and CPA (Panel C, reference compound) displaying the binding affinity to A1 ARs in the absence () and presence (+) of GTP. (A) GTP shift of 1.4 calculated for the A1 AR antagonist compound 1c, (B) GTP shift of 0.9 calculated for the A1 AR antagonist compound 4d, and (C) GTP shift of 15 calculated for the A1 AR agonist CPA.

Benzo-c-pyrone vs benzo-a-pyrone: The position of the ketone (a vs c) in ring C was also explored by comparing 1a (flavone) to compounds 4b (isochromenone) and 4c (coumarin), while compound 3a (flavanone) was compared to 4a (isochromanone). From these comparisons the rearrangement of the ketone and hetero oxygen of ring C of compound 1a (benzo-c-pyrone) afforded 4b (benzoa-pyrone) with a 3-fold decrease in A1 AR activity. On the other hand, compound 4b exhibited a slight 1.5-fold increase in A2A AR binding affinity compared to compound 1a. A further rearrangement in the benzo-a-pyrone scaffold—from an isohromenone (4b) to a coumarine (4c)—resulted in a diminished AR affinity of both AR subtypes. The rearrangement of the ketone and hetero oxygen of 3a (flavanone, benzo-c-pyrone), afforded compound 4a (isochromanone, benzo-a-pyrone), which showed a diminished affinity in A1 AR affinity with A2A AR affinity remaining elusive. Consequently, the above structural modifications allowed a direct comparison of the benzo-c-pyrone and benzo-a-pyrone scaffolds. Overall, it is suggested that the ideal orientation of the ketone to the hetero oxygen in ring C to favour both A1 and A2A AR affinity is the benzo-c-pyrone moiety.

100 lM GTP (Table 2; Fig. 2) [36]. The nonspecific binding was defined by the addition of 10 lM DPCPX (unlabelled) [36]. In the case of an A1 AR agonist the binding curve is expected to significantly shift to the right with the addition of GTP. This may be attributed to an uncoupling of the A1 AR from its Gi protein in the presence of GTP [40]. In literature CPA is recognized as a full agonist and in this study was found to possess a calculated GTP shift of 15 (Table 2). In contrast, an A1 AR antagonist’s binding curve is not influenced by the addition of GTP, as no significant shift to the right is observed [40], and a compound with a calculated GTP shift of approximately 1 is considered a full antagonist [41]. The calculated GTP shift results for CPA and DPCPX (Table 2) was found to correspond with literature values, where CPA act as an agonist (Fig. 2) and DPCPX as an antagonist. Considering the above, compounds 1c, 1e, 1g and 4b are considered antagonists of the A1 AR with calculated GTP shifts of approximately 1 and no significant shift of the competition curves to the right (Table 2; Fig. 2). Based on the above it is expected that the investigated benzo-a-pyrone (4b) and benzo-c-pyrone (1c, 1e and 1g) derivatives will also act as antagonists in vivo. 3. Conclusion

2.3. GTP shift assay In order to determine if the benzo-c-pyrones (1c, 1e and 1g) and benzo-a-pyrones (4b) act as agonists or antagonists at the A1 AR, GTP shift experiments were carried out. The latter compounds were chosen as they exhibited the highest A1 AR binding affinity among the benzo-c-pyrone (1c, 1e and 1g) and benzo-a-pyrone (4b) derivatives. In short, The GTP shift assays were performed as described previously with rat whole brain membranes and [3H] DPCPX (0.1 nM; Kd = 0.36 nM) in the absence and presence of

In conclusion, this investigative study indicated the importance of structural modifications at rings A, B and C of the benzo-cpyrone scaffold for gaining and even losing A1 and/or A2A AR affinity. The structure-activity relationships, identified as vital for benzo-c-pyrone derivatives to govern A1 and A2A AR affinity, were previously unknown. Overall, the flavones (1a–g) were found to possess affinity for both A1 and A2A AR subtypes. Compound 1c exhibited the best A1 AR affinity with a SI of 6.6, while compound 1 g was reported with the second highest A1 AR binding and did

Fig. 3. Structural requirements to enhance the benzo-c-pyrone scaffold’s A1 and A2A AR affinity.

