|Home | About | Journals | Submit | Contact Us | Français|
Spinal Muscular Atrophy (SMA) is an autosomal recessive disorder affecting the expression or function of survival motor neuron protein (SMN) due to the homozygous deletion or rare point mutations in the survival motor neuron gene 1 (SMN1). The human genome includes a second nearly identical gene called SMN2 that is retained in SMA. SMN2 transcripts undergo alternative splicing with reduced levels of SMN. Up-regulation of SMN2 expression, modification of its splicing, or inhibition of proteolysis of the truncated protein derived from SMN2 have been discussed as potential therapeutic strategies for SMA. In this manuscript, we detail the discovery of a series of arylpiperidines as novel modulators of SMN protein. Systematic hit-to-lead efforts significantly improved potency and efficacy of the series in the primary and orthogonal assays. Structure property relationships including microsomal stability, cell permeability and in vivo pharmacokinetics (PK) studies were also investigated. We anticipate that a lead candidate chosen from this series may serve as a useful probe for exploring the therapeutic benefits of SMN protein up-regulation in SMA animal models, and a starting point for clinical development.
Spinal Muscular Atrophy (SMA), an inherited autosomal neurodegenerative disease, is the leading genetic disorder affecting infant mortality.1 SMA is a relatively “common” rare disease, affecting approximately 1 in 6000 newborns. Approximately 1 in 40 individuals are genetic carriers.2 Clinically, there are four types of SMA (Type I, II, III, and IV). The determination of the type of SMA is based upon the physical milestones achieved. Usually, children with the most severe form of the disease (Type 1; Werdnig-Hoffmann disease) die before the age of two years in the absence of supportive respiratory care.3 In fact, SMA is the number one genetic cause of death in children under the age of two and many of those children who make it through infancy are confined to wheelchairs for their entire lives. There is currently no cure or effective therapy for SMA.
SMA is caused by deficiency in survival motor neuron (SMN) protein due to homozygous mutations or deletion of the survival motor neuron 1 telomeric gene (SMN1) that is located on chromosome 5q13.4 The SMN protein is expressed in all tissues, although there are broad variations regarding levels. Deficiencies in SMN lead to degeneration and dysfunction of alpha motor neurons in the anterior horn of the spinal cord. Studies of the correlation between SMA severity and the amount of SMN protein have shown an inverse relationship.5 Humans possess an additional paralogous gene, survival motor neuron 2 (SMN2), a centromeric copy gene that differs from the SMN1 gene at a critical nucleotide at position six in exon 7 (C-T).6 While full length 38 kDa SMN protein can be produced from the transcription of both SMN1 and SMN2 genes, the small variation in the C-T nucleotide has important consequences for alternative splicing of the SMN2 transcript. The pre-mRNA transcripts from SMN1 gene produce full-length mRNA with nine exons encoding full-length of SMN protein, while the C- to T- transition at position six in exon 7 of SMN2 induces mostly alternative pre-mRNA splicing lacking exon 7, which results in the truncated SMN protein (termed as SMNΔ7) as its major product. SMNΔ7 protein has a reduced ability to self-oligomerize, leading to protein instability and rapid degradation.7 SMN2 has been genetically validated as a target for SMA therapy, and there is a striking correlation between SMA type and SMN2 gene copy numbers.8 From the functional point of view, SMN associates with proteins Gemin 2, Gemin 3 and Gemin 4 forming a large complex that plays a role in snRNP assembly, pre-mRNA splicing and transcription.9
There are several SMA therapeutic strategies under investigation, including gene therapy,10 antisense oligonucleotides (ATO)11 and stem cells.12 However, increasing SMN protein levels by altering expression of the SMN2 mRNA by small molecules has been proposed as a promising potential therapeutic strategy for SMA. Over the past several years, several classes of small molecules have been identified that increase SMN transcript and/or protein levels in SMA patient-derived cell lines through a variety of mechanisms.13 Chang et al. reported that sodium butyrate effectively increased the amount of exon 7-containing SMN protein in SMA lymphoid cell lines by changing the alternative splicing pattern of exon 7 in the SMN2 gene. In vivo, sodium butyrate treatment of SMA-like mice resulted in increased expression of SMN protein in motor neurons of the spinal cord and resulted in significant improvement of SMA clinical symptoms.14 Sodium 4-phenylbutyrate (PBA), an analogue of sodium butyrate, was also found to increase SMN expression in vitro. PBA has a much longer half life time (1.5 hours) in human serum compared with sodium butyrate (6 minutes) which suggested that PBA, also due to its favorable pharmacological properties, could be a candidate for the treatment of SMA.15 The FDA approved antineoplastic drug hydroxyurea was recently reported as a candidate SMA therapeutic due to its strong safety profile, high pediatric bioavailability, and its known gene up-regulation.16 Other bioactive agents,17 such as valproic acid,18 aclarubicin (aclacinomycin A),19 tobramycin and amikacin20, Trichostatin A (TSA),21 suberoylanilide hydroxamic acid (SAHA),22 EIPA [5-(N-ethyl-N-isopropyl)-amiloride],23 LBH589,24 and (E)-Resveratrol25 have been shown to increase levels of exon 7–containing SMN transcript and/or overall SMN protein. All of these repurposed agents have significant liabilities that include gastrointestinal bleeding, short half-life in human serum, high toxicity and lack of target specificity; additionally, valproic acid, PBA, and hydroxyurea were not therapeutic in human trials. Lunn et al. identified indoprofen as having an effect on full length SMN2 expression in a cell-based reporter assay that up-regulated the SMN protein through a cyclooxygenase-independent mechanism, but it was not potent (AC50 >1 µM) and increased protein expression by only 13% (p-value <0.0139).26 Recently, a tetracycline compound, PTK-SMA1, was found to have the activity to promote SMN2 exon 7 splicing. However, this compound could not penetrate the blood-brain-barrier or increase SMN protein concentration in the central nervous system.27 Another previous high-throughput screen identified two novel chemical series as HDAC inhibitors that can modulate reporter activity in a cell-based SMN2 promoter assay.28 Another quinazoline series probe, identified as D156844, targets DcpS, which appears to modulate SMN protein levels by stabilizing the SMN2 mRNA transcript.29 Oral administration of D156844 significantly increased the mean lifespan of SMNΔ7 SMA mice by approximately 21–30% when given prior to motor neuron loss.30 After removal of the compound’s off-target activity, reducing its toxicity, and improving its potency and brain permeability, it produced the current clinical candidate, D157495.31 However, it has not been shown whether this series of molecules acts specifically on SMN2 mRNA transcript stability or what the safety profile of this molecule might be.32
We are particularly interested in identifying small molecules that can increase the level of SMN transcript harboring the full length of SMN protein for the therapeutic goal. This may be achieved by modulating splicing, increasing the level of transcript produced, or stabilizing the SMN protein. In order to achieve this goal of looking for molecules able to elevate SMN levels by any of those mechanisms, we developed a new high-throughput screening methodology that allows us to test a large number of compounds in a quantitative manner. In this manuscript, we present some of the results of our screening and optimization campaign.
