Development of New Benzylpiperazine Derivatives as σ1 Receptor Ligands with in Vivo Antinociceptive and Anti-Allodynic Effects

σ-1 receptors (σ1R) modulate nociceptive signaling, driving the search for selective antagonists to take advantage of this promising target to treat pain. In this study, a new series of benzylpiperazinyl derivatives has been designed, synthesized, and characterized for their affinities toward σ1R and selectivity over the σ-2 receptor (σ2R). Notably, 3-cyclohexyl-1-{4-[(4-methoxyphenyl)methyl]piperazin-1-yl}propan-1-one (15) showed the highest σ1R receptor affinity (Ki σ1 = 1.6 nM) among the series with a significant improvement of the σ1R selectivity (Ki σ2/Ki σ1= 886) compared to the lead compound 8 (Ki σ2/Ki σ1= 432). Compound 15 was further tested in a mouse formalin assay of inflammatory pain and chronic nerve constriction injury (CCI) of neuropathic pain, where it produced dose-dependent (3–60 mg/kg, i.p.) antinociception and anti-allodynic effects. Moreover, compound 15 demonstrated no significant effects in a rotarod assay, suggesting that this σ1R antagonist did not produce sedation or impair locomotor responses. Overall, these results encourage the further development of our benzylpiperazine-based σ1R antagonists as potential therapeutics for chronic pain.


■ INTRODUCTION
The σ-1 receptor (σ 1 R) was initially categorized as an opioid receptor subtype because of the binding with the nonselective benzomorphan alazocine (SKF10,047). 1 Subsequent studies have proven that naloxone did not possess antagonism at this receptor, 2 and later molecular cloning and the X-ray crystallographic structure of the human σ 1 R revealed no homology with opioid receptors. 3,4 Indeed, unlike opioid receptors, which possess the seven-transmembrane domain structure characteristics of G protein-coupled receptors, σ 1 R is a transmembrane protein that is present in numerous oligomeric states. 4 The protomer contains one transmembrane domain, five α helices, and 10 β strands forming the ligand-binding pocket. 4,5 Therefore, σ 1 R is now recognized as a unique chaperone protein mostly expressed at the endoplasmatic reticulum and is a highly conserved protein among different species with over 90% identical amino acid sequences. 6 A second σ receptor subtype (named σ 2 R) was discovered and differentiated from the first subtype on the basis of size, tissue distribution, and ligand affinity. 7,8 σ 2 R has been even more challenging to define than σ 1 R, and its crystal structure has not yet been reported. Indeed, the protein has just recently been cloned, with a sequence identical to the transmembrane protein 97 (TMEM97), a protein involved in cholesterol homeostasis. 9 Since then, it is general practice to refer to this protein as σ 2 R/TMEM97. Like σ 1 R, σ 2 R/TMEM97 was initially miscategorized and correlated to the progesterone membrane component 1 (PGRMC1), which was thought to be the σ 2 R binding site. 10 Finally, further studies clarified that σ 2 R/TMEM97 might still interact with PGRMC1 and LDLR (low-density lipoprotein receptor), forming a ternary complex that mediates LDL internalization and trafficking. 11,12 Both σ receptor (σR) subtypes are present in several CNS areas and peripheral tissues such as the spleen and liver, as well as overexpressed on different human tumors. 6,13 σ receptors are known to be expressed in key areas of pain control in the CNS such as the locus coeruleus, periaqueductal gray, and rostroventral medulla. 14,15 In the CNS, these receptors are also very highly expressed in the dorsal root ganglia of the spinal cord, indicating a key role in the function of peripheral pain pathways. 16,17 In agreement with their anatomical distribution, σRs modulate a broad range of body functions. 18 Conversely, dysregulation of the physiological activities of σRs has been observed in several medical conditions, including drug addiction, neuropsychiatric disorders, cancer, and chronic pain. 18−22 With an improving understanding of σRs, σR ligands  have recently attracted increased attention from the scientific community for their potential as new medications to treat unmet medical needs, 23−25 including the novel coronavirus disease 2019 (COVID-19) as recently reported. 26,27 Notably, novel σR selective ligands are currently under evaluation in clinical trials as diagnostic agents (i.e., PET radiotracers) and therapeutic efficacy for some of the diseases mentioned above. 28−30 Consistent with their molecular role as chaperones, the σRs interact with various proteins, modifying their function. 31,32 Concerning their active modulatory activity in pain signaling, these protein targets include the μ-opioid receptor, ion channels, and the NMDA receptors. 