| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
FORUM FOR YOUNG TURKS |
a Stem Cell Biology Program at James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky, USA;
b European Union Stem Cell Therapeutics Excellence Center, CMUJ, Krakow, Poland;
c Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
Key Words. CXCR4 • SDF-1 • Stem cells • Homing • Metastasis
Correspondence: Mariusz Z. Ratajczak, M.D., Ph.D., Director of Stem Cell Biology Program, James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky 40202, USA. Telephone: 502-852-1788; Fax: 502-852-3032; e-mail: mzrata01{at}louisville.edu
 |
ABSTRACT |
|---|
 Top Abstract IntroductionConclusionReferences |
|---|
-chemokine stromal-derived factor (SDF)-1 and the G-protein–coupled seven-span transmembrane receptor CXCR4 axis regulates the trafficking of various cell types. In this review, we present the concept that the SDF-1–CXCR4 axis is a master regulator of trafficking of both normal and cancer stem cells. Supporting this is growing evidence that SDF-1 plays a pivotal role in the regulation of trafficking of normal hematopoietic stem cells (HSCs) and their homing/retention in bone marrow. Moreover, functional CXCR4 is also expressed on nonhematopoietic tissue-committed stem/progenitor cells (TCSCs); hence, the SDF-1–CXCR4 axis emerges as a pivotal regulator of trafficking of various types of stem cells in the body. Furthermore, because most if not all malignancies originate in the stem/progenitor cell compartment, cancer stem cells also express CXCR4 on their surface and, as a result, the SDF-1–CXCR4 axis is also involved in directing their trafficking/metastasis to organs that highly express SDF-1 (e.g., lymph nodes, lungs, liver, and bones). Hence, we postulate that the metastasis of cancer stem cells and trafficking of normal stem cells involve similar mechanisms, and we discuss here the common molecular mechanisms involved in these processes. Finally, the responsiveness of CXCR4+ normal and malignant stem cells to an SDF-1 gradient may be regulated positively/primed by several small molecules related to inflammation which enhance incorporation of CXCR4 into membrane lipid rafts, or may be inhibited/blocked by small CXCR4 antagonist peptides. Consequently, strategies aimed at modulating the SDF-1–CXCR4 axis could have important clinical applications both in regenerative medicine to deliver normal stem cells to the tissues/organs and in clinical hematology/oncology to inhibit metastasis of cancer stem cells.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionConclusionReferences |
|---|
More than 50 different chemokines and 20 different chemo-kine receptors have been cloned [3, 4]. Chemokines usually bind to multiple receptors, and the same receptor may bind to more than one chemokine. However, there is one exception to this rule: the -chemokine stromal-derived factor (SDF)-1, which binds exclusively to CXCR4 and has CXCR4 as its only receptor [5–10]. This fact alone suggests that the SDF-1–CXCR4 axis plays a uniquely important biological role. Supporting this notion are murine knockout data showing that SDF-1 secreted by bone marrow (BM) stromal cells is critical for the colonization of BM by fetal liver-derived hematopoietic stem cells (HSCs) during embryogenesis [5–10]. Furthermore, during adult life, it plays a pivotal role in the retention/homing of these cells into the BM microenvironment [5–10]. Thus, it is not surprising that perturbation of the SDF-1–CXCR4 axis, for example, by mobilizing agents, is essential for the egress of hematopoietic stem/progenitor cells from BM into peripheral blood (PB) [11–13]. On the other hand, proper functioning of the SDF-1–CXCR4 axis is crucial in directing homing/engraftment of HSC into BM after transplantation [14].
However, what is often overlooked is that SDF-1–CXCR4 knockout mice, in addition to defects in colonization of BM by HSCs, also display defects in the development of heart, brain, and large vessels [6–10]. We infer from this that the SDF-1–CXCR4 axis may have a more general role during organogenesis. Compelling evidence is accumulating from our and others’ work that stem cells for different organs and tissues (which we will refer to in this review as tissue-committed stem cells, TCSCs) resemble HSCs in expressing functional CXCR4 on their surfaces and because of this follow an SDF-1 gradient [15, 16]. Thus, from a developmental point of view, SDF-1 seems to be one of the most important motomorphogenic factors and chemoattractants not only for HSCs but for nonhematopoietic CXCR4+ tissue/organ–committed progenitor/stem cells.
