| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print February 14, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Dermatology Branch, Center for Cancer Research, National Cancer Institute, Bethesda, Maryland
1Correspondence: Dermatology Branch, Center for Cancer Research, National Cancer Institute, Bldg. 10/Rm. 12N238, 10 Center Dr., Bethesda, MD 20892. E-mail: hwangs{at}mail.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
Key Words: PI-3K • Akt • carcinoma • melanoma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
Recent evidence (as noted in the following review articles) indicates that members of the chemoattractant cytokines, more popularly known as chemokines, and their receptors may play critical roles in several of the critical steps in tumorigenesis and/or metastasis [3 4 5 6 ]. To some degree, specific chemokines (and their receptors) have been shown to play potentially important roles in many of the critical steps of the metastatic process (Fig. 1 ). The function of chemokines is known in considerable detail for leukocytes, which use these proteins to arrest at inflamed blood vessels and as guides to lead them to specific sites of inflammation. The paradigm of chemokine function for leukocytes in terms of site-specific homing at inflammatory sites also holds for the function of these chemokines in several of the key steps of metastasis. We will not provide an in-depth description of the known function of chemokines in other aspects of biology or human diseases apart from cancer, which has been detailed in several excellent reviews [7 , 8 ]. Instead, this review will emphasize what we believe are striking similarities between the roles of selected chemokines in the metastatic process and the roles of chemokines in the homing of leukocytes to inflamed tissues and/or secondary lymphoid tissues.
|
50). Chemokines interact with cell-surface receptors, which are members of a large superfamily of seven-transmembrane domain, G protein-coupled receptors. So far, 50 chemokines have been discovered, and at least 18 chemokine receptors have been described [9
]. Some chemokine receptors bind to multiple chemokines and vice versa, suggesting that certain redundancies exist in chemokine function. Indeed, when individual chemokines or their receptors have been deleted in mice, defects, with rare exceptions [e.g., CXC chemokine receptor 4 (CXCR4)], have been relatively subtle. To avoid confusing jargon and multiple investigator-specific names, the chemokines (and their receptors) are generally known by their systematic names, consisting of the family of the chemokine followed by the letter L for ligand or R for receptor and a number indicating their order of discovery [9
]. For example, the chemokine originally termed "stromal-derived factor 1" is now termed CXC chemokine ligand 12 (CXCL12), and its receptor is CXCR4. Chemokines were noted initially for their ability to stimulate directional migration of nearly all classes of leukocytes. Neutrophils, for example, migrate strongly in response to chemokines such as CXCL8 (interleukin-8) and eosinophils to CC chemokine ligand 11 (CCL11; eotaxin). T cells, depending on their functional characteristics [T helper cell type 1 (Th1) vs. Th2, naïve vs. memory, regulatory vs. effector, etc.], respond to a wide variety of chemokines. In addition to stimulating migration, chemokines increase the affinity and avidity of ß1 and ß2 integrins on leukocytes for their endothelial counter-receptors, including intercellular adhesion molecule-1 and -2 and vascular cell adhesion molecule-1 (VCAM-1) [10 ]. In vivo, chemokine-dependent integrin activation results in firm arrest of leukocytes on the luminal surfaces of blood vessels, prior to the diapedesis of these cells into inflamed tissues. Once in tissues, leukocytes use directional cues provided by localized sources of chemokines to move toward active areas of inflammation or infection. Limited evidence also suggests that chemokine receptor activation leads to increased resistance to cyclohexamide-induced apoptosis in T cells, possibly via activation of Akt and its downstream effectors [11 ]. The latter observation suggests that prosurvival functions may also play a role in the ability of locally derived chemokines to determine the composition of inflammatory cells at a particular site of inflammation or infection.
