Molecular Machinations: Chemokine Signals in Host-Pathogen Interactions

Abstract

Chemokines and their G-protein-coupled receptors represent an ancient and complex system of cellular communication participating in growth, development, homeostasis and immunity. Chemokine production has been detected in virtually every microbial infection examined; however, the precise role of chemokines is still far from clear. In most cases they appear to promote host resistance by mobilizing leukocytes and activating immune functions that kill, expel, or sequester pathogens. In other cases, the chemokine system has been pirated by pathogens, especially protozoa and viruses, which have exploited host chemokine receptors as modes of cellular invasion or developed chemokine mimics and binding proteins that act as antagonists or inappropriate agonists. Understanding microbial mechanisms of chemokine evasion will potentially lead to novel antimicrobial and anti-inflammatory therapeutic agents.

No area of study reflects the benefit of the molecular biologic and informatics revolution better than the field of chemokine research. The first of these small protein molecules (8 to 17 kDa) with potent neutrophil chemotactic activity, now known as interleukin-8(IL-8), was cloned just over a decade ago (116). Since that time, by exploiting expressed sequence tag libraries, more than 40 human chemokines and nearly as many murine homologues have been described. The term “chemokine” was applied to these molecules since their principal biologic activity was considered to be chemotactic, i.e., directing cellular movement along concentration gradients during inflammatory responses. While chemokines are only one class of many types of known chemotactins that span the molecular spectrum from lipids to nucleotides, they stand out because of their molecular stability and target cell specificity. In recent years it has become apparent that their function extends beyond simply attracting leukocytes to sites of inflammation. Evidence indicates that chemokines participate in organ development, angiogenesis, angiostasis, homeostatic leukocyte recirculation, and immune regulation. Since a number of recent reviews have discussed these topics in detail (21, 25, 84, 86, 93, 99, 103, 110, 143, 151, 176, 183), they will be covered only briefly in this review. Instead, after providing some background, the discussion will focus on chemokines as they relate to different microbial infections and provide recent insight into the dynamic contest between host and pathogen to take advantage of chemokine function.

CHEMOKINES

Chemokines are a homologous superfamily of relatively small proteins ranging from 8 to 17 kDa that probably arose through duplication and modification of an ancestral gene. The superfamily of chemokines is subclassified on the basis of the arrangement of cysteine residues located in the N-terminal region of these molecules. These are designated C, CC, CXC, and CXXXC, where C represents the number of N-terminal region cysteine residues and X represents the number of intervening amino acids. The CXC subfamily is sometimes further classified into ELR and non-ELR types on the basis of the presence or absence of a triplet amino acid motif (Glu-Leu-Arg) that precedes the first cysteine amino acid residue of the primary structure of these chemokines. The presence of this motif imparts angiogenic function to this class of chemokines, while the ELR-negative chemokines have angiostatic properties (83). Table 1 provides a listing of known human chemokines and their nearest mouse homologues, along with a new systematic nomenclature proposed by Zlotnik and Yoshie (183).

TABLE 1.

Human chemokines and mouse homologues

Human	Mouse homologue	Systematic name
C-X-C group or α group
Growth-related oncogenes (GRO) α (MGSA), β (MIP-2α), and γ (MIP-2β)	mKC, mMIP-2	CXCL1, CXCL2, CXCL3
Platelet factor 4 (PF-4)	mPF-4	CXCL4
Epithelial neutrophil-activating protein (ENA-78)	?mLIX (LPS-induced CXC)	CXCL5
Granulocyte chemotactic protein (GCP-2)	mGCP	CXCL6
Neutrophil-activating protein (NAP-2)	Unknown	CXCL7
Interleukin-8 (IL-8)	Unknown	CXCL8
Monokine induced by IFN-γ (MIG)	mMIG	CXCL9
IFN-γ-inducible protein (IP-10)	mIP-10 (mCRG-2)	CXCL10
IFN-induced Tcα chemokine (I-TAC)	Unknown	CXCL11
Stromal cell-derived factor (SDF-1)	mSDF-1	CXCL12
BLC or B-cell attractant (BCA-1)	mBLC	CXCL13
Breast and kidney-derived (BRAK)	mBRAK	CXCL14
Unknown	Lungkine	CXCL15
C-C group or β group
I-309	mTCA3	CCL1
Monocyte chemotactic protein- 1 (MCP-1)	mMCP-1/JE	CCL2
Macrophage inflammatory protein 1α (MIP-1α, LD78)	mMIP-1α	CCL3
Macrophage inflammatory protein 1β (MIP-1β)	mMIP-1β	CCL4
Regulated on activation and normally T-cell expressed (RANTES)	mRANTES	CCL5
Macrophage inflammatory protein 1δ (MIP-1δ)	mC10	CCL6
Monocyte chemotactic protein 3 (MCP-3)	mMCP-3/FIC	CCL7
Monocyte chemotactic protein 2 (MCP-2)	mMCP-2	CCL8
Unknown	mMIP-1γ	CCL9/10
Eotaxin	Eotaxin	CCL11
Unknown	MCP-5	CCL12
Monocyte chemotactic protein 4 (MCP-4)	Unknown	CCL13
Hepatocyte chemokines (HCC-1, HCC-2)	Unknown	CCL14,15
HCC-4		CCL16
Thymus and activation related (TARC)	mTARC	CCL17
Pulmonary and activation related (PARC)	Unknown	CCL18
MIP-3β/ELC/exodus-3	mMIP-3β/ELC/exodus-3	CCL19
MIP-3α/LARC	mMIP-3α/LARC	CCL20
6Ckine/SLC/exodus-2	m6Ckine/SLC	CCL21
Monocyte derived (MDC)	mABCD-1	CCL22
Myeloid progenitor inhibitory factor (MPIF-1 or SCYA23)	Unknown	CCL23
Myeloid progenitor inhibitory factor-2 (MPIF-2), eotaxin-2	Unknown	CCL24
Thymus-expressed chemokine (TECK)	mTECK	CCL25
Eotaxin-3	Unknown	CCL26
Cutaneous T cell-attracting chemokine, CTACK/ILC	mCTACK/ESkine	CCL27
C group or γ group
Lymphotactin	mLymphotactin	XCL1
SCM-1α
SCM-1β	Unknown	XCL2
Lymphotactin	mLymphotactin	XCL1
SCM-1α
C-X-X-X-C or δ group
Fractalkine	mFractalkine, mNeurotactin	CX3CL1

Virtually every tissue and cell type tested to date can be induced to secrete chemokines. While a single cell type, like mononuclear phagocytes, may produce a variety of chemokines, there is known tissue-restricted expression of certain chemokines, suggesting that some chemokines have organ-specific functions.