Using a luciferase reporter gene assay that combines the promoter and splicing based cassettes in tandem with the major portion of the native SMN2 cDNA with luciferase, we were able to screen for modulators of all of these mechanisms in a high-throughput manner. Earlier screening assays for SMA only targeted compounds that either stimulate the SMN2 promoter or decrease exon 7 skipping. In this reporter assay, the SMN2 promoter drives expression of luciferase reporter that is fused to an SMN splicing reporter (Figure 1). By including our SMN splicing cassette,33 luciferase is only expressed when exon 7 is included in the mRNA transcript and the full-length SMN protein is generated. This new assay can identify compounds that increase SMN protein levels by three mechanisms: modulating alternative splicing of SMN2 exon 7, increasing transcription from the SMN2 promoter, or, due to the presence of the native SMN2 cDNA, and stabilizing the SMN protein. The assay measures the luminescent signal produced by luciferase, and compounds that increase luminescence are flagged as potential actives.
We performed quantitative high-throughput screen (qHTS)34 using this cell-based SMN2-luciferase reporter assay to identify compounds that increase expression of the full-length SMN protein. The NIH Molecular Libraries Small Molecule Repository (MLSMR)35 containing 210,386 compounds was screened in the primary assay during the course of this screen campaign (Figure 2). We tested 1,229 1,536-well plates in the primary screen. The signal to background averaged 9 fold and the average Z’ for the screen was 0.47 +/− 0.12, excluding 8 plates which failed visual quality control. 6,128 compounds were found to give significant concentration-responses in the primary screen. Compounds were deprioritized for confirmation if they were suspected to interfere with assay readout, or cellular viability using internal SAR data from other related assays of the MLPCN using the same chemical library.36 Recently, it has been brought to our attention that certain classes of luciferase inhibitors can stabilize the luciferase protein and therefore appear as compounds that are able to increase the signal.37 Compounds that act as luciferase inhibitors could score in the SMN2 assay as false positives due to this phenomenon. Indeed, more than 98% of hits in the primary screen were identified as luciferase inhibitors, which presumably stabilized the SMN2-luciferase fusion protein and prevented cellular degradation, although this has not been explicitly proven. Select samples active in the primary screen were obtained in DMSO stock solutions from the MLSMR and/or as powders from compound vendors to confirm activity in the original assay. A homologous cell line harboring the SMN1 gene fused to luciferase was used as a counter-screen. The SMN1 reporter recapitulates the splicing pattern typically seen with the endogenous SMN1 gene, and predominately includes exon 7. Follow-up compounds that exhibit increased signal in the SMN1 cell line were flagged as non-specific activators (i.e. transcription or protein stability). Purified luciferase enzyme was used in another counterscreen of the active compounds to filter out the potential false positives. A complete cheminformatics analysis of the follow-up screening data included chemotype clustering, singleton identification, and structural considerations, including physical properties and optimization potential. In addition, potency range, maximum response, and the curve class38 were also considered. Ultimately, this analysis led to selection of 21 candidates for further confirmation. The results of this qHTS screening campaign are detailed in Figure 2.
Ideally, compounds modulating SMN protein production in the SMN2-luciferase reporter line would also modulate SMN protein production in primary fibroblasts directly derived from SMA individuals. Therefore, 21 compounds representing the diverse set of chemotypes39 were tested in a SMA-derived patient fibroblast assay for up-regulation of SMN protein production by western blot, in a luciferase-free setting. Ultimately, this campaign led us to focus on two lead series represented by the 5-methoxy-3,6-diphenyl-1,2,4-triazine (1) and 1-(4-(4-bromophenyl)thiazol-2-yl)piperidine-4-carboxamide (2), which were able to increase the luciferase signal in the SMN2 reported assay and increase SMN protein production in a dose dependent manner in the western blot (Figure 3). In addition, although we do not know their specific mechanism of action, we noticed that two active compounds 1 and 2 appear to have distinct mechanisms of action. Compound 1 increased the expression of SMN2 and did not affect the expression of SMN1. On other hand, compound 2 increased the expression of SMN2 but reduced the expression of SMN1. This reduction of SMN1 expression would not be relevant in vivo, due to the fact that the SMN1 gene is not functional at all in SMA patients. The data in Figure 3 also shows the differences in luciferase inhibitory capacity of 1 and 2. It can be seen that compound 1 was able to both increase the production of SMN protein (western) and inhibit luciferase. So without the creation of another cell line with a different reporter format, such as β-lactamase, GFP or Renilla luciferase, it would be impossible to discriminate between SNM up-regulation and luciferase stabilization effects using our firefly luciferase reporter assay. In addition, the first triazine chemotype was found to have a very narrow SAR. Therefore, we chose compound 2 as the prioritized series for further SAR optimization and biological evaluation, since it possessed a more tractable SAR and drug-like properties. Furthermore, as reported in later section, this series displayed high cell permeability in Caco-2 cell permeability assay,40 and it had with no indication of being a glycoprotein (P-gp) substrate.41
Figure 4 shows the initial SAR strategy designed for the study of compound 2, including evaluation of the nature and position of the aromatic substituents, the geometry and substituents at the thiazole core, as well as the modification of the piperidine functional group.
We developed convergent synthetic routes toward the original hit compound 2 and its 4-arylthiazolyl piperidine analogues (Scheme 1). Route (a) was designed to prepare analogues with the modification of the phenyl ring. Starting with 2,4-dibromothiazole, selective amination occurred at the 2-position of the thiazole ring to give the key building block 1-(4-bromothiazol-2-yl)piperidine-4-carboxamide.42 Then, a mixture of an arylboronic acid or its corresponding pinacol ester and 1-(4-bromothiazol-2-yl)piperidine-4-carboxamide was irradiated under microwaves to induce a Suzuki type of cross-coupling reaction43 using tetrakis(triphenylphosphine)palladium(0) as a catalyst to furnish the final 4-arylthiazolyl piperidine analogues. Alternatively, the route (b), a Suzuki coupling at the 4-position of the 2-methylthio-4-bromothiazole, followed by MCPBA oxidation of the methylthioyl group to the methylsulfonyl moiety provided another building block 4-aryl-2-(methylsulfonyl)thiazole. However, the displacement of the methylsulfonyl group of the 4-aryl-2-(methylsulfonyl)thiazole with a secondary amine proved to be very difficult. Thus, all of the analogues described in this article were prepared through the amination-Suzuki coupling approach, route (a). This synthetic route was straightforward and provided for rapid SAR studies.
In a similar reaction fashion, if the starting material, 2,4-dibromothiazole is replaced by 2,5-dibromothiazole,44 2,5-dibromothiadiazole,45 5-bromo-2-chloropyrimidine,46 2,4-dichloropyrimidine,47 4,6-dichloropyrimidine,48 or 2,4-dichloro-1,3,5-triazine,49 the corresponding modifications at the thiazole core template can be accomplished as detailed in Scheme 2. It was also interesting to note that the replacement of thiazole with imidazole or N-methylimidazole was unsuccessful under the same reaction conditions. All aminated building blocks underwent the same type of Suzuki coupling procedure to give the desired final analogues.
