Methoxy substituted 2-benzylidene-1-indanone derivatives as A₁ and/or A_2A AR antagonists for the potential treatment of neurological conditions

Methoxy substituted 2-benzylidene-1-indanones possess improved adenosine A₁ and A_2A receptor affinity in the nanomolar range.

Abstract

A prior study reported on hydroxy substituted 2-benzylidene-1-indanone derivatives as A₁ and/or A_2A antagonists for the potential treatment of neurological conditions. A lead compound (1a) was identified with both A₁ and A_2A affinity in the micromolar range. The current study explored the structurally related methoxy substituted 2-benzylidene-1-indanone derivatives with various substitutions on ring A and B of the benzylidene indanone scaffold in order to enhance A₁ and A_2A affinity. This led to compounds with both A₁ and A_2A affinity in the nanomolar range, namely 2c (A₁K_i (rat) = 41 nM; A_2AK_i (rat) = 97 nM) with C4-OCH₃ substitution on ring A together with meta (3′) hydroxy substitution on ring B and 2e (A₁K_i (rat) = 42 nM; A_2AK_i (rat) = 78 nM) with C4-OCH₃ substitution on ring A together with meta (3′) and para (4′) dihydroxy substitution on ring B. Additionally, 2c is an A₁ antagonist. Consequently, the methoxy substituted 2-benzylidene-1-indanone scaffold is highly promising for the design of novel A₁ and A_2A antagonists.

Introduction

The adenosine system – specifically adenosine A₁ and A_2A receptors – is a promising drug target for the non-dopaminergic treatment of the neurodegenerative disorder Parkinson's disease (PD).1 These adenosine receptors (AR's), together with A_2B and A₃ AR's, are either inhibitory (A₁ and A₃) or stimulatory (A_2A and A_2B) G-protein coupled receptors.2

The spotlight first fell on the A₁ and A_2A AR's when an epidemiological study found that consumption of the non-selective AR antagonist caffeine – present in coffee and tea – is associated with a reduced risk of developing PD.3 Since then, numerous preclinical studies in rodents and non-human primate models of PD have supported the potential of A₁ and A_2A AR antagonists for the treatment of PD.4

The structure of the xanthine derivative caffeine is similar to that of adenosine.5 Caffeine and other A₁ and A_2A AR antagonists exert effects contrary to endogenous adenosine and, in so doing, affects various neurotransmitters, receptors and signalling pathways.4,6

There is growing evidence that selective A_2A AR antagonists, like KW-6002 (istradefylline), may be novel non-dopaminergic treatment for PD; KW-6002 is approved in Japan since 2013 for the treatment of wearing-off phenomenon associated with l-dopa treatment in PD.7,8 Selective A_2A AR antagonists may possibly alleviate parkinsonian motor symptoms due to the close anatomical and functional relationship between A_2A AR's and dopamine D₂ receptors on the indirect striatopallidal GABAergic pathway.7,8 In animal studies, agonists and antagonists of A_2A AR's produce behavioural effects similar to antagonists and agonists of dopamine D₂ receptors, respectively.9 Thus, blockade of A_2A AR's on the indirect striatopallidal GABAergic pathway reduces postsynaptic effects of dopamine depletion (a pathological hallmark of PD) and, subsequently, reduces motor symptoms associated with PD.10 Additionally, KW-6002 may address a non-motor symptom of PD, namely depression, evidenced by a decrease in immobility time during the forced swim test and tail suspension test in rodents (animal models of depression) when the said drug was administered.11,12 Neurodegeneration may be stopped or, at least, slowed by A_2A AR antagonists as KW-6002 attenuated striatal dopamine depletion in the MPTP animal model of PD.13

Synaptic plasticity; the ability of synapses to strengthen or weaken in response to increases or decreases in their activity, is the basis for learning and memory.14 Brain areas associated with cognition are the prefrontal cortex and hippocampus.14 Endogenous adenosine via A₁ AR's—which are abundantly expressed in the prefrontal cortex and hippocampus—modulates synaptic plasticity phenomena long-term depression and long-term potentiation and, in so doing, inhibits learning and memory.15 A₁ AR antagonists block A₁ AR mediated inhibitory modulation of synaptic plasticity in the prefrontal cortex and hippocampus: neurotransmitter release and synaptic transmission is increased, facilitating synaptic plasticity and, thus, learning and memory.14 Pharmacological studies support the above, by means of the selective A₁ AR antagonists BIIP20 and FR194921 which are active in animal models of cognitive deficits.16,17

It proved to date challenging to translate findings on AR function into clinical studies and no other AR antagonists have been approved.18

The investigation of dual A₁/A_2A AR antagonists for the potential treatment of PD has been investigated in the past.19–23 For example, 5-[5-amino-3-(4-fluorophenyl)pyrazin-2-yl]-1-isopropylpyridine-2(1H)-one (ASP5854) (see Fig. 1) showed promise in animal models of PD as well as cognition and has K_i values of 9.03 nM at the human A₁ AR and 1.76 at the human A_2A AR.19 Additionally, 2-amino-8-[2-(4-morpholinyl)ethoxy]-4-phenyl-5H-indeno-[1,2-d]pyrimidin-5-one (JNJ-40255293) (see Fig. 1) is also a dual A₁/A_2A AR antagonist with efficacy in animal models of PD (A₁K_i (human) = 48 nM; A₂AK_i (human) = 6.5 nM).20,21 Of note, this compound is structurally related to the benzylidene indanones currently under investigation. 8-Substituted 1,3-dimethyltetrahydropyrazino[2,1-f]purinedione derivatives were evaluated as dual A₁/A_2A AR antagonist in a multitarget approach for the treatment of neurodegenerative disorders.22 The 7-aminopyrazolo[4,3-d]pyrimidine derivatives were also explored as dual A₁/A_2A AR antagonists with several of these compounds possessing nanomolar affinity for the human A_2A AR and slightly lower human A₁ AR.23

Fig. 1. The structures of ASP-5854 and JNJ-40255293.