CHEMOKINE RECEPTORS

Not surprisingly, great effort has been directed toward gaining an understanding of the mechanisms of chemokine signal transduction. It is well established that chemokines exert their effect through guanosine nucleotide-protein-coupled receptors (GPCR), which are among a superfamily (potentially over 1,000 members, representing 1% of the genome) of related receptors that are involved in transducing a broad spectrum of extracellular stimuli such as hormones, neurotransmitters, chemokines, odorants, and light (14). In general, the chemokine receptors are between 320 and 380 amino acids in length and show significant sequence homology. A list of human chemokine receptors, their murine homologues, and their cellular distribution is presented in Table 2.

TABLE 2.

Chemokine receptors

Receptor	Ligands	Expression	Chromosomea
CXC group
CXCR1	IL-8, KC, MIP-2 GCP-2	Neutrophils, monocytes, basophils, mast cells, endothelial cells, T cells, NK cells	2 (H)
CXCR2	IL-8, GRO-α, NAP-2, ENA-78	Neutrophils, monocytes, basophils, mast cells, endothelial cells, T cells, NK cells, astrocytes, neurons	2, 1 (H, M)
CXCR3	IP-10, MIG, I-TAC	Activated Th1 cells	X (H, M)
CXCR4 (fusin)	SDF-1α	Myeloid cells, T cells, B cells, epithelial cells, endothelial cells, dendritic cells	2, 1 (H, M)
CXCR5 (BLR1)	BLC/BCA-1	B cells	11, 9 (H, M)
DARC (Duffy antigen)	IL-8, GRO-α, RANTES, MCP-1, TARC	Erythrocytes, endothelial cells, T cells	1 (H)
CC group
CCR1	MIP-1α, RANTES, MCP-3, MPIF-1, HCC-1	Neutrophils, monocytes, mast cells, T cells, B cells, NK cells, astrocycytes, neurons	3, 9 (H, M)
CCR2a, CCR2b	MCP-1, MCP-3, MCP-4	Monocytes, T cells, B cells, basophils	3, 9 (H, M)
CCR4	MDC, TARC	Basophils, NK cells, dendritic cells, Th1 cells, Th2 cells	3, 9 (H, M)
CCR3	Eotaxin 1–3, RANTES, MCP-3, MCP-4	Th2 cells, eosinophils, basophils	3, 9 (H, M)
CCR5	RANTES, MIP-1α, MIP-1β	Monocytes, dendritic cells, Th1 cells, epithelial cells	3, 9 (H, M)
CCR6	LARC/MIP-3α/Exodus	Lymphocytes, dendritic cells	6 (H)
CCR7	SLC/6Ckine/Exodus2/TCA4, CKbeta-11/MIP-3beta/ELC	T cells, B cells, monocytes, NK cells, dendritic cells	17 (H)
CCR8	I309/mTCA3, TARC	Monocytes, Th2 cells, smooth muscle cells, thymocytes	3 (H)
CCR9	TECK, RANTES MIP-1α, MIP-1β	Thymocytes, dendritic cells, activated macrophages	3 (H)
CCR10	CTACK/ILC	T cells, fibroblasts, melanocytes, skin Langerhans' cells, placental cells, liver cells	3 (H)
XCR group
XCR (GPR5)	Lymphotactin	T cells, B cells, NK cells	3 (H, M)
CX3CR group
CX3CR1 (V28)	Fractalkine	NK cells, monocytes, T cells	3 (H)

^aH, human; M, mouse. 

Like all known GPCR, the chemokine receptors have seven transmembrane hydrophobic domains with three intracellular and three extracellular hydrophilic loops. A potentially glycosylated extracellular amino-terminal region is involved in chemokine binding, while the intracellular carboxy-terminal region is involved in G-protein linking and is subject to regulatory phosphorylation. Chemokine receptor activation begins with extracellular ligand binding, which triggers interaction with the intracellular quiescent GDP-bearing trimeric G-proteins (99, 109, 179). This results in exchange of GDP for GTP, causing the G-protein to dissociate into G-alpha and G-beta/gamma subunits. The latter subunit in turn activates enzymes such as phospholipase C and phosphoinositide-3-kinase, which convert phosphotidylinositol-4,5-diphosphate (IP₂) into phosphotidylinositol-1,4,5-triphosphate (IP₃) and diacylglycerol (DAG). IP₃ stimulates the influx of calcium ions, and DAG activates protein kinase C (PKC) isoforms. The intracellular environment is thus prepared for a cascade of phosphorylation events involving a series of kinases (e.g., mitogen-activated protein kinase, protein typsine kinase) and small GTPases (e.g., Ras and Rho) that ultimately effect cellular functions such as adhesion, chemotaxis, degranulation, and respiratory burst (48, 69, 106). An interesting aspect of chemokine receptor biology was the finding that different receptors may be linked to different G-proteins, which in themselves represent a large class of signal transduction proteins (3, 90). Differential linkage of chemokine receptors could imply that cellular effector functions can be differentially induced. The basic structural characteristics and activation of chemokine receptors are shown in Fig. 1.

FIG. 1.

Chemokine receptor ligation and activation events. (Top) The GPCR with its seven transmembrane hydrophobic domains and the G-protein α-β/γ complex in the nonactivated GDP-bound state. (Bottom) After ligation with chemokine (CK), GTP replaces GDP and the Gα-β/γ complex subunits dissociate. This leads to activation of phospholipase C (PLC), which generates DAG and IP₃ from the phosphotidylinositol diphosphate membrane substrate (PIP₂). The DAG activates protein kinase C, while IP₃ initiates calcium (Ca++) flux. The Gα subunit also activates protein tyrosine kinase (PTK). These transduction events initiate cellular functions such as chemotaxis, degranulation, respiratory burst, and cell adhesion, as well as activating regulatory protiens that uncouple the chemokine receptor, leading to desensitization.

A mention should be made of chemokine receptor desensitization and regulation. After ligation, chemokine receptors may be internalized and then degraded or recycled, leaving the membrane temporarily unresponsive to further ligand stimulation. In addition, as alluded to above, the C-terminal region contains target residues that may be phosphorylated by GPCR kinases, which allows the binding of regulatory molecules called arrestins, which cause uncoupling and desensitization (23, 56). Another manner by which the signaling can be regulated is by direct inactivation of the G-protein activity by GTPases, known as RGS (regulators of G-protein signaling) proteins (46).

