Chemokine signaling in cancer: Implications on the tumor microenvironment and therapeutic targeting

Summary

Chemokines are soluble factors shown to play important roles in regulating immune cell recruitment during inflammatory responses and defense against foreign pathogens. De-regulated expression and activity of several chemokine signaling pathways have been implicated in cancer progression, including: CCL2, CCL5, CXCL1 and CXCL12. While studies in the past have focused the role of these chemokine signaling pathways in regulating immune responses, emerging studies show that these molecules regulate diverse cellular processes including angiogenesis, and regulation of epithelial cell growth and survival. New evidence indicates that chemokines are critical for cancer progression and indicate complex and diverse functions in the tumor microenvironment. This review will focus on the contributions of chemokine signaling in regulating cancer microvironment and discuss the utility of targeting or delivering chemokines in cancer therapeutics.

I. Introduction

Chemotactic cytokines, also referred as to as chemokines, have long been recognized as critical mediators of the inflammatory response by regulating recruitment of cells from both the innate and adaptive immune systems to the site of injury or infection (Zabel et al, 2006; De Paepe et al, 2008). Early work on chemokines involved studying the effects of Platelet Factor-4 (also known as CXCL4) on the movement of neutrophils and monocytes (Deuel et al, 1981) through chemotaxis, the process by which cells migrate in response to a concentration gradient of chemokine. Since then, chemokines have also been shown regulate other biological processes including angiogenesis, embryonic implantation, and germ and stem cell migration during embryonic development (Rostene et al, 2007; Salamonsen et al, 2007; Beider et al, 2008; Li and Ransohoff, 2009). It is clear that chemokines not only regulate cellular migration of immune cells, but also the migration, proliferation and survival signals in multiple cell types. While chemokine receptors and ligands are expressed in amphibians and mammals, expression of chemokine ligands and receptors is highly conserved between humans and mice, making the mouse model an advantageous and extensively used system to study chemokine function in vivo and in vitro (DeVries et al, 2006; Zlotnick et al, 2006). The mouse model has been used to study the role of chemokines in normal physiologic responses and also during inflammatory diseases contributed in part by deregulated expression or activitiy of chemokines (Gillitzer and Goebeler, 2001; De Paepe et al, 2008). These studies have lead to promising developments in the treatment of inflammatory diseases such as rheumatoid arthritis (Vital and Emery, 2008). More recently, in vitro and in vivo studies have implicated several inflammatory chemokines including: CCL2, CCL5, CXCL1 and CXCL12 in the progression of various cancers. While development of chemokine receptor antagonists to inhibit chemokine signaling is currently a promising avenue in cancer therapeutics, the complex functions and mechanisms of chemokine signaling in the cancer microenvironment may complicate the effectiveness of these agents. It is therefore important to investigate and understand the functions of inflammatory chemokine signaling in cancer progression at the molecular, cellular and whole organism levels. This review will focus on the current knowledge regarding the role of inflammatory chemokine signaling in cancer progression and the current status of chemokines in cancer therapy and prognosis.

II. Structure and signal transduction

Chemokines are a large family of proteins of which over 40 ligands have been identified. The chemokines have been subdivided into classes depending on the spacing of the cysteine amino acid residues at the NH₂ terminus; an X denotes a non- cysteine amino acid residue separating the otherwise conserved cysteine motif. Current classes of the chemokine family are referred to as CC, CXC and CX3C (Tables 1A-C). While the CXC class of chemokines have been shown regulate recruitment of neutrophils and T cells, the CC chemokine regulate both T cells, B cells and recruitment of bone marrow derived cells including-monocytes and dendritic cells (Laing and Secombes, 2004; DeVries et al, 2006).

Table 1.

Systemic and alternative nomenclature for human and mouse chemokineligands and receptors.

A. CXCL family
Chemokine ligand Systemic name	Human	Mouse	Chemokine receptors
CXCL1	GRo-a, MGSA, SCYB1	Gro1, Mgsa, KC, Scyb1	Duffy
CXCL2	GRo-B, MGSA-b, SCYB2	GROb, MIP-2a, KC, Scyb2	Duffy
CXCL3	Gro-g/MGSA-g, SCYB3	GRO, MIP-2, KC, Scyb3	CXCR2
CXCL4	PF4, SCYB4	PF4, Scyb4	unknown
CXCL5	ENA-78, SCYB5	GCP-2, LIX, Scyb5	Duffy
CXCL6	GCP-2, SCYB6	GCP-2, Scyb6	CXCR2
CXCL7	NAP-2, SCYB7	unknown	Duffy
CXCL8	IL-8, SCYB8	unknown	, Duffy
CXCL9	MIg, SCYB9	Mig, Scyb9	CXCR3
CXCL10	IP-10, SCYB10	IP-10, CRG-2, Scyb10	CXCR3
CXCL11	I-TAC, SCYB11	I-TAC, Scyb11	CXCR3
CXCL12	SDF-1 a/b, SCYB12	SDF-1, PBSF, Scyb12	CXCR4
CXCL13	BCA-1, SCYB13	BLC, Scyb13	CXCR5
CXCL14	BRAK, bolekine, MIP-2g, SCYB14	Scyb14	unknown
CXCL15	Lungkine, WECHE, SCYB15	Lungkine, WECHE, Scyb15	unknown
CXCL16	SR-PSOX	Zmynd15	CXCR6