Therefore, dual A₁ and A_2A AR antagonists may be non-dopaminergic drugs for the symptomatic treatment of both PD motor symptoms (for example bradykinesia, rigidity, resting tremor and postural instability) and PD non-motor symptoms (for example cognitive deficits such as cognitive dysfunction and depression) as well as exhibit neuroprotective properties.4

Benzopyrones are a class of compounds with significantly diverse biological activities24 (including antiparkinsonian and neuroprotective properties)25–30 and are, thus, considered a privileged scaffold in medicinal chemistry.24 Benzopyrones constitute the basic framework of flavonoids,31,32 and the structurally related isocoumarins33 and coumarins.24 It is reported that the flavonoid derivative 5,3′-dihydroxyflavone possess an A₁K_i (rat) value of 0.956 μM and an A_2AK_i (rat) value of 1.44 μM (see Fig. 2).34 In fact, the aurone derivatives (specifically hispidol (A₁K_i (rat) = 0.352 μM) and maritimetin (A₁K_i (rat) = 3.47 μM and A_2AK_i (rat) = 9.35 μM) (see Fig. 2)), which are members of the flavonoid family,28 served as inspiration for previous work on the benzylidene tetralones35,36 as well as benzylidene indanones37 – leading to compounds with A₁ and A_2A AR affinity in the micromolar range. For example, the benzylidene tetralone derivative (E)-5-hydroxy-2-(3-hydroxybenzylidene)-3,4-dihydronaphthalen-1(2H)-one has an A₁K_i (rat) value of 1.62 μM and an A_2AK_i (rat) value of 5.46 μM (see Fig. 2).36

Fig. 2. The structures and K_i values of compounds structurally related to benzylidene indanones and the structural modifications to lead compound 1a to determine features essential for dual A₁/A_2A AR affinity.

In recent times, hydroxy substituted 2-benzylidene-1-indanone analogues were explored as A₁ and/or A_2A AR antagonists for the potential treatment of neurological conditions.37 Of note, is compound 1a with K_i values for both the A₁ and A_2A AR below 1 μM (A₁K_i (rat) = 0.435 μM; A_2AK_i (rat) = 0.903 μM). It was found that C4 hydroxy substitution on ring A of the benzylidene indanones in combination with meta (C3′) and para (C4′) dihydroxy substitution on ring B is preferable for both A₁ and A_2A AR binding.

These compounds comprise of a benzylidene indanone scaffold (i.e. fused 6- and 5-membered rings, namely ring A and ring C), where ring C bears a C2-phenyl substituted sidechain (namely ring B). Theoretically, these compounds may be either E- or Z-isomers.38 The E-configuration is favourable for thermodynamic reasons because of steric interaction between the aryl and carbonyl groups in case of the Z-isomers.38,39

Based on the structure of 1a, we investigated the structurally related methoxy substituted 2-benzylidene-1-indanone derivatives as potent A₁ and A_2A AR antagonists.

The 2-benzylidene-1-indanone scaffold was modified to include substituent changes to ring A and ring B (see Fig. 2). Firstly, C4-, C5- and/or C6-methoxy substitution on ring A was made, in order to determine whether hydroxy or methoxy substitution is preferable for A₁ and A_2A AR affinity and, secondly, the optimal position of the methoxy group for both A₁ and A_2A AR affinity was determined. Also, an unsubstituted ring A and methylenedioxy substitution on ring A was incorporated. Substitution at the ortho (C2′), meta (C3′), para (C4′) and/or C5 position(s) of phenyl ring B with hydroxy-, methoxy- or dimethylamino group(s) was investigated.

Accordingly, the 2-benzylidene-1-indanones was evaluated to ascertain which structure activity relationships govern A₁ and A_2A AR affinity.

Results and discussion

Chemistry

Reagents and test compounds were synthesised as depicted in Scheme 1. The key starting material for 2b–g, namely 2a was synthesised by methylation of 4-hydroxy-1-indanone, as described,40 whereas, reagents for 2j, 2l and 2m were commercially available. In turn, test compounds (2b–g, 2i, 2k & 2m) were prepared via an acid catalysed aldol condensation reaction.35–37 Test compounds 2h, 2j and 2l were available from a previous study.41 The novel compounds 2b–g, 2i, 2k and 2m (fair yields) were purified by recrystallization from a suitable solvent (either MeOH or EtOH) and, in each instance, the structure, molecular mass and purity of these compounds were verified by ¹H NMR, ¹³C NMR, MS and/or HPLC analysis.

Scheme 1. Synthesis of 2a–g, 2i, 2k & 2m. Reagents and conditions: a) acetone, K₂CO₃, MeI, 50 °C (18 h); b) MeOH, HCl (32%), 120 °C (24 h).

Biology

Radioligand binding assays

The affinities of the 2-benzylidene-1-indanone analogues (2a–m) at rat A₁ and A_2A AR subtypes were determined with radioligand competition experiments as described previously.42,43 The A₁ and A_2A AR radioligand binding assay results are summarized in Table 1. Two reference compounds, namely CPA and DPCPX, were included in the study and the results are in accordance with literature values (see Table 1).

Table 1. The dissociation constant (K_i) values for the binding of the 2-benyzlidene-1-indanone analogues at rat A₁ and A_2A AR's.