Chemokine receptors were first described on leukocytes. Subsequently, they have been found on endodermal, mesenchymal, ectodermal, and neuroectodermal cells. Thus, chemokines may participate in the growth and migration of epithelial cells of the skin (88, 118), digestive (51, 123) and reproductive (182) tracts, and neuronal and glial cells of the central nervous system (70). Analysis of chemokine receptor expression by leukocytes has revealed that receptor subtypes are expressed to differing degrees by different cell types, thereby dictating their responsiveness to the various chemokines. For example, neutrophils strongly express CXCR1 and CXCR2, making them most responsive to ELR⁺ CXC ligands, whereas eosinophils appear more responsive to CCR3 ligands. An important aspect of these analyses is evidence suggesting that chemokine receptor expression is subject to cytokine-mediated regulation (16, 28, 101, 112, 133), which permits fine-tuning of cellular responses based on organ location and host immune status. Still more intriguing is evidence indicating that chemokine receptor expression by T lymphocytes changes with maturational state, a notion promoted by Sallusto et al. (150–152). As shown in Fig. 2, in this model naîve CD4⁺ T cells express receptors CCR7 and CXCR4, allowing their normal recirculation through lymphoid tissues. On encountering antigen, the primed T cell enters a T-helper (Th) cell memory stage and expresses a particular complement of receptors depending on whether it is of the Th1 or Th2 subtype. The former express CCR1, CCR2, CCR5, and CXCR3, while the latter selectively express CCR2, CCR3, CCR4, and CCR8. Further activation of the memory subtypes results in expression of CCR7 and CXCR5 to facilitate movement to secondary lymphoid tissues. This model is based in large part on in vitro studies of T-cell receptor transgenic cells and T-cell clones; it has yet to be fully tested in vivo and will no doubt undergo modification. In any event, the model offers hope of developing stage-specific receptor targeted interventional therapies.

FIG. 2.

Stage-specific Th-cell chemokine receptor expression. Sallusto et al. (151) propose that after antigen priming, naïve T cells differentiate into Th1 and Th2 cells that display different patterns of chemokine receptors, allowing selective recruitment to sites of inflammation and secondary lymphoid tissues.

CHEMOKINE EVOLUTION

In phylogenetic terms, directed cellular movement in response to external stimuli is an ancient biologic response. Indeed, G-protein-mediated migratory responses are demonstrable in protozoa (126). Comparative amino acid sequence analysis of mammals and primitive vertebrates indicate that the CXC and CC groups of chemokines diverged from an ancestral gene even before the divergence of the various orders of mammals and probably before the emergence of vertebrates (78, 124). Thus, chemokines represent a very ancient system of cellular communication that has undergone extensive refinement over evolutionary time Fig. 3. For example, the CC group of chemokines has diverged into at least three major clusters. The first group is represented by macrophage inflammatory protein 1α MIP-1α, MIP-1β, and RANTES (mostly CCR1 and CCR5 ligands), the second is represented by the monocyte chemotactic proteins and eotaxin (mostly CCR2 and CCR3 ligands), and the third is represented by I-309 (CCR8 ligand). Interestingly, the C-type chemokines fall into a separate but related cluster, while the CCR4 ligand, MDC, defines an independent group that diverged early from the ancestral chemokine gene. In a similar fashion, the major groups of CXC and CC GPCR branched early from an ancestral gene before the divergence of mammalian orders and appear to have coevolved with their chemokine ligands (64), but the mode of receptor gene evolution was possibly different from that of the chemokines (78). As more chemokines and receptors are defined in various species, further phylogenetic analyses may provide critical information that will allow us to match the evolutionary appearance of immune functions with that of chemokines and their receptors.

FIG. 3.

Chemokine evolution. Based on sequence analysis, the different chemokine classes appear to have diverged from a primordial chemokine gene before the divergence of mammalian orders.

CHEMOKINES IN DEVELOPMENT AND HOMEOSTASIS

Initially, the study of chemokines focused on their role in the pathogenesis of inflammation, but in recent years it has become apparent that these molecules participate in organ system development and homeostasis. Transcripts for particular chemokines are constitutively expressed in many tissues under normal physiologic conditions. These presumably are participating in physiologic events. The breadth of involvement remains to be fully determined, but studies to date have established roles in lymphoid tissues, bone marrow, and vascular tissues.

Lymphoid Tissue Development and Leukocyte Homing

The thymus is a critical organ for T-lymphocyte development, and it expresses transcripts for several chemokines with lymphocyte-attracting properties, such as CCL25 (TECK), CCL17 (TARC), CCL21 (SLC), CCL19 (ELC), and CXCL12 (SDF-1α), which have been implicated in the function of this organ (26, 29, 87, 174, 177). Thymic T-cell development involves several migration steps from the influx of immature pre-T cells to movement from the cortex to the medulla to the blood as they achieve maturity. Chemokines are thought to direct this process, through a combination of changing chemokine receptor expression by different stages of maturing T cells and microenvironmental chemokine production. For example, the chemokine TECK is produced by thymic dendritic cells and is a selective chemotactin for immature T cells (174). Thus, TECK could direct the influx of immature T cells as they begin their maturation process in the thymus.

As in the thymus, there are many migration events associated with lymph node function that involve the movement of lymphocytes, dendritic cells, and macrophages through cortical germinal centers and medullary sinuses. Chemokines such as SLC, BLC, MDC, PARC, TARC, and LARC appear to be involved in the formation of secondary lymphoid tissues and lymphocyte recirculation. In particular, the CXCR5 ligand CXCL13 (B lymphocyte chemokine [BLC]), is important to B-cell mobilization (39–41) while the CCR7 ligand CCL21 (secondary lymphoid tissue chemokine [SLC]) may play a significant role in naîve T-cell migration to lymph nodes (178). SLC is expressed by specialized postcapillary high endothelium of the lymphoid tissues and is active on naîve T cells, dendritic cells, and, to a lesser extent, B cells (168). Mice lacking the SLC gene display abnormal lymph node T-cell homing (67) and, complementarily, transgenic ectopic expression of SLC by pancreatic islet cells results in the local organization of lymphoid tissue (54). Thus, SLC participates in both the development and maintenance of peripheral lymphoid tissues. Other chemokines, like MDC (79) and LARC (47), produced in lymph node microenvironments probably contribute to the migration of memory T cells and dendritic cells to cortical follicles.

