Studies on the affinity of 6-[(n-(cyclo)aminoalkyl)oxy]-4H-chromen-4-ones for sigma 1/2 receptors

Abstract

Sigma (σ) receptors represent attractive targets for the development of potential agents for the treatment of several disorders, including Alzheimer's disease and neuropathic pain. In the search for multitarget small molecules (MSMs) against such disorders, we have re-discovered chromenones as new affine σ₁/σ₂ ligands. 6-(4-(Piperidin-1-yl)butoxy)-4H-chromen-4-one (7), a previously identified MSM with potent dual-target activities against acetylcholinesterase and monoamine oxidase B, also exhibited σ₁/σ₂ affinity. 6-(3-(Azepan-1-yl)propoxy)-4H-chromen-4-one (20) showed a K_i value for σ₁ of 27.2 nM (selectivity (σ₁/σ₂) = 28), combining the desired σ₁ receptor affinity with a dual inhibitory capacity against both acetyl- and butyrylcholinesterase. 6-((5-Morpholinopentyl)oxy)-4H-chromen-4-one (12) was almost equipotent to S1RA, an established σ₁ receptor antagonist.

Chromenone derivatives were identified as σ₁ and σ₂ receptor ligands (12, 20) and multitarget small molecules (7).

Sigma (σ) receptors represent a unique receptor class involved in a number of biological processes and classified into two subtypes, sigma-1 (σ₁) and sigma-2 (σ₂). Both receptors have been implicated in several neurological and neurodegenerative disorders, such as Alzheimer's disease (AD), as well as cancer.^1,2 Although σ₁/σ₂ receptors represent structurally different proteins, a number of pharmacologically active compounds possess affinity towards both. Accordingly, receptor-subtype selectivity becomes a crucial aspect in the discovery of σ receptor targeting drugs,^1–3 the design of molecular hybrids,⁴ and the development of compounds for radioligand-based molecular imaging techniques.⁵

The σ₁ receptor is a ligand-regulated molecular chaperone involved in Ca²⁺ signalling at the endoplasmic reticulum (ER), the management of diverse types of ion channels, and the modulation of ER and mitochondrial (dis)functions.⁶ The σ₁ receptors are known to play key roles in the central nervous system modulating cognitive capacities such as memory and learning.⁷ These receptors are addressed by trophic factors and by some anti-amnesic drugs acting as σ₁ receptor agonists by increasing cholinergic and glutamatergic activity.^1,6,8 Antagonists of the σ₁ receptor are valuable pharmacological tools to manage neuropathic pain.⁹ They decrease the capsaicin-induced mechanical allodynia, while σ₁ receptor agonists have the opposite effect, which is currently being used for the functional analysis to assign the agonistic or antagonistic profile of ligands.^10,11

With respect to structure–activity relationships and σ_1/2 agonist/antagonist capacity, a large number of ligands bearing a diverse array of structural motifs have been identified.¹ Among them, piperidines,¹² such as the antipsychotic haloperidol (Fig. 1) or (+)-MR200,¹³ ethylenediamine derivatives such as BD1008 (Fig. 1),¹⁴ piperazines, such as CM156,¹⁵ pyrimidines, such as Lan-0101,¹⁶ as well as aminoalkyl-substituted 1,3-dioxanes¹⁷ and aminoalkyloxy-substituted isoxazoles¹⁸ typically act as σ₁ receptor antagonists. Most, but not all,¹⁹ of such ligands share a common structural feature, a basic nitrogen, that is pivotal for their interaction with the σ₁ receptor. S1RA (Fig. 1),^20,21 a highly selective σ₁ receptor antagonist, reached phase II clinical trials for the treatment of neuropathic pain.²²

Fig. 1. Known σ_1/2 receptor modulators, general structures for (aminoalkoxy)chromones and selective receptor ligands I (R³ = Ar, n = 2–9)²⁴ and representatives of chemotype II, i.e. the multitarget small molecule 7 and the potent σ_1/2 receptor modulator 12.

