Synthesis and Structure–Activity Relationships for the Anti-Mycobacterial Activity of 3-Phenyl-N-(Pyridin-2-ylmethyl)Pyrazolo[1,5-a]Pyrimidin-7-Amines

Pyrazolo[1,5-a]pyrimidines have been reported as potent inhibitors of mycobacterial ATP synthase for the treatment of Mycobacterium tuberculosis (M.tb). In this work, we report the design and synthesis of approximately 70 novel 3,5-diphenyl-N-(pyridin-2-ylmethyl)pyrazolo[1,5-a]pyrimidin-7-amines and their comprehensive structure–activity relationship studies. The most effective pyrazolo[1,5-a]pyrimidin-7-amine analogues contained a 3-(4-fluoro)phenyl group, together with a variety of 5-alkyl, 5-aryl and 5-heteroaryl substituents. A range of substituted 7-(2-pyridylmethylamine) derivatives were also active. Some of these compounds exhibited potent in vitro M.tb growth inhibition, low hERG liability and good mouse/human liver microsomal stabilities, highlighting their potential as inhibitors of M.tb.


Introduction
Tuberculosis (TB) is the deadliest infectious disease around the globe, claiming over a billion lives in the past two hundred years [1][2][3]. According to the World Health Organization (WHO), TB has claimed the lives of 1.3 million people, with an estimated 9.9 million new cases of TB in 2020 [4]. The current major hurdle of the global TB challenge is the fight against the drug-resistant forms of the disease [5,6]. The emergence of drug-resistant strains of Mycobacterium tuberculosis (M.tb), the causative agent of TB, is on the rise, with treatment success rates dropping for patients with multidrug-resistant (MDR) TB. The latest WHO data for drug-resistant TB suggests an estimated 465,000 new cases and 182,000 deaths resulted from MDR-TB in 2019 [7]. In the past two decades, there have also been alarming increases in M.tb strains with resistance to all available TB drugs, resulting in extensively and totally drug-resistant/incurable tuberculosis [8][9][10][11]. The development of novel TB agents to treat these resistant strains of M.tb is urgently needed.
Pyrazolo [1,5-a]pyrimidines have been reported as potential drugs in a number of different areas ( Figure 1); as VEGF/src inhibitors (e.g., 1) [12], as apoptosis inducers (e.g., 2) [13], for treatment of Duchenne muscular dystrophy (e.g., 3) [14] and as cyclin-dependent kinase inhibitors (e.g., 4) [15]. In a recent paper, Tantry et al. [16,17] also discuss 2-phenyl-5-substituted pyrazolo [1,5-a]pyrimidines as inhibitors of mycobacterial ATP synthase. Modelling studies suggest the latter compounds bind between the Atp-a and Atp-c (chain-B) subunits of the enzyme, with the pendant 5-phenyl ring occupying the hydrophobic space between the two Atp-c subunits. Example compounds 5 and 6 were modest inhibitors of both depletion of ATP (using an M. smegmatis inverted membrane vesicle assay to measure ATP inhibition via the oxidative phosphorylation process) and of inhibition of M.tb bacteria in culture, and were active in vivo in an acute mouse model of tuberculosis. Chibale et al. [18] also explored SAR for 7-substituted pyrazolo [1,5-a]pyrimidines as M.tb drugs (e.g., 7) and showed that the optimal C-7 side chain was 2-pyridinemethanamine.
