Structure–Activity Relationships within a Series of σ1 and σ2 Receptor Ligands: Identification of a σ2 Receptor Agonist (BS148) with Selective Toxicity against Metastatic Melanoma

A new series of spirocyclic σ receptor (σR) ligands were prepared and studied. Most were found to have a high affinity and selectivity for σ1R; three compounds were shown to be σ1R agonists, while another proved to be the only σ1R antagonist. Only one of the σ1R agonists (BS148) also exhibited σ2R selectivity and was able to inhibit the growth of metastatic malignant melanoma cell lines without affecting normal human melanocytes. The antiproliferative activity of this compound suggested an σ2R agonist profile. Further, preliminary investigations indicated that the mechanism of metastatic malignant melanoma cell death induced by BS148 is due, at least in part, to apoptosis.


Introduction
Sigma receptors (sRs), initially classified as an additional class of opioid receptors, [1] are now considered au nique entity with no homology to opioid receptors or other mammalian proteins. [2] After their discovery, radioligand binding studies and biochemical analyses classified sRs into two distinct subtypes: s 1 Rand s 2 R. [3,4] The s 1 Rh as been characterized and cloned from different species, including humans. [5][6][7][8][9] It has been shownt hat the endogenous regulators progesterone, dehydroepiandrosterone, N,N-dimethyltryptamine, [10][11][12] and several exogenous ligands such as benzomorphans (SKF10 047 and pentazocine), [13][14][15][16][17] hal-operidol, and NE-100can interact with s 1 Rs. [14,[17][18][19] High-affinity s 1 Rl igandsh ave been considered to play an important role in the treatment of various neurodegenerative disorders, including Parkinson's, Alzheimer's, and Huntington's diseases,d ementias, cognitive aging, depression, schizophrenia, neuropathicp ain, drug addiction, stroke, HIV infection, and cancer. [20][21][22][23][24][25][26][27][28][29] Unlike s 1 Rs, s 2 Rs are less understood and have not yet been cloned. Recently,t he progesterone receptor membrane component 1( PGRMC1) has been proposeda st he s 2 R binding site, [30] although this wasq uestioned by other evidence supporting that the two targets are two different proteins. [31] More recently TMEM97,a ne ndoplasmic reticulum (ER)-resident membrane protein, was shown to possess the full suite of molecular properties that define the s 2 R. [32] Activation of the s 2 Ra ppearst ob ei nvolved in the regulation of cellular proliferation andc ell death. [33,34] Furthermore, it has been reported that s 2 Rl igands can be used as biomarkers in various rapid proliferating tumors and thus could be exploited for tumor imaging. [33,35] The s 1 Ra nd s 2 Rh ave been found to be overexpressed in av ariety of cancerc ells, [34] and it has been described that selective s 1 Ra ntagonists and s 2 Ra gonists are able to modulate cancerc ell viability and growth potential. [36,37] Among the different types of cancer, malignantm etastatic melanoma is by far one of the mostc ommona nd aggressive oncologic diseases. According to estimates for 2017, 87 110n ew melanoma cases will be diagnosed and 9730 people are expected to die by melanoma in the US alone. Advanceds tage melanoma is often associated with an overall median survival period of 2-8 monthsa nd with only 5% patient survival beyond five years. At present, only three compounds have been approved by the US Food and Drug Admin-An ew series of spirocyclic s receptor (sR) ligandsw ere prepared and studied. Most were found to have ah igh affinity and selectivity for s 1 R; three compounds were shown to be s 1 Ra gonists, while another proved to be the only s 1 Ra ntagonist. Only one of the s 1 Ra gonists (BS148) also exhibited s 2 R selectivity and was able to inhibitt he growth of metastatic malignant melanoma cell lines without affecting normal humanm elanocytes. The antiproliferative activity of this compounds uggested an s 2 Ra gonist profile. Further,p reliminary investigations indicated that the mechanism of metastatic malignant melanoma cell deathi nduced by BS148 is due, at least in part, to apoptosis.
