Abstract
Chemokines are a family of low molecular weight proteins with an essential role in leukocyte trafficking during both homeostasis and inflammation. The CC class of chemokines consists of at least 28 members (CCL1-28) that signal through 10 known chemokine receptors (CCR1-10). CC chemokine receptors are expressed predominantly by T cells and monocyte-macrophages, cell types associated predominantly with chronic inflammation occurring over weeks or years. Chronic inflammatory diseases including rheumatoid arthritis, atherosclerosis, and metabolic syndrome are characterized by continued leukocyte infiltration into the inflammatory site, driven in large part by excessive chemokine production. Over years or decades, persistent inflammation may lead to loss of tissue architecture and function, causing severe disability or, in the case of atherosclerosis, fatal outcomes such as myocardial infarction or stroke. Despite the existence of several clinical strategies for targeting chronic inflammation, these diseases remain significant causes of morbidity and mortality globally, with a concomitant economic impact. Thus, the development of novel therapeutic agents for the treatment of chronic inflammatory disease continues to be a priority. In this review we introduce CC chemokine receptors as critical mediators of chronic inflammatory responses and explore their potential role as pharmacological targets. We discuss functions of individual CC chemokine receptors based on in vitro pharmacological data as well as transgenic animal studies. Focusing on three key forms of chronic inflammation—rheumatoid arthritis, atherosclerosis, and metabolic syndrome—we describe the pathologic function of CC chemokine receptors and their possible relevance as therapeutic targets.
I. Introduction
A. The Chemokine Family
Chemokines are a family of around 50 low molecular weight polypeptides with a conserved tertiary structure. They are divided into four classes, C, CC, CXC, and CX3C, on the basis of the location of key cysteine residues that participate in disulfide bonding and are either juxtaposed (CC) or separated by 1 or 3 amino acids (CXC and CX3C, respectively). Virtually all chemokines are secreted from the cell after synthesis, with two exceptions, CX3CL1 (fractalkine) and CXCL16 (SR-PSOX), which can remain tethered to the cell surface by a transmembrane mucin-like stalk (Bazan et al., 1997; Matloubian et al., 2000). Chemokines can be broadly classified as homeostatic or inflammatory depending on whether they have a role in physiologic cell trafficking [e.g., CCL19 (ELC) and CCL21 (SLC)] or are synthesized on demand in response to an inflammatory stimulus [e.g., CCL2 (MCP-1)].
Chemokines have a systematic nomenclature based on the class and a numerical designation e.g., CCL3, CXCL10 (Murphy et al., 2000; Zlotnik and Yoshie, 2000). This greatly simplifies the previous system whereby chemokines were named predominantly by function and could therefore have several different names, e.g., CCL2 was originally named monocyte chemoattractant 1 (MCP-1), small inducible cytokine A2 (SCYA2), and monocyte chemotactic and activating factor (MCAF) (Furutani et al., 1989; Yoshimura et al., 1989b; Mehrabian et al., 1991). Pairings between individual chemokine receptors and their chemokine ligands are shown in Fig. 1.
B. Chemokine Structure and Function
The first chemokine, IL-8 (CXCL8), was originally described for its ability to chemoattract neutrophils, and many other chemokines have since been identified by their role in mediating chemotaxis of specific leukocyte subsets (Yoshimura et al., 1987). All chemokines have a similar tertiary structure comprising a disordered N terminus of 6–10 amino acids followed by a long loop (known as the N loop), a 310 helix, a three-stranded beta-sheet, and finally a C-terminal alpha helix (Allen et al., 2007).
The N terminus has a critical role in receptor activation, and N-terminal truncations can render chemokines inactive or even able to act as antagonists. Deletion of residues 2–8 in the N terminus of CCL2, for example, generates a chemokine that still binds to the receptor but fails to inhibit cAMP synthesis and acts as an antagonist in calcium flux and chemotaxis assays (Jarnagin et al., 1999). In contrast, several of the ligands for the chemokine receptor CCR1 undergo proteolytic processing during inflammation to generate full agonists (Proost et al., 2000; Berahovich et al., 2005). Chemokine processing may also alter receptor usage—CCL4 (MIP-1β) can be processed by CD26 (dipeptidyl peptidase IV) to leave residues 3–69, which generates a protein capable of activating CCR1 and CCR2b in addition to CCR5, which is activated by the wild-type chemokine (Guan et al., 2002). Similarly, N-terminal truncation of the CCR4 ligand CCL22 (MDC) generates a chemokine that loses chemoattractant activity for a T-cell line but retains chemotactic for monocytes, suggesting the presence of an alternative receptor that is differentially activated by the two species (Struyf et al., 1998b). N-terminal processing thus exists as a mechanism to finely control the activity of chemokines, facilitating a wider range of cellular responses to a given spectrum of chemokines. Of note, the ability of chemokines to undergo proteolytic processing means that quantification of absolute chemokine levels by enzyme-linked immunosorbent assay (ELISA) may not accurately reflect the amount of active chemokine present in a given sample—more sensitive bioassays may need to be developed to measure chemokine activity.
Structural studies have demonstrated that many chemokines form dimers or tetramers in solution, and some, e.g., CCL5 (RANTES), may form higher order aggregates of several hundred kilodaltons (reviewed in Allen et al., 2007). CCL2 has been crystallized in two forms: a dimeric form and a tetrameric form of two associated dimers with different structures (Lubkowski et al., 1997). The residues required for aggregation in the chemokines CCL3 (MIP-1α), CCL4, and CCL5 have been identified via mutagenesis studies, and, in the case of CCL3, two acidic residues at positions 26 and 66 are required (Czaplewski et al., 1999). Chemokines can also heteroligomerize with other chemokines of the same or different classes (Allen et al., 2007). For example, CXCL4 (platelet factor 4) and CCL5 are both stored in platelet α granules and were found to heteroligomerize via a specific region of CCL5 required for aggregation (von Hundelshausen et al., 2005). This interaction has a functional consequence: amplifying the effect of CCL5 on monocyte adhesion to activated endothelial cells. However, it is known that chemokines bind their receptors as monomers, because mutants that are unable to oligomerize retain the same receptor binding affinity and in vitro chemotactic activity as wild-type chemokines (Czaplewski et al., 1999; Proudfoot et al., 2003; Allen et al., 2007).
In vivo, most chemokines are thought to bind to glycosaminoglycans (GAGs), proteoglycans such as heparin sulfate and chondroitin sulfate that are expressed on the surface of most cells (Allen et al., 2007). Proudfoot et al. (2003) determined the GAG binding regions of CCL2, CCL3 (MIP-1α), and CCL5 and generated mutants of these chemokines that show reduced heparin-binding activity. In vitro, these chemokines retain equivalent chemoattractant activity to wild-type proteins. However, intraperitoneal administration of the GAG-binding mutants in mice fails to recruit leukocytes, demonstrating the importance of GAG binding for establishment of chemokine gradients in vivo.
C. Receptor Structure and Signaling Mechanisms
All chemokine receptors are class A G protein-coupled seven-transmembrane receptors that induce signal transduction via Gi and occasionally other G proteins. Thus, most chemokine responses can be inhibited by pertussis toxin (PTX) (Wu et al., 1993). By using cells transfected with a variety of G protein-coupled receptors (GPCRs) that couple to Gi, Gs, or Gq heterotrimeric G proteins, Neptune and Bourne (1997) demonstrated that Gi activation is necessary for chemotaxis. In addition, the cotransfection of these cells with Gαtransducin, which sequesters Gβγ subunits, completely prevented cell migration. indicating that Gβγ subunits activate pathways required for chemotaxis (Neptune and Bourne, 1997).
The network of signaling pathways activated by chemokines is complex, allowing multiple functional outcomes including chemotaxis, adhesion, proliferation, and control of gene expression. There is little consensus about canonical pathways for chemokine receptor signaling, because the pathways activated depend on the specific receptor and cell types involved. Many of the early studies focusing on chemokine receptor signaling used transfected cells overexpressing a given receptor, which makes it hard to extrapolate these findings to signaling in primary cells. What is clear is that the majority of chemokines induces calcium flux in cells expressing their cognate receptor, either via calcium influx or release from intracellular stores. This is often induced via direct activation of phospholipase C by Gβγ subunits, causing an increase in intracellular inositol trisphosphate (IP3) and diacylglycerol (DAG) and leading to calcium release from intracellular stores. Numerous other studies have shown an important role for phosphoinositide 3-kinase (PI3K) in coordinating chemokine responses via coupling to downstream effectors, including extracellular regulated kinase (ERK) and Akt (Curnock et al., 2002). It is clear that chemotaxis requires the integration of multiple signaling pathways, culminating in actin polymerization to facilitate movement toward any given chemokine stimulus.
The structural regions required for ligand binding and receptor activation have been defined for several chemokine receptors. For many chemokine receptors, such as CCR2, the N terminus is a critical determinant of ligand binding and is involved in, but not sufficient for, efficient signal transduction (Monteclaro and Charo, 1996; Samson et al., 1997). The N terminus is often glycosylated or tyrosine sulfated, and this may be required for high-affinity chemokine binding, as is the case for CCR5 (Bannert et al., 2001). Other regions of CCR5 required for ligand selectivity and intracellular signaling have been determined, with the second extracellular loop of the receptor having a key role (Samson et al., 1997). In contrast, for another chemokine receptor, CX3CR1, residues in the N terminus and third extracellular loop are required for ligand binding and receptor signaling (Chen et al., 2006). A two-step model for chemokine receptor binding and activation has been postulated, whereby the N terminus and extracellular loops of the receptor are involved in binding the core domain of the chemokine ligand whereas the N terminus of the chemokine penetrates directly into the helical bundle of the receptor (Gupta et al., 2001). Chemokine receptors, in common with other rhodopsin-like GPCRs have a DRY (aspartic acid–arginine–tyrosine) conserved motif at the cytoplasmic end of the third transmembrane segment. This region has been shown to be critical to signaling, because chemokine-like receptors that lack the DRY motif, such as the receptor D6, act as molecular sinks or decoys for chemokines, showing binding but no intracellular signaling in response to chemokines (Bonini et al., 1997; Nibbs et al., 1997). The C terminus of the receptor, as for many GPCRs, contains key serine and threonine residues which can be phosphorylated by G protein-coupled receptor kinases (GRKs) to induce recruitment of arrestin proteins leading to receptor internalization and signal termination (Vroon et al., 2006). It is now known that aside from its role in receptor internalization, beta arrestin recruitment may also enable the formation of a G protein-independent signaling complex (reviewed extensively in Rajagopal et al., 2010). Because many chemokine receptors are promiscuous in their ligand binding, differential activation of G protein- versus β-arrestin-dependent signaling pathways may facilitate functional selectivity of cellular responses. This may also be regulated at the level of phosphorylation, whereby different GRK recruitment may modulate signaling complex formation, as is the case for CCR7, which binds both CCL19 and CCL21 with different outcomes (Zidar et al., 2009). The generalized structural regions of chemokine receptors required for function are summarized in Fig. 2.
To date, only a single chemokine receptor, CXCR4, has been crystallized and the structure determined by generation of a stabilized receptor with a T4 lysozyme fusion inserted between helices 5 and 6 and one or more thermostabilizing point mutations (Wu et al., 2010). The receptor was cocrystallized with small molecule and cyclic peptide antagonists, aiding in identification of the ligand binding site, which seems to differ significantly from other GPCRs. The small molecule isothiourea derivative (IT1t) was cocrystallized and found to contact helices 1, 2, 3, and 7. The authors suggest that the binding of this molecule mimics the binding of the chemokine N terminus, which is believed to protrude into the helical bundle of the receptor in the two-step model of chemokine receptor activation. Furthermore, CXCR4 was found to consistently exist as a homodimer, with the key regions for homodimerization identified as transmembrane helices 5 and 6. These data support multiple studies in transfected cell systems showing that chemokine receptors can homo and heterodimerize with various consequences (Rodriguez-Frade et al., 1999; Mellado et al., 2001; El-Asmar et al., 2005).
D. Chronic Inflammation
Acute inflammation is the normal response of vascularized tissues to injury, irritation, and infection. Tissue injury or the activation of tissue resident macrophages and mast cells causes the release of vasoactive substances that activate endothelial cells in nearby venules, leading to the elaboration of an inflammatory exudate comprised of cells (initially neutrophils) and plasma proteins (including complement, antibodies, and serum albumin). These local changes in the properties of vascular endothelium lead to the cardinal features of inflammation: heat, redness, swelling, and pain. Acute inflammation is an essential physiologic response to injury and infection that rapidly recruits the cells and molecules of the innate immune response to potential sites of microbial infection (Majno and Joris, 1996). The initial wave of neutrophil recruitment is followed by recruitment of monocytes, which differentiate into macrophages that orchestrate the process of tissue repair. Chronic inflammation is a harmful process that can occur through failure to resolve acute inflammation or through persistence of an inflammatory stimulus (Nathan and Ding, 2010). However, many prevalent inflammatory diseases, for example, atherosclerosis, show no progression from acute to chronic but instead have the hallmarks of chronic inflammation (such as monocyte influx and macrophage differentiation) from the outset. Other chronic inflammatory diseases, such as rheumatoid arthritis (RA), show perpetual recruitment of cells characterizing both acute and chronic inflammatory responses.
In this review we have chosen to focus specifically on the function of CC chemokines in chronic inflammation. CC chemokines are known to have a key role in the recruitment of monocytes and macrophages, cell types crucial to the development of atherosclerosis, RA, and adipose inflammation. Although the CXC family of chemokines have essential roles in neutrophil, B-, and T-cell recruitment, they are outside the scope of this review, and we refer the reader to some excellent reviews on the biology of CXC chemokines and their role in various pathologies (Romagnani et al., 2004; Weathington et al., 2005; Bizzarri et al., 2006; Strieter et al., 2007).
II. CC Chemokine Receptors
A. CCR1
CCR1 was cloned in 1993, several years after the identification of two of its ligands, CCL5 and CCL3 (Schall et al., 1988; Sherry et al., 1988; Gao et al., 1993). Both CCL5 and CCL3, but not CCL2, were found to induce calcium flux in human neutrophils with an IC50 of 5 and 50 nM, respectively, suggesting the expression of a CC chemokine receptor in neutrophils (Gao et al., 1993). During cloning of the first chemokine receptor, IL-8R B (CXCR2), four related cDNAs were identified in the HL-60 promyelocytic leukemia cell line (Murphy and Tiffany, 1991). Complementary RNA corresponding to the longest of these cDNAs was injected into Xenopus oocytes, which were then challenged in a calcium flux assay with various chemokines (Gao et al., 1993). The oocytes were found to respond to CCL3 and CCL5, but not CCL2 or any of the CXC chemokines tested. A second group used a degenerate PCR-based approach to clone the receptor, which was then expressed in human embryonic kidney (HEK) 293 cells and found to bind CCL1 (I-309) and CCL5 with high affinity and respond in a calcium flux assay to these chemokines (Neote et al., 1993b). Further study demonstrated that receptor mRNA was expressed in human B lymphocytes and in the promyelocytic cell lines THP-1 and U937 and in differentiated HL-60 cells, which display a more neutrophil-like phenotype. Finally, the receptor was mapped to chromosomal location 3p21. Once a specific antibody against CCR1 had been raised, a more detailed expression analysis was performed that showed that T cells, NK cells, monocytes, and CD34+ bone marrow progenitor cells expressed CCR1 on the cell surface, whereas no expression could be detected on B cells or granulocytes (Su et al., 1996). CCR1 was found to be expressed preferentially on CD45RO+ memory T cells. The murine CCR1 homolog was cloned a few years later by two independent groups on the basis of homology with the human receptor (Gao and Murphy, 1995; Post et al., 1995).
By using GTPγS exchange, calcium flux and chemotaxis assays in HL-60 cells the CCR1 ligands were placed in the following rank order of potency CCL3 > CCL23 (MPIF-1) > CCL5 ≥ CCL4 (Chou et al., 2002). Furthermore, CCL4 was found to antagonize responses to the other ligands, suggesting it could be an endogenous inhibitor of CCR1 activity. Other chemokines, e.g., CCL15 (HCC2), have also been shown to be weak CCR1 ligands. CCL15 is an unusual chemokine with an extended N terminus of 16–20 amino acids, similar to CCL23 and two murine chemokines with no direct human homolog, CCL6 (C10) and CCL9 (MIP-1γ) (described in Berahovich et al., 2005). An important finding was that these ligands can be proteolytically cleaved to truncate the N terminus and generate agonists with increased binding to CCR1 and approximately 1000-fold higher potency compared with intact forms (Berahovich et al., 2005). This could be mediated by inflammatory proteases, supernatants from human cells, and physiologic fluids. Furthermore, CCL15 and CCL23 truncated at the N terminus but not CCL3 or CCL5 were detected in synovial fluid from RA patients. In contrast, truncation of two amino acids from the N terminus of CCL5 makes the chemokine inactive at CCR1 (Struyf et al., 1998a).
A characterization of CCR1 signaling in human monocytes demonstrated that receptor activation by CCL23 led to phospholipase C (PLC) activation and intracellular calcium release as well as some calcium entry from the extracellular medium (Nardelli et al., 1999). CCL23 was not found to alter cAMP levels in the cell, but did activate phospholipase A2 (PLA2) in monocytes, leading to arachidonic acid release. Inhibition of PLA2 and 5-, 12-, and 15-lipoxygenase completely blocked F-actin polymerization induced by CCL23. In the monocytic THP-1 cell line, CCL15, CCL23, and CCL24 (eotaxin 2) were found to induce phosphorylation of NFκB (Lee and Wong, 2009). This activation was PTX insensitive, suggesting the involvement of G proteins other than Gαi in receptor coupling. In transfected cells, CCR1 was found to couple to Gα14/16 to activate this pathway. A more detailed dissection of this pathway in THP-1 cells using pharmacological inhibitors suggested the involvement of several downstream pathways including PLCβ, protein kinase C, Ca2+/calmodulin dependent protein kinase II, Raf-1, MAPK/ERK kinase (MEK1/2), and c-Src.