B. CCL family
Chemokine ligand Systemic name	Human	Mouse	Chemokine receptors
CCL1	I-309, SCYA1	TCA-3, P500, Scya1
CCL2	MCP-1/MCAF, TDCF, SCYA2	JE. Scya2	Duffy,
CCL3	MIP-1a, SCYA3	MIP-1a, Scya3
CCL4	SCYA2, SCYA4, MIP1B	MIP-1b, Scya2, Scya4
CCL5	RANTES, SCYA5	RANTES, Scya5	Duffy,
CCL6	Mrp-1, SCYA6	C10/MRP-1, Scya6	CCR3,
CCL7	MCP-3, MARC, FIC, SCYA7	MARC, Scya7, mcp3	, , CCR3,
CCL8	MCP-2, SCYA8	mcp2, Scya8, HC14	, , CCR3
CCL9/10	unknown	MRP-2, CCF18, MIP-1g, Scya9	CCR3,
CCL11	Eotaxin, SCYA11	Scya11	CCR3
CCL12	SCYA12	MCP-5, Scya12	CCR2
CCL12	MCP-4, SCYA13, NCC-1	unknown	CCR3
CCL14	HCC-1, NCC2, SCYA14	unknown	, CCR3
CCL15	HCC-2, MIP-1d, MIP-5, SCYA15	unknown	CCR3
CCL16	HCC-4, LEC, LCC-1, SCYA16	unknown	,
CCL17	TARC, SCYA17	TARC, ABCD-2, Scya17
CCL18	DC-CK, PARC, AMAC-1, SCYA18	unknown	unknown
CCL19	MIP-3b, ELC, exodus-3, SCYA19	MIP-3b/ELC, exodus-3, Scya19	CCR7
CCL20	MIP-3a, LARC, exodus-1, SCYA20	Scya20	CCR6
CCL21	6CKine, SLC, exodus-2, SCYA21	6CKinase, SLC, exodus-2, TCA-4	CCR7
CCL22	MDC, STCP-1, SCYA22	ABCD-1, Scya22
CCL23	MPIF-1, CKb8, SCYA23	unknown	CCR1
CCL24	Eotaxin2, MPIF-2, SCYA24	MPIF2, Scya24	CCR3
CCL25	TECK, SCYA25	TECK, Scya25	CCR9
CCL26	Eotaxin-3, SCYA26	eotaxin-3, Scya26	CCR3
CCL27	CTACK/ILC, SCYA27	unknown	CCR10
CCL28	MEC, CCK1, SCYA28	MEC, Scya28	CCR3/CCR1

C. XCL and CX3CL families
Chemokine ligand Systemic name	Human	Mouse	Chemokine receptors
XCL1	lymphotactin/SCM-1a/ATAC	lymphotactin	XCR1
XCL2	SCM-1b	unknown	XCR1
CX3CL1	fractaline	neurotactin/ABCD-3	CX3CR

Chemokines signal to seven transmembrane G coupled receptors at the N terminus, resulting in phosphorylation of serine/threonine residues at the C-terminus, conformational changes to the receptor and activation of a heterotrimeric G protein complex bound to the receptor intracellular domain. Activation involves GTP binding to Ga subunit leading to disassociation from its Gb and g subunit partners, and subsequent activation of downstream signaling pathways (Figure 1). These signaling pathways, including: PI-3 Kinase, Rho family of GTPases and MAPK regulate cellular processes such as proliferation, motility and gene expression of matrix metalloproteinases and cytokines (Ganju et al, 1998; Bug et al, 2002; Chinni et al, 2006; Ou et al, 2006). It should be noted that chemokines receptors also activate signaling pathways independent of G proteins, including p38MAPK (Goda et al, 2006) and JAK/Stat (Vila-Coro et al, 1999) to regulate cellular processes such as migration and gene transcription.

Figure 1.

G protein dependent signal transduction through chemokinereceptors: CXCR4 signaling in lymphocytes as a model

The ability of multiple chemokines to bind to the same receptor and the ability of a single chemokine to bind to multiple receptors (Table 1) create the possibility of redundant signaling. In vitro studies have shown that chemokines stimulate migration through common signaling pathways such as G coupled protein dependent mechanisms (Cotton and Claing, 2009). However, there is also evidence supporting the possibility of unique functions for each chemokine ligand/receptor pair. First, chemokine ligands exhibit different binding affinities to the same receptor. For example, chemokines exhibit a greater affinity for CXCR2 versus CXCR1 (Devalaraja and Richmond, 1999; Rajagopalan and Rajarathnam, 2006). In addition, different ligands which bind to the same receptor exert different biological effects in certain cell types. For example, CCL3, CCL4 and CCL5 have been shown stimulate migration of activated T cells, but only CCL5 can stimulate migration of resting T cells (Taub et al, 1993). Moreover, knockout mouse studies of chemokine receptors and ligands such as CCL2 and CCR2 do not show compensatory upregulation of other ligands (Boring et al, 1997; Huang et al, 2001). The purpose of the multiple ligand/receptor binding pairs remains under investigation, but these studies indicate that unique roles for each chemokine/receptor pair may serve as a mechanism to regulate cellular responses to chemokine signaling.

