A New Series of Orally Bioavailable Chemokine ... - ACS Publications

Mar 17, 2016 - Norgine Ltd, Norgine House, Widewater Place, Moorhall Road, Harefield, Uxbridge, UB9 6NS, United Kingdom. ‡. Jubilant Biosys Limited,...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/jmc

A New Series of Orally Bioavailable Chemokine Receptor 9 (CCR9) Antagonists; Possible Agents for the Treatment of Inflammatory Bowel Disease S. Barret Kalindjian,† Sanjay V. Kadnur,‡ Christopher A. Hewson,† Chandregowda Venkateshappa,‡ Suresh Juluri,‡ Rajendra Kristam,‡ Bheemashankar Kulkarni,‡ Zainuddin Mohammed,‡ Rohit Saxena,§ Vellarkad N. Viswanadhan,‡ Jayashree Aiyar,‡ and Donna McVey*,† †

Norgine Ltd, Norgine House, Widewater Place, Moorhall Road, Harefield, Uxbridge, UB9 6NS, United Kingdom Jubilant Biosys Limited, #96, Industrial Suburb, Second Stage, Yeshwanthpur, Bangalore 560022, India § Jubilant Chemsys Limited, B-34 Sector-58, Noida 201301, India ‡

S Supporting Information *

ABSTRACT: Chemokine receptor 9 (CCR9), a cell surface chemokine receptor which belongs to the G protein-coupled receptor, 7-trans-membrane superfamily, is expressed on lymphocytes in the circulation and is the key chemokine receptor that enables these cells to target the intestine. It has been proposed that CCR9 antagonism represents a means to prevent the aberrant immune response of inflammatory bowel disease in a localized and disease specific manner and one which is accessible to small molecule approaches. One possible reason why clinical studies with vercirnon, a prototype CCR9 antagonist, were not successful may be due to a relatively poor pharmacokinetic (PK) profile for the molecule. We wish to describe work aimed at producing new, orally active CCR9 antagonists based on the 1,3dioxoisoindoline skeleton. This study led to a number of compounds that were potent in the nanomolar range and which, on optimization, resulted in several possible preclinical development candidates with excellent PK properties.



INTRODUCTION

(GPCR), 7-trans-membrane (7-TM) superfamily, is expressed on lymphocytes in the circulation and is the key chemokine receptor that enables these cells to target the intestine.3 CCR9 has only one identified ligand, thymus-expressed chemokine (TECK; also known as chemokine ligand 25 (CCL25)).4 The primary site of CCL25 expression is in epithelial cells lining the intestinal lumen,5 and CCR9-positive cells are enriched in both the small intestine and the colon, relative to the circulation.6 Disruption of the CCR9/CCL25 interaction by antibody and small molecule antagonists of CCR9 has been demonstrated to be effective in preventing the inflammation observed in small animal models of IBD.7,8 Consequently, it has been proposed that CCR9 antagonism represents a means to prevent the aberrant immune response of IBD in a more localized and disease-specific manner and one which is accessible to small molecule approaches. This is in contrast to the more systemic and generalized immunosuppression underpinning many of the currently used biological therapies, e.g., anti-tumor necrosis factor−α therapy. As judged by the patent literature,9 a number of groups have been active in the exploration of structure−activity relationships

Inflammatory bowel disease (IBD) is a chronic inflammatory disorder of the gastrointestinal tract in which tissue damage and inflammation lead to long-term, often irreversible impairment of the structure and function of the gastrointestinal tract.1 While there are a number of therapeutic options for treating and maintaining remission in some patients with IBD, none of the currently available medicines provides a cure, and, overall, there remains a significant unmet medical need in this area. Chemokines are a family of structurally related small proteins released from a variety of different cells within the body.2 They have a primary ability to induce chemotaxis and thereby attract multiple cells of the immune system to sites of inflammation or as a part of normal immune function homeostasis. Examples of the types of cells attracted by chemokines include monocytes, T and B lymphocytes, dendritic cells, natural killer cells, eosinophils, basophils, and neutrophils. In addition, chemokines are also able to cause activation of leukocytes at the site of inflammation by a variety of means such as causing degranulation of granulocytes, generation of superoxide anions (oxidative burst), and up-regulation of integrins to cause extravasation. Chemokine receptor 9 (CCR9), a cell surface chemokine receptor which belongs to the G protein-coupled receptor © XXXX American Chemical Society