This review will highlight some of the similarities between the functions of chemokine receptors in physiologic homing of leukocytes and purported roles for these receptors in cancer metastasis. It is difficult to separate some aspects of tumorigenesis from metastasis, as primary tumor growth is correlated with metastasis in many cancers. Angiogenesis, for example, is required by primary as well as secondary tumors for continued growth. Therefore, we will divide this review into several sections, which will include separate descriptions of possible roles for chemokines in tumorigenesis and in metastasis. Within the space of this review, it is not possible to discuss all the available evidence regarding roles for chemokine receptors in organ-selective metastasis. Thus, a referenced summary of possible roles for chemokine receptors [other than CXCR4, CC chemokine receptor 7 (CCR7), and CCR10] in the metastasis of specific cancers is provided in Table 1 . With the development of specific chemokine receptor antagonists, it may be possible to exploit the vulnerability of cancer cells by disrupting chemokine receptor-mediated signaling and in doing so, directly inhibit tumor growth or render tumor cells more susceptible to traditional anticancer treatment modalities.
|
|
CHEMOKINES AND THEIR RECEPTORS IN TUMORIGENESIS |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
Ovarian cancer cells, for example, produce inflammatory chemokines such as CCL2 [16 ] and CCL5 [15 ]. CCL2 localized to epithelial areas within the tumors, and the level of CCL2 expression was correlated with large numbers of lymphocytes and macrophages in the same area [16 ]. Of note, intratumoral macrophages release a variety of factors, including matrix metalloproteinases (MMPs), which may increase the invasiveness of cancers and are clinically associated with poor outcomes in cancer patients [26 ]. Moreover, CCL2 may shift the local cytokine milieu toward one that is Th2-polarized [29 ]. In contrast to CCL2, CCL5 localized to tumor-infiltrating leukocytes, and the presence of CCL5 was correlated with infiltration of CD8+ T cells [15 ].
There is often a predominance of Th2 cells in tumors, and Th2 polarization may be a general strategy to reduce the immune response against tumors [18 ]. Hodgkin’s disease, for example, is characterized by the presence of a relatively small number of Reed-Sternberg (RS) cells within a larger number of inflammatory cells such as lymphocytes, eosinophils, fibroblasts, macrophages, and plasma cells [19 ]. RS cells have been shown to differentially express the chemokine CCL17 [20 ], which binds the lymphocyte receptor CCR4, a receptor associated with Th2 lymphocytes and atopy [21 , 22 ]. It is possible that attraction of Th2-predominant lymphocytes by RS cells contributes to survival of the tumor cells by suppressing Th1-mediated immunity. In addition, in Kaposi’s sarcomas, the human Kaposi sarcoma-associated herpes virus encodes three chemokines [macrophage-inflammatory protein I (MIPI), MIPII, and MIPIII], which selectively attract Th2 cells [23 ]. Thus, tumor cells may evoke an immunosuppressive microenvironment by selectively producing chemokines that attract Th2 cells.
Banchereau and colleagues have described the presence of intratumoral CD1a+, immature ("nonactivated") dendritic cells (DC), which are present in high numbers within lesions of breast cancer [24 ]. CD83+ mature ("activated") DC, however, were found at the perimeter of the tumors. They also observed that many tumor cells produced the chemokine CCL20, which is known to attract immature DC via the chemokine receptor, CCR6. In adherence assays, in vitro-generated, immature DC were demonstrated to preferentially bind within the tumor, whereas mature DC adhered outside these areas. Although the significance of these findings are still unclear, it is possible that attraction of immature (or nonactivated) DC to tumors may increase the likelihood that the host becomes tolerant to the tumor.
Tolerance to tumor cells may also be mediated by a subpopulation of naturally occurring T regulatory cells (Tregs) characterized by expression of CD4, CD25, and the transcription factor, Foxp3 [25 ]. Through contact-dependent mechanisms, Tregs suppress the function and killing ability of CD8+ cytolytic T cells (CTL). Indeed, there is significant evidence to suggest that depletion of Tregs in mice results in enhanced anti-tumor host immunological responses [26 ].