#	Ring A			Ring B				K i ± SEM (μM) a (% displacement) b		SI d
	R 1	R 2	R 3	R 1′	R 2′	R 3′	R 4′	A1 c vs. [3H]DPCPX	A2A c vs. [3H]NECA	(A1/A2A)
Lead compound
1a	OH	H	H	H	OH	OH	H	0.435 ± 0.050 a , e	0.903 ± 0.081 a , e	0.5 d , e
Methoxy substituted 2-benzylidene-1-indanones
2a	OCH3	H	H	—	—	—	—	>100 (78%) b	>100 (72%) b	—
2b	OCH3	H	H	H	H	H	H	1.65 ± 0.15 a	>100 (62%) b	—
2c	OCH3	H	H	H	OH	H	H	0.041 ± 0.002 a	0.097 ± 0.009 a	0.4 d
2d	OCH3	H	H	H	H	OH	H	>100 (31%) b	>100 (65%) b	—
2e	OCH3	H	H	H	OH	OH	H	0.042 ± 0.009 a	0.078 ± 0.002 a	0.5 d
2f	OCH3	H	H	OCH3	H	OCH3	OCH3	>100 (46%) b	>100 (29%) b	—
2g	OCH3	H	H	H	H	N(CH3)2	H	>100 (52%) b	>100 (23%) b	—
2h	H	OCH3	H	H	H	H	H	>100 (42%) b	>100 (82%) b	—
2i	H	OCH3	H	H	OH	OH	H	3.28 ± 0.31 a	6.32 ± 0.32 a	0.5 d
2j	H	OCH3	H	H	H	N(CH3)2	H	>100 (64%) b	>100 (47%) b	—
2k	H	OCH3	OCH3	H	OH	OH	H	4.29 ± 0.18 a	18.02 ± 1.39 a	0.2 d
2l	H	H	H	H	H	N(CH3)2	H	>100 (51%) b	>100 (65%) b	—
2m	—	—	—	—	—	—	—	>100 (70%) b	1.07 ± 0.10 a	—
Reference compounds
CPA (A1 agonist)								0.0068 ± 0.0001 a	0.163 ± 0.001 a	24 d
								(0.0079); f	(0.331) g	(22) g
								(0.015) g
DPCPX (A1 antagonist)								0.0004 ± 0.0002 a	0.545 ± 0.204 a	1363 d
								(0.0005); g	(0.530); g	(958) g
								(0.0003) h	(0.340) h	(1130) h

^aAll K_i values determined in triplicate and expressed as mean ± SEM.

^bPercentage displacement of the radioligand at a maximum tested concentration (100 μM).

^cRat receptors were used (A₁: rat whole brain membranes; A_2A: rat striatal membranes).

^dSelectivity index (SI) for the A_2A receptor isoform calculated as the ratio of A₁K_i/A_2AK_i.

^eLiterature value obtained from reference.37

^fLiterature value obtained from reference.49

^gLiterature value obtained from reference.42

^hLiterature value obtained from reference.50

A prior study identified the 4-hydroxy substituted 2-benzylidene-1-indanone analogue 1a as a lead compound to design novel and potent A₁ and A_2A AR antagonists.37 This study's parent scaffold is the 4-methoxy substituted compound 2a, which is unsubstituted at position 2 and lacks both A₁ and A_2A AR affinity.

Structural modification to ring A

Firstly, in analogy to previous work on the 2-benzylidene-1-tetralones (which determined whether OH- or OCH₃-group substitution on ring A was preferable to attain both A₁ and A_2A AR affinity),36 the current study investigated C4-OH substitution versus C4-OCH₃ substitution on ring A, as well as an unsubstituted ring A and 1,3-dioxolane substitution on ring A by comparing the K_i values of analogous compounds. Secondly, similar to a previous study of the 2-benzylidene-1-indanones – which determined optimal OH-substitution on ring A together with meta (3′) and para (4′) diOH-substitution on ring B37 – the impact of OCH₃-substitution at either position C4, C5 and/or C6 of ring A, in combination with various ring B substitutions were evaluated by comparing the K_i values of analogous compounds.

C4-OH substitution vs. C4-OCH₃ substitution

Comparison of lead compound 1a (A₁K_i (rat) = 0.435 μM; A_2AK_i (rat) = 0.903 μM)37 to its methoxy substituted counterpart 2e (A₁K_i (rat) = 0.042 μM; A_2AK_i (rat) = 0.078 μM) showed that in the 2-benzylidene-1-indanone scaffold C4-OCH₃ substitution is favoured over C4-OH substitution on ring A; where a tenfold increase in A₁ AR affinity and a twelvefold increase in A_2A AR affinity were observed.

Interestingly, the structurally related 2-benzylidene-1-tetralone analogues generally prefer OH-substitution over OCH₃-substitution on ring A.35,36

An unsubstituted ring A in combination with a dimethylamino group in the para (4′) position (2g) diminished both A₁ and A_2A AR affinity as seen with comparison of 2g to unsubstituted 2l.

Methylenedioxy substitution on ring A (2m) reduced both A₁ AR affinity (A₁K_i (rat) = >100 μM) and A_2A AR affinity (A_2AK_i (rat) = >1.07 μM) when compared to 2e, however the said compound retained relatively good A_2A AR affinity.

Optimal position of OCH₃-group

Similar to the hydroxy substituted 2-benzylidene-1-indanone analogues, the position of the OCH₃-group on ring A modulates A₁ and A_2A AR binding affinity and C4-OCH₃ substitution on ring A is preferred over either C5-OCH₃ substitution or C5- and C6-diOCH₃ substitution, as evidenced when C4-OCH₃ substitution on ring A (2e; A₁K_i (rat) = 0.042 μM and A_2AK_i (rat) = 0.078 μM) is compared to C5-OCH₃ substitution (2i; A₁K_i (rat) = 3.28 μM and A_2AK_i (rat) = 6.32 μM) and C5 and C6-diOCH₃ substitution (2k A₁K_i (rat) = 4.29 μM and A_2AK_i (rat) = 18.02 μM).