Mice lacking CCR1 developed normally, showing no histologic differences in lymphoid organs, peripheral blood counts or any altered susceptibility to spontaneous infections (Gao et al., 1997). However, spleens from knockout mice contained fewer granulocyte-macrophage and multipotential progenitor cells, suggesting defective mobilization of these cells from the bone marrow. Upon lipopolysaccharide (LPS) challenge, Ccr1 knockout mice showed normal egress of progenitor cells from the bone marrow but these cells did not migrate normally to the spleen. Cells found in the spleen were also in a slow- or noncycling state in Ccr1 knockout mice, as opposed to the rapid proliferation of these cells in response to LPS seen in wild-type mice. CCR1 was also found to be essential for the response of murine neutrophils to CCL3, both in migration and calcium flux assays. By using an Aspergillus fumigatus infection model (intravenous administration), where neutrophil function is critical, CCR1 knockout mice showed only 30% survival after 30 days compared with 60% survival in control animals. Mortality was also accelerated, with death occurring within 10 days of infection. In a model of granuloma formation (Schistosoma mansoni egg injection), Ccr1-deficient mice showed reduced granuloma formation in the lung, suggesting defective migration of multiple leukocytes required to form granulomatous tissue. Further analysis demonstrated that knockout mice showed a Th1 skewed cytokine profile in the lung, with enhanced interferon-γ (IFN-γ) production that inhibits granuloma formation.
This excessive Th1 response in Ccr1 knockout mice was highlighted in another study using a nephrotoxic nephritis model (Topham et al., 1999). CCR1-deficient mice showed increased renal injury, as measured with histologic and functional parameters, compared with wild-type animals. Ccr1 knockout mice showed enhanced recruitment of macrophages and T cells into the kidney, suggesting that CCR1 is not essential for the migration of these cells. Further analysis showed a Th1 skewed response in these mice with high IgG2a titers, increased IFN-γ production by splenocytes and enhanced tumor necrosis factor α (TNFα) and TNFβ production by mononuclear cells. This suggests that CCR1 has a role in regulating cell-mediated immune responses in mice.
CCR1 deficiency was also detrimental in a model of Leishmania major infection, a pathogen that differentially affects various inbred mouse strains: C57Bl6/J mice (and others) usually develop a Th1 response and control the infection, whereas Th2 polarized Balb/c mice are highly susceptible to infection and produce large amounts of IL-4. In Ccr1-deficient C57Bl6/J mice, the response to Leishmania was enhanced compared with wild-type (WT), smaller lesions were formed, and the infection was cleared more rapidly (Rodriguez-Sosa et al., 2003). In these experiments, both wild-type and Ccr1 knockout mice produced similar levels of Th1 cytokines, but Ccr1-deficient mice generated significantly less IL-4 and IL-10, cytokines that inhibit parasite clearance in this model.
Deletion of CCR1 may have other beneficial effects; for example it reduces airway remodeling in a pulmonary A. fumigatus model (Blease et al., 2000), attenuates the damaging pathologic response to respiratory syncitial virus infection (Miller et al., 2006), protects against enteritis induced by Clostridium difficile toxin in mice (Morteau et al., 2002) and suppresses cardiac allograft rejection (Gao et al., 2000).
Interestingly, CCR1 has recently been shown to have a critical role in bone formation via effects on the differentiation and function of osteoblasts and osteoclasts (Hoshino et al., 2010). Ccr1 knockout animals were found to have fewer and thinner trabecular bones, and osteoblasts showed defective differentiation. Cultured Ccr1−/− bone marrow cells generated fewer osteoclasts due to reduced cell fusion, and these cells showed no osteolytic activity.
Thus, CCR1 has differential effects depending on the pathologic context: it has a critical role in controlling cell-mediated immunity to enable pathogen clearance, but may also generate detrimental pathophysiological responses, e.g., in airway remodeling. Studies analyzing the phenotype of mice deficient for CCR1 and the other CC chemokine receptors discussed in this review are presented in Table 1.
A recently published abstract has identified an association between a single nucleotide polymorphism (SNP) ∼38 kb from the 3′-untranslated region (UTR) of the CCR1 gene and risk of Behcet’s disease, a complex form of systemic vasculitis characterized by recurrent inflammatory attacks throughout the body (Kirino et al., 2011). This SNP was reported to be in a potential regulatory region of the gene, and the protective minor allele was found to be associated with increased CCR1 expression.
B. CCR2
The high-affinity ligand for CCR2-CCL2 had been described several years earlier as a potent monocyte chemoattractant that was purified from the culture supernatant of a glioma cell line (Yoshimura et al., 1989a) and cloned from a HL-60 cDNA library (Furutani et al., 1989). In 1994, CCR2 was cloned from the Monomac 6 monocytic cell line using a PCR-based strategy with degenerate primers based on conserved regions of other chemokine receptors (Charo et al., 1994). This identified a novel PCR product that was then used to screen a cDNA library and identify two clones with predicted transmembrane segments. These two clones differed only in the C terminus and 3′-UTR of the receptor and were designated as splice variants of the CCR2 receptor and named CCR2-A and -B. cRNA of both variants was microinjected into Xenopus oocytes and generated a robust calcium flux in response to CCL2. The murine CCR2 receptor was initially identified by screening cell lines for a response to the murine CCL2 homolog [also known as JE (Boring et al., 1996)]. WEHI 274.1 cells were found to respond to CCL2 by calcium flux assay; a cDNA library was constructed and a full length CCR2 receptor was cloned.
CCR2 is expressed by multiple cell types including monocytes, dendritic cells (DCs), and endothelial cells (Charo et al., 1994; Sozzani et al., 1997; Weber et al., 1999). Indeed, CCR2 is now known to be differentially expressed on the known subsets of both mouse and human monocytes (Geissmann et al., 2003). Aside from CCL2, CCR2 has several other high-affinity ligands including MCP-2 (CCL7), MCP-3 (CCL8), MCP-4 (CCL13), and MCP-5 (CCL12)—a murine chemokine with close homology to human CCL2 (Combadiere et al., 1995; Berkhout et al., 1997; Gong et al., 1997b; Sarafi et al., 1997). However, CCL7, CCL8, and CCL13 all bind other chemokine receptors, whereas CCL2 and CCL12 signal exclusively through CCR2. The rank order of potency of CCR2 ligands at the human receptor is reported to be CCL2 > >CCL13 = CCL8 > CCL7.
Numerous studies have sought to determine the intracellular signaling induced by CCL2 acting on CCR2. In stably transfected HEK-293 cells, CCR2 activation induced calcium release from intracellular stores and inhibited adenylate cyclase with high potency (90 pM), and both responses were blocked by PTX, suggesting coupling to Gαi (Myers et al., 1995). CCL2 was also shown to induce ERK 1/2 (p42/p44 MAPK) activation via a PTX-sensitive mechanism in a T-cell hybridoma (Dubois et al., 1996). Following receptor activation in transfected HEK-293 cells it was shown that the receptor is rapidly phosphorylated (within 1 minute) and internalized (Franci et al., 1996). When the receptor was coexpressed in Xenopus oocytes with various members of the β-adrenergic receptor kinase (βark) family, expression of the βark2 isoform completely blocked CCR2 activation, suggesting that phosphorylation by this kinase induces receptor deactivation, a mechanism now known to be common to many GPCRs. By using site-directed mutagenesis, several serine and threonine residues in the C terminus of the receptor were shown to be essential to receptor internalization and deactivation. A crucial discovery came when it was shown that Gβγ heterodimers released after activation of CCR2 were essential for chemotaxis of a transfected lymphocyte cell line [300.19 (Arai et al., 1997)]. This was achieved by coexpression of the Gαtransducin (Gαt) protein that binds free Gβγ heterodimers with high affinity, preventing their interaction with downstream substrates. In contrast, Gαt expression did not affect ERK phosphorylation in response to CCL2 and only had a small effect on calcium flux, suggesting that this is not mediated via the Gβγ subunit.
CCR2-deficient mice were generated in 1997 and found to be developmentally normal (Boring et al., 1997). The absence of CCR2 impaired recruitment of leukocytes (mainly macrophages) into the peritoneal cavity of mice injected with the inflammatory stimulus thioglycollate. To determine the role of CCR2 in immunity, CCR2 knockout mice were injected with beads coated with purified protein derivative (PPD) antigen from Mycobacterium bovis, which induces granuloma formation in the lung and is associated with a Th1 cytokine profile in the draining lymph node. Ccr2 knockout mice had significantly smaller granulomas containing fewer macrophages in response to PPD challenge. Associated with this, Ccr2-deficient cells isolated from draining lymph nodes produced less IFN-γ and virtually no detectable IL-12 in response to PPD. In other experiments, splenocytes from Ccr2 knockout mice were stimulated with concanavalin A (con A) and found to produce significantly less IFN-γ than wild-type cells, suggesting defective antigen-induced cytokine responses in these animals. Ccr2-deficient mice were generated in a second laboratory, and these animals failed to clear an infection by the intracellular bacteria Listeria monocytogenes, suggesting a direct role for CCR2 in responses to bacterial pathogens (Kurihara et al., 1997).
CCR2 also has a role in myeloid progenitor cell cycling in mice. Mice deficient for Ccr2 were found to have a dramatic increase in the number of myeloid progenitor cells (MPCs) in S phase, indicating increased proliferation of these cells (Reid et al., 1999). However, the absolute number of MPCs in bone marrow or spleen was unaffected in knockout mice. This discrepancy was shown to be due to increased apoptosis of c-kit+ lin− immature cells in the bone marrow, suggesting an important role for this receptor in MPC survival.
CCR2 has been shown to have an essential role in the induction of experimental autoimmune encephalomyelitis (EAE), a mouse model for multiple sclerosis (Fife et al., 2000). The absence of CCR2 completely prevented the development of EAE pathology, and knockout mice showed reduced infiltration of both monocytes and T cells into the central nervous system. T cells from these mice showed similar antigen-induced cytokine production to wild-type controls, and adoptive transfer of wild-type or Ccr2−/− T cells into knockout recipients failed to induce the onset of EAE, indicating that CCR2 expression on host cells is required for disease induction.
CCR2 also has an important role in the development of inflammatory and neuropathic pain (Abbadie et al., 2003). Knockout mice showed a 70% reduction in pain behavior in an inflammatory pain model (intraplantar formalin injection). Intraplantar administration of the ligand CCL2 into wild-type mice also induced a greater degree of mechanical allodynia, a pain response to a stimulus that is not usually painful. By using a model of chronic pain, sciatic nerve injury, Ccr2 knockout mice did not develop the mechanical allodynia seen in wild-type animals. This was associated with reduced infiltration of monocytes/macrophages into the sciatic nerve and dorsal root ganglion and less activation of microglia in the spinal cord.
In 2007, CCR2 was shown to play a critical role in both bone marrow egress of classic monocytes and in monocyte trafficking from the blood to sites of inflammation (Tsou et al., 2007). The key murine ligands required for this process were identified as CCL2 and CCL7.
CCR2 is also important for survival in a murine model of West Nile Virus (Lim et al., 2011). West Nile Virus induces a large degree of monocytosis, and these cells are recruited into the brain, where they limit viral replication. In Ccr2−/− animals, this monocytosis is abolished and infected animals show reduced survival associated with diminished accumulation of classic monocytes in the brain. This is not due to defective recruitment to the brain because transfer of WT and Ccr2−/− monocytes leads to similar accumulation in the brain of infected animals. The authors did not specifically analyze whether failed monocytosis in Ccr2-deficient animals was due to reduced bone marrow egress (as described above) or altered survival in the periphery or possibly defective mobilization from a splenic reservoir, a process that does not require CCR2 after myocardial infarction but that has not been extensively tested in other models (Swirski et al., 2009).
Taken together, these data suggest that CCR2 has a key function in cellular homeostasis, particularly for monocytes, while also playing a critical role in the development of inflammatory responses in response to a wide range of insults.
A nonsynonymous single nucleotide polymorphism (V64I) in the first transmembrane domain of CCR2 has been identified and found to be associated with a faster progression to acquired immunodeficiency syndrome (AIDS) in human immunodeficiency virus (HIV)-positive individuals, although it has no relationship with incidence of HIV infection (Smith et al., 1997). This mutation has not been studied extensively in inflammatory disease, but a few publications have assessed the association of this SNP with pathology. No association was found with the incidence or severity of RA (Bayley et al., 2003) or with systemic lupus erythematosus [SLE (Aguilar et al., 2003)], and similarly the mutation did not correlate with the incidence of lupus nephritis, although the V allele was associated with a less severe disease phenotype as measured by the SLEDAI (SLE disease activity) index (Malafronte et al., 2010). Interestingly, carriers of the I allele with psoriasis were found to be more likely to progress to psoriatic arthritis once the disease was established, but CCR2 genotype did not correlate with the incidence of psoriasis per se (Soto-Sanchez et al., 2010). Finally, the I allele was associated with abdominal aortic aneurysm (Katrancioglu et al., 2011). In summary, these data suggest that the CCR2 V64I mutation may lead to faster disease progression once pathology is established but does not seem to be a strong risk factor for the incidence of inflammatory disease.
C. CCR3
CCR3 was first described by two independent groups (Daugherty et al., 1996; Ponath et al., 1996) and initially referred to as CC CKR3. The receptor, a protein of 355 amino acids in length was isolated and cloned from human eosinophils on the basis of similarity to other chemokine receptors and found to share 63% sequence homology with CCR1 and 51% with CCR2. In contrast to other members, CCR3 was shown to lack sites for N-linked glycosylation (Ponath et al., 1996). The N-terminal region of CCR3 was also shown to contain a leucine residue instead of proline in the proline-cysteine motif conserved in the N terminus of other chemokine receptors.
CCR3 is expressed upon a variety of immune cells, including eosinophils (Fujisawa et al., 2000; Badewa et al., 2002), Th2 lymphocytes (Gerber et al., 1997), mast cells (Brightling et al., 2005; Uguccioni et al., 1997), and basophils (Daugherty et al., 1996). With respect to eosinophils, basophils, and Th2 lymphocytes, CCR3 expression is restricted to the cell surface (Uguccioni et al., 1997). In stark contrast, CCR3 in mast cells is stored in intracellular granules that are mobilized and recruited to the cell surface following Fc receptor engagement/cross-linking with IgE (Price et al., 2003). The murine homolog of CCR3 was also cloned in 1996 and found to be expressed on eosinophils and mediate chemotaxis toward eotaxin (CCL11) (Gao et al., 1996).
CCR3 binds to a wide range of ligands, which can all lead to the activation of the receptor. Examples include CCL5, CCL2, CCL7, CCL11, CCL13, CCL15, CCL24 (eotaxin-2), and CCL26 (eotaxin-3) (Daugherty et al., 1996). Evidence from competitive binding studies has shown that CCR3 displays maximal affinity for CCL11 with a Kd of 0.1 nM. CCL5 and CCL7 were shown to bind at slightly lower affinities with Kd values between 2.7 and 3 nM (Daugherty et al., 1996). Similar trends were observed in functional assays, including intracellular calcium mobilization, with ED50 in the subnanomolar range for all three ligands aforementioned. Chemotaxis assays performed with both primary eosinophils and cells stably transfected with CCR3 revealed CCL11 to possess the most potent migratory effect followed by CCL5 and CCL7, respectively (Daugherty et al., 1996; Ponath et al., 1996).
A study performed by Kitaura et al. (1999) identified CCL26 as a novel functional ligand for CCR3. L1.2 cells (a murine pre B cell line) stably transfected with CCR3 were shown to generate a calcium flux in response to CCL26 with an EC50 of 3 nM. The authors also reported crossdesensitization between CCL26 and other known CCR3 ligands with the rank order of potency reported as: CCL11, CCL13 > CCL26 > CCL24, CCL5. Furthermore, in chemotaxis assays CCL11 induced a typical bell-shaped dose-dependent response with maximal eosinophil migration observed at 100 nM in contrast to CCL26, which only had an effect on migration at 1 µM. This trend was also observed with basophils, with CCL11 inducing migration between 30 and 300 nM, in contrast to CCL26 where chemotaxis was only observed at 300 nM (Kitaura et al., 1999).
A non-CC chemokine ligand, the HIV Tat protein, has also been shown to bind and activate CCR3 on monocytes and macrophages. Albini et al. (1998) generated a synthetic Tat protein and peptide (CysL24-51) that contained the “chemokine-like” region of Tat. Both protein and peptide were shown to induce PTX-sensitive calcium fluxes in monocytes. Receptor desensitization and direct/displacement binding assays were performed with CCR2 and CCR3 transfected CHO-K1 cells. The authors reported both Tat and the CysL24-51 peptide could partially desensitize the response to CCL2, CCL7 and completely block the response to CCL11 in human monocytes. Tat was shown to specifically displace CCL7 from membranes of CCR3 transfected cells. The peptide CysL24-51 was also shown to specifically bind both CCR2 and CCR3 transfected CHO-K1 cells with Kd of 66.4 ± 8.4 nM for CCR3 binding. The results from this study suggest that the HIV Tat protein can mimic the effects of β-chemokines, which may have a role in the recruitment of monocytes/macrophages to sites of HIV producing cells, enhancing the spread of infection within the host.
CCL11 binding to CCR3 has been shown to activate signaling via the mitogen activated protein kinase (MAPK) signaling pathway in eosinophils. Treatment of human eosinophils with CCL11 (10–100 nM) across a time course was shown to induce a rapid phosphorylation and activation of ERK1/2. Pretreatment of eosinophils with a MEK inhibitor (PD980549) reduced kinase activity following stimulation with 10 nM CCL11, highlighting that the phosphorylation and activation of ERK 1/2 occur via a MEK-dependent signaling pathway. Furthermore pretreatment with the MEK inhibitor resulted in a substantial reduction of both CCL11-induced eosinophil rolling in the mouse mesenteric circulation and on CCL11-induced chemotaxis of eosinophils (Boehme et al., 1999). Similar observations were made by another independent group who reported dose-dependent activation of ERK2 and p38 following stimulation with CCL11. In the presence of specific inhibitors against both ERK2 and p38, CCL11-induced eosinophil cationic protein release and chemotaxis were inhibited (Kampen et al., 2000). Collectively these studies highlight the role of MAPK activation in regulating CCL11-induced eosinophil rolling, degranulation and migration via CCR3.
Given that CCR3 is highly expressed upon eosinophils and basophils it is not surprising that the receptor has been highlighted as a leading therapeutic target in diseases that have a strong allergic inflammatory component. The generation of mice lacking CCR3 was first described in 2002 (Humbles et al., 2002). Ccr3−/− mice display normal development with no impairment during and after gestation (Humbles et al., 2002). Under basal conditions Ccr3−/− mice display impaired eosinophil trafficking to the intestinal mucosa with reduced cell numbers observed compared with littermate controls, highlighting the role of the receptor in regulating eosinophil trafficking under homeostatic conditions. In a model of airway hyperresponsiveness, Ccr3−/− mice showed defective eosinophil recruitment, with the majority of cells constricted within the subendothelial space unable to migrate out into the lung parenchyma. Interestingly the authors also reported accumulation of mast cells in the airways of Ccr3−/− mice, unveiling a novel role for the receptor in mast cell homing. Similar observations have been reported by several other groups of defective eosinophil recruitment in ovalbumin (OVA)-induced skin inflammation (Ma et al., 2002) and an OVA-induced experimental asthma model (Pope et al., 2005). More recently, an in vivo study examined the role of epithelial CCR3 in an LPS-induced model of lung inflammation (Li et al., 2011). A specific CCR3 inhibitor [SB-328437, methyl (2S)-2-(naphthalene-1-carbonylamino)-3-(4-nitrophenyl)propanoate] at 5 mg/kg was given via intratracheal instillation to mice with LPS-induced acute lung injury. The inhibitor was shown to reduce neutrophil recruitment into the alveolar space and attenuate IL-8 production in bronchoalveolar lavage (BAL) fluid of these mice when compared with LPS treatment alone. Improved histologic scores were also observed in mice treated with SB-328437 inhibitor. Collectively these findings revealed for the first time a potential role for epithelial expressed CCR3 in promoting LPS-induced lung inflammation through mediating the release of IL-8 (Li et al., 2011).