142

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143

not favour the A2A AR subtype. In general, the benzo-c-pyrone scaffold displayed superior A1 AR affinity over the benzo-a-pyrone scaffold. However, by comparing compound 1a (benzo-c-pyrone) and 4b (benzo-a-pyrone) a similar affinity towards the A2A AR subtype was observed. Substitution with a hydroxy group on ring A and/or B play a key role in modulating the binding affinity at A1 and A2A AR. Specifically, C6 OH-substitution on ring A was preferred over the C5- and C7-positions (1c vs 1b & 1d) for A1 AR affinity. The A2A AR affinity was favoured in the case of di-substitution where the OH-group was substituted at the meta-position on ring B (1e and 1f), while A1 AR affinity was tolerated and led to a nonselective preference. The double bond between C2 and C3 of ring C as well as C2 phenyl-substitution of ring C was shown to be imperative for both A1 and A2A AR affinity, since saturation of ring C (1a vs 3a and 1c vs 3b) and C3 phenyl substitution (1a vs 2a and 1d vs 2b) were found to decrease A1 AR affinity and diminish A2A AR affinity (Fig. 3). Compounds 1c and 1e are two promising drug candidates that both act as A1 AR antagonists in vitro with Ki values below 1 µM. Compound 1c was found to possess the highest A1 AR affinity (A1Ki = 0.592 µM; SI = 6.6) and also possess A2A AR affinity (A2AKi = 3.93 µM). Furthermore, compound 1e possessed the highest A2A AR affinity (A2AKi = 1.44 µM) and the third best A1 AR affinity (A1Ki = 0.956 µM) and acted non-selectively. Therefore, compounds 1c and 1e are ideal drug candidates for future in vivo investigation of the benzo-c-pyrone scaffold as selective A1 (1c) and nonselective A1/A2A (1e) AR antagonists in the treatment of neurological disorders, depression, certain cancers, and potentially improve tumour immunotherapy. 4. Experimental section 4.1. Materials and methods All commercially available reagents were obtained from various manufacturers: radioligands [3H]NECA (specific activity 27.1 Ci/ mmol) procured from PerkinElmer and [3H]DPCPX (specific activity 120 Ci/mmol) from Amersham Biosciences, filter-count from PerkinElmer and Whatman GF/B 25 mm diameter filters from Merck. Radio activity was calculated by a Packard Tri-CARB 2810 TR liquid scintillation counter. 4.2. Radioligand binding assay The collection of tissue samples for the A1 and A2A AR binding studies was approved by the Research Ethics Committee of North-West University (application number NWU-0035-10-A5). The A1 and A2A AR affinities of the test compounds were determined with radioligand binding assays as described in the literature [36,37] In short, the A1 AR radioligand binding assay was performed with rat whole brain membranes in the presence of the radioligand [3H]-8-cylcopentyl-1,3-dipropylxanthine ([3H] DPCPX) [36]. Each incubation, of the A1 AR radioligand binding assay, contained 120 µg of membrane protein suspension (rat whole brain membranes), 0.1 nM [3H]DPCPX, 0.1 units/mL adenosine deaminase, the test compound and 1% DMSO. On the other hand, the A2A AR affinity was measured at rat striatal membranes with 50 -N-[3H]-ethylcarboxamideadenosine ([3H]NECA) as radioligand [36]. The A1 AR agonist, N6-cyclopentyladenosine (CPA), was added to the A2A AR competition experiments to minimize the binding of the radioligand [3H]NECA to A1 AR [36,42]. Each incubation, of the A2A AR radioligand assay, contained 120 µg of membrane protein suspension (rat striatal membranes), 4 nM [3H]NECA, 50 nM CPA, 10 mM MgCl2, 0.2 units/mL adenosine deaminase, the test compound and 1% DMSO. Non-specific binding