[a] Dr.S. Franchini istration (FDA) for the treatment of malignant metastatic melanoma:d acarbazine (DTIC), interleukin-2( IL-2), and vemurafenib (Zelforab). However, only vemurafenib has proven capable of increasing overall patient survival althoughw ith al ow remission rate. [38] Due to the broad diagnostic andt herapeutic potential, the development of potent and selective s 1 Ro r s 2 Rl igands is one of the primary challenges in medicinal chemistry. We recently reported an ew series of conformationally restricted spirodioxolane, among which 1a and 1b showed higha ffinity and good selectivity for s 1 R, behaving as s 1 Ra gonists ( Figure 1). [39] Thes pirocyclic substructure has been shown to be an interesting moiety for sRligands. [40,41] In the present study,w ee xplored an ew set of structurally related analogues of 1a and 1b in an attempt to identify novel and selectivel igandsb yi nvestigatingt he effect of different substitutions on both affinity and selectivity at s 1 Ra nd s 2 R ( Figure 1). The C8 positiono ft he 1,4-dioxaspiro [4.5]decane moiety was selected asastartingp oint, due to its synthetic accessibility and because it could be exploited withouta dding a second stereogenic center.T hus, several substituents were introduced at position8:h ydrogen bond donor/acceptor groups that could establish hydrogen bond contacts, aromatic groups that might be beneficial for additional van der Waals and p-p stacking interactions, and bulky groups that could be favorable for detecting hydrophobic contacts. Furthermore,w eg enerated variants of the leads, focusing our attentiononthe 1,3-dioxolane moiety and replacing it with the corresponding oxathiolane and dithiolane isosteres. This approach led to extensive qualitative structure-activity relationship (SAR) studies and identification of valuable high-affinity sRl igands. Docking studies were performed using the X-ray crystallographic structure of the human s 1 R.
Compounds 15 a, 16 a (BS148), and 15 b, 16 b were similarly prepared, as outlined in Scheme 2, from tosylate 13 or chloride 14, [43] whichw ere reacted with 4-benzylpiperidine or 1-benzylpiperazine to give full assignments of protons andc arbons for all compounds. Thisw as done using 1 Ha nd 13 CNMR analysis through related1 Ds pectra acquisition and the study of key values obtainedf rom 1 H-1 HC OSY, 1 H-13 CH SQC, and HMBC experiments. All final compounds werec onvertedi nto oxalate salts using oxalic acid in acetone. The puritieso ft he salts were confirmed by elemental analysis, andt he obtained values were within AE 0.4 %o ft he calculated values. The exact mass of the salts was confirmed by HPLC-QTOF measurement.
Scheme1.Synthesis of compounds 6-12:a)4-benzylpiperidine or 1-benzylpiperazine, KI, 2-methoxyethanol,120 8C, 12 h, 70 %y ieldfor 6a,85% for 6b,86% for 7a,58% for 7b,82% for 8a,7 7% for 8b,40% for 9a,and 27 %f or 9b;b )TFA,C H 2 Cl 2 ,08Cto258C, 3h,93% yield for 10 a,and 43 %f or 10 b; c) C 6 H 5 SO 2 Cl or C 6 H 5 COCl, Et 3 N, CH 2 Cl 2 ,08Ct o2 58C, 12 h, 32 %y ieldfor 11 a,6 7% for 11 b,5 7% for 12 a,and 73 %f or 12 b. and a 1 adrenoceptors were investigated, as most of the molecules share chemical features with previously published 5-HT 1A Ra nd a 1 ligands. [43][44][45][46] None of the compounds were found to bind to the a 1 adrenoceptors;t herefore, these data were not included in the Tables. Compounds 1a and 1b were included as reference compounds to better understand the SAR. As