Similar procedures were utilized to further explore modifications of the piperidine-4-carboxamide moiety (Scheme 3). In the process of typical amination-Suzuki steps, route (a), piperidine-4-carboxamide can be successfully switched to ethyl piperidine-4-carboxylate, piperazine, tert-butyl piperidin-4-ylcarbamate, piperidine-4-carbonitrile, 1-(piperidin-4-yl)ethanone as well as 4-(1H-imidazol-2-yl)piperidine, but not piperazine-1-carboxamide. Within our SAR studies, we were interested in examining several related substituted amide derivatives. Saponification of the ethyl ester afforded the appropriate acid, which can be readily coupled with amines to provide the final substituted amide modification analogues. We were also interested in replacing the piperidine with N-substituted piperazines. To access these derivatives, the free secondary amine of the piperazine ring can be easily converted into corresponding urea, amides, and sulfonylamides. Finally, a reverse amide bond analogue and a tetrazole analogue were successfully synthesized, as detailed in Scheme 3. Most of the final compounds were purified by preparative scale HPLC with reasonable yields.
Tables 1–7 disclose all of the analogues that were generated for this series to derive a SAR. The activity was reported showing two parameters: AC50 (concentration necessary to reach 50% of maximum luciferase signal) and rate of induction (ROI) (efficacy indicated as maximum percentage of luciferase signal with respect to basal levels), which were used to conduct SAR analysis. Both parameters, ROI and AC50 were evaluated and optimized, as the ideal compound should provide maximum SMN protein production with minimal concentration of a given compound.
The hit compound 2 identified from the primary screen was re-synthesized and found to possess an AC50 value of 1.22 µM with a ROI of 268%. Our early efforts were focused on improving both parameters in order to further explore the SAR for this chemotype. In order to conduct a comprehensive SAR (Table 1) for the aromatic group at the 4-position of the thiazole ring, while exploring the steric and electronic effects of the substituents, the thiazole and piperdine rings were held constant, as shown in Table 1. We first examined the effect of the substituent position at the phenyl ring on potency (AC50) and efficacy (ROI). With the exception of the ortho-fluoro substituted analogue 3r (AC50 = 2.44 µM, ROI = 274%), all other ortho-substituents were either inactive (3b, 3e, 3o, 4a, 4c, 4e, 4g, 4m, and 4p) or were significant less active (3u and 4x) versus compound 2. These data indicated the potential necessity of having a co-planar conformation between the two aromatic rings to maintain good activity. In contrast to ortho-substitutions, substituents at meta or para positions were generally well tolerated. Regarding the electronic nature of the substituents, an electron-donating functional group at the meta or para position tended to provide better activity and efficacy than the corresponding electron-withdrawing group (3c, 3f vs. 4h, 4n, 4q and 3d vs. 4i, 4o, 4r). In addition, halo, methyl or trifluoromethyl substitutions at para and meta position were also tolerated. Several substituents slightly increased or maintained the activity compared to the efficacy of the hit compound 2, such as the meta-methoxy substituted analogue 3c (AC50 = 1.94 µM, ROI = 357%), meta-N,N-dimethylamino substituted analogue 3f (AC50 = 0.97 µM, ROI = 504%), and para-methyl substituted analogue 3m (AC50 = 1.22 µM, ROI = 439%). Regarding hetero-aromatic ring substituents, none of analogues 4s–4u had a positive impact on either parameter compared to the parent compound 2. We also studied several polycyclic aromatic ring analogues (4e, 4f, and 4v–4x). However, most had either no activity or decreased activity and efficacy.
Given the improved potency and efficacy observed with the electron-donating substituents, we chose to screen a large number of phenoxy substitutents (Table 2). Several additional electron-donating substituents were tolerated including meta-isopropoxyphenyl (5c, AC50 = 0.77 µM, ROI = 605%). We also examined the 3,5- or 3,4-disubstituted analogues (5h and 5i, AC50 = 7.72 µM and 12.24 µM, ROI = 45% and 59%, respectively) and 3,4,5-trisubstituted derivative (5m, AC50 = 19.4 µM, ROI = 73%). Surprisingly, these modifications resulted in a huge loss of both activity and efficacy. Interestingly, this loss could be regained by bridging two substituents together. Furthermore, the resulting cyclic poly-phenolic-ethers demonstrated a clear trend of improvement in activity in line with ring size (5 < 6 < 7, 5j, 5k, 5l, AC50 = 2.44 µM, 0.77 µM, and 0.49 µM, respectively).
With several improved analogues in hand (analogues 5c, 5k, and 5l), we further studied the SAR of the piperidine ring moiety as shown in Table 3. In general, mono or disubstituted amide derivatives 6a–6i had worse potency and efficacy, with the exception of the phenylamide derivative (6e, AC50 = 2.44 µM, ROI = 262%) which was about the same as the hit compound 2. It is also important to remark that both the carboxylic acid 6k and the corresponding ethyl ester 6j derivatives were about 4- and 6-fold less active, respectively, compared to compound 2, while both showed improved efficacy (6k, AC50 = 4.87 µM, ROI = 554% and 6j, AC50 = 7.72 µM, ROI = 550%, respectively). This result indicated that only the carbonyl functional group might be involved in a potential hydrogen bond interaction. This was further supported by significant loss of potency of the non-carbonyl bearing analogues 6q–6s. When the amide moiety was replaced by cyano or tetrazole groups, the resulting derivatives (analogues 6l and 6m) also showed a loss of potency. Reverse amide 6o and carbamate 6p analogues were not tolerated either. Surprisingly, a free amine moiety seemed to be tolerated (6n, AC50 = 2.44 µM, ROI = 182%). In contrast to the modification of the amide moiety, the piperidine ring was found to be necessary to maintain potency. Interestingly, a piperazine analogue 6t, which represents another retro-amide variation, displayed reasonable good activity. In contrast, all of the other piperazine derivatives, 6u–6y, were essentially inactive.