Modelling studies suggest the latter compounds bind between the Atp-a and Atp-c (chain-B) subunits of the enzyme, with the pendant 5-phenyl ring occupying the hydrophobic space between the two Atp-c subunits. Example compounds 5 and 6 were modest inhibitors of both depletion of ATP (using an M. smegmatis inverted membrane vesicle assay to measure ATP inhibition via the oxidative phosphorylation process) and of inhibition of M.tb bacteria in culture, and were active in vivo in an acute mouse model of tuberculosis. Chibale et al. [18] also explored SAR for 7-substituted pyrazolo [1,5-a]pyrimidines as M.tb drugs (e.g., 7) and showed that the optimal C-7 side chain was 2-pyridinemethanamine. In the present paper, we extend the latter work with the synthesis of a series of novel 3,5-diphenyl-N-(pyridin-2-ylmethyl)pyrazolo[1,5-a]pyrimidin-7-amines with substituted 3-and/or 5-phenyl rings, exploring relationships between substituent patterns, overall physicochemical properties and their antimycobacterial activity (inhibition of M.tb in culture) (Tables 1-3). In the present paper, we extend the latter work with the synthesis of a series of novel 3,5-diphenyl-N-(pyridin-2-ylmethyl)pyrazolo[1,5-a]pyrimidin-7-amines with substituted 3-and/or 5-phenyl rings, exploring relationships between substituent patterns, overall physicochemical properties and their antimycobacterial activity (inhibition of M.tb in culture) (Tables 1-3).     [19] or non-replicating (LORA) [20] conditions, determined at the Institute for Tuberculosis Research, University of Illinois at Chicago; b IC50 values (μg/mL) for the inhibition of growth of VERO green monkey kidney epithelial cells [21]. c clogP values calculated by ChemDraw Ultra v21.0.0.28 (CambridgeSoft).    Compounds 22-44 of Table 2 (Scheme 2) were prepared from commercial 4-(4-fluorophenyl)-1H-pyrazol-5-amine (8h) by reacting it with diethyl malonate/sodium ethoxide to give the diol (73), which was converted to the dichloride (74). The chloride at the 7 position was selectively displaced with 2-pyridinemethanamine to give 75, BOC-protection of the amine gave 76, which subsequently underwent Suzuki coupling with various aryl and heteroaryl boronic acids to give Boc24-Boc44. The Suzuki coupling was unsuccessful under a variety of conditions when 75 was used, and this necessitated the use of the Boc-protected derivative 76.
Boc24 was hydrogenated using palladium on carbon to give Boc23. An attempt to convert 76 to a boronate ester was unsuccessful, only the product (Boc22) arising from reduction was isolated.
Deprotection of the Boc-protected intermediates (Boc22-Boc44) using trifluoroacetic acid in DCM at reflux gave the compounds 22-44 of Table 2. Compounds 45-71 of Table 3 were prepared as shown in Scheme 3, from 10h and the appropriate substituted pyridine-2-ylmethanamines. The majority of the pyridine-2ylmethanamines were available commercially; the synthesis of pyridylmethanamines used in the synthesis of 57 and 58 is described in the supplementary information.