Hematopoiesis

Regulation of blood-forming elements is another potentially important function of chemokines. In this regard, CXCR4 ligands such as CXCL12 (SDF-1α) appear necessary for normal hematopoiesis since mice with disruption of the CXCR4 gene display defects in B lymphopoiesis and myelopoiesis. Specifically, stromal cell-derived factor (SDF-1) is a product of bone marrow stromal cells and is involved in mobilizing the emigration of hematopoietic precursors to the marrow during embryogenesis (1, 2).

Interestingly, chemokines appear to influence bone marrow function independently of their chemotactic activity. A large number of CC and CXC chemokines have direct myelosuppressive activity on bone marrow cell proliferation when tested in vitro and in vivo (21, 22). One of the best studied of the myelosuppressive chemokines is MIP-1α, which causes a dose-dependent inhibition of cycling status and decreases the absolute numbers of bone marrow progenitor cells when administered to mice (38). This physiologic function was further demonstrated in mice with knockout of CCR1, a major MIP-1α receptor, which displayed enhanced lineage-committed myeloproliferation and leukocyte mobilization to the blood (20). The myelosuppressive chemokines could potentially be exploited as adjuvant agents in the chemotherapeutic treatment of leukemias, where it is beneficial to protect residual normal marrow precursors by temporarily inducing a nonproliferative state.

CYTOKINE-CHEMOKINE NETWORKS

Before the role of chemokines in specific infectious states is discussed, the interactions of chemokines and cytokines should be mentioned. It has long been recognized that many infectious agents and their components are potent inducers of cytokines such as interleukins, tumor necrosis factors (TNF), and interferons (IFN), which appear to contribute to microbial resistance (157). It is not surprising that recent efforts have been directed toward exploring the capacity of cytokines to regulate chemokine expression and vice versa. Among cytokines, TNF-α, an important product of macrophages, is an especially potent inducer of chemokine synthesis in a number of cell types, including neutrophils, fibroblasts, smooth muscle cells, and endothelial and epithelial cells (72, 104, 125, 166, 169). In vivo depletion of TNF-α in a model of mycobacterial granuloma formation indicated that it provided support to a broad spectrum of chemokines (135). Some chemokines such as MIG and IP-10 are more selectively induced by IFN-γ (55). However, under circumstances of infection, cytokines probably likely synergize with each other and with microbial elements to regulate chemokine production. Interestingly, Th1- and Th2-derived cytokines can have antagonistic effects on chemokines. For example, the Th2-related cytokines IL-4 and IL-13 induce monocyte-derived chemokine (MDC) and C10 production in macrophages, but this is inhibited by the Th1 cytokine IFN-γ (17, 128). Cytokine regulation of chemokines also appears to be target cell specific, as illustrated by the observation that IL-4 and IL-13 strongly induce monocyte chemotactic protein 1 (MCP-1) in endothelial cells but inhibit production in epithelial cells (63, 91, 139). As discussed below, many microbial products can directly evoke chemokines, but the spectrum and degree of chemokines produced in response to infectious agents will be modified in part by cytokines produced during host innate and acquired immune responses. As illustrated in Fig. 4, chemokines are probably produced in stages following a microbial insult, resulting in a cascade of amplification.

FIG. 4.

Cascade model of cytokine-chemokine networks. (Top) When an epithelial surface is compromised by microbial invasion, chemokines can be derived from the damaged cells and subepithelial tissue macrophages (MP), which initially respond to microbial components. The macrophages also produce cytokines (e.g., TNF-α and IL-1) that can activate tissue mesenchymal cells such as myofibroblasts (MF), which likewise contribute to the first wave of chemokines. Cytokines also activate local endothelium to express addressins in preparation for leukocyte recruitment. (Middle) After initiation, leukocytes begin to be locally recruited from the blood. The recruited cells, neutrophils and NK cells, responding in an innate fashion provide a further source of amplifying cytokines (e.g., IFN-γ) as well as chemokines. (Bottom) If the initial innate mechanisms are unable to quickly clear the infection, recruited Th cells reactive with microbe-specific antigens will produce cytokines that further amplify the local production of chemokines. Hatched arrows indicate produced factors. Solid arrows indicate amplifying signals.

CHEMOKINES AND INFECTIONS

The greatest focus of chemokine study has been their relationship to infection and inflammation, no doubt because of the central role of leukocyte responses in these conditions. Numerous papers on these topics have been published in recent years, and certain major themes are beginning to emerge. First, chemokine expression has been detected in association with virtually every microbial infection examined (Table 3). Second, different classes of chemokines may aid host resistance to different microbial agents. Specifically, in an environment with pathogens ranging from viruses to multicellular parasites, specific chemokines may be required to elicit the most effective response by mobilizing appropriate leukocyte subpopulations and effector functions. Examples would include neutrophils in response to staphylococci, eosinophils in response to helminths, or cytotoxic T cells in response to viruses. Third, chemokines can be undermined and exploited by pathogenic microbes that neutralize, mimic, and disrupt their activity. The remainder of this review will present some of the important findings in these areas, beginning with a discussion of the role of chemokines in specific types of microbial infections.

TABLE 3.

Chemokine reported in association with infectious organisms

Chemokine	Association of chemokine with:
	Bacteria	Mycoplasma	Mycobacteria	Fungus	Protozoa	Helminths	Viruses
IL-8/GRO	X	X	X		X	X	X
MIP-2	X		X
IP-10			X	X			X
MIG			X				X
MIP-1α	X	X	X	X	X		X
MIP-1β	X	X	X	X			X
RANTES	X		X	X		X	X
MCP-1	X	X	X	X	X	X
MCP-3						X
MCP-5						X
Eotaxin						X	X
TARC	X
MDC	X
TCA-3				X

Bacterial Infections

Mounting evidence supports the notion that chemokines play an important role in innate immunity to bacteria. In a murine model of peritoneal sepsis, Matsukawa et al. demonstrated roles for MCP-1 and MDC in protecting the host from lethality (114, 115). MCP-1 is indirectly responsible for recruitment of neutrophils by way of leukotriene induction, whereas MDC is important to activation and mobilization of peritoneal macrophages. A recent exciting discovery by Krijgsveld et al. (89) seems to further expand the functional role of chemokines in innate immunity. In a survey of human platelet α-granule proteins, truncated variants of CXC chemokines were isolated and found to be directly bactericidal for Bacillus subtilis, Escherichia coli, Staphylococcus aureus, and Lactococcus lactis, as well as fungicidal for Cryptococcus neoformans (89). Thus, chemokine-like molecules may represent a novel class of antimicrobial agents for therapeutic exploitation.

