|Home | About | Journals | Submit | Contact Us | Français|
Donald A. Patrick: 0000-0001-7423-1379
J. Robert Gillespie: 0000-0002-5650-0333
A previous publication from this lab (Patrick, et al. Bioorg. Med. Chem. 2016, 24, 2451–2465) explored the antitrypanosomal activities of novel derivatives of 2-(2-benzamido)ethyl-4-phenylthiazole (1), which had been identified as a hit against Trypanosoma brucei, the causative agent of human African trypanosomiasis. While a number of these compounds, particularly the urea analogues, were quite potent, these molecules as a whole exhibited poor metabolic stability. The present work describes the synthesis of 65 new analogues arising from medicinal chemistry optimization at different sites on the molecule. The most promising compounds were the urea derivatives of 2-aryl-benzothiazol-5-amines. One such analogue, (S)-2-(3,4-difluorophenyl)-5-(3-fluoro-N-pyrrolidylamido)benzothiazole (57) was chosen for in vivo efficacy studies based upon in vitro activity, metabolic stability, and brain penetration. This compound attained 5/5 cures in murine models of both early and late stage human African trypanosomiasis, representing a new lead for the development of drugs to combat this neglected disease.
Human African trypanosomiasis (HAT) is an insect vectored disease caused by infection of either Trypanosoma brucei gambiense or Trypanosoma brucei rhodesiense. Despite the decreased incidence of the disease during the first decade of this century,1 the World Health Organization (WHO) currently estimates 20000 actual cases with 65 million people at risk and that over 98% of cases are due to T. b. gambiense [http://www.who.int/mediacentre/factsheets/fs259/en/]. The disease progresses from a first stage (hemolymphatic) infection to second stage (CNS) infection after the parasites have crossed the blood–brain barrier (BBB). The T. b. gambiense infection is characterized by a slow progression of months or even years from the early to late stage infection, while the T. b. rhodesiense infection progresses more rapidly. Late stage HAT is always fatal in untreated patients regardless of the causative parasite.2,3
The need for new anti-HAT drugs persists, as current drugs are few, antiquated, toxic, prone to resistance, and require parenteral administration. Treatments for T. b. rhodesiense infections are limited to suramin (a polysulfonated naphthylurea) for early stage and melarsoprol (an organoarsenical) for late stage disease. First line treatments for T. b. gambiense infections include pentamidine (an aromatic diamidine) for early stage and nifurtimox–eflornithine combination therapy (NECT) for late stage disease. Neither drug currently used against early stage disease is able to cross the BBB.2,4,5
Pafuramidine,6,7 an oral prodrug of 2,5-bis(4-amidinophenyl)furan (furamidine),8,9 advanced to phase III clinical trials against early stage HAT but was abandoned after the observation of nephrotoxicity in an expanded phase I trial.10 The nitrated imidazole derivative fexinidazole11 and the benzoxaborole SCYX-715812 have also entered clinical trials in recent years. The high attrition rate of clinical trial candidates (around 90% of those entering phase I)13 underscores the need for a pipeline of new trypanocidal compounds.
Following the failure of pafuramidine in clinical trials, the emphasis of this lab has shifted from amidines to other classes of compounds. A phenotypic high-throughput screen (HTS) of a 700000 compound library performed by the Genomics Institute of the Novartis Research Foundation (GNF)3 led to the identification of over 1000 compounds (grouped into over 100 distinct scaffolds) that inhibited growth of T. brucei in vitro at concentrations below 3.6 μM and were nontoxic to mammalian cells (Huh7). Promising results based on the synthesis and in vitro activity of analogues of one of these scaffolds, an oxazolopyridine derivative, have been reported.3 Another hit from this HTS was 2-(2-benzamido)ethyl-4-phenylthiazole (1, Table 1). A number of analogues of this compound were highly potent against the parasite but had poor metabolic stability.14 This paper describes continued work on the latter scaffold.
The previous report describes the syntheses of the initial hit (1, Table 1) and over 70 analogues (including 2–8).14 Traditional medicinal chemistry approaches for hit to lead optimization were followed as the biochemical targets of these compounds were unknown. These derivatives had alterations primarily at either of the two terminal rings, but the SAR of the internal portion of the molecule remained largely unexplored. The present work explores modification of other sites of the molecule with the syntheses of 65 novel compounds. The new compounds 9 and 10 are thiocarbonyl analogues of existing compounds 1 and 4. Analogues 11–19 are derivatives of the highly potent urea 5 bearing various substituents on the piperidine ring. Modification of the ethylene linker, either by alkylation or unsaturation, gave rise to compounds 20–28. Further alteration of the internal portion of the molecule led to fused ring analogues including benzimidazole 29, benzoxazoles 30–32, and benzothiazoles (33–73, Table 2). In all cases except analogues 71–73, the three-carbon bridge between the amide and thiazole nitrogen atoms of 1 is retained.
The treatment of amide 1 with Lawesson’s reagent in THF gave thioamide 9 (Scheme 1). The reaction of amine 74,14 carbon disulfide, and trimethylamine in DCM overnight, followed by the addition of di-tert-butyl dicarbonate and DMAP at 0 °C,15 give the crude isocyanate derivative, which was reacted with piperidine to give thiourea 10, isolated as its hydrochloride salt.
While piperidyl urea 5 was readily prepared by the reaction of the primary amine 75 and piperidine-1-carbonyl chloride,14 alternate strategy was required for substituted piperidine ureas 11–19 where the corresponding N-carbonyl chlorides were not readily available. The use of triphosgene to generate these derivatives from substituted piperidines was generally unreliable as the reactions proceeded very slowly and were difficult to monitor in the absence of UV chromophores. For example, 4-fluoropiperidine was reacted with triphosgene for 4 days before the addition of 75 to afford urea 16. Attempts at preparing the isocyanate derivative of 75 using triphosgene resulted in varying amounts of the undesired symmetric urea, regardless of the reaction conditions employed, due to the high reactivity of the primary amine. The reaction of 75 with 1,1′-carbonyldiimidazole (CDI)45,46 in DCM in the presence of Et3N at 0 °C resulted in the selective formation of the desired carbonylimidazolide derivative in situ. Addition of the appropriate secondary amine to the reaction mixture gave piperidine ureas 11–15, 17, and tetrahydroisoquinoline urea 19. However, this method was not amenable to analogue 18 (presumably due to steric hindrance), which was successfully prepared from of 75 and the N-carbonyl chloride of 1,2,3,4-tetrahydroquinoline generated in situ using triphosgene.
The analogues with modified linkers include thiazole analogues having geminal dimethyl groups on the carbon adjacent to the amide nitrogen (20–21) or on the carbon attached to the thiazole ring (22–23). Analogues 24–27 have the methyl groups on the carbon next to the amide nitrogen as well as a reversed substitution pattern on the thiazole ring. Compound 28 incorporates a vinyl linker.
The syntheses of compounds 20 and 21 (Scheme 2) began with the conjugate addition of hydrazoic acid (NaN3, concd HCl, DCM)16 to 3-methylbut-2-enal (76) to obtain 3-azido-3-methylbutanal (77).17 The aldehyde was reacted with hydroxylamine, and the resulting oxime was dehydrated to give 3-azido-3-methylbutanenitrile using acetic anhydride. The azide underwent catalytic hydrogenation (60 psi, 10% Pd/C, MeOH) in the presence of di-tert-butyl dicarbonate to afford 3-boc-amino-3-methylbutanenitrile. The nitrile was converted to thioamide 78 using sodium hydrosulfide hydrate, magnesium chloride, and 15-crown-5 in DMF.14,18 This intermediate was reacted with 2-bromoacetophenone or 2-bromo-3′-fluoroacetophenone in refluxing ethanol (effecting thiazole ring closure and nitrogen deprotection) to give amines 79 and 80, which precipitated from the reaction mixture as their hydrobromide salts. Benzoylation of the amines (benzoyl chloride, Et3N, THF) gave target compounds 20 and 21.