Received: November 26, 2015

A

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(SAR) for CCR9, and these efforts have been the subject of reviews.10,11 Most prominent among these is the compound discovered by Chemocentryx and licensed by GlaxoSmithKline, N-{4-chloro-2-[(1-oxidopyridin-4-yl)carbonyl]phenyl}-4-(1,1dimethylethyl) benzenesulfonamide (1) (Figure 1), also known

Scheme 1. Synthesis of 3-Chloro-6-nitrophthalic Anhydride (4)a

a

Reagents and conditions: (a) HNO3/H2SO4, rt, 12 h, 74%. (b) Ac2O, 120 °C, 18 h, 99%.

reduction of the nitro group was carried out using hydrogen and 10% palladium on charcoal as a catalyst. The amine iv was reacted with the appropriate sulfonyl chloride in pyridine for 24 h at room temperature, to leave the requisite test compounds. Where R3 was chloro, the nitro reduction reaction was carried out either using Raney nickel and hydrogen or, more usually, iron powder in acetic acid. Furthermore, the sulfonylation reaction required somewhat more forceful conditions. This step was carried out in one of two ways: a one-step process, using pyridine as solvent and heating at 125 °C for 24 h in the presence of a catalytic quantity of DMAP, or a two-step process. This was achieved by heating a pyridine solution to 90 °C for 8 h. This resulted in a mixture of the required compound and a bis-sulfonylated species. The second sulfonyl group was removed by stirring the mixture with 1 M tetra-N-butylammonium fluoride in THF for 3 h at room temperature. Oxidation, as needed to produce pyridine-N-oxides, was achieved by stirring a dichloromethane solution of the substrate with meta-chloroperoxybenzoic acid at room temperature for 8 h. Conversion of an R3 chloro group to the nitrile, as in compound 16, was carried out by heating the substrate with a mixture of zinc cyanide, zinc dust, 1,1′bis(diphenylphosphino)ferrocene, and tris(dibenzylideneacetone)dipalladium(0) in dimethylacetamide to 120 °C in a microwave for 2 h.

Figure 1. Structure of vercirnon, compound 1.

variously as vercirnon, CCX-282, or GSK1605786.8 This compound underwent successful Phase II clinical trials for Crohn’s disease12,13 but ultimately demonstrated a lack of efficacy in the Phase III SHIELD trials.14 It was our conjecture that the reason for failure at Phase III was likely a combination of the patient population selected for study (who were more seriously ill than the Phase II population) and the relatively poor pharmacokinetic (PK) profile, which led to the use of a high, but apparently insufficient, dose of drug. We wish to report on some of the SAR studies underpinning a successful medicinal chemistry program which was aimed at producing an orally available CCR9 antagonist with a superior PK profile to compound 1. Others, too, have recently reported on their efforts to improve on the properties of vercirnon.15 The program started by identifying the common features required by molecules to act as CCR9 antagonists by considering a number of reference templates from the available literature.9 These groups were then introduced onto small, commonly utilized, mainly bicyclic or tricyclic scaffolds, attempting to mimic the steric and electronic chemical space that was a feature of the known ligands. This is effectively a privileged structure approach, a concept that has found substantial utility in discovering nonpeptide ligands for peptide hormone receptors.16 Some 20 distinct series of compounds were devised and examples made, of which three were deemed to be of sufficient interest to investigate in more detail, and one of which, based on the 1,3-dioxoisoindoline skeleton, is presented in this paper. Our approach has given rise to a number of lead optimized molecules and a set of compounds from which a preclinical development candidate can be chosen.



RESULTS AND DISCUSSION Using the approach described in the Introduction section, we have found that the 1,3-dioxoisoindoline skeleton represents an excellent template from which to prepare CCR9 antagonists. The initial assay strategy involved examining the ability of molecules to inhibit the ligand-dependent increase in intracellular Ca2+ initiated through CCR9 interaction with its ligand, CCL25, using the Fluorescent Imaging Plate Reader (FLIPR) platform. This was carried out using MOLT-4 cells (ECACC Catalogue no.: 85011413), which is a cell line of human origin that endogenously expresses both CCR9A and CCR9B. Promising compounds were then examined in a cell line that overexpressed recombinant human CCR9A (Millipore) (CCR9 OE) and in a functional chemotaxis assay, also using MOLT-4 cells. In parallel, a series of tests were carried out in vitro to determine some of the ADME properties of molecules, in order to establish whether they had appropriate drug-like characteristics. PK in mice was used to establish oral bioavailability before the molecules were examined in recognized models of IBD in rodents. CCR9A and CCR9B receptors were first described in 2000.17 However, there have been no reports of any functional differences between them, and so we were not concerned strategically with selectivity in this program. In the event, a cell line expressing the CCR9B receptor was produced, and representative compounds tested in a calcium flux assay. No