Curiel et al. [27
] reported that ovarian cancer cells and tumor-infiltrating macrophages secrete the CCR4 ligand, CCL22, and that CD4+CD25+CD3+ putative Tregs bearing CCR4 (representing an average of 23% of all tumor-infiltrating CD4+ T cells) were present in ovarian tumor masses and in malignant ascites from patients with advanced disease. Nearly all CD25+ T cells expressed Foxp3, currently the most reliable marker for Tregs [28].
Furthermore, isolated CD4+CD25+ T cells migrated in response to patient-derived malignant ascites. Although anti-CXCL12 antibodies did not block Treg chemotaxis in response to malignant ascites fluid, anti-CCL22 antibodies blocked migration by 50%, suggesting that CCL22 was a major chemoattractant for Tregs within the malignant ascites. In vivo, ovarian cancer cells appeared to recruit adoptively transferred Tregs into tumor tissue in a CCL22-dependent manner. On a clinical note, these investigators also showed a striking correlation between the presence of tumor-infiltrating Tregs, advanced cancer stage, and poor patient prognosis [27
]. Thus, CCR4+ Tregs apparently infiltrate into tumors in advanced disease and inhibit the activity of tumor-specific CTL, possibly leading to the poor outcome observed in this subset of patients.
In addition to an immunosuppressive environment at the tumor-bearing site, stromal cells may facilitate tumor progression by producing MMPs [12 ]. CC chemokines such as CCL2, CCL4, and CCL5, as well as CXCL12, can induce MMP9 production, which may contribute to tumor cell migration, invasion, and subsequent metastasis [29 , 30 ].
Chemokines and tumor growth
Chemokines secreted by tumors not only attract infiltrating cells into tumor sites but may also contribute to tumor cell growth. The best-characterized chemokines in this regard are the growth-related (GRO) family chemokines (CXCL1, -2, and -3) and CXCL8, which act as autocrine growth factors for melanoma and other cancers [31
32
33
34
]. These chemokines also stimulate angiogenesis as described in the following section. Melanoma cells differ from normal melanocytes, in part, by the constitutive activation of nuclear factor-
B, leading to constitutive expression of CXCL1 and CXCL8 [34
]. Melanoma cells also consistently express the cognate CXCL1, -2, and -3 receptors CXCR1 and CXCR2 [35
]. Richmond and colleagues [48
] showed that overexpression of MIP-2, a murine GRO family member, under a tyrosinase promoter in the skin of transgenic mice, was sufficient to increase the frequency of cutaneous melanoma by sixfold, suggesting that MIP-2 is capable of stimulating the development of malignant melanoma tumors in vivo [36
].
Chemokines and angiogenesis
Chemokines also regulate angiogenesis in the tumor microenvironment. The N terminus of several CXC chemokines contains three amino acid residues [Glu-Leu-Arg (ELR motif)], which precede the first cysteine amino acid residue of the primary structure of these cytokines [37
, 38
]. Members that contain the ELR motif (ELR+) are angiogenic factors. The angiogenic members of the CXC chemokines include CXCL1, -2, -3, -5, -6, -7, and -8 [37
]. They are directly chemotactic for endothelial cells and can stimulate angiogenesis in vivo [37
].
In contrast, members that lack this motif (ELR–) inhibit angiogenesis through mechanisms that remain poorly defined [39 ]. For example, CXCL9 and CXCL10 exhibit antiangiogenic effects and inhibit tumor progression in vivo [40 ]. Indeed, levels of CXCL10 were inversely correlated with tumor progression in human lung cancer, which may be partly attributable to this mechanism [39 ].
Although CXCL12 is not a ELR+ CXC chemokine, it nonetheless promotes angiogenesis. CXCL12 increased the expression of vascular endothelial growth factor (VEGF) by endothelial cells. A positive-feedback loop may be in place, as VEGF also up-regulated CXCR4 on endothelial cells [41 ]. In addition, blockade of CXCL12/CXCR4 decreased tumor growth in vivo by inhibiting angiogenesis in a VEGF-independent manner [42 ]. It is interesting that this effect was not strictly dependent on the expression of CXCR4 by the tumor cell itself, suggesting that the role of CXCR4 in tumor neovascularization may be sufficient to allow enhanced tumor formation regardless of the expression level of CXCR4 by the tumor.
