Identification of an Orally Bioavailable, Brain-Penetrant Compound with Selectivity for the Cannabinoid Type 2 Receptor

Modulation of the endocannabinoid system (ECS) is of great interest for its therapeutic relevance in several pathophysiological processes. The CB2 subtype is largely localized to immune effectors, including microglia within the central nervous system, where it promotes anti-inflammation. Recently, a rational drug design toward precise modulation of the CB2 active site revealed the novelty of Pyrrolo[2,1-c][1,4]benzodiazepines tricyclic chemotype with a high conformational similarity in comparison to the existing leads. These compounds are structurally unique, confirming their chemotype novelty. In our continuing search for new chemotypes as selective CB2 regulatory molecules, following SAR approaches, a total of 17 selected (S,E)-11-[2-(arylmethylene)hydrazono]-PBD analogs were synthesized and tested for their ability to bind to the CB1 and CB2 receptor orthosteric sites. A competitive [3H]CP-55,940 binding screen revealed five compounds that exhibited >60% displacement at 10 μM concentration. Further concentration-response analysis revealed two compounds, 4k and 4q, as potent and selective CB2 ligands with sub-micromolar activities (Ki = 146 nM and 137 nM, respectively). In order to support the potential efficacy and safety of the analogs, the oral and intravenous pharmacokinetic properties of compound 4k were sought. Compound 4k was orally bioavailable, reaching maximum brain concentrations of 602 ± 162 ng/g (p.o.) with an elimination half-life of 22.9 ± 3.73 h. Whether administered via the oral or intravenous route, the elimination half-lives ranged between 9.3 and 16.7 h in the liver and kidneys. These compounds represent novel chemotypes, which can be further optimized for improved affinity and selectivity toward the CB2 receptor.


Introduction
Neurodegenerative disorders are a widespread cause of morbidity and mortality worldwide, characterized by the slow, progressive damage/loss of neurons in the central nervous system (CNS), which is associated with deficits in function (e.g., movement, memory, cognition) that are related to the affected CNS region(s) [1]. The progression of many neurodegenerative diseases is thought to be driven by the template-directed misfolding, seeded aggregation, and cell-cell transmission of characteristic disease-related proteins, leading to the consecutive spreading of pathological protein aggregates [2]. Modulating the activity of endocannabinoids (ECB) has held therapeutic promise for treating a wide range of diseases associated with inflammation and oxidative stress. The cannabinoid receptors subtype 2 (CB2), which were identified molecularly in 1993, have been the subject of considerable attention, primarily due to their promising anti-inflammatory potential without the adverse psychotropic effects more commonly associated with CB1 receptorbased therapies [3]. Many studies have further corroborated the in vivo link between chronic neuroinflammation and CB2 upregulation in animal models of pain [4] and inflammation [5]. These studies raise the possible logical connection between CB2 receptors and immunological function [6]. Activation of CB2 receptors by ligands favors a range of receptor conformations that can affect different signaling pathways [7].
Administration of the CB2 antagonist was able to decrease the neuroprotective effects of beta-caryophyllene (BCP) and strengthen the need for CB2 activation. The distribution of the CB2 subtype in neural cell types (activated astrocytes, reactive microglia, oligodendrocytes, perivascular microglia, and neural progenitor cells), many of which are integral to blood-brain barrier function, supports its potential role in CNS integrity [7]. CB2 agonists have shown promising neuroprotection using in vitro (BV-2 microglial cell lines and primary microglia cultures) and in vivo models (murine model) of neurodegeneration. The potential of CB2 agonists as neuroprotective agents has been particularly sought given the lack of psychotropic adverse effects generally seen with CB1 agonists [8]. One of the limitations of the small molecule modulators of CB2 has been issues with selectivity and differential activity depending on cell type/species. Several 'selective' CB2 agonists have been shown to regulate orphan cannabinoid receptors and/or opioid receptors or other off-target proteins. In addition, recent evidence suggests that inverse agonists of CB2 may also be neuroprotective. A synthetic CB2 inverse agonist protects neurons from excitotoxic insult in vitro and in vivo [9]. SMM-189 prevented the rat primary microglia-mediated inflammation and was neuroprotective against N-methyl-D-aspartic acid (NMDA)-induced neuronal excitotoxicity in brain-mimicking neuron-glia primary cultures obtained from rat hippocampal tissues [9]. Concurrent development of novel, selective CB2 inverse agonists, partial, and full agonists would allow for a closer comparison of the effects of these two mechanistically distinct pharmacological effects so the role of CB2 in neuroinfectious diseases can be elucidated.
Novel synthetic and natural compounds with neuroprotective potential have emerged as new pharmaceuticals in the management of many neurological conditions, presenting a heterogeneous chemical structure and varied pharmacological activity, which may be useful in the treatment of high-complexity diseases. The search for new drug leads that target endocannabinoid signaling is challenging as endocannabinoids are implicated in various physiological and pathological processes.
Pyrrolo [2,1-c][1,4]benzodiazepines (PBDs) are a class of natural products known to possess anti-tumor and antibiotic activities [10][11][12]. Molecular modeling studies have shown that the PBD tricyclic structure is well-suited to dock within the catalytic cavity of CB receptors, allowing them to interact with the endocannabinoid signaling system. As a part of our continuing efforts to identify novel neuroprotective agents through a structure-based rational design and multi-step synthetic approach [13], the present project focused on the synthesis of seventeen analogs of (S,E)-11-[2-(arylmethylene)hydrazono]-PBD ligands (4a-4q) as potential CB2 selective regulators that may serve as novel and effective neuroprotectants (Figure 1). Discovery of novel regulatory compounds using structure-activity relationship (SAR) studies will advance the field of neuroprotection and neurorestoration by providing information necessary to determine if this class of compound is a viable chemotype for the design of potentially new therapies to treat neurodegenerative disorders in humans. Drug discovery programs routinely perform pharmacokinetic (PK) studies in mice to prioritize lead compounds based on anticipated exposure-efficacy and exposure-toxicity relationships. Promising lead compounds with selective binding affinities were examined for their pharmacokinetics properties in plasma or tissue (brain, liver, or kidney) following oral or intravenous administration.

