Discovery and SAR Evolution of Pyrazole Azabicyclo[3.2.1]octane Sulfonamides as a Novel Class of Non-Covalent N-Acylethanolamine-Hydrolyzing Acid Amidase (NAAA) Inhibitors for Oral Administration

Inhibition of intracellular N-acylethanolamine-hydrolyzing acid amidase (NAAA) activity is a promising approach to manage the inflammatory response under disabling conditions. In fact, NAAA inhibition preserves endogenous palmitoylethanolamide (PEA) from degradation, thus increasing and prolonging its anti-inflammatory and analgesic efficacy at the inflamed site. In the present work, we report the identification of a potent, systemically available, novel class of NAAA inhibitors, featuring a pyrazole azabicyclo[3.2.1]octane structural core. After an initial screening campaign, a careful structure–activity relationship study led to the discovery of endo-ethoxymethyl-pyrazinyloxy-8-azabicyclo[3.2.1]octane-pyrazole sulfonamide 50 (ARN19689), which was found to inhibit human NAAA in the low nanomolar range (IC50 = 0.042 μM) with a non-covalent mechanism of action. In light of its favorable biochemical, in vitro and in vivo drug-like profile, sulfonamide 50 could be regarded as a promising pharmacological tool to be further investigated in the field of inflammatory conditions.


■ INTRODUCTION
Inflammation is a multifaceted, dynamic process inducing a set of reactive modifications that occur in the affected organs and vascular tissue to repair the damage produced by harmful agents or stimuli of various kinds. 1,2 Chronic inflammation is a long-lasting pathological process determined by the persistence of inflammatory stimuli, and is typically characterized by the infiltration of mononuclear cells (macrophages, lymphocytes, and plasma cells), and the simultaneous presence of tissue damage, angiogenesis, and fibrosis. 3 Chronic inflammatory diseases can be considered as a leading cause of death in the world today, with more than 50% of all deaths linked to some degree to inflammationrelated diseases (e.g., ischemic heart disease, cancer, diabetes mellitus, chronic kidney disease, stroke, and autoimmune and neurodegenerative conditions). 4 Among the possible ways to manage inflammation, the inhibition of the cysteine hydrolase N-acylethanolaminehydrolyzing acid amidase (NAAA) has been reported as a promising approach to control pain and inflammatory conditions. 5−8 NAAA, a member of the N-terminal nucleophile family of hydrolases (Ntn hydrolases), 9 is expressed at high levels in immune cells, mainly in macrophages, and localized in the lysosomal compartment. 6,10 The enzyme is produced as a precursor and activated by selfproteolysis at acidic pH (pH = 4.5−5) into two chains: αand β-subunits. In human NAAA (h-NAAA), this process exposes the enzyme's catalytic triad of nucleophilic Cys126, basic Arg142, and acidic Asp145 residues, producing a hydrolysis-competent enzyme. 11 NAAA is involved in the deactivating hydrolysis of Nacylethanolamines (NAEs), a group of endogenous lipid mediators comprising ethanolamine bound to a variable fatty-acid-derived acyl moiety. NAEs play an important role in the control of multiple physiological functions including the regulation of pain, 12 inflammation, 13 feeding behavior, 14,15 and the dopaminergic reward system. 16,17 Previous studies have shown that NAAA exhibits substrate selectivity predominantly for saturated NAEs, such as palmitoylethanolamide (PEA). 10 PEA is an anti-inflammatory, analgesic, and neuroprotective agent, mainly involved in the activation of peroxisome proliferator-activated receptor α (PPAR-α), a nuclear receptor, which plays a crucial role in biological processes enhancing the transcription of various antiinflammatory genes and concomitantly interrupting the activity of proinflammatory transcription factors. 18−20 PEA signaling activity is terminated by its degradation into ethanolamine and palmitic acid catalyzed preferentially by NAAA, and to a lesser extent by fatty acid amide hydrolase (FAAH), 21 a membrane-bound serine amidase biologically related to NAAA. Although NAAA and FAAH share a similar mechanism, catalyzing the hydrolysis of N-acylethanolamines, they have no sequence homology. On the contrary, NAAA displays high homology to acid ceramidases (AC), another lysosomal cysteine amidase, showing 33−35% amino acid identity in their structures. 9,22 From a drug discovery perspective, sustaining PEA levels through the inhibition of intracellular NAAA activity represents a promising approach to modulate the inflammatory response, by increasing and/or prolonging PEA's antiinflammatory and analgesic properties. Therefore, the use of NAAA inhibitors can provide a certain efficacy in the treatment of inflammatory conditions without causing the typical side effects due to generalized stimulation of PPAR-α, as with exogenous agonists. 23 Interestingly, it has been reported that alterations of PEA levels are found in inflammatory diseases, such as rheumatoid arthritis, osteoarthritis, 24 and in cerebrospinal and plasma fluids in MS patients. 25 Despite the encouraging pharmacological benefits achieved by restoring intracellular PEA levels through NAAA inhibition, only a few chemical classes of potent compounds have been identified (Figure 1). 26−31 Over the last decade, few research groups have been involved in the discovery of novel NAAA inhibitors. 7,8 In this context, our work led to the identification of low nanomolar inhibitors, featuring electrophilic warheads, as possible therapeutic agents suited for topical (ARN077 and analogues, Figure 1) 26,32 and/or systemic administration (ARN726 and analogues, Figure 1). 27,33 Depending on the nature of the small heterocyclic reactive moiety, experimental investigations showed for these chemotypes a covalent, partially reversible or irreversible mechanism of action both in vitro and in vivo. Lately, in order to avoid the possible drawbacks of covalent inhibition (i.e., idiosyncratic effects, dosing regimen, etc.), 34,35 the identification of non-covalent and systemically available NAAA inhibitors was recently reported and demonstrated to be beneficial in a MS mouse model (ARN19702, Figure  1). 31,36 Indeed, it was shown that NAAA contributed to disease progression in the experimental autoimmune encephalomyelitis (EAE) mouse model 37 because induction of the enzyme's expression was observed in the spinal cord of EAE-affected mice. 36 In the present work, we report on the discovery of potent, systemically available pyrazole azabicyclooctane sulfonamide derivatives, as a novel class of NAAA inhibitors, endowed with a non-covalent mechanism of action. A biological screening, by the use of a fluorogenic human NAAA assay, allowed identifying few primary hits featuring a similar chemotype. An in-depth structure−activity relationship (SAR) analysis led initially to the discovery of lead compound 39 (ARN16186, Figure 1) 38 with high inhibitory activity and good preliminary pharmacokinetic properties. 39 Structural modifications to improve the physicochemical and drug-like properties of 39 led eventually to the identification of endo-ethoxymethyl-pyrazinyloxy-8-azabicyclo[3.2.1]octanepyrazole sulfonamide 50 (ARN19689 , Table 4), a compound with a superior pharmacological and pharmacokinetic profile.