The source of CXCL12 that drives angiogenesis is likely to be derived from specialized stromal cells within tumors, as demonstrated by two studies that isolated patient-derived, carcinoma-associated stromal fibroblasts (CAF), which expressed significant amounts of this chemokine [43 , 44 ]. Of note, CAF were better able (compared with normal fibroblasts) to stimulate tumor formation when coinjected with MCF-7 human breast cancer cells and stimulated larger numbers of CD31-positive blood vessels within tumors that developed [44 ]. Tumors derived with CAF contained increased numbers of Sca1+CD31+ endothelial progenitor cells, which were subsequently shown to home to the inoculation site, where they were incorporated into new blood vessels within developing tumors.
In addition to CXCL12 supplied by stromal cells, cancer cells can synthesize this chemokine [45 ]. When stimulated by lipopolysaccharide (LPS) or fetal bovine serum, CXCR4-positive glioblastoma cell lines increased basal secretion of CXCL12 [45 ]. LPS-mediated proliferation was inhibited by CXCR4-blocking antibodies, suggesting that autocrine/paracrine secretion of chemokines may contribute to tumor proliferation and possibly, angiogenesis and survival as well (see "CCR10 and cancer survival in skin" below). The ability of CXCR4-expressing cancer cells to secrete CXCL12 (and other chemokines) raises the possibility that autocrine secretion of a chemokine may be able to increase metastasis by cancer cells bearing the appropriate receptor in a variety of tissues in addition to those tissues rich in that particular chemokine. This hypothesis may explain why cancers occasionally metastasize to "uncharacteristic" sites for a particular type of cancer.
|
CHEMOKINES AND THEIR RECEPTORS IN CANCER METASTASIS |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
and ß), is constitutively produced in multiple tissues, including those where metastases develop frequently (i.e., lung, liver, and bone). Expressed on nearly all leukocytes, CXCR4 plays a variety of roles in B and T lymphocyte as well as thymocyte trafficking [49 ]. It is interesting that a truncating mutation of the C-terminal cytoplasmic tail of CXCR4 leads to unchecked activation of this receptor. The resulting activation, surprisingly, results in abnormalities in humoral and cellular immunity in a syndrome known as warts, hypogammaglobulinemia, infections (bacterial), myelokathexis (severe chronic neutropenia with hyperplasia of mature myeloid cells in the bone marrow) [50 ]. CXCR4 is regulated in part by the local production of proteases, including dipeptidyl peptidase/CD26 [51 ], cathepsin G, and MMPs, which cleave short sequences of amino acids from the N terminus of the CXCL12, rendering this chemokine inactive [52 ]. Following binding with CXCL12, CXCR4 (as well as other chemokine receptors) activates several key migratory, proliferative, and survival signaling pathways, including the mitogen-activated protein kinase pathway, phosphatidylinositol-3 kinase (PI-3K)-Akt pathway, and possibly certain Janus tyrosine kinases 2 and 3 [52 ]. CXCR4 is critical during embryologic development, as CXCR4 knockout mice have severe defects in vascular development, which result in the death of mice in utero [53 ].
Zlotnik and colleagues [47 ] first reported that human breast cancer and melanoma cell lines expressed a limited number of chemokine receptors that appeared to be characteristic for each type of cancer. For example, breast cancer cells expressed CXCR4, CXCR2, and CCR7 in descending levels, and melanoma cell lines typically expressed CCR10, CXCR4, and CCR7. It is interesting that both types of cancer cell lines expressed CCR7, a receptor associated with LN metastasis (see below), as well as CXCR4. Furthermore, in confirmation of prior results [54 ], Zlotnik and co-workers [58 ] confirmed that CCL21 (a CCR7 ligand) was expressed highly in LN tissue, and CXCL12 (the CXCR4 ligand) was expressed highly in lung, LN, and liver and at lower levels in many other tissues [47 ].