Comparison of 4-methoxy substituted 2b (A₁K_i (rat) = 1.65 μM; A_2AK_i (rat) = >100 μM) to 5-methoxy substituted 2h (A₁K_i & A_2AK_i (rat) = >100 μM), showed that C4-OCH₃ substitution on ring A, instead of C5-OCH₃ substitution, lead to increased A₁ AR affinity – yet, A_2A AR affinity was unaffected.

Additionally, both the A₁ and A_2A AR's favour C4-OCH₃ substitution on ring A to either C5-OCH₃ substitution or C5 and C6-diOCH₃ substitution when 4-methoxy substituted 2e (A₁K_i (rat) = 0.042 μM; A_2AK_i (rat) = 0.078 μM) was compared to 5-methoxy substituted 2i (A₁K_i (rat) = 3.28 μM; A_2AK_i (rat) = 6.32 μM) and 5,6-dimethoxy substituted 2k (A₁K_i (rat) = 4.29 μM; A_2AK_i (rat) = 18.02 μM); a dramatic increase in both A₁ and A_2A AR affinity was observed – seeing as 2e possess A₁ and A_2A AR K_i values in the nanomolar range. In fact, the A₁ AR affinity of 2e is 78 times better than that of 2i and 102 times better than that of 2k, whereas as the A_2A AR affinity of 2e is 81 times better than that of 2i and 231 times better than that of 2k. However, comparison of 4-methoxy substituted 2g, 5-methoxy substituted 2j and unsubstituted 2l – all with K_i values >100 μM – such a trend could not be observed. All in all, it may be said that 4-methoxy substitution on ring A is preferred over either 5-methoxy substitution, 5,6-dimethoxy substitution or no substitution on ring A.

Additionally, when compound 2m – containing a methylenedioxy on ring A – was compared to compounds 2e, 2i and 2k, all with meta (3′) and para (4′) diOH-substitution on ring B, it was seen that A₁ AR affinity (>100 μM) was diminished, yet the compound retained relatively good A_2A AR affinity (A_2AK_i (rat) = 1.07 μM).

An unsubstituted ring A in combination with para (4′) N(CH₃)₂-substitution on ring B (2l) was compared to C4-OCH₃ substitution on ring A (2g) and C5-OCH₃ substitution on ring A (2j) in combination with para (4′) N(CH₃)₂-substitution on ring B of the 2-benzylidene-1-indanones and, although, none of these compounds possess micromolar affinity for either the A₁ or A_2A AR, it seems that the A₁ AR prefers no substitution on ring A in combination with N(CH₃)₂-substitution (2l) on ring B, rather than OCH₃-substitution at position C4 or C5 of ring A. The A_2A AR again favours C4-OCH₃ substitution over C5-OCH₃ and then no substitution on ring A.

Structural modification to ring B

Comparison of compound 2a (A₁K_i & A_2AK_i (rat) = >100 μM) to compound 2b (A₁K_i (rat) = 1.65 μM; A_2AK_i (rat) = >100 μM) showed that the benzylidene linked at position C2 on ring C increased both A₁ and A_2A AR affinity – conveying the necessity of the benzylidene at position C2 on ring C – this trend was also observed with regards to 5-substituted 2-benzylidene-1-tetralones and hydroxy substituted 2-benzylidene-1-indanones.36,37

The A₁K_i (rat) value of compound 2b (1.65 μM) when compared to the A₁K_i (rat) value of compound 2a (>100 μM) suggests that phenyl ring B is valuable to A₁ AR affinity.

In correlation with previous studies35–37 OH-group substitution on ring B favours A₁ and A_2A AR binding, especially meta (3′) hydroxy-group substitution (2c) and meta (3′) and para (4′) diOH-group substitution (2e).

It seems that both A₁ and A_2A AR's prefer either meta (3′) OH-group substitution (giving an A₁K_i (rat) value of 0.041 μM and an A_2AK_i (rat) value of 0.097 μM) or meta (3′) and para (4′) diOH-group substitution (giving an A₁K_i (rat) value of 0.042 μM and an A_2AK_i (rat) value of 0.078 μM); seeing as these K_i values are quite similar.

Interestingly, para (4′) hydroxy-substitution diminished both A₁ and A_2A AR affinity (2d: A₁ & A_2A (rat) K_i = >100 μM).

Other substitutions that diminish both A₁ and A_2A AR affinity include 2′,4′,5′-trimethoxy substitution on ring B (2f), as well as dimethylamino substitution (2g & 2j) when compared to its unsubstituted counterpart 2b (in the case of 2f and 2g) and 2h (in the case of 2j) (Fig. 3).

Fig. 3. Structural requirements of the 2-benzylidene-1-indanone scaffold for dual A₁/A_2A affinity.

These compounds were found to be E-isomers. The E-configuration is favourable for thermodynamic reasons because of steric interaction between the aryl and carbonyl groups in case of the Z-isomers.38,39

The benzylidene indanone scaffold (1a, 2b–m) contains an α,β-unsaturated ketone group perceived as a potential Michael acceptor.44,45 Although, compounds that act as Michael acceptors are generally biologically active,44,45 Michael acceptors are also notoriously reactive compound substructures.46,47 These compound substructures, containing an electrophile, might show reactivity towards nucleophiles such as thiols.47 Various biologically-relevant nucleophiles are thiols, for example glutathione, coenzyme A and protein cysteines.47 Off-target effects are often due to such compound substructures reactivity.47 Yet, a compound that acts as a Michael acceptor (and contains an electrophile) may still be useful after on-reaction with a suitable nucleophile to yield a lower energy compound.48

Compounds containing a catechol group, like 2e, 2i, 2k & 2m, are widespread in the literature as potential starting points to further explore structure activity relationships – yet, compounds that contain known reactive moieties, such as a catechol group, are also considered a liability;46–48 as catechols potentially are chelators, redox-active and oxidizes to form protein-reactive quinones.47,48 For example, the activity of 2e, with relatively good A₁ and A_2A AR affinity, may be due to oxidation of the ortho-hydroquinone compound substructure (or catechol group) to ortho-quinone – a protein-reactive quinone.

