Exploring a pocket for polycycloaliphatic groups in the CXCR3 receptor with the aid of a modular synthetic strategy Maikel Wijtmans, Dennis Verzijl, Cindy M. E. van Dam, Leontien Bosch, Martine J. Smit, Rob Leurs, Iwan J. P. de Esch * Leiden/Amsterdam Center for Drug Research, Division of Medicinal Chemistry, Faculty of Sciences, VU University Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands article info Article history: Received 10 January 2009 Revised 22 February 2009 Accepted 24 February 2009 Available online 27 February 2009 Keywords: CXCR3 Reductive amination GPCR Polycycloaliphatic abstract A CXCR3 pocket capable of accommodating polycycloaliphatics was explored using a modular synthetic strategy. The systematic studies reveal that the tricyclic 2-adamantane and bicyclic (iso)bornyl group are efficiently recognized by CXCR3. Ó 2009 Elsevier Ltd. All rights reserved. CXCR3 is a G protein-coupled chemokine receptor involved in a variety of inflammatory and infectious diseases and in certain metastasis processes. 1,2 Several series of small CXCR3 antagonists have been developed to explore the associated physiological and clinical roles of CXCR3. 1–10 Of these, azaquinazolinone AMG487 (1, Fig. 1) is the most widely described in the public domain, while some of the highest affinities reside with the broad class of piper- azinyl-piperidines 2 (e.g., 2a/b: pIC 50 = 9.5/9.7, respectively). 2,3,10 The diversity amongst reported CXCR3 antagonists is high and it remains a challenge to identify ‘CXCR3-priviliged’ molecular motives. 2 Our attention was drawn to a few bicycloaliphatic groups that seem to be readily accommodated by CXCR3. 8,9 Archetypical is the unique ()myrtenyl group, which was found in SAR studies on 1-aryl-3-piperidin-4-yl-ureas (e.g., 3). 8 In general, several examples are known in which chemokine receptor antagonists pick up considerable affinity by binding to hydrophobic domains of the chemokine receptor. 11,12 Therefore, with the goal of contrib- uting to the understanding of CXCR3–ligand binding, we decided to explore a hypothesized CXCR3 pocket for polycycloaliphatics through a modular synthetic approach. Our design strategy involved equipping the benzyl-aminopi- peridine part of piperazinyl-piperidine 2 with the polycycloali- phatic groups of choice (Fig. 2). The rationale for this strategy was twofold: (1) We envisioned a highly modular synthetic proto- col in which a reductive amination using simple building blocks is key, paving the way for efficient identification and exploration of the pocket. (2) The reported picomolar-affinity of 2 may allow the removal of a substantial molecular portion without absolute loss of affinity. In the synthesis of virtually all products, two or three subse- quent reductive aminations were used to install the benzyl, poly- cycloaliphatic and methyl groups (Scheme 1). Routes A and B both start from a mono-protected dibasic core and carbonyl build- ing blocks but the peripheral groups are installed in different order. Routes C and D are essentially the same as routes A and B but use amine-building blocks for the polycycloaliphatic unit. Last, route E adds an N–Me through an Eschweiler–Clark reductive amination. Almost all reductive aminations proceeded smoothly with func- tionalities such as double bonds, azides and pentafluorophenyl groups left intact. Many non-commercial building blocks were pre- pared by methods described in the literature (see Supplementary data). A few building blocks and final compounds required alterna- tive approaches (Scheme 2). 1-Adamantyl substituted pyridine 4 was benzylated to salt 5 followed by a swift reduction to 22. Benzylation of 4 by 4-ClBnBr was preferred instead of by 4-Cl,2- F–BnBr. Sequential treatment of keto-ester 6 with 4-chloro-2- fluorobenzaldehyde and then 2-adamantaneamine gave enamine 31. This enamine resisted NaBH(OAc) 3 , but NaCNBH 3 in AcOH provided a separable mixture of diastereomeric esters 32 and 33. The relative stereochemistry of these could not be unambigu- ously determined. Aniline 57 was obtained from azide 56 by a Zn- based reduction. Urea 37 was prepared by corresponding isocya- nate-chemistry. 0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.bmcl.2009.02.093 * Corresponding author. Tel.: +31 20 5987600; fax: +31 20 5987610. E-mail address: ideesch@few.vu.nl (I.J.P. de Esch). Bioorganic & Medicinal Chemistry Letters 19 (2009) 2252–2257 Contents lists available at ScienceDirect Bioorganic & Medicinal Chemistry Letters journal homepage: www.elsevier.com/locate/bmcl