Exploring N-acyl-4-azatetracyclo[5.3.2.02,6.08,10]dodec-11-enes as 11β-HSD1 Inhibitors

We recently found that a cyclohexanecarboxamide derived from 4-azatetracyclo[5.3.2.02,6.08,10]dodec-11-ene displayed low nanomolar inhibition of 11β-HSD1. In continuation of our efforts to discover potent and selective 11β-HSD1 inhibitors, herein we explored several replacements for the cyclohexane ring. Some derivatives exhibited potent inhibitory activity against human 11β-HSD1, although with low selectivity over the isoenzyme 11β-HSD2, and poor microsomal stability.


Introduction
Glucocorticoids (GCs) are hormones that play a major role in the modulation of inflammatory and immune responses, metabolism regulation, cardiovascular homeostasis, and the body's response to stress [1,2]. It is well accepted that the local GC concentration in peripheral tissues depends not only on the circulating levels from adrenal secretion but also on the intracellular metabolism performed by activating and deactivating enzymes. 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) catalyzes cortisol regeneration from its inactive form cortisone [3]. In contrast, 11β-HSD2 catalyzes the opposite reaction by oxidizing cortisol to cortisone [4]. 11β-HSD1 predominates in tissues mainly expressing glucocorticoids receptors, such as liver, adipose, and brain, whereas 11β-HSD2 is found in tissues mainly expressing mineralocorticoid receptors, such as kidney, colon, and salivary glands [5,6]. Selectivity against the desired 11β-HSD isoform is a key factor to avoid side effects of novel 11β-HSD1 inhibitors in development.
In recent years, both academia and industry have made great efforts to determine the role of this enzyme in diseases in which elevated cortisol plays an important role [7]. As a result, 11β-HSD1 activity has been found to be important in type 2 diabetes and metabolic syndrome [8], in Alzheimer's disease (AD) [9], in osteoporosis [10], and in glaucoma [11]. In light of this evidence, 11β-HSD1 has been explored as a therapeutic target to decrease cortisol concentrations in target tissues.
In the case of AD, it has been demonstrated that aged mice with cognitive deficits show increased 11β-HSD1 expression in the hippocampus and forebrain, and that overexpression of 11β-HSD1 leads to a similar premature memory decline [9]. Conversely, 11β-HSD1 knock-out mice perform better in behavioral tests, suggesting resistance to cognitive decline through a neuroprotective effect [12]. Accordingly, this protection correlates with loss of the age-associated rise in intrahippocampal corticosterone [9]. As matter of fact, 11β-HSD1 inhibitors in acute and short-term treatments have shown memory consolidation and improvements in cognitive function in aged mice and AD models [13][14][15]. Overall, these data support the 11β-HSD1 inhibition as a novel approach through a non-cholinergic mechanism to deal with these cognitive disorders.

Design of New Inhibitors
Our previous work on polycyclic substituent optimization of N-(2-adamantyl)amide 1 led to the identification of pyrrolidine-based amides 2 and 3 as potent 11β-HSD1 inhibitors (Table 1) [16]. When tested against the 11β-HSD2 isoform, 2 had a selectivity index of at least 50-fold (IC 50 = 1-10 µM), while 3 showed poor selectivity (IC 50 = 0.1-1 µM). However, 3 possessed high metabolic stability in human liver microsomes (HLM, 94% of remaining compound after 30-min incubation), whereas 2 was rapidly metabolized (27%). In light of these results, and with the aim of prioritizing the microsomal stability, 3 was further in vitro characterized in terms of murine enzyme inhibition (mHSD1 IC 50 = 81 nM) and metabolic stability in murine liver microsomes (MLM, 93%). Subsequently, we performed an in vivo study with 3 in the Senescence-Accelerated Mouse Prone 8 (SAMP8) model of cognitive dysfunction in order to support the neuroprotective effect of 11β-HSD1 inhibition in cognitive decline related to the aging process. We found that 3, administered to 12-month-old SAMP8 mice for four weeks, prevented memory deficits and displayed a neuroprotective action through reduction of neuroinflammation and oxidative stress, in cognitive decline related to the aging process [17]. These promising results with an early lead without optimal selectivity and DMPK (drug metabolism and pharmacokinetics) properties led us to investigate additional potent 11β-HSD1 inhibitors that maintained the optimized polycycle of compound 2 while modifying the right-hand side (RHS) substituent of the structure.  [13][14][15]. Overall, these data support the 11β-HSD1 inhibition as a novel approach through a noncholinergic mechanism to deal with these cognitive disorders.