CHEMISTRY Scheme 1 shows the preparation of 3-chloro-6-nitrophthalic anhydride (4) from 3-chlorophthalic anhydride (2). Compound 2 was nitrated in 74% yield by treatment for 12 h at room temperature with a mixture of nitric and sulfuric acids. The resulting phthalic acid (3) was then treated with acetic anhydride at a temperature of 120 °C for 18 h to produce the desired anhydride (4) in almost quantitative yield. This material was used in the preparation of compounds 11−16 and 19−71. The preparation of the compounds for testing followed the generic procedures illustrated in Scheme 2. The appropriate anhydride i was condensed with the amine ii to yield the imide iii. This reaction was performed by heating the reagents at a reflux in acetic acid for 18 h. When R3 was hydrogen or methyl, B

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Scheme 2. General Reaction Scheme for the Synthesis of Compounds 5−71a

a Reagents and conditions: (a) HOAc, reflux, 18 h. (b) 10% Pd/C/H2, MeOH. (c) Fe, HOAc, 6 h or H2/Raney Ni, ethanol, 6 h. (d) Pyridine, rt, 24 h. (e) cat DMAP, pyridine 125 °C, 24 h or (1) Pyridine, 90 °C, 8 h; (2) 1 M TBAF in THF, rt, 3 h. (f) mCPBA, CH2Cl2, rt, 8 h. (g) Zn(CN)2, dppf, Pd2(dba)3, Zn dust, N2, DMA, 120 °C, 2 h, microwave.

significant differences in the behavior of the ligands was seen (data not shown). In the first phase of the work, careful consideration of the reference templates showed that all the molecules contained an aryl-sulfonamide and that this was generally attached through a spacer to another aromatic ring. Optimally, this second aromatic ring was a pyridine moiety. With this putative pharmacophore established, a series of molecular alignments were carried out using molecular modeling tools to determine the best positions for substitution of the various skeletons under consideration. In this way, several 1,3-dioxoisoindoline molecules were designed and tested. Compound 5, a 4-tbutylsubstituted aryl sulfonamide and containing a pyridine ring attached at its 3-position, was found to possess submicromolar levels of CCR9 activity (Ki 479 nM). With this compound in hand, a number of close analogues were made in order to survey the initial SAR (Table 1). It was found that other substituents could be introduced onto the aryl sulfonamide ring in place of a 4-tbutyl group (e.g., compounds 6−9) some of which, such as in compounds 7 and 8, showed slightly improved primary potency in the MOLT-4 calcium flux assay. However, other data generated with these molecules in the CCR9 OE calcium flux and MOLT-4 chemotaxis assays (data not shown) showed unpredictable and wide discrepancies. This coupled with some stability concerns did not make the alternative substitutions attractive for further elaboration either initially or later in the program. The most important change made at this point was the introduction of a substituent into the 7-position of the 1,3-dioxoisoindoline ring. Methyl (compound 17), cyano (compound 16), and, especially, chloro (compound 11) substituents improved the primary potency. However, larger groups in this position were generally detrimental. Molecular docking studies (Figure 2) into a

Table 1. FLIPR Data Obtained for Compounds 5−18 in MOLT-4 Cells: Initial Surveya

R1

R2

R3

n

B

MOLT-4 Ki (nM)

Bu H OCF3 Cl F t Bu t Bu t Bu t Bu t Bu t Bu t Bu t Bu t Bu

H CF3 H CF3 CF3 H H H H H H H H H

H H H H H H Cl Cl Cl Cl Cl CN Me Me

0 0 0 0 0 1 0 0 1 1 2 1 0 1

N N N N N N N N+-O− N N+-O− N N N N

479 2713 123 206 1322 618 25 16 22 13 100 19 189 456

cpd 5 6 7 8 9 10 11 12 13 14 15 16 17 18

t

a

Values quoted are geometric means obtained from at least two separate experiments. For n = 2 individual replicates are within 3-fold and for n ≥ 3, SEM does not exceed 0.2 log units. The assay was carried out as outlined in the Experimental section. Compound 1, used as a reference, produced a Ki of 1.1 nM.