of [3H]DPCPX and [3H]NECA was defined as binding in the presence of 100 µM CPA [36,42]. In turn, specific binding was defined as the total binding minus the non-specific binding [36,42]. The protein content of all membrane preparations was determined with Bradford reagent according to the method described by Bradford [43]. The A1 and A2A AR radioligand binding assays were validated with CPA (agonist) and DPCPX (antagonist) as reference compounds and the results were found to be in accordance with literature values (Table 1). Furthermore, some previously reported benzo-c-pyrone derivatives (1a, 1b, 1d and 2c) were tested in our radioligand binding experiments for comparison with literature data. Generally our reported Ki values were in agreement with literature data reported by other research groups, with the only exception being compound 1b. Please see Section 2.2 for additional information. 4.3. GTP shift assay Compounds 1c, 1e, 1 g and 4b were explored via GTP (guanosine triphosphate) shift assays to determine the agonistic or antagonistic functionality of the investigated benzopyrone derivatives towards the A1 AR. The GTP shift assays were performed as described previously with rat whole brain membranes and [3H] DPCPX (0.1 nM; Kd = 0.36 nM) in the absence and presence of 100 lM GTP (Table 2; Fig. 2) [36]. The non-specific binding was defined by the addition of 10 lM DPCPX (unlabelled) [36]. 4.4. Data analysis The data analyses of the test and reference compounds were determined by plotting the specific binding vs. the logarithm of the test compounds’ concentrations to obtain a sigmoidal doseresponse curve via the Prism software package (GraphPad Software Inc.) [36]. The Ki values for the competitive inhibition of [3H]DPCPX (0.1 nM; Kd = 0.36 nM) [38] or [3H]NECA (4 nM; Kd = 15.3 nM) [41] by the investigated compounds were calculated from the IC50 values (that was calculated according to the Cheng-Prusoff equation) [44]. The calculated Ki values are expressed as mean ± standard error of the mean (SEM) after triplicate determinations [36]. The GTP shifts were calculated by dividing the Ki value of a compound reported in the presence of GTP by the corresponding value obtained in the absence of GTP [36,41]. A compound with a calculated GTP shift of approximately 1 is considered an antagonist; the presence of GTP affects the competition curves of an agonist and shifts the curve to the right [36,40]. Acknowledgements We are grateful to Dr. J. Jordaan of the SASOL Centre for Chemistry, North-West University, for recording the NMR spectra of 1b. Financial support for this work was provided by the North-West University, the National Research Foundation and the Medical Research Council, South Africa. References [1] T.M. Palmer, G.L. Stiles, Adenosine receptors, Neuropharmacolgy 34 (1995) 683–694. [2] S. Cies´lak, M. Komoszyn´ski, A. Wojtczak, Adenosine A2A receptors in Parkinson’s disease treatment, Purinergic Signal. 4 (2008) 305–312. [3] A. Espinosa, M. Alegret, S. Valero, G. Vinyes-Junque, I. Hernandez, A. Mauleon, M. Rosende-Roca, A. Ruiz, O. Lopez, L. Tarraga, M. Boada, A longitudinal followup of 550 mild cognitive impairment patients: evidence for large conversion to dementia rates and detection of major risk factors involved, J. Alzheimer’s Disease 34 (2013) 769–780. [4] M.E. Yacoubi, C. Ledent, M. Parmentier, R. Bertorelli, E. Onginni, J. Costentin, J. Vaugeois, Adenosine A2A receptor antagonists are potential antidepressants: evidence based on pharmacology and A2A receptor knockout mice, Br. J. Pharmacol. 134 (2001) 68–77.