Ethyl cyanoacetate (81, Scheme 3) underwent methylation (CH3I, Ce2CO3, DMF) to its 2,2-dimethyl derivative19 followed by treatment with ammonia in methanol to give cyano-amide 82.20 Catalytic hydrogenation of the nitrile (H2, 60 psi, 5% Rh/Al2O3, concd NH4OH, EtOH) to the amine (avoiding the use of Raney nickel21 or circuitous pathways22) followed by protection with di-tert-butyl dicarbonate gave the amido-carbamate 83, which was reacted with Lawesson’s reagent to afford thioamide 84. This intermediate was reacted with 2-bromoacetophenone as above to give amine 85 as the hydrobromide salt, which underwent acylation as above to target compounds 22 and 23.
The synthesis of target compounds 24–27 (Scheme 4) began with the conjugate addition of hydrazoic acid (NaN3, concd HCl, DCM)16 to mesityl oxide (86) to give 4-azido-4-methylpentan-2-one.23 Bromination of the α-methyl carbon followed by a Hantzsch thiazole synthesis involving thiobenzamide and then catalytic hydrogenation of the azido group gave amine 87, isolated as its HCl salt. The reaction of 87 with the appropriate acyl halide as above gave benzamide 24, and ureas 25–26. The decreased reactivity of the sterically hindered amine 87 allowed the use of triphosgene for selective formation of its isocyanate derivative, which was then quenched with azepane to prepare urea 27 in the absence of the less readily available azepane-1-carbonyl chloride.
The preparation of compound 28 (Scheme 5) began with a Hantzsch thiazole synthesis involving thioamide 88 (prepared from the corresponding amide and P2S5)24 and α-bromo ketone 89 followed by acidic hydrolysis of the acetal. The resulting aldehyde was reacted with triethyl phosphonoacetate and excess sodium hydride in THF by a modification of a known procedure25 to give the vinyl acid 90. The acid was treated with ethyl chloroformate in acetone in the presence of Et3N followed by the addition of sodium azide26 to give acyl azide 91. A Curtius rearrangement of the azide to the vinyl isocyanate was effected in refluxing toluene (in place of benzene)26 followed by the addition of piperidine to give the vinyl urea 28.
The syntheses of the benzimidazole and benzoxazole derivatives are depicted in Scheme 6. The reaction of 4- nitrobenzene-1,2-diamine (92) with benzoyl chloride (THF at −10 °C) resulted in the selective formation of amide 93, which underwent ring closure to benzimidazole 94 in the presence of boron trifluoride etherate in refluxing dioxane. Reduction of the nitro group using tin(II) dichloride in refluxing concentrated HCl gave amine 95, which underwent benzoylation in DCM to target compound 29, the HCl salt of the known free base.27 4-Hydroxybenzene-1,3-diamine dihydrochloride (96) was reacted with benzoic acid in PPA to give 2-phenyl-5-aminobenzoxazole (97). Amine 97 underwent benzoylation to give amide 30 or reaction with triphosgene and triethylamine in DCM followed by addition of the appropriate secondary amine to obtain ureas 31 and 32.
The syntheses of benzothiazoles 33–70 (Scheme 7) began with the acylation of 2-chloro-5-nitroaniline (98a) with the appropriate benzoyl chloride in pyridine to obtain N-phenyl-benzamides 99a–g. Formamide 99h28,29 was prepared from bromoaniline 98b in refluxing formic acid. Amides 99a–h were reacted with sodium sulfide nonahydrate and sulfur in refluxing ethanol30 to give 5-nitrobenzothiazoles 100a–h. These intermediates plus commercially available 100i were reduced to the corresponding amines 101a–i using iron powder and ammonium chloride in refluxing aqueous ethanol30 after the failure to obtain 101a either by catalytic hydrogenation or stannous chloride reduction of 100a. The target benzothiazole amides 33–35, 37, 42, and 45 were prepared from amines 101a–c and the appropriate acyl chloride in the presence of Et3N in DCM. Thiazole 2-carboxylic acid was treated with 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]-pyridinium 3-oxide hexafluorophosphate) (HATU) in the presence of Et3N in THF followed by the addition of amine 101a to obtain amide 36. Urea analogues 38–41, 43–44, and 46–70 were prepared by reaction the isocyanate derivatives of primary amines 101a–i (generated in situ using triphosgene) with the appropriate secondary amine. Initially (for 38, 40, 41, and 43), the isocyantes were prepared by dropwise addition of a solution of triphosgene (in DCM) to a solution of the primary amine and triethylamine in DCM at −5 °C in order to avoid formation of the undesired symmetric urea. In subsequent reactions, solid triphosgene was added to the reaction mixture in a single portion, producing the same result provided the concentration of the primary amine was below 20 mM. The 3-aminopyrrolidinyl urea 62 was prepared from 3-Boc-aminopyrrolidine and 101g, followed by amine deprotection using TFA. The hydrochloride salt of (3R,4R)-3,4-difluoropyrrolidine,31 the secondary amine precursor to analogue 59, was prepared in five steps from L-tartaric acid by modification of known procedures.32–37
Compound 71 (Scheme 8) is a regioisomer of 55 in which the urea function is attached to the 6-position of the benzothiazole system. A pathway to intermediate 103 analogous to that of 100g beginning with 2-chloro-4-nitroanaline was unsuccessful with difficulties encountered at the ring closure step. Intermediate 103 was prepared by diazotization–bromination of 2-amino-6-nitro-benzothiazole (102) followed by a Suzuki coupling with 3,4-difluorophenylboronic acid catalyzed by tetrakis(triphenylphosphine)palladium(0). The two remaining steps in the pathway are analogous to the preparation of 55.
Compounds 72 and 73 (Scheme 9) are regioisomers of 71 and 55, respectively, in which the substitution patterns on the benzothiazole system are reversed. A Suzuki coupling between 2-amino-6-bromobenzothiazole (105) and 3.4-difluorophenylboronic acid catalyzed by Pd(dppf)Cl2 gave intermediate 106, after the failure of reaction conditions using Pd(PPh3)4 analogous to those employed to prepare 103. Amine 106 was reacted with CDI, followed by treatment of the resulting intermediate with 3-fluoropyrrolidine hydrochloride and trimethylamine in DCM to give urea 72. While the preparation of intermediate 108 from 107 was analogous to that of 106 from 105, different conditions were required for the formation of urea 73. Amine 108 was treated with phenyl chloroformate, and the resulting carbamate was reacted with 3-fluoropyrrolidine hydrochloride in refluxing pyridine to give the desired product 73.
Compound 1 (Table 1) and over 70 analogues (including 2–8) were tested previously against T. b. rhodesiense.14 The potency of 1 was enhanced 30-fold by replacement of the aromatic ring attached to the carbonyl group (R1) with piperidine (urea analogue 4). Smaller enhancements of activity were observed upon fluorination of the opposite aromatic ring (R2) at any position or in the 2,4- and 3,4-difluorophenyl derivatives. Incremental enhancements were achieved by combining favorable modifications at both sites (e.g., 5, the most potent overall). Little change in activity resulted when the substitution pattern on the thiazole ring was reversed, as observed in multiple pairs of regioisomers.
The new compounds were tested against T. b. brucei (a different parasite). To provide a more accurate comparison of activities of the new and existing compounds, a small number of the existing compounds (1–8) were also tested against T. b. brucei. A greater than 3-fold disparity between the EC50 values against the two different parasites was observed for 1, but these differences were less than 2-fold for 2–8.
