Evaluation of structural effects on 5-HT_2A receptor antagonism by aporphines: identification of a new aporphine with 5-HT_2A antagonist activity

Abstract

A set of aporphine analogs related to nantenine was evaluated for antagonist activity at 5-HT_2A and α_1A adrenergic receptors.

With regards to 5-HT_2A receptor antagonism, a C2 allyl group is detrimental to activity. The chiral center of nantenine is not important for 5-HT_2A antagonist activity, however the N6 nitrogen atom is a critical feature for 5-HT_2A antagonism.

Compound 12b was the most potent 5-HT_2A aporphine antagonist identified in this study and has similar potency to previously identified aporphine antagonists 2 and 3. The ring A and N6 modifications examined were detrimental to α_1A antagonism. A slight eutomeric preference for the R enantiomer of nantenine was observed in relation to α_1A antagonism.

The tetracyclic aporphine template is a privileged scaffold that is endowed with several biological activities.^1–8 As central nervous system (CNS) receptor ligands, aporphines have been found to possess high affinity for a number of dopamine receptors (predominantly D₁ and D₂),^9–12 serotonin (5-HT) receptors^13–15 and α-adrenergic receptors.^6,16 Furthermore, aporphines are known with both agonist and antagonist activity at neuroreceptor sites. The continued exploration of the aporphine template over the last few decades has been driven to some extent by the opportunity/promise for discovery of new ligands with high potency and selectivity for subtypes of the above-mentioned receptors. Such molecules will continue to supplement our toolbox of CNS receptor ligands that will be useful as novel biological probes, as new imaging agents and as leads for drug discovery efforts relevant to psychiatric disorders and drug abuse.

We are primarily interested in evaluating the potential of aporphines as ligands for 5-HT_2A receptors. 5-HT_2A receptors are implicated in several neuropsychiatric maladies including schizophrenia, depression, anxiety and insomnia.^17,18 5-HT_2A receptors are also involved in the actions of some stimulant drugs as recent reports have revealed.^19–21

Several potent 5-HT_2A receptor antagonists are known; in particular compounds with a mixed D₂/5-HT_2A antagonist profile (eg risperidone, clozapine) are quite prominent and are used clinically to manage schizophrenic symptoms.^22–24 However, there are no highly selective 5-HT_2A antagonists (>100-fold selectivity vs all other common neuroreceptor targets) clinically available. Nevertheless, such promising compounds (eg eplivanserin) have recently been or are currently being investigated in clinical trials as anti-insomnia medications. Thus, the identification of new highly selective and therapeutically useful 5-HT_2A receptor antagonists is still of topical interest.

Our research team has engaged structure-activity relationship (SAR) studies on the aporphine alkaloid nantenine (1, Figure 1) and have identified a number of new nantenine analogs with antagonist activity at the 5-HT_2A receptor.^25–28 We were guided by previous SAR studies on nantenine for the present study. Nantenine itself is a high affinity α_1A adrenergic receptor antagonist with moderate 5-HT_2A receptor antagonist potency (see Table 1). Compounds 2 and 3 (Figure 1) are two of the most potent 5-HT_2A antagonists we have obtained to date. These compounds lack affinity for the α_1A adrenergic receptor. Our prior investigations have mainly focused on the ring A portion of nantenine and in general have indicated a reasonable degree of tolerance for other types and patterns of substitution in the A ring in obtaining high 5-HT_2A potency and selectivity vs the α_1A receptor. However, the identity and optimal placement of substituents requires further research for maximal potency and selectivity.

Figure 1.

Structures of nantenine (1), compound 2 and compound 3

Table 1.

K_e values for analogs at 5-HT_2A and α_1A receptors

	Compd.	R1	R2	X	Ke ± SEMa (nM)		Selectivity 5-HT2A/α1A
					5-HT2A	α1A
	12a	H	H	NMe	>3000	2950 ± 457	<1.01
	12b	allyl	H	NMe	47 ± 5	744 ± 74	0.06
	14	H	allyl	NMe	>3000	>3000	-
	15a	Me	allyl	NMe	485 ± 123	566 ± 112	0.85
	15b	allyl	allyl	NMe	1374 ± 405	>3000	<0.45
	15c	cyclopropylmethyl	allyl	NMe	963 ± 103	>3000	<0.32
	16a	Me	OMe	O	>3000	>3000	-
	16b	allyl	OMe	O	>3000	>3000	-
	(R)-1	Me	OMe	NMe	946 ± 61	70 ± 10	13.5
	(S)-1	Me	OMe	NMe	657 ± 89	196 ± 3	3.4
	±-(1)b	Me	OMe	NMe	850 ± 6	36 ± 7	23.6
	2c	allyl	OMe	NMe	70 ± 15	>10000	<0.007
	3d	cyclopropylmethyl	OMe	NMe	68 ± 8	>10000	<0.007
	prazosin	-	-	-	-	1.1 ± 0.4	-
	ketanserinb,e	-	-	-	32	-	-