previously reported, 1a and 1b bound preferentially to the s 1 R, with a1 0-and 47-fold preference for the s 1 Ro ver the s 2 R, respectively.W ei nitially studied isosterics ubstitution of the carbon atom at position 8o ft he 1,4-dioxaspiro [4.5]decane with sulfur,o xygen, or nitrogen atoms. Increasing the polarity on 4-benzylpiperidine derivatives( 6a< 7a< 10 a)w as shown to have al imited effect on both s 1 Ra nd s 2 Ra ffinity.C ompound 10 a,t he most polar of the compounds, was the only compound to show am arked decrease in affinity for both the s 1 Ra nd s 2 Rw ith reverse s 1 /s 2 selectivity.T he observed decreaseda ffinity was probably due to the protonation of the nitrogen atom, as this positivec hargea ppeared to be an unfavorable factor for binding. For 1-benzylpiperazine derivatives, the increasing polarity (6b< 7b< 10 b)h ad an even stronger effect on affinity at the s 1 Ra nd s 2 R, and compound 10 b was nearly unable to bind at both receptors. On the other hand, the benzylated derivatives of 10 a and 10 b,c ompounds 8a and 8b,d isplayed increased affinity at s 1 Ra nd s 2 R, although protonation of the same nitrogen atom may have occurreda sw ell. In particular,c ompound 8a,b earing a4 -benzylpiperidine moiety,e xhibited ap K is1R value (9.08) higher than lead compound 1a with as imilars electivityr atio (s 1 R/s 2 R = 9). This may be explained by the fact that the introductiono fabenzyl moiety increased lipophilicity, which could overcome the negative effect of the positive charge on then itrogena tom. Conversely,a ll amidic ands ulfonamidic derivatives coupled with a4 -benzylpiperidine group (9a, 11 a, 12 a)o rw ith 1-benzylpiperazine( 9b, 11 b, 12 b) showedl ower affinity values, both at s 1 Ra nd s 2 R . Subsequently,i sosteric substitution of the annular oxygen atoms of the 1,4-dioxaspiro [4.5]decane moiety of compounds 1a and 1b was evaluated. By replacing the oxygen with as ulfur atom at Table 1. Affinity constants( pK i )and selectivity of benzylpiperidine derivatives 6-12 a, 15-16 a,a nd reference compound 1a at s 1 R, s 2 R, and 5-HT 1A R.

Compd
Rp K i s 1  [47] at one or two concentrations and agreedw ithin 10 %.
[f] Antilog of the difference between the pK i values for s receptors( higher value) and the 5-HT 1A R. ChemMedChem 2017ChemMedChem , 12,1893ChemMedChem -1905 www.chemmedchem.org 2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim the 4-position to give 1-oxa-4-thiaspiro [4.5]decane derivatives 15 a and 15 b,t he affinity at the s 1 Ra nd s 2 Rw ere diversely affected. The 1-benzylpiperazine derivative 15 b showedasmall decreasei na ffinity,w hile this decrease was much more marked for 4-benzylpiperidine 15 a.U pon replacing both oxygens by sulfur atoms to give 1,4-dithiaspiro [4.5]decane derivatives, ad ifferent behavior was again observed for 4-benzylpiperidine 16 a (BS148) and 1-benzylpiperazine 16 b.T he latter compound highlighted af urther decrease in affinity at s 1 R while the affinity at s 2 Rw as barely affected. Adversely,f or compound 16 a,t he affinity at the s 2 Rr emained unchanged, whereas at s 1 R, it decreased up to 100-fold with respectt o compound 1a,t hus giving ar eversal of selectivity (selectivity value = 0.04). Therefore, compounds, 1b, 7b, 15 b, 16 b,a nd 16 a were selected for furtherp harmacological characterization; 1b, 7b, 15 b,a nd 16 b for their affinity and selectivity toward s 1 R, and 16 a for its affinity and selectivity toward s 2 R.