M.M. van der Walt, G. Terre’Blanche / Bioorganic Chemistry 77 (2018) 136–143 [5] S. Gessi, S. Merighi, V. Sacchetto, C. Simioni, P.A. Borea, Adenosine receptors and cancer, Biochim. Biophys. Acta 2011 (1808) 1400–1412. [6] P.J. Richardson, H. Kase, P.G. Jenner, Adenosine A2A receptor antagonists as new agents for the treatment of Parkinson’s disease, Trends Pharmacol Sci. 18 (1997) 338–344. [7] T. Mihara, K. Mihara, J. Yarimizu, Y. Mitani, R. Matsuda, H. Yamamoto, S. Aoki, A. Akahane, A. Iwashita, N. Matsuoka, Pharmacological characterization of a novel, potent adenosine A1 and A2A receptor dual antagonist, 5-[5-Amino-3-(4fluorophenyl)pyrazin-2-yl]-1-isopropylpyridine-2(1H)-one (ASP5854), in models of Parkinson’s disease and cognition, J Pharmacol. Exp. Ther. 323 (2007) 708–719. [8] E.R. Dorsey, R. Constantinescu, J.P. Thompson, K.M. Biglan, R.G. Holloway, K. Kieburtz, F.J. Marshall, B.M. Ravina, G. Schifitto, A. Siderowf, C.M. Tanner, Projected number of people with Parkinson’s disease in the most populous nations, 2005 through 2030, Neurology 68 (2007) 384–386. [9] A. Brunschweiger, P. Koch, M. Schlenk, F. Pineda, P. Küppers, S. Hinz, M. Köse, S. Ullrich, J. Hockemeyer, M. Wiese, J. Heer, C.E. Müller, 8Benzyltetrahydropyrazino[2,1-f]purinediones: Water-soluble tricyclic xanthine derivatives as multitarget drugs for neurodegenerative diseases, Chemmedchem. 9 (2014) 1704–1724. [10] W. Bara-Jimenez, A. Sherzai, T. Dimitrova, A. Favit, F. Bibbiani, M. Gillespie, M.J. Morris, M.M. Mouradian, T.N. Chase, Adenosine A2A receptor antagonist treatment of Parkinson’s disease, Neurology 61 (2003) 293–296. [11] K. Ikeda, M. Kurokawa, S. Aoyama, Y. Kuwana, Neuroprotection by adenosine A2A receptor blockade in experimental models of Parkinson’s disease, J. Neurochem. 80 (2002) 262–270. [12] D. Xiao, E. Bastia, Y.-H. Xu, C.L. Benn, J.-H.J. Cha, T.S. Peterson, J.-F. Chen, M.A. Schwarzschild, Forebrain adenosine A2A receptors contribute to l-3,4dihydroxyphenylalanine-induced dyskinesia in hemiparkinsonian mice, J. Neurosci. 26 (2006) 13548–13555. [13] A. Pinna, Adenosine A2A receptor antagonists in Parkinson’s disease: Progress in clinical trials from the newly approved istradefylline to drugs in early development and those already discontinued, CNS Drugs 28 (2014) 455–474. [14] R. Dhanasekaran, A. Limaye, R. Cabrera, Hepatocellular carcinoma: current trends in worldwide epidemiology, risk factors, diagnosis, and therapeutics, Hepatic Med. 4 (2012) 19–37. [15] K.A. McGlynn, W.T. London, Epidemiology and natural history of hepatocellular carcinoma, Best Pract. Res. Clin. Gastroenterol. 19 (2005) 3–23. [16] J.W. Fisher, J. Brookins, Adenosine A2A and A2B receptor activation of erythropoietin production, Am. J. Physiol. Renal Physiol. 281 (2001) F826– F832. [17] K. Nagashima, A. Karasawa, Modulation of erythropoietin production by selective adenosine agonists and antagonists in normal and anemic rats, Life Sci. 59 (2006) 761–771. [18] P. Fishman, S. Bar-Yehuda, M. Synowitz, J.D. Powell, K.N. Klotz, S. Gessi, P.A. Borea, Adenosine receptors and cancer, Handbook Exp. Pharmacol. 193 (2009) 399–441. [19] A. Gaspar, J. Reis, M.J. Matos, E. Uriarte, F. Borges, In search for new chemical entities as adenosine receptor ligands: development of agents based on benzoc-pyrone skeleton, Eur. J. Med. Chem. 54 (2012) 914–918. [20] K.E. Heim, A.R. Tagliaferro, D.J. Bobilya, Flavonoid antioxidants: chemistry, metabolism and structure-activity relationships, J. Nutr. Biochem. 13 (2002) 572–584. [21] S. Kumar, A.K. Pandey, Chemistry and biological activities of flavonoids: an overview, Sci. World J. 2013 (2013) 162750. [22] D. Matosiuk, D. Gralak, M. Ryznar, Lipophilicity of natural coumarines, Annales UMCS Sectio DDD. 24 (2011) 19–25. [23] K.B. Magalingam, A.K. Radhakrishnan, N. Haleagrahara, Protective mechanisms of flavonoids in parkinson’s disease, Oxid. Med. Cell Longev. 2015 (2015) 314560.