In general, replacement of the 2,4-disubstituted thiazole ring with a 2,4-disubstituted pyrimidine maintained or improved potency and efficacy (Table 4). Placement of the piperidine-4-carboxamide substituent between the two hetero-aromatic nitrogen atoms and aromatic substituent at the 4-position, increased the potency of the molecule by 2–3 fold (7d, AC50 = 0.387 µM, ROI = 387% and 7i, AC50 = 0.772 µM, ROI = 446%). The aromatic substituent at the 5 position of the pyrimidine was also tolerated as represented by analogue 7b (AC50 = 0.613 µM, ROI = 236%), although its efficacy was only half in comparison with the above described pyrimidine analogues. The meta-isopropoxyphenyl substituted derivative 7g was found to be the best analogue in this series with about 9-fold increase in potency (7g, AC50 = 0.154 µM, ROI = 298%) versus the hit compound 2. Derivatives bearing a monofluoro substituent at 5 position of the pyrimidine gave an additional 2–4 fold improvement in potency (7j vs. 7i and 7l vs. 7n). Replacement of the pyrimidine for 1,3,5-triazine resulted in analogues with decreased potency (7i vs. 7k and 7h vs. 7p). These results indicated that the potency of these compounds would not be improved by the introduction of additional nitrogens in the pyrimidine core ring. An ortho-methyl substituted pyrimidine derivative 7o was found to be totally inactive which further supported an earlier observation for requirement of a co-planar conformation for this portion of the molecule for activity. On the other hand, when the pyrimidine substitution patterns were reversed, with the aromatic substituent between the two hetero-aromatic nitrogen atoms and the piperidine-4-carboxamide substituent at the position 4, the resulting derivatives decreased activity by 8–16 fold (7i vs. 7n and 7j vs. 7l). In addition, moving two nitrogen atoms around the pyrimidine also decreased the potency by 5-fold (7i vs. 7m and 7h vs. 7q). These findings lead to the hypothesis that the sterically accessible nitrogen atom adjacent to the piperidine group is involved in a hydrogen bond interaction, serving as a hydrogen bond acceptor. Table 5 shows modifications for the thiazole portion of the molecule. Switching the aromatic substituent from the 4 position to 5 position of the thiazole core ring increased the potency by 5-fold and efficacy by 2-fold (8g, AC50 = 0.224 µM, ROI = 599%). We speculated that this increment was due to accessibility of the thiazole nitrogen for involvement in hydrogen bond interactions with the putative molecular target. Replacement of a CH with a N atom in the thiazole core resulting in 1,3,4-thiadiazole led to a 3–10 fold loss of potency (8a vs. 8b; 8c vs. 8d; 8e vs. 8f; and 8g vs. 8h). These results are analogous to replacement of the pyrimidine with the triazine and are due to the electronic effect of the core ring.
With the rationale that the SAR of the aromatic substituents from the 2,4-substituted thiazole could be translated to a 2,5-substituted thiazole series, the SAR of the different electron-donating groups on the phenyl moiety was explored (Table 5 and Table 6). Gratifyingly, several electron-donating groups on the phenyl ring indeed produced a synergistic effects on both potency and efficacy (8a, AC50 = 0.038 µM, ROI = 370%; 8e, AC50 = 0.039 µM, ROI = 550%; 8l, AC50 = 0.077 µM, ROI = 709%; and 8m, AC50 = 0.031 µM, ROI = 576%). Again, the 3,4-disubstituted derivative 8n again was less active than the corresponding ring closed analogues 8a and 8o–8q. Several cyclic meta-N,N-disubstituted aminophenyl derivatives, 8r–8t, were examined and were found to be less active than the parent meta-N,N-dimethylaminophenyl derivative 8e.
The modification of the amide moiety in the 2,5-substituted thiazole system was further explored, as shown in Table 7. For these studies, 3,4-dihydro-2H-benzo[b][1,4]dioxepin-7-yl or meta-N,N-dimethylaminophenyl were chosen as the two aromatic substituents at the C5 position of thiazole core ring. The results demonstrated that the replacement of amide with an acid had a clear advantage for potency with a 3-fold improvement, but with a slight loss of efficacy (9a with best AC50 = 0.012 µM, ROI = 325%). Ethyl ester or phenylamide derivatives were also tolerated (9b, AC50 = 0.031 µM, ROI = 320% and 9c, AC50 = 0.049 µM, ROI = 353%). However, the imidazole and methylketone replacements were not tolerated (9d, 9e, 9f, and 9g).
With the SAR assessment established for this series, a number of selected analogues featuring desirable potency below 150 nM in the reporter assay were then examined to evaluate the effect on the human SMN protein expression using fibroblasts from SMA patients. We incubated a fibroblast cell line from a type I SMA patient (3813 cell line) with different doses of analogues and assessed the SMN protein level by quantitative western blotting (Figure 5). We observed that analogue 8m at 37 nM increased the SMN protein level by 2–3 fold. The SMN protein level was decreased with increasing drug concentration. This result matched with the bell-shaped dose-response curve observed in the SMN2-luciferase reporter assay. This could be due to decreased compound solubility at higher concentrations, or it may reflect the intrinsic mechanism of action. Analogue 9a showed a dose-dependent trend between 37 nM to 1 µM concentration. Other selected agents (9c, 8c and 8l) all showed an up-regulation of SMN protein at concentrations ranging between 37 nM and 333 nM. Furthermore, the increase in SMN protein production led us to explore whether the arylpiperidine analogues had effect on the overall number of SMN positive foci or gems in the nucleus. This would corroborate that the increased protein amounts represent functional SMN protein. SMN protein is predominately a cytoplasmic protein, but in the nucleus, SMN localizes to distinct punctuate bodies often referred to as gems (gemini of coiled [Cajal] bodies). There is a direct correlation between the number of gems and total SMN protein production in the cell.50 Gem counts are commonly used as another metric to score the amount of total SMN protein expressed on a cell-to-cell basis. The number of nuclei with gems and the number of gems per cell were both significantly reduced in type I SMA cells. Only 3.3% of nuclei have gems in fibroblast cells from SMA type I patients (cell line 3813), while 24.8% of nuclei have gems in fibroblast from a carrier parent (3814 cell line). We treated human type I SMA fibroblasts with increasing concentrations (37 nM – 3000 nM) of arylpiperidine analogues 8a, 8m, and 9a for 3 days, and the number of gems per 100 nuclei were examined (Figure 6). Analogue 8m treatment yielded more than a 2-fold of increase of the number of the gems at the low 37 nM dose. With increasing concentration, the number of gems was reduced, which agreed with the previous findings on our reporter and westerns experiments. Analogue 9a also showed an 80% increase of the gem numbers at 37 nM to 1000 nM concentrations, and a slightly decreased at 3000 nM. However, analogue 8a did not show any activity in this assay.
Compounds were initially characterized using RT-PCR to measure RNA expression levels and exon 7 inclusion within our reporter cell line (not endogenous SMN). Some analogues (2, 3f, and 8e) showed a slight increase in total SMN-luciferase transcripts, but there was almost no increase in exon 7 mRNA (data not shown). Therefore, the mechanism of action of our series is likely post-transcriptional, in contrast to the HDAC and DcpS inhibitors previously identified. Perhaps this mechanism includes increasing the stability of the SMN protein and/or affecting its degradation, but it is distinct from D156844, a deCode compound,29a which had no activity in the present screening assay. So far, no other SMN modulator has been described as having this type of activity, and therefore it will be interesting for the scientific community to have a better understanding of its properties, mechanism and potential pharmaceutical application. The SMNΔ7 protein, the primary product from SMN2, is functional in the context of low levels of exon-7 included SMN protein, but it is proteolytically unstable. It will be interesting to explore whether the present series has an impact on the stability of the SMNΔ7 derived truncated SMN protein, thereby establishing a new paradigm for the treatment of SMA.