14 Significantly, σ 1 R antagonists block the activity of the σ 1 Rs, producing increased opioid and decreased NMDA receptor signaling, thereby enhancing antinociception by opioids and decreasing the hypersensitivity commonly associated with pathological pain. 33 Interestingly, an increased number of ligands possessing different chemotypes and heterogeneous σRs binding profiles showed a significant antinociception effect in different preclinical in vivo pain models ( Figure 1). 34−41 Continuing our efforts to discover selective σ 1 R ligands, 42,43 in this paper, a set of new benzylpiperazines was synthesized and pharmacologically characterized for their analgesic effects in mice models of pain. Previously, we developed a series of bifunctional σRs ligands with in vitro antioxidant properties. 44 These ligands were obtained by combining a preferred σR cyclic amino moiety, such as benzylpiperazine, with the 1,2dithiolan-3-yl moiety belonging to the natural antioxidant compound α-lipoic acid (Figure 2A). The 4-methoxybenzylpiperazinyl derivative 8 was previously identified as a potent and selective ligand for the σ 1 R over σ 2 R/TMEM97 (Figure 2A). Moreover, previous structure−affinity relationships (SAfiRs) suggested that the introduction of a para-substituent at the secondary hydrophobic domain (HYD2) improved the affinity and selectivity at σ 1 R. 44  both the chain linker and the primary hydrophobic domain (HYD1) was performed. With this in mind, we used compound 8 as our lead molecule to develop new benzylpiperazine derivatives as potentially more potent and selective σ 1 R ligands over σ 2 R ( Figure 2B). Similar to 8, the newly synthesized compounds fulfilled Glennon's pharmacophore model in which two distal hydrophobic regions and a central positive ionizable nitrogen give the essential features for the binding at σ 1 R (Figure 2A,B). 45 Due to the promising outcomes obtained with previous benzylpiperazines, in this new series, we maintained the 4methoxybenzylpiperazinyl moiety as the HYD2 and we modified the other distal hydrophobic region and the linker portion (i.e., 13−16 and 20−22). Specifically, the substitution of a phenyl, phenoxy, or cyclohexyl group in place of a lipoyl one as the HYD1 was carried out. Additionally, to explore the importance of an additional H-bond donor group for the target binding, the 4-methoxy substituent was replaced with a hydroxymethyl one (24). according to the two pathways reported in Scheme 1. In the first case, acids 9−12 were activated by a reaction with 1,1′carbonyldiimidazole (CDI) in dry dichloromethane (DCM) to give acyl imidazole intermediates (not isolated), which were then reacted with 1-(4-methoxybenzyl)piperazine in a protective nitrogen atmosphere to afford final amides 13−16 in good yields (36−68%). Amides 20−22 were obtained directly by the coupling of 1-(4-methoxybenzyl)piperazine with the corresponding organic halides (17−19) in dry tetrahydrofuran (THF) and using triethylamine as a base catalyst.
The final compound 24 was prepared following the two-step reaction depicted in Scheme 2. The acid derivative 9 was activated by CDI and converted into amide intermediate 23 by reaction with piperazine. Subsequently, treatment with the commercially available 4-(chloromethyl)benzyl alcohol, with K 2 CO 3 and KI, and using microwaves (MW) irradiation gave the final benzylpiperazine derivative 24 in excellent yield (90%).
Final compounds were characterized as a free base and submitted as such for the in vitro binding assay. Compound 15 has been converted into oxalate salt for in vivo behavioral studies.
Regarding the σ 1 R affinity, a clear trend based on the length of the linker chain between the distal phenyl ring and the central amide group can be observed in analogues 13−14, 16, 21, and 22 (i.e., ethylene > vinylene ≃ propylene > butylene ≫ methylene). Indeed, chain shortening from four (13) to two methylene units (21) led to an increase in σ 1 R affinity (K i σ 1 = 18.1 and 8.8 nM, respectively); however, a further shortening to only one methylene unit (22) was detrimental (K i σ 1 = 145 nM). Thus, the length of the ethylene chain in compound 21 produced optimal σ 1 R affinity (K i = 8.8 nM) and selectivity (370-fold) among this set. A phenyl ring instead of a lipoyl one was tolerated concerning the HYD1 domain, although a loss of selectivity was observed (8 vs 13), whereas the introduction of a phenoxy group (20) gave the worst result. Further replacement with a cyclohexyl ring resulted in a remarkable improvement of both the affinity and selectivity for σ 1 R (15 vs 21).