Reactive compound substructures are a risk factor in early drug discovery and development; as the results of biological assays are subject to interference from reactive moieties,47,48 like the α,β-unsaturated ketone group (present in all compounds in the current study, except 2a) or the catechol group (present in 2e, 2i, 2k & 2m). Failure to identify these moieties may lead to wasted resources, as such, compounds have poor development potential.47,48 Therefore a lack of reactivity interference, must be demonstrated,47 before further development.

GTP shift assay

GTP shift experiments are performed to determine whether the test compounds that exhibit A₁ AR affinity function as agonists or antagonists. For this purpose, compound 2c was selected as it exhibited the highest A₁ AR binding affinity among the investigated compounds. The affinities of the reference (CPA and DPCPX) and test compound 2c were determined in the absence and presence of 100 μM GTP and are reported with the calculated GTP shifts in Table 2. The calculated GTP shift results for CPA and DPCPX (see Table 2) were found to correspond with literature values, where CPA acts as an agonist (see Fig. 4) and DPCPX as an antagonist. Generally, a rightward shift of the binding curve in the presence of GTP (due to an uncoupling of the A₁ AR from its G_i protein) is expected for an A₁ AR agonist.42,51 In the case of an A₁ AR antagonist no significant shift is anticipated in the presence of GTP.42,51 The results suggest that compound 2c act as A₁ AR antagonist – as no significant rightward shift of the binding curve was observed in the presence of GTP (see Fig. 4).

Table 2. The A₁ AR affinities (in the absence and presence of GTP) and calculated GTP shifts of selected 2-benyzlidene-1-indanone analogues.

#	K i ± SEM (μM) a		GTP shift d
	A1 b vs. [3H]DPCPX	A1 b + GTP c vs. [3H]DPCPX
Lead compound
1a	0.435 ± 0.050 e	0.339 ± 0.071 e	0.90 e
Methoxy substituted 2-benzylidene-1-indanones
2c	0.041 ± 0.002 a , b	0.060 ± 0.002 a , c	1.46 d
Reference compounds
CPA (A1 agonist)	0.0068 ± 0.0001 a	0.099 ± 0.015 a	15
	(0.0079); f	(0.099) g	(14) g
	(0.015) g
DPCPX (A1 antagonist)	0.0004 ± 0.0002 a	0.0004 ± 0.0002 a	1.0
	(0.0005); g	(0.0004) g	(1.0) g
	(0.0003) h

^aAll K_i values determined in triplicate and expressed as mean ± SEM.

^bRat receptors were used (A₁: rat whole brain membranes).

^cGTP shift assay, where the 100 μM GTP was added to the A₁ AR radioligand binding assay.

^dGTP shifts calculated by dividing the K_i in the presence of GTP by the K_i in the absence of GTP.

^eLiterature value obtained from reference.37

^fLiterature value obtained from reference.49

^gLiterature value obtained from reference.42

^hLiterature value obtained from reference.50

Fig. 4. The binding curves of compounds 2c and CPA (reference compound) are examples of A₁ AR antagonistic action (A) and A₁ AR agonistic action (B), respectively, determined via a GTP shift assays (with and without 100 μM GTP) in rat whole brain membranes expressing A₁ ARs with [³H]DPCPX as radioligand. (A) GTP shift of 1.46 calculated for compound 2c, (B) GTP shift of 15 calculated for compound CPA.

Conclusions

In summary, this study involved the synthesis, characterization and evaluation of methoxy substituted 2-benzylidene-1-indanone derivatives to understand the importance of structural modifications to ring A and B of the 2-benzylidene-1-indanone scaffold in gaining or even losing A₁ and/or A_2A AR affinity. Upon analysis, it was found that C4-OCH₃ substitution on ring A (2e) is preferred to C4-OH substitution (1a). Additionally, meta (3′) OH-group substitution (2c; A₁K_i (rat) = 41 nM; A_2AK_i (rat) = 97 nM) and meta (3′) and para (4′) diOH-group substitution (2e; A₁K_i (rat) = 42 nM; A_2AK_i (rat) = 78 nM) on ring B along with C4 OCH₃-substitution on ring A is complimentary to A₁ and A_2A AR affinity, affording these non-selective compounds K_i values in the nanomolar range for both the A₁ and A_2A AR. To reiterate the significance of the aforementioned, 2c and 2e showed an approximately tenfold increase in A₁ and A_2A AR affinity when compared to lead compound 1a. Other compounds showed A₁ affinity (2b) and A_2A affinity (2m) in the micromolar range, whereas compounds 2i and 2k showed dual A₁ and A_2A AR affinity. Yet, none of the other target compounds come close to 2c and 2e with regards to A₁ and A_2A AR affinity. Functional characterization of 2c proved this compound to be an A₁ AR antagonist. In view of these findings, compounds 2c and 2e present the potential starting points to further explore the structure activity relationships for affinities of this class of compounds as ligands for A₁ and A_2A AR. However, since both 2c and 2e contain potential reactive compound substructures, experiments that demonstrate the absence of reactivity interference using a variety of methods must be performed. If necessary, the undesirable reactive compound substructures may be removed or modified during medicinal chemistry optimization.