^aValues represent mean ± SEM for at least three independent experiments;

^bK_i, Data from ref. 28;

^cData from ref. 27;

^dData from ref. 26;

^eIC₅₀ determined in the presence of 5-HT EC₈₀

To continue to expand our understanding of the structural tolerance of aporphines as 5-HT_2A receptor antagonists as well as selectivity vs the α_1A receptor, we have synthesized and evaluated a new set of aporphine analogs. These compounds were prepared in order to probe three regions in the northern portion of the aporphine template, namely: ring A, the chirality center and the nitrogen atom. As alluded to earlier, our previous studies have indicated that substituents on the ring A moiety are important in controlling the 5-HT_2A receptor antagonist activity and selectivity. However, we have not investigated the effect of the chiral center on 5-HT_2A antagonism before. This is an important task especially since the existing literature suggests that aporphines exhibit a stereochemical preference for activation of dopamine D₁ and D₂^29,30 receptors as well as the related 5-HT_1A receptor³¹ R enantiomers being predominantly agonists and S enantiomers being predominantly antagonists in both cases. With regards to the effect of the N6 nitrogen atom, our prior studies suggest that a basic nitrogen atom is required since the N-acetamide and N-methylsulfonamide derivatives were devoid of activity.²⁵ We sought herein to obtain experimental proof of the absolute requirement for a nitrogen atom for antagonist activity.

Along those lines we have: 1) synthesized and evaluated new ring A analogs containing features of compounds 1, 2 or 3; 2) evaluated nantenine enantiomers - in order to begin to probe the effect of the chirality center of aporphines on receptor antagonism ; and 3) synthesized and evaluated compounds that possess a nitrogen - oxygen isosteric replacement - to investigate the importance of the nitrogen atom on receptor antagonism. Details of these studies are described henceforth.

We were interested in examining the extent to which an allyl group would be accommodated at the C2 position since an allyloxy group seems to be beneficial for antagonism (see data for compound 2, Table 1). The synthetic practicability of obtaining this structural feature via a Claisen rearrangement (a reaction rarely employed with aporphines) supported this impetus.³²

To obtain the required ring A analogs Scheme 1 was engaged. Commercially available amine 4 was coupled to bromoacid 5 to give amide 6. Bischler-Napieralski cyclization of 6 was followed immediately by reduction of the dihydroisoquinoline thus formed to the secondary amine 7. Protection of the amine as the N-ethyl carbamate gave compound 8. Microwave-assisted direct arylation³³ on 8 afforded compound 9. The benzyl ether 9 was deprotected revealing the phenol functionality in the key intermediate 10. Reduction of 10 with lithium aluminium hydride (LAH) gave compound 12a. Compound 12b was prepared from 10 via allylation to afford compound 11 and subsequent LAH reduction. Claisen rearrangement of the allyl ether 11 provided compound 13. Reduction of 13 gave the phenol analog 14. Analogs 15a–15c were prepared from 13 in two steps via etherification and reduction as shown. Nantenine enantiomers (R)-1 and (S)-1 were prepared by resolution of racemic nantenine as described previously.²⁸ The isochroman analogs 16a and 16b were prepared via a sequence involving oxa Pictet-Spengler cyclization and direct arylation as presented in a recent report.³⁴

Scheme 1.

Synthesis of ring A analogs

Reagents and conditions: (a) 1,1′-carbonyldiimidazole (CDI), THF, 0 °C - rt, 5 h, 80% ; (b) trifluoromethanesulfonic acid, pyridine, DCM, 0 °C - rt, 4 h; (c) NaBH₄, MeOH, 0 °C, 2 h, 88% over two steps; (d) Ethyl chloroformate, K₂CO₃, DCM, rt, 3 h, 85% ; (e) Pd(OAc)₂, di-tert-butyl(methyl)phosphonium tetrafluoroborate, K₂CO₃, (CH₃)₃CCOOH, DMSO, 135 °C, microwaves, 6 min, 50% ; (f) H₂/Pd, rt, 8 h, 95% ; (g) alkyl bromide, KI, K₂CO₃, acetone, 70 °C, 6 h, 60–70% ; (h) LAH, THF, 0 °C, 10 h, 50–60% ; (i) N,N-diethylaniline, 215 °C, microwaves, 6 min, 90%

All analogs were screened at 10 μM in multi-well format for intrinsic (agonist) and antagonist activity at the human 5-HT_2A receptor using Fluorescence Imaging Plate Reader (FLIPR) -based (Molecular devices, Sunnydale, CA) functional assays that detect receptor-mediated mobilization of internal calcium with a calcium sensitive fluorescent dye as reported previously.^25–27 A similar set of assays was performed for the α_1A - adrenergic receptor. Data from these evaluations are presented in Table 1.