It is well known that resistance to many gram-positive and gram-negative bacterial infections is dependent on the efficient mobilization of polymorphonuclear neutrophilic leukocytes. These phagocytic cells are normally mobilized within minutes to hours of tissue disruption and consequently play a crucial role in preventing dissemination of infectious bacteria. Many bacteria release products such as formylated peptides, which are directly chemotactic for neutrophils by way of GPCR. In addition to this adaptation, host-generated chemokines provide further amplification to the recruitment of neutrophils. These cells are highly responsive to ELR⁺ CXC chemokines (8), which are potently induced in host cells by bacterial and bacterial products such as S. aureus, endotoxin, peptidoglycan, or Pseudomonas aeruginosa coenzyme S (53, 175). Indeed, using cDNA array analysis, Wang et al. monitored the expression of 600 genes in human monocytes stimulated with bacterial products and found that genes encoding the chemokines IL-8, MIP-2α, MIP-1α, and MIP-1β represented the most strongly induced of the cytokine genes (175). Of these, IL-8 and MIP-2α are both potent ELR⁺ CXC neutrophil chemotactins. The circumstantial evidence suggesting a role for these CXCR1 and CXCR2 ligands in bacterial resistance has been further supported by direct demonstration. Tsai et al. reported that neutralization of the CXCR2 receptor in mice caused striking mortality due to P. aeruginosa pneumonia that was associated with reduced neutrophil recruitment and bacterial clearance (171). In an animal model of corneal P. aeruginosa keratitis, Kernaki et al. showed a critical role for MIP-2α in neutrophil recruitment (85). An important role for epithelial cell-derived IL-8 has been demonstrated in the intestine. Teleologically, it is reasonable to predict that epithelial cells residing at the interface between the host and a potentially microbe-infested environment should have the capacity to secrete chemokines, and many investigators have shown this to be the case (107, 144, 156, 163, 164, 167). IL-8 is expressed strongly by gut epithelial cells. In a rabbit model of Shigella flexneri infection, Sansonetti et al. demonstrated that neutralization of IL-8 decreased the influx of neutrophils into the gut lamina propria and attenuated the severity of epithelial lesions (153). However, bacteria displayed increased transepithelial translocation and growth in the lamina propria with increased passage into the mesenteric blood. Thus, IL-8 was essential in preventing bacterial invasion but did so at the expense of epithelial bystander damage by activated neutrophils.

In addition to gram-positive and -negative bacteria, other types of bacteria and bacterium-related microorganisms have the capacity to induce host chemokines. Consistent with the prominent interstitial leukocyte recruitment associated with Mycoplasma infection, this organism induces the expression of the neutrophil-attracting CXC chemokines. IL-8 and growth-related oncogene α GRO-α, as well as the mononuclear leukocyte-attracting CC chemokines MCP-1, MIP-1α, and MIP-1β (44, 80, 160). A major stimulating agent was mycoplasma membrane-derived macrophage-activating lipopeptide 2 (MALP-2), which appears to be a primary molecular element recognized by the host.

Chemokine production may not be beneficial in all types of bacterial infections. The spirochetal etiologic agent of Lyme disease, Borrelia burgdorferi, strongly stimulated MIP-1α expression as well as IL-8, GRO-α, MCP-1, and RANTES expression in human monocytes, but it is unclear if these products effectively promote the elimination of this organism (162). In fact, in Borrelia infection the chemokines may prove harmful by contributing to tissue damage and chronic arthritis. In a similar vein, Yoneyama et al. (181) demonstrated that fulminant hepatitis induced by Propionibacterium acnes infection and endotoxin was dependent on TARC, a CCR4 ligand that was responsible for hepatic recruitment of host CCR4⁺ CD4⁺ T cells. Immune cell recruitment was associated with lethal hepatic failure.

Bacteria that manifest predominantly as chronic intracellular infections, such as Mycobacteria and Listeria spp., provide special difficulty for the host since chemokines may not be consistently effective in eliminating these agents. It is recognized that mycobacterial organisms induce chemokine production. Compared with controls, bronchoalveolar lavage fluid from patients with active pulmonary tuberculosis contained increased levels of RANTES, MCP-1, and IL-8 (147). In vitro studies indicate that the degree of chemokine production may vary with the bacterial strain and is not necessarily related to resistance (138). Macrophages are key targets of intracellular bacterial colonization. Ligands of CCR2, like MCP-1, can promote macrophage activation and contribute to macrophage recruitment during the T-cell-mediated granulomatous inflammation in response to mycobacterial antigens (19). However, with suboptimal activation of intracellular killing mechanisms, chemokine-mediated inflammation may not be sufficient for protective immunity, which may rely more on cytokines such as IFN-α and TNF-α, as suggested by Orme and Cooper (129). In this model, macrophage-rich, chemokine-mediated granulomas are generated but represent a primitive, less optimal sequestration response when intracellular bacterial destruction fails.

As with mycobacteria, Listeria infection induces a spectrum of chemokines including MCP-1, MIP-1α, MIP-2, RANTES, IP-10, and IL-8 (9, 58). However, in contrast to mycobacteria, Listeria-induced chemokines such as MIP-1α and MCP-1 may contribute more significantly to bacterial elimination, since interference with these chemokines did result in impaired elimination of organisms (35, 94). Specifically, unlike wild-type CD8⁺ T cells, those from MIP-1α knockout mice failed to confer protection when adoptively transferred to Listeria-infected recipients, and similarly disruption of the MCP-1 receptor, CCR2, profoundly compromised the capacity of mice to eliminate this organism. Thus, the relative contribution of chemokines may vary significantly with the virulence of the organism and its adaptive countermeasures.

Fungal Infections

A role for chemokines in fungal resistance was first suggested by Huffnagle et al., who demonstrated by chemokine neutralization that MIP-1α was required during the efferent phase of Th1-cell-mediated immunity to the yeast-like fungus C. neoformans in mice (77). Specifically, MIP-1α promoted the recruitment of phagocytic effectors, most notably neutrophils and macrophages, into the lungs. It was subsequently demonstrated, using MIP-1α knockout mice, that resistance to central nervous system cryptococcal infection was also dependent on MIP-1α (76). Likewise, mice with knockout of CCR1, a MIP-1α receptor, showed impaired resistance to C. neoformans (62). Another chemokine, TCA-3, also appears to contribute to cryptococcal immunity, but this was shown to act indirectly by inducing mononuclear cell chemoattractants MIP-1α and MCP-1 (50).