Experimental

Chemistry

General remarks

Unless otherwise noted, all starting materials and solvents were procured from Sigma-Aldrich and used without further purification. Proton (¹H) and carbon (¹³C) nuclear magnetic resonance (NMR) spectra were recorded on a Bruker Avance III 600 spectrometer at frequencies of 600 MHz and 151 MHz, respectively, with deuterated dimethylsulfoxide (DMSO-d₆) as solvent. Chemical shifts are reported in parts per million (δ) in relation to the signal of tetramethylsilane (Si(CH₃)₄). Spin multiplicities are indicated as: s (singlet), d (doublet), dd (doublet of doublets), td (triplet of doublets), t (triplet), q (quartet) and m (multiplet). High resolution mass spectra (HRMS) were recorded on a Bruker micrOTOF-Q II mass spectrometer in atmospheric pressure chemical ionisation (APCI) mode. High performance liquid chromatography (HPLC) analyses were determined on an Agilent 1100 HPLC system. Melting points (mp) for 2a–g were measured with a Buchi B545 melting point apparatus, whereas mp's for 2i, 2k & 2m were measured by means of differential scanning calorimetry (DSC) with a Mettler DSC 3 Star System (Mettler Toledo, Greifensee, Switzerland) and are uncorrected and are uncorrected. Thin layer chromatography (TLC) was done using silica gel 60 (Merck) with UV254 fluorescent indicator.

Synthesis of 2a

4-Methoxy-2,3-dihydro-1H-inden-1-one (2a)

To a solution of 4-hydroxy-1-indanone (2.00 g, 13.5 mmol) in acetone (70.0 mL), K₂CO₃ (7.46 g, 54.0 mmol) and then MeI (11.5 g, 81.0 mmol) were added and mechanically stirred at 50 °C under reflux for 18 h. Next, the reaction mixture was concentrated, extracted with EtOAc (3 × 100 mL), combined organic extracts dried (MgSO₄), filtered and concentrated to yield compound 2a as brown crystals (2.00 g, 91%): R_f: 0.51 (PE/EtOAc 4 : 1); mp: 105.3–107.1 °C; ¹H NMR (600 MHz, DMSO) δ 2.61 (dd, J = 6.7, 4.6 Hz, 2H), 2.97–2.92 (m, 2H), 3.87 (s, 3H), 7.22 (dd, J = 18.8, 7.7 Hz, 2H), 7.40 (t, J = 7.7 Hz, 1H); ¹³C NMR (151 MHz, DMSO) δ 22.31, 35.75, 55.52, 114.48, 115.48, 129.09, 138.20, 143.52, 156.89, 206.39. APCI-HRMS m/z calculated for C₁₀H₁₀O₂ (MH⁺): 163.0753, found: 163.0754. Purity (HPLC): 98.6%.

Synthesis of 2b–g, 2i, 2k and 2m

(E)-2-Benzylidene-4-methoxy-2,3-dihydro-1H-inden-1-one (2b)

Compound 2a (0.200 g, 1.233 mmol) and benzaldehyde (0.131 g, 1.233 mmol) were suspended in MeOH (4 mL) and HCl (32%, 6 mL) and mechanically stirred at 120 °C under reflux for 24 h. Thereafter, the reaction mixture was cooled to room temperature, ice (20 g) was added and the resulting precipitate was filtered, dried (60 °C) and recrystallized from a suitable solvent (either MeOH or EtOH) to yield 2b as light brown crystals (0.22 g, 71%): R_f: 0.68 (PE/EtOAc 3 : 1); mp: 385.3–385.5 °C; ¹H NMR (600 MHz, DMSO) δ 3.96–3.90 (m, 5H), 7.29 (d, J = 7.9 Hz, 1H), 7.36 (d, J = 7.5 Hz, 1H), 7.57–7.43 (m, 5H), 7.78 (d, J = 7.4 Hz, 2H); ¹³C NMR (151 MHz, DMSO) δ 28.90, 55.56, 115.20, 116.00, 129.05, 129.44, 129.86, 130.79, 133.08, 134.71, 134.78, 138.05, 138.59, 156.50, 193.30. APCI-HRMS m/z calculated for C₁₇H₁₄O₂ (MH⁺): 251.1067, found: 251.1088. Purity (HPLC): 98.1%.

(E)-2-(3-Hydroxybenzylidene)-4-methoxy-2,3-dihydro-1H-inden-1-one (2c)

Prepared as for 2b from compound 2a (0.300 g, 1.850 mmol) and 3-hydroxybenzaldehyde (0.226 g, 1.850 mmol) to yield compound 2c as light brown crystals (0.44 g, 90%): R_f: 0.29 (PE/EtOAc 3 : 1); mp: 104.3–104.4 °C; ¹H NMR (600 MHz, DMSO) δ 3.90 (d, J = 14.8 Hz, 5H), 6.86 (dd, J = 8.0, 1.9 Hz, 1H), 7.20 (dd, J = 10.1, 4.7 Hz, 2H), 7.38–7.26 (m, 3H), 7.45 (dd, J = 15.6, 7.9 Hz, 2H), 9.73 (s, 1H); ¹³C NMR (151 MHz, DMSO) δ 29.01, 55.56, 115.20, 115.98, 116.59, 117.24, 122.39, 129.45, 130.07, 133.30, 134.43, 135.95, 137.97, 138.65, 156.51, 157.73, 193.32. APCI-HRMS m/z calculated for C₁₇H₁₄O₃ (MH⁺): 267.1016, found: 267.1017. Purity (HPLC): 100%.

