Identification of C10 nitrogen-containing aporphines with dopamine D₁ versus D₅ receptor selectivity

Abstract

New aporphines containing C10 nitrogen substituents (viz. nitro, aniline or amide moieties), were synthesized and evaluated for affinity at human serotonin 5-HT_1A and 5-HT_2A receptors and at human dopamine D₁, D₂ and D₅ receptors. Two series of analogs were investigated: series A which contain a sole C10 nitrogen substituent on the tetracyclic aporphine core and series B which are 1,2,10-trisubstituted aporphines. Remarkably, compounds from both series lacked affinity for the D₅ receptor, thus attaining D₁ versus D₅ selectivity. Compound 20c was the most potent D₁ ligand identified. Docking studies at D₁ and D₅ receptors indicate that the binding mode of 20c at the D₁ receptor allows for stronger hydrophobic contacts, (primarily with Phe residues) as compared to the D₅ receptor, accounting for its D₁ versus D₅ selectivity. Considering the lack of affinity for the D₅ receptor (and low affinity at other receptors tested), compound 20c represents an interesting starting point for further structural diversification of aporphines as sub-type selective D₁ receptor tools.

Graphical Abstract

There are two families of dopamine receptors – D₁-like and D₂-like. The D₁-like family includes D₁ and D₅ receptors while the D₂-like family comprises D₂, D₃ and D₄ receptors. D₁-like receptors are involved in a variety of physiological processes including locomotion, cognition and reward.^1-3 There is a dearth of ligands that can discriminate between D₁ and D₅ receptors. Currently known D₁-like receptor ligands tend to bind to both D₁ and D₅ receptors (often with similar affinities), a consequence of the close structural homology between D₁ and D₅ receptors (>80% sequence homology in transmembrane regions).⁴ Compounds that are selective for the D₁ receptor versus the D₅ receptor will be useful pharmacological tools to more fully distinguish the physiological functions of D₁ and D₅ receptors. This fundamental knowledge could also be useful for the discovery and development of therapeutics where the D₁ receptor is implicated such as Parkinson’s disease, schizophrenia and substance use disorders.^5-10 Thus, there is a continued interest in the identification of D₁ receptor ligands that are selective versus the D₅ receptor.

Aporphine alkaloids are known to display affinity for serotonin and dopamine receptors.^11-21 For example, we have identified ligands structurally related to nantenine (1) with affinity for serotonin receptors.^{22, 23} The prototypical aporphine, (R)-apomorphine (2) is a clinically approved drug for the treatment of Parkinson’s disease.^24-27

The anti-Parkinsonian effects apomorphine are attributable, at least in part to its affinity and agonist actions as a dopamine mimic at dopamine D₁-like and D₂-like receptors.^28-30

We considered that structurally modified apomorphine derivatives may be identified with selective D₁ receptor affinity. To date there have been no structure-activity studies on aporphines that contain a C10 nitrogen substituent at dopamine receptors. To begin to fill this current deficit in knowledge, we embarked on a project to prepare and evaluate via a structure-affinity relationship (SAR) study, 2 series of aporphines that contain 10-substituted (series A) and 1,2,10-trisubstituted patterns (series B), in which the C10 functional group is nitrogenous (either nitro, aniline or amide). The compounds were prepared as racemates and were evaluated across a subset of serotonin and dopamine receptors in order to gauge their D₁ receptor selectivity.

Synthesis of the series A analogs (Scheme 1) commenced with the readily available 2-(2-bromo-4-nitrophenyl) acetic acid (3) and followed similar synthetic strategies to aporphines as previously reported by us and others.^31-34 Coupling of 3 with phenethylamine (4) gave amide 5. Cyclization of 5 under Bischler-Napieralski conditions gave an intermediate imine which was subsequently reduced to afford an intermediate tetrahydroisoquinoline. The secondary amine functionality of this intermediate was Boc-protected and the aryl bromide (6) thus produced was cyclized via palladium-catalyzed direct arylation to afford aporphine 7.^{31, 33, 34} The Boc group of 7 was removed with acid and the resulting secondary amine 8 was then N-methylated to give 9. Reduction of the nitro group of 9 to give amine 10 was followed by acylation to give the C10 acetamide and butamide analogs 11a and 11b respectively.