Surprisingly in spite of being a privileged motif in medicinal chemistry,²³ chromones have been scarcely investigated as σ_1/2 receptor modulators. In fact, to the best of our knowledge, only one report of Erickson and co-workers on (aminoalkoxy)chromones (I) (Fig. 1) with potent (16–100 nM) binding affinities has been published.²⁴

In search for multitarget small molecules (MSMs)²⁵ and prompted by the previously described 6-(4-(piperidin-1-yl)butoxy)-4H-chromen-4-one (7),²⁶ a chromenone of type II (Fig. 1) showing potent dual-target activity towards human acetylcholinesterase (hAChE) and human monoamine oxidase B (hMAO-B) (Table S1, ESI†), we initiated the σ_1/2 receptor binding analysis of a recently synthesized chromenone library, designed as a combination of different linker lengths and different terminal amine components.²⁶ We discovered in this study that (i) the MSM 7 exhibited moderate dual σ_1/2 affinity, (ii) 6-(3-(azepan-1-yl)propoxy)-4H-chromen-4-one (20) combined the desired σ₁ receptor affinity with a dual inhibitory capacity against both cholinesterases, and (iii) 6-((5-morpholinopentyl)oxy)-4H-chromen-4-one (12) was almost equipotent to S1RA,²¹ an established σ₁ receptor antagonist.

By using a validated radioligand assay (see ESI,† Table S2), we evaluated the chromenone library of compounds 1–22 (ref. 26) as ligands of the human σ₁ (hσ₁) and rat σ₂ (rσ₂) receptor. Haloperidol, SA4503 (cutamesine), a synthetic σ₁ receptor agonist, and S1RA were included as reference compounds. The results are shown in Table 1. A structure–activity-relationship analysis of the σ_1/2 receptor binding affinity was possible in view of the multitude combinations of the terminal amine motifs and the lengths of the linker (Table 1). These observations could be summarized as follows. The majority of investigated chromenones were potent and selective σ₁ receptor ligands, with affinity values ranging from K_i = 19.6 nM (12) to K_i = 12.1 μM (16). An increased linker length caused, by trend, an improved σ₁ receptor affinity and pentamethylene compounds with n = 4, such as 4, 12, 15, and 18 were the most potent in each corresponding subseries.

Affinity values (K_i, nM) of selected chromenones at the human σ (hσ₁) and rat σ₂ (rσ₂) receptors.

Compd	R	n	hσ1	rσ2	Selectivity hσ1//rσ2
1		1	2870	12 500	4.4
2		2	499	23 900	48
3		3	227	3110	14
4		4	87.0	1450	17
5		1	825 ± 146	9070	11
6		2	65.5 ± 4.5	2890	44
7		3	309 ± 45	1970 ± 1400	6.4
8		4	213	490	2.3
9		1	>10 000	>10 000	n.d.
10		2	245	22 500	92
11		3	23.9 ± 9.7	14 600	610
12		4	19.6 ± 3.2	2630 ± 1200	130
13		1	604	14 200	24
14		2	809	17 100	21
15		4	82.5	4100	50
16		1	12 100	30 000	2.5
17		2	1220	8360	6.9
18		4	218	2500	11
19		1	113 ± 28	1620	14
20		2	27.2 ± 6.8	750 ± 250	28
21		3	104	1090	10
22		4	134	583	4.4
Haloperidol			3.53 ± 0.47	20.5 ± 1.0	5.8
SA4503			3.77 ± 0.11	n.d.a	n.d.
S1RA b			17.0	9300	550

^aNot determined.

^bData from ref. 21.

Hence, in the first subset (ligands 1–4), the affinity gradually increased with the linker length, from n = 1 to 4. Chromenone 4 (n = 4) was the most potent σ₁ receptor ligand containing a pyrrolidine as basic center (K_i = 87.0 nM, selectivity ratio (sel.) (σ₂/σ₁) = 17) and showed modest hAChE (IC₅₀ = 7.20 μM) with no remarkable MAO inhibition (Table S1, ESI†). An attractive MSM was previously identified in this subseries.²⁶ This compound (3; hAChE: IC₅₀ = 3.74 μM; hMAO-B: IC₅₀ = 19.6 μM) has now been discovered as a σ₁ receptor ligand with good affinity (K_i = 227 nM). The incorporation of a (R)-2-methyl group in the pyrrolidine ring of ligands 3 and 4 led to the enantiomerically pure chromenones 21 and 22, with n = 3 and 4, respectively. This simple modification afforded either a 2.2 more (21) or 1.5-fold less (22) potent σ₁ receptor agent (Table 1).