(E)-2-(4-Hydroxybenzylidene)-4-methoxy-2,3-dihydro-1H-inden-1-one (2d)

Prepared as for 2b from compound 2a (0.300 g, 1.850 mmol) and 4-hydroxybenzaldehyde (0.226 g, 1.850 mmol) to yield compound 2d as mustard powder (0.45 g, 92%): R_f: 0.14 (PE : EtOAc 3 : 1); mp: 275.3–396.2 °C; ¹H NMR (600 MHz, DMSO) δ 3.86 (d, J = 0.4 Hz, 2H), 3.91 (s, 3H), 6.91 (d, J = 8.6 Hz, 2H), 7.27 (d, J = 7.9 Hz, 1H), 7.33 (d, J = 7.5 Hz, 1H), 7.44 (dd, J = 12.7, 4.8 Hz, 2H), 7.64 (d, J = 8.7 Hz, 2H), 10.15 (s, 1H); ¹³C NMR (151 MHz, DMSO) δ 28.97, 55.53, 115.06, 115.63, 116.10, 125.89, 129.30, 131.13, 133.03, 133.61, 137.69, 139.03, 156.48, 159.50, 193.19. APCI-HRMS m/z calculated for C₁₇H₁₄O₃ (MH⁺): 267.1016, found: 267.1020. Purity (HPLC): 100%.

(E)-2-(3,4-Dihydroxybenzylidene)-4-methoxy-2,3-dihydro-1H-inden-1-one (2e)

Prepared as for 2b from compound 2a (0.300 g, 1.850 mmol) and 3,4-dihydroxybenzaldehyde (0.255 g, 1.850 mmol) to yield compound 2e as light yellow crystals (0.47 g, 90%): R_f: 0.29 (PE/EtOAc 3 : 1); mp: 294.0–295.1 °C; ¹H NMR (600 MHz, DMSO) δ 3.84 (s, 2H), 3.92 (s, 3H), 6.87 (d, J = 8.2 Hz, 1H), 7.10 (dd, J = 8.3, 1.9 Hz, 1H), 7.27 (t, J = 5.1 Hz, 2H), 7.33 (d, J = 7.5 Hz, 1H), 7.37 (s, 1H), 7.44 (t, J = 7.7 Hz, 1H); ¹³C NMR (151 MHz, DMSO) δ 29.05, 55.54, 115.07, 115.63, 116.13, 117.08, 124.73, 126.32, 129.33, 130.89, 134.08, 137.62, 139.11, 145.69, 148.19, 156.49, 193.17. APCI-HRMS m/z calculated for C₁₇H₁₄O₄ (MH⁺): 283.0965, found: 283.0959. Purity (HPLC): 100%.

(E)-2-(2,4,5-Trimethoxybenzylidene)-4-methoxy-2,3-dihydro-1H-inden-1-one (2f)

Prepared as for 2b from compound 2a (0.300 g, 1.850 mmol) and 2,4,5-trimethoxybenzaldehyde (0.363 g, 1.850 mmol) to yield compound 2f as green powder (0.34 g, 54%): R_f: 0.14 (PE/EtOAc 4 : 1); mp: 217.6–218.0 °C; ¹H NMR (600 MHz, DMSO) δ 3.82 (s, 3H), 3.94–3.87 (m, 9H), 3.95 (d, J = 1.3 Hz, 2H), 6.79 (s, 1H), 7.37–7.27 (m, 3H), 7.45 (t, J = 7.8 Hz, 1H), 7.45 (t, J = 7.8 Hz, 1H), 7.89 (t, J = 1.9 Hz, 1H); ¹³C NMR (151 MHz, DMSO) δ 28.52, 55.60, 55.86, 56.47, 97.64, 113.20, 114.49, 115.08, 115.66, 127.30, 129.29, 131.31, 137.70, 139.08, 142.73, 152.39, 154.91, 156.45, 167.00, 193.20. APCI-HRMS m/z calculated for C₂₀H₂₀O₅ (MH⁺): 341.1384, found: 341.1393. Purity (HPLC): 95.1%.

(E)-2-(4-(Dimethylamino)benzylidene)-4-methoxy-2,3-dihydro-1H-inden-1-one (2g)

Prepared as for 2b from compound 2a (0.300 g, 1.850 mmol) and 4-(dimethylamino)benzaldehyde (0.276 g, 1.850 mmol) to yield compound 2g as brown crystals (0.33 g, 61%): R_f: 0.46 (PE/EtOAc 4 : 1); mp: 190.7–191.7 °C; ¹H NMR (600 MHz, DMSO) δ 3.01 (s, 6H), 3.85 (s, 2H), 3.92 (s, 3H), 6.85–6.77 (m, 2H), 7.26 (d, J = 7.9 Hz, 1H), 7.32 (d, J = 7.4 Hz, 1H), 7.44 (dd, J = 9.6, 5.0 Hz, 2H), 7.62 (d, J = 8.9 Hz, 2H); ¹³C NMR (151 MHz, DMSO) δ 29.16, 55.53, 112.01, 114.93, 115.32, 122.13, 129.09, 129.20, 132.72, 134.31, 137.35, 139.45, 151.23, 156.43, 192.86. APCI-HRMS m/z calculated for C₁₉H₁₉NO₂ (MH⁺): 294.1489, found: 294.1499 Purity (HPLC): 97.9%.

(E)-2-(3,4-Dihydroxybenzylidene)-5-methoxy-2,3-dihydro-1H-inden-1-one (2i)

Prepared as for 2b from 5-methoxy-1-indanone (0.300 g, 1.850 mmol) and 3,4-dihydroxybenzaldehyde (0.255 g, 1.850 mmol) to yield compound 2i as brown powder (0.33 g, 63%): R_f: 0.13 (DCM/EtOAc 10 : 1); mp: 281.26 °C; ¹H NMR (600 MHz, DMSO) δ 3.88 (s, 3H), 3.96 (s, 2H), 6.85 (d, J = 8.2 Hz, 1H), 7.01 (dd, J = 8.5, 2.2 Hz, 1H), 7.08 (dd, J = 8.3, 2.0 Hz, 1H), 7.17 (dd, J = 12.5, 2.0 Hz, 2H), 7.29 (t, J = 1.7 Hz, 1H), 7.69 (d, J = 8.5 Hz, 1H), 9.24 (s, 1H), 9.64 (s, 1H); ¹³C NMR (151 MHz, DMSO) δ 32.09, 55.76, 110.23, 115.15, 116.02, 117.44, 123.87, 125.20, 126.57, 130.91, 131.91, 132.43, 145.58, 147.75, 152.55, 164.61, 191.59. APCI-HRMS m/z calculated for C₁₇H₁₄O₄ (MH⁺): 283.0965, found: 283.0972. Purity (HPLC): 95.4%.