Molecularm odeling studies
The rational design of novel s 1 Rl igands wase fficiently driven by deepening computational methods, including homology modeling of the biological target, [37, 42-46, 48, 49] as well as pharmacophore-based criteria. Indeed, severals eries of derivatives fulfilling specific patternso fc hemical features have been described in the literature. [49] Most were characterized by positive ionizableg roup and hydrogen bond acceptor functionalities, connecting at least two hydrophobic cores. Furthermore, docking studies revealed ar ecurrent binding mode for s 1 Rl igands, based on key contacts with an aspartic acid residue (Asp 126), as confirmed by mutagenesis data. [50] More recently,t he X-ray crystallographic structure of the human s 1 Rb ecame available, paving the way for the rational design of new compounds (Protein Data Bank [PDB] ID:5 HK1;r esolution = 2.51 ). [51] In particular, these data allowed exploration of the binding mode of the co-crystallized ligand PD144418( 1,2,3,6-tetrahydro-5-[3-(4-methylphenyl)-5-isoxazolyl]-1-propylpyridine). As shown in Figure 2, PD144418 exhibited salt bridges between the protonated nitrogen atom of the tetrahydropyridine group and Asp 126 and Glu 172, validating the aforementioned homology modelsa nd mutagenesis data. In addition, the ligand was stabilized within the receptor binding site by hydrophobic contacts involving the propyl chain and phenyl ring and the surrounding amino acids Leu 105, Ile 178, and Leu 182. The recently available X-ray crystallographic data of the human s 1 Ra llowed us to perform more reliable docking calculations.Thus, we had the opportunity to gain more detailed insight into the SAR of the newly synthetized compounds through docking studies focusedo nt he s1b inding site (Table 1S,S upporting Information). The main issue to be addressedw as to clarify the role played by the Rm oiety in terms of steric and electrostatic properties when it is coupled with the piperidine or piperazine moiety.A no verall perspective of the derived docking poses supported the ability of all compounds to display the required contact with Glu 172, thus being able to anchor the s 1 R. In particular, within the piperidine series (a series), the Rsubstituent and the benzylm oiety were shown to properlym imic the propylp iperidine and the phenyl-oxazole moiety of PD144418, respectively.T he docking mode of 1a is shown in Figure 3, highlighting the S enantiomer as the most probable. Indeed, all compounds maintained a commonp ositioning,exhibiting as alt bridge between the protonatedn itrogen atom and Glu 172, while the benzyl moiety occupied ac avity including Leu 105, Ile 178, and Leu 182.O n the other hand, the Rgroup was projected toward ad eep receptor pocket delineated by Phe 107, Tyr120, and Trp164,d etecting hydrophobic contacts. Notably,t he presence of as mall hydrogen bonding heteroatom, such as oxygen, for compounds 1a-12 a allowed them to hydrogen bond more efficiently to Tyr120 than the corresponding oxathiolane or dithiolane. Bulkier groups provedt ob ew ell-suited for the piperidine  Conversely,t he piperazine analogues (b series) exhibited a reversed ocking mode in the presence of small and hydrophobic Rgroups, as for 1b, 6b, 15 b,a nd 16 b,m oving the Nbenzylp iperazine within much closer proximity to the piperidine ring of PD144418, while the Rsubstituent overlapped the reference phenyl-oxazole, featuring van der Waals contacts with Met 93, Tyr103,I le 178, and Ala 185. This behavior allowed this series of compounds to better stabilizeb oth protonated nitrogen atoms with the Glu 172 sidec hain, while no hydrogen bonds involved the Rgroup (Figure 4; docking mode of 1b is shown as the R enantiomer,w hich is the most probable).
In addition, moving from the dioxolane to oxathiolane and dithiolanel ed to piperazines in all cases endowed with adequate affinity values for s 1 R, while the binding ability of the related piperidines proved to be deeply impaired, lacking additional hydrogen bonds with Tyr120. When hindered substituents were placed at the Rposition, as for 8b, 11 b,a nd 12 b (pK i s 1 = 6.90-7.72), the piperazines displayed as imilard ocking mode with respectt ot he previously discussed piperidine analogues, leading to detectiono fo nly one of the two mentioned salt bridgesa nd exhibiting lower affinity valuest han the smaller analogues 1b, 6b, 15 b,a nd 16 b (pK i s 1 = 8.21-9.13). In conclusion,w ec an summarize that, with the exceptiono fc ompounds 7a and 7b, bulky Rgroups are required for the piperidine series, while smaller Rgroups are preferred within the piperazineseries.

