From 2-Triethylammonium Ethyl Ether of 4-Stilbenol (MG624) to Selective Small-Molecule Antagonists of Human α9α10 Nicotinic Receptor by Modifications at the Ammonium Ethyl Residue

Nicotinic acetylcholine receptors containing α9 subunits (α9*-nAChRs) are potential druggable targets arousing great interest for pain treatment alternative to opioids. Nonpeptidic small molecules selectively acting as α9*-nAChRs antagonists still remain an unattained goal. Here, through modifications of the cationic head and the ethylene linker, we have converted the 2-triethylammonium ethyl ether of 4-stilbenol (MG624), a well-known α7- and α9*-nAChRs antagonist, into some selective antagonists of human α9*-nAChR. Among these, the compound with cyclohexyldimethylammonium head (7) stands out for having no α7-nAChR agonist or antagonist effect along with very low affinity at both α7- and α3β4-nAChRs. At supra-micromolar concentrations, 7 and the other selective α9* antagonists behaved as partial agonists at α9*-nAChRs with a very brief response, followed by rebound current once the application is stopped and the channel is disengaged. The small or null postapplication activity of ACh seems to be related to the slow recovery of the rebound current.


■ INTRODUCTION
Mammalian nonmuscle nicotinic acetylcholine receptors (nAChRs) form as pentameric assemblies of protein subunits (named α2-α7, α9, α10, and β2−β4), with particular subunit combinations termed "subtypes". 1 The majority of nonmuscle nAChR subtypes contain at least one α and one β subunit (i.e., are heteromeric), and agonists bind sites formed at interfaces between the (+)-faces of α-subunits and (−)-faces of βsubunits. 1 However, α7 and α9 subunits appear to be uniquely capable of forming functional homomeric nAChR subtypes, in which a single subunit is capable of providing both the (+)and (−)-faces needed to bind agonists. 1 A further similarity is that compounds such as MLA and α-bungarotoxin (α-Bgtx), originally thought to selectively antagonize α7-nAChR, also are similarly potent antagonists of α9-nAChR homopentamers. 2 Heteropentameric α9α10-nAChRs are often formed and show very similar antagonist pharmacological profiles to α7and α9-nAChR homopentamers. 3 Nevertheless, some important features distinguish α7from α9*-nAChRs (where * denotes the possible presence of additional nAChR subunits, in this case α10). 4 First, while both are expressed in both neuronal and non-neuronal cells, the expression of α9*-nAChRs is restricted to the periphery, whereas α7-nAChRs expression is widely distributed across CNS and peripheral locations. 5 Second, nicotine is an α7-nAChR agonist, whereas it is an antagonist of α9*-nAChRs. 3 Third, α9*-nAChRs exhibit a Ca 2+ permeability 2-fold higher than α7-nAChRs, with an outstanding fractional Ca 2+ current of about 22%. 6 Forth, the physiological effects of α7vs α9*-nAChR antagonists can be opposing. For example, α9*-nAChR antagonists have been explored as novel, nonopioid, analgesics to treat neuropathic and inflammatory pain. 7 In contrast, the agonism of peripheral α7-nAChR produces analgesic and anti-inflammatory effects. 7 This property recently has been extended to α7-nAChR "silent agonists" (ligands that produce very little or no ionotropic agonist activity (if not coapplied with a positive allosteric modulator; PAM), but that instead activate metabotropic signaling pathways through α7-nAChR). 7,8 However, opposing effects are not the rule. For example, α7and α9*-nAChR are overexpressed in multiple tumor types, where activation of either promotes tumor cell growth, and inhibition has antiproliferative effects. 9−11 For instance, analogues of the 2-triethylammonium ethyl ether of 4-stilbenol (1, MG624) with a lengthened alkylene linker between charged nitrogen and ethereal oxygen (2 and 3) (Chart 1) display potent antiadenocarcinoma and antiglioblastoma activity, paralleled by increased α7and α9α10-nAChR antagonism. 12 More-recent results have suggested dimerization of α-conotoxins capable of inhibiting α9α10-nAChRs and, with lower potency, α7-nAChRs as a strategy to obtain more potent α7 and α9α10 dual inhibitors, which could be useful probes and/or drug leads to investigate the role of these receptors in tumorigenesis. 13 Indeed, α7and α9*-nAChR involvement in tumor cell proliferation is consistently supported across a range of published studies, 9,10 although it is important to note that additional non-nicotinic mechanisms could also contribute to the antitumor activity of some nicotinic antagonists, as demonstrated by a very recent pharmacological investigation on the above stilbenol derivatives 2 and 3. 10 Within this context, antagonists selectively targeting α9α10-nAChRs, indispensable for dissecting function of α9α10vs α7-nAChRs, are very few and those eligible for the development of druglike leads are still lacking. Indeed, when excluding some potent and selective bis-, tris-, and tetrakis-azaaromatic quaternary ammonium analogues 14 for poor drug-likeness and, for similar reasons, α-conotoxins and related peptides, the literature offers no examples of small molecules that can be considered promising hits for the development of selective α9α10 antagonists.