Interestingly, a nonsynonymous SNP of CCR3 (L324P) in the C terminus of the receptor was found to ablate transport of the receptor to the cell surface when expressed in transfected cells (Wise et al., 2010). The receptor was found to be expressed intracellularly, and these cells were unable to migrate toward CCL11. However, this polymorphism is found at very low frequency in the population, and thus the significance of this finding remains unclear.
D. CCR4
CCR4 was cloned from a basophilic cell line in 1995 using an RT-PCR strategy with degenerate primers based on conserved regions in the CXCR2 and CCR1 receptors (Power et al., 1995). The murine homolog was subsequently cloned the following year (Hoogewerf et al., 1996). The human receptor mRNA was highly expressed in thymus and peripheral blood leukocytes and at a low level in spleen (Power et al., 1995). CCR4 cRNA was transiently expressed in Xenopus oocytes, and various chemokines were assayed for their ability to induce calcium flux and opening of a voltage-gated calcium channel using a patch clamp assay. CCL3, CCL2, and CCL5 were found to induce a signal in CCR4-transfected cells but very high doses were used (1 μM).
Subsequently, TARC (CCL17) and CCL22 have been identified as the high-affinity ligands of CCR4, which were not displaced by other CC chemokines in a competitive binding assay (Imai et al., 1997, 1998). Imai et al. (1997) confirmed that CCR4 was expressed in T cells (mainly CD4+), but was undetectable in any other leukocytes. CCR4 was subsequently shown to be expressed by Th2-polarized lymphocytes (Bonecchi et al., 1998a).
Although intracellular signaling pathways downstream of CCR4 have not been as extensively studied as for other chemokine receptors, two papers have explored CCR4 signaling in a T-cell line and in vitro differentiated Th2 cells. In the human leukemic cell line CEM, CCL22 was found to induce a rapid and transient accumulation of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) as well as a slower phosphorylation of protein kinase B (PKB/Akt) (Cronshaw et al., 2004). This effect was blocked by PI3K inhibitors, although these had no effect on chemotaxis in response to CCL22, indicating that PI3K activation is not required for CCR4-dependent migration of these cells. In contrast, a ROCK (Rho-associated coiled coil-forming protein kinase) inhibitor did block CCL22-induced chemotaxis in CEM cells. The induction of PKB phosphorylation by CCL17 and CCL22 was replicated in Th2 cells differentiated in vitro from primary human peripheral blood leukocytes (PBLs), as was the inability of PI3K inhibitors to block CCR4-mediated chemotaxis.
In a subsequent study using CEM cells, both CCL17 and CCL22 were found to induce calcium flux in a PLC-dependent manner (Cronshaw et al., 2006). Calcium release was found to be dependent on inositol trisphosphate (IP3)-sensitive store operation. However, CCL22-induced chemotaxis required PLC activation but not IP3-mediated calcium release. An inhibitor against the novel protein kinase C (PKC) isoforms that are activated by diacylglycerol (DAG) blocked chemotaxis induced by CCL22. Furthermore, CCL22 was found to induce phosphorylation of PKC δ. Inhibitors of PKC and PLC were also confirmed to block Th2 cell migration toward CCR4 ligands.
Ccr4−/− mice were developmentally normal, and splenocytes from these mice were shown to be unresponsive to the high-affinity ligands CCL17 and CCL22 and also to CCL3 (Chvatchko et al., 2000). Thus, although CCL3 is thought to be a low affinity ligand at CCR4, loss of the CCR4 receptor abrogates the functional response of splenocytes to CCL3. In vivo, deletion of CCR4 was found to have no effect on generation of a Th2 immune response in an OVA-induced airway inflammation model (Chvatchko et al., 2000). However, loss of CCR4 was found to protect mice in a lethal endotoxemia model, with a concomitant decrease in peritoneal macrophage numbers and plasma TNFα and IL-1β levels in response to LPS (Chvatchko et al., 2000). This initial observation was confirmed and extended by Ness et al. (2006) who analyzed the responses of CCR4-deficient mice backcrossed for 10 generations to C57BL/6J challenged with a range of bacterial Toll-like receptor (TLR) ligands and defense against intraperitoneal infection with Escherichia coli. The authors showed that decreased numbers of viable bacteria in Ccr4−/− animals were associated with enhanced recruitment of CD11b+ neutrophils and monocytes into the peritoneal cavity at early time points following inoculation (Ness et al., 2006). Intriguingly, detailed characterization of macrophages from Ccr4−/− animals showed an alteration in gene expression pattern skewed toward that seen in alternatively activated (M2) macrophages.
Traeger et al. (2008) further confirmed the importance of CCR4 in host defense against bacterial infection by studying the effect of CCR4 gene deletion in a model of polymicrobial sepsis following abdominal surgery. Ccr4−/− mice showed decreased mortality and greatly enhanced clearance of bacteria from major organs (lung, liver, kidney, spleen) but no change in viable bacteria numbers in the peritoneal cavity or blood (Traeger et al., 2008). The authors speculated that CCR4 deficiency protects mice from lethal sequelae of systemic bacterial infection by reducing the recruitment of anti-inflammatory regulatory T (Treg) cells into sites of bacterial infection.
Recently Guabiraba and colleagues (2010) compared the effects of dengue virus infection in three different strains of chemokine receptor deficient mice (Ccr1−/−, Ccr2−/−, and Ccr4−/− animals). The mouse model of dengue used shares many features of severe dengue infection in humans, including liver damage, thrombocytopenia, cytokine storm, systemic inflammation, and death. Dengue infection of Ccr4−/− animals was characterized by reduced lethality and reduced tissue damage despite a similar viral load to that observed in wild-type and CCR1- and CCR2-deficient animals (Guabiraba et al., 2011). Taken together with the decreased lethality observed in Ccr4−/− mice in endotoxemia and sepsis models, a picture emerges of CCR4 as an important chemokine receptor in the systemic response to overwhelming innate immune activation.
The role of CCR4 in allergic inflammation has been the subject of extensive investigation. This stems from original observations that expression of CCL17 and CCL22 is potently upregulated by Th2 cytokines and the preferential expression of CCR4 on Th2 CD4+ T lymphocytes (Bonecchi et al., 1998b; Imai et al., 1999; Greaves et al., 2001). The association between CCR4 and allergic inflammation was strengthened by observations made with clinical material. For instance, Panina-Bordignon et al. (2001) reported upregulation of CCL17 and CCL22 expression in airway epithelial cells of lung biopsy samples from atopic asthmatic patients immediately following allergen challenge and Nakatani et al. (2001) reported elevated numbers of CCR4+ T cells in chronic skin lesions of patients with chronic dermatitis. These studies suggesting the involvement of CCR4 in human atopic allergy were reinforced by studies undertaken using animal models of allergic inflammation. By using a murine T-cell transfer model of asthmatic airway disease (AAD) Lloyd et al. (2000) demonstrated that T cells polarized in vitro to Th2 differentiation, but not cells polarized to Th1, could induce airway hyper-responsiveness (AHR) following allergen inhalation in sensitized recipient mice. Transferred Th2 lymphocytes preferentially expressed the CCR3 and CCR4 chemokine receptors in vitro and in the lung after transfer. By using validated anti-CCL11 and anti-CCL22 polyclonal antibodies, the authors showed decreased recruitment of Th2 donor lymphocytes in the early stages of lung disease by blocking the activity of the CCR3 ligand CCL11, whereas antibodies against the CCR4 ligand CCL22 blocked the later stages of AAD. Lung inflammation has been studied in Ccr4−/− mice using a range of different experimental models. In the original report describing the generation of Ccr4−/− mice Chvatchko et al. (2000) showed no difference between wild-type and CCR4-deficient mice in the development of AHR following OVA sensitization. However, a subsequent report by Schuh et al. (2002b) using the same Ccr4−/− mice in a different model of lung inflammation that used sensitization and subsequent challenge with Aspergillus conidia showed reduced eosinophil recruitment and reduced AHR in CCR4−/− mice compared with wild-type animals.
Various explanations for the conflicting results in different animal models of allergic inflammation have been proposed, but optimism about targeting of CCR4 for therapeutic benefit in allergic asthma has come from more recent studies in a humanized model of asthma. NOD/SCID (nonobese diabetic/severe combined immunodeficient) mice reconstituted with peripheral blood mononuclear cells (PBMCs) from normal and atopic human volunteers underwent bronchial challenge with house dust mite antigen. Allergic inflammation, goblet cell hyperplasia, and AHR were only seen in mice reconstituted with PBMCs from atopic donors, and this inflammatory response was abolished by administration of a CCR4 blocking monoclonal antibody (Perros et al., 2009). Further validation of CCR4 as a useful therapeutic target in human asthma has come from Vijayanand et al. (2010) who showed that CCR4+ but not CCR4- T cells from asthmatics are the major source of Th2 cytokines, including CCL17. By using explanted bronchial tissue from asthmatic donors the authors went on to show that a small molecule CCR4 antagonist markedly reduced Th2 cell chemoattractant activity in culture medium of such explants (Vijayanand et al., 2010).
Most of the foregoing discussion of CCR4 has focused on the proinflammatory role of CCR4+ cell recruitment in response to upregulation of CCL17 and CCL22 expression. However, in common with several other CC chemokine receptors, CCR4 also plays a homeostatic role in leukocyte trafficking in addition to its well documented effects in innate immune cell activation and Th2 immunopathology. In 1999, Campbell et al. (1999) showed that many CCR4+ memory T cells in blood coexpressed CCR4 and cutaneous lymphocyte antigen. The authors went on to demonstrate that these cells underwent CCL17-triggered tight adhesion to intercellular adhesion molecule 1 (ICAM-1) under conditions of physiologic flow (Campbell et al., 1999). Later experiments confirmed the importance of the CCR4 receptor rather than the CCR10 receptor for skin homing and recruitment of T cells to sites of skin inflammation (Soler et al., 2003). Campbell et al. (2007) elegantly demonstrated the importance of CCR4 for the development of appropriate skin-homing T lymphocyte populations in competitive repopulation experiments using equal numbers of Ccr4−/− and Ccr4+/+ cells in Rag−/− (recombinase activating gene) mice and in experiments using T-cell receptor (TCR) transgenic models (Baekkevold et al., 2005).
A deeper appreciation of the central role of CCR4 in the biology of skin-homing T lymphocyte biology should not prevent the exploration and hopefully the exploitation of CCR4 as a therapeutic target in diseases such as asthma where there is clear evidence linking CCL17 and CCL22 as major players in disease pathology.
One study has examined the potential association of a CCR4 SNP (C1014T) with atopic dermatitis (Tsunemi et al., 2004). There was no association between CCR4 genotype and disease incidence, blood IgE levels, or eosinophil counts.
E. CCR5
The CCR5 receptor was described by three independent groups at around the same time (Combadiere et al., 1996; Raport et al., 1996; Samson et al., 1996a). The receptor was found to be expressed in monocytes and macrophages, the THP-1 monocyte cell line, CD4+ and CD8+ T cells, but not in eosinophils or neutrophils (Combadiere et al., 1996; Raport et al., 1996). In HEK-293 cells transfected with CCR5, several chemokines were found to induce calcium flux with the rank order of potency CCL3 > CCL5 > CCL4 (Combadiere et al., 1996). In CCR5 transfected COS-7 cells, inositol phosphate production was elicited by low nanomolar concentrations of chemokine and was found to have a different rank order: CCL5 > CCL4 > CCL3 (Raport et al., 1996). By using competitive radioligand binding assays in transfected HEK-293, all three ligands were found to have relatively low affinity for CCR5 (IC50 ∼100 nM); instead both CCL5 and CCL3 were found to bind with higher affinity at CCR1 (Combadiere et al., 1996). In CCR5 transfected COS-7 cells, all three chemokines bound with higher affinity, with competition binding IC50 values all around 7 nM (Raport et al., 1996). The different values reported by different groups are likely to represent different levels of expression and/or receptor coupling in the various transfected cell lines used. The receptor was mapped to chromosomal location 3p21, very close to the CCR2 gene (Raport et al., 1996). The murine homolog of CCR5 was described a few months prior to the identification of the human receptor and was also found to bind CCL3, CCL4, and CCL5 (Boring et al., 1996).
Subsequent studies have shown that CCR5 can also bind other CC chemokines, notably CCL8 and CCL13, although both are less potent than CCL5 in inducing migration of transfected cells—doses of 100 nM CCL13 are required to elicit migration (Ruffing et al., 1998). These data were confirmed by another group who also showed that CCR5 could bind the chemokines CCL7 and HCC-1 (CCL14), although the biologic activity of the chemokines was not assessed (Napier et al., 2005). CCL11 also has some biologic activity at CCR5, although very high concentrations (1 μM) were needed to induce a response (Ogilvie et al., 2001).
Just before the identification of CCR5, its ligands CCL5, CCL3, and CCL4 were found to be produced by cytotoxic CD8+ T cells and were capable of suppressing HIV infection (Cocchi et al., 1995). Subsequently, CCR5 was found to be the major coreceptor for entry of macrophage-tropic strains of the virus (Alkhatib et al., 1996; Deng et al., 1996) . Indeed, individuals homozygous for a 32-bp deletion in the coding region of CCR5 (CCR5 Δ32) are resistant to HIV infection (Samson et al., 1996b).
The regions of the CCR5 receptor important for ligand binding and signaling have been extensively studied. Samson et al. (1997) generated chimeric receptors with different combinations of transmembrane and extracellular domains of the CCR2 and CCR5 receptors. By using competitive binding assays and a microphysiometer to assess biologic responses (via a change in extracellular pH following signaling), the second extracellular loop of the receptor was found to be the key region determining ligand specificity. However, using a similar approach with chimeric receptors, it was the N-terminal domain and the first extracellular loop of CCR5 that were found to be important in infection with HIV (Rucker et al., 1996). Furthermore, it was shown that binding of the HIV envelope protein to CCR5 induces a signaling response that is not required for entry in vivo but instead may induce chemotaxis of T cells or enhance viral replication (Weissman et al., 1997).
CCR5, like all chemokine receptors, is Gαi coupled, leading to a reduction in intracellular cAMP levels (Aramori et al., 1997). Following activation, the receptor was found to be desensitized via phosphorylation by GRKs 2, 3, 5, and 6, leading to β-arrestin recruitment and receptor internalization (Aramori et al., 1997). Binding of CCL4 to CCR5 was also shown to induce activation of RAFTK/Pyk2 (a member of the focal adhesion kinase family), leading to activation of the Jun N-terminal kinase (JNK) and p38 MAPK pathways and the cytoskeletal protein paxillin (Ganju et al., 1998). This provided the first evidence of how chemokine receptors might link ligand binding to cytoskeletal rearrangement and migration. CCL4 was subsequently also shown to activate Syk via RAFTK and recruit the phosphatases SHP1 (Src homology region 2 domain-containing phosphatase-1) and SHP2 associated with the adapter protein Grb2 (Ganju et al., 2000).
The generation and phenotyping of a mouse lacking CCR5 was first described in 1998 (Zhou et al., 1998). Macrophages isolated from Ccr5−/− mice generated reduced amounts of inflammatory cytokines (including IL-6, IL-1β ,and TNFα) when classically activated with LPS and IFN-γ ex vivo. By using a Listeria monocytogenes infection model, Ccr5−/− mice showed defective macrophage-dependent clearance, with higher titers of bacteria found in the liver. CCR5 deficiency also offered protection in a lethal endotoxemia model in which CCR5 ligands are known to play an important role in macrophage recruitment, contributing to pathology. In wild-type mice, expression of CCR5 in T cells was shown to be dramatically upregulated following in vitro activation with anti-CD3 and anti-CD28 antibodies. Activated Ccr5−/− T cells produced more IFN-γ, GM-CSF (granulocyte macrophage colony stimulating factor) and IL-4 compared with wild-type cells. This enhanced cytokine production was mirrored in an enhanced delayed-type hypersensitivity (DTH) reaction and increased antibody production following antigen challenge in CCR5 deficient mice.
Ccr5−/− mice were also found to have drastically reduced survival in a model of intratracheal delivery of Cryptococcus neoformans, an AIDS-associated pathogen (Huffnagle et al., 1999). Knockout mice showed normal recruitment of leukocytes to the lung but defective accumulation of leukocytes in the brain, associated with increased accumulation of capsule polysaccharide and brain edema. Thus, CCR5 appears to play a critical role in the generation of both innate and adaptive immune responses to control a variety of pathogens.
Other studies have shown that CCR5 deficiency has no effect in an EAE model or in antiviral immunity but protects mice from a dextran sodium sulfate (DSS)-induced colitis model and from chronic fungal asthma (Andres et al., 2000; Tran et al., 2000; Nansen et al., 2002; Schuh et al., 2002a).
Several studies have analyzed the importance of CCR5 expression Treg cells. In murine models of graft versus host disease (GVHD), Tregs are required to inhibit rejection but suppress effector responses. Wysocki et al. (2005) demonstrated a requirement for CCR5 in this process since irradiated mice given transplants supplemented with CCR5-deficient Tregs showed reduced survival compared with mice receiving WT Tregs. Ccr5−/− Tregs showed a normal initial accumulation in lymphoid tissues, but a later defect in recruitment to both lymphoid tissues and GVHD target organs. A second study by Moreira et al. (2008) used the fungal pathogen Paracoccidiodes brasiliensis, which induces the formation of granulomas containing viable yeast particles, allowing disease reactivation. This process is driven by recruitment of Treg cells to the infected side, leading to inhibition of effector responses and fungal persistence. CCR5 was shown to have a critical role in the process because CCR5 knockout Tregs showed a reduced accumulation in granulomas, leading to better control of fungal growth and dissemination. Finally, CCR5 expression on Tregs has been shown to be important for evasion of the immune response by tumors. By using a CCR5 antagonist, Tan et al. (2009) demonstrated that CCR5 blockade reduced Treg migration to tumors, thus leading to diminished tumor growth in a pancreatic cancer model.