The treatment of CXCR4-expressing breast cancer cells with neutralizing anti-CXCR4 antibody reduced metastasis to lungs in an intravenous (i.v.) tail-vein injection model and an orthotopic implantation model [47 ]. The contributory role of CXCR4 in lung metastasis was supported by a number of other observations. Overexpression of CXCR4 pancreatic cancer cells [55 ] and in murine melanoma cells [56 ] resulted in significant increases in lung metastasis following i.v. tail injection (see Fig. 2 ), as well as in an orthotopic melanoma metastasis model [56 ]. It is interesting that in the former study, liver metastases were also observed in greater numbers in CXCR4-transfected cells, consistent with the elevated levels of CXCL12 noted in liver tissue.
|
(TNF-)-stimulated lung endothelial cells, which expressed VCAM-1 [57
]. These results suggest that endothelial cell-derived CXCR12 may trigger the arrest of circulating cancer cells on endothelial cells by activating adhesion molecules in a manner analogous to the process described in leukocytes. Second, in several types of cancer such as melanoma, glioma, ovarian, small cell lung, basal cell carcinoma, and renal cancer, stimulation by CXCL12 resulted in increased proliferation and survival of CXCR4-expressing cancer cells when they were cultured under suboptimal conditions, such as low serum concentration [56 , 58 59 60 ], suggesting survival and/or growth advantages of CXCR4 activation for these cells under stress. Several groups have demonstrated that CXCL12 stimulated cancer cell proliferation [44 , 56 ], migration, and invasion [43 , 61 ]. Enhanced survival is likely to be mediated by activation of PI-3K and its downstream effector, Akt [62 ].
The up-regulation of chemokine receptor expression by cancer cells may be triggered by local environmental factors. Cancer cells are exposed to various physiological stress, including hypoxia [63
, 64
]. Even under such conditions, these cells are able to survive and proliferate, whereas normal tissues cannot do so. Recent findings show that CXCR4 expression is regulated (in part) by a hypoxia response element within the CXCR4 promoter through binding of hypoxia-inducible factor-1 (HIF-1) [65
, 66
]. Renal cell carcinomas with mutations in the von Hippel Lindau (VHL) protein, which functions as a member of the protein complex targeting HIF-1 for degradation, were shown to have marked increases in CXCR4 expression [65
]. Re-expression of wild-type VHL in these cells resulted in down-regulation of CXCR4 [75
]. The up-regulation of CXCR4 under hypoxic condition was dependent on activation of the PI-3K signaling pathway, which also inhibits apoptosis [67
]. These data suggest that the up-regulation of CXCR4 is one of many cellular responses, which allow tumor cells to withstand hypoxic conditions.
CXCR4 is likely to be regulated by a number of other cellular pathways. For example, CXCR4 up-regulation can result from HER2 receptor tyrosine kinase-mediated signaling through inhibition of ligand-induced CXCR4 degradation [68 ]. Moreover, kisspeptin-10, a product of the KISS1 metastatic suppressor gene [69 ], has recently been shown to markedly inhibit CXCL12-stimulated chemotaxis through CXCR4 [70 ]. Kisspeptin-10 binds to the Gq-coupled receptor GPR54 and was shown to unilaterally prevent CXCR4 activation from inducing calcium flux and stimulating Akt phosphorylation in HeLa and Chinese hamster ovary cells in vitro [70 ]. Although in vivo experiments were not performed [70 ], interplay between different classes of G protein-coupled receptors may impact on the functional ability of tumor-associated chemokine receptors to facilitate metastasis.
One key step in the formation of clinically significant metastases is the conversion of disseminated single cancer cells to proliferating, proangiogenic, secondary tumors capable of disrupting normal organ function [71 ]. This potentially rate-liming step has been termed metastatic colonization [72 ]. Roos and colleagues used overexpression of a CXCL12 sequence containing an endoplasmic reticulum retention signal (KDEL) to block newly synthesized CXCR4 from reaching the cell surface, thus abrogating CXCR4 function in CT26 colon carcinoma cells [73 ]. It is surprising that CT26 cells grown in vitro expressed low levels of CXCR4, but tumors developing from experimental lung metastases showed high levels of this receptor. CT26 cells expressing CXCL12-KDEL formed far fewer metastases than wild-type CT26 cells, although both cell types were able to extravasate into lung tissue and survive in similar (small) numbers immediately after inoculation and for up to 6 days.