Finally, an additional H-bond donor group (24) in the secondary hydrophobic domain improved neither the σ 1 R affinity nor the selectivity significantly (24 vs 8). Compound 15, which showed the best binding profile among the series (K i σ 1 = 1.6 nM; K i σ 2 = 1,418 nM; K i σ 2 /K i σ 1 = 886), was selected for a more in-depth in vivo pharmacological evaluation.
Focally Induced Inflammatory Nociception. Due to the lack of reliable in vitro assays to establish the agonist/ antagonist properties of σRs ligands, the intrinsic functional activity of 15 was assessed by using a behavioral model of nociception. Consistent with the evidence of σ 1 R modulation of nociceptive signaling, 46 σ 1 R antagonists ameliorate pain responses in a focally induced inflammatory nociception model such as the formalin assay. 15,47,48 Compound 15 showed significant dose-dependent antinociception in this assay ( Figure 3A), consistent with action as a putative σ 1 R antagonist.
Compound 15 showed a similar efficacy at the highest dose in reducing time spent licking the injected paw compared to the positive control CM304 (1), a well-characterized selective Focally Induced Neuropathy. On the basis of the observed antinociceptive effect exerted by 15 in the formalin assay, we further characterized 15 in a representative model of neuropathic pain. We selected the chronic constriction injury (CCI) model as a widely used and validated assay to produce   Figure 4), with significant increases in withdrawal thresholds at 20, 60, and 80 min post-injection of a 60 mg/kg, i.p. dose (p < 0.05; Tukey's post hoc test). These effects were comparable to the results of the positive control gabapentin ( Figure 4). CM304, the reference selective σ 1 R antagonist, demonstrated anti-allodynic effects that are comparable to those of gabapentin in a timedependent manner. The anti-allodynic effects of CM304 peaked at 40 min but began to diminish at 60 min. The current results are consistent with a mechanistic interpretation of anti-allodynia through the σ 1 R antagonism. Notably, CCI produces a focal injury of the sciatic nerve that has been demonstrated to enhance the labeling of spinal σ 1 Rs in a manner enhancing nociceptive signaling. 55 Currently, these studies with compound 15 verify previous findings stating that noxious stimuli are attenuated by σ 1 R antagonists. 35,48,56 Induced Locomotor Activity. Treatments for neuropathic pain may be complicated by concordant sedation and impairment of motor function, as demonstrated by gabapentin. 57 To eliminate the potential complication of impaired locomotion or sedation, the effect of compound 15 on elicited locomotor activity was assessed using the rotarod assay. 35 The positive control and κ-opioid receptor agonist trans-(±)-3,4dichloro-N-methyl-N-[2-(1-pyrrolidinyl)cyclohexyl]benzeneacetamide hydrochloride (U50,488, 10 mg/kg, i.p.) significantly impaired evoked locomotor activity compared to the vehicle control (F (4, 301) = 26.02; p < 0.0001; two-way ANOVA with Tukey's post hoc test; Figure 5) up to 60 min after administration. In contrast, at doses proving effective in the pain assays, compound 15 did not significantly impair evoked locomotor activity, although the 30 mg/kg, i.p. dose produced a singular increase in locomotor performance 60 min post-administration (p = 0.003). Although the mechanism of σRs involvement in motor coordination and sedation has not yet been fully defined, the current results confirm the recent finding by Cirino et al., 35 showing that selective σ 1 R antagonists fail to produce sedative effects or impair evoked locomotor activity in rodents, confirming their analgesic properties.