Thus CCR5 expressed on leukocytes has disparate roles that both facilitate effective immune responses against pathogens, but may also promote immune tolerance. This may be beneficial in certain scenarios, such as transplantation, but may allow tumors to evade the host immune system.
In terms of epidemiologic studies, the CCR5 Δ32 polymorphism mentioned above has also been extensively studied for potential association with inflammatory disease, in particular RA. A meta-analysis of five case-control studies with European participants demonstrated a mean allele frequency of 6% for Δ32 in RA patients and 10% in controls (Prahalad, 2006). A significant negative association of the Δ32 allele with RA was identified, indicating that this polymorphism is protective. Similarly, a case-control study from the UK replicated the negative association of the Δ32 allele with juvenile idiopathic arthritis [JIA (Hinks et al., 2010)]. These authors also performed a meta-analysis of three studies including >2000 patients and confirmed the protective effect of the polymorphism. One of the studies included in this meta-analysis, however, failed to independently show an association of Δ32 with JIA (Lindner et al., 2007). Another recent small study showed no correlation of the Δ32 polymorphism with either the incidence of RA or SLE or disease severity in these patients (Martens et al., 2010). These studies suggest that the influence of CCR5 polymorphism on arthritis is relatively subtle, and large-scale studies are needed to demonstrate an involvement of this receptor with disease. The recent failure of the licensed CCR5 antagonist maraviroc (Selzentry; Pfizer, Sandwich, UK) to reduce arthritis severity beyond established therapies (see IV. Chemokine Receptor Drugs in Clinical Trials), also suggests that CCR5 may not be a viable therapeutic target in RA.
F. CCR6
Several groups independently identified an open reading frame located on chromosome 6q27 as a potential GPCR with homology to known chemoattractant receptors (Zaballos et al., 1996; Baba et al., 1997; Greaves et al., 1997; Liao et al., 1997a, b; Power et al., 1997). The formal demonstration that these cDNA clones encoded the CCR6 chemokine receptor only became possible with the availability of recombinant protein for the CCL20 chemokine, variously known as MIP-3α, Exodus, and LARC (Hieshima et al., 1997; Hromas et al., 1997; Rossi et al., 1997). CCR6 transfected cells were shown to generate a calcium flux only when exposed to recombinant CCL20 protein and labeled CCL20 bound to CCR6 transfected cells with high affinity (∼0.1 nM Kd) (Greaves et al., 1997). CCL20 remains the sole high-affinity chemokine ligand of the CCR6 receptor but low affinity binding of human beta defensin-1 and -2 to the CCR6 receptor has been reported (Hoover et al., 2002; Yang et al., 1999a).
Initial mRNA expression data across panels of human tissues and primary cells suggested that CCR6 was expressed predominantly by DCs and T cells. This observation was confirmed and extended using CCR6-specific antibodies that showed surface expression of CCR6 on immature DCs that underwent dose-dependent chemotaxis in response to CCL20 (Carramolino et al., 1999; Dieu-Nosjean et al., 2000). Maturation of DCs derived from CD34+ cord blood progenitors in vitro following treatment with IL-4 or TNFα led to a marked decrease in CCR6 expression and a corresponding decrease in DC responsiveness to CCL20 (Carramolino et al., 1999). Immature DCs generated by culturing human monocytes with GM-CSF, IL-4 and TGF-β1 (transforming growth factor β1) showed good levels of CCR6 expression and chemotaxis to CCL20 (Yang et al., 1999b). Further evidence that CCL20 is a potent chemoattractant of immature DCs came from experiments using CD1a+ Langerhans cell precursors generated from human skin (Dieu-Nosjean et al., 2000). In keeping with the initial reports of CCR6 mRNA expression patterns, use of CCR6-specific antibodies in flow cytometry showed that resting memory T cells in human blood both expressed CCR6 and were responsive to CCL20 (Liao et al., 1999). More recently CCR6 was shown to be selectively expressed by Th17 CD4+ T cells, a subset of effector cells associated with protection against extracellular microbes. Moreover, these cells showed functional responses to CCL20 in vitro (Annunziato et al., 2007; Singh et al., 2008). CCR6 expression on human B lymphocytes and activated neutrophils has been reported, but CCR6 expression is not always associated with chemoattractant effects of CCL20 in these cell types (Yamashiro et al., 2000; Liao et al., 2002).
Relatively few papers have addressed the signaling pathways downstream of the CCR6 receptor in primary cells. Keates et al. (2007) demonstrated CCR6 expression by the epithelial cell lines Caco-2 and HT-29 and then demonstrated that CCL20 addition to these cells leads to epidermal growth factor receptor (EGFR) transactivation via shedding of the cell-associated EGFR ligand amphiregulin by metalloproteinase activation. More recently Lin et al. (2010) undertook a detailed proteomics profiling of proteins mobilized to lipid rafts in CCR6 transfected Jurkat cells soon after addition of CCL20. The authors identified 85 proteins that are rapidly mobilized to lipid rafts containing CCR6 and siRNA knockdown of one of these proteins that is associated with actin cytoskeletal rearrangement (L-Plastin) decreased CCL20-mediated chemotaxis but did not affect calcium mobilization in response to CCL20. The authors additionally demonstrated a role for the chaperone protein heat shock protein 90 (HSP90) in CCR6 signaling in transfected Jurkat cells.
Cloning of the murine homolog of CCR6, which shares 74% homology with the human CCR6 receptor, was reported in 1998 along with the demonstration that murine CCL20 caused calcium mobilization in mCCR6 transfected cells (Varona et al., 1998). Soon afterward two independent groups reported the generation of Ccr6−/− mice. Consistent with the reported expression pattern of CCL20 in the intestinal epithelium of mice and humans (Tanaka et al., 1999) and a specific role for CCR6 in intestinal T-cell homeostasis and DC recruitment, Ccr6−/− mice displayed altered T-cell populations in mucosal associated lymphoid tissue (MALT) and impaired IgA responses to orally administered antigens and viral pathogens (Cook et al., 2000). Ccr6−/− mice displayed no defect in humoral immune responses to subcutaneously delivered antigens. Varona et al. (2001) independently generated a second line of Ccr6-deficient mice and confirmed the previous observation of altered T-cell numbers and T-cell subsets within the intestinal mucosa. Additionally, these authors showed that Ccr6−/− mice displayed altered immune responses in DTH and contact hypersensitivity. The involvement of CCR6 expressing T cells in contact hypersensitivity to hapten sensitization and challenge was confirmed in a subsequent study (Paradis et al., 2008). Further evidence for the involvement of CCR6+ cells in intestinal immunity and inflammation came from the demonstration that Ccr6−/− mice showed less severe intestinal pathology in the DSS model of colitis, whereas the absence of CCR6 conferred susceptibility to the trinitrobenzene sulfonic acid-induced (TNBS) model of colitis in C57BL/6J mice, a usually resistant strain (Varona et al., 2003). These results were complemented by later studies showing that a CCL20 blocking antibody reduced intestinal pathology, decreased PMN recruitment, and reduced the number of CCR6+ T cells in the TNBS model of colitis (Katchar et al., 2007).
Ccr6-deficient mouse strains have been used to study the role of CCR6 expression in a range of animal disease models in which DC mobilization and T-cell responses are known to be important in pathogenesis. Lukacs et al. (2001) reported that expression of the CCR6 receptor is required for the allergic immune response that drives allergic pulmonary inflammation in mice sensitized to cockroach antigen. Compared with sensitized wild-type mice, sensitized Ccr6−/− mice displayed reduced airway resistance, fewer eosinophils, and less IL-5 in lung tissue following antigen challenge. Taken together with the initial description of Ccr6−/− mouse IgA responses to rotavirus infection, this study suggests a key role for CCR6-expressing cells in the generation of adaptive immune responses to mucosal antigens. Further investigation of the role of CCR6 expressing cells in a cigarette smoke-induced lung injury model was reported by Bracke et al. (2006), who showed reduced numbers of DCs, activated CD8+ T cells, and neutrophils in Ccr6-deficient mice leading to decreased emphysema in this model of chronic obstructive pulmonary disease (COPD).
Liston et al. (2009) examined the role of CCR6-expressing cells in both the induction and effector phase of the EAE model of demyelination and central nervous system (CNS) inflammation. Ccr6−/− mice had reduced disease scores compared with wild-type mice following immunization with myelin oligodendrocyte glycoprotein (MOG) peptides (Liston et al., 2009). The authors extended this observation by showing that peak EAE disease scores and spinal cord CD4+ T-cell numbers in wild-type mice could be significantly reduced by use of a CCR6 antagonist [a truncated form of the CCR6 ligand, CCL20(6-70)] and a rabbit anti-mouse CCR6 antiserum. In contrast Elhofy et al. (2009) demonstrated more severe CNS inflammation in Ccr6−/− mice compared with wild-type animals using a similar, but not identical, EAE induction protocol. By using adoptive transfer of different cell populations, the authors mapped the increased severity of EAE in Ccr6-deficient animals to an absence of Ccr6+ DCs rather than Ccr6+ T cells. Another explanation for the contrasting results obtained with CCR6-deficient mice in the EAE model of CNS inflammation comes from the work of Yamazaki et al. who demonstrated that CCR6 is expressed on both Th17 effector cells and Treg populations. Taken together these results suggest a complex role for CCR6 mediated T cell and DC recruitment in autoimmune disease.
Genome-wide association studies (GWAS) have identified single nucleotide polymorphisms near the CCR6 gene on chromosome 6q27 as novel risk loci for autoimmune Graves’ disease, generalized vitiligo, and RA (Chu et al., 2011; Jin et al., 2010a; Quan et al., 2010; Stahl et al., 2010). Of note Kochi et al. (2010) identified a dinucleotide polymorphism in the CCR6 gene that was associated with RA susceptibility in two independent cohort studies and was correlated with the level of CCR6 expression and the presence of IL-17 in patient sera. The demonstration that a functional polymorphism within the human CCR6 gene is associated with an increased risk of RA and increased IL-17 cytokine levels in the sera of RA patients has fueled increased interest in Th17 cell biology in a range of autoimmune diseases (Lubberts, 2010).
G. CCR7
CCR7 was first identified in 1993 by Birkenbach and colleagues (1993) and was soon after assigned its nomenclature in a study by Yoshida et al. (1997), who carried out molecular cloning and site-specific mapping of a novel ligand for the receptor that would later be designated CCL19. CCR7 was initially referred to as EBI1 (Epstein Barr virus-induced gene 1) because the gene that encoded the receptor was induced by EBV infection (Birkenbach et al., 1993; Schweickart et al., 1994). The CCR7 gene is located on human chromosome 17q12-q21.2. Expression of CCR7 has been detected on T lymphocytes (Bardi et al., 2001; Kim et al., 2005), B lymphocytes (Birkenbach et al., 1993), DC (Walker et al., 2005), and natural killer (NK) cells (Inngjerdingen et al., 2001). In activated B cells, CCR7 is highly upregulated following EBV infection (Birkenbach et al., 1993); similar observations were made with CD4+ T cells following infection with human herpes virus 6 (HHV6) and HHV7 (Hasegawa et al., 1994). The mouse homolog of CCR7 was cloned in 1994 and found to display 86% homology with the human receptor (Schweickart et al., 1994).
CCL19 (also known as EBI-1-ligand chemokine; ELC, MIP-3β, CK β-11) and CCL21 (also known as secondary lymphoid-tissue chemokine; SLC, TCA-4, 6Ckine) are the primary ligands for CCR7. Studies from the group led by Osamu Yoshie were the first to describe the isolation, cloning, and characterization of both these ligands (Nagira et al., 1997, 1998; Yoshida et al., 1997, 1998). CCL19 was shown to share close homology with other CC chemokines, including CCL3, CCL7, and CCL5. Binding studies utilizing K562 cells stably transfected with a variety of CC chemokine receptors revealed only specific and high-affinity binding of CCL19 to CCR7 (Yoshida et al., 1997). This binding was also shown to induce calcium mobilization with an EC50 of 0.9 nM (Yoshida et al., 1997). Furthermore, 293/EBNA-1 cells stably transfected with CCR7 and HUT78 cells (human T-cell line) expressing endogenous CCR7 were shown to undergo chemotaxis toward CCL19, with maximal effects observed in both cell types at 300 ng/ml (Yoshida et al., 1997).
CCL21, a polypeptide of 134 amino acids in length, shares approximately 21–33% homology with other human CC chemokines. High levels of CCL21 mRNA expression have been reported in human lymph nodes and appendix, with intermediate levels detected in spleen and low levels in the thymus. Similar to CCL19, CCL21 expression is virtually absent on peripheral blood leukocytes (Nagira et al., 1997). However, more recent evidence has shown CCL19 to be expressed by mature DCs and CCL21 upon endothelial cells of afferent lymphatics (Ansel et al., 2000; Baekkevold et al., 2001) and high endothelial venules (Gunn et al., 1998; Stein et al., 2000; Warnock et al., 2000). Nagira et al. (1997) demonstrated the ability of CCL21 to induce PTX-sensitive chemotaxis of T-cell lines Hut78 and Hut102 at doses of 10–100 nM. A similar chemotactic effect was observed with freshly isolated peripheral blood lymphocytes (PBL) with maximal chemotaxis at 0.1 nM. In contrast, chemotaxis was absent when testing CCL21 with neutrophils, monocytes, or monocytic cell lines. Interestingly, CCL21 was not found to induce calcium flux in PBL; however, calcium mobilization was observed with cultured T cells with a reported EC50 of 1 nM (Nagira et al., 1997). These initial observations were confirmed and expanded upon in a subsequent study by Sullivan et al. (1999), whereby the authors used stably transfected CCR7 HEK-293 clones and demonstrated high-affinity (pM) binding of both CCL19 and CCL21. Furthermore, the authors also reported an enhanced chemotactic response to both ligands in the subnanomolar range (0.1–1 nM) in addition to rapid calcium mobilization and MAPK activation (Sullivan et al., 1999).
During the course of DC maturation, CCR7 cell surface expression is upregulated (Dieu et al., 1998; Sallusto et al., 1998). Both CCL19 and CCL21 are constitutively expressed in lymph nodes and play a central role in the trafficking of CCR7 expressing DCs toward these highly organized lymphoid structures where antigen presentation to naive T cells takes place (Ngo et al., 1998). CCR7 has been shown to activate two independent signaling modules in human DCs. The first involves Gαi-dependent MAPK activation of ERK1/2, JNK, and p38 that regulate chemotaxis of DCs. The second involves activation of the GTPase Rho, the tyrosine kinase Pyk2, and the inactivation of cofilin, with all three components shown to have a role in DC migratory speed (Riol-Blanco et al., 2005). CCR7 has also been shown to play an important role in the survival of mature DCs via anti-apoptotic signaling (Sanchez-Sanchez et al., 2004). Stimulating DCs with either CCL21 or CCL19 inhibited apoptosis by 50–60% with the maximal effect observed at 200 ng/ml. When pretreating DCs with a neutralizing antibody against CCR7, the protective effects of CCL19 and CCL21 were abolished. These effects observed with CCL19 and CCL21 were partly attributed to the rapid activation of PI3K and Akt, which are both key regulators of survival in a variety of cell types. NFκB was also shown to be involved in promoting the antiapoptotic effects of both chemokines, because inhibition of NFκB resulted in attenuated effects of CCL19 and CCL21 on DC survival (Sanchez-Sanchez et al., 2004).
The generation and phenotyping of CCR7-deficient mice (Ccr7−/−) was first described by Forster et al. (1999). In the absence of CCR7, elevated levels of CD4+ T cells were found in the peripheral blood, spleen, and bone marrow of knockout mice. In contrast, lower levels of CD4+ T cells were detected in the mesenteric lymph nodes (LN), peripheral LNs, and Peyer's patches. Specific immunostaining of LNs from Ccr7−/− mice revealed an impaired distribution and in some cases reduced numbers of both B and T cells within the outer cortex, paracortex, and marginal sinus areas of LNs when compared with WT mice. Another key observation in the Ccr7−/− mice was that increased numbers of activated B cells were recovered from LNs, whereas migration of activated skin DCs into draining LNs was impaired. One of the most significant observations reported in this study was the impaired ability of Ccr7−/− mice to mount T-cell responses in both models of contact hypersensitivity and DTH with no ear swelling observed following rechallenge with OVA at 24 or 48 hours. Impaired humoral responses were also apparent in the Ccr7−/− mice following application of the T-cell-dependent antigen DNP-KLH (Forster et al., 1999). Other in vivo evidence indicates a critical association between CCR7 and CXCR5 in the development and organization of secondary lymphoid organs. Mice deficient in both receptors (Ccr7−/− Cxcr5−/−) were shown to lack all peripheral LNs (Ohl et al., 2003). Furthermore, CD3-CD4+IL-7Rαhi cells which play a critical role during the early phase of secondary lymphoid organ development were shown to co-express both receptors (Ohl et al., 2003). More recently, CCR7 deficiency was shown to promote Th2 polarization and B cell activation, characterized by elevated IL-4 levels in lymph nodes (Moschovakis et al., 2012). Collectively these studies highlight a dual functionality for CCR7, first in mounting effective primary immune responses and, second, in homeostatic mechanisms that include developing and maintaining the architecture and activity of local microenvironments in secondary lymphoid tissues.
Several SNPs in the CCR7 gene have been identified but are found at very low frequency and were shown to have no association with several diseases including SLE (Kahlmann et al., 2007).
H. CCR8
Molecular cloning of the CCR8 receptor (formerly known as TER1, ChemR1, CY6, CKR-L1) was first described by three independent groups at around the same time (Napolitano et al., 1996; Samson et al., 1996c; Zaballos et al., 1996). The gene that encodes the receptor was mapped onto chromosome 3p21 and shown to share close homology with other known chemokine receptors (Napolitano et al., 1996; Samson et al., 1996c; Zaballos et al., 1996). The human and murine CCR8 genes have been shown to share approximately 71% homology, with both receptors shown to be activated with either murine or human CCL1 (I-308) (Roos et al., 1997; Goya et al., 1998).
CCR8 mRNA expression appears to be restricted to lymphoid tissue and several cell lines. Zaballos et al. (1996) reported mRNA expression of CCR8 in PBMCs, more specifically monocytes/macrophages, CD4+ and CD8+ T cells, and to a lesser extent in CD19+ B cells. CCR8 was also shown to be selectively expressed upon Th2 polarized cells and clones (Zingoni et al., 1998). Human NK cells have been reported to express CCR8 upon activation with IL-2 (Inngjerdingen et al., 2000). However, several other independent studies have been unable to show detectable CCR8 mRNA levels in human PBMC subsets (Napolitano et al., 1996; Roos et al., 1997), therefore data relating to the expression profile of CCR8 in primary human immune cells should be interpreted with caution.