Results and Discussion
The compounds of Tables 1-3 were tested for their ability to inhibit the growth of Mycobacterium tuberculosis (strain H37Rv) when cultured under either aerobic (MABA) [19] or low-oxygen (LORA) [20] conditions, by determining the minimum inhibitory concentrations (MIC 90 ; µg/mL) needed to reduce growth by 90%. The compounds were also assessed for their ability to inhibit the growth of mammalian cells (VERO green monkey kidney cells) by determining IC 50 values [21]. The majority of the compounds were non-toxic in this assay (IC 50 of >32 µg/mL).

Overall Lipophilicity Structure-Activity Relationships
It has been consistently observed across many classes of tuberculosis inhibitors that MIC potency usually increases for more lipophilic compounds; a phenomenon that has been attributed [22,23] to the usually lipophilic cell wall of Mycobacterium tuberculosis restricting the passive diffusion of large hydrophilic compounds. While no statistically significant correlation was seen across the whole set of quite diverse compounds in Tables 1-3, there was a modest correlation of higher MIC potency with increasing lipophilicity (Equation (1)) for the small but more tightly-defined set of eleven 5-(substituted phenyl) compounds 28-38 of Table 2 Table 1 records MABA and LORA data (MIC 90 , µg/mL) for a series of compounds 11-21 bearing a range of differing substituents off the 3-phenyl ring. The 4-F and 4-OMe substituted compounds were clearly the most potent in both the MABA (>5.5-fold) and LORA (>7.5-fold) assays, possibly by blocking metabolism [24], and the 4-F substitution in this ring was thus employed in all the later compounds in Tables 2 and 3. The lack of activity for compounds with an ortho substituent could be the result of an unfavourable change in torsion angle between the pendant aryl group and the pyrazolopyrimidine core due to increased steric hindrance. However, electronic or steric arguments cannot be used to rationalise the difference in activity between active (18,20) and inactive (19, 21) para substituted compounds.
With the 3-phenyl ring substituent fixed as 4-F, a more extensive SAR study was carried out on the 5-substituent ( Table 2). The compounds in Table 2 explore structure-activity relationships for pyrazolopyrimidine 5-substituents, from hydrogen (compound 22) and simple alkyls or alkenyls (compounds 23-27), and for a series of additional substituted phenyl (compounds 28-38) and other heteroaromatic substituents (compounds 39-44). Compounds 22-27 show that those with no 5-substituent, or with a series of linear saturated and unsaturated alkyl/alkenyl groups, retain good in vitro anti-mycobacterial activities (MABA and LORA MICs) from 0.2 to 1.5 µg/mL. The two 2-(substituted phenyl) analogues (28, 29) suggested (not unexpectedly) that only very small substituents (F) were permitted at this site, and bulky substituents at this position may induce an unfavourable change in torsion angle between the pendant 5-aryl group and the pyrazolopyrimidine core. The set of 4-(substituted phenyl) compounds (32-38), despite covering substituents with a wide variation in bulk, lipophilic and electronic properties, showed similar and quite potent MICs (from 0.2-3.8 ug/mL), suggesting significant bulk tolerance at this position. Compounds 39-44 carry aromatic rings other than phenyl at the 5-position, resulting in some compounds of generally lower lipophilicity (clogP below about 4) while retaining activity, with the exception of 40.
Going forward, it was therefore decided to retain a 3-(4-fluoro)phenyl group and an unsubstituted 5-phenyl group for the subsequent SAR study of the 7-(2-pyridylmethylamine) group structure-activity relationships (Table 3; compounds 45-72). In this series, small substituents with varied electronic and lipophilicity characteristics were used in each available position. Compounds 45-48 showed that neutral groups with varying physicochemical properties at the 6 -position were not favoured, whereas powerful electron donors resulted in active compounds (49-51).
The compounds of Table 3 explore the SAR for substituents on the 7-(2-pyridylmethylamine) benzyl group, while holding the 3-and 5-substituents constant as 4-fluorophenyl and phenyl, respectively. Compounds 45-51 explore substituents on the 6 -position, ortho to the pyridine nitrogen. Only compounds 49-51, with a variety of aliphatic amine substituents, were active, with MICs of around 1 µg/mL. In contrast, a series of compounds (52-59) bearing 5 -substituents of different electronic, lipophilic and steric properties all showed modest to good activity (MICs of 1-7 µg/mL).
Finally, a small group of 3 -substituted analogues (compounds 69-72) showed a similar dependence of activity on substituent electronic properties, with electron-donating substituents providing more potent anti-mycobacterial inhibition, while those with halogen or electron-withdrawing substituents at the 3 -position were inactive.
A representative subset of the more active compounds from the cell culture assays were evaluated for their microsomal stability and hERG inhibitory properties, and the results are provided in Table 4.  were considerably more stable (T 1 ⁄2 values from 109 to >145 min) and were also much more lipophilic (average clogP 5.88 compared to 4.39). Apart from compounds 28, 39 and 68, there was very little inhibition of the hERG potassium ion channel observed by the pyrazolopyrimidines, even at 1 µM. While contributing factors for drugs to express significant hERG inhibition are high logP, high basicity and drug flexibility [25], the above three compounds do not stand out from the others in these properties.