(E)-2-(3,4-Dihydroxybenzylidene)-5,6-dimethoxy-2,3-dihydro-1H-inden-1-one (2k)

Prepared as for 2b from 5,6-dimethoxy-1-indanone (0.300 g, 1.561 mmol) and 3,4-dihydroxybenzaldehyde (0.216 g, 1.561 mmol) to yield compound 2k as dark green powder (0.41 g, 84%): R_f: 0.12 (DCM/EtOAc/PE 8 : 1 : 1); mp: 262.66 °C; ¹H NMR (600 MHz, DMSO) δ 3.83 (s, 12H), 3.90 (s, 20H), 6.85 (d, J = 8.2 Hz, 4H), 7.07 (dd, J = 8.3, 2.0 Hz, 4H), 7.18 (dd, J = 12.3, 4.1 Hz, 12H), 7.27 (s, 4H), 9.43 (s, 7H); ¹³C NMR (151 MHz, DMSO) δ 31.69, 55.65, 55.96, 104.50, 108.07, 116.01, 117.41, 123.75, 126.62, 130.32, 131.98, 132.19, 144.61, 145.58, 147.67, 149.25, 154.94, 191.88. APCI-HRMS m/z calculated for C₁₈H₁₆O₅ (MH⁺): 313.1071, found: 313.0993. Purity (HPLC): 98.6%.

(E)-6-(3,4-Dihydroxybenzylidene)-6,7-dihydro-5H-indeno[5,6-d][1,3]dioxol-5-one (2m)

Prepared as for 2b from 5,6-methylenedioxy-1-indanone (0.300 g, 1.703 mmol) and 3,4-dihydroxybenzaldehyde (0.235 g, 1.703 mmol) to yield compound 2m as green powder (0.45 g, 90%): R_f: 0.09 (DCM/EtOAc 10 : 1); mp: 317.96 °C; ¹H NMR (600 MHz, DMSO) δ 3.88 (s, 2H), 6.17 (s, 2H), 6.84 (d, J = 8.2 Hz, 1H), 7.06 (dd, J = 8.2, 1.7 Hz, 1H), 7.15 (dd, J = 9.0, 2.4 Hz, 3H), 7.25 (s, 1H), 9.25 (s, 1H), 9.63 (s, 1H); ¹³C NMR (151 MHz, DMSO) δ 31.97, 101.97, 102.37, 105.92, 116.01, 117.33, 123.93, 126.49, 131.95, 132.18, 132.28, 145.59, 146.99, 147.77, 148.03, 153.48, 191.34. APCI-HRMS m/z calculated for C₁₇H₁₂O₅ (MH⁺): 297.0758, found: 297.0755. Purity (HPLC): 87.2%.

Biology

General remarks

All commercially available reagents were obtained from various manufacturers: radioligands [³H]NECA (specific activity 27.1 Ci mmol^–1) procured from PerkinElmer and [³H]DPCPX (specific activity 120 Ci mmol^–1) 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.

Radioligand binding assays

The collection of tissue samples for the A₁ and A_2A AR binding studies were approved by the Research Ethics Committee of the North-West University (application number NWU-0035-10-A5). The affinities of the 2-benzylidene-1-indanone analogues (2a–m) at rat A₁ and A_2A AR subtypes were determined with radioligand competition experiments as described previously.22,23 The competition experiments were carried out in the presence of the radioligands [³H]-8-cylcopentyl-1,3-dipropylxanthine ([³H]DPCPX; 0.1 nM; K_d = 0.36 nM) and 5′-N-[³H]-ethylcarboxamideadenosine ([³H]NECA; 4 nM; K_d = 15.3 nM) for the A₁ and A_2A AR radioligand binding assays, respectively.22,24 In addition, the A_2A AR binding studies were determined in the presence of N⁶-cyclopentyladenosine (CPA) to minimize the binding of [³H]NECA to A₁ AR's. Non-specific binding was defined by the addition of a final concentration of 100 μM CPA. The sigmoidal-dose response curves, via Graphpad Software Inc. package, were obtained by plotting the specific binding versus the logarithm of the test compound's concentrations. Subsequently, the K_i values were obtained by using the IC₅₀ values that were determined from sigmoidal-dose response curves. All incubations were carried out in triplicate and the K_i values are expressed as the mean ± standard error of mean (SEM). CPA and DPCPX (unlabelled) were used as reference compounds and their assay results confirmed validity of the radioligand binding assays (see Table 1).

GTP shift assays

The GTP shift assay was performed as described previously with rat whole brain membranes and [³H]DPCPX (0.1 nM; K_d = 0.36 nM) in the absence and presence of a final concentration of 100 μM GTP (see Table 2).22,23 Non-specific binding was defined by the addition of 10 μM DPCPX (unlabelled). If a calculated GTP shift of approximately 1 is obtained, that compound is considered to function as an antagonist. On the other hand, the presence of GTP affects the competition curve of an agonist and shifts the curve to the right, as previously demonstrated by the A₁ AR agonist CPA.22 The sigmoidal-dose response curves were obtained via the Graphpad Software Inc. package and the K_i values determined as described above. The GTP shift was calculated by dividing the K_i value of a compound reported in the presence of GTP by the K_i value obtained in the absence of GTP.22

Conflicts of interest

There are no conflicts to declare.