As shown in Table 1, compound 12a lacked any appreciable activity at either receptor. The placement of an allyloxysubstitutent at C2 (ie compound 12b) resulted in a significant increase in 5-HT_2A antagonist activity (> 60-fold as compared to the parent phenol 12a). Antagonism of the α_1A receptor also increased although the magnitude of this increase was less (4-fold increase as compared to 12b). However, when 12b is compared to nantenine (1), a significant decrease in α_1A antagonist activity was seen. The above data tends to suggest that the C2 methoxyl substituent of nantenine is not required for 5-HT_2A antagonist activity.

Compound 14 lacked antagonist activity for both receptors indicating that an allyl substituent at C2 is not well tolerated at either receptor. A comparison of 15a with 1 reveals a slight improvement in 5-HT_2A antagonism but a decrease in α_1A antagonist activity upon replacement of the C2 methoxyl group with an allyl substituent. The antagonist activity of 15a was higher at both receptors than compound 14 which is indicative of a greater tolerance for an alkoxy substituent than a phenol at C1. Compound 15b had a diminished affinity at both receptors as compared to 15a. When 15b is compared to compound 12b, a significant decrease in antagonist activity of 15b at both receptors manifests. This suggests that an allyl substituent at C2 is not well accommodated at either receptor. Compound 15c had activity and selectivity that was similar to 15b. If a comparison of the 15b/15c pair is made with the 2/3 pair of compounds it may be surmised that C1 allyloxy and C1 cyclopropylmethyloxy groups endow the aporphine template with very similar 5-HT_2A antagonist potency irrespective of the identity of the C2 substituent. That is, it appears that the allyl and cyclopropylmethyl functionalities are bioisosteric with respect to 5-HT_2A receptor antagonism. From our previous studies, an allyloxy or cyclopropylmethyloxy substituent (ie compounds 2 and 3 respectively) imparted high 5-HT_2A antagonist activity and selectivity to the nantenine template. The analysis of compounds 15a, 15b and 15c showed a reversal in this trend and again points to a considerable lack of tolerance for a C2 allyl group at the 5-HT_2A receptor.

Both compounds 16a and 16b (that lack the N6 moiety) were devoid of antagonist activity. This supports previous SAR evidence that a basic N6 atom is critical for affinity to both receptors. This is also in line with previous molecular docking studies which suggest that the protonated N6 atom is involved in a H-bonding interaction with an aspartate residue in the 5-HT_2A receptor binding pocket.²⁵

Evaluation of (R)-1 and (S)-1 indicates that the chiral center of nantenine is not critical for 5-HT_2A antagonism although the (S) enantiomer is slightly more potent. Interestingly, there seems to be a reversal of this trend at the α_1A receptor; the (R)-enantiomer seems to be slightly more potent than the (S)-enantiomer at the α_1A receptor (approximately 3-fold).

In summary, this study has revealed some useful qualitative information concerning the antagonism of aporphines at the 5-HT_2A receptor. The data suggest that the C2 position is not tolerant of an allyl moiety. However, a C1 allyloxy substituent is well tolerated when the C2 substituent is hydrogen implying that the C2 methoxyl group of nantenine is not required for high 5-HT_2A antagonist potency. This modification also improves selectivity vs the α_1A receptor (though this selectivity is moderate as compared to that seen in 2 and 3). Of note, the most potent 5-HT_2A aporphine antagonist identified in this study was compound 12b which rivals 2 and 3 in terms of 5-HT_2A antagonist potency.

The chiral center of nantenine does not engender any significant preference for either enantiomer towards 5-HT_2A antagonism. Somewhat unsurprisingly, the N6 nitrogen is critical for antagonist activity of nantenine analogs at both receptors.

The evaluation of this set of compounds has further expanded our fundamental knowledge concerning the viability of the aporphine template for development as selective 5-HT_2A receptor antagonists. For certain, evaluation of larger series of analogs will enable a better understanding of the extent to which the SAR information extracted up to this point may be generalized. Compound 12b identified herein as well as compounds 2 and 3 identified earlier, are useful starting points for further SAR exploration and optimization studies. We are continuing in this vein and will report our findings in due course.