In addition, several of the most potent analogues were profiled for mouse liver microsomal stability and cell permeability in Caco-2 cells (Table 8). For the microsomal stability tests, we decided to evaluate the amount of parent compound remaining following incubation at three different time periods: 10, 15 and 60 minutes. This allowed us to both calculate both intrinsic clearance and model half lives. The original hit compound 2 was found to have a relatively short half life (12 minutes). Analogues 8c, 8m, and 9c all had a very short half lives < 10 minutes. Analogues 8b, 8l, and 9a had half lives of more than 30 minutes. In the bidirectional Caco-2 cell permeability assay, analogues 7g, 8b, and 9a were found to exhibit high permeability, while analogues 8b and 9a had efflux ratios of 2.4 and 1.3 respectively, and analogue 8l had a promising efflux ratio of 0.05. The cell permeability of analogue 8m was not great, though the efflux ratio was acceptable (0.125).
In consideration of a balance of potency, efficacy, and ADME properties of all current analogues, we selected analogues 8l and 9a for full mouse in vivo pharmacokinetic (PK) studies to evaluate the central nervous system (CNS) penetration of this series (Table 9). The plasma and brain concentration of analogues 8l and 9a in male Swiss Albino mice after a single oral gavage administration at a dose of 30 mg/kg were measured. As shown in Table 9, the plasma to brain ratio of analogues 8l and 9a in male Swiss Albino mice was found to be 0.29 and 0.028, respectively. Both analogues 8l and 9a possessed a reasonable long half life in brain (t1/2 = 5.44 h and 13.69 h), as well as in plasma (t1/2 = 5.78 h and 11.18 h). The concentration of analogue 8l in brain reached at 1.28 µmol/kg (Cmax) in only 5 minutes, and a concentration above its AC50 of 0.077 µM was maintained for more than 4 hours. Despite its high plasma protein binging (94.9%) for analogue 9a, its concentration in brain reached 1.23 µmol/kg (Cmax) in 15 minutes and a concentration above AC50 concentration of 0.012 µM for over 24 hours. In addition, no behavioral disturbance was observed in animals administered with both tested compounds throughout the whole study period. Oral administration with good brain exposure will be preferred for demonstrating in vivo activity, as many of the current animal models require chronic administration of the drug.
In summary, after several rounds of medicinal chemistry SAR studies, we were able to improve the potency of the hit compound by ~ 100-fold from µM to low double digit nM in our cell-based SMN2-luciferase reporter assay, we were also able to increase the rate of SMN protein production 3–7 fold. From this study, analogues 8a (AC50 = 0.038 µM, ROI = 370%), 8e (AC50 = 0.039 µM, ROI = 550%), 8l (AC50 = 0.077 µM, ROI = 709%), 8m (AC50 = 0.031 µM, ROI = 576%), and 9a (AC50 = 0.012 µM, ROI = 325%) were identified as novel SMN protein modulators. In the field, it is generally believed that an up-regulation of SNM production of at least 20% above patient basal levels will be necessary to have a therapeutic impact. Our best molecules clearly surpass that mark, in vitro, increasing levels of SMN protein between 100 to 300% at low nanomolar concentrations when they were tested in our primary reporter assay, as well as in relevant orthogonal assays such as western blot and gems assays with patient fibroblasts. In consideration of all the aspects including ADME properties, analogues 8l and 9a possessed the best combination of potency, efficacy, mouse liver microsomal stability and cell permeability of all the analogues presented in this study. Moreover, analogues 8l and 9a showed good oral absorption and CNS penetration upon oral gavage administration at a reasonable dose of 30 mg/kg providing levels of exposure above their AC50 for more than 4 hours in brain, with no sign of toxicity or behavior disturbance in animals.
All cell culture reagents were purchased from Invitrogen. The white solid bottom 1,536 well, tissue culture treated plates came from Greiner. The OneGlo luciferase detection reagent was from Promega. The luciferase enzyme and buffers came from Sigma Aldrich. The 4f11 and 4B7 SMN antibodies were provided to EJA as a generous gift from Christian Lorson (University of Missouri). Goat anti-mouse HRP-conjugated secondary and α-tubulin antibodies (DM1α) were supplied by Sigma.
HEK-293 cells stably expressing the SMN1 or SMN2 reporter construct were maintained and passaged in complete DMEM media containing phenol red, glutamax, sodium pyruvate, 10 % FCS, 1x pen/strep, and 200 µg/mL hygromycin. Cells were assayed in the same media lacking only hygromycin. For the HTS, cells were plated at a density of 2,000 cells/well in 5 µL of media and were allowed to adhere overnight at 37 °C in 90% humidity with 5% CO2. After incubation, cells received 23 nL of compound library (in 100% DMSO) by pintool addition (Kalypsys). The final concentration of DMSO was 0.46%. The screen was conducted in a 7-point qHTS titration with a final concentration of 50 µM to 25 nM. The positive control compound used was sodium butyrate (4.5 mM final concentration) and the negative control compound was DMSO. Plates were incubated in the presence of compounds for 30 to 36 hours in the humidified incubator. After incubation, cells were treated with 3 µL of OneGlo luciferase detection reagent and incubated at room temperature for 5 to 15 minutes. Luminescence was detected on a Viewlux CCD based instrument (Perkin Elmer) with settings of 60 second integration and high speed 2x binning. Data were normalized against sodium butyrate (100%) and DMSO (0%) treatments.
The firefly luciferase enzyme was assayed at 5 nM in buffer consisting of 50 mM Tris-acetate pH 7.6, 10 mM Mg acetate 0.05% BSA, 0.01% Tween. Enzyme (2.5 µL/well) was dispensed in a white solid bottom 1,536 well plate and compounds (23 nL) were added using a pintool. Plates were incubated at room temperature for 5 minutes followed by the addition of 2.5 µL of OneGlo luciferase detection reagent. Luminescence was detected in the Viewlux with a 1 second integration time, no binning and medium camera sensitivity.
For detection of SMN protein in patient fibroblasts, 8,000 cells per cm2 were plated 24 hours prior to drug addition. Fresh media and compound were added every 24 hours. After 72 hours of incubation, cells were harvested, washed with cold PBS, and lysed as above. We have determined that 10 µg total protein per lane is within the linear range for immunoblot detection of SMN and α-tubulin. Western blots were probed for SMN with the 4f11 mouse monoclonal antibody and α-tubulin. Quantification of protein was performed with Fujifilm LAS-4000 Multifunctional Imaging System. The signal intensity was measured for each band on an immunoblot, normalized to the loading control, and the fold increase was determined in relation to the appropriate DMSO treated control.