None of the new thiazoles 9–28 showed enhanced potencies relative to their existing counterparts. The replacement of the carbonyl oxygen with sulfur resulted in an 8-fold loss of activity (1 vs 9) or similar potency (4 vs 10). The potency of 5 was not enhanced by the introduction of substituents on the piperidine ring (11–19). Small losses of activity (around 3-fold) were observed in compounds having of 2-methyl (11), 3-fluoro (14), and 4-fluoro substituents (16) on the piperidine ring, all having EC50 values below 50 nM. Greater losses of activity were observed with the 2,6-dimethyl (12), 4-oxo (13), 3,3- and 4,4-difluoro (15 and 17) analogues, with EC50 values between 100 and 700 nM. The tetrahydroquinoline (18) and -isoquinoline (19) derivatives were the least active of this subset, with EC50 values above 5 μM. The presence of geminal dimethyl groups on the ethylene linker (20–27) generally resulted in no detectable activity, regardless of the position of the methyl groups or thiazole geometry. The introduction of a trans-double bond in the linker resulted in an 8-fold loss of potency (5 vs 28).
Activities of the fused ring analogues 29–73 are shown in Table 2. Neither the 5-benzamidobenzimidazole (29) nor -benzoxazole (30) derivatives showed detectable activity, while the corresponding benzothiazole 33 exhibited an EC50 value of 2.38 μM. The benzothiazole urea derivatives 38 (EC50 = 325 nM) and 40 (EC50 = 366 nM) were around seven times more potent than amide 33 as well as 8–16 times more potent than the corresponding benzoxazoles 31 and 32 (EC50 ≥ 2.8 μM). After these findings, all further efforts were focused on the benzothiazole derivatives. The thiophenyl amides 34 and 35 were slightly more potent than benzamide 33, but the thiazolyl amides 36 and 37 were less potent, Of the amides 33–37, only 34 exhibited submicromolar potency but was still less potent than urea 38. Insertion of a sulfur atom into the pyrrolidine ring or a double bond into the piperidine ring resulted in 3–4-fold losses of potency (38 vs 39, 40 vs 41).
The effect of fluorination of the aromatic ring attached to the 2-position of benzothiazole (R2) upon activity depended upon the position of the fluorine atom(s). The potency of amide 33 was only slightly enhanced (less than 2-fold) by introduction of either a 3-fluoro (42) or 4-fluoro (45) substituent. The potency of pyrrolidyl urea 38 was either slightly diminished, enhanced 2-fold, or diminished 3-fold by introduction of a 4-, 3-, or 2-fluoro substituent (43, 46 (EC50 = 158 nM), and 50, respectively). A similar pattern was observed for piperidyl urea 40, as seen in analogues 44, 47, and 51. The potency of pyrrolidyl urea 46 bearing a 3-fluorophenyl group was enhanced by the addition of a second fluorine atom at the 4-position (54, EC50 = 91.8 nM). However, the 2,3-difluoro analogue 52 was similar in potency to 46, while the 2,4-difluoro isomer 53 was nearly six times less potent.
Further modifications were made to the urea portion of the molecule. The replacement of piperidine with N-methylpiperazine resulted in an 18-fold loss of activity (47 vs 48). No activity was detected with the piperazine-1,4-dicarboxamide derivative 49.
The potency of pyrrolidyl urea 54 was slightly enhanced by the introduction of a 3-fluoro substituent on the pyrrolidine ring (55, EC50 = 51.9 nM). The (S)-enantiomer 57 (EC50 = 34.8 nM) proved to be the active stereoisomer, being nearly 10 times more potent than (R)-enantiomer 56 and the most potent overall in this study. The 3,3-difluoro analogue 58 (EC50 = 934 nM was over 25 times less active than 57 and over 10 times less active than 54. The trans-3,4-difluoro derivative 59 (in which either of the two aliphatic fluorines has the same spatial orientation as the aliphatic fluorine of 57) had an EC50 value of 114 nM, being about three times less potent than 57 but about eight times more potent than 58.
Replacement of the 3-pyrrolidyl fluorine atom of 55 with other substituents led to decreased activity in all cases. The trifluoromethyl analogue 60 was over 30 times less active. The electron withdrawing cyano derivative 61 as well as the electron donating amino, dimethylamino, hydroxyl, and methoxy analogues (62–65) were all substantially less potent than 55, with EC50 values between 1.5 and 8 μM.
The activity of piperidyl urea 66 was slightly diminished by the introduction of a fluorine atom at the 4-position of the piperidine ring (67), in contrast to the results noted above with the pyrrolidyl ureas (54 vs 55). The 4,4-difluoro derivative 68 was over 10 times less active than 66, similar to the disparity between pyrrolidyl urea 54 and its 3,3-difluoro analogue 58.
The presence of an aromatic substituent on the benzothiazole system, as well as the spatial orientation of both substituents, proved to be essential to activity. Removal of the aromatic ring from the 2-position (69) or replacement by a methyl group (70) resulted in loss of activity. The relocation of the urea moiety from the 5- to the 6-position also resulted in no detectable activity (55 vs 71). The reversal of the substitution patterns on the benzothiazole system (71 vs 72, 55 vs 73) resulted in weak activity (EC50 > 2 μM). Thus, the spatial orientation of the thiazole and amide nitrogens proved to be essential to activity.
Similarities and differences existed between the SAR of the thiazole14 and benzothiazole derivatives. In both groups, the urea derivatives were more potent than the corresponding benzamides. Piperidyl ureas were more potent than the corresponding pyrrolidyl ureas among the thiazoles, but the opposite was true for the benzothiazoles. Monofluorination of the aromatic ring at any position was generally beneficial in the thiazoles, but monofluorination was beneficial only at the 3-position for the benzothiazoles. Aromatic 3,4-difluorination was beneficial in both groups, while 2,4-difluorination was beneficial only to the thiazoles. The introduction of any substituents on the piperidine ring resulted in losses of activity regardless of the internal core of the molecule.
Cytotoxicity data for compounds 1–8 were reported previously.14 All new compounds with EC50 values below 1 μM (except 53) were tested for toxicity against human lymphocytes CRL-8155 and human hepatocytes HepG2 (Table 3). Most of the compounds (including 57) exhibited no detectable toxicity to either cell line. Only compounds 54 and 59 had detectable toxicity to the CRL-8155 cells (EC50 = 33.9 and 27.1 μM respectively), and none had detectable toxicity to the HepG2 cells. The selectivity of each compound for the parasite over each of the two cell lines was calculated. Compounds 10, 11, 14, 16, and 57 exhibited selectivity indices above 1000 against either cell line. Seventeen analogues (12, 15, 17, 28, 38, 40, 43, 44, 46, 47, 52, 54, 55, 56, 59, 66, and 67) had selectivity indices above 100 against either cell line. Within this group, compound 28 had a selectivity index above 500 in the CRL-8155 assay, while 54 and 55 had selectivity indices above 500 in the HepG2 assay.
Select compounds were assayed for stability to mouse and human liver microsomes (Table 4). The short half-lives of compounds 4 and 5 are consistent with previous results obtained using a different protocol.14 The introduction of an additional aromatic fluorine atom (5 vs 6, 5 vs 7) offered no advantage regarding metabolic stability.
None of the new thiazole derivatives (9–28) that were tested were metabolically stable. Only analogue 28 (t1/2 = 13 min) had a half-life above 10 min in the mouse microsomes, and all others in this group had half-lives below 5 min with 0–21% of the test compounds remaining after the 5 min time point. These compounds were also unstable to the human micro-somes, where only analogue 14 had a half-life as high as 10 min. All four compounds in this group having potencies below 50 nM (10, 11, 14, and 16) were unstable to both mouse and human microsomes. No modifications of the urea portion of the molecule or of the ethylene bridge were found to improve metabolic stability. Thus, none of the compounds 9–28 offered any advantage regarding either potency or metabolic stability.