Design of New Inhibitors
Our previous work on polycyclic substituent optimization of N-(2-adamantyl)amide 1 led to the identification of pyrrolidine-based amides 2 and 3 as potent 11β-HSD1 inhibitors (Table 1) [16]. When tested against the 11β-HSD2 isoform, 2 had a selectivity index of at least 50-fold (IC50 = 1-10 µM), while 3 showed poor selectivity (IC50 = 0.1-1 µM). However, 3 possessed high metabolic stability in human liver microsomes (HLM, 94% of remaining compound after 30-min incubation), whereas 2 was rapidly metabolized (27%). In light of these results, and with the aim of prioritizing the microsomal stability, 3 was further in vitro characterized in terms of murine enzyme inhibition (mHSD1 IC50 = 81 nM) and metabolic stability in murine liver microsomes (MLM, 93%). Subsequently, we performed an in vivo study with 3 in the Senescence-Accelerated Mouse Prone 8 (SAMP8) model of cognitive dysfunction in order to support the neuroprotective effect of 11β-HSD1 inhibition in cognitive decline related to the aging process. We found that 3, administered to 12month-old SAMP8 mice for four weeks, prevented memory deficits and displayed a neuroprotective action through reduction of neuroinflammation and oxidative stress, in cognitive decline related to the aging process [17]. These promising results with an early lead without optimal selectivity and DMPK (drug metabolism and pharmacokinetics) properties led us to investigate additional potent 11β-HSD1 inhibitors that maintained the optimized polycycle of compound 2 while modifying the right-hand side (RHS) substituent of the structure. A series of different substituents were integrated into the RHS moiety, while the amido linker was retained to enable the key hydrogen bonds in the binding site [17]. A diversity of substituents was generated including aromatic, heteroaromatic (electron-rich and deficient rings), branched alkyl, cycloalkenyl, heterocycloalkyl-and other groups inspired in previously reported 11β-HSD1 inhibitors from Abbott (a series of dichoroaniline amides [18], and ABT-384, which contains a 4-(pyridin-2-yl)piperazin-1-yl ring system) [19].

In Vitro Pharmacological Evaluation
A preliminary screen was performed using a human microsome assay with compounds at 10 µM in order to assess their potential inhibition of the target enzyme ( Table 2). Percentage of inhibition was determined by measuring the conversion of 3 H-cortisone to 3 H-cortisol by capturing liberated 3 H-cortisol on anti-cortisol (HyTest Ltd., Turku, Finland)-coated scintillation proximity assay beads. The assay uses human liver microsomes (HLM), where the enzyme is expressed, and NADPH as the cofactor needed by the enzyme. Eight of the thirteen new compounds presented 100% inhibition of the human 11β-HSD1 in this single concentration assay, so dose-response curves were performed to get their IC50 values.

In Vitro Pharmacological Evaluation
A preliminary screen was performed using a human microsome assay with compounds at 10 µM in order to assess their potential inhibition of the target enzyme ( Table 2). Percentage of inhibition was determined by measuring the conversion of 3 H-cortisone to 3 H-cortisol by capturing liberated 3 H-cortisol on anti-cortisol (HyTest Ltd., Turku, Finland)-coated scintillation proximity assay beads.
The assay uses human liver microsomes (HLM), where the enzyme is expressed, and NADPH as the cofactor needed by the enzyme. Eight of the thirteen new compounds presented 100% inhibition of the human 11β-HSD1 in this single concentration assay, so dose-response curves were performed to get their IC 50 values.
The analysis of these potencies showed some structure-activity relationships (SAR). First, the introduction of a double bond in the cyclohexyl substituent of 2 delivered compound 9 which maintained nanomolar potency (IC 50 = 0.056 µM) comparable to compound 2. Second, introduction of a phenyl group or few other aryl groups (either electron rich or electron-deficient, see compounds 7 or 5, 10, 11, respectively) on the RHS of the molecule did not improve or was deleterious for the activity (IC 50 = 0.546 µM for RHS = phenyl, 4; 49% inhibition at 10 µM for RHS = 2-thiophenyl, 7; IC 50 = 4.3 µM for RHS = 2-pyridinyl, 5; and 0 and 23% inhibition at 10 µM for RHS = 4-chloro-3-pyridinyl, 10, and RHS = 3-chloro-4-pyridinyl, 11, respectively). Fortunately, when the phenyl group was substituted by a previously reported dichloroaniline group [17], the potency was substantially increased to deliver a low nanomolar inhibitor (8, IC 50 = 0.045 µM). Third, introduction of N-substituted piperidinyl groups was again deleterious for the 11β-HSD1 inhibitory activity (12 and 13, 3% and 0% inhibition at 10 µM, respectively). Fourth, a branched alkyl substituent, such as the tert-butyl group, delivered amide 6 with a moderate potency (IC 50 = 0.666 µM). Finally, compounds 14-16 containing a 6-(4-phenylpiperazin-1-yl)pyridin-3-yl system showed interesting SAR while completely inhibiting the target enzyme at 10 µM. Compound 14, featuring a terminal non-substituted phenyl ring in its structure, exhibited an IC 50 of 5.44 µM. Surprisingly, introduction of the trifluoromethyl group in the para position mimicking the ABT-384 structure [18] reduced considerably the activity (compound 15, IC 50 = 11.60 µM) while a cyano group increased the potency (compound 16, IC 50 = 0.377 µM).  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-  Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,   Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,    Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,    Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,   Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,   Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,   Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,   Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,    Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2,    Those compounds with submicromolar IC50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell- Those compounds with submicromolar IC 50 values (4, 6, 8, 9 and 16) were further evaluated in terms of cellular potency, selectivity over 11β-HSD2, and metabolic stability as follows ( Table 2, 5th, 6th, and 7th column, respectively). Cellular potency was assessed using Human Embryonic Kidney 293 (HEK293) cells stably transfected with the 11β-HSD1 gene. Cells were incubated with substrate (cortisone) and percentage of inhibition was determined by measuring the conversion of cortisone to cortisol by LC/MS. The results were in line with the previous results obtained in the microsomal assay. The most potent compounds, 8, 9 and 16, presented complete inhibition at 10 µM in the cell-based assay, whereas compounds with IC 50 values between 0.5 and 1 µM, i.e., 4 and 6, showed a moderate cellular potency (77% and 41%, respectively).