Due to its high inhibitory activity and selectivity for NAAA, supported by the in vitro biochemical and in vivo pharmacological data, novel azabicyclic compound 50 could represent a valuable tool to be further evaluated in the management of inflammatory conditions.

■ RESULTS AND DISCUSSION
In early drug discovery, the screening of mid-large compound collections is a key activity to help identifying and/or expanding the portfolio of novel, active chemotypes on biological targets. In an effort to discover novel non-covalent human NAAA (h-NAAA) inhibitors, a small tailored set of 1000 compounds was selected from our internal collection of commercially available small molecules. These compounds were identified based on diversity in structural and physicochemical properties, and were tested via a medium throughput screening (MTS) fluorogenic h-NAAA assay ( Figure 2).
The MTS campaign identified a few hits belonging to structurally diverse chemical classes, showing promising inhibitory activity in the micromolar range. An accurate analysis highlighted the presence of compounds featuring a similar chemotype, lacking highly reactive/electrophilic chemical moieties, and therefore potentially matching our purpose to discover non-covalent inhibitors. Ultimately, we decided to focus our optimization efforts on hit compound 1, 41 which showed, along with a promising, initial single-digit micromolar inhibitory activity against h-NAAA (IC 50 = 1.09 μM, Table 1), structural novelty, synthetic feasibility, and an encouraging lipophilic efficiency (LipE: 6.3) 42,43 (Figure 3). From a preliminary structural analysis, hit 1 was envisaged as a reasonable starting point for the identification of novel h-NAAA inhibitors, featuring a non-covalent mechanism of action, as it does not contain known reactive moieties. Driven by these initial observations, we developed a structure− activity relationship (SAR) study by means of iterative designing, synthesis, and biological characterization cycles of analogues of sulfonamide 1. Chemical investigations were undertaken exploring rationally the key structural regions in primary hit 1, that is, the type and the substitution pattern of heteroaromatic moiety (A), the functionalization and structural changes of piperidine (B), and finally the modifications of the heteroaryl group (C) (Figure 3).
We started our SAR investigation exploring the pyrazole region A of compound 1 (Table 1). To gain broad structural information, a variety of substituted aromatic and hetero-aromatic moieties were introduced as an alternative to the 3,5-dimethylpyrazole group.
Initially, we focused our study on 5-membered heteroaromatic rings having different substitution patterns. The effect of the 3,5-dialkyl substitution on the pyrazole ring resulted to be important for activity because the corresponding mono- (2) or unsubstituted (3) analogues turned out to be devoid of any activity with respect to dimethyl-pyrazole derivative 1 (Table 1). Similarly, the absence of any hydrogen bond donor, as for 1,3,5-trimethyl-(4), 1,2-isoxazolyl- (5), and regioisomeric 1,3-dimethyl-(6) analogues led to a complete loss of inhibitory effect, thus suggesting the crucial requirement of a hydrogen bond donor feature in that specific region of the protein pocket ( Table 1). The low inhibitory activity of the 1,3-dimethyl phenyl analog (7) could be ascribed to either the lack of a hydrogen bond donor or alternatively due to the different stereoelectronic properties of the phenyl ring with respect to the fivemembered heterocycles.
The influence on activity of aliphatic substituents with increasing length or bulkiness was also investigated. Extending the size of one of the alkyl groups was reasonably well tolerated, the corresponding 5-ethyl (8, IC 50 = 0.62 μM), 5-n-butyl (10, IC 50 = 0.91 μM), 5-iso-propyl (11,IC 50 = 0.64 μM), and 5-tert-butyl (12, IC 50 = 0.78 μM) derivatives being active in the submicromolar range. Interestingly, among this set of sulfonamides, an n-propyl chain in position 5 of the pyrazole, as in compound 9, furnished a new analogue with a 3-fold higher potency (IC 50 = 0.33 μM) compared to hit 1 (Table 1). These data could indicate the presence of a lipophilic pocket accommodating an aliphatic side chain of a specific size and bulkiness, in either the 3-or 5-position of the pyrazole. To further elucidate whether an increase of lipophilicity in both positions would favor the inhibitory effect against NAAA, the corresponding 3,5-diethyl substitution, as in pyrazole sulfonamide 15, was investigated. However, the compound showed an about 2-fold drop in activity (IC 50 = 1.11 μM) with respect to the corresponding 3-methyl-5-ethyl derivative 8.
Electronic properties of the pyrazole substituents also seemed to have a significant impact on potency. Indeed, while still keeping a 5-methyl residue, an electron-withdrawing trifluoromethyl (14) or an electron-donating methoxy (13) group in the 3-position resulted in either a drop in efficacy with inhibition in the micromolar range (IC 50 = 3.29 μM) or no detectable activity, respectively (Table 1).