C

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2. (a) Overlay of two CCR9 antagonists, compounds 5 (pink) and 11 (cyan) docked into the binding pocket of a model of the CCR9 receptor. (b) Detail of the interactions of the chloro atom in compound 11 with residues in the binding pocket responsible for the enhanced affinity of this molecule. All distances quoted are in angstroms.

within its active site. The tbutylphenylsulfonamide portions from the two molecules overlay in close proximity, but the space occupied by the 1,3-dioxoisoindoline ring in each molecule is subtly different. First, these rings are not quite coplanar. Second and more obviously, one molecule appears to be significantly rotated around an axis, perpendicular to the plane of the ring, relative to the other. These changes allow the extra carbon atom in the linker of compound 14 to be accommodated, while simultaneously permitting the pyridine rings in both the molecules to fit to about the same depth within the aromatic binding pocket. This model suggests that the mouth of the binding pocket is of sufficient size to permit its access from two different approach vectors. As mentioned above, alterations to the substituent or the substitution pattern of the aryl group attached to the sulfonamide were ultimately unproductive. The next step was to examine structure activity dependencies for changes around the pyridine ring, with this group either attached directly to the 1,3-dioxoisoindoline ring or with a one carbon atom separation. For both groups of compounds reported here, the rest of the molecule was held constant, with a 4-tbutylphenylsulfonamide group in the 4-position and a chloro group substitution in the 7-position of the 1,3-dioxoisoindoline ring. Looking first at compounds with a direct attachment between the pyridine and the 1,3-dioxoisoindoline rings (Table 2), a series of pyridines, attached at the 3-position and substituted in the 5-position with a modest variety of groups, were prepared. Among these, data from compounds 19−22, suggested that as long as a substituent was not too large, there was little preference for either modestly electrondonating or -withdrawing substituents for this combination. Thus, it was decided to probe further substitutional dependencies with a combination of methyl, methoxy, ethoxy, and nitrile groups with halo groups used additionally, if the chemistry readily permitted. This strategy allowed compounds

model of the receptor suggest that the chloro group in compound 11 picked up a number of extra receptor residues, relative to the situation with compound 5. There were interactions of both molecules with a number of side-chains, such as Tyr 126 and Ile 197, which help to form a hydrophobic pocket in which to accommodate the tbutylphenyl group and with several polar amino acids on the receptor such as Lys 127, Lys 218, and Thr 199, which are thought to bind with the sulfonyl oxygen and pyridine nitrogen. However, contacts were seen with compound 11 alone (Figure 2b) between the chloro group and Ile 279, Phe 130, and Val 276. These extra interactions, none particularly dominant, may be giving rise to the approximate 20-fold enhancement in binding potency observed. Two other SAR trends were noted in this initial study. First, compounds such as 5 and 11, where the pyridine ring is attached directly to the 1,3-dioxoisoindoline ring, are of similar potencies to analogues such as 10 and 13, where there is a single carbon linker. Comparing 13 to 15, which has a two carbon linker, sees the longer chain start to become less potent. Hence, it was decided to concentrate future efforts on directly linked and C1-bridged molecules. Second, compounds were prepared (12 and 14) where the pyridine nitrogen had been oxidized to a pyridine N-oxide moiety. This change led to compounds with similar potency to the parent pyridines and with generally equivalent or superior ADME-related properties such as solubility. On this basis and in future studies, a significant number of pyridine N-oxides were prepared from some of the more interesting pyridine compounds, in an attempt to fine-tune drug-like properties in the evolving series. Figure 3 shows a comparison of the binding modes of compounds 12 (Table 1, n = 0) and 14 (Table 1, n = 1), predicted by docking in the CCR9 receptor model and potentially demonstrates how the receptor can accommodate both directly linked and one carbon-separated compounds D

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 2. FLIPR Data Obtained in MOLT-4 Cells for Compounds 19−54 Directly Linking the Pyridine to the 1,3Dioxoisoindoline Skeletona

cpd

A

B

C

D

E

MOLT-4 Ki (nM)

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

CH CH CH CH C-Me C-Me C-OMe C-OMe C-F CH CH CH CH CH CH C-Me C-CN C-OMe C-Me C-OMe C-Me C-OMe CH CH N N N N N N N N N N CH C-Me