Cellular 11β-HSD2 Enzyme Inhibition Assay
For measurement of inhibition of 11β-HSD2, HEK293 cells stably transfected with the full-length gene coding for human 11β-HSD2 were used. The protocol was the same as for the cellular 11β-HSD1 enzyme inhibition assay, only changing the substrate, this time cortisol, and the concentrations of the tested compounds, 10, 1, and 0.1 µM. Reported values are the average of 1-3 measurements.

Microsomal Stability Assay
The microsomal stability of each compound was determined using human liver microsomes (HLM, Celsis In-vitro Technologies, Baltimore, MD, USA). Microsomes were thawed and diluted to a concentration of 2 mg/mL in 50 mM sodium phosphate buffer pH 7.4. Each compound was diluted in 4 mM NADPH (made in the phosphate buffer above) to a concentration of 10 µM. Two identical incubation plates were prepared to act as a 0 min and a 30 min time point assay. 30 µL of each compound dilution was added in duplicate to the wells of a U-bottom 96-well plate and warmed at 37 • C for approximately 5 min. Verapamil, lidocaine, and propranolol at 10 µM concentration were utilized as reference compounds in this experiment. Microsomes were also pre-warmed at 37 • C before the addition of 30 µL to each well of the plate resulting in a final concentration of 1 mg/mL. The reaction was terminated at the appropriate time point (0 or 30 min) by addition of 60 µL of ice-cold 0.3 M trichloroacetic acid (TCA) per well. The plates were centrifuged for 10 min at 112× g and the supernatant fraction transferred to a fresh U-bottom 96-well plate. Plates were sealed and frozen at −20 • C prior to MS analysis. LC-MS/MS was used to quantify the peak area response of each compound before and after incubation with HLM using MS tune settings established and validated for each compound. These peak intensity measurements were used to calculate the percentage remaining after incubation with HLM for each hit compound. Reported values are the average of 1-3 measurements.

Conclusions
In summary, we designed, synthesized and described SAR for a novel series of 11β-HSD1 inhibitors featuring the optimized polycyclic substituent 4-azapentacyclo[5.3.2.0 2,6 .0 8,10 ]dodec-11-ene. Nanomolar potencies were achieved for compounds 8 and 9, although selectivities and metabolic stabilities were suboptimal. The discovery of inhibitors with desirable selectivity and DMPK properties is a key step for the development of successful 11β-HSD1 inhibitors for the treatment of GC-related disorders such as diabetes and AD. Clear SAR in this new family of 11β-HSD1 inhibitors was found; a double bond was tolerated in the initial cyclohexyl unit (9, IC 50 = 0.056 µM), but the inclusion of heterocycloalkyl and heteroaromatic groups reduced considerably or were detrimental for the inhibitory activity (5, IC 50 = 4.265 µM, and 7, 10, 11, 12 and 13, <50% inhibition at 10 µM). The introduction of a phenyl group as RHS of the molecule was also detrimental for the potency (4, IC 50 = 0.546 µM); however, the introduction of a previously reported substitution pattern on the aryl unit delivered again a low nanomolar inhibitor (8, IC 50 = 0.045 µM). Future efforts will be focused on rational design of the substitution pattern of this aryl group to identify optimized compounds addressing the weaknesses of those described in this work.

Patents
A PCT patent application has been filed. See PCT WO2017/182464A1 (priority data 19 April 2016).

Supplementary Materials:
The following are available online: copies of the 1 H-and 13 C-NMR spectra of the new compounds.