Analgesic activity
It is known that s 1 Ra ffects the analgesic activity of opioid analgesics,a sp reviously reported. [52] It has been demonstrated that while agonists decreases ystemic and supramaximala nalgesia, antagonists increasei t. Compounds 1b, 7b, 15 b, 16 a (BS148), and 16 b,a ts ubcutaneous (s.c.) doses of 1.0, 2.0, and 5.0 mg kg À1 ,d id not affect basal tail flick latencies (TFLs) during the entire time of observation (data not shown). Injection of mopioid peptide (MOP)a gonist morphine, at ad ose of 4.0 mg kg À1 s.c.,s ignificantly increased the nociceptive latency following thermal stimulation, which demonstrated ac lear analgesic effect (*p < 0.05 vs. saline-treated mice;F igure 5a). Compared with the group of mice treated withs aline, the changes from basal TFL level (mean area under the curve [MAUC] over 120 min of observation) was increased from 2.69 to 117.3 % ( Figure 5b).
Pre-treatment with compounds 1b, 7b, 15 b,a nd 16 a at 1.0 mg kg À1 s.c.,f ollowed by morphine (4.0 mg kg À1 s.c.), caused as ignificant decreasei nt he opioid analgesic effect. The MAUCv alue was significantly lower than that in mice treated with opioid alone:5 3.3, 4.1, 0.95, and 7.85 vs. 117.3 for compounds 1b, 7b, 15 b and 16 a,r espectively (**p < 0.01 vs. morphine-treatedm ice;F igure 5b). Ad ecrease in the analgesic effect was observed for the entire period of observation. These resultsa re consistent with agonistic behavior of these compounds at s 1 receptors, confirming results previously reported  after the last injection, over the 120 min testing period. Data are expressed as the mean AE SEM from 8-10 mice.* p < 0.05 vs. saline-treatedmice; **p < 0.05 vs. morphine-treatedm ice; # p < 0.05 vs. morphine-treated mice. ChemMedChem 2017ChemMedChem , 12,1893ChemMedChem -1905 www.chemmedchem.org with 1b on ar at preparation. [39] In contrast, pre-treatment with 16 b at 1.0 mg kg À1 s.c.,f ollowed by morphine injection (4.0 mg kg À1 s.c.), caused an increasei nt he opioid analgesic effect, and the MAUC value was significantly highert han that of mice receiving morphinea lone (192.4 %c ompared with 117.3 %; # p < 0.05 vs. morphine-treatedm ice) (Figure 5b), suggesting an antagonist profile at s 1 R. This result is of certain interest, as enhancemento ft he analgesic effect could decrease the necessary dose of opioid agonists in pain therapy,m inimizing the adverse effects of opiate treatments. [53] Furthermore, it has been shown that s 1 Rantagonists may have promisinga ntinociceptive properties in differentn eurogenic pain models. [54] It has also been shownt hat high-affinity and selective s 1 Ra ntagonists dose-dependently inhibitm echanicala llodynia, which was reversed completely by the application of as elective s 1 Ra gonist. [55] Therefore, compound 16 b deserves further investigation.