Concept of Cancer Stem Cells
The evidence accumulates that quiescent TCSCs or cells developmentally closely related to them distributed in various organs may be a cellular origin of cancer development. To support this notion, stem cells are long-lived cells and thus become the subject of accumulating mutations that are crucial for initiation/progression of cancer. Thus, mutations that occur in normal stem cells lead to their malignant transformation and tumor initiation [17–19]. The concept of cancer stem cells has been postulated by several investigators but was first experimentally documented for human leukemias [20]. Recently, a stem cell origin of cancer was demonstrated for several solid tumors such as brain and breast cancers [21, 22]. Of note, because cancer stem cells, much like normal stem cells, exist in a quiescent state, they are relatively resistant to most cytostatics that target only dividing cells. Therefore, in a growing tumor, cancer stem cells represent a subpopulation of tumor cells that are capable of initiating metastasis and regrowth of new tumors after unsuccessful treatment.
The fact that CXCR4 is expressed on normal stem cells for different organs/tissues may help to explain why several tumors that derived from these cells usually express CXCR4 on their surfaces (Table 1
). This also implies that the SDF-1–CXCR4 axis may influence the biology of tumors and direct the metastasis of CXCR4+ tumor cells by chemoattracting them to organs that highly express SDF-1 (e.g., lymph nodes, lungs, liver, or bones). Supporting this notion, it has been recently reported that several CXCR4+ cancers (e.g., breast, ovarian, prostate cancers, rhabdomyosarcoma, and neuroblastoma) metastasize to the bones from the bloodstream in an SDF-1–dependent manner [22–29]. In this review, we will focus on the role of the SDF–1–CXCR4 axis in regulating the trafficking/homing of normal stem cells and the metastasis of tumor stem cells and discuss the similarities of the molecular mechanisms involved in these processes. We postulate that the metastasis of cancer stem cells and trafficking/circulation of normal stem cells involve very similar processes.
|
). Hence, CXCR4 can be considered a universal marker of versatile stem cells beginning with the population of pluripotent embryonic stem cells and ending with the population of TCSCs.
|
B [47], hypoxia-inducible factor (HIF-1) [48], glucocorticoids [49], lysophosphatidylcholine [50], TGF-ß1 [51], VEGF [52], IFN- [53], and several interleukins (IL-2, IL-4, and IL-7) [54, 55]. Thus, it is likely that stress-related conditions may upregulate CXCR4 expression on both normal and malignant stem cells.
SDF-1 as a Developmental and Postnatal Chemoattractant for Stem Cells
The ligand for CXCR4, SDF-1 that is playing a crucial role in trafficking of normal and malignant cells has been reported in the literature to be expressed by BM endothelium, fibroblasts/osteoblasts, and functions in the homing/retention of CXCR4+ HSCs in the BM microenvironment [14, 30–33, 56]. It is, however, also secreted by stromal and endothelial cells of other organs such as heart [57], skeletal muscle [35], liver [41, 58], brain [7–9], and kidney [59]. Moreover, its secretion increases during tissue damage such as heart infarct [40, 60], limb ischemia [61, 62], toxic liver damage [58], excessive bleeding [15], total body irradiation, and after-tissue damage related to chemotherapy [63, 64]. This suggests that SDF-1 may play a pivotal role in chemoattracting CXCR4+ TCSCs necessary for organ/tissue regeneration [16, 58, 65–67], implicating this chemokine as an important maintenance factor in postnatal tissue repair. Because a strong correlation exists between inflammation and tumor progression/metastasis, inflammation-driven expression of SDF-1 may also play an important role in dissemination/metastasis of cancer stem cells.
In humans there are two identified splice variant forms of SDF-1: SDF-1 and SDF-1 ß. The former is more abundant than the latter, yet both are derived from a single gene. Studies at the molecular level show that SDF-1 expression is upregulated in endothelial cells by HIF-1 [68]. Supporting this is the finding that the SDF-1 promoter contains two HIF-1 binding sites, and HIF-1 binds to these sequences, as revealed by chromatin immunoprecipitation analysis [68]. Thus, HIF-1 elevates SDF-1 expression in hypoxic/damaged tissues, which leads to chemoattraction of CXCR4+ TCSCs that participate in tissue regeneration. Expression of SDF-1 has also been found to be positively regulated by NF-B [47]. Therefore, SDF-1 appears to be an important indicator of tissue/organ hypoxia/injury and is involved in the active recruitment of TCSCs for repair of damaged tissue. Because, as stated previously, HIF-1 and NF-B also upregulate the expression of CXCR4 [47, 68], these factors would appear to regulate the SDF-1–CXCR4 axis at both the ligand and receptor levels. This confers another important argument for a pivotal role of the SDF-1–CXCR4 axis in maintaining postnatal tissue repair.