General Experimental Procedures
The 1 H and 13 C NMR spectra were recorded in DMSO-d6 and CDCl3 on Bruker 400 and 500 MHz spectrometer operating at 400, 500 MHz for 1 H and 100, 150 MHz for 13 C NMR. Chemical shift (δ) values are presented in ppm and in reference to the residual solvent signals of DMSO-d6 and CDCl3 at δH/δC 2.50/39.5 and 7.25/70.2, respectively. The coupling constants value (J) reported in Hz.
HPLC-grade acetonitrile and water solvents were used. Acetic acid was added as a modifier to achieve a final concentration of 0.1% in each solvent.
Preparative HPLC purification was carried out using an Agilent 1100 HPLC system, RP-C18 column (250 × 10 mm; particle size 10 μm; Luna) with column oven temperature set at 25 °C and a gradient system of eluent water (A) and acetonitrile (B). The gradient Drug discovery programs routinely perform pharmacokinetic (PK) studies in mice to prioritize lead compounds based on anticipated exposure-efficacy and exposure-toxicity relationships. Promising lead compounds with selective binding affinities were examined for their pharmacokinetics properties in plasma or tissue (brain, liver, or kidney) following oral or intravenous administration.

General Experimental Procedures
The 1 H and 13 C NMR spectra were recorded in DMSO-d 6 and CDCl 3 on Bruker 400 and 500 MHz spectrometer operating at 400, 500 MHz for 1 H and 100, 150 MHz for 13 C NMR. Chemical shift (δ) values are presented in ppm and in reference to the residual solvent signals of DMSO-d 6 and CDCl 3 at δ H /δ C 2.50/39.5 and 7.25/70.2, respectively. The coupling constants value (J) reported in Hz.
LC analysis was conducted using an Agilent 1100 HPLC system, RP-C18 column (150 × 4.6 mm; particle size 5 µm; Luna) with column oven temperature set at 25 • C and a gradient system of eluent water (A) and acetonitrile (B). The gradient condition was as follows: 0-2 min (5% B), 2-5 min (5% B→50% B), 5-10 min (50% B→100% B), 10-15 min (100% B). The flow rate of the solvent was 1.0 mL/min, and the injection volume was 25 µL. All analysis was carried out at wavelength of 254 nm with a run time of 15 min. HPLC-grade acetonitrile and water solvents were used. Acetic acid was added as a modifier to achieve a final concentration of 0.1% in each solvent.
Preparative HPLC purification was carried out using an Agilent 1100 HPLC system, RP-C18 column (250 × 10 mm; particle size 10 µm; Luna) with column oven temperature set at 25 • C and a gradient system of eluent water (A) and acetonitrile (B). The gradient condition was as follows: 0-2 min (5% B), 2-5 min (5% B→50% B), 5-10 min (50% B→100% Molecules 2022, 27, 509 4 of 16 B), 10-15 min (100% B). The flow rate of the solvent was 3.0 mL/min, and the injection volume was 25 µL. All analysis was carried out at wavelength of 254 nm with a run time of 15 min. HPLC-grade acetonitrile and water solvents were used. Acetic acid was added as a modifier to achieve a final concentration of 0.1% in each solvent.
Other common chromatographic techniques, such as thin layer chromatography (TLC) on precoated silica gel G 254 aluminum plates and silica gel flash column chromatography, were also engaged in the purification of the synthesized compounds.  Starting material 1-methylimidazole-2-carboxaldehyde (95.8 mg, 0.87 mmol) was used and the reaction mixture was stirred under nitrogen gas overnight for 15 h. Extraction was performed using chloroform/isopropanol (2:1) (3 × 20 mL). The combined organic layers were dried over anhydrous Na 2 SO 4 . The solvent mixture was then removed in vacuo and washed with diethyl ether, filtered off, and dried to afford an off-white solid of 4b.