All solvents were of anhydrous quality, purchased from Aldrich Chemical Co., and used as received. Commercially available starting materials and reagents were purchased from Aldrich, TCI and Acros, and were used as received. Analytical thin layer chromatography (TLC) was performed with Sigma Aldrich TLC plates (5 × 20 cm, 60 Å, 250 µm). Visualization was accomplished by irradiation under a 254 nm UV lamp. Chromatography on silica gel was performed using forced flow (liquid) of the indicated solvent system on Biotage KPSil pre-packed cartridges and using the Biotage SP-1 automated chromatography system. 1H NMR spectra were recorded on a Varian Inova 400 MHz spectrometer. Chemical shifts are reported in ppm with the solvent resonance as the internal standard (CDCl3 7.27 ppm, DMSO-d6 2.49 ppm, for 1H NMR). Data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, sep = septet, quin = quintet, br = broad, m = multiplet), coupling constants, and number of protons. Low resolution mass spectra (electrospray ionization) were acquired on an Agilent Technologies 6130 quadrupole spectrometer coupled to an Agilent Technologies 1200 series HPLC. The HPLC retention time were recorded through standard gradient 4% to 100% acetonitrile (0.05% TFA) over 7 minutes using Luna C18 (3 µm, 3 mm × 75 mm) column with a flow rate of 0.800 mL/min. High resolution mass spectral data were collected in-house using an Agilent 6210 time-of-flight mass spectrometer coupled to an Agilent Technologies 1200 series HPLC system. If needed, products were purified via a Waters preparative HPLC equipped with a Phenomenex Luna or Gemini C18 reverse phase (5 µm, 30 mm × 75 mm) column having a flow rate of 45 mL/min. The mobile phase was a mixture of acetonitrile (0.1% TFA) and H2O (0.1% TFA) for acidic condition and a mixture of acetonitrile and H2O (0.1% NH4OH) for basic condition. Samples were analyzed for purity on an Agilent 1200 series LC/MS equipped with a Luna C18 reverse phase (3 µm, 3 mm × 75 mm) column having a flow rate of 0.800 mL/min over a 7.0 minutes gradient and a run time of 8.5 minutes. Purity of final compounds was determined to be > 95%, using a 3 µL injection with quantitation by AUC at 220 and 254 nm (Agilent diode array detector).
A mixture of bromo- or chloro-substituted building block (0.100 mmol), boronic acid or pinacol ester (0.200 mmol) and tetrakis(triphenylphosphine)palladium (5.00 µmol) in DMF (1.50 mL) or CH3CN (1.50 mL) and 2.0 M Na2CO3 aqueous solution (0.50 mL) was heated in a microwave at 100 °C for 30 minutes. The reaction mixture was cooled to room temperature, treated with a small portion of Si-THIOL to get rid of palladium. The mixture was filtered through a frit to give a light yellow solution. The crude material was purified by preparative HPLC under acidic or basic condition to give the final product.
A mixture of carboxylic acid (0.082 mmol), EDC (0.082 mmol), HOBt (0.082 mmol), DMAP (0.082 mmol) and amine (0.163 mmol) in DMF (1.50 mL) was heated in a microwave at 100 °C for 1.5 hours. The crude material was purified by preparative HPLC to give the final product.
A solution of 4-(4-bromophenyl)-2-(piperazin-1-yl)thiazole (0.093 mmol), Et3N (0.139 mmol) in CH2Cl2 (2.00 mL) was treated at 0 °C with carboxylic chloride or sulfonyl chloride (0.111 mmol). The reaction mixture was allowed to warm to room temperature, stirred for 1 hour, and concentrated in vacuo. The crude material was purified by preparative HPLC to give the final product.
A general protocol for amination: A mixture of 2,4-dibromothiazole (2.06 g, 8.48 mmol), piperidine-4-carboxamide (1.30 g, 10.1 mmol) and Et3N (2.50 mL) in ethanol (5.00 mL) was heated in a microwave at 100 °C for 1 hour. The reaction mixture was cooled to room temperature, diluted with water, and extracted with a mixture of MeOH and CH2Cl2. The organic layer was separated, dried with Na2SO4, and concentrated as a light brown solid. The crude mixture was purified by Biotage with 0–10% of MeOH in CH2Cl2 with 1% Et3N to give 2.08 g (85%) of 1-(4-bromothiazol-2-yl)piperidine-4-carboxamide as a white solid: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.31 (br. s., 1 H), 6.85 (s, 1 H), 6.82 (br. s., 1 H), 3.77 – 3.90 (m, 2 H), 3.03 (td, J=12.5, 2.8 Hz, 2 H), 2.29 – 2.39 (m, 1 H), 1.78 (dd, J=13.5, 3.1 Hz, 2 H), 1.50 – 1.63 (m, 2 H); LCMS RT = 4.13 min, m/z 289.9 [M+H+]. The title compound was prepared according to the general protocol A: 1H NMR (400 MHz, DMSO-d6) δ 7.83-7.79 (m, 2 H), 7.59-7.55 (m, 2 H), 7.33 (s, 1 H), 7.32 (br. s., 1 H), 6.82 (br. s., 1 H), 3.96 (br. d., J=12.8 Hz, 2 H), 3.06 (td, J=12.6, 2.9 Hz, 2 H), 2.36 (tt, J=11.6, 3.8 Hz, 1 H), 1.82 (dd, J=13.0, 2.6 Hz, 2 H), 1.61 (qd, J=12.4, 4.0 Hz, 2 H); LCMS RT = 5.16 min, m/z 366.0 [M+H+]; HRMS (ESI) m/z calcd for C15H1779BrN3OS [M+H+] 366.0276, found 366.0272.
The compound was prepared according to the general protocol B as a TFA salt: 1H NMR (400 MHz, CDCl3) δ ppm 7.60 – 7.71 (m, 2 H), 7.47 – 7.57 (m, 2 H), 6.72 (s, 1 H), 5.78 (br. s., 1 H), 4.15 (dt, J=13.3, 3.5 Hz, 2 H), 3.10 – 3.30 (m, 2 H), 2.84 (d, J=5.1 Hz, 3 H), 2.41 (tt, J=11.1, 3.8 Hz, 1 H), 1.95 – 2.05 (m, 2 H), 1.81 – 1.95 (m, 2 H); 19F NMR (376 MHz, CDCl3) δ ppm −75.97 (s); LCMS RT = 5.34 min, m/z 380.0 [M+H+]; HRMS (ESI) m/z calcd for C16H1979BrN3OS [M+H+] 380.0432, found 380.0425.
A solution of ethyl 1-(4-(4-bromophenyl)thiazol-2-yl)piperidine-4-carboxylate (3.34 g, 8.45 mmol) in THF (24.0 mL) and H2O (8.00 mL) was treated at room temperature with LiOH (1.01 g, 42.2 mmol). The reaction mixture was stirred at room temperature for 24 hours, diluted with 100 mL of CH2Cl2 and washed with 2.0 N HCl (25.0 mL). The organic layer was separated, dried with Na2SO4, and concentrated to give a yellow oil. The crude product was purified by Biotage with 0–15% of MeOH in CH2Cl2 to give 2.79 g (90%) of product as a white solid. A small amount of sample was re-purified by preparative HPLC under acidic condition to give the product as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.74 – 7.88 (m, 2 H), 7.50 – 7.64 (m, 2 H), 7.33 (s, 1 H), 3.90 (dt, J=12.7, 3.5 Hz, 2 H), 3.04 – 3.23 (m, 2 H), 2.51 – 2.59 (m, 1 H), 1.94 (dd, J=13.1, 3.7 Hz, 2 H), 1.51 – 1.75 (m, 2 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.87 (s); LCMS RT = 5.84 min, m/z 367.0 [M+H+]; HRMS (ESI) m/z calcd for C15H1679BrN2O2S [M+H+] 367.0116, found 367.0103.