General Information
Final products were analysed by reverse-phase HPLC (Alltima C18 5 µm column, 15 × 3.2 mm; Alltech Associated, Inc., Deerfield, IL, USA) using an Agilent HP1100 equipped with a diode-array detector. It was run using mobile phases with 80% CH 3 CN/20% H 2 O (v/v) in 45 mM NH 4 HCO 2 at pH 3.5 and 0.5 mL/min. The purity level was determined by monitoring at 330 ± 50 nm and was ≥95% for all final products. NMR spectra were obtained on a Bruker Avance 400 spectrometer at 400 MHz for 1 H and 13 C. Low-resolution mass spectra (LRMS), using atmospheric pressure chemical ionisation (APCI), were measured on a ThermoFinnigan Surveyor MSQ mass spectrometer, connected to a Gilson autosampler. High resolution mass spectra (HRMS) were obtained using an Agilent G6530B Q-TOF spectrometer and are reported as M + H. Melting points were determined on an Electrothermal 9100 melting point apparatus. Copies of the 1 H and 13 C Spectra for compounds that progressed to advanced testing in Table 4  . Apart from compounds 28, 39 and 68, there was very little inhibition of the hERG potassium ion channel observed by the pyrazolopyrimidines, even at 1 μM. While contributing factors for drugs to express significant hERG inhibition are high logP, high basicity and drug flexibility [25], the above three compounds do not stand out from the others in these properties.

General Information
Final products were analysed by reverse-phase HPLC (Alltima C18 5 μm column, 15 × 3.2 mm; Alltech Associated, Inc., Deerfield, IL, USA) using an Agilent HP1100 equipped with a diode-array detector. It was run using mobile phases with 80% CH3CN/20% H2O (v/v) in 45 mM NH4HCO2 at pH 3.5 and 0.5 mL/min. The purity level was determined by monitoring at 330 ± 50 nm and was ≥95% for all final products. NMR spectra were obtained on a Bruker Avance 400 spectrometer at 400 MHz for 1 H and 13 C. Low-resolution mass spectra (LRMS), using atmospheric pressure chemical ionisation (APCI), were measured on a ThermoFinnigan Surveyor MSQ mass spectrometer, connected to a Gilson autosampler. High resolution mass spectra (HRMS) were obtained using an Agilent G6530B Q-TOF spectrometer and are reported as M + H. Melting points were determined on an Electrothermal 9100 melting point apparatus. Copies of the 1 H and 13 C Spectra for compounds that progressed to advanced testing in Table 4 are available in the Supplementary Materials.
Preparation of compounds 11-21 of Table 1.   13 13 (19.63 g, 0.111 mol). The mixture was refluxed for 16 h, then cooled, the solvent was removed in vacuo and the residue was dissolved in water (400 mL). The solution was stirred rapidly and acidified with concentrated HCl to pH 2. The off-white precipitate was filtered and dried to give crude 3-(4-fluorophenyl)pyrazolo[1,5-a]pyrimidine-5,7-diol (73) which was used directly without further purification. 1

General Suzuki Procedure
A mixture of 76 (0.100 g, 0.22 mmol), boronic acid (0.88 mmol, 4 eq.) and Na 2 CO 3 (0.140 g, 1.32 mmol) in toluene (5.5 mL) and water (1.5 mL) was purged with nitrogen in a sealable tube. Pd(PPh 3 ) 4 (0.051 g, 0.044 mmol) was added, and the mixture was purged with nitrogen, sealed and then heated to reflux under nitrogen for 4 h. The mixture was partitioned between EtOAc and water, and the organic fraction was dried (MgSO 4 ) and evaporated onto silica gel. Column chromatography on silica gel gave the Boc-protected products.

General Deprotection Procedure
A solution of the Boc-protected product (0.12 mmol) in TFA (10 mL) and DCM (10 mL) was refluxed for 3 h. The solvent was evaporated, and the residue was partitioned between EtOAc, water and saturated aqueous NaHCO 3 (until basic). The organic fraction was dried (MgSO 4 ) and evaporated onto silica gel. Column chromatography on silica gel gave the product, which was recrystallised from DCM/heptane by evaporation of DCM to give compounds 22-44.