The failure of CXCL12-KDEL-transfected cells to proliferate in the lungs suggests that CXCR4 may be required for outgrowth or metastatic colonization of CT26 cells. Moreover, liver metastases were reduced markedly following inoculation of CXCL12-KDEL-transfected CT26 cells, indicating that other tissues (besides lung) develop metastases, which are highly dependent on CXCR4 expression by tumor cells.
Additional experiments by Cardones et al. demonstrated that CXCR4-expressing B16 melanoma cells showed dramatic decreases in experimental lung metastases if the tumor cells were treated with T22, an 18-amino acid, CXCR4-specific inhibitor, just prior to tail vein inoculation [57 ]. Yet, tumor outgrowth was not inhibited by systemic treatment with T22 beginning at Day 7 (halfway through the course of the experiment), suggesting that T22 is only effective if it is present at an early step in metastasis. Similar studies conducted with highly metastatic murine breast cancer cells demonstrated that RNA interference-mediated down-regulation of CXCR4 in these cells prevented tumor formation and reduced experimental lung metastases [74 ]. However, these and other studies have generally failed to establish whether CXCR4 inhibition can effectively inhibit established metastatic lesions—an important point when considering the potential use of CXCR4 antagonists as therapeutic agents in metastatic disease.
Clinically, there is evidence to suggest that expression of CXCR4 by human cancers is correlated with poor patient prognosis in breast cancer [68 ] and in melanoma [75 ]. Although CXCR4 is included as a component of a lung metastasis gene signature in breast cancer patients [76 ], it is likely that CXCR4 affects cancer cell metastasis, not only to the lung but also to other organs. In primary neuroblastoma, CXCR4 expression is associated with bone and bone marrow metastasis [77 ]. In breast cancer, CXCR4 has been reported as one of several genes, which constitutes a bone metastasis gene signature [78 ]. In addition to CCR7, CXCR4 has been associated with LN metastasis [47 , 79 ]. Of note, however, Weigelt et al. [80 ] did not detect an increase in CXCR4 expression between primary breast cancer tumor samples and nodal metastases taken from the same patients in a small clinical, microarray-based study. Taken together, up-regulation of CXCR4 by some cancer cells leads to the potential to metastasize to various organs, not only the lung.
The role of CXCR4 in homing of neuroblastoma and breast carcinoma to bone is reminiscent of the role of CXCR4 in the well-documented homing of hematopoietic stem cells to the bone marrow [81 ]. Moreover, some have proposed the concept that CXCR4 may be a broad marker for tissue stem cells as well as cancer stem cells [82 ]. Supporting this concept is the finding that members of the paired-box transcription factor family, which are critical for organ development, are able to bind the CXCR4 promoters and may promote tissue-selective expression of CXCR4 in tissue stem cells [83 ].
CCR7 and regional LN metastasis
Regional draining LN are the most common sites of metastasis for solid and hematopoietic cancers. For example, 15% of 260 melanoma patients (all stages) at a large melanoma treatment center were sentinel LN-positive [84
]. When LN metastasis occurs, it is associated strongly with clinically poor outcomes in melanoma [85
] as well as in nearly all other solid tumors, including those of the breast and colon. Superficially, the process of tumor cells invading lymphatic vessels and migrating to regional LN resembles the trafficking pattern of antigen-bearing DC as they become activated in peripheral tissues, enter lymphatics, and accumulate in LN, where they potentially interact with T cells bearing cognate antigen receptors. Our understanding of the factors that regulate nodal metastasis are limited, but recent studies suggest that CCR7, a chemokine receptor involved in the emigration of activated DC from the periphery to draining LN, is likely to play a critical role in nodal metastasis.