Recent studies of chemokines in cryptococcal infection of human cells have revealed interesting findings. Incubation of primary human endothelial cells with C. neoformans did not induce chemokine synthesis but resulted in differential inhibition of cytokine-induced IL-8, IP-10, and MCP-1 (122). Chemokine suppression may be an important virulence factor for C. neoformans and may explain the often sparse inflammation associated with opportunistic infections. For example, mice infected intratracheally with a highly virulent strain of C. neoformans demonstrated active proliferation of organisms in the alveolar spaces and a poor cellular inflammatory response (82). This was associated with little or no production of MCP-1, RANTES, MIP-1α, MIP-1β, and IP-10. This could be reversed by administration of IL-12, a Th1- and natural killer (NK)-cell-stimulatory agent, which induced the synthesis of these chemokines along with cellular infiltration involving histiocytes and CD4⁺ lymphocytes. A study by Huang and Levitz, showed that infection of human peripheral blood mononuclear cells with either Cryptococcus or Candida caused the release of MIP-1α, MIP-1β and RANTES (75). However, this study also showed that production was not impaired in mononuclear cells from immunocompromised patients with human immunodeficiency virus (HIV) infection, thus pointing to the critical role of Th cells in ultimate fungal resistance. Taken together, the studies suggest that chemokines are important to the mobilization of leukocytes but that Th1-cell-derived factors are needed to efficiently kill and eliminate fungal organisms.

Chemokines also contribute to resistance in nonyeast fungal infections. The ELR⁺ CXC chemokines MIP-2 and KC/GROα were induced in response to intratracheal administration of conidia of Aspergillis fumigatus (117). Antibody-mediated neutralization of the CXC chemokine receptor CXCR2 resulted in invasive aspergillosis that was associated with reduced lung neutrophil influx and a marked increase in mortality. In contrast, animals provided with constitutive lung-specific transgenic expression of KC/GROα were resistant to the organism, with reduced mortality and a lower fungal burden in the lungs. Thus, CXCR2 ligands appear to be critical mediators of host resistance to invasive A. fumigatus.

Another aspect of Aspergillis infection is the potential development of chronic aspergillosis associated with allergic responses that can result in hypersensitivity pneumonitis and pulmonary fibrosis. This condition may normally be kept in check by CCR2 ligands such as MCP-1, since CCR2 knockout mice display exacerbated Aspergillus hypersensitive responses and lung damage (12). Ironically, other types of chemokines may ultimately contribute to the lung damage. Specifically, the late-stage fibrotic remodeling of lung airways caused by Aspergillus hypersensitivity is in part mediated by CCR1 ligands, since disruption of the CCR1 gene ameliorates this pathology (13).

Protozoal Infections

The relationship of protozoa to chemokines was discovered by Horuk et al. in 1993, when it was reported that the CXC chemokines IL-8 and GROα could block binding of the malarial agent Plasmodium knowlesi to the erythrocyte Duffy blood group antigen (74). It has since been established that the Duffy antigen, a receptor for P. knowlesi and P. vivax, is a promiscuous chemokine receptor also known as DARC (Duffy antigen-related chemokine receptor). DARC is expressed by erythrocytes, capillary endothelial cells, renal collecting-duct epithelial cells, lung alveoli, and cerebellar Purkinje cells (134). Analogous to HIV, some Plasmodium species have exploited chemokine receptors as a means of penetrating host cell membranes. Interestingly, DARC is also among the spectrum of chemokine receptors used by HIV-1, allowing erythrocytes to potentially act as a retroviral reservoir (95).

While host chemokine receptors may abet malarial infection, it is unclear if the chemokine ligands promote elimination. Human P. falciparum infection is associated with IL-8 and MIP-1α release into the circulation (24). Circumstantially, the chemokines are at their highest circulating levels when parasites are at their lowest, but a direct antimalarial effect has yet to be established. However, it is conceivable that chemokine blockade of receptors could inhibit cellular penetration by malarial parasites, thereby limiting subsequent rounds of cell penetration.

Leishmaniasis, another well-studied protozoan infection, is also associated with chemokine induction that appears to promote immunity. Human monocytes infected with promastigotes of the causative agent of cutaneous leishmaniasis, Leishmania major, display rapid induction of MCP-1 and IL-8 (7). Similarly, mouse bone marrow-derived macrophages exposed to L. major promastigotes displayed rapid and transient expression of transcripts of MCP-1 and GRO-α, but MIP-1α, C10, and RANTES were not induced (136). Of the induced chemokines, MCP-1 was involved in macrophage recruitment to the cutaneous lesions and microbial killing (121). Further evidence for chemokine-mediated resistance to L. major was demonstrated in mice with knockout of CCR2 (MCP-1 receptor), which displayed defective dendritic cell migration to the spleen and subsequent impaired Th1-cell-mediated immunity (155).

Helminthic Infections

Adaptive resistance to parasitic worms is a complex biologic issue. Many parasites have learned to fall short of rapidly killing their host and are often difficult to eliminate, as a result of either modulated host immunity or direct evasive measures used by the parasite. When generated, the specific mode of parasite-directed immunity is dependent on the particular maturation stage of the parasite and its primary site of infestation. For example, tissue-penetrating stages of parasites generally induce destructive interstitial inflammatory responses whereas intestinal parasites often induce expulsive responses in the gut in addition to some degree of inflammation. Both Th1 (IFN-γ-mediated) and Th2 (IL-4, IL-5, and IL-13-mediated) immune responses participate in various antihelminthic reactions, but particular attention has been paid to the role of the latter. It is widely recognized that Th2-mediated responses, characterized by eosinophil-rich inflammation, immunoglobulin E (IgE) and IgA production, and mast cell degranulation, are commonly induced during parasitic infections. Moreover, this complex of responses seems to be directed toward worm elimination since it involves the delivery of helminthicidal proteins and stimulation of intestinal expulsion. Not surprisingly, chemokines appear to participate in various aspects of this response.