Scheme 1. Reagents and conditions:

(a) HBTU, THF, 4, DIPEA, rt, 96% (b) POCl₃, ACN, reflux, 2h (c) NaBH₄, MeOH, 0 °C (d) (Boc)₂O, K₂CO₃, ACN, rt, 60% over 3 steps (e) Pd(OAc)₂, tris(4-fluorophenyl)phosphine, K₂CO₃, (CH₃)₃CCOOH, DMA, 100 °C, 70% (f) TFA, CHCl₃, rt, 78% (g) formaldehyde, NaBH(OAc)₃, DCM, rt, 75% (h) HSiCl₃, ACN, 0°-rt, 18 h, 80% (i) acetic anhydride/butyric anhydride, CHCl₃, rt, 3h, 45% and 34%

The series B analogs (18a-e, 19a-e and 20a-c) were prepared as outlined in Scheme 2 using similar chemistry as described above for the series A analogs. Thus, coupling of amine 12 with acid 3 gave the amide 13 which was subjected to sequential Bischler-Napieralski cyclization, imine reduction, O-debenzylation and di-N,O-Boc protection thus affording 14. Cyclization of 14 to aporphine 15 was followed by selective O-Boc deprotection to reveal phenol 16. Alkylation of phenol 16 with various alkyl halides gave the O-alkylated, N-Boc protected derivatives 17a-d. Compound 16 was transformed into 18a in two steps via sequential N-Boc cleavage and N-methylation. A similar two-step sequence was used to prepare 18b – e from compounds 17a – d respectively. Reduction of the nitro group in 18a-e provided the corresponding aniline analogs 19a-e. Thereafter, acylation of compounds 19a, 19d and 19e with acetic anhydride gave analogs 20a, 20b and 20c respectively.

Scheme 2. Reagents and conditions:

(a) HBTU, THF, 12, DIPEA, rt , 96% (b) POCl₃, DCM, 0 °C - rt, 16h (c) NaBH₄, MeOH, 0 °C (d) HCl (conc.), EtOH, reflux, 2h (e) (Boc)₂O, K₂CO₃, ACN, rt , 30% over 4 steps (f) Pd(OAc)₂, tris(4-fluorophenyl)phosphine, K₂CO₃, (CH₃)₃CCOOH, DMA, 60 °C, overnight, 66% (g) 1M NaOH, dioxane, 100 °C, 3h, 80% (h) alkyl bromide, K₂CO₃, ACN, reflux, overnight, 90% (i) TFA, DCM, rt (j) formaldehyde, NaBH(OAc)₃, DCM, rt (50%-80%) (k) HSiCl₃, ACN, 0 °C-rt, 18h, 62-95% (l) acetic anhydride, CHCl₃, rt, 2h, (45%-65%)

Analogs from series A (compounds 8 – 10, 11a and 11b) and series B (compounds 18a-e, 19a-e and 20a-c) were evaluated for binding affinity at human serotonin 5-HT_1A, 5-HT_2A and dopamine D₁, D₂ and D₅ receptors. The data from these evaluations are reported as K_i values in Table 1.

Table 1.

Binding affinity of series A and series B analogs at human serotonin and dopamine receptors

				Ki(nM)a
Cmpd #	R1	R2	R3	5-HT1A	5-HT2A	D1	D2	D5
8	NO2	-	H	350±45	nab	na	na	na
9	NO2	-	Me	100±13	na	928±120	na	na
10	NH2	-	Me	269±0.35	672±87	na	679±88	na
11a	NHCOMe	-	Me	93±12	1062±140	na	na	na
11b	NHCO(CH2)2Me	-	Me	83±1.1	na	na	na	na
18a	NO2	H	Me	548±71	908±120	316±41	320±41	na
18b	NO2	Me	Me	458±59	1347±170	726±94	478±62	na
18c	NO2	cyMe	Me	439±57	na	na	na	na
18d	NO2	allyl	Me	1676±220	na	na	na	na
18e	NO2	propargyl	Me	206±27	na	536±69	2048±260	na
19a	NH2	H	Me	na	850+110	301±39	231±30	418±54
19b	NH2	Me	Me	156±20	585±75	691±89	260±34	na
19c	NH2	cyMe	Me	na	110±14	1111±140	963±120	na
19d	NH2	allyl	Me	na	234±30	830±110	759±98	na
19e	NH2	propargyl	Me	428±55	378±49	709±91	431±56	na
20a	NHCOMe	Me	Me	na	na	756±98	978±130	na
20b	NHCOMe	allyl	Me	na	1220±160	155±20	na	na
20c	NHCOMe	propargyl	Me	na	1272±160	58±7.5	854±110	na
8-OH DPAT				0.8±0.01
Clozapine					3.03±0.39
(+)-butaclamol						2.0±0.26
Haloperidol							1.97±0.25
SKF 83566								2.3±0.3
Apomorphinec				117	120	372	82	15

^aExperiments carried out in triplicate

^bna – not active (compounds displayed < 50% inhibition in a primary assay)

^cData from reference 30

The compounds in series A had weak or no affinity at the serotonin 5-HT_2A receptor and at dopamine receptors. Interestingly, all compounds in this series lacked affinity for the D₅ receptor. Compounds 8 and 11b had similar selectivity profiles in that both have exclusive affinity for the 5-HT_1A receptor. Of these, compound 11b stands out given its higher affinity.