CCL1 is the principal functional ligand for CCR8. CCL1 binds with high affinity to CCR8 (∼1.2 nM Kd) and induces transient intracellular calcium mobilization (EC50 2 nM) (Roos et al., 1997; Tiffany et al., 1997). Cells transfected with the CCR8 receptor have also been shown to migrate toward CCL1 in a dose-dependent manner, with maximal activity observed at 10 nM (Tiffany et al., 1997). Other ligands, including CCL17 and CCL4, have been reported to induce chemotaxis in Jurkat cells stably expressing CCR8 (Bernardini et al., 1998). However, a recent study tested all previously described ligands for CCR8 and could only confirm CCL1 and vMIP-1 (viral macrophage inflammatory protein-1: a viral chemokine mimic from Kaposi’s sarcoma-associated herpes virus) as potent activators of the receptor in chemotaxis and calcium flux assays (Fox et al., 2006).
CCL1 protects murine thymic lymphoma cell lines from dexamethasone-induced apoptosis (Van Snick et al., 1996). This initial observation was confirmed and extended by Spinetti et al. (2003) who demonstrated that both CCL1 and vMIP-1 induce CCR8-mediated rescue from dexamethasone-induced apoptosis via an ERK dependent pathway. The use of a specific antagonist (MC148/MCC-I) of CCR8 was shown to inhibit the rescue effect of CCL1 and vMIP-1, highlighting the involvement of CCR8 in cell survival (Spinetti et al., 2003).
A study performed by Haskell et al. (2006) was the first to describe a novel nonpeptide chemokine receptor agonist for CCR8. The agonist 2-{2-[4-(3-phenoxybenzyl)piperalzin-1]ethoxy}ethanol (ZK 756326) was shown to inhibit binding of CCL1 to U87 cells expressing CCR8 with an IC50 of 1.8 µM. ZK 756326 induced PTX-sensitive calcium mobilization in CCR8-expressing cells in a dose-dependent manner. ZK 756326 was also shown to activate murine cells expressing CCR8, promoting their chemotaxis and inducing ERK1/2 phosphorylation. Akin to CCL1, ZK 756326 at 10 µM was shown to inhibit HIV fusion of cells expressing CD4 and CD8 (Haskell et al., 2006). Collectively these observations highlighted the use of this CCR8 agonist as a potential tool to study CCR8 biology.
CCR8-deficient mice have been instrumental in deciphering the role the receptor plays in the recruitment of specific leukocyte subsets, in regulating immune responses, and as a potential therapeutic target. Phenotypically, Ccr8−/− mice were shown to develop normally, to be fertile, and not to develop spontaneous disease (Chensue et al., 2001). CCR8 deletion was shown to specifically impair local Th2 cytokine responses in models of S. mansoni soluble egg antigen (SEA)-induced granuloma formation and OVA and cockroach antigen-induced allergic airway inflammation. This impaired Th2 response was associated with a significant reduction in eosinophil levels in the lung (50–80%) and Th2 cytokine (IL-5 and IL-13) production in the lungs and draining LNs of Ccr8−/− mice compared with WT mice. Interestingly, Th1 specific immune responses elicited with M. bovis PPD remained unaffected in Ccr8−/− mice, emphasizing the specificity of CCR8 in Th2 functional responses (Chensue et al., 2001). In a follow-up study from the same group, CCR8 expression was shown to be restricted to SEA elicited (specific type 2 response) IL-10+ CD4+ T cells (Freeman et al., 2005). These cells were further characterized and shown to express CD25+ and CD44+ but lacked Foxp3 expression, indicating that they were not Treg cells. Collectively these observations highlight the importance of a defined subset of IL-10-producing CCR8+ CD4+ CD25+ CD44+ T cells in mediating Th2 type responses to parasitic antigens.
The importance of CCR8 in Th2 responses was also highlighted in an atopic dermatitis model in which Ccr8-deficient animals showed less eosinophilic inflammation, and adoptive transfer studies showed CCR8 was critical for homing of Th2 cells to inflamed skin (Islam et al., 2011). This process was shown to depend on murine CCL8, which was highly expressed in inflamed skin.
Other interesting observations made with CCR8-deficient mice include leukocyte trafficking studies that implicate a role for the receptor in the migration of monocyte-derived dendritic cells to draining lymph nodes (Qu et al., 2004). In a model of chronic fungal asthma, the absence of CCR8 receptor was shown to promote the clearance of fungal material in the lungs of these animals, highlighting the receptor as a potential therapeutic target in fungal associated pulmonary conditions (Buckland et al., 2007).
I. CCR9
In 1997, a novel CC chemokine, thymus-expressed chemokine (TECK, CCL25), was identified from analysis of a RAG-1 deficient mouse thymus cDNA library and found to be expressed at the mRNA level in thymus and the small intestine (Vicari et al., 1997). Recombinant murine CCL25 was shown to be chemotactic for murine macrophages, DCs, and thymocytes. Chemotaxis could be blocked by pretreatment with PTX, implicating a Gαi-coupled receptor as mediating the chemoattractant effects of CCL25 (Vicari et al., 1997). In 1999, Zabellos et al. (1999) demonstrated that a previously identified orphan GPCR, GPR-9-6, encoded the CCL25 receptor (designated CCR9) as HEK293 cells transfected with full length versions of the GPR-9-6 open reading frame showed CCL25 mediated calcium flux and dose-dependent chemotaxis to CCL25. Other groups confirmed the assignment of GPR-9-6 as the receptor for CCL25 and further demonstrated that human thymocytes expressing CCR9 displayed chemotaxis to CCL25 (Youn et al., 1999; Norment et al., 2000). Additionally, Yu et al. (2000) demonstrated that alternative splicing of the CCR9 gene generated two different CCR9 mRNAs encoding two different versions of CCR9. CCR9A contains 12 additional amino acids at its N terminus compared with CCR9B and had a lower EC50 for CCL25-mediated calcium flux compared with CCR9B in transfected cells (Yu et al., 2000). CCL25 remains the only identified high-affinity ligand for CCR9. The murine homolog of CCR9 was identified in 2000 (Norment et al., 2000).
Generation of an anti-CCR9 monoclonal antibody allowed a detailed examination of the cells expressing CCR9 in different anatomic sites (Kunkel et al., 2000). Butcher and co-workers demonstrated that CD4+ and CD8+ T cells expressed the highest level of CCR9 in human peripheral blood and that CCR9+ T cells coexpressed high levels of the gut-homing α4β7 integrin. Consistent with this observation, the authors used flow cytometry to demonstrate expression of CCR9 by all intraepithelial T cells and lamina propria lymphocytes in the small intestine (Staton et al., 2006). These findings were confirmed in a complementary study using a CCR9-specific polyclonal antibody, which also showed that a significant fraction of peripheral γδ T cells express CCR9 (Uehara et al., 2002).
Possibly because of the limited availability of CCR9+ primary cells, relatively little information is available on the signaling pathways downstream of CCR9. The few published studies of CCR9 signaling have come from CCR9+ T-cell lineage acute lymphocytic leukemia (T-ALL) cells. In a survey of 38 T-cell leukemia cases, Qiuping et al. (2003) showed that CCR9 was highly expressed on peripheral blood CD4+ T-ALL cells but was less well expressed and expressed on fewer CD4+ cells from patients with T-cell lineage chronic lymphoblastic leukemia (T-CLL). CCR9 expression was associated with chemotaxis of T-ALL cells but not T-CLL cells to CCL25 (Qiuping et al., 2003). The same group subsequently reported that CCL25 treatment of CCR9+ T-ALL cells enhanced resistance to TNFα-induced apoptosis, and this protection was associated with CCL25-mediated upregulation of the inhibitor of apoptosis protein (IAP) family member Livin (Qiuping et al., 2004). The authors demonstrated that CCL25-mediated Livin induction was via a JNK1 pathway and that shRNA inhibition of Livin specifically blocked CCL25-mediated resistance to TNFα-induced apoptosis. Other CCR9 signaling studies undertaken using the MOLT4 T-ALL cell line revealed a role for Rho-ROCK and Ezrin in mediating cytoskeletal rearrangement following CCL25 treatment (Zhou et al., 2010; Zhang et al., 2011).
The first description of Ccr9−/− mice reported no major developmental or immunologic abnormalities, and the fact that Ccr9−/− thymocytes showed no chemotaxis to CCL25 strongly suggested that CCR9 is the sole functional receptor for CCL25 (Wurbel et al., 2001). Parenthetically, it should be noted that a nonsignaling orphan chemokine receptor (CCX CKR) with ∼30% sequence homology with CCR9 was identified by Gosling et al. (2000) and shown to bind the CCR7 ligands CCL19 and CCL21 and the CCR9 ligand CCL25 with high affinity (Comerford et al., 2006). Further detailed analysis of lymphocyte homing in wild-type and CCR9-deficient mice revealed CCR9 expression on IgA+ plasma cells in mesenteric lymph nodes and Peyer’s patches that was downregulated on migration to the small intestine (Pabst et al., 2004). Ccr9−/− animals failed to mount a normal IgA response to orally administered antigens (Pabst et al., 2004), although Ccr9−/− animals showed no gross defect in rotavirus-specific plasmablast recruitment to the intestine during rotavirus infection (Feng et al., 2006).
Given the expression of the CCR9 receptor on gut-homing lymphocytes in the blood and lymphocytes in the gut, a number of laboratories looked at the effects of CCR9 ablation and CCL25 blockade in mouse models of intestinal inflammation. Apostolaki et al. (2008) generated Ccr9- and Ccl25-deficient mice on the TNFΔARE/+ mouse background. TNFΔARE/+ mice have elevated levels of TNFα and develop a Crohn’s disease-like inflammation of the ileum, which is characterized by increased CD8+ T-cell recruitment. Ablation of either the Ccr9 gene or the Ccl25 gene did not alter the disease pathogenesis in this study (Apostolaki et al., 2008). A similar study using TNFΔARE/+ Ccr9-deficient mice generated by a second group showed increased disease severity in Ccr9−/− mice compared with Ccr9+/+ mice at 4, 8, and 20 weeks of age that was accompanied by an increased number of CD8+ effector cells and a decreased number of Treg cells in the lamina propria of Ccr9-deficient animals (Wermers et al., 2011). In this second study, the authors also treated TNFΔARE/+ animals with a CCR9 blocking monoclonal antibody and observed a similar increase in intestinal inflammation to that seen in their TNFΔARE/+ Ccr9-deficient mice. Given that both genetic deletion of CCR9 and antibody blockade of CCR6 exacerbated ileitis in this model, the authors concluded that CCR9 has beneficial effects in this spontaneous model of intestinal inflammation. In a subsequent study, Walters et al. (2010) reported that treatment of TNFΔARE/+ mice with subcutaneous injections of the CCR9 antagonist CCX282-B reduced disease scores judged by histology. This study seems difficult to reconcile with the two previous reports on the role of CCR9 in the TNFΔARE/+ model. Further clarification of the role of CCR9 in the development and chronicity of ileitis could come from testing CCX282-B in Ccr9−/−, TNFΔARE/+ mice, and from testing CCX282-B in other models of intestinal inflammation.
Wurbel et al. (2011) studied the role of CCR9 in inflammation of the colon by testing the effects of CCR9 and CCL25 deficiency in the DSS model of colitis. The authors showed that Ccr9−/− and Ccl25−/− mice are more susceptible to DSS colitis than wild-type littermate controls and that these animals take longer to recover from colitis upon withdrawal of DSS from the drinking water. Taken together, the published work using the TNFΔARE/+ and DSS models are consistent with a homeostatic rather than a proinflammatory role for CCR9+ cells in murine models of intestinal inflammation.
Interestingly, a recent report implicated activated macrophages expressing CCR9 in the pathogenesis of murine acute liver injury following con A injection. Ccr9−/− mice did not develop hepatitis following con A injection unless they also received CCR9+-activated macrophages from a con A-injected wild-type donor. Moreover, con A-induced hepatitis could be ameliorated by injection of CCL25 blocking antibodies (Nakamoto et al., 2012). Another interesting CCR9+ myeloid cell population was identified by Zeyda et al. (2010) in their analysis of adipose tissue macrophages obtained from high fat diet-induced obese mice. Using F4/80, Mannose Receptor (MR), and CD11c, the authors showed differential expression of CCR9 in F4/80+ MR- CD11c− cells. The relationship of this myeloid cell population to the CCR9+ tolerogenic immature plasmacytoid DCs (pDCs) described by the group of Eugene Butcher remains to be defined (Hadeiba et al., 2008).
Given the phenotype of CCR9-deficient mice in models of intestinal inflammation it is a little surprising that no genetic association between CCR9 or CCL25 has been seen in GWAS studies and case-control cohorts for inflammatory bowel disease. However, Inamoto et al. (2010) described an SNP (rs12721497, G926A) within the human CCR9 gene, which alters the amino acid sequence within the third extracellular domain loop from valine to methionine (V272M). Stably transfected Jurkat cells expressing both variants of CCR9 at similar levels were generated and compared for their chemotactic response to CCL25 response. Jurkat cells expressing the CCR9-926G variant were more responsive to CCL25 than Jurkat cells expressing the CCR9-926A variant (Inamoto et al., 2010).
J. CCR10
The first citation of a receptor named CCR10 was in 1997 (Bonini et al., 1997; Bonini and Steiner, 1997). The receptor was found to be highly expressed in placenta and fetal liver and bound several CC chemokines, including CCL2, CCL7, CCL13, and CCL5. Sequence analysis showed the receptor did not have a DRY box motif and was unable to induce calcium flux in transfected cells. In the same year, this receptor was also identified by a second group who named the receptor D6, which is now known to be a decoy receptor for chemokines without signaling capacity (Nibbs et al., 1997).
The true CCR10 receptor was identified in 2000 when the orphan receptor GPR2 was cloned in humans and mice and found to bind a chemokine, ESkine/CTACK, now named CCL27 (Homey et al., 2000; Jarmin et al., 2000). The receptor had been previously localized to chromosomal locus 17q21 in a cluster with several CC chemokines (Marchese et al., 1994). High levels of human CCR10 mRNA were found in testis and small intestine and foetal lung and kidney, whereas many other tissues including spleen, thymus, lymph node, and colon showed low level CCR10 expression (Jarmin et al., 2000). In the mouse, CCR10 mRNA was not found in the testis, but high levels were detected in small intestine, colon, lymph node, and Peyer’s patches, with lower levels found in spleen and thymus. This expression pattern suggested a potential role for the receptor in leukocyte trafficking. Functionally, CCR10-transfected L1.2 cells were found to release intracellular calcium and migrate in response to CCL27 but not to 17 other chemokines tested (Jarmin et al., 2000). A second group also demonstrated that CCR10 mediated migration and calcium flux of transfected Baf/3 cells toward CCL27 (Homey et al., 2000). These authors further demonstrated that the receptor is expressed in T cells and Langerhans cells but not monocytes or DCs. Finally, CCR10 was found to be expressed by several cell types in normal skin and was upregulated by the inflammatory cytokines IL-1β and TNFα in melanocytes (Homey et al., 2000).
A second chemokine ligand for CCR10, CCL28 (MEC), was subsequently identified in humans and mice and found to chemoattract both CD4+ and CD8+ T cells (Wang et al., 2000). Another group demonstrated that this chemokine could also induce migration of eosinophils via the CCR3 receptor (Pan et al., 2000).
CCL27 has been shown to have a key role in T-cell migration during cutaneous inflammation, and CCR10 is expressed on most skin-homing T cells in patients with psoriasis and atopic and allergic dermatitis (Reiss et al., 2001; Homey et al., 2002).
The CCL28-CCR10 axis is believed to be important in migration of IgA antibody-secreting cells (ASCs) to mucosal surfaces (Wilson and Butcher, 2004). By using CCR10 knockout mice (which are developmentally normal), this was shown to be crucial to migration of IgA ASCs to the lactating mammary gland but not to intestinal surfaces (Morteau et al., 2008).
A very recent study suggests that CCL28-mediated Treg recruitment plays a role in tumor development (Facciabene et al., 2011). Hypoxia occurs in many tumors and contributes to angiogenesis. However, this process releases damage-associated molecular patterns and could therefore drive tumor rejection without the induction of tolerance via Treg cell recruitment. Facciabene et al. (2011) demonstrated that hypoxia induces CCL28 expression, leading to Treg cell recruitment, tolerance induction, and angiogenesis, hence promoting tumor survival.
K. Atypical CC Chemokine Receptors
In addition to the receptors described above, a number of nonsignaling receptors are expressed by mammalian cells that bind CC chemokines with high affinity without generating a functional response. There are three nonsignaling or “atypical” receptors known to be CC chemokine receptors: DARC, D6, and CCX-CKR.
DARC (the Duffy antigen receptor for chemokines) is an antigenic determinant expressed on the surface of red blood cells, where it is also the entry receptor for the malarial parasite Plasmodium vivax (Cutbush and Mollison, 1950; Miller et al., 1976). DARC binds a wide range of both CXC and CC chemokines, including CCL2 and CCL5 (Neote et al., 1993a; Neote et al., 1994). DARC is also expressed on the surface of venular endothelial cells, where it is known to mediate transcytosis of chemokines from the basolateral to apical surface of the cell (Pruenster et al., 2009). Rather than being degraded, transported chemokine remains intact and promotes migration of leukocytes across the endothelial monolayer. On red blood cells, DARC may function as a sink for chemokines in blood, which may prevent excessive inflammatory reactions as has been shown to occur in DARC-knockout mice in an endotoxemia model (Dawson et al., 2000).
D6 was cloned in 1997 and initially thought to be a functional chemokine receptor designated as CCR10 (Bonini et al., 1997). D6 was subsequently shown to be an atypical receptor with a mutated DRY motif and the ability to bind a broad range of inflammatory CC chemokines (Nibbs et al., 1997). D6 is expressed on tissues, including placenta, skin, gut, and lung, where it is primarily found on lymphatic endothelial cells. On lymphatic endothelium, D6 has been shown to mediate chemokine scavenging of chemokines, leading to internalization and degradation (Fra et al., 2003). Interestingly, D6 can selectively recognize only the active forms of some chemokines, e.g., CCL14 without binding inactive molecules that have been truncated by CD26 (Savino et al., 2009). Numerous studies have demonstrated that D6 knockout mice display excessive inflammatory responses in several experimental models, supporting a chemokine scavenging role for D6 in vivo (Jamieson et al., 2005).