Our recent investigations proved that pretreatment of α7and α9α10-nAChRs with 1, first reported as a selective α7 antagonist in a pioneering study dating from 1998, 15 potently inhibits the activation of both receptor subtypes by subsequently applied ACh and that elongation of the alkylene chain between the ammonium head and the oxygen by addition of further methylene units (compounds 2 and 3) results in more potent antagonism. 9,12 These observations indicate that 1 is a good hit, susceptible to useful changes of pharmacological profile by structural modification. Accordingly, we have extended our structure− activity relationship (SAR) studies in the hope of differentiating its activities at α7and α9α10-nAChR, from each other. We considered a wide number of modifications of the ammonium head and of the linker on one side, and of alternative scaffolds in replacement of the stilbene substructure on the other. Here, we report the synthesis and binding affinities at the α4β2, α3β4, and α7 subtypes of the analogues of 1 modified at the ammonium head (compounds 4−17) and at the linker (compounds 18−27) (Chart 2). The ammonium head was modified by the increase of its steric bulk through gradual replacement of methyls with ethyls (compounds 4−6), inclusion of two of the three alkyls in increasingly larger cycles (compounds 13−16), replacement of one or two alkyls with cyclic substituents (compounds 7−11), change in shape and positive charge distribution (compounds 12 and 17). The linker was modified maintaining the interposition of two carbons between oxygen and nitrogen, but enclosing, partially or totally, this fragment into a four or five-membered nitrogen heterocycle (compounds 18 −22, 24, and 25) or rigidifying it into cyclopropane (compounds 26 and 27). In compound 23, three carbons were interposed, but conformationally constrained in the piperidine cycle. A large selection of these analogues, including compounds proved to have higher or remarkably lower α7 affinity than 1, was then screened for α7and α9α10-nAChR antagonist activity. Lastly, the four compounds showing the best profiles in terms of potency and subtype selectivity were further characterized to better understand the mechanism by which they exert antagonism. ■ RESULTS Chemistry. Compounds 4−17 were synthesized according to Scheme 1. The commercially available building block (E)-4hydroxystilbene was alkylated using dibromoethane under basic conditions affording the intermediate 28. Upon conversion by Finkelstein reaction to the iodo-derivative 29, treatment with a selected variety of secondary amines provided the tertiary amines 30−34 and 37−40, which were quaternarized by treatment with either methyl or ethyl iodide or benzyl bromide to the corresponding quaternary ammonium iodide or bromide salts (4−10, 13−16). Likewise, the reaction between the primary 1-adamantylamine and the intermediate 29 afforded the secondary amine 35, which was methylated to the tertiary amine 36 by reductive amination, further methylated to 11 by treatment with methyl iodide. Similarly, quinuclidine and pyridine were coupled with 29 to obtain the quaternary ammonium iodide salts 12 and 17. . Additionally, (R)-42 was also boc-deprotected to afford the secondary amine (R)-51, that was further ethylated to (R)-21.
Biology. Binding Studies. We tested the affinity of all of the synthesized compounds for human α7-nAChR using competition binding assays. A wide selection of the synthesized compounds (primarily those with high affinity for α7-nAChR) were also assessed for competitive binding affinity at human α4β2and α3β4-nAChR (Table 1). SH-EP1 cells stably transfected with α3β4-nAChR were those described in an earlier publication. 16 HEK 293 cells stably transfected with α4β2-nAChR were a generous gift from Lindstrom, 17 whereas the α7 subtype was transiently expressed in SH-SY5Y cells as previously described. 12 Radiolabeling of α7 nAChRs was performed using [ 125 I]-α-Bgtx at the saturating concentration of 2−3 nM, while for labeling of the two heteromeric subtypes, we used 0.1 nM [ 3 H]-epibatidine for the α4β2 nAChR subtype and 0.25 nM [ 3 H]-epibatidine for the α3β4 nAChR.
On the other hand, affinity determination at α4β2-nAChR was limited to most of the compounds with high α7 affinity and, for completeness of SAR analysis, to the trans cyclopropane (±)-26. All of the tested compounds showed low α4β2 affinity (K i > 1 μM).