Cannabinoid Receptor Binding Assay
The affinities of the compounds for CB1 and CB2 receptors were examined using displacement assays, as previously described [13]. Briefly, cell membranes from CHO cells expressing human CB1 or human CB2 receptors were isolated using differential centrifugation. Test compounds reconstituted in DMSO and were incubated with the isolated membrane in binding buffer (50 mM Tris-HCl, 1 mM EDTA, 3 mM MgCl 2 , 5 mg/mL BSA, pH 7.4) along with 2.5 nM [ 3 H]CP-55,940. Total binding was assessed in the presence of equal concentration of DMSO, while non-specific binding was determined in the presence of 10 µM CP-55,940, and background binding was determined in wells lacking membrane. Following incubation at 30 • C for 60 min, the binding reactions were terminated by vacuum filtration through Whatman GF/C filters. The filters were then washed twice with ice-cold buffer (50 mM Tris-HCl, 1 mg/mL BSA). Liquid scintillation cocktail was added to each well and the total tritiated counts per minute were analyzed using a TopCount scintillation counter. Background counts were subtracted from all wells and the percent displacement from total binding was calculated.
The compounds were initially screened at 10 µM concentrations. If they produced at least ±30% displacement of the radioligand, then full competition curves were constructed. K i values were calculated using GraphPad Prism (San Diego, CA, USA) and K d values determined using a 1 site fit. All assays were run in technical and biological replicates so that the n = 5-6.

Preparation of Stock Solutions, Calibration Standards, Quality Control Samples and Internal Standard Solution
The primary stock solutions of 4k and phenacetin (internal standard, IS) were prepared in methanol at a concentration of 1.0 mg/mL. Working solutions of calibration standards and quality control (QC) samples were prepared by dilution with methanol and stored at −20 • C. A working stock of the IS solution (20 ng/mL) was prepared in methanol and stored at −20 • C.

Instrument and Analytical Conditions
Chromatography was performed on an Acquity TM UPLC system (Waters Corp, Milford, MA, USA) with an autosampler temperature at 10 • C. Waters Acquity UPLC ® HSS C18 column (3.0 × 50 mm, 1.8 µm particle size) was used for chromatographic separation with linear gradient elution consisting of (A) 90% acetonitrile and (B) 10% 0.2% formic acid in Milli-Q water as mobile phases. The flow rate was set at 0.30 mL/min, and the injection volume was 2 µL.
An Acquity Tandem Quadrupole Mass Detector (Xevo TQ-S; Waters Corp, Milford, MA, USA) in positive electrospray ionization mode was used for mass spectrometric detection. For collision-induced dissociation, argon was used as collision gas. The cone voltage and collision energy were set at 60 V and 34 V for 4k and 46 V and 26 V for the IS, respectively. Quantification was performed using the monitoring of multiple reaction monitoring (MRM) of following transitions: m/z 409.0/220.9 for 4k and m/z 180.0/92.7 for IS. Retention times of 4k and IS were 2.85 and 1.81 min, respectively.

Sample Preparation
A simple protein precipitation method was followed for extraction of 4k from mice plasma. To an aliquot of 50 µL of plasma or tissue (brain, liver, or kidney) samples, IS solution (5 µL of 20 ng/mL) was added and mixed for 15 s on a cyclomixer (Thermo Scientific, Indianapolis, IN, USA). After precipitation with 200 µL of acetonitrile, the mixture was vortexed for 2 min, followed by centrifugation for 10 min at 14,000 rpm on an accuSpin Micro 17R (Fisher Scientific, Suwanee, GA, USA) at 5 • C. An aliquot of~150 µL of clear supernatant was transferred into vials and 2 µL was injected onto LC-MS/MS system for analysis.