The metabolic stabilities of the fused ring analogues 29–73 were more promising. All compounds in this group that were tested had half-lives greater than 10 min in the mouse microsomes and at least 30 min (except for 47) in the human microsomes. The inactive benzoxazole amide 30 and its weakly active piperidyl urea analogue 32 had half-lives of between 10 and 20 min in mouse microsomes and at least 60 min in human microsomes. The benzothiazole amide 33, although only weakly active against the parasite, had half-lives of 50 and >60 min in the mouse and human microsomes, respectively. The corresponding pyrrolidyl and piperidyl ureas 38 and 40, although more potent had 33, had shorter half-lives (>20 min) in in mouse microsomes, and 38 had a shorter half-live (30 min) in human microsomes. Fluorination of the phenyl ring at the 3-position (38 vs 46, 40 vs 47) led to enhanced potency of both ureas but not enhanced stability, as both fluorinated derivatives 46 and 47 were less stable to mouse microsomes and only 46 was at least as stable to human microsomes. The pyrrolidyl urea 46 (R2 = 3-fluorophenyl, t1/2 = 37 min) was four times more stable to human microsomes than corresponding piperidyl urea 47. Urea 54, the (more highly potent) 3,4-difluorophenyl analogue of 46, showed similar stability (t1/2 = 14 min) to mouse microsomes but enhanced stability (t1/2 > 60 min) to human microsomes. The addition of a 3-fluoro substituent on the pyrrolidine ring (55) resulted in greater potency, greater than 4-fold increased stability to mouse microsomes, and similar stability to human microsomes compared to 54. High metabolic stability (t1/2 > 60 min in both mouse and human microsomes) was also observed in the (S)-enantiomer 57 and the trans-3,4-difluoropyrrolidine derivative 59.
Compounds 54, 55, and 57, and 59 were administered to groups of three mice in single doses of 50 mg/kg by oral gavage in a dosing vehicle consisting of Tween 80 (7%), EtOH (3%), and DMSO (5%) in 0.9% sodium chloride solution (Table 5). Compound 57 (the most potent in vitro) clearly showed the highest blood levels of the four compounds tested in the same vehicle, with a maximum blood concentration of 23.3 μM and an AUC of 13216 min·μM. Compound 59 had a lower Cmax but a similar AUC compared to 57. Less than 1% of the Cmax of 57 remained at the 24 h time point compared to 34% of the Cmax of 59. More than the 17% of the Cmax of 59 remained at the 32 h time point. Compound 57 gave a PK profile similar to that obtained using the original vehicle when administered in a vehicle consisting of Phosal 53 MCT (60%), PEG400 (30%), and EtOH (10%); however, the use of a vehicle consisting of methylcellulose cP 400 (0.5%) and Tween 80 (0.5%) in water resulted in a lower Cmax and AUC values (lines 3–5 of Table 5).
Brain penetration studies (Table 6) were initially performed upon compounds 54, 55, and 57 but were expanded to include other compounds in order to explore the SAR of brain penetration. Groups of three mice were given single 5 mg/kg ip doses of the test compounds and were sacrificed 1 h postdose. The benzothiazole pyrrolidyl urea 38 exhibited a mean brain concentration of 4.2 μM (over 10 times greater than that of the corresponding benzoxazole analogue 31) and a mean brain to plasma ratio (BPR) of 1.2. Fluorination of the terminal aromatic ring had relatively little effect as 38, and the 3,4-difluorophenyl derivative 54 had nearly identical mean brain concentrations, although the latter did have a slightly higher BPR. By contrast, the mean brain concentrations of the 3-fluoro (46) and 2,3-difluoro (52) analogues were somewhat lower than that of 38. The brain levels of 54 were enhanced by the introduction of a 3-fluoro substituent on the pyrrolidine ring (55–57). Of these compounds, the greatest enhancement was seen in the (S)-enantiomer 57, with a mean brain concentration of 9.1 μM and a BPR of 4.0. The (R)-enantiomer 56 also exhibited the ability to cross the BBB. The highest brain levels were seen in the difluoropyrrolidyl derivatives 58 and 59, each having a mean brain concentration exceeding 16 μM and a BPR above 7.5. Replacement of the fluorine atom with a trifluoromethyl group (55 vs 60) resulted in lower brain levels. Similar decreases also resulted when the pyrrolidine ring was replaced with piperidine (54 vs 66). Introduction of a 4-fluoro substituent on the piperidine ring also resulted in enhanced brain levels (66 vs 67), consistent with the results of fluorination of the pyrrolidine ring.
The brain and plasma concentrations of compound 57 were studied in a time-course experiment following oral dosing at 50 mg/kg (Table 7). The data show that 57 partitions to the brain compartment with BPRs of 1.48, 4.43, and 5.48 at 1, 4, and 8 h postdose, respectively. At 24 h, the plasma concentrations are undetectable and brain levels are down to submicromolar concentrations. Compared to the previous experiment (Table 6), the BPR ratio for 57 is lower at 1 h postdose (1.48 vs 4.0), probably due to the oral route of administration (as opposed to the ip route used before), leading to slower systemic absorption and distribution. The time-course experiment demonstrates that brain concentrations of 57 are sustained at high levels for at least 8 h postdose and help account for the successful results in the late-stage efficacy model discussed below.
Upon the basis of its promising in vitro activity, cytotoxicity, metabolic stability, PK, and brain penetration data, compound 57 was selected as a candidate for in vivo efficacy studies. An acute model (requiring 60 days to complete) was employed, followed by a chronic model (requiring 180 days to complete).
In a model of the acute phase of HAT, five mice infected with T. b. rhodesiense STIB900 were given the test compound at 50 mg/kg po b.i.d. × 4 days, beginning 2 days postinfection. Efficacy PK blood samples were collected from three of the five mice prior to dose 7, and at 1 and 6 h postdose. Average plasma levels of the test compound were 9.1 ± 7.6 μM (predose), 18.6 ± 7.2 μM (1 h), and 15.0 ± 4.2 μM (6 h). Compound 57 attained 5/5 cures as determined by the absence of detectable parasitemia in any of the treated mice 60 days postinfection (Figure 1). All mice receiving only vehicle showed high parasitemia on the last day of dosing, with all concentrations >1.5 × 107 parasites/mL of blood, and were euthanized.
In a model of the chronic phase of HAT, compound 57 was administered to five mice infected with T. b. brucei TREU667 at 50 mg/kg/po b.i.d. × 10 days, beginning 21 days postinfection. This dose was chosen because earlier PK studies demonstrated good plasma exposure at 50 mg/kg (Table 5). Efficacy PK blood samples were collected from three of the five mice prior to dose 15, and at 1 and 6 h postdose. Average plasma levels of the test compound were 15.8 ± 8.7 μM (predose), 23.1 ± 603 μM (1 h), and 26.7 ± 7.1 μM (6 h). Compound 57 attained 5/5 cures as determined by the absence of detectable parasitemia in any of the treated mice through 180 days postinfection (Figure 2). No mice receiving vehicle alone had spontaneous cures. The mice receiving diminazene aceturate (which does not cross the BBB)38 had temporary clearance of parasitemia but subsequently relapsed, most likely from parasites leaving the brain and returning to the hemolymphatic system. By observing mice for 180 days, more than ample time was allowed for mice to show signs of illness or parasites to become visible on blood films (as was observed with the diminazene-treated mice), thus providing assurance that the 57-treated mice were cured. Experiments evaluating lower doses of 57 are underway will be published at a later time.