We may assume that the secretion of chemoattractants such as SDF-1 in or around injured tissue is a crucial event that creates an environment facilitating the homing of circulating TCSCs for endothelium and the affected tissue necessary for organ regeneration/tissue repair [37, 39–41, 58, 61, 62, 69]. A similar mechanism, however, may in our opinion induce a pro-metastatic environment for CXCR4+ tumor stem cells in damaged organs. There is the possibility that this could lead, for example, to unwanted side effects of radiochemotherapy, and we will discuss this topic later. Conversely, the expression of SDF-1 was found to be down-regulated by steroids [70] and TGF-ß1 [71]. The downregulation of SDF-1 expression by steroids inhibits the accumulation of CXCR4+ inflammatory cells and contributes to the anti-inflammatory effect of these compounds [70, 71]. Thus, steroid therapy may be envisioned as a two-edged sword that not only inhibits trafficking of inflammatory cells but may inhibit homing of stem cells to the damaged organ.
Negative and Positive Modulators of the SDF-1–CXCR4 Axis, Implications for Trafficking of Normal and Malignant Stem Cells
It is evident that the function of the SDF-1–CXCR4 axis must be tightly regulated in vivo by various biological mechanisms, a fact that we are only now beginning to understand. Some of these mechanisms that may be relevant for trafficking of normal stem cells and metastasis of cancer cells will be discussed below.
Both tissue damage that is a signal for chemoattracting normal stem cells for regeneration and chronic inflammation promoting the metastasis of cancer stem cells lead to expression/accumulation of several molecules/factors in the damaged tissues. We and others have recently identified several of these factors that positively affect the sensitivity/responsiveness of CXCR4+ cells to an SDF-1 gradient (Fig. 2). Factors such as (a) anaphylatoxin C3a (C3 complement protein cleavage fragment), (b) des-Arg C3a (product of C3a degradation by carboxypeptidase), (c) platelet-derived membrane microvesicles, (d) hyaluronic acid, and (e) sphingosine-1 phosphate were found to significantly increase the chemotaxis of CXCR4+ cells to low/threshold dosages of SDF-1 [72–74]. Similarly, we found that several other molecules such as fibronectin, fibrinogen, thrombin, soluble uPAR, and VCAM-1 sensitize/increase the chemotactic responses of cells to low dosages of SDF-1 as well [75]. These observations support the concept that the SDF-1–CXCR4 axis may be modulated by various molecules related to inflammation/tissue damage (e.g., C3a anaphylatoxin, des-ArgC3a, fibronectin, hyaluronic acid), coagulation (e.g., fibrinogen, uPAR, thrombin) or cell activation (e.g., s-VCAM-1, s-ICAM-1, membrane-derived vesicles) [75, 76]. Because inflammation plays an important role in tumor progression, molecules generated during inflammation (e.g., C3a, fibrinogen, fibronectin, and hyaluronic acid) could have an enhancing effect on the metastatic behavior of CXCR4+ tumor cells. Our preliminary data lend support to this notion [75].