■ CONCLUSIONS
This work has described the design and synthesis of new benzylpiperazinyl derivatives possessing high affinities for σ 1 R   (K i σ 1 = 1.6−145 nM) and selectivity over σ 2 R (K i σ 2 /K i σ 1 = 43−886). Following Glennon's structural features criteria necessary for σ 1 R binding, we discovered compound 15 as a potent and selective σ 1 R ligand. Especially, the use of hydrophobic cyclohexyl or phenyl groups and the 4methoxybenzylpiperazinyl moiety (HYD1 and HYD2, respectively) linked by three-carbon units linker (i.e., 15, 16, and 21) was an excellent combination to obtain optimal σRs binding profiles. Importantly, behavioral pharmacology studies showed that 15 produced significant antinociceptive and anti-allodynic effects in preclinical mouse models of pain without impaired locomotor activity, supporting the development of benzylpiperazine-based σ 1 R antagonists as potential therapeutics for chronic pain. The signal multiplicities are characterized as s (singlet), d (doublet), t (triplet), or m (multiplet). Microwave irradiation experiments were carried out with a CEM Discovery instrument using closed Pyrex glass tubes with Teflon-coated septa. Thin-layer chromatography (TLC) on Merck plates (aluminum sheet coated with silica gel 60 F 254 ) was used to monitor the progress of reactions and to test the purity (≥95%) of all the synthesized compounds, and spots were visualized under UV (λ = 254 and 366 nm) or in an iodine chamber. The purification of synthesized compounds by column chromatography was performed using Merck silica-gel 60 (230−400 mesh). All chemicals and solvents were purchased from commercial vendors and were of reagent grade.
General Procedure for the Synthesis of 4-(Methoxyphenyl)methylpiperazine Derivatives 13−16. 1,1′-Carbonyldiimidazole (1.0 equiv) was mixed to a stirred solution of suitable acid 9−12 (1.0 equiv) in dry DCM (6 mL) at room temperature. Then, after no gas evolution was observed, the mixture thus obtained was added dropwise to a stirred solution of 1-(4-methoxybenzyl)piperazine (1.1 equiv) in dry DCM (6 mL) at 0°C under a nitrogen atmosphere. The reaction was carried out for 30 min at 0°C and then for 1−2 h at room temperature. The mixture was washed with 10% aqueous NaCl solution (4 × 10 mL) and H 2 O (2 × 10 mL). The organic layer was dried over anhydrous sodium sulfate and then evaporated under reduced pressure to obtain a crude, which was purified as specified for each final product.
1-{4-[(4-Methoxyphenyl)methyl]piperazin-1-yl}-5-phenylpentan-1-one (13). The yellow crude oil was purified by flash column chromatography using ethyl acetate/methanol (9.5:0.5, v/v) as an eluent to afford 13 (0.547 g, 60 General Procedure for the Synthesis of 4-(Methoxyphenyl)methylpiperazine Derivatives 20−22. A mixture of 1-(4methoxybenzyl)piperazine (1.0 equiv) and triethylamine (1.0 equiv) in dry THF (5 mL) was prepared and left under stirring for 10 min at 0°C. Subsequently, the appropriate acyl chloride (17−19, 1.0 equiv) was added to the obtained solution, and the reaction was carried out at 0°C for 30 min and then at room temperature for 1−3 days. At the end of the reaction time, the solvent was evaporated to dryness under a vacuum. The crude product thus obtained was solubilized in DCM and then washed with a water solution of Na 2 CO 3 0.1 M (2 × 20 mL) and NaCl 10% (20 mL). The organic layer was dried over anhydrous sodium sulfate and evaporated under reduced pressure to obtain a residue, which was purified as specified for each final product.
Behavioral Pharmacology. Animals. Adult male C57BL/6J and CD-1 mice housed five to a cage (8−12 weeks of age) were used. C57BL/6J mice were used for evoked locomotor rotarod and formalin assays. 35,60,61 Antinociception was confirmed with the use of CD-1 mice in the CCI nerve assay. The CD-1 strain has been well-validated for antinociceptive 62 and mechanical anti-allodynic testing. 63,64 All test compounds were administered using the intraperitoneal (i.p.) route. All animal studies reported herein adhere to ARRIVE guidelines. 65 Animals were randomly assigned, and researchers were blinded to group treatments. Animals were housed on a 12:12 h light/ dark cycle (lights off at 7:00 pm) with ad libitum access to food and water except during experimental sessions. All procedures were preapproved by the Institutional Animal Care and Use Committee (University of Florida) and conducted in accordance with the 2011 NIH Guide for the Care and Use of Laboratory Animals.