In Vitro Functional Activity on α7 and α9α10-nAChR Subtypes. We have previously characterized 1 as a selective α7-nAChR antagonist with an IC 50 of 109 nM at oocyteexpressed chick α7-nAChR. This value was determined by preapplication of increasing concentrations of 1 for 30 s, before the application of ACh (100 μM). 15 Oocytes expressing α7-nAChRs did not show a detectable response to 1 applied alone at 10 nM−1 μM concentrations. Recently, we have reported that 1 reduces the ACh activation of both human α7and α9α10-nAChRs expressed in oocytes with IC 50 values of 41 and 10 nM, respectively. 12 The experiments were performed applying 1-s pulses of 10 μM ACh (α9α10-nAChR) or 200 μM ACh (α7-nAChR) at regular time intervals to oocytes perfused or bath-applied with 1 at varying concentrations. 18 Of the compounds reported in this manuscript, 12 (6, 7, 12−15, 18, (S)-19, (R)-19, (±)-22, 23, and (±)-24) were selected to test their ability to inhibit currents induced by coapplied 1 mM ACh in Xenopus laevis oocytes heterologously expressing human α7and α9α10-nAChRs subtypes. The selection was done so as to include compounds representative of structural modifications both at the ammonium head (compounds 6, 7, 12−15) and at the linker (compounds 18, (S)-19, (R)-19, (±)-22, 23, and (±)-24) and of both high (compounds 12−15, 18, (S)-19, (R)-19) and moderate or modest (compounds 6, 7, (±)-22, 23, and (±)-24) α7 affinity. The equipment and technique used were essentially identical to those previously described for α7-nAChR expressed from unlinked subunits, 19 except that α9α10-nAChRs were also expressed (also from unlinked subunit constructs, using a 9:1   Figure 1. Experimental details are provided in the caption of Figure 1 and in the Experimental Section. Values in parentheses represent the 95% confidence interval of the mean value. "NA" = not applicable, and denotes instances where no inhibition of agonist-induced function was observed, even in the presence of 100 μM test compound. ratio of α9 to α10 cRNAs). The magnitude of the expression of α9α10-nAChR function was found to be highly dependent on the ratio of subunit cRNAs injected into the oocytes, and the use of a 9:1 ratio of α9 to α10 cRNAs was chosen throughout this study because it produced the most function. A previous study indicates that these conditions likely produce α9α10-nAChR with a mix of both (α9) 2 (α10) 3 and (α9) 3 (α10) 2 stoichiometries. 20 The resulting IC 50 values are reported in Table 2, together with those of the lead 1 for comparison. The corresponding concentration−response curves are shown in Figure 1. Eight of the tested compounds, namely, 1, 12, 13, 14, 15, 18, (S)-19, and (R)-19, inhibited ACh-induced currents at both the subtypes (although with higher potency at the α7-nAChR subtype), whereas the other five compounds 6, 7, (±)-22, 23, and (±)-24 had inhibitory effects on ACh-induced function at the α9α10-nAChR, but had no effect at the α7-subtype.
Biphasic inhibition of α9α10-nAChR function was not shown by any of the compounds. This strongly suggests that none of these compounds distinguish between alternate α9α10-nAChR stoichiometries. As can be seen, the observed inhibition of ACh-induced function was not always complete due either to reaching a plateau of inhibition or to the use of test concentrations that were too low to obtain maximal inhibition. At the α9α10-nAChR, 100% inhibition was observed for compounds 7, 12, 14, 15, and 23 and it was almost complete for (±)-22 and (±)-24, whereas it was near 80% for 1 and, notably, largely incomplete for 6, 13, 18, (S)-19, and (R)-19, which are characterized by a smaller ammonium head or a more rigidified linker. On the other hand, 100% inhibition at the α7-nAChR was never observed.
Compounds 7, (±)-22, 23, and (±)24, which showed no inhibition effect at the α7-nAChR and complete or nearly complete inhibition of ACh-induced responses at α9α10- Figure 1. Antagonist concentration−response profiles for α7or α9α10-nAChRs. Oocytes were injected with mRNA encoding human α7-nAChR subunits (together with NACHO to increase functional expression; •), or human α9and α10-nAChR subunits (9:1 ratio; ○). At 1 week after injection, the function of the corresponding nAChR populations was assessed using two-electrode voltage clamp electrophysiology. Before antagonists were applied to each oocyte, a train of five ACh stimulations was applied to ensure that stable function could be observed and to provide a positive control (1 s duration with 60 s wash periods between applications). Test compounds were co-applied with 1 mM ACh (1 s duration, 60 s wash periods between applications), starting at the lowest concentrations shown and increasing to 100 μM test compound in half-log increments. Magnitudes of responses in the presence of test compounds were normalized to the mean magnitude of the preceding positive control responses. Data points represent the mean ± standard error of the mean (S.E.M.) of 5−6 responses, recorded from individual oocytes�note that error bars are in some cases smaller than the associated points. For compounds 6, 7, (±)-22, 23, and (±)-24, application at α7-nAChR had no effect on ACh-induced function; corresponding data are not shown for the sake of clarity. nAChR, were tested for intrinsic agonist activity at the two nAChR subtypes. First, repeated control stimulations with a maximally effective dose of ACh (1 mM) were applied to establish the magnitude and stability of ACh control responses. Next, the test compounds were applied alone at 100 μM to oocytes expressing α9α10-nAChR (corresponding to the maximum concentration co-applied with ACh in Figure 1), or 10 μM to oocytes expressing α7-nAChR. When applied alone at a single, high, concentration the four compounds (7, (±)-22, 23, and (±)24) behaved as partial agonists at α9α10-nAChRs (20−60% efficacy compared to ACh control) ( Figure  2). In contrast, intrinsic agonist activity of compounds 7, 23, and (±)-24 at α7-nAChRs was much lower (23 and (±)- 24) compared to that at the α9α10-nAChR or zero (7) (Figure 2). It is to be noted that 1 has been previously reported as a partial agonist at human α7-nAChR (40% of the response of 200 μM ACh) at high concentrations (100 μM). 