DC have long been noted to enter lymphatic vessels as they transit from tissues such as skin to draining LN via afferent lymphatic vessels [86
]. Little was known about the molecular processes that guided these cells to the lymphatic vessels, however, until Gunn et al. described a novel chemokine, CCL21, which was expressed by lymphatic endothelial cells in the gut and liver [54
]. Naïve T cells were also shown to express CCR7 and use this receptor to arrest on CCL21-expressing, high endothelial venules within secondary lymphoid organs [54
]. Subsequent work showed that CCL21 was expressed by skin lymphatic endothelial cells and that skin DC, including Langerhans cells, up-regulated the CCL21 receptor, CCR7, upon activation with TNF- [87
]. CCL21 is required for efficient migration of DC from skin to regional LN, as accumulation of DC in the draining LN was impaired by antibodies that neutralized CCL21. The requirement of CCR7 for DC emigration to regional LN was demonstrated through the use of CCR7-deficient mice, whose skin DC fail to enter dermal lymphatic vessels [88
] or to accumulate in regional LN [89
].
Breast cancer cell lines as well as melanoma cell lines were shown to express CCR7 at the transcriptional level [47
]. To determine if expression of CCR7 was sufficient to alter the metastatic phenotype of murine melanoma cells, Wiley et al. overexpressed CCR7 in B16 murine melanoma cells. The resulting cells were able to metastasize far more efficiently (700-fold) to the draining LN following initial inoculation of the tumor cells in the footpad [90
]. Moreover, LN metastasis could be blocked with anti-CCL21 monoclonal antibody. Of note, CCR7 mediated specific metastasis to the LN, as CCR7-expressing B16 cells showed no increase in experimental lung metastasis following i.v. injection of tumor cells [90
]. The involvement of CCR7 in nodal metastasis is supported by subsequent clinical findings that CCR7 is expressed in a variety of solid cancers (breast, gastric, non-small cell lung, melanoma, and esophageal squamous cell cancer) and hematopoietic cancers (CLL) and is correlated with nodal metastasis and poor clinical outcome [91
92
93
94
95
]. Thus, cancer cells appear to use the same receptor, CCR7, as do DC to facilitate their entry into lymphatics and subsequent retention within CCL21-rich secondary lymphoid organs (Fig. 3
).
|
CCR10 was expressed frequently by human melanoma cell lines [47 ] and was detected at moderate to high levels in 10 of 13 primary melanomas by immunohistochemistry [62 ]. B16 cells transduced with luciferase only could not form tumors, as luciferase acted as a neoantigen that permitted the host to suppress tumor growth [62 ]. Overexpression of CCR10 (in addition to luciferase) in B16 cells, however, allowed these cells to establish tumors in skin. Furthermore, mice injected in ears with luciferase-B16 cells were protected from subsequent rechallenge when these cells were implanted at a secondary skin site, which could be overcome if the B16 cells expressed CCR10 in addition to the luciferase.
CCR10-expressing B16 cells exposed to CCL27 were resistant to killing by cell-surface Fas cross-linking and by GP100 peptide-specific CD8+ cytolytic T cells [62 ]. A key to this phenomenon was activation of the PI-3K/protein kinase B (Akt) signaling pathway upon CCR10 engagement (Fig. 4 ) [62 ]. PI-3K and its downstream effector, Akt, prevent cellular apoptosis via a number of antiapoptotic mechanisms, including phosphorylation of BAD and forkhead transcription factors [102 ]. It is important that inhibition of the chemokine receptor with PTX or of its downstream effector, PI-3K, with wortmannin or LY294002 restored the sensitivity of CCR10-B16 cells to apoptotic cell death even in the presence of the CCR10 ligand [62 ].