Within series B, compounds with a C10 nitro functionality (18a-e) displayed weak to moderate affinity for the 5-HT_1A receptor, ranging in K_i values from approximately 200 to 1700 nM. The aniline counterparts of 18a-e (i.e. 19a-e), generally displayed lower affinity for the 5-HT_1A receptor. Unlike the case with series A, acylation of the aniline functionality (i.e. 20a-c) did not improve 5-HT_1A affinity. Therefore, it appears that of the C10 nitrogen functionalities examined, a nitro group at C10 is generally best tolerated for 5-HT_1A affinity. No clear trend was apparent with regards to the steric or electronic effects of the C1 alkoxy groups in 18a-e on 5-HT_1A receptor affinity, a situation which was echoed for analogs 19a-e. However for both the nitro containing (18a-e) and the aniline containing groups (19a-e), it would seem that a C1 propargyloxy functionality has better tolerance (18e, K_i = 206 nM; 19e K_i = 430 nM) than a C1 allyloxy group (18d, K_i = 1670 nM; 19d, no affinity).

At the 5-HT_2A receptor, compounds containing a C10 nitro functionality displayed weak or no affinity. The aniline derivatives 19a-e, all had improved 5-HT_2A receptor affinity as compared to their C10 nitro counterparts, while the amide derivatives 20a-c, had at least 3-fold lower 5-HT_2A receptor affinities than the corresponding aniline precursors. This indicates that the C10 aniline functionality is better tolerated than either C10 nitro or C10 amide functionalities for 5HT_2A receptor affinity.

At the dopamine D₁ receptor, compounds 18a, 18b and 18e had moderate affinity (320, 730 and 540 nM respectively). The other compounds in the nitro containing sub-group lacked D₁ receptor affinity. On examination of the data for 18a-d, it would appear that smaller groups are better tolerated at the C1 position with the nitro-containing compounds. However, the moderate affinity of compound 18e with the C1 propargyloxy moiety does not fit this trend. It is possible that the presence of the C1 propargyloxy group in 18e allows the molecule to make unique interactions with the D₁ receptor that contribute positively to its affinity. The C10 aniline containing compounds 19a-e had D₁ receptor affinities ranging from approximately 300 - 1100 nM. Generally, as compared to the C10 nitro containing compounds, the C10 aniline containing congeners displayed similar or improved D₁ receptor affinities (except for 19e where D₁ receptor affinity was lower when compared to 18e). There was no clear SAR trend in comparing the amide derivatives 20a-c with their aniline precursors. In the case of 20a, N-acylation led to a slight decrease in affinity, while with 20b and 20c N-acylation gave up to a 12-fold increase in D₁ receptor affinity (e.g. 19e, K_i = 710 nM; 20c, K_i= 58 nM). Compound 20c had the highest D1 receptor affinity of any compound measured from series A and series B. SAR trends at the D₂ receptor were similar to those seen at the D₁ receptor where compounds with smaller C1 alkoxy (or propargyloxy) substituents displayed moderate affinity within the C10 nitro containing sub-group of compounds. C10 aniline and C10 amide containing compounds also had moderate D₂ receptor affinity, except for the C1 allyloxy derivative 20b, which had no D₂ receptor affinity. The D₂ receptor affinities of the series B compounds fell in a K_i range of approximately 300 - 2000 nM. Compounds 19a-e had marginally higher D₂ receptor affinity versus D₁ receptor affinity.

At the D₅ receptor compound 19a was the only compound to show affinity. In fact, it was the only compound from either series A or series B that had affinity for the D₅ receptor, highlighting the fact that as a group, these 1,2,10-trisubstituted and 10-substituted analogs have preference for D₁ over D₅ receptors.

Apomorphine is known to suffer from poor pharmacokinetic properties, which is attributable to the presence of the readily biotransformed catechol functionality. As compound 20c lacks this catechol moiety, we were interested to see the extent to which 20c would be more metabolically stable than apomorphine. Thus, the metabolic stability of compound 20c was assessed in human liver microsomes (HLMs). The data (Table 2) show that as compared to apomorphine, compound 20c is significantly more metabolically stable. This increased metabolic stability is a promising physicochemical characteristic that supports further development of aporphine-derived C10 amides such as compound 20c as in vivo tools. Furthermore, 20c has calculated properties (ChemBioDraw Ultra version 11) which are indicative of good blood-brain barrier (BBB) penetrability and “drug-likeness”.³⁵ The clogP for 20c is 3.08, it contains 5 (N+O) groups, its total polar surface area is 50.8 Ų and its molecular mass is 377 amu. All of these parameters are within range for good BBB penetration.³⁶

Table 2.

