Chemokine Receptors as Drug Targets

Martine J. Smit, Sergio A. Lira, Rob Leurs, Raimund Mannhold, Hugo Kubinyi, Gerd Folkers

• Publisher: Wiley-VCH
• Year: 2010
• Print ISBN:
  9783527321186
• eBook ISBN:
  9783527632343
• Edition: 1
• Topic: Medicine
• Subtopic: Pharmacology

Chemokines are hormone-like signaling molecules secreted by cells to signal infection and guide the immune response. Following a decade of basic chemokine research, the pharmaceutical industry has now begun to exploit this crucial signaling pathway for the development of innovative drugs against AIDS, cancer, neural and autoimmune diseases. Here is the first reference focusing on these novel drug development opportunities.
Opening with a general introduction on chemokine function and chemokine receptor biology, the second part covers the known implications of these signaling molecules in human diseases, such as cancer, neural disorders, and viral infection, including AIDS. The third part systematically surveys current drug development efforts at targeting individual chemokine receptors, as well as other chemokine interaction partners, including up-to-date reports from the pharmaceutical industry.

Part One

Fundamentals of Chemokines and Chemokine Receptors

1

Structural Aspects of Chemokines and their Interactions with Receptors and Glycosaminoglycans

Amanda E. I. Proudfoot, India Severin, Damon Hamel, and Tracy M. Handel

1.1 Introduction

Chemokines are a large subfamily of cytokines (∼50 in humans) that can be distinguished from other cytokines due to several features. They share a common biological activity, which is the control of the directional migration of leukocytes, hence their name, chemoattractant cytokines. They are all small proteins (approx. 8 kDa) that are highly basic, with two exceptions (MIP-1α, MIP-1β). Also, they have a highly conserved monomeric fold, constrained by 1–3 disulfides which are formed from a conserved pattern of cysteine residues (the majority of chemokines have four cysteines). The pattern of cysteine residues is used as the basis of their division into subclasses and for their nomenclature. The first class, referred to as CXC or α-chemokines, have a single residue between the first N-terminal Cys residues, whereas in the CC class, or β-chemokines, these two Cys residues are adjacent. While most chemokines have two disulfides, the CC subclass also has three members that contain three. Subsequent to the CC and CXC families, two additional subclasses were identified, the CX₃C subclass [1, 2], which has three amino acids separating the N-terminal Cys pair, and the C subclass, which has a single disulfide.

The first chemokine, PF-4, was identified in 1977 [3] but it was not for almost a decade that other members of the family started to emerge, with the discovery of the proinflammatory chemokines: IP-10 was identified in 1985 as a protein showing homology to PF4 [4], while IL-8 and the MIP-1 proteins were isolated in the late 1980s as active protein from tissues or culture supernatants. The neutrophil chemoattractant, IL-8, was purified from culture supernatant of stimulated blood monocytes [5] and the monocyte chemoattractants MIP-1α and MIP-1β were purified from LPS-stimulated mouse macrophages [6]. The primary amino acid sequence of these chemokines rapidly led to the identification of the highly conserved four-cysteine motif described above and also allowed their classification into the two principal subclasses. The number of chemokines then grew rapidly through homology cloning using the conserved motifs, but the real explosion in the identification of members came from EST database searches [7]. Initially, chemokines were given names usually associated with their activity; for example, the MIP-1 proteins were discovered as “macrophage inflammatory proteins”. Similarly, PF-4 (platelet factor IV) was a factor produced from platelets. However, since members of the family were often identified concomitantly by different laboratories resulting in different names, a systemic nomenclature was introduced in 2000 in order to introduce harmonization [8]. In this nomenclature, the ligands are named according to subclass (CC, CXC, C, CX3C) followed by L for ligand and a number. Under this nomenclature IL-8 became CXCL8 while MIP-1α became CCL3. This nomenclature was created for human chemokines based on their genomic localization, but was rapidly “pirated” for the mouse chemokines, since even prestigious journals insisted that the new nomenclature be applied to the mouse chemokines! Interestingly certain chemokines are not found in both the human and mouse systems. For instance CXCL8 does not exist in the mouse, and the equivalent of several mouse chemokines such as lungkine and MCP-5 (CXCL15 and CCL12, respectively), have not been identified in humans (as shown in Table 1.1), which shows the old and new nomenclatures for human chemokines. In the rest of the chapter, we refer to chemokines by their new nomenclature.