10 Interestingly, responses induced by the test compounds at α9α10-nAChR were always shorter in duration than those induced by preceding ACh control applications. Further, responses to the test compounds were followed by rebound currents that were longer in duration than either the initial responses to the test compound, or the ACh control responses. In addition, responses to a further control application of ACh (following the application of the test compounds to each oocyte) produced a response much smaller in amplitude than the initial ACh control applications. An example trace depicting such behavior at α9α10-nAChR is shown in Figure  3A for compound 23. In Figure 3B, the magnitude of the poststimulation rebound currents of 7, (±)-22, (±)-24, and 23 at α9α10-nAChRs is represented, relative to that of the preceding ACh-induced control responses (note that no rebound currents were observed at α7-nAChR following application of 7, (±)-24, and 23). Figure 3C shows the residual activities induced by the final ACh (1 mM) control stimulation, following the test application of 7, (±)-22, (±)-24, or 23 at α9α10-nAChRs, and of 7, (±)-24, and 23 at α7-nAChRs. As can be seen, consistent with a profile of no efficacy at α7-nAChR, 7, (±)-24, and 23 produced little block of subsequent ACh-induced function at this subtype, whereas the same compounds significantly blocked subsequent AChinduced function at α9α10-nAChRs. Such a behavior is most evident for 7, which induced a profound inhibition of the subsequent α9α10-nAChR response to ACh (1 mM) while showing essentially no effect on responses of α7-nAChRs (no intrinsic activity, no poststimulation rebound current, and very little decrease of ACh control response following application of 7).

■ DISCUSSION
As in our previous investigation on onium-alkyloxy-stilbenebased compounds, the pharmacological characterization of this new series of stilbenol ammonium alkyl ethers, formally derived from the lead 1 by modification of the cationic head (4−17) or rigidification of the O−N linker (18−27), started from the evaluation of the α7-nAChR binding affinity by α-Bgtx displacement. This is indicative of competition for the same ACh orthosteric binding sites of this receptor and, in the development of our investigation, a useful guiding criterion in selecting candidates for functional study at both α7and α9α10-nAChR.
As for cationic head modifications (compounds 4−17), some SARs can be established by comparison with 1 (104 nM K i ). The increase of steric bulk of the ammonium head by the gradual replacement of methyls with ethyls (in order, compounds 6, 5, 4, and 1) or by inclusion of two of the three alkyls in increasingly larger cycles (in order, compounds 13, 14, 15, and 16) enhances and then decreases the α7 affinity, reaching a maximum in the two series for 4 (34 nM K i ) and 14 (94 nM K i ). Compound 12, with a quinuclidinium head, formally comparable to a diethylpropylammonium head, but downsized by conformational restrainment, has a high affinity (37 nM K i ), similar to the diethylmethyl ammonium analogue 4 (34 nM K i ). Conversely, cyclic substituent in place of one alkyl (compounds 7, 9, and 11) is detrimental (1.6−3.3 μM K i ) and replacement of two alkyls with cycles (compounds 8 and 10) is not tolerated (>10 μM K i ), as well as the radical change in shape and positive charge distribution resulting from replacement with a pyridinium head (compound 17; 5.9 μM K i ). As for the selectivity over the ganglionic α3β4-nAChR subtype, which is a critical issue in the development of ligands with high α7 affinity, the present cationic head modifications do not increase the α7over α3β4-nAChR selectivity of the lead 1. They maintain it or, in some cases, reduce or even invert it.
Modifications of the linker (compounds 18−27) have also resulted in some compounds having similar or higher α7-nAChR affinity compared with 1. This is the case of the compounds that maintain the two-carbon distance between oxygen and nitrogen enclosing, partially or totally, this fragment into a four or five-membered nitrogen heterocycle with two methyls quaternarizing the nitrogen atom (compounds 18, 19, and 25). Notably, compared with 1, Figure 2. Intrinsic agonist activity of selected compounds at human α7and α9α10-nAChRs. The ability of compounds of most interest (see text for details) to activate α9α10-nAChRs was assessed using two-electrode voltage clamp electrophysiology. As for antagonist concentration−response profiles, oocytes were first assessed using a set of five ACh-evoked applications (positive control; 1 s application, 60 s wash periods between applications). This was followed 60 s later by a 1 s application of the test compound at 100 μM (corresponding to the highest concentration used in the earlier antagonist concentration−response experiments). For each individual oocyte, the magnitude of the response evoked by the test compound was normalized to the mean of the magnitudes of the positive control responses. Data were collected from three individual oocytes, for each test compound, at each nAChR subtype and are presented as mean ± S.E.M. (histograms), with each individual response additionally shown as an individual point. compounds 18 (79 nM K i ) and especially (R)-19 (23 nM K i ) display not only higher α7-nAChR affinity, but also a remarkably increased selectivity over the α3β4-nAChR subtype. This is in sharp contrast to what we observed for the two most ameliorative modifications, in terms of α7 affinity, of the cationic head which either did not increase (compound 4) or nullified (compound 12) the selectivity over the α3β4-nAChR subtype. On the other hand, as exemplified by the just mentioned 12, 18, 19, and 25, selectivity over the α4β2-nAChR subtype seems much less critical for these new stilbenol ammonium alkyl ethers as for those previously reported. 