While a number of the initial analogues14 were highly potent against the parasite in vitro, they exhibited poor metabolic stability, which precluded their further advancement as drug candidates. The goal of the present study was to improve metabolic stability while maintaining potency. Neither replacing the carbonyl oxygen with sulfur, the introduction of various substituents at the urea end of the molecule, nor the alkylation or unsaturation of the ethylene linker provided any advantage regarding either antitrypanosomal activity or metabolic stability. Enhanced metabolic stability was observed only in the fused ring analogues where thiazol-2-ethylamino fragment was replaced with a benzoxazol- or benzothiazol-5-amino moiety, but antitrypanosomal activity was retained only in the benzothiazole derivatives. These results strongly suggest that the ethylene linker played a role in the poor metabolic stability of the initial thiazole analogues. Brain penetration studies demonstrated the ability of the benzoxazole and benzothiazole derivatives, as a whole, to cross the BBB and even concentrate in the brain.
The most significant findings in this paper were the promising biological data for compound 57. This molecule exhibited good activity in vitro (EC50 = 34.8 nM), no detectable toxicity to two human cell lines (CRL-8155 and HepG2), and half-lives of at least 60 min in both mouse and human liver microsomes. It also demonstrated promising oral PK (Cmax = 23.3 μM and AUC = 13216 min × μM) and excellent brain penetration (brain concentration = 9.07 μM and BPR = 4.00) at 1 h after administration of a 5 mg/kg ip dose in mice. These results led to the selection of compound 57 for in vivo efficacy studies. Oral dosing of this molecule achieved 5/5 cures in both acute and chronic murine models of HAT. The efficacy PK experiments in both studies demonstrated that the compound accumulated in plasma after dosing. Synthesis of additional analogues, as well as further testing of compound 57 (and 59), including dose optimization and comprehensive pharmacokinetic studies, is either planned or currently in progress. Experiments to identify the biochemical target(s) of action in T. brucei are also underway.
Compounds were tested for antitrypanosomal activity against T. b. brucei (strain BF427) in HMI-9 media as previously described.39 Cells were tested in triplicate against serial dilutions of compounds and quantified with Alamar Blue (ThermoFisher Scientific, Waltham, MA) at 48 h.40 EC50 and EC90 values were calculated by nonlinear regression using software by the Collaborative Drug Database (Burlingame, CA. www.collaborativedrug.com). Growth assays for T. b. brucei included the control compound, pentamidine isethionate (Sigma-Aldrich, St. Louis, MO.
Compounds were assayed for cytotoxicity against CRL-8155 (human lymphoblasts) and HepG2 cells (human hepatocellular carcinoma) as previously described.41 Cells were exposed to serial dilutions of compounds for 48 h, and cytotoxicity was quantified using Alamar Blue41 (ThermoFisher Scientific, Waltham, MA). Compounds were assayed in quadruplicate, and CC50 and CC90 values were calculated by nonlinear regression using software by the Collaborative Drug Database (Burlingame, CA. www.collaborativedrug.com). Growth assays for CRL-8155 and HepG2 included the control compound, quinacrine (Sigma-Aldrich.
Pooled liver microsomes from mouse and human sources were purchased from BD Biosciences (San Jose, CA). Microsome reactions were preincubated for 5 min at 37 °C with 0.16 M KH2PO4, 1 mM NADPH, and 0.5 mg/mL of microsomes. At t = 0, 1.5 μM of test compounds were added and the t = 0 time point was immediately quenched with the addition of 4× the sample volume of 100% acetonitrile. Subsequent time points were quenched at 5, 10, 15, 30, and 60 min post addition of test compounds, then processed using liquid chromatography/tandem mass spectrometry analysis. Assays were performed with control compounds, testosterone, and dextromethorphan (purchased from Sigma-Aldrich).
Compounds were administered orally to mice at a concentration of 50 mg/kg in a vehicle consisting of either: (1) DMSO (5%), Tween 80 (7%), and EtOH (3%) in physiological saline (0.9%) solution, (2) Phosal 53 MCT (60%), PEG400 (30%), and EtOH (10%), or (3) methylcellulose cP400 (0.5%) and Tween 80 (0.5%) in water. Phosal 53 MCT is a lecithin emulsifier purchased from American Lecithin Co. Three mice were used per group, and 40 μL of blood was collected from the tail at time points extending out to as much as 32 h postdose into heparinized capillary tubes. Whole blood was then spotted onto FTA DMPK-C cards (GE Healthcare, Boston, MA) and was measured by liquid chromatography/tandem mass spectrometry. For the efficacy PK experiments, samples were collected from experimental mice (n = 3) before the stated day of dose, then at 1 and 6 h postdose using the same collection method as above.
Test compounds were administered at 5 mg/kg ip to three mice in a vehicle consisting of DMSO (5%), Tween 80 (7%), and EtOH (3%) in physiological saline (0.9%) solution.3 At 1 h post injection, blood was collected, plasma was separated by brief centrifugation, and the brain was removed and homogenized in acetonitrile. Levels of compound in the plasma and brain were determined via liquid chromatography/tandem mass spectrometry. Calculations of brain concentrations accounted for 3% volume/weight of blood in the brain. For the brain time-course experiment of compound 57, compounds were administered at 50 mg/kg po in the same vehicle listed above. Mice were sacrificed in groups of three at 1, 4, 8, and 24 h postdosing for collection of brains and plasma for analysis as mentioned above.
Experiments were carried out using the standard operating procedure used by WHO screening centers42 and done in compliance with the University of Washington Institutional Animal Care and Use Committee (IACUC) approved protocol. Groups of five female Swiss-Webster mice (ND4 outbred, ages 6–8 weeks) were infected on day 0 with 1 × 104 T. b. rhodesiense (strain STIB900) parasites. The test compound was administered by oral gavage at 50 mg/kg every 12 h from day 2 to day 5, for a total of eight doses in a 200 μL volume of a vehicle consisting of DMSO (5%), Tween 80 (7%), and EtOH (3%) in physiological saline (0.9%) solution. Parasitemia was monitored via microscopic analysis of tail blood for 60 d postinfection or until parasites were detected. Mice were removed from the experiment once parasites were detected in the blood, at which point the mouse was euthanized.
Again, experiments were done in compliance with the University of Washington IACUC approved protocol. According to published procedures,38 groups of five mice were infected with 1 × 104 T. b. brucei (strain TREU667) at day 0 to establish a chronic infection. Treatment began on day 21 postinfection, and mice received 50 mg/kg test compound orally b.i.d. for 10 d (total of 20 doses) in a 200 μL volume of a vehicle composed of Phosal 53 MCT (60%), PEG400 (30%), and EtOH (10%). A control group received vehicle with no compound and another control group received a single intraperitoneal dose of diminazene aceturate at 10 mg/kg in water on day 21. The diminazene aceturate temporarily clears parasites from the blood, but because it does not cross the BBB, the blood is later repopulated from parasites in the CNS.38 Post dosing, parasitemia was monitored via microscopic examination of tail blood slides until 180 days postinfection. Mice were removed from the experiment once parasites were detected in the blood.