Table 1.1 The old names and the new systematic nomenclature of the human chemokines are listed side by side and grouped into their respective CXC, XC, CX3C and CC families. For example, I-309 (old) is now referred to as CCL1 (new). Mouse chemokines for which no human homologs has been identified are shown in parentheses

Initial support for the division of chemokines into the α (CXC) and β (CC) subclasses was not only structural, but also based on biological activity as it described leukocyte specificity. The discovery that chemokine receptors were seven transmembrane spanning G protein-coupled receptors (GPCR) in the early 1990s [9, 10] was extremely important for the pharmaceutical industry as it presented a novel target class in the GPCR family which represent up to 60% of the targets of marketed medicines. The initial hope was that individual leukocyte populations would express a single chemokine receptor, which held firm until the cloning of the third CXC receptor, CXCR3 [11]. Until this point, the CXC chemokines were thought to be responsible principally for neutrophil recruitment and were therefore implicated in acute inflammation, while CC chemokines recruited other leukocyte types and were thus involved in chronic inflammation. However CXCR3 is mainly expressed on activated T cells, and its ligands were initially identified as IFNγ inducible polypeptides and are therefore pivotal in chronic inflammatory disorders. The subsequent identification of CXCR4 and CXCR5, as well as several CC chemokine receptors and their respective ligands, then introduced yet another concept in chemokine biology – that chemokines could be further subdivided into two broad classes on the basis of: (i) those that are inducible and therefore involved in inflammation and (ii) those that are constitutively expressed and are involved in leukocyte homing.

This chapter concentrates on the structure of chemokines and their receptors and how these aspects may be related to their biology. Understanding the relationship between the structure and function of chemokines has lead to ideas of how chemokines can be modified to produce analogs that are useful for modifying disease, in animal models and perhaps in man in the future.

1.2 Receptor–Ligand Interactions

The classification of chemokine receptors is based on the ligands they bind, in other words CXC receptors bind CXC ligands, CC receptors bind CC ligands and so on, as shown in Figure 1.1. Chemokine receptors have been identified that bind chemokines but do not signal. One of these, the Duffy antigen receptor for chemokines (DARC) is a promiscuous chemokine receptor expressed on erythrocytes that binds both CC and CXC ligands [12]. In contrast, the decoy receptor D6 only binds CC chemokines [13]. Thus, with some exceptions like DARC, the chemokine system is specific with respect to the binding pattern, in that chemokines in each class do not bind to receptors of another class. However, binding across classes has been demonstrated with antagonists. Thus CXCL9, CXCL10 and CXCL11, the agonists of CXCR3, also bind to CCR3 as antagonists and inhibit Th2 cell migration [14]. Interestingly a chimera consisting of the first eight residues of CCL11 (and the remainder consisting of CXCL11) bound CCR3 more strongly than the parental chemokines. CCR3 can also be antagonized by CCL18, despite the fact that the receptor for CCL18 remains unidentified to date [15]. Viruses have also adopted deviations from classical chemokine receptor pharmacology. For example the virally encoded chemokine receptor, US28, does not demonstrate reciprocal heterologous competition like most chemokines/receptors. Instead, one of its ligands, CX3CL1, cannot be competed by certain CC chemokines, whereas these CC chemokines are all displaced by CX3CL1 [16]. We believe that the study of the virally encoded members of the chemokine system will teach us a great deal about the intricacies involved in chemokine/receptor interactions, since viruses have produced chemokine ligands, such as vMIP-II, that can bind across chemokine receptor subclasses.

Figure 1.1 This diagram shows the pairing of chemokine receptors with their respective ligands. For example, CCL2, CCL8, CCL7 and CCL13 are all ligands of the receptor, CCR2. Some receptors like CXCR4 are much more selective and have a single ligand.