Other multiple mechanisms also exist to regulate chemokine signaling. Continuous chemokine signaling has been shown to lead to receptor desensitization, internalization through b arrestin and clathrin dependent mechanisms, resulting in downregulation of chemokine signaling (Aramori et al, 1997; Oppermann, 2004). In addition, the D6 and Duffy receptors which bind multiple ligands (Table 1), do not appear to activate signaling pathways, indicating that these receptors may act to sequester chemokine ligands as additional regulatory mechanisms for down regulation of chemokine signaling (Locati et al, 2005; Comerford et al, 2007). In fully developed mammalian organisms, these mechanisms are vital to controlling immune responses during inflammation to restore normal tissue homeostasis.

III. Role of chemokine signaling in inflammation and cancer

A. Phenotypes of knockout mouse models

Chemokines appear to play dual roles in the inflammatory process. On the one hand, targeted deletion of CC and CXC chemokine receptors or ligands in mice leads to increased susceptibility to viral or bacterial infections and decreased clearance of pathogens due to diminished immune cell recruitment. On the other hand, targeted deletion of chemokines and chemokine receptors have been shown to alleviate the phenotypes of inflammatory diseases including rheumatoid arthritis, multiple sclerosis, autoimmune encephalitis and macular degeneration (Table 2).

Table 2.

Targeted deletion of inflammatory chemokineand chemokinereceptors in mice

Gene	phenotype	citation
CCL2	diminished macrophage and TH1 T cell responses inexperimental autoimmune encephalomyelitis (EAE), no compensatory upregulation of CCR2 binding ligands	Huang et al, 2001
CCR2	decreased formation of lung granulomas induced by mycobacterium bovis, similar phenotype in EAE as ccl2-/-, decreased monocyte recruitment to inflamed tissues	Boring et al, 1997; Fife et al, 2000; Tsou et al, 2007
CCL5	defects in T cell proliferation, an overall reduction in T cell activation and recruitment in cutaneous delayed-type hypersensitivity assays	Makino et al, 2002
CCR5	decreased NK cell mobilization in mice infected with herpes simplex virus 2 leading to decreased survival, increased NK cell infiltration in experimental mouse model of colitis correlating with increased resistance to disease	Yamaoka et al, 1998; Andres et al, 2000; Thapa et al, 2007
CXCL1	artherosclerotic lesions leads to a reduction in macrophage recruitment associated with decreased lesion formation	Boisvert et al, 2006
CXCR2	decreased neutrophil recruitment and delayed wound healing responses, abnormal granulocyte differentiation, mice develop splenomegaly	Cacalano et al, 1994; Devalaraja et al, 2000
CXCL12	perinatal lethality, defects in B cell development in in fetal liver and bone marrow, reduction in myeloid progenitors in bone marrow, defects in cardiac development	Nagasawa et al, 1996
CXCR4	perinatal lethality due to multiple defects including decreased bone marrow cell and B cell development abnormal cerebellum morphology	Ma et al, 1998

Chronic chemokine signaling is associated with macrophage and T cell accumulation at the inflammatory site, and studies indicate that chronic activation of macrophages may lead to alterations in normal tissue architecture, abnormal angiogenesis and DNA damage due to excess secretion of reactive oxygen species (ROS); (Gillitzer and Goebeler, 2001; Moll et al, 2009).

While the causes for chronic chemokine signaling are still under investigation, gene mutations in chemokine receptors in immune cells (Hernandez et al, 2003) and increased expression of cytokines which regulate chemokine expression at the site of inflammation (Firestein, 2004) may be important contributing factors.

These studies indicate that that loss of control over chemokine signaling results in long range consequences with damaging changes in the tissue microenvironment. Emerging studies indicate that de-regulation of the activity and expression of chemokine ligands and receptors in cancer may also alter the tumor microenvironment with long term consequences.

B. Expression patterns in cancer

Many studies indicate a significant correlation of elevated expression of chemokines and signaling receptors with poor cancer prognosis and lymph node metastases in various cancers as determined by immunohistochemistry and RNA in situ analyses (Table 3). Interestingly, the same chemokine receptors that show chronic signaling inflammatory diseases are upregulated in cancer, including CCL2, CCL5, CXCL1 and CXCL12.

Table 3.

Protein expression of commonly found chemokinesand chemokine receptors in cancer epithelium.

	Breast	colorectal	brain	melanoma	prostate	non-small cell lung	ovarian	thyroid	pancreatic	renal cell
CCL2	+1	+	+6	+	vl	nd	-	nd	+	nd
CCR2	nd	vl	+6	nd	+	+	nd	nd	vl	nd
CCL5	+1,2	nc	nd	+	+	+	nd	+9	+	nd
CCR5	+3	nc	+7	+	nd	+	nd	nd	nd	nd
CXCL1	+4	+	+6,7	+	+	nd	+	nd	nd	+
CXCR2	nd	+	+7	+	+	+	+	+10	+	+
CXCL12	v11,4	+	+7	nd	+	nd	+	nd	+	+
CXCR4	+1,5	+	+6	+	+	+	+	+9	+	+

(-): reduced expression, (+): increased expression, nd: not determined, vl: very low or undetectable; nc: no change