Selection of a diversity set of small molecules from our compound library and medium throughput screening (MTS) outcome. A hierarchical agglomerative cluster analysis procedure on our internal compound collection of 13,289 entries was carried out. Molecules are expressed as ECFP4 40 fingerprints and distances between fingerprints were estimated in terms of Tanimoto similarity. The agglomerative process was arbitrarily terminated when 1000 clusters were generated (corresponding to an intracluster Tanimoto distance of 0.462). Each centroid was selected as representative of its entire cluster.
We then conveyed our attention to the modification and functionalization of the piperidine heterocycle connecting the pyrazole ring to the pyrazine (B, Figure 3). While keeping unmodified the rest of the molecule in hit 1, the key role of the tertiary sulfonamide was confirmed by the complete loss of activity shown by the amide analogue 16. A similar outcome was found with the secondary sulfonamide 17, featuring a functionalization of an exocyclic amino group on a cyclohexyl moiety (Table 2).
Next, we explored alternatives to the piperidinyl moiety via ring morphing and scaffold hopping. We first prepared and evaluated the azetidine sulfonamide 18, as a ring contraction analogue of 1, which resulted in a complete loss of enzyme inhibition. The ring opening of the piperidine moiety was also investigated to determine the effect of more flexible substituents, leading to the corresponding acyclic tertiary sulfonamide 19. This modification produced a drop of activity (IC 50 = 4.29 μM), presumably caused by an increased entropic penalty of binding to the biological target (Table 2).
Finally, we tried to constrain the piperidine core into bridged bicyclic systems by designing, synthesizing, and profiling a series of aliphatic heterocyclic replacements. While ring opening or ring contraction was detrimental for inhibition, constraining the piperidine ring into a more conformationally rigid aza-bridged bicyclic scaffold was found to be beneficial. Sulfonamide analogue 20, featuring an azabicyclo[3.2.1]octane core, showed submicromolar activity (h-NAAA IC 50 = 0.23 μM, Table 2), with approximately 5fold boost in potency compared to parent hit 1. At this stage, the stereochemistry of the ether substitution at the pseudoasymmetric carbon in position 3 of the azabicyclic scaffold was evaluated. Notably, contrary to the beneficial conformation effect seen for the endo-isomer 20, the corresponding exo-diastereoisomer 21 turned out to be devoid of any activity toward human NAAA.
As a natural development of our SAR investigation, we replaced the oxygen linker connecting the piperidine core to the pyrazine ring with a nitrogen (22) or a methylene unit (23). Unfortunately, removal (24) or modification (22,23) of the ether linker abolished the inhibitory activity (Table 2).
Encouraged by the increase in potency shown by endosubstituted tropyl sulfonamide 20, we decided to prepare and profile additional bridged aza-bicyclic systems, varying the size, the lipophilic nature, and the position of the bridge on the piperidine ring. While expanding the bridge up to three methylene units, the activity remained in the submicromolar range (25, h-NAAA IC 50 = 0.37 μM), the introduction of an oxygen caused a significant 10-fold drop of the inhibitory effect (26, h-NAAA IC 50 = 2.33 μM). This outcome may be rationalized by hypothesizing that the bridged aza-bicyclic Table 1. Structure and h-NAAA Inhibitory Activity (IC 50 ) of Compounds 1−15 a system is well accommodated into a lipophilic cleft within the h-NAAA active site (vide infra, the docking section).
Moving the bridge in proximity to the pyrazinyloxy substituent, the endo-isomer lost efficacy, providing sulfonamide 27 with only double-digit micromolar activity (IC 50 = 15.4 μM). Bridge expansion to three carbon atoms was again tolerated, depending on the stereochemistry. Interestingly, in this case, the endo-isomer 28 was found to be inactive, while the corresponding exo-diastereoisomer 29 maintained some inhibitory effect (h-NAAA IC 50 = 0.69 μM), with a marginal loss compared to endo-substituted tropyl analogue 20 ( Table  2).
Taken together, these data suggest that conformational restrictions of the piperidine ring could improve h-NAAA inhibition, most likely by minimizing the entropic penalty of binding and increasing lipophilicity. In addition, the bridge seems to be well suited at either the vicinal or distal bridgeheads at the piperidine nitrogen atom, but the geometry of the substituents is crucial for activity. In general, N-vicinal bridges require endo-stereochemistry to produce the most favorable inhibitory effect, while N-distal bridges seem to be more effective in binding to the h-NAAA active site in their exo-configuration.
After exploration of the possible modifications on both the pyrazole ring and the piperidine moiety of sulfonamide 1, having also identified new h-NAAA inhibitors with submicromolar affinity for the target, as a final step of our program of structural manipulations, we investigated the relevance of the terminal heteroaryl group (C, Figure 3).
Initially, the importance of the nitrogen atoms in the pyrazine ring was assessed by preparing and screening against h-NAAA both the corresponding pyridyl (30) and phenyl (31) analogues of endo-substituted sulfonamide 20. Removal of one nitrogen from the pyrazine core, as for pyridine analogue 30, showed a slightly improved activity (h-NAAA IC 50 = 0.16 μM), compared to its more polar parent pyrazine 20. Introducing a more lipophilic moiety, such as a phenyl ring (31), improved the potency further, furnishing a new h-NAAA inhibitor in the double-digit nanomolar range (IC 50 = 0.093 μM, Table 3).
Consistent with the preferred endo-geometry observed for the pyrazine-matched pair endo-20/exo-21, the phenoxy derivative 31 was ca. 7-fold more active than its exodiastereoisomer 32 (h-NAAA IC 50 = 0.655 μM). A methyl scan around the phenyl ring was then explored, highlighting the para position as the preferred vector for further growing. In fact, while the para-methyl-substituted phenoxy sulfonamide 33 displayed a very promising inhibitory potency in the low double-digit nanomolar range (IC 50 = 0.036 μM), the corresponding ortho-(36, IC 50 = 0.291 μM) and meta-(35, IC 50 = 0.614 μM) methyl-phenoxy analogues showed from 8to 17-fold drop in activity, respectively ( Table 3). The stereochemistry effect in terms of h-NAAA inhibition at the pseudoasymmetric carbon in position 3 on the azabicyclic system was further confirmed for the para-methyl phenoxy sulfonamide 34. This compound, featuring an exo-configuration, was shown to inhibit h-NAAA only with a modest micromolar efficacy (IC 50 = 8.71 μM).