Uncorrected melting points were measured on a Thermo Scientific 9200 melting point apparatus. 1H NMR spectra were recorded on a Varian Inova 400 MHz, a Bruker AVANCE 400 MHz, or a Varian Inova 600 MHz spectrometer. Anhydrous solvents were purchased from Aldrich Chemical Co., Milwaukee, WI, or from Fisher Scientific, Waltham, MA, in Sure-Seal or AcroSeal containers and were used without further purification. Organic starting materials were purchased from the same sources or were prepared by published procedures as noted. Reaction mixtures were monitored by TLC on silica gel or by reverse phase HPLC. Organic layers of extraction mixtures were neutralized as necessary with acidic or basic washes, washed with saturated NaCl solution, and dried over MgSO4 before being evaporated under reduced pressure. Normal phase flash column chromatography was performed using Davisil grade 633, type 60A silica gel (200–425 mesh). Analytical HPLC chromatograms were recorded on an Agilent 1100 or 1200 series chromatograph using a Zorbax Rx C8 column (4.6 mm × 75 mm, 3.5 μm) maintained at 40 °C and UV photodiode array detection at 230, 254, 265, 290, and 320 nm. Area % values are reported at the wavelengths where the strongest signals of the products were observed. Mobile phases consisted of mixtures of MeOH (0–95%) in water containing formic acid (80 mM), ammonium formate (20 mM), and Et3N (15 mM). Samples were eluted at appropriate gradients at a flow rate of 1.5 mL/min. Low resolution ESI mass spectra were recorded on an Agilent Technologies 1100 series LC/MSD trap mass spectrometer or at the North Carolina State University Mass Spectrometry Facility located in the Department of Chemistry. In cases of hydrochloride salts, the m/z values reported are those of the free bases. Elemental analyses were measured by Atlantic Microlab, Norcross, GA, and unless stated otherwise, were within ±0.4% of calculated values. All target compounds are judged to be >95% pure by elemental analysis and analytical HPLC. Representative syntheses for all target compounds as well as for intermediates of benzothiazoles 33–70 are given herein. Synthetic details for all compounds are given in Supporting Information.
Lawesson’s reagent (493 mg, 1.22 mmol) was added to a solution of 2-(2-benzamido)ethyl-4-phenylthiazole (1,14 353 mg, 1.15 mmol) in THF (20 mL). The mixture was stirred overnight before being diluted with saturated NaHCO3 solution and extracted into DCM. The oily residue was recrystallized from MeOH as light-yellow crystals (184 mg, 50%): mp 124–126 °C. 1H NMR (400 MHz, DMSO-d6) δ 10.43 (t, J = 5.5 Hz, 1H), 8.00 (d, J = 0.7 Hz, 1H), 7.99−7.92 (m, 2H), 7.80−7.72 (m, 2H), 7.54−7.46 (m, 1H), 7.46−7.37 (m, 4H), 7.37−7.29 (m, 1H), 4.12 (td, J = 6.9, 5.2 Hz, 2H), 3.49 (t, J = 6.9 Hz, 2H). EIMS m/z 324.9 (M + 1)+. HPLC 100 area % (265 nm). Anal. Calcd for C18H16N2S2: C, 66.63; H, 4.97; N, 8.63; S, 10.40. Found: C, 66.36; H, 5.16; N, 8.38.
A suspension of 4-phenylthiazol-2-ethylamine hydrobromide (74,14 500 mg, 1.75 mL) in DCM (8 mL) was treated with Et3N (0.6 mL, 4.30 mmol) to give a homogeneous solution. Carbon disulfide (1.2 mL, 19.3 mmol) was added, and the mixture was stirred overnight. After the mixture cooled in an ice bath, DMAP (10 mg, 4.8 mol %) and a solution of di-tert-butyl dicarbonate (383 mg, 1.75 mmol) in DCM (1 mL) were added.15 The ice bath was removed, and the mixture was stirred for 2 h. The mixture was concentrated under reduced pressure, diluted with ether, and filtered. The filtrate was evaporated to give 2-(2-isothiocyanatoethyl)-4-phenylthiazole as an oil. 1H NMR (400 MHz, DMSO-d6) δ 8.04 (s, 1H), 8.02−7.91 (m, 2H), 7.49−7.39 (m, 2H), 7.42−7.29 (m, 1H), 4.11 (dd, J = 6.7, 5.9 Hz, 2H), 3.44 (dd, J = 6.7, 5.8 Hz, 2H).
A solution of the 2-(2-isothiocyanatoethyl)-4-phenylthiazole, piperidine (200 μL, 2.02 mmol), and Et3N (300 μL, 2.15 mmol) in DCM (10 mL) was stirred overnight at room temperature. After aqueous workup, the crude product was converted to the HCl salt using EtOH and concentrated HCl, followed by recrystallization from EtOH/ether to give white crystals (249 mg, 39%): mp 138–140 °C. 1H NMR (400 MHz, DMSO-d6) δ 7.98 (d, J = 0.6 Hz, 1H), 7.96−7.92 (m, 2H), 7.80 (s, 1H), 7.48−7.39 (m, 2H), 7.37−7.30 (m, 1H), 3.88 (td, J = 6.8, 4.3 Hz, 2H), 3.79−3.72 (m, 4H), 3.34 (t, J = 7.0 Hz, 2H), 1.63−1.55 (m, 2H), 1.52−1.41 (m, 4H). EIMS m/z 332.0 (M + 1)+. HPLC 96.9 area % (254 nm). Anal. Calcd for C17H21N3S2·HCl: C, 55.49; H, 6.03; N, 11.42. Found: C, 55.59; H, 6.12; N, 11.31.
A mixture of 4-(3-fluorophenyl)thiazol-2-ethylamine hydrobromide (75,14 0.4–0.7 mmol) and Et3N (min of 3 equiv) in DCM was cooled to −5 °C before the addition of 1,1′-carbonyldiimidazole (CDI, 1.2–1.3 equiv). After 1 h, the secondary amine or its hydrochloride salt was added. The mixture was stirred at room temperature overnight or until complete by HPLC before being diluted with water and extracted into DCM. The crude products were then recrystallized from appropriate solvents.
11 was prepared by the general procedure from 75 (153 mg, 0.506 mmol) and 2-methylpiperidine (100 μL, 0.851 mmol). The crude product was recrystallized from DCM/hexanes as a white solid (75.1 mg, 43%): mp 81–83 °C. 1H NMR (400 MHz, DMSO-d6) δ 8.09 (s, 1H), 7.80 (ddd, J = 7.8, 1.6, 0.9 Hz, 1H), 7.74 (ddd, J = 10.7, 2.6, 1.5 Hz, 1H), 7.48 (td, J = 8.0, 6.2 Hz, 1H), 7.16 (dddd, J = 9.0, 8.3, 2.7, 0.9 Hz, 1H), 6.57 (t, J = 5.4 Hz, 1H), 4.23 (d, J = 4.3 Hz, 1H), 3.79−3.70 (m, 1H), 3.41 (dt, J = 12.6, 6.5 Hz, 2H), 3.17 (t, J = 6.9 Hz, 2H), 2.73 (td, J = 13.1, 2.8 Hz, 1H), 1.60−1.40 (m, 5H), 1.30−1.10 (m, 1H), 1.04 (d, J = 6.8 Hz, 3H). EIMS m/z 348.0 (M + 1)+. HPLC 100 area % (254 nm). Anal. Calcd for C18H22FN3OS·0.25H2O: C, 61.43; H, 6.44; N, 11.94. Found: C, 61.22; H, 6.69; N, 11.82.
Benzoyl chloride (200 μL, 1.72 mmol) was added to a stirred mixture of 2-(2-amino-2-methylpropyl)-4-phenylthiazole hydrobromide (79, 258 mg, 0.824 mmol) and Et3N (500 μL, 3.59 mmol) in THF (10 mL). The mixture was stirred overnight before being diluted with H2O and extracted into EtOAc. The product was recrystallized from EtOAc/hexanes as white crystals (191 mg, 69%): mp 107–108 °C. 1H NMR (400 MHz, DMSO-d6) δ 7.94 (s, 1H), 7.93−7.88 (m, 3H), 7.83−7.79 (m, 2H), 7.54−7.48 (m, 1H), 7.46−7.37 (m, 4H), 7.34−7.28 (m, 1H), 3.60 (s, 2H), 1.45 (s, 6H). EIMS m/z 337.1 (M + 1)+. HPLC 98.4 area % (254 nm). Anal. Calcd for C20H20N2OS: C, 71.40; H, 5.99; N, 8.33. Found: C, 71.49; H, 6.04; N, 8.22.