|
), allowing HSCs to better sense the SDF-1 gradient, this may explain in part why HSCs from mobilized PB engraft faster in patients than those from BM or cord blood [75]. Our recent data show that similarly as observed for normal hematopoietic cells [75], depletion of membrane cholesterol severely affects responsiveness of CXCR4+ cancer cells to an SDF-1 gradient (manuscript in preparation). Thus, drugs that perturb lipid raft formation by depleting cholesterol from cells such as polyene antibiotics (e.g., amphotericin B, nystatin) could potentially negatively affect the meta-static properties of cancer cells. To support this notion further, we envision that one of the noted anticancer effects of statins [77, 78], compounds that inhibit intracellular cholesterol synthesis, could also be explained by decreasing membrane cholesterol concentrations and as a result of this the formation of lipid rafts [79]. This, however, requires further studies in vivo in animal models. Conversely, it has been reported that CXCR4 signaling may be desensitized in B lymphocytes and T lymphocytes by MIP-1ß or RANTES, which activate another G-protein–coupled chemokine receptor, CCR5 [80, 81]. The molecular mechanism of this heterologous cross-desensitization between chemokine receptors is still unclear, but it is likely that it involves RGS proteins (regulators of G-protein signaling). Whether a similar desensitization mechanism also occurs on normal and malignant CXCR4+ HSCs and TCSCs requires more investigation. Interestingly, we found that the SDF-1–CXCR4 axis in HSCs in addition to polyene antibiotics (e.g., amphotericin B, nystatin) and statins may also be negatively modulated by soluble heparin and lipopolysaccharide (LPS). It is likely that, whereas heparin restricts SDF-1 availability in the intercellular environment, LPS probably interferes directly with the binding of SDF-1 to CXCR4 [82, 83].
Furthermore, the N-terminus of CXCR4 and its first extracellular loop are crucial for SDF-1 binding, and this explains why this part of the receptor is modulated by several mechanisms. Figure 1 shows different mechanisms by which the SDF-1–CXCR4 axis is modulated. Hyposulfation of N-terminal tyrosine residues [84] or enzymatic processing/cleavage of the CXCR4 N-terminus by leukocyte-derived proteases inhibits CXCR4 signaling [85]. Like CXCR4, SDF-1 may also be N-terminally truncated by cell-surface–expressed CD26/dipeptidylpeptidase IV [86, 87]. As a result, truncated (3–68 aa) SDF-1, in contrast to full-length SDF-1 (1–68 aa), does not possess chemotactic activity and may even act as an antagonist of CXCR4 [86]. Thus, leukocyte-derived proteases released during inflammation may negatively affect the function of the SDF-1–CXCR4 axis in CXCR4+ cells such as monocytes and lymphocytes, which accumulate in tissues affected by inflammation. The same process, however, may also inhibit the homing responses of normal CXCR4+ stem cells. Thus, tissue damage/inflammation may exert pleiotropic effects on trafficking of stem cells. Accordingly, whereas damaged tissues on one hand secrete SDF-1 and other potential chemoattractants for stem cells, on the other hand their highly proteolytic environment may affect proper homing of circulating normal stem cells and affect their regenerative potential.
Based on all these investigations, a new and more complex picture of the mechanisms that regulate the activity of the SDF-1–CXCR4 axis is emerging. The functionality of the CXCR4 receptor on normal and malignant cells is modulated by factors such as (a) the level of receptor expression on the cell surface, (b) the sulfation status of its N-terminus, (c) the expression and biological availability of SDF-1 in tissues, (d) cleavage of the CXCR4 N-terminus on the cells and SDF-1 in the extracellular space by serine proteases and matrix metalloproteinases (MMPs), (e) heterologous desensitization by the CCR5 receptor, and (f) the modulation of CXCR4 incorporation into membrane lipid rafts positively by molecules related to inflammation/tissue remodeling and negatively by polyene antibiotics (Fig. 2). Thus, all of these versatile mechanisms may modulate the responsiveness of CXCR4+ normal and malignant stem cells to an SDF-1 gradient and affect their trafficking or metastasis, respectively.
The SDF-1–CXCR4 Axis Regulates Cell Trafficking: Biological Effects of SDF-1–Mediated Signaling
Extensive studies have been done on the biological effects and signaling pathways of the SDF-1–CXCR4 axis in normal and malignant cells (Fig. 2). As reported, SDF-1’s major biological effects are related to the ability of this chemokine to induce (a) motility, (b) chemotactic responses, (c) adhesion, and (d) secretion of MMPs and angiopoietic factors (e.g., VEGF) in cells bearing cognate CXCR4 [88–91]. SDF-1 also increases adhesion of cells to VCAM-1, ICAM-1, fibronectin, and fibrinogen by activating/modulating the function of several cell-surface integrins [28, 74, 88, 90, 92]. There is some controversy regarding whether SDF-1, besides regulating the trafficking of cells, also directly affects their proliferation and survival. In some studies, SDF-1 has been reported to increase the proliferation of astrocytes and myeloid progenitor cells [2, 93]. Moreover, it has been suggested that SDF-1 may be secreted by hematopoietic stem/progenitor cells and be involved in autocrine/paracrine regulation of their development and survival [94]. Interestingly, SDF-1 was also found to be expressed in solid tumors, for example, in some of human rhabdomyosarcoma cell lines [28].