Beyond the basic rule of subclass selectivity (with the exceptions noted), the binding patterns of the chemokine system is far from simple! First, the assignment of receptor-ligand pairs arises from in vitro assays, and one should be aware that the situation in vivo may be different due to factors that cannot be captured in vitro. Second, the majority of receptor/ligand interactions are not specific in that several receptors bind more than one ligand – in fact only about one-third are specific single ligand binders to date. Third, the reason that this statement is qualified by “to date,” is that as the identification of new ligands continued, absolute specificity has tended to disappear, although the question remains as to whether there are additional ligands to be identified. CXCR1 was classified as a specific receptor for CXCL8 for seven years, until CXCL6 was identified as a ligand [17]. As an extreme example, CCR1 binds at least eight ligands. The situation is further complicated by the fact that certain chemokines are ligands of more than one receptor, which is best exemplified by CCL5 which binds to CCR1, CCR3 and CCR5.

However, the biology that has emerged over the past decade or so has identified a broad definition which supports the classification of selective versus shared receptors. The selective receptors have been shown to generally correspond to those which are constitutively expressed and are involved in development and homeostasis. In contrast, the shared receptors are those which are inducible and associated with inflammatory disease [18]. The fact that the shared receptors are the “villains” in disease makes the task of understanding how to target them a challenge, particularly if one is interested in using neutralizing antibodies against the ligands. Intuitively one would suggest that a small molecule inhibitor of the receptor would be the chosen strategy, or alternatively a neutralizing receptor antibody, but neither of these strategies is that simple. Therefore neutralization of a prominent ligand could be a successful strategy – one could suggest CCL2 for CCR2 or CXCL10 for CXCR3? However, it is not always easy to establish which ligand is the most potent and has the highest affinity for a certain receptor. This is well illustrated by CCR5, and a comparison of the rank order of the published potencies of its ligands. Using a calcium mobilization assay, the rank order potency of ligands on CCR5 expressed in CHO cells was reported as CCL5 > CCL4 > CCL3 [19] whereas in RBL cells stably transfected with CCR5, the rank order was CCL5 > CCL4 = CCL3 [20] and in HEK293/CCR5 transfectants the order was different again, with CCL3 > CCL5 > CCL4 [21]. However in the third example, the form of CCL3 used was the allelic variant, CCL3L1 (LD78β instead of LD78α), which has a Pro instead of a Ser residue at position 2 and two S/G switches (Figure 1.2a). Thus although CCL3 is often described as being a ligand for CCR5, its affinity is approximately 100 nM, whereas CCL3L1 has an affinity of 1 nM.

Figure 1.2 Alignment of the allelic CCL3 variants (a) and the splice variants of CXCL12 (b).

Another complexity arises from the fact that although certain chemokines can bind to several different receptors, the induced biological activity may differ significantly and can even vary depending on the cell type on which the receptor is expressed. CCL5 induced downregulation of three of its receptors and the ensuing recycling illustrates this phenomenon nicely. On incubation with CCL5 in vitro, the surface expression is reduced by approximately 80% in each case of CCR1 and CCR5 from the surface of PBMC [22, 23] and CCR3 from eosinophils [22, 24]. However on removal of the chemokine from the culture medium, very different patterns of receptor recycling are observed. In the case of CCR5, receptor density returns to that observed initially [23]. With CCR3, only 70–80% of the initial receptor density is observed, but with CCR1, no recycling is observed [22]. While the CCR3 receptors that do not recycle have been shown to traffic to the lysosomal compartment where they are degraded, the fate of CCR1 remains to be established. Therefore the apparent redundancy of a chemokine binding to more than one receptor may not be as redundant as meets the eye.

An additional layer of complexity has been found for the chemokine CXCL12, where six splice variants have been identified (Figure 1.2b) [25]. The main difference is the extended C-termini of the δ and γ isoforms. The γ isoform has an extremely large number of basic residues resulting in a significantly increased a...

Table of contents

• Cover
• Title Page
• Copyright
• List of Contributors
• Preface
• A Personal Foreword
• Part One Fundamentals of Chemokines and Chemokine Receptors
• Part Two Chemokine Receptors in Disease
• Part Three Targeting Chemokine Receptors
• Index