In addition to simple hydrophobic substituents, other small functionalities were investigated at the para position of the phenyl group, featuring an endo-configuration at the 3 position in the tropyl scaffold. While a polar 4-cyano residue (45, IC 50 = 0.757 μM) resulted to be quite detrimental in terms of h-NAAA inhibition, the corresponding 4-trifluoromethyl (43, IC 50 = 0.212 μM), 4-methoxy (44, IC 50 = 0.175 μM), or 4-fluoro (46, IC 50 = 0.143 μM) substitutions were found to be tolerated with only a slight loss in activity compared to the double-digit nanomolar methyl derivative 33.
Having now in hand highly potent inhibitors, we focused our optimization efforts on improving their overall drug-like profile by carefully controlling lipophilicity, increasing lipophilic efficiency (LipE), 42,43 and trying to balance potency with physicochemical and ADME properties.
Being an oxygen atom tolerated at the para position, as seen for methoxy analogue 44 (Table 3), we explored this substitution further by making a couple of extended O-alkyl derivatives. First, we hybridized the methoxy residue with the potency-builder n-butyl alkyl chain (as in 39) to obtain a new  n-propoxy derivative 47 (Table 4). Unfortunately, this modification determined an overall drop in inhibitory activity in the submicromolar range (h-NAAA, IC 50 = 0.45 μM). To further explore the effect of a polar atom in the 4-carbon alkyl tail, we moved the oxygen along the aliphatic chain leading to the ethoxymethyl analogue 48. This change resulted to be beneficial both in terms of gained affinity toward h-NAAA (IC 50 = 0.016 μM) and improved overall polarity.
Taking into account the positive effect on activity shown by the pyrazine ring, as seen for endo-substituted tropyl sulfonamide 20, and aiming to reduce lipophilicity while keeping good h-NAAA inhibition (as proven by the n-butyl side chain on the aryloxy moiety), we tried to combine both features in an additive manner. Notably, azabicyclic sulfonamide 49 bearing an endo-5-n-butyl-pyrazyn-2-yloxy substitution was synthesized and tested to display low double-digit nanomolar activity (h-NAAA, IC 50 = 0.017 μM). Intrigued by this last outcome, we ultimately sought to further modulate lipophilicity by incorporating an ethoxymethyl side chain on the pyrazine ring. Along with a substantial contribution to the polarity of the molecule, this modification led to optimized 5-ethoxymethyl-pyrazinyloxy-8azabicyclo[3.2.1]octane pyrazol-sulfonamide 50 (ARN19689, h-NAAA, IC 50 = 0.042 μM, Table 4), as an excellent compromise between reduced lipophilicity and sustained activity compared to endo-substituted tropyl derivative 39.
According to these promising biological data, sulfonamide 50 was further characterized for its in vitro profile. Interestingly, in biological assays, the compound showed a very high selectivity toward both human FAAH and human acid ceramidase (AC) 22 (25 and 34% inhibition at 30 μM, respectively). Along with its NAAA inhibitory activity and selectivity toward FAAH and AC, an evaluation of the mode of action of compound 50 was also investigated. A competitive activity-based protein profiling (ABPP) 44 biochemical analysis revealed also for this new azabicyclic compound, featuring no reactive chemical moieties toward the catalytic cysteine, a non-covalent interaction with h-NAAA. Compound 50 prevented the binding of the activitybased probe (ABP) ARN14686 45 ( Figure S1) to h-NAAA in cell lysosomal extracts in a 15 min incubation experiment. However, its binding was almost totally reverted after 4 h, demonstrating a relatively transient interaction with the enzyme. Conversely, as expected, a known β-lactam covalent NAAA inhibitor (ARN15393) 46 ( Figure S1) stably antagonized h-NAAA labeling by the ABP at both short (15 min) and long (4 h) incubation times ( Figure 4). In addition to demonstrating its reversibility as a NAAA inhibitor, this experiment allows speculations on the putative site of interaction of compound 50 with NAAA. Because the ABP ARN14686 has been reported to bind the catalytic cysteine of h-NAAA (Cys-126), 45 the prevention of its interaction with NAAA by preincubation with 50 can be explained by the compound's binding to the same pocket or to an allosteric site that, upon engagement, induces a conformational change of the enzyme structure that precludes the interaction with the covalent inhibitor.
To support these outcomes, flexible ligand-docking studies in the NAAA active site were also performed on lead compound 39 and optimized analogue 50. Endo-substituted azabicyclooctane 39 was shown to bind at the binding pocket of NAAA with an orientation similar to that of the cocrystallized non-covalent inhibitor ARN19702 ( Figure  5A). 11,31 In particular, the common sulfonamide group was predicted to occupy approximately the same region as in the cocrystal, although with a different orientation. The substituted pyrazole ring established a double H-bond Table 4. h-NAAA Inhibitory Activity (IC 50   interaction with the side chain and backbone of E195. The endo-isomer of the substituted azabicyclic core positioned the phenyl ring in the same region occupied by the 1,3-thiazole ring of the benzothiazole group in ARN19702. In this way, the aliphatic n-butyl chain could be lodged in a deeply hydrophobic region of the binding site, slightly displacing the side chain of M64 ( Figure 5B). Additionally, desolvation penalty also contributes positively to the higher binding affinity observed for the more lipophilic moieties.
The proposed binding mode explained why preventing one of the two H-bond interactions (i.e., introducing a substituent on the N-1 nitrogen of the pyrazole or replacing the pyrazole itself with an isoxazole ring) is detrimental for activity.