A solution of 1-(4-(4-bromophenyl)thiazol-2-yl)piperidine-4-carbonitrile (59.5 mg, 0.171 mmol) in water (1.00 mL) and 1,4-dioxane (0.58 mL) was treated at room temperature with sodium azide (33.3 mg, 0.513 mmol) and zinc bromide (57.7 mg, 0.256 mmol). The pH value of the solution was adjusted to about 7 with 1 N NaOH (~ 6 drops). The reaction mixture was heated in oil bath at 120 °C for 60 hours. Another aliquot of reagents was added to the reaction solution and the mixture was re-heated at 120 °C for additional 60 hours. The reaction mixture was cooled to room temperature, diluted with EtOAc, and washed with H2O. The organic layer was separated, dried with Na2SO4, and concentrated in vacuo. The crude product was purified by preparative HPLC under basic condition to give 15.0 mg (22%) of product: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.76 – 7.87 (m, 2 H), 7.53 – 7.62 (m, 2 H), 7.33 (s, 1 H), 3.97 (dt, J=13.2, 3.7 Hz, 2 H), 3.20 – 3.30 (m, 2 H), 3.14 (tt, J=11.1, 3.6 Hz, 1 H), 2.04 (dd, J=13.3, 3.1 Hz, 2 H), 1.71 – 1.87 (m, 2 H); LCMS RT = 5.62 min, m/z 391.0 [M+H+]; HRMS (ESI) m/z calcd for C15H1679BrN6S [M+H+] 391.0341, found 391.0338.
A solution of tert-butyl 1-(4-(4-bromophenyl)thiazol-2-yl)piperidin-4-ylcarbamate (78.7 mg, 0.180 mmol) in CH2Cl2 (2.00 mL) was treated at 0 °C with TFA (1.00 mL). The reaction mixture was stirred at 0 °C for 30 minutes and at room temperature for additional 10 minutes. The solvents were removed in vacuo. The crude product was filtered through a short cartridge column to remove TFA to give 43.4 mg (72%) of product as a white solid. Small portion of sample was re-purified by preparative HPLC under acidic condition to give the final product as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.93 (br. s., 2 H), 7.72 – 7.85 (m, 2 H), 7.50 – 7.65 (m, 2 H), 7.37 (s, 1 H), 4.00 (d, 2 H), 3.25 – 3.35 (m, 1 H), 3.14 (td, J=12.8, 2.6 Hz, 2 H), 1.99 (dd, J=12.6, 3.0 Hz, 2 H), 1.57 (qd, J=12.4, 4.4 Hz, 2 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −73.54 (s); LCMS RT = 4.66 min, m/z 338.0 [M+H+]; HRMS (ESI) m/z calcd for C14H1779BrN3S [M+H+] 338.0327, found 338.0323.
The compound was prepared according to the general protocol C as a TFA salt: 1H NMR (400 MHz, CDCl3) δ ppm 7.64 – 7.74 (m, 2 H), 7.47 – 7.55 (m, 2 H), 6.82 (s, 1 H), 3.77 – 3.92 (m, 2 H), 3.66 (br. s., 4 H), 3.48 – 3.58 (m, 2 H), 2.20 (s, 3 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −76.04 (s); LCMS RT = 5.89 min, m/z 366.0 [M+H+]; HRMS (ESI) m/z calcd for C15H1779BrN3OS [M+H+] 366.0276, found 366.0274.
A solution of 4-(4-bromophenyl)-2-(piperazin-1-yl)thiazole (40.0 mg, 0.123 mmol) in H2O (2.00 mL) was treated at room temperature with KOCN (20.0 mg, 0.247 mmol). The reaction mixture was stirred at room temperature overnight and extracted with EtOAc. The organic layer was separated, dried with Na2SO4, and concentrated in vacuo. The crude mixture was purified by preparative HPLC under basic condition to give 14.2 mg (31%) of product: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.82 (d, J=8.4 Hz, 2 H), 7.57 (d, J=8.4 Hz, 2 H), 7.37 (s, 1 H), 6.11 (br. s., 2 H), 3.37 – 3.57 (m, 8 H); LCMS RT = 5.33 min, m/z 367.0 [M+H+]; HRMS (ESI) m/z calcd for C14H1679BrN4OS [M+H+] 367.0228, found 367.0215.
The compound was prepared according to the general protocol A as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 8.41 (d, J=5.1 Hz, 1 H), 7.66 (ddd, J=8.1, 1.3, 1.0 Hz, 1 H), 7.57 – 7.63 (m, 1 H), 7.41 (t, J=7.9 Hz, 1 H), 7.29 (br. s., 1 H), 7.19 (d, J=5.3 Hz, 1 H), 7.03 – 7.13 (m, 1 H), 6.78 (br. s., 1 H), 4.55 – 4.84 (m, 3 H), 2.88 – 3.10 (m, 2 H), 2.41 (tt, J=11.5, 3.8 Hz, 1 H), 1.80 (dd, J=12.7, 2.9 Hz, 2 H), 1.50 (qd, J=12.4, 3.9 Hz, 2 H), 1.30 (d, J=6.1 Hz, 6 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.96 (s); LCMS RT = 4.45 min, m/z 341.2 [M+H+]; HRMS (ESI) m/z calcd for C19H25N4O2 [M+H+] 341.1978, found 341.1983.
The compound was prepared according to the general protocol A: 1H NMR (400 MHz, DMSO-d6) δ ppm 8.36 (d, J=5.1 Hz, 1 H), 7.62 – 7.80 (m, 2 H), 7.29 (br. s., 1 H), 7.01 – 7.14 (m, 2 H), 6.78 (br. s., 1 H), 4.63 – 4.84 (m, 2 H), 4.10 – 4.26 (m, 4 H), 2.93 (td, J=12.7, 2.5 Hz, 2 H), 2.40 (tt, J=11.5, 3.7 Hz, 1 H), 2.14 (quin, J=5.6 Hz, 2 H), 1.79 (dd, J=12.9, 2.7 Hz, 2 H), 1.49 (qd, J=12.4, 4.3 Hz, 2 H); LCMS RT = 3.84 min, m/z 355.2 [M+H+]; HRMS (ESI) m/z calcd for C19H23N4O3 [M+H+] 355.1770, found 355.1770.
The compound was prepared according to the general protocol A: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.47 (s, 1 H), 7.31 (br. s., 1 H), 7.07 (d, J=2.3 Hz, 1 H), 6.97 – 7.04 (m, 1 H), 6.89 – 6.96 (m, 1 H), 6.82 (br. s., 1 H), 4.12 (dt, J=9.3, 5.5 Hz, 4 H), 3.89 (dt, J=12.8, 3.3 Hz, 2 H), 3.04 (td, J=12.5, 2.9 Hz, 2 H), 2.36 (tt, J=11.5, 3.7 Hz, 1 H), 2.02 – 2.15 (m, 2 H), 1.79 (dd, J=13.2, 2.8 Hz, 2 H), 1.50 – 1.68 (m, 2 H); LCMS RT = 3.74 min, m/z 360.1 [M+H+]; HRMS (ESI) m/z calcd for C18H22N3O3S [M+H+] 360.1382, found 360.1379.