The most potent σ₁ receptor ligand within the piperidine subset (ligands 5–8) was 6 (K_i = 65.5 nM; sel. (σ₂/σ₁) = 44), showing very modest hAChE (IC₅₀ = 15.9 μM), but no noteworthy MAO inhibition. Compound 7 combines pronounced hAChE and hMAO-B inhibition (Table 2; IC₅₀ = 5.58 μM and 7.20 μM, respectively) with σ₁ receptor (K_i = 309 nM) and modest σ₂ receptor (K_i = 1.97 μM) affinity.

Inhibition of hAChE, hBuChE and hMAO-B (IC₅₀) and affinity values (K_i) at the human σ₁ (hσ₁) and rat σ₂ (rσ₂) receptors.

Compd	hAChEa	hBuChEa	hMAO-Ba	hσ1	rσ2
	IC50 ± SE (μM)			K i ± SE (nM)
7	5.58 ± 0.49	87.5 ± 11.8	7.20 ± 0.41	309 ± 45	1970 ± 1400
12	96.2 ± 7.7	n.i.b	v.n.d.c,d	19.6 ± 3.2	2630 ± 1200
20	10.6 ± 2.2	25.0 ± 4.7	v.n.d.c,e	27.2 ± 6.8	750 ± 250

^aData from ref. 26.

^bn.i. (no inhibition) refers to IC₅₀ >100 μM.

^cv.n.d. (values not determined). IC₅₀ values have not been determined.

^dResidual activity of MAO-B @ an inhibitor concentration of 10 μM was 58%.

^eResidual activity of MAO-B @ an inhibitor concentration of 10 μM was 54%.

The morpholine subset (ligands 9–12) deserved particular attention, in view of the structure of S1RA (Fig. 1) and the general pharmacological properties attributed to this moiety.²⁷ Again, chromenone 12 with the longest linker (n = 4) was the most potent σ₁ receptor ligand out of this set (K_i = 19.6 nM; sel. (σ₂/σ₁) = 130), but showed poor cholinesterase and MAO inhibition (Table 2). Chromenone 10 was previously reported not to affect the cholinesterases, but to inhibit both MAO enzymes with IC₅₀ values of less than 2 μM.²⁶ This activity profile was further characterized by the observed affinity for the human σ₁ receptor (K_i = 245 nM).

The most potent σ₁ receptor ligand in the piperazine subset (15; K_i = 82.5 nM; sel. (σ₂/σ₁) = 50) was weaker than the analogous morpholine derivative 12, and in the same range than the pyrrolidine counterpart 4.

The diethylamine-containing ligands 16–18 exhibited weaker σ₁ receptor affinity, e.g. when compared to the corresponding piperazine analogs, indicating the disadvantage of the acyclic amine portion.

Out of the two azepane ligands, 20 with the longer linker performed better and gave a K_i value of 27.2 nM (sel. (σ₂/σ₁) = 28). This compound combined the desired σ₁ receptor affinity with a dual inhibitory capacity against both cholinesterases (Table 2).

In conclusion, the σ_1/2 receptor radioligand binding assay of chromenones has allowed us to identify three highly potent σ₁ receptor ligands, 11, 12 and 20, with K_i values lower than 30 nM. They were selective for the human σ₁ receptor over the rat σ₂ receptor and exhibited selectivity ratios higher than 25. Importantly, our best ligand, chromenone 12 was equipotent to the well-known morpholine-containing ligand S1RA, an advanced agent for possible therapeutic treatment of neuropathic pain.²¹

It is well known that σ₁ receptor antagonists are highly cytotoxic, whereas σ₁ receptor agonists preserve cell survival.²⁸ This feature was employed to address the functional agonist/antagonist profile of selected chromenones. Hence, the cytotoxic activity of ligands 7, 12 and 20 was determined by the MTT assay with the SH-SY5Y neuroblastoma cell line as described.²⁹ The results were compared with those obtained for reference compounds NE-100 and S1RA, accepted σ₁ antagonists, and SA4503, a selective σ₁ agonist. The outcome of the cell viability assay after 48 h compound treatment (Fig. S1, ESI†) indicated that 7 behaved as a σ₁ receptor antagonist, while 12 and 20 acted as σ₁ receptor agonists.