Generally, the SDF-1–CXCR4 axis activates several signaling pathways in target cells crucial for both their trafficking and interaction with the intercellular environment. Having discussed in the previous section the biological consequences of CXCR4 incorporation into membrane lipid rafts, we will now describe the signaling pathways in cells stimulated by SDF-1 which we summarized in Figure 3. It is believed that after binding to CXCR4, SDF-1 triggers CXCR4 dimerization [95]. Shortly after being activated, CXCR4 becomes physically associated with Gi protein [95, 96], and tyrosine residues within the C-terminus become phosphorylated, probably through activation and association with the receptor of JAK2 and JAK3 kinases [97]. Subsequently, the activated CXCR4–SDF-1 complex is rapidly internalized from the cell surface in a mechanism involving G protein-coupled receptor kinases (GRKs) followed by the binding of ß-arrestin [98]. Internalized CXCR4 may again be re-expressed on the cell surface. We recently found, however, that the internalization of CXCR4 from the surface of HSCs could be inhibited by heparin (not published). Likewise, a recent study reported inhibition of internalization of CXCR4 after crosslinking of L-selectin when specific antibodies were used or after the addition of L-selectin–binding ligands such as fucoidan or sulfatide [99]. Besides receptor internalization during desensitization, which terminates receptor signaling, endocytosis of CXCR4 may be necessary to activate several pathways and functions such as chemotaxis and activation of the MEK kinase–MAPK p42/44 cascade [90, 96]. Recent experiments generated with mutated CXCR4 signaling domains do not, however, support this notion. Surprisingly, it was found that SDF-1–mediated phosphorylation of MAPK p42/44 also occurs in cells transfected with a mutant receptor that does not undergo internalization [97].
|
(PKC ), the MAPK p42/44–ELK-1 and PI-3K–AKT–NF-B axes [47, 91, 96, 97, 100, 101]. Recently, the major role of atypical isoform of PKC (PKC
) has been proposed to play a crucial role in stem cell trafficking [102]. Strong phosphorylation of focal adhesion components and MAPK p42/44 and serine-threonine kinase AKT is observed within seconds of CXCR4 activation [28, 103]. CXCR4 signaling also involves the Ras-activated signaling pathway, several src-related kinases such as Src, Lyn, Fyn, Lck, T-cell activating molecule ZAP-70, and vav [104]. Recent evidence from other [105–107] and our [75] laboratories indicates the role of small GTPases in CXCR4 signaling. Activation of these GTPases is crucial for cell migration to an SDF-1 gradient [75]. On the other hand, a recent report indicates that the RasGAP-associated docking protein p62Dok-1 negatively regulates chemotaxis of the Jurkat T-cell line [108].
JAK2, JAK3 [96], and Tyk-2 [95] may also associate in some cell types with CXCR4 and are activated, probably by transphosphorylation, in a Gi-independent manner [95]. Consequently, several members of the STAT family of transcription factors may become recruited and phosphorylated. However, involvement of STAT proteins in signaling from activated CXCR4 may depend on the cell type. For example, whereas STAT 2 and 4 but not STAT 1, 3, 5, and 6 become activated in the hematopoietic CTS progenitor cell line [97], STAT 1, 2, 3, and 5 but not STAT 4 or 6 become activated in MOLT4 cells [95]. Surprisingly, in our work we did not find that SDF-1 induces tyrosine phosphorylation of STAT 1, 3, 5 and 6, either in human CD34+ primary hematopoietic cells [103] or in many established human lymphohematopoietic cell lines [90].
Strong evidence exists that the protein-tyrosine phosphatases SHIP1 and SHIP2, as well as membrane-expressed hematopoietic phosphatase CD45, are also involved in the modulation of CXCR4 signaling [109]. Hematopoietic cells derived from mice lacking SHIP1 showed altered patterns of chemotactic response to SDF-1 [110]. SHIP2 also associated with CXCR4 and regulated SDF-1-induced migration of T and pre-B cells [110]. Similarly, lymphocytes negative for CD45 phosphatase expression also show reduced chemotaxis to SDF-1. Moreover, after stimulation by SDF-1, CD45 was found to associate with CXCR4 within lipid rafts [109], an interaction that could be inhibited by pretreatment of cells with ß-cyclodextrin, an inhibitor of lipid raft formation.