¹invasive ductal carcinoma

²estrogen receptor negative (ER negative)/progesterone receptor negative stage II breast cancer

³infiltrating ductal carcinoma and infiltrating lobular carcinoma

⁴ER negative breast cancer

⁵invasive lobular carcinoma

⁶glioblastoma

⁷glioma,

⁸astrocytoma

⁹papillary thyroid carcinoma

¹⁰medullary thyroid carcinoma

Epithelial specific expression of chemokines and chemokine receptors have been validated by protein and RNA expression analysis of breast, renal, prostate cancer cell lines (Schrader et al, 2002; Vaday et al, 2006; Wente et al, 2008). In addition, decreased expression of Duffy and D6 decoy receptors in breast cancer inversely correlate with lymph node metastases and increased survival rates, indicating the cancer cells showed impaired ability to down regulate chemokine signaling (Ou et al, 2006; Wu et al, 2008). These studies associate increased expression of chemokines and decreased expression of decoy receptor with invasive cancer. However, recent studies show loss of CCL2 in ovarian cancer (Arnold et al, 2005) and reduced CCR2 expression in patients with myeloma (Van de Broek et al, 2006) indicate a possible tumor suppressive role for chemokine signaling which may in part be tissue dependent. Studies have also found associations between the patterns of chemokine expression and levels of immune cells in the primary tumor, as demonstrated by studies of CCL2. In breast cancer, increased CCL2 expression is associated with increased levels of tumor associated macrophages, which have been found to correlate with the invasive phenotype and poor cancer prognosis (Valkovic et al, 1998; Ueno et al, 2000). However, in cervival cancer, loss of CCL2 expression is associated with decreased levels of macrophages in cervical cancer, and correlates with poor cancer prognosis (Kleine-Lowinski et al, 1999). Taken together, these studies demonstrate a tissue dependent pattern of expression and a tissue dependent role for chemokines and immune cells in cancer progression.

Recent studies have shown that the surrounding tumor stroma also show significant changes to chemokine expression. In particular, increased expression of CXCL1 was consistently observed in the stroma of multiple types of breast cancers by microarray analysis of tumor associated stroma, correlating with lymph node metastases, invasiveness and poor patient survival (Finak et al, 2008). Recent studies have shown that CXCL1 and CXCR2 may play a role in regulating replicative senescence in fibroblasts through a p53 dependent mechanism as a possible means to suppress tumor formation. Furthermore, gene mutations in components of the CXCL1 signaling pathway may allow tumor cells to escape this tumor suppressive mechanism (Acosta et al, 2008). In addition to CXCL1, increased expression of CXCL12 have been reported in the stroma of basal and invasive mammary ductal adenocarcinoma, head and neck cancer, papillary thyroid carcinoma and squamous cell carcinoma by SAGE, differential display and immunohistochemistry studies (Frederick et al, 2000; Shellenberger et al, 2004).

The increased expression of these chemokines in the stroma correlates with tumor size and lymph node invasion (Kleer et al, 2008; Oliveira-Neto et al, 2008). These studies indicate that stromal specific expression of chemokines may be an important factor to consider when determining cancer prognosis.

While changes in protein or RNA expression of chemokine receptors and ligands are strongly associated with progression of various cancer types, few studies of serum levels and polymorphisms have currently yielded significant links to cancer. Elevated serum levels of CCL5 correlate with poor patient prognosis in ovarian cancer (Tsukishiro et al, 2006) and breast cancer (Niwa et al, 2001). While serum levels of CCL5 had no significant association with patient survival in gastric cancer, high serum levels of CCL5 correlated significantly with invasive disease.

Serum level studies have involved comparisons between healthy vs. cancer patient samples, as well as malignant vs. benign samples. The differences in results in serum level studies may be due to differences in patient samples in terms age, disease stage, and whether or not disease was treated prior to sample collection. Future comparative analyses must consider such variables, as well as the possibility that changes in chemokine levels may be restricted to the local tumor environment. Therefore the prognostic value of serum analyses might be lower than analysis of expression within the tumor/stroma itself.

C. Regulation of chemokine expression

As a number of studies have reported alterations in the expression of chemokines and chemokine receptors in various cancers, current studies are underway to determine the possible causes for the changes in expression during cancer progression. Certain environmental and soluble factors have been shown to induce expression of chemokines and chemokine receptors in epithelial and mesenchymal cells. Hypoxic regulated factors and hepatocyte growth factor (HGF) were shown to induce expression of CXCR4 in MCF-7 cells and MDA-MB-231 cells (Matteucci et al, 2007), while TNF-a enhanced CXCR4 expression in ovarian cancer cells (Kulbe et al, 2005). Multiple cytokines and growth factors including, TNF-a, IL-6, MSP, LMP1, CD40 were shown to induce expression of CCL2 and CCL5 in epithelial cells, macrophages, fibroblasts and endothelial cells (Biswas et al, 1998; Buettner et al, 2007). The possibility of genetic mutations as a possible cause for abnormal chemokine expression is currently under investigation. Loss of heterozygosity of CCL2 in cervical carcinoma at 17q11.2 has been associated with diminished progression, increased survival, and macrophage accumulation (Zijlmans et al, 2006). Gene variants of chemokines and receptors have been detected in one study in the form of base substitutions or base deletions: CCL5 -403(G>A), CXCL12 +801(G>A), CCR2 V64I (G>A), CCR5 (Delta32), though only CXCL12 +801(G>A) has been found to be associated with increased prostate cancer risk (Petersen et al, 2008). In addition, the polymorphism CCL5 -403(G/A) has been associated with increased susceptibility of prostate cancer (Saenz-Lopez et al, 2008).