Evidence suggests that Th2 cytokines elicited by helminth antigens are often associated with strong elicitation of CC chemokines such as MCP-1, MCP-3, MCP-5 and eotaxin, which potentially recruit helminthicidal leukocytes (32, 98, 102, 120, 146, 154). Specifically, since inflammatory responses to helminth parasites often involve eosinophils, it has raised the possibility that helminth infections elicit a specific complement of chemokines designed to recruit eosinophils. Indeed, CCR3 agonists, which are potent chemotactins for eosinophils, are commonly induced during Th2 responses (36, 120, 146, 169). Moreover, using specific anti-CCR3 antibodies, Grimaldi et al. showed this receptor to be important to eosinophil mobilization during Nippostrongylus brasiliensis infection of mice (66). Thus, hosts have probably adapted specific chemokine responses to deal with helminth parasites that are induced or augmented by Th2-cell-derived cytokines.

Chemokines appear to have regulatory effects during immune responses to helminths. Infection with the etiologic agent of river blindness, Onchocerca volvulus, causes a destructive corneal inflammation (keratitis) that is probably related to the host's attempt to kill the worms. Treatment with the helminthicidal agent ivermectin induces parasite destruction that is associated with the production of IL-8, eotaxin, and RANTES (36, 37). Eotaxin correlated with eosinophil mobilization, but RANTES was seemingly tempering eosinophil recruitment. A similar cross-regulatory effect of RANTES was shown in the response to egg antigens of Schistosoma mansoni that was possibly related to direct RANTES-mediated inhibition of Th2-cell-derived IL-4 production (31). Interestingly, RANTES is among a group of chemokines, including MIP-1α and MIP-1β, which appear to be more selectively associated with Th1 responses (158). Therefore, the balance of Th1- and Th2-associated chemokines and cytokines generated during helminth infection probably dictates the final nature of the host response. For example, in a murine model of onchocercal antigen-elicited keratitis, exogenously administered IL-12, an IFN-α-promoting chemokine, impaired Th2 cytokines while augmenting the local production of chemokines (IP-10, MIP-1α, MIP-1β, MCP-1/JE, RANTES, and eotaxin) and exacerbating corneal damage (131). Thus, the Th2 component may protect bystander host tissues from excessive Th1-mediated damage. In a similar fashion, this cross-regulatory effect of the Th2 response associated with helminth infection has been demonstrated to temper coexistent Th1-mediated responses to other microbes, specifically blunting responses to Helicobacter pylori bacteria and bacillus Calmette-Guérin vaccination (59, 111).

Viral Infections

Much attention has been directed at the relationship of viral infection and chemokines. This is especially true since the discovery that HIV utilizes chemokine receptors for attachment to and penetration of host cells. Since this topic has been widely reviewed (11, 33, 71, 73, 100, 142, 145), it will not be discussed further here. Rather, the role of chemokines in antiviral responses will be presented. Like other microbes, viruses represent a diverse class of microorganisms varying in their target tissue preferences and ultimate pathogenic manifestations. The optimal response mechanisms used by the host must therefore also vary. Viral immunity usually employs induction of IFNs, antibody-mediated neutralization, cell-mediated cytotoxicity, or combinations of these mechanisms. Cytotoxicity would potentially involve chemokines for local recruitment of immune effector cells such as cytotoxic lymphocytes and NK cells. Due to cytopathic effects, virally infected cells probably provide the initital source of chemokines that initially recruit and activate inflammatory leukocytes, which in turn begin a cascade of events involving induction of cytokines such as IFN-γ and TNF-α that further amplify chemokine synthesis and leukocyte recruitment with ultimate destruction of virally infected cells. Unfortunately, the latter is sometimes associated with significant bystander damage and potentially permanent organ impairment, e.g., hepatitis-induced cirrhosis. Since their survival depends on viable target tissues, it is in the interest of viruses to undermine host resistance, at least until another host can be infected. Recent reports suggest that eluding chemokine function is one mode used by viruses to achieve this end.

A survey of the literature reveals numerous reports of various modes of chemokine evasion used by viruses (discussed below), but there is surprisingly limited evidence demonstrating a direct role for chemokines in viral resistance. Studies of influenza virus and respiratory syncitial virus infection have provided some insights in this regard. Chemokines are clearly produced during these infections. Upper respiratory tract secretions contain increased levels of IL-8, MCP-1, MIP-1α, MIP-1β, and RANTES in response to influenza virus or respiratory syncytial virus infection that correlate statistically with symptoms and virus release (18, 60, 127, 170). Interestingly, cotaxin, which is constitutively expressed at low levels in respiratory epithelial cells, is also augmented by influenza virus infection and has been shown to potentially contribute to virus-associated asthmatic responses (81). Mice with knockout of the MIP-1α gene had reduced pneumonitis and delayed clearance of influenza virus compared to controls, suggesting that this chemokine promotes but is not essential to viral resistance (34). However, in a model of paramyxovirus infection, either MIP-1α or CCR1 knockout resulted in reduced pneumonitis and impaired viral clearance (49). A more recent analysis of influenza virus-infected mice with CCR2 (MCP-1 receptor) or CCR5 (RANTES, MIP-1β receptor) knockout revealed increased mortality in the latter that was associated with exacerbated pneumonitis whereas the former showed reduced pathology but augmented viral titers (43). These studies illustrate the delicate balance that must be achieved for viral resistance. While chemokines may promote the elimination of infected cells, they also can contribute to inflammation-related morbidity.

The double-edged nature of the chemokine sword is also apparent in herpesvirus infections of MIP-1α knockout mice. In a model of murine hepatic cytomegalovirus infection, MIP-1α was needed for NK cell recruitment to the liver and early virus elimination (149). However, in experimental herpes simplex virus-induced keratitis, MIP-1α-evoked inflammation led to blindness but failed to contribute to viral elimination (172).

Another approach that has been used to test the role of chemokines and cytokines in viral resistance is by genetic engineering of poxviruses to express chemokine and cytokine genes. Such engineered viruses become the source of active mediators upon infection of the host. These studies demonstrated the contribution of MIG and IP-10 to host leukocyte mobilization, but the chemokines lacked direct antipoxvirus activity (137).

CHEMOKINE EVASION BY VIRUSES

While more studies are needed to explore the role of chemokines in viral resistance, it would seem from a teleological viewpoint that it is in the survival interest of viruses to exploit host cells while minimizing the host inflammatory responses. For viruses with limited cytopathic effects, such mechanisms may also be of benefit to the host by minimizing bystander organ damage. Therefore, it is not surprising that viruses have acquired numerous mechanisms to undermine multiple host mediators (for a recent review, see reference 5), but the present discussion will focus on those involving chemokines. It is generally considered that viral chemokine-related elements were derived from acquired host genes during the course of evolution. Therefore, DNA viruses such as herpesviruses and poxviruses appear to be particularly well endowed with regard to chemokine evasive mechanisms, but it should be recognized that this field is still emerging and that undoubtedly other viral and nonviral infectious agents will be found to have developed chemokine countermeasures.