Second, it accounts for the marked decrease in potency generally displayed by the exo-diastereoisomers. In the exoseries, the interaction with E195 is partially compromised ( Figure 5C). Furthermore, pointing toward the backbone of A63, the vector for the aliphatic substituent in the phenyl ring displays a less-favorable orientation with respect to the endo-series ( Figure 5D). Given the tightness of the pocket, the ca. 450-fold potency gap between the paired isomers 39 and 40 could be reasonably justified. While docking score does not usually correlate with activity within congeneric series, advanced docking protocols, that take receptor flexibility into account, have been reported to be able to correctly characterize activity cliffs, that is, significant changes  The newly identified optimized endo-substituted azabicyclic sulfonamide 50 was also investigated in docking studies in the NAAA active site, displaying a binding mode almost perfectly overlapping with the one predicted for lead compound 39 (see Figure S2, Supporting Information).
Altogether, these data could indicate that the novel compound 50 may interact with NAAA by forming tight/ non-covalent contacts via H-bond and hydrophobic interactions, as seen for lead compound 39 ( Figure 5).
Although these studies support the binding of compound 50 to the NAAA active site, we cannot rule out that the compound inhibits the enzyme by binding to an allosteric site, thus inducing a conformational change of the enzyme structure that prevents the interaction with the substrate.
The excellent in vitro inhibitory activity, selectivity, and biochemical data of optimized compound 50 prompted us also to assess its physicochemical properties and in vivo pharmacokinetic profile, compared to its early analogue 39 (Table 5). While both compounds showed a high plasma (mouse and rat t 1/2 > 120 min) and liver microsomal (mouse and human t 1/2 > 60 min) stability, the structural modifications leading to compound 50 contributed to improve not only resistance to oxidative transformations in rat microsomes (t 1/2 = 48 min) but also, more importantly, lipophilicity (c Log P = 0.54) and kinetic solubility in buffer (>250 μM in PBS at pH 7.4). Interestingly, the more polar character of NAAA inhibitor 50, together with its favorable biological activity, resulted in an overall optimal lipophilic ligand efficiency value (LipE = 6.83), which is expected to positively impact on the developability of the compound as a possible preclinical candidate (Table 5).
Based on both its biological characterization, showing a good efficacy and biochemical profile, and preliminary in vitro ADME properties, the novel NAAA inhibitor 50 was further evaluated in vivo. The compound was administered to male C57BL/6 mice by intravenous (i.v.) infusion, at a dose of 3 mg/kg, and by oral gavage (p.o.), at a dose of 10 mg/kg, to determine its pharmacokinetic parameters (Table 5 and Figure S3). Compound 50 showed much higher plasma concentrations (C max and AUC) following i.v. and p.o. administration, a lower volume of distribution (V D = 0.53 L/kg) and clearance (CL = 13 mL/min/kg), and a 2-fold increase in overall oral bioavailability (F = ca. 60%) with respect to its earlier analogue 39 (Table 5).
Thanks to its encouraging in vitro and in vivo pharmacological data, the newly identified non-covalent azabicyclooctane sulfonamide 50 represents a valuable tool to be further investigated in animal models of inflammatory conditions.
Sulfonamide 27 was obtained in a straightforward manner starting from commercially available benzyl-azabicyclooctanol 60, which was initially used in a substitution reaction with 2chloropyrazine (73a) to give intermediate 61. Next, the debenzylation with ammonium formate in the presence of Pd/C (10%), followed by coupling with sulfonyl chloride 52k gave the desired compound in a 30% overall yield (Scheme 3).

■ CONCLUSIONS
The endogenous lipid mediator, PEA, has been extensively reported to have anti-inflammatory and analgesic properties. 52 Among the potential ways to manage the inflammatory response, sustaining PEA levels by inhibiting its deactivating hydrolysis represents a promising therapeutic approach. 53 Cellular breakdown of PEA is mainly mediated by the cysteine hydrolase NAAA, whose inhibition can prevent endogenous PEA hydrolysis, thereby eliciting an antiinflammatory response.
In the present work, we report the identification of a potent, systemically available, novel class of NAAA inhibitors, endowed with a non-covalent mechanism of action. Starting from sulfonamide hit 1, identified in an MTS campaign, an in-depth SAR exploration led us to the discovery of highly potent h-NAAA inhibitors, featuring an azabicyclo[3.2.1]octane-pyrazole sulfonamide, as a novel chemotype. Among them, optimized compound 50 (ARN19689) was found to inhibit h-NAAA with a median inhibitory concentration in the low nanomolar range, showing more than 25-fold activity improvement compared to the starting hit (1 vs 50, IC 50 = 1.09 and 0.042 μM, respectively). The evolution of hit 1 led to the azabicyclo[3.2.1]octane-pyrazole lead compound 39, a potent h-NAAA inhibitor with suboptimal drug-like properties. Structural modifications of the latter compound were undertaken to improve the physicochemical and drug-like properties, while retaining a good inhibitory activity against h-NAAA. This work ultimately allowed identifying endoethoxymethyl-pyrazinyloxy-8-azabicyclo[3.2.1]octane-pyrazole sulfonamide 50 (ARN19689), a compound with a good lipophilic efficiency (LipE = 6.83) and a superior pharmacological and pharmacokinetic properties.
The overall excellent profile of sulfonamide 50, in terms of NAAA inhibitory activity and biochemical/drug-like properties, makes this compound an interesting candidate to be tested in animal models of inflammatory diseases.