Studies of CXCL1 signaling have suggested that inactivating point mutations to CXCR2 (CXCR2G354W allele) in lung adenocarcinoma may be a means to escape CXCL1 induction of cellular senescence, and consequently tumor suppression (Acosta et al, 2008)

While these few studies have detected a significant association with cancer risk or invasiveness, the functional significance to these gene variants remain largely unclear. Further studies should be conducted to investigate and understand the functional mechanisms of these gene variants during cancer progression.

In summary, studies analyzing differences in local expression of chemokines, compared to those measuring chemokine levels in the serum, offer a stronger, more consistent association is between cancer prognosis and changes in patterns of chemokines ligands and receptors in the primary tumor and metastatic lesions. Currently, the genetic link between chemokine expression patterns and cancer remains unclear. As a number of soluble factors can regulate expression of chemokine ligands and receptors, it is more likely that alterations in expression of these regulators combined with possible genetic mutations may contribute to de-regulated expression of chemokines and their receptors.

D. Role of chemokines in tumor progression in the mouse model

The use of transgenic and transplantion mouse models of cancer have proven invaluable in understandng the functions and significance of chemokine signaling during disease progression. A tumor promoting role for inflammatory chemokines is best demonstrated in studies of breast cancer. Over expression of CXCL1 in breast cancer cells results in increased tumor growth and lung metastasis when these over expressing cells are grafted in the mammary fat pads of nude mice (Li and Sidell, 2005; Vazquez-Martin et al, 2007). In addition, siRNA mediated knockdown of CXCR4 expression in mouse and human mammary carcinoma cells inhibits tumor growth and metastatic spread in nude mice (Smith et al, 2004; Liang et al, 2005). CCL2 knockout mice bred to HER2/neu transgenic mice produce progeny exhibiting slower mammary tumor growth and longer tumor latency (Conti et al, 2004). While these studies indicate that expression of chemokines in the tumor epithelial is sufficient to promote breast cancer progression in mice, mesenchymal stem cells have been shown to enhance cancer progression in nude mice through CCL5, CXCL12, CXCR4 and CCL2 dependent mechanisms (Karnoub et al, 2007) indicating these chemokine also regulate breast cancer progression through mediation of stromal: epithelial interactions.

The same chemokines shown to play tumor promoting roles in breast cancer affect other types of cancer differently. Targeted deficiency in expression of CCR2 in the XK14-HPVR(2) mouse model of cervical carcinoma does not significantly affect tumor angiogenesis or end stage progression even though decreased numbers of macrophages have been observed in the primary tumor. The modest phenotype may be due to increased presence of tumor associated neutrophils observed in the CCR2 deficient mice; these cells have been proposed to exert tumor promoting effects to balance out any tumor suppressive effects caused by CCR2 deficiency (Pahler et al, 2008). In addition, CCL2 over expression in colon carcinoma cells and rat gliosarcoma cells suppress tumor formation when these tumor cells are injected into rodents (Hoshino et al, 1995). This anti-tumor phenotype correlates with increased accumulation and activation of macrophages at the site of injection in immunocompromised mice (Rollins and Sunday, 1991) and increased T cell responses in immunocompetent mice (Manome et al, 1995). Moreover, over expression of CCL5 in the thymoma cell lines EL4 or EG 7 has been shown to inhibit tumor growth when injected in mice, characterized by increased recruitment of T cells, natural killer (NK) cells, and dendritic cells. These studies indicate that enhanced chemokine signaling in cancer in part drives immune cell recruitment to the primary tumor. The ability of immune cells to mount an anti-tumor response may in part be tissue type dependent, and dependent on the mouse model used, as immune cells appear to be more effective in preventing establishment of the primary tumor in mice with intact immune systems as opposed to immunocompromised mice. In addition, studies have reported tumor-promoting roles for antigen presenting cells including macrophages and dendritic cells in immunocompromised mouse models and transgenic mouse models of cancer, which may in part be regulated by signals in the tumor microenvironment (Pollard, 2004; Melief, 2005; Allavena et al, 2008).

E. Role of chemokine signaling in epithelial and tumor associated stromal cells

Studies have revealed that chemokines regulate growth and migratory signals in multiple types of epithelial cells. Induction of chemotaxis has been observed in a number of carcinoma cell lines including those of breast, prostate, melanoma and lung (Prest et al, 1999; Woodward et al, 2002, Loberg et al, 2006; Huang et al, 2008). Chemokines were also shown to regulate growth of certain carcinoma cell lines. CXCL12, CCL2, CCL5 and CXCL1 were shown to stimulate proliferation of melanoma, glioma, and prostate cancer cell proliferation (Payne and Cornelius, 2002; Darash-Yahana et al, 2004; Loberg et al, 2006; Vaday et al, 2006). These same chemokines appear to mainly stimulate migratory signals in breast cancer cells; however, CXCL12 has been shown to stimulate proliferation of breast cancer cells (Allinen et al, 2004). These studies indicate that the abilty of chemokines to stimulate cell proliferation may be cell type specific.