|
It is interesting that mice injected in ears with CCR10-B16 cells showed massive metastasis of tumor cells to the draining cervical LN in addition to growth of the tumors at the site of inoculation [62 ]. In vivo luciferase imaging revealed that CCR10-B16 cells accumulated with minutes following ear skin inoculation (T. Murakami and S. T. Hwang, unpublished data). It is possible that CCR10-B16 may have been activated through keratinocyte-derived CCL27, which was transported in lymphatic fluid to the draining LN, as CCL27 itself is not expressed in LN. Others have shown convincingly, however, that chemokines injected into skin are transported rapidly to draining LN [105 ]. Thus, an alternative hypothesis to explain the growth of CCR10-B16 tumors in ear skin may be a failure of immunological priming in the presence of early nodal metastasis [62 ]. Testing this hypothesis may reveal a mechanism whereby nodal metastasis is associated with poor outcome in patients with many forms of cancer, including melanoma.
|
CHEMOKINES AND CANCER METASTASIS: THERAPEUTIC IMPLICATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
One possible use of chemokine receptor antagonists will be to block angiogenesis, a critical step in the conversion of micrometastases to macroscopic secondary tumors. Guleng et al. demonstrated that neutralization of CXCR4 with blocking antibodies resulted in a delay of tumor formation by CXCR4-expressing Colon38 tumor cells [42 ]. It is surprising that these antibodies also effectively delayed tumor formation and prevented angiogenesis by Colon38 tumor cells in which CXCR4 expression was silenced by CXCR4-specific, inhibitory RNAs. Thus, CXCR4 inhibitors may be useful in preventing tumor formation by blocking CXCR4-dependent processes, including blood vessel formation, independent of CXCR4 expressed by the tumor cells.
Results from our in vitro studies suggested that inhibition of chemokine receptor signaling pathways (i.e., PI-3K) renders cancer cells susceptible to apoptosis induced by B16 melanoma-specific CTL [62 ]. Although it has been shown that chemokine receptor activation increases the resistance of cancer cells to death induced by specific CTL, it is unclear whether activation of chemokine receptors leads to resistance of cancer cells to apoptosis induced by other cancer treatment modalities (e.g., cytotoxic chemotherapeutic agents and radiation therapy). One can imagine that once these apoptotic triggers are identified, chemokine receptor antagonist pretreatment combined with the specific treatment may kill tumor cells better than the single treatment alone. It is interesting that chemokine receptor antagonists (even those that target receptors with broad distribution such as CXCR4) appear to have relatively little toxicity in clinical trials. Unlike traditional anticancer drugs, they do not have demonstrable cytotoxic effects in vitro. Already, a large number of small molecule or peptide inhibitors of chemokine receptors (mostly targeted at CXCR4) have been shown to have effects on tumor growth in animals models (Table 2 ). It should also be noted that the use of chemokine receptor antagonists in treating metastatic tumors may rely on the up-regulation of specific chemokine receptors (e.g., CXCR4) by tumors in the presence of hypoxia or other cellular stress and on the dependence on these receptors for angiogenesis or survival. Thus, the chemokine receptors may be likened to an Achilles’ heel of cancer. When activated, they protect the tumor against a variety of apoptotic triggers, but inhibition of these chemokine receptors may leave cancer cells vulnerable to a number of existing anticancer treatments.
|
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
|
|
ACKNOWLEDGEMENTS |
|---|
Received November 5, 2005; revised December 12, 2005; accepted December 13, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND THEIR RECEPTORS...CHEMOKINES AND CANCER...SUMMARYREFERENCES |
|---|
Eur. J. Immunol. 32,404-412[CrossRef][Medline]
-inducible protein 10 (IP-10) is an angiostatic factor that inhibits human non-small cell lung cancer (NSCLC) tumorigenesis and spontaneous metastases J. Exp. Med. 184,981-992 Am. J. Pathol. 154,1125-1135 stimulates human glioblastoma cell growth through the activation of both extracellular signal-regulated kinases 1/2 and Akt Cancer Res. 63,1969-1974-mediated chemotaxis of human cord blood CD34+ progenitor cells J. Immunol. 169,7000-7008 J. Biol. Chem. 280,22473-22481