To better elucidate the role of CXCR4 domains in SDF-1–mediated signaling, studies were performed on a human embryonic kidney (HEK)-293 cell line stably expressing wild-type or mutated forms of three different intracellular loops or the C-terminus of CXCR4 [111]. These data show that the third intracellular loop of CXCR4 alone supports the binding of Gi proteins and is involved in calcium mobilization and MAPK p42/44 activation but does not trigger CXCR4 internalization after SDF-1 binding. This latter observation indicates that MAPK p42/44 phosphorylation is independent of CXCR4 internalization. Furthermore, the second intracellular loop was found to be dispensable for Gi signaling. Interestingly, the second, third, and C-terminal intracellular domains were essential to transduce SDF-1–mediated chemotaxis [111]. Finally, involvement/cooperation of several CXCR4 intracellular domains in this process suggests that chemotaxis is mediated by several complementary signaling events and not by activation of a single signaling pathway [111].
No CXCR4/SDF-1 deficiencies have been described in humans. Based on murine knockout data, we assume such defects to be lethal in the embryo. In contrast, however, an activating mutation of CXCR4 in patients suffering from warts, hypogammaglobulinemia, immunodeficiency, and myelokathexis (WHIM) syndrome has been reported recently [112]. The molecular basis of all these mutations are truncations in the C-terminus of CXCR4 [112, 113]. Lymphoblastoid cells derived from these patients have greater calcium flux than cells from normal donors, consistent with dysregulated signaling by the mutant receptor. This latter supports the idea that the C-terminal portion of CXCR4 is crucial for receptor desensitization and protects its hyperactivity [114]. It would be interesting to determine whether the WHIM mutation has any impact on trafficking of HSCs and TCSCs.
A New Perspective on BM as the Home of Circulating Normal CXCR4+ HSCs and CXCR4+ Tissue-Committed Stem/Progenitor Cells
Below, we will present a concept that normal CXCR4+ stem/progenitor cells accumulate during ontogenesis in BM, circulate in PB, and compete for SDF-1–positive niches in peripheral tissues.
Role of the SDF-1–CXCR4 Axis in Accumulation of CXCR4+ Stem/Progenitors in BM The BM tissue itself develops relatively late during ontogenesis at a time when fetal marrow replaces fetal liver as the hematopoietic organ. This developmental process establishes hematopoiesis in the human BM by the end of the second trimester of gestation. At this time, fibroblasts and osteoblasts in early bones begin to express/secrete SDF-1 that chemoattracts CXCR4+ HSCs from the fetal liver into BM [6, 9, 10, 14]. The important message from developmental studies is that CXCR4+ HSCs from the fetal liver colonize the BM microenvironment (to which they are chemoattracted by an SDF-1 gradient) as the first wave of CXCR4+ stem cells. Our recent data suggest that during ontogenesis, in addition to HSCs, other TCSCs (e.g., for muscles, neurons, liver, heart, endocrine pancreas, and kidney tubular epithelium) and perhaps even more primitive PSCs (precursors for various TCSCs) accumulate gradually in a SDF-1–dependent manner in the BM environment, where they find an environment conducive to their survival [16, 115]. Thus, BM can be envisioned not only as the home of HSCs but as the home of a small population of versatile CXCR4+ TCSCs that may serve in adult life as a reserve/mobile pool of stem cells for tissue/organ regeneration [15, 16, 115, 116].