A solution of (E)-3-[4-(3-fluorophenyl)thiazol-2-yl]acryloyl azide (91, 140 mg, 0.514 mmol) in toluene (10 mL) was refluxed for 1.5 h, resulting in a Curtius rearrangement to the vinyl isocyanate.26 After the reaction mixture cooled to rt, piperidine (80 μL, 0.810 mmol) and Et3N (80 μL, 0.810 mmol) were added and the mixture was stirred for 2 h before being diluted with water and extracted into EtOAc. The product was purified on a column of silica eluting with hexanes/EtOAc (2:1) followed by recrystallization from EtOAc/hexanes to give a pale-yellow solid (56.1 mg, 33%): mp 101.5–103 °C. 1H NMR (400 MHz, DMSO-d6) δ 10.89 (d, J = 10.5 Hz, 1H), 8.04 (s, 1H), 7.75 (dt, J = 7.7, 1.2 Hz, 1H), 7.71 (ddd, J = 10.5, 2.6, 1.5 Hz, 1H), 7.52 (td, J = 8.0, 6.1 Hz, 1H), 7.29 (dd, J = 10.5, 8.7 Hz, 1H), 7.26−7.17 (m, 1H), 5.91 (d, J = 8.7 Hz, 1H), 3.50 (t, J = 5.4 Hz, 4H), 1.67−1.52 (m, 6H). EIMS m/z 332.1 (M + 1)+. HPLC 100 area % (290 nm). Anal. Calcd for C17H18FN3OS: C, 61.61; H, 5.47; N, 12.68. Found: C, 61.37; H, 5.48; N, 12.40.
Et3N (0.5 mL, 3.59 mmol) was added to a solution of 5-amino-2-phenylbenzothiazole (101a, 340.3 mg, 1.50 mmol) and benzoyl chloride (250 μL, 2.16 mmol) in DCM. The mixture was stirred at rt overnight before being diluted with water and extracted into DCM. Recrystallization of the product from EtOH gave white crystals (406 mg, 82%): mp 221–222 °C. 1H NMR (400 MHz, DMSO-d6) δ 10.50 (s, 1H), 8.62 (d, J = 2.0 Hz, 1H), 8.18−8.07 (m, 3H), 8.06−7.97 (m, 2H), 7.85 (dd, J = 8.7, 2.0 Hz, 1H), 7.68−7.52 (m, 6H). EIMS m/z 331.0 (M + 1)+. HPLC 99.2 area % (290 nm). Anal. Calcd for C20H14N2OS: C, 72.69; H, 4.27; N, 8.48. Found: C, 72.45; H, 4.50; N, 8.42.
A stirred solution of the appropriate primary amine (0.5–2 mmol) and Et3N (min of 2 equiv) in DCM was chilled to maximum internal temperature of −5 °C (ice–salt bath). Triphosgene (0.35–0.55 equiv) was added either dropwise as a solution in DCM (for 38, 40, 41, and 43) or all at once in solid form for all others. The method of introduction was noncritical provided the final concentration of the primary amine was below 20 mM. The mixture was maintained at −5 °C until the starting material was no longer detectable by HPLC (typically 1 h or less). The secondary amine (min of 1 equiv) or its HCl salt (in which cases an additional equivalent of Et3N was employed) was added, and the solution was allowed to warm to room temperature overnight. The reaction mixture was diluted with water and extracted into DCM. Unless stated otherwise, the product was directly recrystallized from the appropriate solvent(s).
57 was prepared from 5-amino-2-(3,4-difluorophenyl)benzothiazole (101g, 0.63 mmol) and (S)-3-fluorpyrrolidine hydrochloride. The product was purified on a silica gel column eluting with hexanes/EtOAc (1:2) and recrystallized from EtOAc/hexanes as white crystals (86 mg, 36%): mp 208–209 °C. 1H NMR (400 MHz, DMSO-d6) δ 8.50 (s, 1H), 8.33 (d, J = 2.0 Hz, 1H), 8.13 (ddd, J = 11.3, 7.7, 2.2 Hz, 1H), 7.99 (d, J = 8.8 Hz, 1H), 7.96−7.89 (m, 1H), 7.69−7.58 (m, 2H), 5.39 (dm, J = 53.3 Hz, 1H), 3.80−3.52 (m, 3H), 3.47 (td, J = 10.3, 6.9 Hz, 1H), 2.27−2.00 (m, 2H). EIMS m/z 378.0 (M + 1)+. HPLC 97.0 area % (254 nm). Anal. Calcd for C18H14F3N3OS·0.3H2O: C, 56.48; H, 3.84; N, 10.98. Found: C, 56.21; H, 3.77; N, 10.89.
A scale-up synthesis from a total of 6.6 mmol of 101g was performed in three batches. The combined product was chromatographed as above and recrystallized from EtOH as white crystals (1.59 g, 64%): mp 210–211.55 °C. HPLC 97.8% (254 nm). Anal. Calcd for C18H14F3N3OS: C, 57.29; H, 3.74; N, 11.13. Found: C, 57.32; H, 3.81; N, 11.02.
1,1′-Carbonyldiimidazole (163 mg, 1.01 mmol) was added to a solution of 2-amino-6-(3,4-difluorophenyl)-benzothiazole (106, 200 mg, 0.761 mmol) in CH3CN (10 mL). The solution was stirred overnight with the formation of a precipitate, which was filtered off and rinsed with acetone. This material was added to a solution of 3-fluoropyrrolidine hydrochloride (87.0 mg, 0.693 mmol) and Et3N (250 μL, 1.79 mmol) in DCM (10 mL). The suspension was stirred overnight with the dissolution of the newly formed product. The resulting solution was diluted with water and extracted into DCM. The crude product was recrystallized from EtOH/H2O as a beige powder (177 mg, 62%): mp 171.5–173 °C. 1H NMR (400 MHz, DMSO-d6) δ 11.36 (s, 1H), 8.26−8.20 (m, 1H), 7.81 (ddd, J = 12.4, 7.8, 2.2 Hz, 1H), 7.70 (dd, J = 8.4, 1.9 Hz, 1H), 7.65 (d, J = 8.6 Hz, 1H), 7.57 (dddd, J = 7.7, 4.7, 2.3, 0.9 Hz, 1H), 7.52 (dt, J = 10.6, 8.3 Hz, 1H), 5.39 (d, J = 53.1 Hz, 1H), 3.87−3.42 (m, 4H), 2.28−2.00 (m, 2H). EIMS m/z 378.1 (M + 1)+. HPLC 96.0 area % (290 nm). Anal. Calcd for C18H14F3N3OS·0.4H2O: C, 56.21; H, 3.88; N, 10.93. Found: C, 56.17; H, 4.00; N, 10.75.
Phenyl chloroformate (200 μL, 1.59 mmol) was added dropwise by syringe to a solution of 2-amino-5-(3,4-difluorophenyl)benzothiazole (108, 132 mg, 0.505 mmol) and pyridine (200 μL, 2.47 mmol) in THF (10 mL) maintained at −5 °C, accompanied by the formation of a precipitate. The reaction mixture was diluted with water and extracted with EtOAc. The extract was evaporated to near dryness followed by dilution with hexane to give crude phenyl (5-(3,4-difluorophenyl)benzothiazol-2-yl)carbamate (174 mg, 90%), which was used in the next step without further purification.