■ EXPERIMENTAL SECTION
Chemistry. Synthetic Materials and Methods. All manipulations of air-or moisture-sensitive materials were carried out in oven-or flame-dried glassware under an inert atmosphere of nitrogen or argon. Syringes, which were used to transfer reagents and solvents, were purged with nitrogen prior to use. Reaction solvents were obtained anhydrous from commercial suppliers. All reagents were obtained from commercial suppliers and used without further purification unless indicated otherwise. Automated column chromatography purifications were performed on a Teledyne ISCO apparatus (CombiFlash Rf) with prepacked silica gel columns of different sizes (Redisep). NMR experiments were run at 300 K on a Bruker Avance III 400 system (400.13 MHz for 1 H and 100. 62 MHz for 13 C), equipped with a BBI probe and Z-gradients, and a Bruker FT NMR Avance III 600 MHz spectrometer equipped with a 5 mm CryoProbeTM QCI 1 H/ 19 F− 13 C/ 15 N−D quadruple resonance, a shielded z-gradient coil, and the automatic sample changer SampleJet NMR system (600 MHz for 1 H, 151 MHz for 13 C and 565 MHz for 19 F). Chemical shifts for the 1 H and 13 C spectra were reported in parts per million (ppm), calibrating the residual non-deuterated solvent peak for the 1 H and 13 C, respectively, to 7.26 and 77.16 ppm for CDCl 3 18 OBD column (100 × 19 mm ID, particle size 5 μm) with a XBridge Prep C 18 (10 × 19 mm ID, particle size 5 μm) guard cartridge with a flow rate = 20 mL/min. High-resolution mass spectrometry (HRMS) measurements were performed on a Waters Synapt G2 Q-ToF mass spectrometer equipped with an electrospray ionization interface and coupled to a Waters ACQUITY UPLC from Waters Inc. (Milford, MA, USA). Leucine enkephalin (2 ng/mL) was used as lock mass reference compound for spectral recalibration. The analyses were run on an ACQUITY UPLC BEH C 18 column (100 × 2.1 mm ID, particle size 1.7 μm) with a VanGuard BEH C 18 precolumn (5 × 2.1 mm ID, particle size 1.7 μm). The mobile phase was H 2 O + 0.1% HCOOH (A) and CH 3 CN + 0.1% HCOOH (B) with 0.5 mL/min as a flow rate. A linear gradient was applied: 0−0.2 min: 10% B; 0.2−6.2 min: 10−90% B; 6.2−6.3 min: 90−100% B; and 6.3−7.0 min: 100% B. The synthesis and characterization of all final compounds 1-50 is reported below. Purity of the final compounds was determined by UPLC-MS and quantitative 1 H NMR (qNMR, see the Supporting Information) and was equal or greater than 95% for all of the compounds, except for analogue 3 (91% purity).
General Procedure (GP) for the Synthesis of Amines 53, 56, B.
General Procedure for the Synthesis of Pyrazoles 51a,b,d−g,i− j. To a solution of diketone (1.0 equiv) in EtOH (2.0 mL/equiv), hydrazine hydrate (1.2 equiv) at 0°C was added. The reaction mixture was warmed up to room temperature and stirred for 1 h. The solution was concentrated under vacuo and extracted with EtOAc (3 × 10 mL). The organic phases were dried over Na 2 SO 4 , filtered, and evaporated under reduced pressure to obtain the desired pyrazoles, which were used in the next step without any further purification. 5-Ethyl-3-methyl-1H-pyrazole (51c) 57 and 5-methoxy-3-methyl-1H-pyrazole (51h) 58 were synthesized as previously reported.
General Procedure for the Synthesis of Pyrazole-4-Sulfonyl Chlorides 52a−j. Chlorosulfonic acid (5.0 equiv) was added to the appropriate pyrazole (1.0 equiv) at 0°C and the reaction mixture was heated at 100°C for 3 h. The solution was cooled at room temperature and added to a stirrer solution of ice and DCM (15 mL). The organic phase was separated, dried over Na 2 SO 4 , and evaporated under reduced pressure to obtain the desired sulfonyl chlorides, which were used in the next step without any further purification.
Synthesis of Compounds 1−50 (Schemes 1−5). General Procedure (GP1) for the Synthesis of Sulfonamides. To a solution of proper amine (1.1 equiv) in dry THF or DCM (6.0 mL/equiv), TEA (3.0 equiv) and the appropriate sulfonyl chloride (1.0 equiv) were added, and the reaction mixture was stirred at room temperature overnight. The mixture was quenched by the addition of aq. HCl 2 N (5.0 mL/equiv) and extracted with EtOAc (3 × 10 mL). The combined organic phases were dried over Na 2 SO 4 , filtered, and evaporated under vacuo. The crude product was purified by flash chromatography or by recrystallization from Et 2 O to give the pure title compound.
General Procedure (GP2) for the Mitsunobu Reaction. To a solution of tert-butyl-hydroxy-azabicyclo carboxylate (1.05 equiv), substituted phenol (1.0 equiv), and PPh 3 (1.05 equiv) in dry THF (10 mL/equiv), under a nitrogen atmosphere, at 0°C, DIAD (1.05 equiv) was added dropwise. The reaction mixture was allowed to warm to room temperature and stirred overnight. Then, the mixture was quenched with aq. HCl 2 N (10 mL) and extracted with EtOAc (2 × 10 mL). The organic extracts were washed with brine, dried over Na 2 SO 4 , and concentrated under vacuo to give a crude residue, which was purified by flash chromatography to give the title compound.
General Procedure (GP3) for O-Substitution. To a solution of tert-butyl-hydroxy-azabicyclo carboxylate (1.0 equiv) and substituted heteroaryl chloride (1.0 equiv) in dry THF (10 mL/equiv), t-BuOK (1.0 equiv) was added. The reaction mixture was refluxed overnight, then quenched with water (10 mL) and extracted with DCM (2 × 10 mL). The organic extracts were washed with brine, dried over Na 2 SO 4 , and concentrated under vacuo to give a crude residue, which was purified by flash chromatography to give the title compound.
General Procedure (GP4) for the Boc-Deprotection. The appropriate N-Boc-substituted azabicyclo derivative (1.0 equiv) was treated at 0°C with a 3:1 DCM/TFA mixture (4.0 mL/ equiv) and the reaction was stirred at room temperature for 2 h. The crude mixture was concentrated under vacuo, and the subjected to three cycles of suspension (DCM, 10 mL)/concentration to obtain the desired product, which was used in the next step without any further purification.