Formalin Test. The efficacy of compound 15 to ameliorate inflammatory nociception was achieved with the use of C57BL/6J mice in the formalin assay as previously described. 53 After a 10 min pretreatment (i.p.) of vehicle control (saline), CM304 (3−30 mg/kg, i.p.), or compound 15 (3−30 mg/kg, i.p.), an intraplantar (i.pl.) injection of 5% formalin (2.5 μg in 15 μL) was administered into the right hind paw. Time spent licking the right hind paw was recorded in 5 min intervals for 60 min following injection. The last 55 min of assessment was used to determine the inflammatory response stimulus. Data were analyzed as the summed duration of licking hind paw.
CCI Assay. CCI was introduced in CD-1 mice that were first anesthetized with isoflurane as described by Hoot et al. 66 and Cirino et al. 35 to induce mechanical allodynia. 51−54 After anesthetization, mice were subjected to surgery where an incision was made along the surface of the biceps femoris of the right hind paw. 66 Blunt forceps were used to split the muscle and expose the right sciatic nerve. The tips of two 0.1−10 μL pipet tips facing opposite directions were passed under the sciatic nerve to allow for easy passing of two sutures under the nerve, 1 mm apart. The sutures were tied loosely around the nerve and knotted twice, and the skin was closed with 29 mm skin staples. Mice were given a 7 day recovery period prior to the baseline von Frey testing as described below to confirm the induction of hyperalgesia in each mouse. Animals demonstrating allodynia or a response to lower pressure were deemed to have neuropathic pain. Allodynic mice were then administered (i.p.) either vehicle (saline), morphine (10 mg/kg, i.p.), gabapentin (50 mg/kg, i.p.), CM304 (45 mg/kg, i.p.), or compound 15 (10−60 mg/kg, i.p.). Note that gabapentin was tested 1 h postinjection to circumvent known sedative effects that may confound the assay. 67 Each mouse was then tested for the modulation of tactile allodynia every 20 min up until 80 min postinjection with the use of von Frey testing. The assessment of mechanical allodynia was performed to measure compound 15's efficacy against CCI-induced allodynia as described. 51−54 Mice were habituated on a mesh platform for 1 h prior to testing. Filaments of increasing pressure (0.4−6 g) were applied and then held to the plantar surface of both the injured and uninjured hind paws of mice for approximately 1−2 s prior to drug administration to record baseline responses to a peripheral stimulus. The filaments were applied with increasing strengths, and threshold responses were defined as two hind paw responses per trial of the same filament strength.
Control or test compounds were administered (i.p.), and pawwithdrawal thresholds were again recorded from 20 to 80 min postinjection. Each hind paw was tested in a counterbalanced manner. Each time was measured in triplicate and then averaged.
Responsiveness was a clear withdrawal, shaking, or licking of the paw. To account for variability between mice, data are presented as the percent of baseline paw withdrawal thresholds following filament stimulation of the ipsilateral hind paw. The following equation was used: % anti-allodynia = 100 × ([mean paw withdrawal force {g} in control group − paw withdrawal force {g} of each mouse]/mean paw withdrawal force [g] in control group).
Rotarod Assay. The rotarod coordination assay was used to assess effects on evoked locomotor activity in C57BL/6J mice administered vehicle (saline, i.p.), morphine (10 mg/kg, i.p.), U50,488 (10 mg/kg, i.p.), CM304 (45 mg/kg, i.p.), or compound 15 (30−60 mg/kg, i.p.) using methods described previously. 68,69 Seven habituation trials were performed where the last habituation trial was used as an initial baseline of performance. The mice were administered (i.p.) test agents and then evaluated every 10 min in accelerated speed trials (180 s max latency at 0−20 rpm) over a 60 min period. The latency to fall was measured in seconds. Data are reported as the mean percent change from each mouse's initial baseline latency to fall. Decreased latencies to fall in the rotarod test indicate impaired motor coordination or sedation Statistical Analysis. All data are presented as mean ± SEM. Significance is indicated as *p < 0.05 and was analyzed using two-way ANOVA with Tukey's post hoc analysis. Statistical analysis was performed with the use of GraphPad Prism 9.0 software. Dose response lines were analyzed by linear or nonlinear regression modeling and ED 50 values (dose yielding 50% effect) along with 95% confidence limits using each individual data points. The rotarod data are expressed as the % change from baseline performance for each animal's baseline response.