N N N N N N+-O− N N+-O− N N N N N N N N N N N N N N N N CH CH CH CH CH C-Me C-CN CH CH C-Me CH CH

CH CH CH CH CH CH CH CH CH C-Me C-OMe C-F CH CH CH CH CH CH C-OMe C-OMe C-Me CH CH CH CH CH CH CH C-Me CH CH C-F CH C-OMe N N

C-OMe C-OEt C-Me C-CN CH CH CH CH CH CH CH CH CH CH CH C-Me C-Me C-Me CH CH CH CH C-OMe C-OMe CH CH C-Me C-CN CH CH CH CH C-F CH CH CH

CH CH CH CH CH CH CH CH CH CH CH CH C-Me C-OMe C-OEt CH CH CH CH CH CH C-Me C-Me C-Et CH C-OMe CH CH CH CH CH CH CH CH CH CH

22 128 30 73 3 6 3 4 7 19 78 19 7 2 9 11 18 16 14 8 5 28 18 25 6 22 54 45 21 5 30 10 22 16 11 6

Figure 3. Alternate views of the overlay of two CCR9 antagonists, compounds 12 (cyan) and 14 (green) docked into the binding pocket of a model of the CCR9 receptor. All distances quoted are in angstroms.

to retain a molecular weight close to a notional limit of approximately 500. Continuing the exploration of the pyridines attached at the 3position, substituents were introduced into the 2-position, ortho to both the nitrogen group and the attachment point of the pyridine. Of the molecules examined, it was found that the methyl (23, 3 nM), methoxy (25, 3 nM) and fluoro (27, 7 nM)) derivatives provided excellent primary activity, in the nanomolar range. Pyridine N-oxides corresponding to compounds 23 (compound 24) and 25 (compound 26) maintained these levels of potency. Several compounds (e.g., 28−30) were prepared with substitutions at the 6-position of the pyridine, but the potency was found to be reduced relative to the molecules substituted at the 2-position. Finally, monosubstitutions in the 4-position were considered. Here, again, by substituting ortho to the point of attachment, compounds 31− 33 all showed potency in the nanomolar range (7, 2, and 9 nM, respectively). Although only a relatively small effect, in the 3-pyridine direct linked series, ortho substitution to the point of attachment was found to be optimal. This is presumably due to the substituent acting to slow down any torsional

a

Values quoted are geometric means obtained from at least two separate experiments. For n = 2 individual replicates are within 3-fold and for n ≥ 3, SEM does not exceed 0.2 log units. The assay was carried out as outlined in the Experimental section. Compound 1, used as a reference, produced a Ki of 1.1 nM.

movements of the two rings and hence reinforcing conformations, whereby the pyridine nitrogen is in position to interact with its conjugate partner, Lys 218, on the receptor. Further analogues were prepared with additional substituents on the 3-pyridine ring (compounds 34−42). Although two compounds, 38 (8 nM) and 39 (5 nM), where two methoxy and two methyl groups, respectively, flanked the pyridine nitrogen, gave potency in the nanomolar range, none exceeded E

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

in compounds 56 and 14, respectively, both slightly more active than the corresponding parent in the primary assay. The pyridine N-oxide of 69 (compound 70) is of note, as it loses activity by about 4-fold relative to the pyridine itself, an unusual observation for this normally very conservative modification. In contrast to the compounds in Table 2, where substitution in the ortho position to the point of attachment was generally beneficial to potency, all the compounds with a single carbonatom linker did not improve with substituents added. Indeed, with one exception, potency was significantly reduced on introduction of the candidate substituents, especially when the sterically most demanding methoxy groups were used. The exception is compound 59 where the small, fluorine substituent had been introduced and where potency was slightly improved. As would be expected, the extra carbon in the linker between the pyridine and the 1,3-dioxoisoindoline rings allowed the pyridine a greater locus in which to move, which probably takes it to the extremes of the binding pocket. Extra substituents with any significant steric demand would be difficult to accommodate in such a situation. Given the relatively flat SAR, we were in the fortunate position of having many compounds to progress through our screening cascade and examine for drug-like properties. A considerable number of compounds were progressed for further in vitro assessment, some of which are shown in Table 4. The