In Vivo Studies in CD-1 Mice
All work involving animal subjects was pre-approved by the University of Mississippi Institutional Animal Care and Use Committee and was conducted in accordance with ethical guidelines defined by the National Institutes of Health (NIH Publication No. 85-23).

Subjects
Male CD-1 mice (N = 48) were obtained from Envigo (St. Louis, MO, USA) and maintained on a 12:12 h reversed dark/light cycle (lights off at 07:00 h) with ad libitum access to food and water. Following a~4 h fast, animals were randomly assigned to one of two groups. Group I and II animals (n = 24/group, weight range 25-30 g) received compound 4k (5 mg/kg) orally (in the form of a solution, prepared using 10% absolute alcohol, 10% cremophor, and 80% Milli-Q water) or intravenously (i.v., using solution formulation comprising 10% absolute alcohol, 10% cremophor, and 80% normal saline), respectively. At 0.5, 1, 2, 4, 8, and 24 h post administration, animals were euthanized, blood was collected in K 2 . EDTA-containing polypropylene tubes and, tissues were collected (brain, liver, and kidney). Plasma was harvested by centrifuging the blood using Eppendorf 5430R Centrifuge (Germany) at 5000 rpm for 5 min and stored frozen at −80 • C until analysis.

Tissue Preparation
Brain, liver, and kidney tissues were homogenized in separate 15 mL round-bottom screw-capped vials in phosphate buffered saline (5 volumes of each tissue mass) with a homogenizer (Polytron ® ) and stored at −80 • C until analysis. Plasma or tissue homogenates (50 µL) were spiked with IS and processed as mentioned in sample preparation section.

Pharmacokinetic Assessments
Plasma and tissue concentration and time data of compound 4k were analyzed by a non-compartmental method using WinNonlin Version 5.3 (Pharsight Corporation, Mountain View, CA, USA).

Chemistry
To gain a better understanding of the structure-activity relationships of 4, we extended the range of PBD-11-hydrazinyl derivatives (4a-4q) by alkenylation of the primary amine group of the cycloamidine 3 via the formation of a Schiff's base with various carboxaldehydes [13]. Synthetic approaches were initiated based on the stepwise synthesis of dilactam 1 by the cyclocondensation of an equimolar mixture of L-proline and readily available basic structure of PBD natural product, isatoic anhydride (or substituted compound) in DMF at 155 • C. Thionation of compound 1 with 0.5 eq. of 2,4-bis-(4-methoxyphenyl)-1,3dithia-2,4-diphosphetane-2,4-disulfide (Lawesson's reagent) in THF at room temperature produced thiolactam 2. The subsequent treatment of 2 with 98% hydrazine monohydrate in ethanol at room temperature generated cycloamindine 3. The title compound 3 was further subjected to condensation with several aldehydes in anhydrous MeOH and molecular sieves (3 Å) at room temperature to afford a cluster of highly conjugated Schiff's base 4 in high yield. This procedure involves the initial formation of the intermediate carbinol amine, which dehydrates to form an imine. The reaction time was varied from a few minutes to a couple of hours and gave the desired products in good yields. The alkenylation of 3 with aldehydes was best performed under optimized conditions in which the parent molecule was exposed to five equivalents of aldehyde in methanol for varied reaction times at room temperature. Upon further crystallization of crude product from EtOAc/hexanes, pure (S,E)-11-[2-(arylmethylene)hydrazono]-PBD analogs (4a-4q) were formed as crystalline solids. The chemical structures are shown in Figure 2. These compounds are structurally unique from any reported small-molecule CB ligands, confirming the chemotype novelty of all the generated compounds.