Microsomal stability assay in HLMs^a

Test article	Percent remaining (%)
	0 min	15 min	30 min	60 min	120 min
propanthelineb	100	65.33	36.32	6.40	0.16
Apomorphine (2)	100	89.33	99.73	80.43	36.68
20c	100	99.76	97.80	95.60	102.69

^aExperiments performed in duplicate

^bcontrol

Computational simulations were conducted in order to rationalize the D₁ versus D₅ selectivity of compound 20c. The D₁ receptor homology model was generated from the high-resolution crystal structure of the human β₂-adrenergic G protein-coupled receptor with pdb code 2RH1 ³⁷ followed by induced fit docking involving several benzazepine analogs. Similarly, the D₅ receptor homology model was developed from the high-resolution crystal structure of the β₁-adrenergic G protein-coupled receptor with pdb codae 6H7J³⁷ followed by induced fit docking with the benzazepine analogs. Models for the D₁ and D₅ receptor structures were, thus, prepared with suitable backbone and side-chain orientations within the binding site. This process was conducted by application of the Schrödinger Prime Structure Prediction and Glide software modules and manual intervention to support the generation of known key receptor-ligand interactions. The docking investigations of compound 20c into the D₁ and D₅ receptor binding sites exploited the Schrödinger Glide methodology in Standard Precision (SP) mode ³⁸. In this context, the Glidescore scoring function was used to give an estimate of the ligand binding affinities and the values of the highest ranked pose for 20c in the D₁ and D₅ target receptors.

The docked pose for compound 20c in the dopamine D₁ receptor shown in Figure 2A illustrates the key salt bridge formed between the quaternary N atom and Asp103 and the π - cation interaction between the same quaternary N and Trp99. There are also important hydrophobic interactions involving the fused rings of the ligand and Phe288, Phe289 and Phe313. At the dopamine D₅ receptor, compound 20c displays similar receptor-ligand interactions as depicted in Figure 2B, involving the quaternary N – Asp120 salt bridge and the Phe312 π – quaternary N cation interaction. The predicted energy is slightly less favorable for the D₅ receptor (−6.7 kcal/mol) compared to D₁ (−7.5 kcal/mol) mainly due to the formation of fewer complementary hydrophobic interactions involving the receptor Phe residues.

Figure 2.

Docked poses of the compound 20c (green carbon atoms) in A. the dopamine D₁ receptor target (grey carbon atoms) and B. the dopamine D₅ receptor target (grey carbon atoms). Key quaternary N – Asp salt bridges are depicted by the pink dashed lines and corresponding π – cation interactions by the green dashed lines. Docking was performed with the R enantiomeric form of compound 20c.

In conclusion, this SAR study has provided important information concerning the structural tolerance for affinity of aporphines at serotonin and dopamine receptors, particularly as it relates to the presence of nitrogen containing motifs at C10, which has not been previously studied. The tolerance for C10 nitro, aniline and amide functionalities at the various receptors was found to vary depending on the specific substituents appended to the aporphine core.

The lack of D₅ receptor affinity of the compounds is quite interesting given the fact that previously identified D₁ receptor ligands also tend to display comparable D₅ receptor affinity. In fact, as alluded to earlier, commercially available D₁ receptor ligands cannot adequately discern D₁ and D₅ receptors in pharmacological studies. D₁ versus D₅ receptor selectivity has been observed in only a handful of compounds from the arylbenzazepine class³⁹ and in a few compounds reported as allosteric D₁ receptor binders.⁴⁰ As evident from Table 1, several of the compounds examined had D₁ receptor affinity but no D₅ receptor affinity making for quite an unusual pharmacology. This study reaffirms our notion that aporphines may be used as templates for the identification of D₁ receptor selective ligands. Clearly, the D₁ receptor affinity of the compounds herein needs refinement in order to make them useful D1 receptor tools, but their D₁ versus D₅ selectivity is noteworthy and valuable. In this regard, compound 20c appears to be an interesting starting point from which to extrapolate in future work towards the discovery of novel high affinity, D₁ receptor sub-type selective ligands.

Figure 1.

Structures of nantenine and apomorphine and general structures of racemic series A and series B analogs in this study