In addition to signaling in epithelial cells, chemokines have also been shown to regulate the functions of various mesenchymal cell types. The ability of chemokines to stimulate chemotactic responses of T cells and antigen presenting cells using in vitro assays is well established (Zabel et al, 2006). Additional studies have shown that CCL2, CXCL1, CCL5, CXCL12 function as potent angiogenic factors, stimulating endothelial cell migration and tube formation, as well as angiogenesis in corneal and chick cam angiogenesis assays (Goede et al, 1999; Salcedo et al, 2000; Azenshtein 2002; Dhawan and Richmond, 2002). Recent studies have also shown that chemokines can also regulate migration of lung fibroblasts through a CXCR2 dependent mechanism (Wislez et al, 2006); however the significance of chemokines on the regulation of fibroblast behavior currently remains unclear. These studies demonstrate that chemokines can regulate the functions of multiple types of mesenchymal cells with important implications on the functions on these cells in the tumor microenvironment.

Expression of chemokines in the tumor-associated stroma indicates that chemokines may regulate stromal: epithelial interactions in the primary tumor microenvironment. This hypothesis is supported by studies of carcinoma cells interactions with various stromal cell types. Increased expression of CXCL1 in melanoma-associated fibroblasts indicated that CXCL1 may regulate paracrine signaling interactions between fibroblasts and melanoma cancer cells to promote cancer cell proliferation and migration (Gallagher et al, 2005). In lung adenocarcinoma cells that showed increased expression of CXCR2, co-culture with lung stromal cells including macrophages, endothelial cells and fibroblasts increased carcinoma cell proliferation and enhanced tumor growth in mice, through a CXCL1/CXCR2 dependent mechanism (Zhong et al, 2008).

Studies have also shown that stromal: epithelial interactions regulated by chemokines signaling may also have important implications in metastatic spread. To study the mechanisms of bone metastases during prostate cancer, human prostate cancer cells were co-cultured with normal human bone fibroblasts in a 3-D co-culture model. These experiments resulted in long term changes in gene expression patterns in the bone fibroblasts including increased expression of CCL5, CXCL5 and CXCL15. Moreover, co-culture of these carcinoma-associated bone fibroblasts enhanced tumor progression of a benign prostate cell line in mice, associated with increased chemokine expression (Sung et al, 2008), indicating that carcinoma cells can exert long term tumor promoting changes in stromal cells associated with enhanced chemokine signaling. In other studies, increased expression of CCL2 in human bone marrow endothelial cells was shown to enhance prostate cancer cell proliferation and migration associated with increased Rac1 signaling when co-cultured in vitro (van Golen et al, 2008) with important implications in the mechanism of transendothelial migration of prostate cancer cells to the bone. In other studies, mesenchymal stem cells co-cultured with human breast cancer cell lines including: MCF-7, T47D and MDA-MB-231 cells enhanced cancer cell migration in vitro through CCL5, CCL2 and CXCL12 dependent mechanisms with implications in bone metastases as well as invasiveness of the primary tumor (Karnoub et al, 2007; Corcoran et al, 2008; Molloy et al, 2009). Furthermore, CXCL12 and its receptor CXCR4 have been shown to regulate homing of receptor expressing cancer cells to tissues where non-malignant stromal cells express CXCL12 (Muller et al, 2001). These studies demonstrate that tumor cell metastasis is not random, but guided by the expression of chemokine receptors and adhesion molecules on the neoplastic cells.

IV. Current use of chemokines in cancer therapy

As the importance of chemokine expression becomes more recognized in the growth, invasion and metastasis of cancer, it also becomes necessary to develop inhibitors of these chemokines or their related receptors. There have been many studies performed using various inhibitors of chemokines and chemokine receptors and recently have shown efficacy in the treatment of cancer of solid tissues in preclinical models. Currently, a few of these studies have proceeded to clinical trials towards patient therapy.

One promising avenue in the treatment of metastatic disease in solid tumors is in the targeting of the CXCL12 and its receptor CXCR4, which have been shown to regulate homing of cancer cells to distant sites as best demonstrated in breast cancer (Muller et al, 2001). One of the most widely studied compounds is AMD3100 which is thought to specifically block CXCR4 signaling (Burger et al, 1999). This compound has shown efficacy in murine models as it delays pulmonary metastases of mammary carcinoma cells (Smith et al, 2004), reduces dissemination of ovarian cancer cells (Kajiyama et al, 2008) and reduces gastric cancer tumor growth (Yasumoto et al, 2006). While the mechanism of action for AMD3100 is currently under investigation, AMD3100 has been shown to inhibit CXCL12 stimulated migration of breast (Cabioglu et al, 2005), ovarian (Scotton et al, 2002) and gastric cancer cells (Ohira et al, 2006) as well as decrease the invasiveness of prostate cancer cells (PC3 cell line) (Zhang et al, 2008). This compound has been also been tested in vitro studies with other cell types including, pancreatic, colorectal, osteosarcoma and, malignant melanoma (Burger and Peled, 2009). These studies indicate that AMD3100 inhibits tumor progression, in part by inhibiting tumor epithelial cell proliferation, migration and invasion. Recent studies have shown that AMD3100 has mobilizes hematopoietic cells into the peripheral blood (Azab et al, 2009). The increased mobilization renders leukemia and myeloma cells more sensitive to chemotherapeutic intervention by decreasing their association with the bone marrow (Nervi et al, 2008). Thus, the potential benefits of AMD3100 are currently being evaluated in combination with other chemotherapies in patients with blood cancers (Cashen et al, 2008; Stewart et al, 2009). However, the effects of AMD3100 on hematopoietic cells in solid tumors remain unclear. Given the contribution of hematopoietic cells to the progression of solid tumors (Pollard, 2004; Karnoub et al, 2007), the functional consequences of AMD3100 in the tumor microenvironment should be further investigated.