12 As previously explained, in vitro functional activity at the α7 and α9α10-nAChRs was determined for 1 and 12 analogues, selected, among the initially tested twenty-seven, for structural representativeness and not a priori excluding compounds with moderate or modest α7 affinity. Given these selection criteria and the restricted number of compounds, a SAR analysis is arduous. Nevertheless, one cannot but notice two things. First, all 13 compounds antagonize ACh activity at α9α10-nAChR, but only some of them also at the α7-nAChR. Indeed, too small or too bulky cationic heads (6 and 7) and inclusion of the two-carbon linker into larger heterocycles than pyrrolidine (22, 23, and 24) lead to complete loss of α7-nAChR antagonism. Second, 100% inhibition of ACh-induced function at the α7-nAChR was never observed, whereas complete or nearly complete inhibition of the α9α10-nAChR to ACh was produced by most of the 13 tested compounds. Incomplete inhibition of the α7-ACh response could be due to partial agonism at high supra-micromolar concentrations of these compounds, as recently shown for 1. This will be a matter for future investigation and discussion on analogues of 1 modified at the stilbene fragment, some of which resulted in selective and complete inhibition of ACh-induced function at the α7-nAChR. Here, we focused on compounds exhibiting only α9α10-nAChR antagonism and with complete or nearly complete inhibition of ACh response, namely, 7, (±)-22, 23, and (±)-24. The IC 50 values of these four compounds range between 5 and 17 μM. We were interested in studying their behavior at both α7and α9α10-nAChRs at high concentrations, those producing maximal inhibition of ACh responses  Figure 2), followed by a "rebound" current that appears after test compound application is stopped, and diminished ACh response following application of the test compound. Compound 23 was chosen for this illustration since it produces the largest-magnitude rebound currents. (B) Magnitudes of rebound currents recorded from α9α10and α7-nAChRs, normalized to the mean magnitude of the preceding control ACh traces (1 mM, 1 s application times). (C) Magnitudes of ACh control stimulation (1 mM, 1 s) applied 60 s following test compound applications. For (B) and (C), data were collected from three individual oocytes, for each test compound, at each nAChR subtype and are presented as mean ± S.E.M. (histograms), with individual responses in each case also shown as individual points.
(10 and 100 μM, respectively), and in the absence of ACh to better understand their mechanisms of action.
These further insights revealed common features of the above compounds, but with diversified profiles. Applied at high concentrations, all four of them have significant intrinsic activity at α9α10-nAChRs (20−60% efficacy compared to 1 mM ACh) and very low α7-nAChR activity, which becomes null in the case of 7 (which is also the compound with the lowest α7-nAChR binding affinity; 2730 nM K i ). The very short duration of the α9α10-nAChR response, visible in the trace of 23 shown in Figure 3A, reflects an abrupt truncation of the induced passage of current through the channel most likely due to a rapid occupation and block of the open channel by the test compound after ligand-induced channel opening. Openchannel block would also explain the following occurrence of the rebound currents, which have a relatively long and variable duration due to gradual increase and subsequent slow baseline recovery. Once the compound engages the open channel, this would be held in an open position with no current passage, a mechanism demonstrated in a classic single-channel publication. 21 Once compound application is stopped, the compound will start to wash out allowing current to flow again through the channel and then to slowly end when the channel closes by itself. Consistent with such an explanation, rebound currents were not observed at the α7-nAChR, which are not activated by these compounds.
A third consideration, consistent with those previously discussed concerning intrinsic activity and rebound current, can be advanced about residual ACh function, which was smaller in amplitude than the ACh control traces and after application of the test compounds alone ( Figure 3C). Inspection of the traces of the compounds tested at α9α10-nAChR shows that the slower the recovery of the rebound current at the α9α10-nAChR, that is to say, the slower the disengagement of the test compounds from the channel, the smaller the residual ACh activity is. Indeed, plotting the amplitude of residual ACh-induced current, as a percent of ACh control, against the percentage recovery of the preapplication baseline just before the final ACh stimulation resulted in a good linear correlation. This would confirm that slow and incomplete recovery of the rebound current before the final ACh stimulation is applied results in extensive suppression of the final ACh peak. Again, compound 7 stands out for the lowest rebound current and the most extensive block of subsequent ACh-induced function at α9α10-nAChR and, on the other hand, for the minimum decrease of poststimulation residual ACh activity at α7-nAChR.
As for the selectivity against the α3β4 nAChR subtype, this is, in our experience, a crucial issue for the nicotinoids we have so far developed, namely, for the stilbenol ammonium ethyl ethers as α7-α9α10 ligands 12,15 and for pyrrolidinyl-benzodioxanes as α4β2 ligands. 22,23 Otherwise, the selectivity between these two latter subtypes is less challenging. The present results indicate that dissecting the α9α10-nAChR antagonist activity from the α7-nAChR antagonism can coincide with very low α7and α3β4-nAChR affinities. This is the case, unique in the present series of stilbenol ammonium alkyl ethers, for compound 7.