Antiproliferative activity
In the current study,t he effect on cell viability of compounds 1b, 7b, 10 a, 15 b, 16 a (BS148), and 16 b was primarily tested on two melanoma cell lines:S K-MEL-28, derived from primary melanoma, and SK-MEL-2, derived from metastatic melanoma. [56,57] Siramesine,apotent s 2 Ra gonista nd ap rovenp otent cell toxicityi nducer,w as used as ap ositive control for cell death. [58,59] In vitro toxicity of the compounds was determined by MTT cell viabilitya ssay.E ach experiment was performed in triplicatei nt hree independente xperiments,a tc oncentrations of 20, 40, 80, and 100 mm,i no rdert oo btain dose-dependent curves. As shown in Figure 6, almost all tested compounds decreasedc ell viabilityt ov arying percentages.I nt he primary melanoma model (SK-MEL-  The results of this study provide additional information about both the sigma pharmacology of these molecules and the employed melanoma cancerm odels. Based on the results of the binding experiments (Tables1 and 2), BS148 was found to be the only compound to show marked s 2 Rs electivity.A ccording to the MTT assay,t he selective toxicityo fB S148 revealed a s 2 Ra gonistp rofile, whichw as consistent with current literaturei ndicating that s 2 Ra ctivation is responsible for cell death. [36] The in vivo analgesia studies demonstrated that BS148 behaves as a s 1 Ra gonist, thus excluding the involvement of the s 1 Ri nc ell death, as it is already known that only antagonists at s 1 Ra ffect cell proliferation. [37] Furthermore, the selectived eath of metastatic SK-MEL-2 cells caused by BS148 may also unveild ifferent sRe xpression patterns, suggesting that the s 2 Rs ubtypec ould be over expressed in metastatic melanoma cells. Regardless of the exact mechanism of action, the lack of toxicityo nh ealthy melanocytes suggestst hat BS148 is ap romising candidate to be developed as an ovel drug in melanoma chemotherapy.I na ni nitial attemptt ou nderstand, whenever possible, the mechanism of action of BS148, DNA fragmentation was measured by terminal deoxynucleotidyl transfer-mediated dUTP nick-endl abeling (TUNEL) for the detection of apoptotic cells. Normal human melanocytes, SK-MEL-2, and SK-MEL-28 cell lines were treatedw ith BS148 at doses of 20 and 40 mm at 24 (data not shown)a nd 48 hours.B oth negative controls (untreated cells) and positive controls (siramesinea t4 0mm,9 8% of apoptotic cells, data not shown)w ere employed.
More than 20 %o fp ositive cells were detectedb yf luorescence microscopy in SK-MEL-2 treated samples (16 a,4 0mm)a t 48 hours,w hile almostn oa poptotic cells (< 3%)w ere present in normalh uman melanocytes and SK-MEL-28 cell lines (Figure 8a). The number of apoptotic cells proportionally increases   (Figure 8b). These preliminary results suggest that the apoptotic pathway is involved in s 2 Ri nduced cell death,b ut apoptosis is only partly responsible for cell death. Recently,o ther authors have shown that s 2 Rl igandsi nducet umor cell death in the MDA-MB-435 human melanoma cell lines by multiple signalingp athways, such as autophagy and cell cycle deregulation. [36] However,m ore research is needed to better define the functional activity of BS148, and future investigations into cell death mechanisms should focus on oxidative stress induction. Further research should also consider recent findings on tumor cell lines, whichs trongly suggest the loss of mitochondrial integrity as ac ritical factor for cell death induction. [58] Conclusions Startingf rom the 1,4-dioxaspiro [4.5]decaned erivatives 1a and 1b,1 8n ew compoundsw ere synthetized and studied. Most were found to have affinity for s 1 Ri nt he nanomolar range.
The most interesting compounds were furthers tudied to determineafunctionalb ehavior.I nt he tail flick test, 7b, 15 b, and 16 a (BS148) decreased the analgesic effect of morphine, which allowed them to be defined as s 1 Ra gonists. Conversely, 16 b was able to potentiate this effect, defining it as s 1 Rantagonist. In thea ntiproliferative study,B S148, the only compound to exhibit s 2 Rs electivity,w as able to induce selective toxicity toward SK-MEL-2, am etastatic malignant melanoma cell line, while not affecting normal human melanocytes.T hese results suggested an agonisticp rofile at s 2 R. Malignant melanoma cell death induced by BS148 was due at least in part to apoptosis, as indicated by TUNEL assay. The current study suggestst hat BS148 should be furtheri nvestigated as ap romising candidate for novel metastatic melanoma therapy development.  [47] at one or two concentrations and agreedw ithin 10 %.