Finally, molecular modelling was carried out to shed light on the binding mode of the highly active compounds 12 (Fig. 2 and 3) and 20 (see ESI†) with the σ₁ receptor protein. A validated three-dimensional model of the σ₁ receptor based on the crystal structure of the human σ₁ receptor in complex with the antagonist PD144418 was used for this analysis.³⁰ The docking process was executed using AutoDock Vina software.³¹ The binding pocket primarily consists of hydrophobic residues of Val84, Trp89, Met93, Leu95, Leu105, Leu182, Phe107, Ile124, Trp164, and Tyr103, and the two acidic residues of Glu172 and Asp126. Receptor flexibility was introduced with respect to the rearrangement of these side chains allowing a ligand to properly adapt to the target.

Fig. 2. Modeled complex of the σ₁ receptor with compound 12 (yellow), showing the key interactions.

Fig. 3. 2D schematic view of the interactions between chromenone 12 and the σ₁ receptor. Intermolecular interactions are indicated with colors: green, van der Waals; orange, salt bridge and attractive charge interactions; pink, π–alkyl contacts.

Compound 12 fully occupies the active site and established a number of interactions with the σ₁ receptor, as shown in Fig. 2 and 3. The morpholine moiety is perfectly encased in the hydrophobic cavity formed by residues, Ser117, Ile124, Phe133, His154, Val162, Trp164 and Tyr103. The linker alkyl chain resides in the hydrophobic region lined by residues Val84, Trp89 Phe107, Tyr120 and Phe184. The protonated nitrogen is engaged in a key bidentate salt bridge with Glu172 and in an attractive charge interaction with Asp126. On the other hand, a network of π–alkyl interactions stabilizes the position of the chromenone moiety in proximity to the residues Met93, Leu95, Leu105, Leu182 and Ala185.

The observed binding of compound 12 in the active site of the σ₁ receptor fulfilled Glennon's pharmacophore model,³² preserving the central protonated group, i.e. the basic morpholine nitrogen, located between the two hydrophobic sites, i.e. the chromenone moiety and the carbon chain of the morpholine ring, at appropriate distances of 8.6 and 2.5 Å, respectively.

In order to evaluate the drug likeness and the potential ability to cross the BBB, chromenones 12 and 20 were scored in silico for their physiochemical and pharmacokinetic parameters (ADME). The results are summarized in Table S2 (ESI†). All calculated descriptors and properties are within the expected thresholds. Chromenones were found to have no Lipinski rule violations,³³ thus proving their drug-likeness properties (see ESI†).

To sum up, in our search for MSMs as possible therapeutics against AD, we have successfully re-discovered and addressed the potential of chromenones as ligands of the σ₁ and σ₂ receptors. Noteworthy, we have identified 6-(4-(piperidin-1-yl)butoxy)-4H-chromen-4-one (7), a previously identified MSM, showing dual-target anti-hAChE and anti-hMAO-B activities,²⁶ as a good σ₁/σ₂ receptor modulator. We identified 6-((5-morpholinopentyl)oxy)-4H-chromen-4-one (12) and 6-(3-(azepan-1-yl)propoxy)-4H-chromen-4-one (20) as potent σ₁ receptor modulators, almost equipotent to S1RA. This promising result may pave the way for future developments on this field in our laboratories.

This work was supported by pre-doctoral FPU grants to D. D. I. from the University of Alcalá and the Spanish Ministry of Science, Innovation and Universities. JMC thanks Eli Lilly OIDD (Open Innovation Drug Discovery) program, and Universidad Camilo José Cela (Grant UCJC 2020-33) for support. W. D. C. thanks Tina Spalholz (HZDR) for supporting the radioligand binding experiments.

Conflicts of interest

There are no conflicts to declare.