CCX-CKR was originally identified as a signaling receptor and named CCR11, although subsequent studies failed to replicate this finding, and it is now known to be a nonsignaling receptor (Schweickart et al., 2000). CCX-CKR binds the homeostatic CC chemokines CCL19, CCL21, and CCL25 (Gosling et al., 2000). Like D6, CCX-CKR is believed to have a scavenger role, binding chemokines for internalization and degradation. The function of CCX-CKR in vivo has not been as extensively explored as for the other decoy receptors, but two studies have suggested a role in DC trafficking and in the regulation of immune responses, because CCX-CKR knockout mice have exaggerated responses in the EAE model (Heinzel et al., 2007; Comerford et al., 2010).
Clearly the atypical chemokine receptors have an important role in the regulation of inflammatory responses in vivo. Further discussion of these receptors is outside the scope of this review, and we refer the reader to an excellent recent review exploring the biology of these receptors in more detail (Graham et al., 2012).
III. Role of Chemokines in Chronic Inflammatory Diseases
In this section, we review the role of chemokines and their possible utility as therapeutic targets in three key disease areas, namely, atherosclerosis, rheumatoid arthritis, and metabolic syndrome. Although disparate in their clinical manifestations, these three forms of chronic inflammation share several common features: they often persist for several decades, have similar underlying pathologic mechanisms (notably monocyte recruitment and macrophage activation), and all constitute a significant healthcare burden. Furthermore, these diseases often exist as comorbid pathologies in the same individual: patients with metabolic syndrome are at increased risk of atherosclerosis, as are individuals with RA.
A. Atherosclerosis
1. Summary of Pathology
Atherosclerosis is a chronic inflammatory disease of the major arteries. Atherosclerotic plaques composed largely of modified lipids, recruited macrophages, T cells, and smooth muscle cells (SMCs) accumulate in the arterial wall over many decades. Eventually these plaques may occlude the lumen of the artery, leading to angina (if sited in the coronary arteries) or transient ischemic attacks (when localized in the carotid or cerebral arteries). Acute plaque rupture can lead to potentially fatal outcomes, including myocardial infarction (MI) and stroke. In the UK and the US, cardiovascular disease accounts for one in three of all deaths, with half of these occurring as a result of coronary heart disease. Cardiovascular disease also has a significant economic impact, with direct costs of £14.4 billion to the UK economy in 2006, mainly for hospital-based treatment. Indirect costs, which are more difficult to measure, include production losses as a result of mortality and morbidity, particularly for stroke.
The histologic features of atherosclerosis have been well described for over a century, and multiple models to explain pathogenesis have been proposed (Williams et al., 2012). The key initiating event in atherogenesis is believed to be endothelial dysfunction, which can occur as a result of hypertension, diabetes, smoking, or elevated plasma low-density lipoprotein (LDL) levels. Endothelial dysfunction leads to decreased nitric oxide (NO) production and expression of adhesion molecules capable of mediating monocyte adherence, production of inflammatory cytokines, and enhanced permeability of the endothelium. Apolipoprotein B (Apo B)-containing lipoproteins enter the vessel wall by diffusion, where they can become modified to become oxidized LDL (oxLDL) or minimally modified LDL (mmLDL), pro-inflammatory species that are likely detected by resident macrophages in the vessel wall. Monocytes are recruited from the blood into the intima in response to chemokines such as CCL2 expressed on the endothelium. In the subendothelial space of the tunica intima, monocytes differentiate into macrophages that then phagocytose the oxLDL in the vessel wall via scavenger receptors. This nonsaturable pathway of modified LDL uptake leads to accumulation of cholesterol droplets in the cytoplasm of the macrophage, generating the canonical “foam cells” that are typical of early atherosclerotic lesions. T cells, particularly CD4+ Th1 cells, are also recruited into early atherosclerotic lesions and have been found to recognize self antigens, including oxLDL and HSP60. Th1 cells produce large quantities of IFN-γ that activate macrophages, leading to further cytokine and chemokine production.
Continued recruitment of inflammatory cells and accumulation of modified lipid leads to the generation of a necrotic core to the plaque composed of dead and dying cells as well as extracellular cholesterol. As the plaque continues to develop, SMCs are recruited from the tunica media into the tunica intima where they proliferate and secrete extracellular matrix, forming a “fibrous cap” that covers the inflammatory necrotic core of the plaque. This process may continue for many decades, leading to arterial stenosis and gradual loss of the vessel lumen. Furthermore, apoptosis of SMCs or degradation of extracellular matrix proteins in the cap weakens this structure, leading to plaque rupture and release of thrombogenic material into the bloodstream. An arterial thrombus can form rapidly, leading to cessation of blood flow and ischemia manifested as MI or stroke.
2. Current Treatments
Primary and secondary prevention strategies for atherosclerotic disease currently target modifiable risk factors. Lifestyle and dietary advice are offered to help patients stop smoking, reduce plasma cholesterol, exercise more, and lose weight. Pharmacological interventions include the use of antihypertensive medications, lipid lowering, and antiplatelet drugs to reduce the likelihood of thrombus formation and MI or ischemic stroke. In patients at high risk of an acute cardiovascular event, or immediately following such an event, anticoagulants or antiplatelet drugs are used routinely to limit a repeat event.
Several agents exist to modify plasma cholesterol, including bile acid sequestrants, niacin and ezetimibe (Zetia), but statins are by far the most widely prescribed, accounting for approximately 1 million prescriptions per week in the UK. Statins competitively inhibit the rate-limiting enzyme in cholesterol bio-synthesis, HMG-CoA-reductase, which converts HMG-CoA to mevalonic acid. A fall in cholesterol synthesis leads to an upregulation of hepatocyte LDL-R expression, increasing clearance of LDL from the plasma. It is now known that statins have numerous actions that are independent of their cholesterol-lowering activity, some of which may be mediated by products of the mevalonate pathway that prenylate or farnesylate several membrane-bound enzymes.
3. Evidence Supporting a Role for CC Chemokines in Development of Pathology
Evidence supporting a role for CC chemokines in atherogenesis is derived from three main sources: hypercholesterolemic mouse models, epidemiology of human chemokine or chemokine receptor polymorphisms, and histologic evidence for chemokine expression within human atherosclerotic lesions.
The majority of atherosclerosis research in animals has used two well characterized models of accelerated atherogenesis: the Apoe−/− and Ldlr−/− mouse models that develop severe hypercholesterolemia and plaque formation when fed a high-fat diet [reviewed in (McNeill et al., 2010)]. The development of atherosclerosis in these animals may be further accelerated by the use of mechanical stimuli, such as arterial wire injury, which causes significant endothelial damage.
The first evidence for a causal role for chemokines in atherogenesis in this model came from the generation of Ccr2−/− Apoe−/− mice (Boring et al., 1998). Quantitative histologic analysis of lesions in the aorta of these animals revealed that the absence of CCR2 inhibited plaque formation and reduced macrophage infiltration into the vessel wall. Subsequent data generated in Ccl2−/− Ldlr−/− mice confirmed that absence of the CCR2 ligand (JE) also significantly reduced lesion formation (by 83%) in mice fed a high-fat diet (Gu et al., 1998). Since these seminal studies, numerous reports utilizing chemokine receptor knockout animals crossed onto Apoe- or Ldlr-deficient mouse strains have been published. The results of these studies are presented in Table 2.
Studies utilizing CCR1 knockout animals have provided conflicting data. Deletion of CCR1 in the bone marrow of LDLR knockout animals led to a dramatic 70% increase in plaque area in the thoracic aorta after 12 weeks of fat feeding (Potteaux et al., 2005). In common with studies described above (see II: CC Chemokine Receptor, A: CCR1), the absence of CCR1 in bone marrow led to an enhanced Th1 response to con A and to increased levels of IFN-γ in the spleen, specifically localized to areas of oxidized lipid accumulation, suggesting an enhanced adaptive immune response in these animals. Other studies have shown no effect of CCR1 deficiency on neointima formation following wire injury (Zernecke et al., 2006) and increased plaque formation and T-cell content in Ccr1−/− Apoe−/− mice fed a high-fat diet (Braunersreuther et al., 2007b).
Studies aimed at demonstrating a nonredundant role for CCR5 in mouse models of atherogenesis have met with limited success. Initial studies reported no significant effect of deleting CCR5 on lesion size in Apoe−/− mice on chow diet (Kuziel et al., 2003) and smaller lesions in Ccr5−/− Apoe−/− animals fed a high-fat diet (Braunersreuther et al., 2007b). Bone marrow transfer of Ccr5−/− stem cells was reported to have no effect on lesion size but was associated with a reduction in plaque macrophage content (Potteaux et al., 2006). Taken together, the published data on CCR5 in mouse models of atherosclerosis is not as impressive as the results obtained by ablation or targeting of CCR2 (or CX3CR1), and currently CCR5 is not seen as a promising target for therapeutic intervention in cardiovascular disease.
The role of CCR7 and its ligands CCL19 and CCL21 in lymphocyte trafficking, DC migration, and lymphocyte motility in the context of mounting adaptive immune responses is well established (Gunn et al., 1999; Robbiani et al., 2000; Worbs et al., 2007). The role of CCR7 in atherogenesis is, however, more controversial due to the demonstration that CCR7 is important for macrophage emigration out of established atherosclerotic plaques in an aortic transplantation model of atherosclerosis (Trogan et al., 2006; Feig et al., 2010, 2011a). Ablation of the CCR7 gene in Ldlr−/− mice followed by 12 weeks of a high-fat diet was associated with reduced plaque size and reduced numbers of both CD68+ cells (macrophages) and CD11c+ cells [macrophages and DCs (Luchtefeld et al., 2010)]. An intriguing recent publication from the Fisher and Garabedian laboratories demonstrated that Apoe−/− mice a high-fat diet and treated with either atorvastatin or rosuvastatin demonstrated no change in atherosclerotic plaque area but did show significant reductions in CD68+ macrophage staining area (Feig et al., 2011b). Plaques from statin-treated animals had reduced CCL2 expression compared with control animals but showed increased CCR7 mRNA expression. The statin-induced decrease in CD68+ macrophages was not seen in Ccr7−/− Apoe−/− mice treated with atorvastatin or rosuvastatin. The authors went on to show that statin treatment increased macrophage CCR7 expression via a sterol regulatory element (SRE) within the CCR7 promoter.
Somewhat surprisingly given the focus on CCR9 expressing cells in mucosal immunity, Abd Alla et al. (2010) demonstrated that RNA-interference mediated knockdown of CCR9 in hematopoietic progenitor cells of Apoe-deficient mice retarded atherosclerotic lesion development. Supporting evidence for a role for the CCL25/CCR9 axis in atherogenesis came from demonstration of CCR9+ and CCL25+ cells in human atherosclerotic lesions. It will be interesting to determine more precisely which cell types in atherosclerotic plaques express CCR9 and whether genetic deletion or antibody blockade of CCL25 and CCR9 similarly affects atherogenesis in Apoe−/− and Ldlr−/− mouse models.
The functions of key CC chemokine receptors that have been well defined in atherogenesis are summarized in Fig. 3.
4. Key Chemokine Receptors as Drug Targets
Many of the CC chemokine receptors discussed previously as promising candidates for therapeutic intervention in atherogenesis on the basis of knockout mouse studies have been targeted in animal models of atherosclerosis using small molecule antagonists, N-terminal deleted or modified chemokines, monoclonal antibodies, and viral CC chemokine binding proteins. CC chemokine receptors targeted in animal models of atherosclerosis using small molecules include targeting of CCR2 using GSK 134486B in Apoe−/− mice fed a high-fat diet and implanted with angiotensin II-containing osmotic mini pumps (Olzinski et al., 2010), oral delivery of the CCR2 antagonist INCB-3344 in Apoe−/− mice (Aiello et al., 2010), targeting of both CCR2 and CCR5 with a nonpeptide small molecule in Apoe−/− mice with an angiotensin II mini pump (Major et al., 2009), and TAK-779 targeting of both CCR5 and CXCR3 in Ldlr−/− mice (van Wanrooij et al., 2005). Peptides used to antagonize CC chemokine activity in animal models of atherosclerosis include a peptide that disrupts CXCL4-CCL5 heterodimerization (Koenen et al., 2009). Mutated chemokines tested for their antiatherogenic properties in mouse models of atherosclerosis include the CCL5 variant [44AANA47] (Braunersreuther et al., 2008), met-RANTES (Veillard et al., 2004), and N-terminally deleted CCL2 (Inoue et al., 2002). Blocking monoclonal antibodies (mAbs) tested in animal models of atherogenesis include a reagent that blocks murine chemokines CCL2 and MCP-5 (Lutgens et al., 2005). Broad spectrum CC chemokine blockade in Apoe−/− mice has been tested using in vivo delivery of a viral CC chemokine binding protein derived from poxvirus (known as 35K or vCCI) (Bursill et al., 2004; Bursill et al., 2009).
Despite the wealth of intervention studies in preclinical models so far there is only one reported example of CC chemokine receptor blockade conducted in patients at increased risk of cardiovascular disease using the anti CCR2 blocking monoclonal antibody MLN1202 generated by Millennium (Gilbert et al., 2011). The double-blind placebo trial of MLN1202 recruited 112 patients with two or more risk factors for coronary artery disease who had elevated levels of C-reactive protein. Patients receiving a single dose of the anti CCR2 humanized monoclonal antibody displayed significantly decreased levels of plasma CRP between 4 and 12 weeks post treatment and a transient 70% decrease in circulating monocyte numbers. This trial showed that a single dose of MLN1202 is generally well tolerated but did not allow conclusions to be drawn about the efficacy of this reagent to reduce monocyte recruitment into atherosclerotic lesions or decrease the incidence of cardiovascular disease.
Thus, although preclinical animal models of atherosclerosis have been critical in determining pathologic mechanisms and in identifying the key chemokines involved, this is yet to translate into clinical benefit. It remains to be seen whether chemokine blockade, in a disease spanning several decades, could be a viable therapeutic strategy. Any future therapy would need to add value to existing treatments, in particular the use of statins, which may prove challenging. Short-term targeting of chemokines, for example to prevent the accelerated atherogenic processes often occurring following angioplasty, stenting, or organ transplant, may be a more realistic aim.
B. Rheumatoid Arthritis
1. Summary of Pathology
RA is a chronic inflammatory autoimmune condition characterized by synovitis, pannus formation, joint tenderness, and structural damage/destruction of bone, cartilage, and ligaments. The inflammation of the synovial tissue lining the joints leads to the development of symmetrical polyarthritis, which tends to manifest in the wrist and small joints of the hands. RA affects 0.2–1% of the global population, which illustrates the need for the development of effective therapeutic interventions (Imboden, 2009). The severity of RA can vary among patients with some displaying a mild self-limiting disease while in others a more chronic progressive disease may persist, which can eventually lead to total loss of joint function and permanent disability (Cho et al., 2007). In addition to joint inflammation/destruction, other systems, including the cardiovascular and respiratory, may also be affected, which, taken together, can account for the increased morbidity/mortality associated with RA (van Vollenhoven, 2009).
Chronic inflammation of the synovium drives the process of cartilage and bone destruction through the release of a variety of mediators, including chemokines, matrix metalloproteinases (MMPs), cytokines, and growth factors. This persistent inflammation continues as a result of dysregulation in mechanisms involved in the resolution of inflammation, thus leading to continual activation of both innate and adaptive immune systems.
Animal models of autoimmune arthritis have been an invaluable tool in furthering our understanding of some of the key mechanisms involved in the pathogenesis of human RA, such as the involvement of specific cellular subsets and pro inflammatory mediators. Furthermore, animal models have proven to be fundamental in testing and developing novel therapeutics used in the treatment of RA, for example, the development of biologics such as those directed against TNFα (van den Berg, 2001, 2009). A variety of experimental arthritis models are currently used with some better characterized then others. Examples include those that require immunization with antigen/proteoglycan such as collagen-induced arthritis (CIA) (Courtenay et al., 1980), aggrecan-induced arthritis (Finnegan et al., 1999), and streptococcal cell wall arthritis (Koga et al., 1985) and those that are induced with adjuvant such as complete Freund’s adjuvant (CFA) and pristane (Hopkins et al., 1984) and those induced by serum transfer such as the K/BxN model (Kouskoff et al., 1996).
2. Current Treatments
RA along with other inflammatory autoimmune conditions has been traditionally treated with a combination of glucocorticoids (e.g., prednisolone), gold, and nonsteroidal anti-inflammatory drugs [NSAIDs (Malemud, 2009)]. However, treatment with NSAIDs often results in gastric and cardiac side effects and provides only symptomatic relief rather than targeting disease progression (Scott et al., 2010). The introduction of disease-modifying antirheumatic drugs (DMARDs) overhauled existing drug therapies, and to date these remain the leading choice for the treatment of RA. They have been shown to reduce joint swelling and pain and limit joint damage, thereby improving overall function (Malemud, 2009). When first diagnosed with active RA, approximately 50% of patients are started on methotrexate, which still remains the most widely used DMARD (Suarez-Almazor et al., 2000). In some patients, combinational therapy with other DMARDs including hydroxychloraquine, sulfasalazine, and leflunomide has proven to be highly effective (Strand et al., 1999). However, many patients who have been on a single DMARD or a combination over a long period of time can develop adverse side effects and/or become unresponsive.
More recently, the advent of biologic therapies (biologics), which target proinflammatory cytokines such as TNFα have revolutionized the current treatments available for RA. In the case of RA, biologics are prescribed only when treatment with NSAIDs or DMARDs has failed. There are three anti-TNF therapies currently approved in the UK for treatment of RA: etanercept, a fusion of soluble TNFR with the Fc domain of human IgG1; infliximab, a chimeric monoclonal antibody (mAb) against TNFα; and adalimumab, a fully human mAb [reviewed in (Khraishi, 2009]. Two more recently developed anti-TNF drugs—certolizumab, a PEGylated Fab′ fragment of a humanized mAb, and golimumab, a humanized mAb administered monthly—can also be used if other anti-TNFs have failed. Several studies have shown that patients treated with an anti-TNFα concomitantly with MTX display more effective disease suppression compared with patients treated with MTX alone (Maini et al., 1999; Weinblatt et al., 1999; Keystone et al., 2004). Moreover, treatment with the anti-TNFα was also shown to inhibit the progression of joint bone erosion in RA patients (Keystone et al., 2004).
Other cytokines targeted by biologics include IL-1 and IL-6, which are inhibited by the soluble antagonist anakinra and the mAb toculizumab, respectively, which both block receptors for these cytokines (Bresnihan, 1999; Salliot et al., 2011). Other biologics approved for use in the UK include abatacept, an Fc fusion protein of CTLA-4 that binds CD80-CD86 to prevent T-cell costimulation (Ostor, 2008) and rituximab, a chimeric monoclonal antibody that binds CD20 on B cells to induce depletion of these cells via antibody-dependent cellular cytotoxicity and complement dependent cytotoxicity (Dass et al., 2006; Nakou et al., 2009). Severe side effects of biologics can include susceptibility to serious infections such as tuberculosis and rare allergic reactions.