To investigate this idea further, we hypothesized that CXCR4+ stem/progenitor cells residing in BM can be recovered, like CXCR4+ HSCs, from a suspension of BM mononuclear cells (BMMNCs) using chemotactic isolation to an SDF-1 gradient [15, 16, 116]. To confirm that cells isolated by a chemotactic SDF-1 gradient [15] are enriched in TCSCs, real-time reverse transcription–polymerase chain reaction (RT-PCR) and immunohistochemical staining were used to identify the isolated population of cells. We found that cells isolated in this way are enriched in mRNA for markers for early skeletal muscle (Myf-5, MyoD, myogenin), heart muscle (Nkx2.5/Csx, GATA-4, MEF-2C), neural (Nestin, GFAP), liver (CK19, -fetoprotein), and endocrine pancreas (Nkx6.1, Pdx1, Ptf1) [15, 16, 37, 40]. Furthermore, using immunohistochemical staining, we also detected in these cells the presence of various proteins characteristic of TCSCs, such as Myf-5, nestin, Nkx2.5/Csx, and GATA-4 [37, 40, 117]. More important, the tissue/organ commitment and stem/progenitor nature of BM-derived CXCR4+ TCSCs was demonstrated in in vitro cultures, where these cells were found to be able to differentiate into cardiomyocytes [37], grow neurospheres [64], and form endothelial colonies [69, 118]. Of note, CXCR4+ BMMNCs also highly expressed mRNA for transcription factors of pluripotent stem cells (PSCs) such as mRNA for Oct-4, Nanog, and Rex-1, suggesting that some PSCs could possibly be present among cells enriched for TCSCs or TCSCs may express some markers typical of PSCs [16, 115].
These data not only explain a pivotal role of the SDF-1–CXCR4 axis in accumulating these cells in BM, but provide another explanation of the phenomenon of stem cell plasticity. Accordingly, because TCSCs have been found in BM aspirates, it may explain why it is possible to grow endothelial [61, 67, 118, 119], skeletal muscle [120, 121], heart [60, 121], neural [122–124], liver [41, 125], and endocrine pancreatic cells from BMMNCs [126]. This may also explain the phenomenon of sex-mismatched BM transplants in humans that results in donor-derived hepatocytes and cholangiocytes [125], Purkinje cells [123], epidermal cells [127], or even intestinal epithelial cells [128]. We envision that TCSCs accumulate during ontogenesis in BM as a reserve pool of stem cells for regeneration [16, 66].
The Role of the SDF-1–CXCR4 Axis in Postnatal Circulation of CXCR4+ Stem/Progenitor Cells However circulation of stem cells plays an important role during development, we believe that versatile stem cells may also circulate under physiological conditions at low levels in PB, thereby maintaining a pool of cells in distant parts of the body during postnatal life. This is especially important for hematopoietic, muscle, and neural stem cells that are distributed/reside in niches dispersed at distant locations around the body. The phenomenon of stem cell circulation in blood is well documented for HSCs [56, 129], endothelial TCSCs [118], and recently for skeletal muscle satellite [15, 120] and skeleton TCSCs [130]. However, the percentage of nonhematopoietic TCSCs in PB is significantly lower than that of early HSCs.
Supporting this concept are our recent data that mRNA for several early markers for muscle (Myf-5, Myo-D), neural (GFAP, nestin), and liver (CK19, fetoprotein) can be detected in circulating (adherent cell-depleted) PBMNCs [15]. Furthermore, we recently observed that the number of circulating CXCR4+ TCSCs in PB is increased by the administration of agents similar to those used for mobilization of HSCs (e.g., G-CSF) [15, 117] or triggered by stress factors related to tissue/organ injury (e.g., heart infarct or stroke) [37, 40, 131]. Thus, pharmacological mobilization could be envisioned not only as a tool for increasing the number of HSCs in PB, but as a process that mobilizes TCSCs into PB. TCSCs, in addition to BM, are probably also mobilized into PB from various tissue/organ–specific niches. Furthermore, an increase in the number of these cells in PB during tissue/organ injury supports the idea that they play an important role in tissue repair. Therefore, it is likely that regeneration processes involve the recruitment not only of local stem/progenitor cells that reside in the damaged organ, but of TCSCs (specific for the damaged organ) that reside in distant niches of the same tissue/organ (e.g., in remote healthy muscle tissue) or, as we recently postulated, TCSCs that reside as a reserve pool in BM [16, 64, 115].
The Role of the SDF-1–CXCR4 Axis in Postnatal Tissue Distribution of CXCR4+ Stem/Progenitor Cells Because TCSCs are CXCR4+, these cells respond to an SDF-1 gradient, and there is growing evidence that the SDF-1–CXCR4 axis plays an important role in regulating their trafficking. SDF-1 is not only constitutively expressed by BM fibroblasts/osteoblasts and enriched in BM stem cell niches [5, 56, 129], but expressed in stem cell niches located in various tissues/organs; this may explain the presence of heterogenous populations of CXCR4+ TCSCs in various organs such as HSCs in muscle or neural tissue and, conversely, muscle or neural progenitors in BM [15, 16, 35].