Other inhibitors to the CXCL12/CXCR4 pathway have shown promising results in pre-clinical models and in early clinical trials. CTCE-9908, which is a peptide analog of CXCL12 and an active inhibitor of the ligand, has shown promising results as a well tolerated drug that stabilized disease in early clinical trials for late stage cancer patients (Hotte et al, 2007; Evans, 2008). In pre-clinical studies, this compound has been used in the treatment of osteogenic sarcoma in mice and has been found to decrease growth, adhesion, migration, and invasion in osteosarcoma cells in vitro (Kim et al, 2008). When these cells were then injected into the tail veins of mice, treatment with CTCE-9908 resulted in a 50% reduction in the number of gross metastatic lung nodules and a marked decrease in micro-metastatic disease. Similar results have also seen with melanoma cells, but only when the cells were pre-treated with the inhibitor before injection (Kim et al, 2008). While CTCE-9908 specifically targets CXCL12 and AMD3100 is inhibits CXCR4, both compounds inhibited metastatic spread of various cancers in preclinical models, underscoring the importance of CXCL12/CXCR4 signaling in metastatic disease.

Another chemokine receptor/ligand pair that has been studied in the growth and metastasis of cancer is CCL5/CCR5. Like CXCL12/CXCR4, this pair of proteins has been implicated in the growth and metastasis of many types of cancers, including breast cancer and multiple myeloma. One study showed that monoclonal antibodies to CCR5 significantly blocked CCL5 signaling in MDA-MB-231 breast cancer cells enhanced by mesenchymal stem cells in vitro. Furthemore, systemic treatment of tumor bearing mice with anti -CCR5 inhibited metastatic spread of MDA-MB-231 cells (Karnoub et al, 2007). These studies indicate that CCR5 antagonists significantly block interactions between tumor epithelial cells and mesenchymal stem cells during tumor progression. In other studies, daily systemic treatment of a functional antagonist to CCL5 (met-RANTES or met-CCL5) slowed tumor growth of 410.4 breast cancer cells transplanted into mice, and also reduced macrophage infiltration into the tumor (Robinson et al, 2003). However, Met-CCL5 was unable to reduce the infiltration of other immune cells such as neutrophils (Robinson et al, 2003), which have also been shown to play tumor promoting roles in breast cancer (Queen et al, 2005). One possibility is that these cells express other chemokine or cytokine receptors that could respond to other ligands in the tumor microenvironment while another possibility is that other CCR5 binding ligands could compensate for lack of CCL5 function. Thus, it is possible that redundant signaling of chemokine receptors could reduce the efficacy of these chemokine antagonists in cancer. Indeed, these factors may be partly responsible for the lack of improvement reported in Phase IIA clinical trials for the treatment of rheumatoid arthritis (Vergunst et al, 2008). These studies indicate that inhibiting stromal: epithelial interactions mediated by CCL5 represent one potential approach to treating metastatic disease. Another CCR5 antagonist, TAK-779 has been found to inhibit CCL5-induced prostate cancer cell invasion in a concentration dependent manner (Vaday et al, 2006), indicating that CCR5 antagonists also affects autocrine signaling in cancer cells. However, the mechanism of this compound and efficacy in other types of cancers in in vitro and in vivo studies still remains unclear. Taken together, these studies indicate that the CCL5/CCR5 represents another potentially significant chemokine signaling pathway to target in cancer therapeutics.

Other chemokine antagonists have also shown potential for clinical application in cancer treatment. For example, the CXCR2-selective antagonist AZ10397767 has been shown to decrease resistance of androgen-independent prostate cancer cells to oxaliplatin through an NF-kB dependent mechanism (Wilson et al, 2008). Co-administration of this inhibitor with TRAIL increased the sensitivity of PC3 cells, and was most likely due to the ability of AZ10397767 to block TRAIL and IL-8-induced upregulation of c-FLIP (Wilson et al, 2008). These studies indicate potential benefits in combining chemokine antagonists with other therapies.