Chemokine function can be undermined in several ways: (i) production of chemokine mimics that act as receptor antagonists, (ii) production of chemokine mimics that act as inappropriate agonists, (iii) production of receptor mimics, and (iv) synthesis of binding proteins that neutralize chemokine activity. Table 4 provides a listing of some of these viral agents and their known activities. As will be illustrated by the discussion below, the precise physiologic function of many of these factors is unknown and largely speculative.

TABLE 4.

Viral chemokines and chemokine-binding factors

Viral chemokine or chemokine binding factor	Virusa	Actions	Reference(s)
Chemokines
vMIP-I	HHV-8	CCR8 agonist	42
vMIP-II	HHV-8	Broad CK antagonist and CCR8 agonist	42, 104, 160
vMIP-III	HHV-8	CCR4 agonist, angiogenic, Th2 attractant	164
U83	HHV-6	CC-CK agonist, monocyte attractant	182
MCK-1/2, m131/129	Mouse CMV	CC-CK agonist, monocyte attractant	57, 107, 147
vCXC-1	Human CMV	CXC-CK agonist	131
vCXC-2	Human CMV	Unknown	95
MC148	MC virus	CCR8 antagonist	104
Tat	HIV	CK-like agonist, induces CXCR4	4, 158
Chemokine receptors
ORF74 (vGPCR)	HHV-8, herpesvirus saimiri, MHV-68, EHV-2	Functional growth-stimulating receptor, promotes Kaposi's angioproliferative lesions	68, 96, 140, 180
US28, E1	Human CMV, EHV-2	Binds CC-CKs, can decrease local CK levels	61, 92
US27, E6	Human CMV, EHV-2	Unknown	15, 113
U12, UL33, M33, R33	HHV-6, HHV-7, human CMV, mouse CMV, rat CMV	Binds CC-CKs, promotes viral replication	96
U51, UL78, M78	HHV-6, HHV-7, human CMV, mouse CMV	Binds CXC and CC-CKs, downregulates RANTES	119
K2R	Swinepox virus	Binds CXC (IL-8 receptor homologue)	5
Q2/3L	Capripox virus	Binds CC-CKs (CC chemokine receptor homologue)	27
Chemokine-binding factors
vCKBP-I, M-T7	Myxoma virus	Secreted broad-spectrum binder of all CK types	5, 6
vCKBP-II, B29R, G3R, CCI, H5R, MT1, ST1	Mousepox virus, smallpox virus, cowpox virus, vaccinia myxoma virus, Shope fibroma virus	Secreted CC-CK binder	65, 97
vCKBP-III, M3	Mouse gamma-herpesvirus	Secreted broad-spectrum binder of all CK types	130, 173

^aMHV, mouse herpesvirus, EHV, equine herpesvirus. 

Viral chemokine evasive mechanisms may be related to the different modes of infection and dissemination utilized by viral species and their subtypes. Human herpesvirus 8 (HHV-8), an agent associated with Kaposi's sarcoma, is the source of a number of chemokine mimics, vMIP-I, MIP-II, and MIP-III (42, 161, 165). It is particularly intriguing that several groups have reported that vMIP chemokines have agonistic interactions with CCR8 or CCR4, chemokine receptors expressed by Th2 cells (42, 52, 161, 165). This activity has provoked the notion that HHV-8 may interfere with the Th1/Th2 immune balance, specifically impeding Th1 (IFN-γ-dominated) antiviral responses. However, in conflicting reports, Chen et al. (30) showed anti-inflammatory effects of vMIP-II in an rat model of glomerulonephritis and Luttichau et al. (105) demonstrated vMIP-II to be a broad-spectrum chemokine receptor antagonist, suggesting that HHV-8 directly inhibited chemokine function. In addition, the latter group showed that the skin-infecting Molluscum contagiosum poxvirus produced a highly selective CCR8 antagonist, raising the possibility that this virus tempers antiviral immunity by impeding the recruitment of dendritic cells bearing CCR8.

Chemokines associated with herpesviruses (HHV-6 and cytomegalovirus) and HIV-1 (4, 57, 108, 132, 148, 184) have agonist activity that appears to promote the recruitment of leukocytes that can serve as a means of viral dissemination. Specifically, it was shown that induced mutations of viral chemokine genes resulted in viral forms causing reduced inflammation and impaired capacity for dissemination (57, 148). In the case of the chemokine-like agonist Tat, produced during HIV-1 infection, not only did it have monocyte chemoattractant properties but also it could promote the expression of CXCR4 in bystander CD4⁺ cells (159). Consequently, the host T cells would be more susceptible to infection by HIV types utilizing CXCR4 as a coreceptor.

Among the viral chemokine receptors, ORF74 is a functional GPCR of HHV-8 with significant homology to the high-affinity IL-8 receptor, and it appears to promote cell growth. Particularly, it appears to promote the formation of Kaposi's sarcoma-like angioproliferative tumors, as demonstrated by transgenic expression in mice (141, 180). This activity is fully consistent with the known angiogenic properties of the ELR⁺ CXC chemokines like IL-8 (10, 45), and it may serve as a means of viral proliferation or survival. Other described chemokine receptors and binding proteins appear to behave primarily as promiscuous chemokine “sponges,” reducing local chemokine concentrations and presumably limiting inflammation (see Table 4 for references). As these viral products become better characterized, they may ultimately serve as the basis of novel therapeutic anti-inflammatory agents.

CLOSING REMARKS

For over a decade, chemokine research has been an exciting and rapidly burgeoning field. We are just beginning to understand the meaning of this ancient mode of chemical communication that has been expanded and refined during evolution to participate in virtually every aspect of physiology. With regard to the interaction of host and pathogen, chemokines are undoubtedly central to immunity, serving both as the sparks that initiate responses to infection and as the fuel that feeds subsequent inflammation. The importance of the chemokine system to host defense is reflected by its extensive built-in redundancy and the elaborate mechanisms used by microbes to undermine or exploit its function. A detailed knowledge of this system will probably provide novel means of manipulating host immunity to promote efficient elimination of pathogens and temper organ-damaging inflammatory conditions.