The compound was prepared according to the general protocol A as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.23 – 7.52 (m, 3 H), 6.95 – 7.11 (m, 1 H), 6.84 (br. s., 1 H), 4.19 (t, J=5.6 Hz, 4 H), 3.88 (dt, J=12.8, 3.5 Hz, 2 H), 3.19 (td, J=12.4, 2.9 Hz, 2 H), 2.35 (tt, J=11.4, 4.0 Hz, 1 H), 2.14 (dt, J=11.2, 5.6 Hz, 2 H), 1.82 (dd, J=13.5, 2.9 Hz, 2 H), 1.65 (qd, J=12.4, 3.9 Hz, 2 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.79 (s); LCMS RT = 4.21 min, m/z 361.1 [M+H+]; HRMS (ESI) m/z calcd for C17H21N4O3S [M+H+] 361.1334, found 361.1332.
The compound was prepared according to the general protocol A as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.58 (s, 1 H), 7.34 – 7.45 (m, 2 H), 7.32 (br. s., 1 H), 7.18 (d, J=8.0 Hz, 2 H), 6.84 (br. s., 1 H), 3.91 (dt, J=12.9, 3.2 Hz, 2 H), 3.12 (td, J=12.6, 2.8 Hz, 2 H), 2.38 (tt, J=11.4, 3.8 Hz, 1 H), 2.29 (s, 3 H), 1.82 (dd, J=13.2, 2.8 Hz, 2 H), 1.48 – 1.72 (m, 3 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.86 (s); LCMS RT = 3.91 min, m/z 302.1 [M+H+]; HRMS (ESI) m/z calcd for C16H20NOS [M+H+] 302.1327, found 302.1328.
The compound was prepared according to the general protocol A as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.65 (s, 1 H), 7.32 (br. s., 1 H), 7.24 (t, J=8.0 Hz, 1 H), 6.93 – 7.03 (m, 2 H), 6.83 (br. s., 1 H), 6.76 – 6.81 (m, 1 H), 4.55 – 4.75 (m, 1 H), 3.91 (dt, J=12.8, 3.4 Hz, 2 H), 3.11 (td, J=12.5, 2.9 Hz, 2 H), 2.37 (tt, J=11.4, 3.7 Hz, 1 H), 1.81 (dd, J=13.3, 2.9 Hz, 2 H), 1.52 – 1.68 (m, 2 H), 1.26 (d, J=6.1 Hz, 6 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.91 (s); LCMS RT = 4.30 min, m/z 346.1 [M+H+]; HRMS (ESI) m/z calcd for C18H24N3O2S [M+H+] 346.1589, found 346.1593.
The compound was prepared according to the general protocol A as a TFA salt: 1H NMR (400 MHz, DMSO-d6) δ ppm 7.58 (s, 1 H), 7.37 – 7.45 (m, 2 H), 7.32 (br. s., 1 H), 7.21 – 7.29 (m, 2 H), 6.83 (br. s., 1 H), 3.91 (dt, J=12.9, 3.0 Hz, 2 H), 3.09 (td, J=12.5, 2.8 Hz, 2 H), 2.48 (s, 3 H), 2.37 (tt, J=11.6, 4.0 Hz, 1 H), 1.81 (dd, J=13.5, 3.5 Hz, 2 H), 1.54 – 1.67 (m, 2 H); 19F NMR (376 MHz, DMSO-d6) δ ppm −74.56 (s); LCMS RT = 4.07 min, m/z 361.1 [M+H+]; HRMS (ESI) m/z calcd for C16H20N3OS2 [M+H+] 334.1048, found 334.1048.
A solution of ethyl 1-(5-(3,4-dihydro-2H-benzo[b][1,4]dioxepin-7-yl)thiazol-2-yl)piperidine-4-carboxylate (284 mg, 0.72 mmol) in THF (6.00 mL) and H2O (2.00 mL) was treated at room temperature with LiOH (86.0 mg, 3.60 mmol). The reaction mixture was stirred at room temperature for 24 hours, diluted with 100 mL CH2Cl2, and washed with 2 N HCl (25.0 mL). The organic layer was separated, dried with Na2SO4, and concentrated as a yellow oil. The crude product was purified by Biotage with 0–15% of MeOH in CH2Cl2 to give 233 mg (90%) of product as a white solid: 1H NMR (400 MHz, DMSO-d6) δ ppm 12.32 (br. s., 1 H), 7.48 (s, 1 H), 7.07 (d, J=2.2 Hz, 1 H), 6.97 – 7.03 (m, 1 H), 6.89 – 6.96 (m, 1 H), 4.12 (dt, J=9.2, 5.5 Hz, 4 H), 3.83 (ddd, J=13.2, 3.7, 3.5 Hz, 2 H), 2.98 – 3.20 (m, 2 H), 2.51 – 2.57 (m, 1 H), 2.04 – 2.17 (m, 2 H), 1.87 – 1.98 (m, 2 H), 1.49 – 1.69 (m, 2 H); LCMS RT = 4.18 min, m/z 361.1 [M+H+]; HRMS (ESI) m/z calcd for C18H21N2O4S [M+H+] 361.1222, found 361.1221.
The compound was prepared according to the general protocol A: 1H NMR (400 MHz, DMSO-d6) δ ppm 9.95 (s, 1 H), 7.60 (dd, J=8.7, 1.1 Hz, 2 H), 7.49 (s, 1 H), 7.24 – 7.33 (m, 2 H), 6.97 – 7.11 (m, 3 H), 6.92 – 6.97 (m, 1 H), 4.13 (dt, J=9.4, 5.5 Hz, 4 H), 3.90 – 4.03 (m, 2 H), 3.10 (td, J=12.6, 2.7 Hz, 2 H), 2.62 (tt, J=11.5, 3.7 Hz, 1 H), 2.10 (quin, J=5.4 Hz, 2 H), 1.90 (dd, J=13.0, 2.8 Hz, 2 H), 1.71 (qd, J=12.4, 4.3 Hz, 2 H); LCMS RT = 4.94 min, m/z 436.2 [M+H+]; HRMS (ESI) m/z calcd for C24H26N3O3S [M+H+] 436.1695, found 436.1699.
The authors thank William Leister, Jim Bougie, Chris LeClair, Paul Shinn, Danielle van Leer, Thomas Daniel, and Jeremy Smith for assistance with compound purification and management. We thank Dr. Marc Ferrer for helpful suggestions and Ms. Allison Mandich for critical review of the manuscript. This research was supported by the Molecular Libraries Program of the National Institutes of Health Roadmap for Medical Research (U54MH084681), the Intramural Research Program of the National Human Genome Research Institute, National Institutes of Health, and R03MH084179-01 to EJA.
Assay protocols, experimental procedures and spectroscopic data (1H-NMR, LCMS and HRMS) are listed for all compounds. This material is available free of charge via the Internet at http://pubs.acs.org.