In addition to pharmacologic inhibitors that have been designed to specifically target chemokines and their receptors, recent studies have also found that pharmacologic inhibitors used to target other molecules also act as chemokine inhibitors. For example, Etanercept (Enbrel), a TNF-a inhibitor normally used to treat arthritis, has been shown to significantly decrease systemic levels of CCL2 in patients with metastatic breast cancer (Madhusudan et al, 2004). In addition, tamoxifen treatment has been found to inhibit CCL2 mRNA and protein expression in endometrial cancer cell line (EFE184) (Wang et al, 2006). These studies reveal that therapies intended to target specific molecules also indirectly affect the expression of other molecules. However, these unintended side effects may be exploited towards the treatment of cancer, as CCL2 as well as many other chemokines have been shown to play tumor promoting roles in different cancers (Table 3). It is possible that current drug therapies used to treat inflammatory diseases or particular types of cancers could function as inhibitors of chemokine signaling and be redirected towards the treatment of other cancers. However, this avenue of treatment requires further investigation and understanding how conventional therapies affect expression and function of chemokines at the molecular and cellular levels. In summary, while few chemokine antagonists have proceeded to clinical trials (Table 4), the development of chemokine antagonists for the treatment of cancer remains a promising avenue to explore.

Table 4.

Current use of chemokinesin cancer therapy

Chemokine Target	Antagonist/Agonist	Cancer	Clinical Trial Status
CXCR4	AMD3100	Leukemia and Lymphoma	Phase II
CXCL12	CTCE-9908	Late stage ovarian, breast, lung, colorectal, melanoma, gastric	Phase I/II
CCL5	Met-CCL5	Breast Cancer	Pre-clinical
CCR5	TAK-779	Multiple Myeloma, Prostate	Pre-clinical
CXCR2	AZ10397767	Prostate	Pre-clinical
	SB225002	Esophageal
CCR2	MLN1202	Breast	Pre-clinical

V. Summary and Conclusions: Future role of chemokines in the clinical setting

It is apparent that chemokine signaling regulates multiple processes during tumor progression including primary tumor growth, tumor angiogenesis and metastatic spread. Based on current studies, chemokines mediate stromal: epithelial interactions in the primary tumor microenvironment to regulate tumor growth and invasion (Figure 2A). In addition, studies also indicate that chemokines also regulate homing of chemokine receptor expressing cancer cells to distant sites in metastatic disease (Figure 2B).

Figure 2.

Model for the functional roles of inflammatory chemokinesignaling in cancer. (A) Stromal and epithelial cells secrete inflammatory chemokines, contributing to the bioavailability of chemokinesin the local tisuemicroenvironment. Inflammatory chemokinesregulate autocrineand paracrinesignaling interactions between stromal and epithelial cells to regulate cellular proliferation, migration and invasion. (B) Expression of Inflammatory chemokinesin normal tissues, as demonstrated by CXC112 and CCL5 create a chemokinegradient which regulate homing of chemokinereceptor expressing cancer cells to other organs.

Studies of chemokine expression in the cancer stroma and cancer stage, grade and patient survival and indicate these expression patterns may be indicative of host responses towards the tumor.

Chemokine expression has correlated with the presence of immune cells in some cancers representing either a positive or poor progonosis in cancers (Kleine-Lowinski et al, 1999; Ueno et al, 2000) and may be dependent on the tissue type.

The ability of immune cells to mount anti- tumor response may be in part dependent on the soluble factors expressed in the tumor microenvironment (Lewis and Pollard, 2006). Thus, by regulating the recruitment of immune cells to the primary tumor, the expression of chemokines in cancer may have multi-level consequences on cancer progression.

While further studies should be investigated on the functional consequences of chemokine signaling on immune cells recruitment in cancer, the strong correlations of cancer stromal specific expression of chemokines and tumor malignancy indicate that a practical application for chemokines may be as markers for cancer prognosis.

While studies have shown that chemokine antagonists are effective in treating invasive cancer in mouse models, there are many issues associated with therapeutic targeting of chemokines that have to be addressed and fully understood prior to their use in the treatment of cancer. The possibility of cytotoxic effects in normal tissues is always a consideration in the delivery of chemotherapeutic drugs which target specific molecules (Chari, 2008; Wysocki et al, 2008). This issue is an important consideration particularly for chemokines and their receptors, which are expressed in normal and cancer tissues. The promiscuous binding of ligands to multiple receptors may further complicate the ability of chemokine antagonists to specifically target cancer tissues. While some chemokine ligand/receptor pairs exhibit unique functions, as demonstrated by CCL2/CCR2 signaling, the possibility of redundant signaling may further complicate the effectiveness of chemokine antagonists in cancer.

In summary, while current studies indicate promising roles for chemokines in the clnical setting, the sheer number of chemokines, pleiotropic effects of chemokines, and complex mechanisms of signal transduction add further questions regarding the functions of chemokines in cancer. What signaling pathways do chemokines activate to regulate cellular proliferation and migration, and are these signaling pathways activated in a cell type dependent context? How do chemokines regulate the dynamic interactions between different types of immune cells in the tumor microenvironment? How do chemokines coordinate with other signaling pathways in the tumor microenvironment to regulate tumor progression? Answering these and other questions through further study will lead to a greater understanding of the role of chemokines in tumor progression, and will enable the design of more effective applications for chemokines in the clinical setting.

Abbreviations

c-FLIP

Caspase-8 homologue FLICE-Inhibitory Protein

LMP1

Latent Membrane Protein 1

MSP

Macrophage Stimulating Protein

NK

natural killer

ROS

reactive oxygen species

RANTES

Regulated upon Activation Normal T cell Expressed and Secreted

siRNA

small interfering RNA

TNF-α

Tumor Necrosis Factor α