Preparation of Enzyme-Enriched Lysate (h-NAAA and h-AC).
HEK-293 cells stably transfected with the human NAAA or human AC coding sequences were used as the enzyme source. Cells were suspended in 20 mM Tris-HCl (pH 7.4) with 0.32 M sucrose, sonicated, and centrifuged at 800g for 30 min at 4°C. Supernatants were then centrifuged at 12,000g for 30 min at 4°C. Pellets were resuspended in PBS buffer (pH 7.4) and subjected to three freezethaw cycles at −80°C. The suspension was finally centrifuged at 105,000g for 1 h at 4°C, supernatants were collected, protein concentration was measured, and samples were aliquoted and stored at −80°C until use.
Preparation of Membrane-Enriched Lysate (h-FAAH-1). Cell pellets, obtained by centrifugation at 300g for 7 min at 4°C, were resuspended in 20 mM Tris-HCl pH 7.4, 0.32 M sucrose, disrupted by sonication (10 pulses, 5 times), and centrifuged (1000g, 10 min, 4°C). Supernatants were then centrifuged at 12,000g for 10 min at 4°C and then at 105,000g for 1 h at 4°C. Membrane pellets were resuspended in PBS to obtain h-FAAH-1 preparation; the protein concentration was measured by the Bradford Protein Assay (BioRad) and samples were aliquoted and stored at −80°C until use.
In Vitro ADMET. Aqueous Kinetic Solubility Assay. The aqueous kinetic solubility was determined from a 10 mM DMSO stock solution of test compound (39 or 50) in PBS buffer at pH 7.4. The study was performed by incubating an aliquot of 10 mM DMSO stock solution in PBS (pH 7.4) to a target concentration of 250 μM (2.5% DMSO). The incubation was carried out under shaking at 25°C for 24 h, followed by centrifugation at 21,100g for 30 min. The supernatant was further diluted (4:1) with CH 3 CN, and the dissolved test compound was quantified by UV at 215 nm on a Waters ACQUITY UPLC-MS system consisting of a single quadrupole detector (SQD) mass spectrometer equipped with an electrospray ionization interface and a photodiode array detector (PDA) from Waters Inc. (Milford, MA, USA). Electrospray ionization in positive mode was used in the mass scan range 100−500 Da. The PDA range was 210−400 nm. The analyses were run on an ACQUITY UPLC BEH C18 column (50 × 2.1 mm ID, particle size 1.7 μm) with a VanGuard BEH C18 precolumn (5 × 2.1 mm ID, particle size 1.7 μm), using 10 mM NH 4 OAc in H 2 O at pH 5 adjusted with AcOH (A) and 10 mM NH 4 OAc in CN−H 2 O (95:5) at pH 5 (B) as the mobile phase. The aqueous kinetic solubility (in μM) was calculated by dividing the peak areas of dissolved test compound and test compound in the reference (250 μM of test compound in CH 3 CN) and multiplying by the target concentration and dilution factor.
Plasma Stability Assay. 10 mM DMSO stock solution of test compound (39 or 50) was diluted 50-fold with DMSO/H 2 O (1:1) and incubated at 37°C for 2 h with mouse or rat plasma added with 5% DMSO (preheated at 37°C for 10 min). The final compound's concentration was 2.0 μM. At each time point (0, 5, 15, 30, 60, and 120 min), 50 μL of incubation mixture was diluted with 150 μL cold CH 3 CN spiked with 200 nM of warfarin, as the internal standard, followed by centrifugation at 3.270g for 20 min. The supernatant was further diluted with H 2 O (1:1) and analyzed by LC−MS/MS on a Waters ACQUITY UPLC-MS TQD system consisting of a triple quadrupole detector (TQD) mass spectrometer equipped with an electrospray ionization interface and a photodiode array eλ detector (PDA) from Waters Inc. (Milford, MA, USA). Electrospray ionization was applied in positive mode. Compounddependent parameters as MRM transitions and collision energy were developed for each compound. The analyses were run on an ACQUITY UPLC BEH C 18 (50 × 2.1 mm ID, particle size 1.7 μm) with a VanGuard BEH C 18 precolumn (5 × 2.1 mm ID, particle size 1.7 μm) at 40°C, using H 2 O + 0.1% HCOOH (A) and CH 3 CN + 0.1% HCOOH (B) as the mobile phase. The percentage of the test compound remaining at each time point relative to t = 0 was calculated by the response factor on the basis of the internal standard peak area. The percentage of test compound versus time was plotted and fitted by GraphPad Prism (GraphPad Software, Version 5 for Windows, CA, USA, www.graphpad.com) to estimate the compound's half-life (t 1/2 ), which was reported as mean value along with the standard deviation (n = 3).
Liver Microsomal Stability Assay. Pooled CD1 mouse (M1000, male), IGS Sprague-Dawley rat (R1000, male), and human (H1000, male) liver microsomes were purchased from SEKISUI XenoTech. 10 mM DMSO stock solution of the test compound (39 or 50) was preincubated at 37°C for 15 min with mouse, rat, or human liver microsomes in 0.1 M Tris-HCl buffer (pH 7.5) with 10% DMSO. The final concentration was 4.6 μM. After preincubation, the cofactor (NADPH) was added to the incubation mixture, and the incubation was continued at 37°C for 1 h. At each time point (0, 5, 15, 30, and 60 min), 30 μL of incubation mixture was diluted with 200 μL cold CH 3 CN spiked with 200 nM of an appropriate internal standard, followed by centrifugation at 3.270g for 15 min. The supernatant was further diluted with H 2 O (1:1) for analysis. A reference incubation mixture of each specie (microsomes without cofactors) was prepared for each test compound and analyzed at t = 0 and 60 min in order to verify the compound stability in the matrix. The two time points were diluted as for the time points of the incubation mixture above. The supernatants were analyzed by LC−MS/MS on a Waters ACQUITY UPLC-MS TQD system consisting of a triple quadrupole detector (TQD) mass spectrometer equipped with an electrospray ionization interface and a photodiode array eλ detector (PDA) from Waters Inc. (Milford, MA, USA). Electrospray ionization was applied in positive mode. Compounddependent parameters as MRM transitions and collision energy were developed for each compound. The analyses were run on an ACQUITY UPLC BEH C 18 (50 × 2.1 mm ID, particle size 1.7 μm) with a VanGuard BEH C 18 precolumn (5 × 2.1 mm ID, particle size 1.7 μm) at 40°C, using H 2 O + 0.1% HCOOH (A) and CH 3 CN + 0.1% HCOOH (B) as the mobile phase. The percentage of test compound remaining at each time point relative to t = 0 was calculated by the response factor on the basis of the internal standard peak area. The percentage of test compound versus time was plotted and fitted by GraphPad Prism (GraphPad Software, Version 5 for Windows, CA, USA, www.graphpad.com) to estimate the compound's half-life (t 1/2 ), which was reported as mean value along with the standard deviation (n = 3).