■ CONCLUSIONS
A series of structural modifications restricting the flexibility of the ethylene linker or varying the bulk of the ammonium head of the 2-triethylammonium ethyl ether of 4-stilbenol (1), an α7 and α9α10-nAChR antagonist exhibiting partial agonism at high supra-micromolar concentrations, resulted in significant variations of its α7-nAChR affinity (104 nM K i ). The modifications maintaining or increasing the α7-nAChR affinity of the lead did not substantially improve its profile in terms of potency as antagonist and/or of α7-/α9α10-nAChR selectivity (compounds 12, 13, 14, 15, 18, and 19). Otherwise, some modifications, detrimental for the α7-nAChR affinity, such as oversized increase or decrease of the ammonium head volume or inclusion of the linker in six-membered nitrogen heterocycles, resulted in α9α10-nAChR antagonists devoid of any antagonist activity at the α7-nAChR (compounds 6, 7, (±)-22, 23, and (±)-24). Further characterization of these selective α9α10-nAChR antagonists for intrinsic activity at high concentrations at both the subtype receptors allowed us to get insight into the mechanism of their selective α9α10-nAChR antagonism, most likely consisting of opening and rapidly engaging the channel, then blocking it in an open and nonconducting state. Among these analogues of 1, compound 7 stands out for its highest potency as an α9α10-nAChR antagonist and by exhibiting nearly complete block of poststimulation ACh activity at the α9α10-nAChR. The additional lack of an α7-nAChR response combined with very low α3β4-nAChR affinity makes compound 7 an invaluable and, to our knowledge, unique tool to define the highly debated potential of α9α10-nAChR antagonists as new therapeutics for the treatment of inflammatory and neuropathic pain. ■ EXPERIMENTAL SECTION Chemistry. All chemicals and solvents were used as received from commercial sources or prepared as described in the literature. Flash chromatography purifications were performed using KP-Sil 32−63 μm 60 Å cartridges. Thin-layer chromatography (TLC) analyses were carried out on alumina sheets precoated with silica gel 60 F254 and visualized with UV light. Content of saturated aqueous solution of ammonia in eluent mixtures is given as v/v percentage. R f values are given for guidance. 1 H and 13 C NMR spectra were recorded at 300 and 75 MHz using an FT-NMR spectrometer. Chemical shifts are reported in ppm relative to residual solvent (CHCl 3 , MeOH, or DMSO) as internal standard. Melting points were determined by Buchi Melting Point B-540 apparatus. Optical rotations were determined using a Jasco P-1010 polarimeter. Chiral high-performance liquid chromatography (HPLC) analyses were performed using Hewlett Packard 1050 instrument.
Liquid chromatography−mass spectrometry (LC-MS) analysis was performed using an Agilent 1200 series solvent delivery system equipped with an autoinjector coupled to a PDA and an Agilent 6400 series triple quadrupole electrospray ionization detector. Gradients of 5% aqueous MeCN + 0.1% HCO 2 H (solvent A), and 95% aqueous MeCN + 0.05% HCO 2 H (solvent B) were employed. Purity was measured by analytical HPLC on an UltiMate HPLC system (Thermo Scientific) consisting of an LPG-3400A pump (1 mL/ min), a WPS-3000SL autosampler, and a DAD-3000D diode array detector using a Gemini-NX C18 column (4.6 mm × 250 mm, 3 μm, METHOD E: General Procedure for the Preparation of (±)-55 and (±)-56. Intermediates (±)-53 or (±)-54 (1.53 mmol, 1 equiv) were dissolved in absolute ethanol and a solution of NaOH 1 M (5 equiv) was added. The reaction mixture was stirred at 60°C for 3 h. Upon cooling to room temperature, the solvent was evaporated under reduced pressure, the resulting crude was diluted with water and acidified to pH 3 by dropwise addition of a solution of HCl 1 M solution, and the product was extracted with EtOAc. The organic layer was dried over anhydrous Na 2 SO 4 , filtrated, and evaporated under reduced pressure to afford the desired compound in very good yields (78−95%).
METHOD F: General Procedure for the Preparation of (±)-57 and (±)-58. The reaction was carried out in three successive steps.
(1) In an oven-dried round-bottom flask under N 2 atmosphere, (±)-55 or (±)-56 (0.7 mmol, 1 equiv) was suspended in anhydrous Et 2 O (4 mL). Upon addition of triethylamine (1.1 equiv), the reaction mixture was cooled to −10°C, and isobutyl chloroformate (1.1 equiv) was added dropwise. The suspension was stirred at −10°C for 2 h. Upon completion, monitored by TLC, the solvent was evaporated under reduced pressure and the crude containing the correspondent isobutyl anhydride was utilized in the next step without any further purification. (2) Under inert atmosphere, the crude was dissolved in anhydrous THF (10 mL). After the addition of sodium azide (12 equiv) and of a catalytic amount of tetrabutyl ammonium bromide (0.1 equiv), the mixture was stirred at 40°C for 4 h. Upon cooling to room temperature, the solvent was concentrated under reduced pressure and the resulting residue was (3) dissolved in tertbutylalcohol (10 mL) and refluxed for 48 h, after which the solvent was removed under reduced pressure providing a residue that was purified by silica gel chromatography providing the desired compounds (±)-57 or (±)-58 as white solids in 65−91% yields.