3. Evidence Supporting a Role for CC Chemokines in Development of Pathology
Several CC chemokines (summarized in Fig. 4) can readily be detected in RA synovial fluid and tissue biopsies. CCL2, a potent chemoattractant for monocytes/macrophages, is highly elevated in synovial fluid and sera of RA patients (Koch et al., 1992; Akahoshi et al., 1993). Ex vivo culture of synovial fibroblasts stimulated with IL-1, TNF-α, or IFN-γ results in the production of CCL2 (Koch et al., 1992; Villiger et al., 1992; Hachicha et al., 1993). CIA carried out in rats was shown to generate increased levels of CCL2 in lavage fluid collected from the joints, in addition to increased CCL2 mRNA expression in synovial tissue. Administration of a neutralizing mAb targeting CCL2 was shown to reduce ankle swelling by 30% when compared with controls. This reduction in swelling was associated with a decrease in the number of monocyte/macrophages recruited to the joints (Ogata et al., 1997). In the MRL-lpr mouse model of arthritis, a strain of mice that spontaneously develop arthritis resembling human RA, daily administration of a CCL2 antagonist, MCP-1(9-76), prevented the onset of disease in these animals. When administered therapeutically, a reduction in clinical scores was also observed both macroscopically and histologically (Gong et al., 1997a).
CCL3, a potent chemoattractant of lymphocytes, monocytes, and eosinophils, can be readily detected in synovial fluid and tissue of RA patients (Matsui et al., 2001; Ruth et al., 2003). RA fibroblasts stimulated with IL-1, TNF-α, and IL-18 induce CCL3 production, with peak levels observed after 48 hours. Neutrophils isolated from synovial fluid have also been reported to contain increased levels of CCL3 protein and mRNA compared with peripheral blood neutrophils (Hatano et al., 1999). Infusion of CCL3 antibodies in mice with CIA was shown to delay the onset and reduce the severity of disease (Kasama et al., 1995).
CCL20, a chemoattractant of lymphocytes and monocytes, is detected in large quantities in both RA synovial fluid and tissue (Matsui et al., 2001; Ruth et al., 2003). Synovial tissue fibroblasts produce CCL20 in response to a range of cytokines, including TNF-α and IL-17 (Chabaud et al., 2001; Matsui et al., 2001; Ruth et al., 2003).
CCL5 mRNA expression in synovial fibroblasts is increased following stimulation with IL-1 or TNF-α (Rathanaswami et al., 1993; Hosaka et al., 1994). These effects can be augmented with INF-γ or suppressed by IL-4 (Rathanaswami et al., 1993). Immunohistochemical analysis also revealed elevated CCL5 expression by macrophages present within RA synovial tissue (Volin et al., 1998). Treatment with an anti-CCL5 antibody ameliorates rat adjuvant- induced arthritis and was shown to be as effective as the NSAID indomethacin in this model (Barnes et al., 1998).
In human RA a range of CC chemokine receptors including CCR1-7 have been reported as expressed by leukocytes found in synovial fluid, tissue, and peripheral blood (Loetscher et al., 1998; Qin et al., 1998; Katschke et al., 2001; Ruth et al., 2001; Haringman et al., 2006b; Szekanecz et al., 2006, 2010). The use of antibodies against CCR2 and CCR5 was recently reported to have failed in inhibiting synovial fluid-induced monocyte chemotaxis. Interestingly the use of either an anti-CCR1 Ab or a CCR1 antagonist (BX471) inhibited synovial fluid-induced monocyte chemotaxis, highlighting this receptor as a potential target in reducing monocyte recruitment to RA synovium (Lebre et al., 2011). In terms of animal models, conflicting evidence surrounding CC chemokine receptors has been reported. Treatment with Met-RANTES, a CCR1/CCR5 antagonist, was reported to reduce the incidence of murine CIA in a dose-dependent manner. Animals that did go on to develop disease were shown to have reduced clinical scores following treatment (Plater-Zyberk et al., 1997). A nonpeptide antagonist for CCR5 and CXCR3, TAK-779, administered before the onset of clinical signs of CIA, was shown to reduce disease incidence and severity in addition to a reduction in histologic scores (Yang et al., 2002). CCR2 blockade with a neutralizing Ab during the initiation of CIA was reported to improve clinical scores; however, when administered therapeutically exacerbation in clinical and histologic scores was observed (Bruhl et al., 2004). CIA performed in Ccr2-deficient mice resulted in a more severe disease profile when compared with WT controls (Quinones et al., 2004). In another CIA study, treatment with a small molecule inhibitor of CCR2 (MK-0812) was shown not to have any effect on the disease severity (Min et al., 2010). A recent study from Jacobs et al. (2010) systematically reviewed a range of chemokine receptors and their involvement, if any, in the K/BxN serum transfer model. The results reported showed that mice deficient in a range of CC chemokine receptors (CCR1-7) displayed no difference in disease profile when compared with littermate controls (Jacobs et al., 2010). Collectively these studies highlight the need for caution when interpreting data relating to CC chemokine receptor blockade in animal models of arthritis.
4. Key Chemokine Receptors as Drug Targets
To date, a limited number of studies have investigated the effects of chemokine and chemokine receptor blockade in human RA. A CCR1 antagonist, CP481,715, that showed promising preclinical efficacy through its ability to inhibit synovial fluid-induced monocyte chemotaxis, was tested in a small phase Ib trial (Gladue et al., 2003; Clucas et al., 2007). Patients with active RA who were administered every 8 hours with CP481,715 over a course of 2 weeks showed a reduction in the total number of macrophages within the synovial intimal lining and an overall reduction in CCR1+ cells within the synovium. Furthermore, one-third of all patients showed improved clinical scores by reaching ACR20 criteria following treatment (Gladue et al., 2003). A human CCR2 blocking antibody (MLN1202) tested in phase IIa clinical trial was shown to reduce circulating levels of CD14+ monocytes; however, no clinical improvements were observed in active RA patients on the basis of ACR criteria (Vergunst et al., 2008). Similarly, in another randomized controlled trial, no clinical benefit was observed in RA patients following monoclonal antibody treatment (ABN912) against the CCR2 ligand CCL2 (Haringman et al., 2006a). Maraviroc, a CCR5 antagonist recently used in a phase IIa randomized, double-blind placebo-controlled trial, showed no efficacy in patients with active RA on background methotrexate (Fleishaker et al., 2012). Similar negative outcomes in smaller clinical trials have also been previously reported with other CCR5 antagonists, AZD5672 and SCH351125, in RA (Gerlag et al., 2010; van Kuijk et al., 2010). Ongoing or completed clinical trials targeting chemokine receptors in RA and other chronic inflammatory diseases are summarized in Table 3.
C. Obesity, Metabolic Syndrome, and Type 2 Diabetes
1. Summary of Pathology
Over the last few decades, the incidence of obesity worldwide has increased dramatically, with almost 1 in 4 adults in the UK classed as obese in 2009 (http://www.dh.gov.uk/en/Publichealth/Obesity/DH_078098). Direct costs to the UK healthcare system are estimated at £4.2 billion per year when costs for treatment of obesity-related diseases, including type 2 diabetes, cardiovascular disease, and some cancers, are included. Metabolic syndrome describes the coincidence of multiple cardiovascular risk factors, including obesity, hyperglycemia, insulin resistance, hypertension, and hyperlipidemia.
It is now recognized that persistent, low-grade inflammation provides a causal link between obesity and metabolic disorders such as insulin resistance and type 2 diabetes. In 2003, two studies demonstrated that obesity due to genetic factors or high-fat feeding in mice is associated with macrophage infiltration into adipose tissue (Weisberg et al., 2003; Xu et al., 2003). Furthermore, these macrophages produce large amounts of TNFα, IL-6, and inducible nitric oxide synthase [iNOS (Weisberg et al., 2003)]. Further profiling has demonstrated that macrophages in lean adipose express genes associated with alternative “M2” activation and produce the anti-inflammatory cytokine IL-10 to protect adipocytes from TNFα-induced insulin resistance (Lumeng et al., 2007). Diet-induced obesity causes these cells to undergo a phenotypic switch to classically activated “M1” macrophages capable of producing high levels of proinflammatory cytokines that induce insulin resistance in adipocytes (Lumeng et al., 2007).
Insulin resistance describes the inability of metabolic tissues to respond to the actions of insulin and is seen as impaired glucose uptake into muscle and enhanced lipolysis in adipose tissue. This results in hyperglycemia and hyperlipidemia, with multiple pathologic consequences. Subsequently, peripheral insulin resistance results in increased insulin secretion by β cells of the pancreas, causing hyperinsulinemia. Eventually, β-cell exhaustion may result in sustained hyperglycemia and type 2 diabetes.
2. Current Treatments
Therapy for metabolic syndrome generally involves lifestyle and dietary advice to reduce obesity as well as drug treatment to reduce risk factors, such as hypertension, hypercoagulation, and high plasma cholesterol. More drastic measures to reduce weight include gastric surgery or use of orlistat (GlaxoSmithKline), an oral drug that irreversibly inhibits gastric and pancreatic lipases, thereby reducing absorption of dietary fat. Unfortunately, unpleasant side effects are common, potentially reducing patient compliance. The only other pharmacological treatments licensed for obesity treatment, sibutramine, a serotonin and noradrenaline reuptake inhibitor, and rimonabant, a CB1 antagonist that acts centrally to reduce appetite, were withdrawn from the market due to an increased rate of adverse cardiovascular events and psychiatric side effects, respectively.
In patients who have progressed to type 2 diabetes, there are several drug treatments available. Metformin lowers blood glucose via numerous mechanisms, including increasing glucose uptake and utilization in muscle as well as reducing hepatic glucose production via a mechanism involving activation of AMP activated protein kinase (AMPK) (Zhou et al., 2001). It is usually well tolerated and is the first drug of choice in the majority of obese type 2 diabetic patients who fail to control dietary intake and lose weight. Sulfonylureas such as glibenclamide act directly on β cells in the pancreas to stimulate insulin release and therefore are only useful if residual β cell activity is present. They are generally well tolerated but can lead to prolonged and severe hypoglycemia, particularly when long-acting sulfonylureas are prescribed. Thiazolidinediones such as pioglitazone reduce peripheral insulin resistance via their action on the peroxisome proliferator-activated receptor γ (PPARγ), a nuclear receptor that regulates glucose and lipid metabolism. In adipocytes, activation of PPARγ by endogenous agonists or pioglitazone causes adipocyte differentiation, increased lipogenesis, and uptake of fatty acids and glucose. Importantly, PPARγ also controls alternative activation of macrophages, thus reducing adipose inflammation and improving insulin resistance (Odegaard et al., 2007). Another widely prescribed thiazolidinedione, rosiglitazone, was withdrawn in 2010 due to an association with increased cardiovascular events. If oral medications fail to sufficiently control blood glucose levels, insulin may be prescribed.
Despite the existence of several therapies to treat diabetes, there remains a significant unmet clinical need. As diabetes is a chronic, progressive condition, most of the established therapies eventually become less efficacious over time and new medications or increased doses must be prescribed. Also, apart from metformin, all the antidiabetic medications cause weight gain, meaning that the underlying inflammatory pathology occurring in obese adipose tissue will continue to worsen.
3. Evidence Supporting a Role for CC Chemokines in Development of Pathology
There is strong evidence suggesting a role for the CCL2-CCR2 axis in adipose inflammation and obesity. CCL2 is constitutively expressed by primary human preadipocytes and is upregulated by TNFα treatment in mature adipocytes (Gerhardt et al., 2001). Levels of CCL2, both systemically and in white adipose tissue, are increased in fat-fed mice and plasma CCL2 levels correlate with body weight (Takahashi et al., 2003). In diabetic and nondiabetic obese patients, treatment with the thiazolidinedione rosiglitazone leads to a reduction in plasma CCL2 levels, suggesting a mechanistic basis for the efficacy of these drugs (Mohanty et al., 2004). CCL2 may have a direct role in insulin resistance because CCL2 added to mature adipocytes inhibits insulin-stimulated glucose uptake and expression of several genes required for adipogenesis (Sartipy and Loskutoff, 2003).
To confirm a role for CCR2 in obesity and adipose inflammation, Ccr2−/− mice were fat fed for 24 weeks, and their metabolic phenotype was studied (Weisberg et al., 2006). Ccr2−/− mice had a 15% lower body mass than wild-type animals after 24 weeks fat feeding and were observed to consume fewer calories than Ccr2+/+ animals. Furthermore, obese Ccr2−/− mice had lower fasting blood glucose and insulin levels and were more insulin sensitive than control mice. CCR2 also seemed to have a direct role in macrophage recruitment, because knockout animals had significantly fewer adipose tissue macrophages than wild-type mice. Finally, 14-day treatment of wild-type obese mice with a CCR2 small molecule antagonist (INCB-3344) reduced fasting glucose and insulin concentrations and increased insulin sensitivity. Adipose macrophage numbers were also significantly reduced by CCR2 antagonism, suggesting that relatively short treatment could alter adipose content.
Another chemokine with a known role in obesity, adipose inflammation, and diabetes is CCL5. Both CCL5 and one of its receptors CCR5 were found to be upregulated at the mRNA level in adipose tissue of male fat-fed mice (Wu et al., 2007). This was associated with increased numbers of T cells in obese adipose tissue, and the authors demonstrated that conditioned medium from explanted adipose tissue of fat-fed mice could chemoattract T cells and this activity was blocked by a neutralizing CCL5 antibody. In vitro, both murine and human adipocytes stimulated with TNFα produced abundant CCL5 mRNA. Finally, subcutaneous adipose tissue from obese individuals with metabolic syndrome had significantly higher levels of CCL5 mRNA compared with lean subjects, and the level of CCL5 was positively correlated with body mass index. CCL5 expression was increased further in visceral adipose tissue of morbidly obese patients and correlated with expression of the T cell marker CD3 and the macrophage marker CD11b.
Serum levels of CCL5 were also found to be elevated in cohorts of patients with either impaired glucose tolerance or type 2 diabetes compared with control subjects (Herder et al., 2005). To assess whether CCL5 has a causative role in diabetes, the association between CCL5 serum levels, CCL5 polymorphisms, and the incidence of type 2 diabetes was investigated (Herder et al., 2008). Six SNPs in noncoding regions of the CCL5 gene (promoter, intronic, and 3′ flanking region) were analyzed for any correlation with serum CCL5 levels. The rarer alleles of four SNPs were associated with significantly lower serum CCL5 levels, but after adjustment for confounding factors, neither the CCL5 serum level or genotype were correlated with the incidence of type 2 diabetes. The authors concluded that CCL5 was elevated as a consequence of hyperglycemia, but could not exclude a role for CCL5 in the pathogenesis of adipose inflammation.
A potential pathogenic role for CCL5 in human adipose tissue was investigated by Keophiphath et al. (2010). The authors hypothesized that CCL5 may have a role in mononuclear cell recruitment and in promoting the survival of adipose tissue macrophages from apoptosis induced by free cholesterol. CCL5 was found to induce both adhesion and transmigration of monocytes through an endothelial cell monolayer. Furthermore, macrophages in obese adipose tissue were found in crown-like structures around adipocytes and were shown to be apoptotic by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining. In vitro induction of apoptosis in macrophages by free cholesterol loading [involving incubation of cells with free cholesterol and an acyl-CoA cholesterol acyl transferase (ACAT) inhibitor to block cholesterol esterification] could be reduced by treatment with CCL5. Taken together, this suggests that a potential mechanism for the action of CCL5 in obese adipose tissue is to both recruit and promote the survival of macrophages, thus contributing to ongoing inflammation. The function of CC chemokine receptors in adipose inflammation is summarized in Fig. 5.
4. Key Chemokine Receptors as Drug Targets
There is strong evidence from humans and animal models that CCR2 may be a potential therapeutic target in metabolic syndrome and diabetes. Several small molecule antagonists of CCR2 have been developed and tested in preclinical animal models (Shin et al., 2009; Kang et al., 2010; Sayyed et al., 2011). One of these, CCX140-B, developed by ChemoCentryx has completed phase II clinical trials (clinicaltrials.gov study identifier NCT01028963). The data are not currently published in a peer-reviewed journal, but the manufacturers cite a dose-dependent decrease in plasma glucose and a significant decrease in HbA1c levels in patients receiving CCX140-B. This compound is now entering phase II trials for diabetic nephropathy (clinical trials.gov study identifiers NCT01447147 and NCT01440257). The status of this and other CCR2 antagonists in clinical trials is summarized in Table 3.
Data from humans and mice suggest that CCR5 may have a pathogenic role in metabolic syndrome and type 2 diabetes, and licensed medications targeting CCR5 already exist. Indeed, it has also been shown that the CCR5-Δ32 mutation is associated with better survival in type 2 diabetic patients, suggesting that targeting of CCR5 may be beneficial (Muntinghe et al., 2009). To our knowledge, there are currently no studies assessing CCR5 antagonists in type 2 diabetes.
IV. Chemokine Receptor Drugs in Clinical Trials
A number of CC chemokine receptors have been targeted for potential clinical benefit in inflammatory disease. A summary of compounds currently in clinical trials is presented in Table 3.
V. Critical Assessment of Chemokine Receptors as Anti-inflammatory Drug Targets
In 1998, the publication of a report showing that deletion of the chemokine receptor CCR2 in a mouse model of atherosclerosis could dramatically reduce the initiation of disease provided the first unequivocal evidence that chemokine receptors could be key targets of new anti-inflammatory drugs (Boring et al., 1997). Since then, a plethora of studies, some of which we presented above, have shown that chemokines play a nonredundant role in the development of multiple chronic inflammatory diseases. In this section, we review the successes and failures of chemokine receptor drug development to date and consider future prospects for the field.
Since the discovery that HIV cellular entry requires a chemokine coreceptor (either CCR5 or CXCR4), it is no surprise that a large proportion of basic research and clinical drug development has focused on these two targets. Drugs targeting CCR5 (maraviroc) and CXCR4 [plerixafor (Mozobil)/AMD3100] have been approved for use in HIV infection and for mobilization of hematopoietic stem cells for transplantation, respectively, thus proving that drugs targeting chemokine receptors can be well tolerated and demonstrate therapeutic efficacy. Furthermore, the development of other orally administered drugs targeting inflammatory GPCRs, e.g., the CysLt1 leukotriene receptor antagonist montelukast orally used to treat asthma, demonstrates that in principle, targeting single inflammatory GPCRs is a viable therapeutic strategy. However, 20 years since the description of receptors for the first known chemokine (IL-8, CXCL8), there are still no marketed drugs targeting chemokine receptors for their presumed primary indication—as anti-inflammatories (Murphy and Tiffany, 1991). Indeed, a recently reported phase II trial utilizing maraviroc in patients with RA already receiving methotrexate failed to demonstrate any improvement in clinical score compared with placebo (Fleishaker et al., 2012). This is perhaps not surprising, because different regions of the CCR5 receptor are known to be involved in HIV binding and intracellular signaling, and indeed maraviroc is reported to have no effect on CCR5 signaling (Dorr et al., 2005).