Cannabinoid Receptors Displacement Assay
The CB1/CB2 receptor binding activities of analogs 4a-4q were assessed via radioactivitybased competitive binding assays using [ 3 H]CP55,940, an extensively studied radioligand that is frequently used for cannabinoid assays [14][15][16]. The highly potent and nonselective CB agonist CP55,940 binds to the same orthosteric active site where known CB agonists are known to bind. In preliminary screening, all 17 synthetic PBD compounds (4a-4q) were subjected to in vitro CB1 and CB2 binding assays at a single concentration of 10 µM. The observed percentage displacement (%) of the radioligand at the CB receptors of these analogs are summarized in Table 1. From the 17 compounds evaluated in the competitive radioligand assay, compounds 4h, 4j-4l, and 4q showed significant displacement (more than 60%). Among them, two structurally distinct PBD analogs (4k and 4q) displayed the greatest [ 3 H]CP-55,940 displacement and selectivity for CB2 over CB1. Therefore, the most promising ligands (4k and 4q) were selected for assessment in full competition curves against CB1 and CB2. The binding assays revealed selective binding affinity of 4k and 4q with K i values of 146 and 137 nM, respectively, toward CB2 receptors (Table 2).  Drug discovery research programs routinely execute mouse PK studies to characterize the PK properties of compounds as a filtering tool to advance drug candidates through the pipeline as well as to support efficacy and toxicology studies. Stability studies on two selected analogs has revealed a better tolerance of 4k under physiological pH when compared to 4q compound (not reported). Consequently, 4k was selected for pharmacokinetic studies in CD1 male mice at 5 mg/kg body weight (Table 3, Figure 3).  Drug discovery research programs routinely execute mouse PK studies to characterize the PK properties of compounds as a filtering tool to advance drug candidates through the pipeline as well as to support efficacy and toxicology studies. Stability studies on two selected analogs has revealed a better tolerance of 4k under physiological pH when compared to 4q compound (not reported). Consequently, 4k was selected for pharmacokinetic studies in CD1 male mice at 5 mg/kg body weight (Table 3, Figure 3). Drug discovery research programs routinely execute mouse PK studies to characterize the PK properties of compounds as a filtering tool to advance drug candidates through the pipeline as well as to support efficacy and toxicology studies. Stability studies on two selected analogs has revealed a better tolerance of 4k under physiological pH when compared to 4q compound (not reported). Consequently, 4k was selected for pharmacokinetic studies in CD1 male mice at 5 mg/kg body weight (Table 3, Figure 3). Table 3. Pharmacokinetic parameters of compound 4k in CD1 mice following intravenous and oral dosing of 4k in mice at 5 mg/kg.  In the kidney, the concentrations of compound 4k decreased mono-exponentially after 5.0 mg/kg intravenous administration. The mean kidney clearance (CL) was found to be 6.38 ± 0.49 mL/min/kg. Compound 4k had a high volume of distribution of 5.63 ± 0.37 L/kg in the kidney. The terminal half-life (t1/2) was found to be 10.2 ± 0.49 h. Post oral administration, the maximum kidney concentrations (Cmax: 795 ± 22.65 ng/g) were attained at 0.75 ± 0.15 h (Tmax). The apparent half-life was found to be 16.7 ± 1.61 h. The AUC0-∞ attained post oral dose was 5223 ± 44.75 ng × h/g. The apparent oral bioavailability in the kidney at 5 mg/kg was 37 ± 3.78% (Table 1). The mean kidney to plasma ratios after intravenous and oral administration were 14.5 and 7.25, respectively.

PK Parameters Intravenous Oral
In the liver, the mean liver clearance (CL) was found to be 4.28 ± 0.38 mL/min/kg. Compound 4k had a high volume of distribution of 3.38 ± 0.18 L/kg in liver. The terminal half-life (t1/2) was found to be 9.26 ± 0.67 h. Post oral administration, the maximum liver concentrations (Cmax:1891 ± 282 ng/g) were attained at 0.75 ± 0.15 h (Tmax). The apparent In plasma, the concentrations of compound 4k decreased in a mono-exponential manner after 5.0 mg/kg intravenous administration. The mean clearance (CL) was found to be 23.5 ± 5.35 mL/min/kg, which is 26% of hepatic blood flow in mice. Compound 4k had a high volume of distribution of 14.3 ± 3.11 L/kg. The terminal half-life (t 1/2 ) was found to be 6.91 ± 0.15 h. The post oral administration, maximum plasma concentrations (C max : 198 ± 11.51 ng/mL) were attained at 0.88 ± 0.13 h (T max ), indicating rapid absorption from the gastrointestinal tract. The apparent half-life (16.4 ± 1.66 h) determined after oral administration was longer than that after intravenous administration (6.91 h), which may indicate multiple sites of absorption. The AUC 0-∞ attained post oral dose was 739 ± 6.75 ng × h/mL. The oral bioavailability in mice at 5 mg/kg was 18 ± 2.43% (Table 3).