METHOD G: General Procedure for the Preparation of Compounds (±)-59 and (±)-60. Intermediates (±)-57 or (±)-58 (0.1 mmol, 1 equiv) were dissolved in MeOH (2 mL), and the solution was cooled to 0°C. Under vigorous stirring, a methanolic solution of HCl 1.25 M (6.25 equiv) was added dropwise. The reaction mixture was heated at reflux temperature for 3 h. Upon cooling at room temperature, the organic solvent was concentrated under reduced pressure providing a crude solid, that was suspended in EtOAc, stirred for 1 h, and filtered, providing the desired compounds (±)-59 and (±)-60 as white solids in 58−75% yields.
METHOD H: General Procedure for the Preparation of (±)-61 and (±)-62. Intermediates (±)-59 or (±)-60 (0.38 mmol, 1 equiv) were dissolved in DCM and washed with a saturated solution of NaHCO 3 to generate the corresponding free base. The organic layer was dried over anhydrous Na 2 SO 4 , filtered, and the solvent was removed under vacuum. The residue was dissolved in ethyl iodide (4 mL) and K 2 CO 3 (1.5 equiv) was added. The reaction mixture was stirred at 40°C for 20 h and then concentrated under reduced pressure. The residue was purified by silica gel flash column chromatography (gradient from cyclohexane to cyclohexane/EtOAc 8:2 + 0.5% N,N-diisopropylethylamine (DIPEA)) providing the desired compounds (±)-61 and (±)-62 as white solids in 32−53% yields.
(E)-4-(2-Bromoethoxy)-stilbene (28). A suspension of anhydrous K 2 CO 3 (5.28 g, 45.9 mmol, 2.5 equiv), (E)-4-hydroxystilbene (3.00 g, 15.3 mmol, 1.0 equiv), and KI (0.19 g, 1.15 mmol, 0.075 equiv) in 30 mL of 2-methylethylketone was stirred for 30 min and then 1,2dibromoethane was added (5.6 mL, 12.15 g, 64.5 mmol, 4.2 equiv). The reaction mixture was refluxed under nitrogen atmosphere for 48 h. Upon cooling at room temperature, the inorganic salts were removed by filtration and the solvent was evaporated under reduced pressure. The residue was diluted with DCM and washed with an aqueous solution of NaOH 1 M. The organic layer was dried over anhydrous Na 2 SO 4 , filtered, and the solvent was evaporated under reduced pressure. The crude was recrystallized from EtOAc to yield the desired product as a white powder in 61% yield. (E)-4-(2-Iodoethoxy)-stilbene (29). (E)-4-(2-bromoethoxy)-stilbene 28 (1.50 g, 4.95 mmol) was dissolved in 30 mL of a saturated solution of NaI in acetone and the reaction mixture was refluxed overnight. Afterward, the solvent was evaporated under vacuum, and the residue was diluted with diethyl ether, washed with a 10% solution of Na 2 S 2 O 5 , and then washed with brine. The organic phase was dried over anhydrous Na 2 SO 4 , filtered, and evaporated under vacuum, affording the desired product as a white powder in quantitative yield.  (30). Obtained from 186 mg (0.53 mmol, 1 equiv) of (E)-4-(2-iodoethoxy)-stilbene 29 and diethylamine (10 equiv) in THF (3 mL), at reflux temperature, overnight according to METHOD A. The crude was concentrated under reduced pressure, the residue was dissolved in AcOEt and extracted three times with HCl. The water layer was basified to pH 10 with NaOH and extracted three times with AcOEt. The organic layer was dried over anhydrous Na 2 SO 4 , filtered, and evaporated under vacuum, providing the desired compound 30 as a white solid in 94% yield. Mp  (

E)-4-(2-(N,N-Dimethylamino)ethyloxy)stilbene (31).
Obtained from 280 mg (0.80 mmol, 1 equiv) of (E)-4-(2-iodoethoxy)-stilbene 29 and a 2 M solution of dimethylamine in THF (10 equiv) in THF, at 40°C, overnight, according to METHOD A. The crude was concentrated under reduced pressure, the residue was dissolved in AcOEt and extracted three times with HCl. The water layer was basified to pH 10 with NaOH and extracted three times with AcOEt. The organic layer was dried over anhydrous Na 2 SO 4 , filtered, and evaporated under vacuum, providing the desired compound 31 as a white solid in 92% yield. Mp = 103−105°C (coherent with the literature 26 ). (
Biological Assays. Affinity to α7, α3β4, and α4β2 Nicotinic Receptors. For (±)-[ 3 H]epibatidine (specific activity of 56−60 Ci/ mmol; PerkinElmer, Boston MA), saturation binding studies were carried out on membrane homogenates. These were prepared from either SH-EP1 cells stably transfected with α3and β4-nAChR subunit cDNAs, 16 or HEK 293 cells stably transfected with the α4 and β2 cDNAs (generous gift of Dr. Jon Lindstrom). 17 For saturation experiments, the membrane homogenate aliquots were incubated overnight at 4°C with 0.01−5 nM concentrations of (±)-[ 3 H]epibatidine. Nonspecific binding was determined in parallel by adding to the incubation solutions 100 nM unlabeled epibatidine (Sigma-Aldrich) as described previously. 31 At the end of the incubation, the samples were filtered on a GFC filter soaked in 0.5% polyethylenimine and washed with 10 mL of ice-cold phosphatebuffered saline (PBS), and the filters were counted in a β counter.