the results obtained with analogous mono 2- or 6-substituted materials. The unsubstituted analogues with the point of attachment in the 2-position (43) and the 4-position (53) of the pyridine were prepared, and both were found to be modestly more potent than compound 11, the equivalent 3-pyridine. A number of substituents were introduced to compound 43, but in general this was found to be a detrimental tactic for the 2-attached pyridines, and, with the exception of the 6-methyl derivative (48), which was equipotent with the parent, a several fold loss of potency was observed (compounds 44−47, 49−52). A similar pattern (data not shown) was observed with the 4position attached pyridine analogs, with only the introduction of a 3-methyl group (54) leading to materially enhanced potency. It is likely, given the relative lack of dependence of activity on the nature or position of substitution on the pyridine ring, that the role of these groups, providing that they are relatively small, is to effect the precise conformation of the ring relative to the isoindoline, rather than to interact directly with the receptor. Analogues (Table 3) were also prepared of the prototype, C1-bridged molecule (13, 22 nM). The 2-attached pyridine compound (55) maintained a similar potency, whereas the 4attached pyridine analogue (69) improved primary activity by about 3-fold (8 nM). N-oxide formation of 55 and 13 resulted

Table 4. FLIPR and Chemotaxis Data Obtained Respectively in CCR9 Overexpressing Cells (CCR9 OE) and MOLT-4 Cells for Selected Compoundsa

Table 3. FLIPR Data Obtained in MOLT-4 Cells for Compounds 55−71 with a One Carbon Atom Spacer between the Pyridine to the 1,3-Dioxoisoindoline Skeletona

cpd

A

B

C

D

E

MOLT-4 Ki (nM)

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71

N N+-O− N N+-O− N N CH CH CH CH CH CH CH C-OMe CH CH CH

CH CH CH CH CH C-Me N N N N N+-O− N N N CH CH C-OMe

CH CH C-Me C-Me CH CH CH CH CH CH CH C-Me C-OMe CH N N+-O− N

CH CH CH CH CH CH C-CN C-Me C-OMe CH CH CH CH CH CH CH CH

CH CH CH CH C-F CH CH CH CH C-Me C-Me CH CH CH CH CH CH

20 6 51 13 13 58 31 23 59 24 12 85 123 78 8 30 108

cpd

FLIPR CCR9 OE Ki (nM)b

MOLT-4 chemotaxis Ki (nM)c

13 14 23 24 25 26 27 32 39 43 48 56 69

27 74 12 13 3 6 26 35 6 12 6 10 17

357 38 14 24 32 113 91 3 >300 9 19 40 >100

a

Values quoted are geometric means obtained from at least two separate experiments. The assays were carried out as outlined in the Experimental section. bFor n = 2, individual replicates are within 4fold, and for n ≥ 3, SEM does not exceed 0.3 log units. Compound 1, used as a reference, produced a Ki of 3.7 nM in this assay. cFor n ≥ 3, SEM does not exceed 0.5 log units. Compound 1, used as a reference, produced a Ki of 10.0 nM in this assay.

most promising compounds showed a good congruence in potency between the two FLIPR-based assays and the functional chemotaxis measured in MOLT-4 cells. In our hands, this last assay was too variable to be considered as a reliable primary data source, but was used to give confidence that compounds could elicit a functional response. Examining these results, alongside information on solubility, metabolic stability as measured in microsomes (Table 5) and general selectivity considerations allowed for a group of compounds to be identified as candidates for PK studies, the data for some of which are reported in Table 5. The excellent PK profiles

a

Values quoted are geometric means obtained from at least two separate experiments. For n = 2 individual replicates are within 3-fold and for n ≥ 3, SEM does not exceed 0.2 log units. The assay was carried out as outlined in the Experimental section. Compound 1, used as a reference, produced a Ki of 1.1 nM. F

DOI: 10.1021/acs.jmedchem.5b01840 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 5. ADME and PK Data for Selected Compounds cpd

solubility pH 7.4 (μM)a

human liver microsomes % remainingb

mouse liver microsomes % remainingc

mouse iv t1/2 (h)d

clearance (mL/min/kg)d

mouse AUC po (ng·h/mL)e

mouse Fabs (%)f

1 14 23 24 25 32 48 56

6 126 81 114 26 10 27 69

73 15 89 71 67 81 61 81

9 8 62 76 64 60 42 79

0.16 0.62 3.26 1.28 5.37 1.03 1.32 1.05

103 3.92 2.60 4.69 1.73 6.86 2.34 2.57

340 13892 46837 16230 23998 35594 67859 35999

22 33 73 48 25 100 95 56

a

Solubility of test compounds in pH7.4 buffer. Mean based on three replicates. bHuman liver microsomes incubated with the test compound at 37 °C. The data are the mean of 3 replicates. Compound disappearance monitored over 30 min period. SEM is