Experimental Section
Chemistry General methods:A ll reagents, solvents, and other chemicals were used as purchased from Sigma-Aldrich without further purification unless otherwise specified. Reactions using air-or moisture-sensitive reactants and solvents were carried out under nitrogen atmosphere unless otherwise noted. Microwave-assisted synthesis was conducted using aD iscover Labmate reactor (CEM Corp.,M atthews, NC). Flash column chromatography purifications (medium pressure liquid chromatography) were carried out using Merck silica gel 60 (230-400 mesh, ASTM). The purity of compounds was determined by elemental analysis (C, H, N) General procedure A:T he selected halogen derivatives 2-5 [42] and 13 and 14 [43] (1.0 mmol) were added portionwise to as olution of 4-benzylpiperidine or 1-benzylpiperazine (3.0 mmol) and KI (0.2 mmol) in anhydrous 2-methoxyethanol (10 mL). The resulting mixture was heated at reflux for 12 ho ru sing microwave irradiation (160 8Cf or 30 min), then cooled to room temperature and concentrated in vacuo. The residue was partitioned between EtOAc and H 2 O. The organic layer was separated, and the aqueous layer was extracted with EtOAc. The organic layers were combined, washed with H 2 O, dried over anhydrous Na 2 SO 4 ,f iltered, and concentrated in vacuo. The residue was purified by flash chromatography to yield the desired compound. General procedure B:T rifluoroacetic acid (2.6 mmol, 0.2 mL) was added to as tirred solution of 9a or 9b (0.035 g, 0.081 mmol) in CH 2 Cl 2 (2 mL). The resulting mixture was stirred at room temperature for 3h;s ubsequently,t he organic phase was washed with a solution of NaOH (5 %), and dried over anhydrous Na 2 SO 4 .T he suspension obtained was filtered, and the solution was concentrated under vacuum to give the desired compound. General procedure C:T he selected benzenesulfonyl chloride and/ or benzoyl chloride (1.2 mmol) was added at 0 8Ct oas olution of compound 10 a or 10 b (1 mmol) and triethylamine (2 mmol) in dry CH 2 Cl 2 (5 mL). The mixture was stirred at room temperature for 12 h. Ice and water were then added, and the mixture was extracted with CH 2 Cl 2 .T he organic layers were combined and dried over anhydrous Na 2 SO 4 .T he resulting suspension was filtered, and the solution was concentrated under vacuum to give as olid residue, which was purified by column chromatography,p roviding the expected compound.

Biology
Radioligand binding assay at sR:I nv itro s-binding experiments were carried out as previously reported. [60] s 1 Binding assays were performed on guinea pig brain membranes according to an experimental protocol described by De Haven et al. [61]  Molecular modeling:A ll compounds were built, parameterized (Gasteiger-Hückel method), and energy-minimized within MOE (Chemical Computing Group, Montreal) using an MMFF94 force field. For the newly synthesized compounds, the two R and S enantiomers were taken into account and built in silico. Docking calculations within the X-ray structure of human s 1 R( PDB ID:5 HK1) were performed using the Surflex docking module implemented in Sybyl-X1.0 (Tripos, Inc.). Surflex-Dock uses an empirically derived scoring function based on the binding affinities of X-ray proteinligand complexes. The Surflex-Dock scoring function is aw eighted sum of nonlinear functions involving van der Waals surface distances between the appropriate pairs of exposed protein and ligand atoms, including hydrophobic, polar,r epulsive, entropic, and solvation and crash terms, represented in terms of at otal score conferred to any calculated conformer.T he best docking geometry (se-lected on the basis of the Surflex scoring functions) was refined by ligand-protein complex energy minimization (CHARMM27) using the MOE software. Finally,t he protein-ligand complex stability was successfully assessed using as hort ( % 1ps) molecular dynamics (MD) run at constant temperature, followed by an all-atom energy minimization (LowModeMD using MOE software). This module allowed us to perform an exhaustive