For competition studies, the inhibition of [ 3 H]epibatidine and [ 125 I] αBgtx binding was measured by incubating the membranes transfected with the appropriate subtype with increasing concentrations of the compounds (1 nM−1 mM) 5 min followed by overnight incubation at 4°C, with [ 3 H]epibatidine 0.1 nM for the α4β2 subtype or [ 3 H]epibatidine 0.25 nM for the α3β4 subtype or at r.t. with [ 125 I]αBgtx 2−3 nM in the case of the α7 subtype. At the end of the incubation time, the samples were processed as described for the saturation studies.
[ 3 H]epibatidine binding was determined by liquid scintillation counting in a β counter, and [ 125 I] αBgtx binding by direct counting in a γ counter. Saturation binding data were evaluated by one-site competitive binding curve-fitting procedures using GraphPad Prism version 6 (GraphPad Software, CA). In the saturation binding assay, the maximum specific binding (B max ) and the equilibrium binding constant (K d ) values were calculated using one site-specific binding with Hill slope−model. K i values were obtained by fitting three independent competition binding experiments, each performed in duplicate for each compound on each subtype. Inhibition constants (K i ) were estimated by reference to the K d of the radioligand, according to the Cheng−Prusoff equation and are expressed as nM values.
Two-Electrode Voltage Clamp (TEVC) Recording of α7-and α9α10-nAChR Function. For functional pharmacology studies, twoelectrode voltage clamp recordings were performed, using human nAChR subunits heterologously expressed in X. laevis oocytes. Approaches were closely related to those previously detailed. 19 Briefly, X. laevis oocytes were purchased from Ecocyte Bioscience US (Austin, TX), and the incubation temperature was 13°C. Harvesting of oocytes from X. laevis by EcoCyte follows the guidelines of the National Institute of Health's Office of Laboratory Animal Welfare and was authorized under IACUC number #1019-1 (valid through December 2022). Injections of nAChR subunit mRNA were made using glass micropipettes (outer diameter ≈40 μm, resistance 2−6 MΩ), and mRNA was injected in a total volume of 40 nL. For α7-nAChR, 1.25 ng of α7-nAChR subunit mRNA was injected per oocyte, along with 0.125 ng of NACHO mRNA to improve functional expression. 32 For α9α10-nAChR, a total of 10 ng of nAChR subunit mRNA was injected using α9 to α10 cRNAs in a 9:1 ratio by mass.
TEVC recordings were made in oocyte saline solution (82.5 mM NaCl, 2.5 mM KCl, 5 mM HEPES, 1.8 mM CaCl 2 · 2 H 2 O, and 1 mM MgCl 2 · 6 H 2 O, pH 7.4), and were performed at room temperature (20°C ). One week after injection, oocytes were voltage-clamped (−70 mV; Axoclamp 900A amplifier, Molecular Devices, Sunnyvale, CA). Recordings were sampled at 10 kHz (low-pass Bessel filter, 40 Hz; high-pass filter, DC), and saved to disk (Clampex v10.2; Molecular Devices). To ensure quality of recordings, oocytes with leak currents (I leak ) > 50 nA were discarded without being recorded from. In all cases, initial control stimulations (ACh, 1 mM, applied for 1 s) were performed, with 60 s washout (no drug) between control stimulations (total of 5 stimulations). This allowed us to define a 100% response control, and to ascertain that run-down or desensitization was not occurring due to repeated ACh stimulation.
For antagonist concentration−response curves, test compounds were applied simultaneously with 1 mM ACh, starting with the lowest concentration of test compound and increasing in half-log steps to a maximum concentration of 100 μM. The standard 1 min spacing between stimulation was maintained. Data for each oocyte were normalized by expressing peak function in the presence of test compounds as % of control function (the mean peak function measured across the initial control stimulations was defined as 100% for each oocyte). IC 50 values were calculated from these normalized nAChR-mediated currents through nonlinear least-squares curve fitting (GraphPad Prism 5.0; GraphPad Software, Inc., La Jolla, CA).
Intrinsic agonist efficacy of test compounds was measured by applying them (alone at 100 μM, 1 s application time, no ACh coapplication) 1 min following the last initial control stimulation. Peak function following addition of the test compound was normalized for each oocyte in the same way just described for antagonist concentration curves. The same normalization was applied to the peak of any rebound current observed during the 60 s washout period following application of the test compound, and to the peak function induced by a final control application of ACh (1 mM, 1 s application time).