3-Fluoropyrrolidine hydrochloride (56.8 mg, 0.524 mmol) was added to a solution of crude (5-(3,4-difluorophenyl)benzothiazol-2-yl)carbamate (153 mg, 0.401 mmol) in pyridine (10 mL). The mixture was refluxed for 1 h. The cooled mixture was diluted with H2O (10 mL), and the resulting suspension was stored overnight at 4 °C. The precipitate was filtered off and dried to give an off-white powder (95.7 mg, 57% overall): mp >250 °C (dec). 1H NMR (400 MHz, DMSO-d6) δ 11.21 (s, 1H), 7.97 (d, J = 8.2 Hz, 1H), 7.91−7.73 (m, 2H), 7.64−7.45 (m, 3H), 5.39 (d, J = 53.0 Hz, 1H), 3.91−3.40 (m, 4H), 2.29−1.99 (m, 2H)). EIMS m/z 378.1 (M + 1)+. HPLC 96.0 area % (290 nm). Anal. Calcd for C18H14F3N3OS: C, 57.29; H, 3.74; N, 11.13. Found: C, 57.13; H, 3.83; N, 11.15.
Benzoyl chloride (2.63 g, 18.71 mmol) was added dropwise by syringe to a solution of 2-chloro-5-nitroaniline (98a, 3.06 g, 17.72 mmol) in dry pyridine (7.5 mL) under argon, resulting in the formation of a precipitate. More pyridine (7.5 mL) was added, and the mixture was stirred overnight. The reaction mixture was poured into water. The precipitated product was recrystallized from EtOH as white crystals (4.62 g, 95%): mp 171–172 °C. 1H NMR (400 MHz, DMSO-d6) δ 10.35 (s, 1H), 8.58 (d, J = 2.8 Hz, 1H), 8.13 (dd, J = 8.9, 2.8 Hz, 1H), 8.08−7.97 (m, 2H), 7.88 (d, J = 8.8 Hz, 1H), 7.69−7.61 (m, 1H), 7.61−7.52 (m, 2H). HPLC 100 area % (265 nm). Anal. Calcd for C13H9ClN2O3: C, 56.44; H, 3.28; N, 10.13. Found: C, 56.43; H, 3.19; N, 10.15.
99g was prepared analogously to 99a from 3,4-difluorobenzoyl chloride (2.77 g, 15.69 mmol) and 2-chloro-5-nitroaniline (98a, 2.60 g, 15.07 mmol). The product was recrystallized from EtOH as white crystals (3.84 g, 82%): mp 149–51 °C. 1H NMR (400 MHz, DMSO-d6) δ 10.50 (s, 1H), 8.53 (d, J = 2.7 Hz, 1H), 8.15 (dd, J = 8.9, 2.8 Hz, 1H), 8.06 (ddd, J = 11.4, 7.8, 2.2 Hz, 1H), 7.95−7.89 (m, 1H), 7.89 (d, J = 8.9 Hz, 1H), 7.67 (dt, J = 10.5, 8.3 Hz, 1H). HPLC 99.1 area % (265 nm). Anal. Calcd for C13H7ClF7N2O3: C, 49.94; H, 2.26; N, 8.96. Found: C, 49.87; H, 2.29; N, 8.98.
A mixture of N-(2-chloro-5-nitrophenyl)benzamide (99a, 3.00 g, 10.85 mmol), sodium sulfide nonahydrate (2.96 g, 12.34 mmol), and sulfur (397 mg, 12.37 mmol) in EtOH (100 mL) was refluxed for 2.5 h. The cooled reaction mixture was evaporated under reduced pressure, diluted with 1 M HCl (100 mL), and extracted into EtOAc. The product was purified on a column of silica gel eluting in DCM, followed by recrystallization from EtOH to give white needles (1.57 g, 57%): mp 195–196 °C. 1H NMR (400 MHz, DMSO-d6) δ 8.82 (d, J = 2.3 Hz, 1H), 8.45 (d, J = 8.8 Hz, 1H), 8.30 (dd, J = 8.8, 2.3 Hz, 1H), 8.15 (ddt, J = 8.4, 3.1, 1.9 Hz, 2H), 7.74−7.57 (m, 3H). HPLC 100 area % (290 nm). Anal. Calcd for C13H8N2O2S: C, 60.93; H, 3.15; N, 10.93. Found: C, 90.95; H, 2.98; N, 11.03.
100g was prepared analogously to 100a from N-(2-chloro-5-nitrophenyl)-3,4-difluorobenzamide (99g, 3.78 g, 12.09 mmol), but the crude product was directly recrystallized from EtOH as a white solid (3.00 g, 85%): mp 175–196 °C. 1H NMR (400 MHz, DMSO-d6) δ 8.84−8.79 (m, 1H), 8.48 (dt, J = 8.8, 0.4 Hz, 1H), 8.31 (dd, J = 8.8, 2.2 Hz, 1H), 8.20 (ddd, J = 11.3, 7.6, 2.3 Hz, 1H), 8.02 (dddd, J = 8.7, 4.3, 2.3, 1.4 Hz, 1H), 7.69 (dt, J = 10.5, 8.4 Hz, 1H). HPLC 97.6 area % (290 nm). Anal. Calcd for C13H6F2N2O2S: C, 53.43; H, 2.07; N, 9.59. Found: C, 53.18; H, 2.09; N, 9.59.
A mechanically stirred mixture of 5-nitro-2-phenylbenzothiazole (100a, 1.87 g, 7.30 mmol), iron powder (2.21 g, 39.57 mmol), and ammonium chloride (818 mg, 15.30 mmol) in EtOH (100 mL) and water (50 mL) was stirred at reflux for 2 h until the reaction was complete by HPLC. The hot mixture was filtered through Celite, and the filtrate was evaporated to near dryness under reduced pressure. The residue was diluted with water and extracted into DCM. The product was recrystallized from EtOH as yellow crystals (1.47 g, 89%): mp 205–206 °C. 1H NMR (400 MHz, DMSO-d6) δ 8.08−7.97 (m, 2H), 7.71 (dd, J = 8.6, 0.5 Hz, 1H), 7.61−7.49 (m, 3H), 7.17 (dd, J = 2.2, 0.5 Hz, 1H), 6.79 (dd, J = 8.6, 2.2 Hz, 1H), 5.33 (s, 2H). HPLC 100 area % (254 nm). Anal. Calcd for C13H10N2S: C, 69.00; H, 4.45; N, 12.38. Found: C, 69.07; H, 4.59; N, 12.19.
101g was prepared analogously to 101a from 5-nitro-2-(3,4-difluorophenyl)-benzothiazole (100g, 2.96 g, 10.1 mmol) as a yellow solid (1.60 g, 60%): mp >174 °C dec. 1H NMR (400 MHz, DMSO-d6) δ 8.06 (ddd, J = 11.4, 7.7, 2.2 Hz, 1H), 7.87 (dddd, J = 8.6, 4.3, 2.2, 1.4 Hz, 1H), 7.73 (dd, J = 8.6, 0.4 Hz, 1H), 7.61 (dt, J = 10.5, 8.4 Hz, 1H), 7.18 (dd, J = 2.2, 0.5 Hz, 1H), 6.81 (dd, J = 8.6, 2.2 Hz, 1H), 5.37 (s, 2H). HPLC 100 area % (254 nm). Anal. Calcd for C13H8F2N2S: C, 59.53; H, 3.07; N, 10.68. Found: C, 59.29; H, 2.94; N, 10.62.
This work was supported by the National Instituted of Health (grant 5R01AI106850). We thank Patrick T. Weiser for his constructive comments and assistance in editing this manuscript.
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jmedchem.6b01163.
General experimental procedures, synthesis and characterization of target compounds and intermediates (PDF) Molecular formula strings (CSV)
Author ContributionsThe manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.
The authors declare no competing financial interest.