conformational analysis of the ligand-receptor binding site complex, as discussed in our other case studies, where it proved to be useful for preliminary evaluation of docking poses. [42,64,65] In vivo biologicala ssays Nociceptive test:N ociception was evaluated by the radiant heat tail flick test, consisting of irradiation of the lower third of the tail with an I.R. source. Experiments were performed at room temperature (25 AE 1 8C). The basal predrug latency was established between 3a nd 4s,w hich was calculated as the average of the first three measurements performed at 5min intervals. Ac utoff latency of 20 sw as established to minimize damage to the tail. Post-treatment tail flick latencies (TFLs) were determined at 20, 30, 40, 60, 90, and 120 min after subcutaneous (s.c.) injection. Compounds under study were administered at different doses (1.0, 2.0, and 5.0 mg kg À1 s.c.). Mice were pretreated with compounds at 1mgkg À1 (s.c.) 45 min prior to morphine (4.0 mg kg À1 s.c.);t ail flick latencies were measured after 20, 30, 40, 60, 90, and 120 min from time of opioid administration. Behavioral tests were conducted by researchers blinded to treatment group. [66] Statistical analysis:D ata are expressed as the mean AE SE. The intergroup comparisons were assessed using an initial two-way analysis of variance (ANOVA), followed by Student's t-test. Any differences were considered significant at p < 0.05.
Antiproliferative activity: Reagents:T rypan blue was purchased from Biochrom AG (Berlin, DE), penicillin/streptomycin and trypsin from Lonza (Verviers, BE), and fetal calf serum (FCS), medium 254, and human melanocyte growth supplement (HMGS) from Gibco by Life Te chnologies (Grand Island, NY). EMEM was purchased from American Type Culture Collection;M anassas, VA). For the TUNEL assay,a ni nsitu Cell Death Detection Kit was purchased from Roche (Indianapolis, IN). MTT and siramesine were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Evaluation of cytotoxic activity:C ytotoxic effects of the tested compounds were evaluated by MTT dye test (cell viability). Normal ChemMedChem 2017ChemMedChem , 12,1893ChemMedChem -1905 www.chemmedchem.org 2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim human melanocytes and SK-MEL-2 and SK-MEL-28 were seeded in a96-well culture plate at 15 000, 10 000, and 7000 cells per well, respectively,a nd incubated for 24 ha t37 8Ct op romote adhesion. The culture medium was removed, and increasing concentrations (20,40,80, and 100 mm)o ft est compounds 1b, 7b, 10 a, 15 b, 16 a, 16 b,a nd siramesine were added to the wells and incubated without serum for 24 and 48 h. Untreated cells were used as av iability control. After 24 and 48 h, tetrazolium salt in phosphate-buffered saline (PBS;2 0mL, 5mgmL À1 )w ere added to every well, and the plates were incubated again for 4h.T he medium was removed, and the formazan salts, precipitated on the well bottom after oxidation, were dissolved in DMSO (200 mL). The solubilized formazan was quantified with am icroplate spectrophotometer at a wavelength of 540 nm, using ar eference wavelength of 690 nm. The results were expressed as av iability percentage with respect to the control. Results were calculated as the mean AE SD of three independent experiments performed in triplicate. [67] TUNEL assay:As tandard protocol was provided by the manufacturer within the in situ Cell Death Detection kit. The cells were seeded in slide flasks at densities of 40 000, 36 000, and 27 000 cm À2 for normal human melanocytes and SK-MEL-2 and SK-MEL-28 cells, respectively.T he cells were incubated with aT UNEL kit (50 mL) for 1hat 37 8Ci nah umidified dark slide box. The apoptosis ratio was examined under af luorescence microscope (NIKON TE 2000) by counting positively stained cells in five different visual fields, selected randomly.
Statistical analysis:T he data are expressed as the mean AE SE. The intergroup comparisons were assessed by Student's two-tailed ttest. Any differences were considered significant at p < 0.05.