Despite strong evidence from preclinical models demonstrating a nonredundant role for CC chemokine receptors in disease pathogenesis as well as the development of many potent receptor antagonists, why have drugs that target CC chemokine receptors so far failed to yield any viable anti-inflammatory drugs?
A. CCR1
A number of clinical trials have targeted CCR1 in RA, but to date no compound has shown enhanced benefit compared with established treatments (Szekanecz et al., 2011). The biology of CCR1 is complex, not least because the most potent ligands for the receptor require proteolytic activation in vivo at sites of inflammation. Several studies have shown that CCR1 knockout animals show an enhanced inflammatory response with a Th1 biased cytokine profile that often exacerbates disease progression—as is the case for the Apoe−/− model of diet-induced atherosclerosis (Potteaux et al., 2005). CCR1 also seems to have a critical role in neutrophil recruitment, and blockade of the receptor may inhibit pathogen clearance. Although CCR1-mediated leukocyte recruitment seems to be associated with chronic inflammation rather than leukocyte homeostasis, it remains to be seen if CCR1 can be targeted for therapeutic benefit without skewing Th1/Th2 immune responses.
B. CCR2
Antagonism of CCR2 in type 2 diabetes and related pathologies seems to have shown promise in early clinical trials, but several compounds have failed in trials targeting RA or MS. In mice, CCR2 is known to have a critical role in regulating monocyte egress from the bone marrow, which could make it a less attractive target for therapy. However, at least one CCR2 antagonist that has shown clinical benefit in type 2 diabetes (CCX140-B) had no effect on blood monocyte levels in phase I trials. Thus, CCR2 remains a key target in metabolic inflammation although anti-CCR2 trials in the setting of RA have been disappointing (Szekanecz et al., 2011).
C. CCR3
A large number of small molecule antagonists against CCR3 have been developed and tested primarily in patients with asthma. However, no encouraging clinical data have been published so far. We do not know whether this stems from a failure of animal models of allergic airway to predict good targets in human asthma, whether CCR3 blockade will only be efficacious in a subgroup of patients with asthma, or if CCR3 blockade needs to be applied much earlier in the human disease process.
D. CCR4
CCR4 remains an interesting molecular target. A recent study has suggested that CCR4 antagonists may prove useful as vaccine adjuvants as a means of breaking peripheral tolerance induced by Treg cells that can hamper the efficacy of anticancer vaccines (Bayry et al., 2008; Pere et al., 2011). A monoclonal antibody against CCR4 (mogamulizumab) also exists, and trials are ongoing in patients with adult T-cell leukemia/lymphoma. This drug may also be developed in future by Amgen for the treatment of asthma, but as yet no clinical trials have been initiated (Antoniu, 2010). However, in common with several other CC chemokine receptors, e.g., CCR9, the homeostatic role of this chemokine receptor in leukocyte homing to the skin may limit long-term targeting of this receptor outside of oncology.
E. CCR5
As discussed previously, CCR5 has been blocked as a coreceptor for HIV entry by maraviroc, showing that it is possible to target CC chemokines for therapeutic benefit in the infectious disease arena if not in the context of chronic inflammation. It should be noted that at least two other anti-CCR5 drugs failed in clinical trials for HIV due to lack of efficacy (vicriviroc) or liver toxicity (apliviroc).
F. CCR6 and CCR7
Evidence from animal models suggests that the CCR6 and CCR7 receptors have critical roles in immune homeostasis, particularly in intestinal immunity and lymphoid organization, respectively. Evidence that these receptors are important in inflammation is scarce, and it seems unlikely that they will prove useful therapeutic targets. However, it is theoretically possible that agonism of the CCR7 receptor could prove beneficial in certain pathologies, e.g., in atherosclerosis where CCR7 has a key role in plaque regression (Trogan et al., 2006; Feig et al., 2011b).
G. CCR8
The CCR8 receptor has a single known ligand, giving a simple one ligand–one receptor interaction, unlike most chemokines. However, evidence for the role of CCR8 in inflammation is weak, and this receptor remains relatively understudied. To our knowledge, no CCR8 antagonists have been developed for potential therapeutic application.
H. CCR9
A small molecule orally dosed antagonist targeting CCR9 [Traficet-EN (vercirnon)] has been developed by ChemoCentryx (Walters et al., 2010). This compound is currently in phase III clinical trials for Crohn’s disease after promising data were obtained from a phase II trial (although these data have not been published in a peer-reviewed journal). As detailed in II. CC Chemokine Receptors, I. CCR9, results obtained with CCR9 blockade in animal models of colitis have provided conflicting results and debate as to the suitability of CCR9 as a target for treatment of inflammatory bowel disease (IBD). At the heart of the debate is whether CCR9 ligands are solely proinflammatory in colitis or whether therapeutic blockade of CCR9 modifies the behavior of gut-homing T lymphocyte populations to effectively reduce the recruitment of effector T-cell populations while sparing gut-homing Treg cells. One great advantage of oral dosing CCR9 antagonists over using therapeutic antibodies to inflammatory cytokines or chemokines is that the dosage of the CCR9 blockade can be modified and curtailed with “washout” of the anti-CCR9 drug within 48 hours. Given that the design of reported phase II clinical trials was aimed at patients with moderate to severe Crohn’s disease and that one of the readouts was maintenance of clinical remission, CCR9 blockade may represent an important therapeutic paradigm shift, i.e., long-term modification of leukocyte homeostasis with an antichemokine drug rather than antibody-mediated anti-inflammatory targeting of TNF or integrins. Future developments in this area are eagerly anticipated by chemokine biologists, immunologists, and gasteroenterologists.
I. CCR10
As yet, no evidence suggests an inflammatory role for this receptor.
Despite the ongoing development of multiple chemokine receptor drugs and ongoing clinical trials, notably with CCR2 and CCR9 antagonists, it is important to consider why there has been a general failure to translate promising data from in vitro and animal studies into useful clinical therapies. One possible reason is that we have overlooked the importance of chemokine receptors in regulation and resolution in the immune response. As seen with CCR4, Treg cells—not just effector cells—require chemokine receptors for efficient migration and hence maintenance of immune tolerance. In atherosclerosis, the CCR7 receptor has been shown to be critical for exit of inflammatory cells from sites of inflammation and for efficient dendritic cell homing and antigen presentation. We perhaps have underestimated the importance of other chemokine receptors in these processes, meaning that antagonism of these receptors could hamper inflammation resolution.
Another potential reason for clinical failures of chemokine receptor drugs is the reliance on animal models, particularly mice, for identification of key targets. Species differences in CC chemokine receptor binding specificities have led to development of human CC chemokine knock-in mice to test therapeutic efficacy and toxicity of small molecules active at human but not mouse or rat CC chemokine receptors. Additionally, the use of transfected cells overexpressing chemokine receptors for initial stages of drug screening may make it hard to translate findings obtained in these cells to primary cell assays or preclinical models. The fact that many chemokine receptors may function on primary cells as homo- or heterodimers may mean that transfected cells lack key signaling partners. It seems sensible to ensure that early drug screens should incorporate human primary cell assays to increase the likelihood of later success and assays should be conducted in the presence of human plasma to better predict in vivo potency (Schall and Proudfoot, 2011).
As discussed above another reason why CC chemokine receptor blockade may have provided no new anti-inflammatory drugs may reside in the underappreciated role of chemokine receptors in immune homeostasis. It is clear that CCR4, CCR6, and CCR9 mediate homeostatic leukocyte recruitment, and it could be argued that CCR2 plays an important role in homeostatic regulation of monocytes numbers in peripheral blood by mobilization of monocytes from bone marrow (Tsou et al., 2007). It is perhaps germane to reflect on the fact that one of the few successful drugs that targets human chemokine receptors is plerixafor/AMD3100, which disrupts CXCR4 signaling that is required for hematopoietic stem cell sequestration in bone marrow niches.
Another key consideration is that chemokine receptor knockout animals have been shown to have raised circulating levels of their respective chemokine ligands (Cardona et al., 2008), suggesting some scavenging roles of these receptors under homeostatic conditions. Thus, it is possible that antagonism of chemokine receptors that prevents ligand binding may increase chemokine ligand levels. Because chemokines generally bind to multiple receptors, these ligands may then activate alternative receptors, abrogating the effect of the antagonist. Targeting of multiple chemokine receptors may be required to overcome this potential compensatory mechanism. To date, however, no strategies targeting multiple chemokine receptors have approached clinical use, despite positive data in animal models (Bursill et al., 2004; Shahrara et al., 2005; Millward et al., 2010).
Finally another important issue that needs to be considered when thinking about the paucity of novel anti-inflammatory drugs arising from research into CC chemokine receptors is the important issue of achieving an effective therapeutic dose of chemokine blockade in randomized clinical trials. This issue was recently reviewed by Schall and Proudfoot through consideration of three different clinical trials of three different anti CCR1 drugs (Schall and Proudfoot, 2011). The authors’ conclusion that anti-inflammatory effects of CCR1 blockade will only be seen when antagonist occupancy of the target receptor is greater than 98% over a 24-hour period sets the bar very high in terms of receptor “coverage” and, if true, will have important consequences for the development of drugs against this class of receptors.
Whatever the reason for a lack of therapeutic success to date, there remains a clear unmet clinical need for development of novel anti-inflammatory strategies. It seems likely that with a wealth of available chemokine receptor antagonists, useful drugs will be developed, but the examples of CCR5 and CXCR4 suggest that these may not be for the indications initially predicted from studies performed in animal models. A more mature understanding of the role of chemokines in both homeostasis and inflammation resolution is needed to inform future drug development and clinical trial design.
Authorship Contributions
Wrote or contributed to writing of the manuscript: White, Iqbal, Greaves.
Footnotes
The work in the Greaves laboratory is funded by the British Heart Foundation, Programme Grant Number RG/10/15/28578. G.E.W. is funded by British Heart Foundation project Grant Number PG/10/60/28496.
Abbreviations
- AAD
- asthmatic airway disease
- ABN912
- high-affinity human anti-human CCL2/MCP-1 monoclonal antibody of the IgG4/κ isotype (MW 145 kd)
- ACAT
- acyl-CoA cholesterol acyl transferase
- AHR
- airway hyperresponsiveness
- AIDS
- acquired immune deficiency syndrome
- AMD3100
- plerixafor
- AMPK
- AMP-activated protein kinase
- Apo
- apolipoprotein
- ASC
- antibody secreting cell
- AZD4818
- (1R,9S,12S,13S,14S,17R,20S,21R,23S,24R,25S,27R)-17-Ethyl-1,14,20-trihydroxy-12-{(1E)-1-[(1R,3R,4R)-4-hydroxy-3-methoxycyclohexyl]-1-propen-2-yl}-23,25-dimethoxy-13,19,21,27-tetramethyl-11,28-dioxa-4-a zatricyclo[22.3.1.04,9]octacos-18-ene-2,3,10,16-tetrone
- AZD5672
- 3-[1-[(4-chlorophenyl)methyl]-3-(3,3-dimethylbutanoyl)-5-(quinolin-2-ylmethoxy)indol-2-yl]-2,2-dimethylpropanoic acid
- BAL
- bronchoalveolar lavage
- CCX354
- 1-[4-(4-chloro-3-methoxyphenyl)piperazin-1-yl]-2-[3-(1H-imidazol-2-yl)-1H-pyrazolo[3,4-b]pyridin-1-yl]ethan-1-one
- BX471
- (R-N-[5-chloro-2-[2-[4-[(4-fluorophenyl)methyl]-2-methyl-1-piperazinyl]-2-oxoethoxy]phenyl]urea hydrochloric salt)
- CFA
- complete Freund’s adjuvant
- CIA
- collagen-induced arthritis
- CNS
- central nervous system
- con A
- concanavalin A
- COPD
- chronic obstructive pulmonary disease
- CP481,715
- 2-Quinoxalinecarboxamide, N-((1S,2S,4R)-4-(aminocarbonyl)-1-((3-fluorophenyl)methyl)-2,7-dihydroxy-7-methyloctyl)-, 212790-31-3
- DAG
- diacylglycerol
- DARC
- Duffy antigen receptor for chemokine
- DC
- dendritic cell
- DMARD
- disease-modifying antirheumatic drug
- DSS
- dextran sulfate sodium
- DTH
- delayed type hypersensitivity
- EAE
- experimental autoimmune encephalomyelitis
- EBV
- Epstein Barr virus
- EGFR
- epidermal growth factor receptor
- ELISA
- enzyme-linked immunosorbent assay
- ERK
- extracellular regulated kinase
- GAG
- glycosaminoglycan
- GM-CSF
- granulocyte macrophage colony-stimulating factor
- GPCR
- G protein-coupled receptor
- GRK
- G protein-coupled receptor kinase
- GVHD
- graft versus host disease
- GWAS
- genome-wide association study
- HEK
- human embryonic kidney
- HHV
- human herpes virus
- HIV
- human immunodeficiency virus
- HSP
- heat shock protein
- IAP
- inhibitor of apoptosis protein
- ICAM-1
- intercellular adhesion molecule 1
- INCB3284
- 1-hydroxy-4-[[(1R,3S)-3-propan-2-yl-3-[3-(trifluoromethyl)-7,8-dihydro-5H-1,6-naphthyridine-6-carbonyl]cyclopentyl]amino]cyclohexane-1-carbonitrile
- INCB-3344
- N-[2-[[(3S,4S)-1-[4-(1,3-benzodioxol-5-yl)-4-hydroxycyclohexyl]-4-ethoxypyrrolidin-3-yl]amino]-2-oxoethyl]-3-(trifluoromethyl)benzamide
- INCB8761
- N-[2-[(3S)-3-[[4-hydroxy-4-(5-pyrimidin-2-ylpyridin-2-yl)cyclohexyl]amino]pyrrolidin-1-yl]-2-oxoethyl]-3-(trifluoromethyl)benzamide
- IFN
- interferon
- iNOS
- inducible nitric oxide synthase
- IP3
- inositol trisphosphate
- JIA
- juvenile idiopathic arthritis
- JNK
- Jun N-terminal kinase
- LDL
- low-density lipoprotein
- LN
- lymph node
- LPS
- lipopolysaccharide
- mAb
- monoclonal antibody
- MALT
- mucosal-associated lymphoid tissue
- MAPK
- mitogen-activated protein kinase
- Maraviroc
- 4,4-difluoro-N-[(1S)-3-[(1S,5R)-3-(3-methyl-5-propan-2-yl-1,2,4-triazol-4-yl)-8-azabicyclo[3.2.1]octan-8-yl]-1-phenylpropyl]cyclohexane-1-carboxamide
- MEK
- MAPK/ERK kinase
- MI
- myocardial infarction
- MK 0812
- 1,5-Anhydro-2,3-dideoxy-3-{[(1S,3R)-3-isopropyl-3-{[3-(trifluoromethyl)-7,8-dihydro-1,6-naphthyridin-6(5H)-yl]carbonyl}cyclopentyl]amino}-4-O-methyl-D-erythro-pentitol
- MK-0812
- 1,5-Anhydro-2,3-dideoxy-3-[[(1R,3S)-3-[[7,8-dihydro-3-(trifluoromethyl)-1,6-naphthyridin-6(5H)-yl]carbonyl]-3-(1-methylethyl)cyclopentyl]amino]-4-O-methyl-D-erythro-pentitol
- MLN1202
- Humanised mAb against CCR2
- MMP
- matrix metalloproteinase
- MOG
- myelin oligodendrocyte glycoprotein
- MPC
- myeloid progenitor cell
- MR
- mannose receptor
- NK
- natural killer
- NO
- nitric oxide
- NOD/SCID
- nonobese diabetic/severe combined immunodeficiency
- NSAID
- nonsteroidal anti-inflammatory drug
- OVA
- ovalbumin
- PBL
- peripheral blood leukocyte
- PBMC
- peripheral blood mononuclear cell
- PD980549
- 2-(2-Amino-3-methoxyphenyl)-4H-1-benzopyran-4-one
- pDC
- plasmacytoid dendritic cell
- PI3K
- phosphoinositode 3-kinase
- PIP3
- phosphatidylinositol (3,4,5)-trisphosphate
- PKB
- protein kinase B
- PKC
- protein kinase C
- PLA2
- phospholipase A2
- PLC
- phospholipase C
- PPAR
- peroxisome proliferator-activated receptor
- PPD
- purified protein derivative
- PPD
- purified protein derivative
- PTX
- pertussis toxin
- RA
- rheumatoid arthritis
- RAG
- recombinase activating gene
- RANTES
- regulated on activation normal T cell expressed and secreted
- ROCK
- Rho-associated coiled coil-forming protein kinase
- SB-328437
- methyl (2S)-2-(naphthalene-1-carbonylamino)-3-(4-nitrophenyl)propanoate
- SCH351125
- [4-[4-[(Z)-C-(4-bromophenyl)-N-ethoxycarbonimidoyl]piperidin-1-yl]-4-methylpiperidin-1-yl]-(2,4-dimethyl-1-oxidopyridin-1-ium-3-yl)methanone
- SEA
- Schistosoma mansoni egg antigen
- SHP
- Src homology region 2 domain-containing phosphatase
- SLE
- systemic lupus erythematosus
- SLEDAI
- SLE disease activity index
- SMC
- smooth muscle cell
- SNP
- single nucleotide polymorphism
- SRE
- sterol regulatory element
- SR-PSOX
- CXCL16
- TAK-779
- dimethyl-[[4-[[3-(4-methylphenyl)-8,9-dihydro-7H-benzo[7]annulene-6-carbonyl]amino]phenyl]methyl]-(oxan-4-yl)azanium chloride
- T-ALL
- T-cell acute lymphoblastic leukemia
- T-CLL
- T-cell chronic lymphoblastic leukemia
- TCR
- T-cell receptor
- TGF
- transforming growth factor
- TGF-β
- transforming growth factor-β
- TLR
- Toll-like receptor
- TNBS
- trinitrobenzene sulfonic acid
- TNF
- tumor necrosis factor
- Treg
- regulatory T cell
- TUNEL
- terminal deoxynucleotidyl transferase dUTP nick end labeling
- UTR
- untranslated region
- WT
- wild type
- ZK 756326
- 2-{2-[4-(3-phenoxybenzyl)piperalzin-1]ethoxy}ethanol
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