143

[24] R. Koes, W. Verweij, F. Quattrocchio, Flavonoids: a colorful model for the regulation and evolution of biochemical pathways, Trends Plant Sci. 10 (2005) 236–242. [25] T.P. Cushnie, A.J. Lamb, Antimicrobial activity of flavonoids, Int. J. Antimicrob. Agents. 26 (2005) 343–356. [26] A.R. Tapas, D.M. Sakarkar, R.B. Kakde, Flavonoids as nutraceuticals: a review, Trop. J. Pharm. Res. 7 (2008) 1089–1099. [27] E. Middleton Jr, C. Kandaswami, Effects of flavonoids on immune and inflammatory cell functions, Biochem. Pharmacol. 43 (1992) 1167–1179. [28] J.E. Brown, H. Khodr, R.C. Hider, C.A. Rice-Evans, Structural dependence of flavonoid interactions with Cu2+ ions: implications for their antioxidant properties, Biochem. J. 330 (1998) 1173–1178. [29] M. Friedman, Overview of antibacterial, antitoxin, antiviral, and antifungal activities of tea flavonoids and teas, Mol. Nutr. Food Res. 51 (2007) 116–134. [30] M.K. Chahar, N. Sharma, M.P. Dobhal, Y.C. Joshi, Flavonoids: a versatile source of anticancer drugs, Pharmacogn. Rev. 5 (2011) 1–12. [31] S. Moro, A.M. van Rhee, L.H. Sanders, K.A. Jacobson, Flavonoid derivatives as adenosine receptor antagonists: a comparison of the hypothetical receptor binding site based on a comparative molecular field analysis model, J. Med. Chem. 41 (1998) 46–52. [32] S.P. Alexander, Flavonoids as antagonists at A1 adenosine receptors, Phytother. Res. 20 (2006) 1009–1012. [33] Y. Karton, J.L. Jiang, X.D. Ji, N. Melman, M.E. Olah, G.L. Stiles, K.A. Jacobson, Synthesis and biological activities of flavonoid derivatives as A3 adenosine receptor antagonists, J. Med. Chem. 39 (1996) 2293–2301. [34] K.A. Jacobson, S. Moro, J.A. Manthey, P.L. West, X.D. Ji, Interactions of flavones and other phytochemicals with adenosine receptors, Adv. Exp. Med. Biol. 505 (2002) 163–171. [35] X. Ji, N. Melman, K.A. Jacobson, Interactions of flavonoids and other phytochemicals with adenosine receptors, J. Med. Chem. 39 (1996) 781–788. [36] M.M. Van der Walt, G. Terre’Blanche, 1,3,7-Triethyl-substituted xanthines – possess nanomolar affinity for the adenosine A1 receptor, Bioorg. Med. Chem. 23 (2015) 6641–6649. [37] M.M. Van der Walt, G. Terre’Blanche, Selected C8 two chain linkers enhance the adenosine A1/A2A receptor affinity and selectivity of Caffeine, Eur. J. Med. Chem. 125 (2017) 652–656. [38] R.F. Bruns, J.H. Fergus, E. Badger, J.A. Bristol, L.A. Santay, J.D. Hartman, S.J. Hays, C.C. Huang, Binding of the A1-selective adenosine antagonist 8-cyclopentyl1,3-dipropylxanthine to rat brain membranes Naunyn-Schmiedeberg’s Arch, Pharmacology 335 (1987) 59–63. [39] T.Y. Wang, Q. Li, K.S. Bi, Bioactive flavonoids in medicinal plants: structure, activity and biological fate, Asian J. Pharma. Sci. (2017), https://doi.org/ 10.1016/j.ajps.2017.08.004 (in press). [40] M. Gutschow, M. Schlenk, J. Gab, M. Paskaleva, M. Wassam Alnouri, S. Scolari, J. Iqbal, C.E. Muller, Benzothiazinones: a novel class of adenosine receptor antagonists structurally unrelated to xanthine and adenine derivatives, J. Med. Chem. 55 (2012) 3331–3341. [41] E.M. Van der Wenden, H.R. Hartog-Witte, H.C.P.F. Roelen, J.K. Von Frijtag Drabbe Kunzel, I.M. Pirovano, R.A.A. Mathot, M. Danhof, A. Van Aerschot, J.M. Lidaks, A.P. Ijzerman, W. Soudijn, 8-Substituted adenosine and theophylline-7riboside analogues as potential partial agonists for the adenosine A1 recepto, Eur. J. Pharmacol-Mol. Pharmacol. 290 (1995) 189–199. [42] R.F. Bruns, G.H. Lu, T.A. Pugsley, Characterization of the A2 adenosine receptor labeled by [3H]NECA in rat striatal membranes, Mol. Pharmacol. 29 (1986) 331–346. [43] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72 (1976) 248–254. [44] Y. Cheng, W.H. Prusoff, Relationship between the inhibition constant (K1) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction, Biochem. Pharmacol. 22 (1973) 3099.