In Vivo Pharmacology. Animals. Male C57BL/6 mice, weighing 22−24 g, were used (Charles River). All procedures were performed in accordance with the Ethical Guidelines of European Communities Council (Directive 2010/63/EU of 22 September 2010) and accepted by the Italian Ministry of Health. All efforts were made to minimize animal suffering and to use the minimal number of animals required to produce reliable results, according to the "3Rs concept". Animals were group-housed in ventilated cages and had free access to food and water. They were maintained under a 12 h light/dark cycle (lights on at 8:00 am) at controlled temperature (21 ± 1°C) and relative humidity (55 ± 10%).
Pharmacokinetics Methods. Compound 39 or 50 was administered p.o. and i.v. to C57BL/6 male mice at 10 and 3 mg/kg. The vehicle used was PEG400/Tween 80/saline solution at 10/10/80% in volume, respectively. Three animals per each time point were treated. Blood samples at 0, 15, 30, 60, 120, 240, and 480 min after administration were collected for the p.o. arm. Blood samples at 0, 5, 15, 30, 60, 120, and 240 min after administration were collected for the i.v. arm. Plasma was separated from blood by centrifugation for 15 min at 1500 rpm a 4°C, transferred to Eppendorf tubes, and frozen (−80°C). Control animals treated with vehicle only were also included in the experimental protocol.
Plasma samples were centrifuged at 21,100g for 15 min at 4°C. An aliquot of each sample was extracted (1:3) with cold CH 3 CN containing 200 nM of an appropriate internal standard as a close analogue of the parent compound (for compound 39: (1R,3r,5S)-3-(4-phenylphenoxy)-8-((3,5-dimethyl-1H-pyrazol-4-yl)sulfonyl)-8 azabicyclo[3.2.1]octane; for compound 50: (1R,3r,5S)-3-((5-methylpyrazin-2-yl)oxy)-8-((3,5-dimethyl-1H-pyrazol-4-yl)sulfonyl)-8azabicyclo[3.2.1]octane). A calibration curve was obtained in blank mouse plasma over a 1 nM to 10 μM range. Three quality controls were prepared by spiking the parent compound in blank mouse plasma to 20, 200, and 2.000 nM, as final concentrations. The calibrators and quality controls were extracted (1:3) with the same extraction solution as the plasma sample. The plasma samples, the calibrators, and quality controls were centrifuged at 3.270g for 15 min at 4°C. The supernatants were further diluted (1:1) with H 2 O + 0.1% HCOOH, and analyzed by LC−MS/MS on a Waters ACQUITY UPLC-MS TQD system consisting of a Triple Quadrupole Detector (TQD) Mass Spectrometer equipped with an Electrospray Ionization interface and a Photodiode Array eλ Detector from Waters Inc. (Milford, MA, USA). Electrospray ionization was applied in positive mode. Compound-dependent parameters such as MRM transitions and collision energy were developed for the parent compound and the internal standard. The analyses were run on an ACQUITY UPLC BEH C 18 (50 × 2.1 mm ID, particle size 1.7 μm) with a VanGuard BEH C 18 precolumn (5 × 2.1 mm ID, particle size 1.7 μm) at 40°C. The mobile phase was H 2 O + 0.1% HCOOH (A) and CH 3 CN + 0.1% HCOOH (B) at a flow rate = 0.5 mL/min. A linear gradient was applied starting at 10% B with an initial hold for 0.2 min, then 10−100% B in 2 min. All samples (plasma samples, calibrators, and quality controls) were quantified by the MRM peak area response factor in order to determine the levels of the parent compound in plasma. The concentrations versus time were plotted and the profiles were fitted using PK Solutions Excel Application (Summit Research Service, USA) in order to determine the pharmacokinetic parameters.
Computational Methods (Docking Study). Ligand docking simulations were carried out with the induced fit docking (IFD) protocol 63 as implemented in Schrodinger 2019-3 (Schrodinger LLC, New York, USA). The coordinates of the enzyme in complex with the non-covalent inhibitor ARN19702 ( Figure 1) were processed with the protein preparation routine. Default parameters were used. The coordinates of the cocrystallized ligand (PDB ID: 6DXX) 11 were used to define the size and the position of the binding box and then the ligand was deleted from the system. In the first step of the IFD protocol, each ligand was docked scaling by a factor of 0.5 the van der Waals radii of all the protein and ligand atoms with partial charges ≤0.25. The standard Glide SP protocol was used for docking. 64,65 Up to 20 poses were retained and progressed to the next step. The van der Waals scaling factor was removed, and each complex was refined using Prime. 66 All residues within 5 Å from the ligand were optimized. Finally, ligands were redocked in each complex using the optimized receptor coordinates, standard van der Waals radii, and the Glide XP protocol. 67 For each ligand, the best scoring complex was retained. All calculations were carried out using the OPLS3e force field.