In support of this concept is our recent finding that both CXCR4+ muscle stem/progenitor satellite cells as well as CXCR4+ HSCs could be chemoattracted in an SDF-1–dependent manner by media conditioned by either BM or muscle fibroblasts [35]. The effect of these media was less visible in our study when (a) the media were precleared by anti-SDF-1 antibodies, (b) CXCR4 on HSPCs or muscle satellite cells was blocked by antiCXCR4 antibodies, or (c) cells were preincubated with the small peptide CXCR4 antagonist T140 before chemotaxis [35]. These data provide evidence that both HSCs and muscle satellite TCSCs follow an SDF-1 gradient. Because neural, oval liver, and myocardium TCSCs are also chemoattracted by SDF-1, this chemokine alone or in combination with other factors appears to affect the homing of circulating stem cells into tissue-specific niches. Thus, the concept of circulating CXCR4+ stem cells and their competition for SDF-1–positive niches explains why stem cells specific to other organs are detectable in certain organs/tissues in addition to the organ-specific stem cells [15, 16, 34, 35]. Of course, this does not mean that other tissue-specific chemoattractants are not involved, which would account for why circulating stem cells have a higher affinity for niches in organs to which they are committed [16, 132–134].
The SDF-1–CXCR4 Axis Regulates Crucial Steps Involved in Mobilization, Trafficking, and Homing of Normal and Metastasis of Malignant Stem Cells
Mobilization/trafficking/homing of normal stem cells and metastasis of malignant stem cells is a multistep process [11, 129, 135] (Fig. 4). First, normal stem cells that egress from their tissue/organ niches or cancer stem cells that will establish distant metastases have to detach from the primary tumor mass and enter the PB or lymph. At a certain point, they sense a chemoattracting gradient that directs their homing (normal stem cell) or metastasis (cancer stem cell). On receiving this signal, they actively attach to the endothelium and subsequently penetrate the microvessel wall in a mechanism involving the secretion of metalloproteinases (MMPs). At the new distant site, stem cells may engraft and find an environmental niche that will protect them from apoptosis and allow them to expand to regenerate damaged tissue or establish a metastatic tumor (normal and cancer stem cells, respectively) (Fig. 4). The role of the SDF-1–CXCR4 axis in the consecutive steps of stem cell trafficking will be discussed below.
|
The SDF-1–CXCR4 Axis and Mobilization/Egress of Stem Cells
To get into circulation, a normal stem cell has to egress from its niche in the organ or, as in the case of a cancer stem cell, detach from the primary tumor (Fig. 4). The influence of SDF-1 on detachment/egress of cells from their tissue niches was studied best in a model of mobilization of normal HSPCs [11, 56, 129]. This intriguing process, however, is not yet fully understood. Mobilization may be enhanced by both decreasing the concentration of endogenous SDF-1 (e.g., after infusion of G-CSF or cyclophosphamide) in BM [11, 56] or, in contrast, by stimulating BM with exogenous SDF-1 [135]. In the first event, proteolytic enzymes released by marrow accessory cells may degrade SDF-1 expressed in BM and inactivate CXCR4 by partial proteolysis. As a result, cells will be released from BM into PB. In the second scenario, exogenous SDF-1 may stimulate the release of MMP-9 by HSPCs and consequently enhance their translocation from the endosteal niche to the endothelial niche within the BM environment [56, 69, 129]. HSPCs residing in the endothelial niche could subsequently be easily mobilized into PB by responding to a transendothelial SDF-1 gradient. To further support this last possibility, small peptide analogs of SDF-1 (e.g., CTCE0021) were recently used for mobilization [136]. An interesting phenomenon was also recently described, so called chemofugotaxis/reverse chemotaxis to SDF-1 [137]. Whether these mechanisms are also involved in the release/egress of the metastasizing cancer cells from the primary tumor is an open question.
The SDF-1–CXCR4 Axis, Cell Motility, and Directed Chemotaxis After they reach circulation, stem cells respond to chemoattractants that direct them to a new stem cell niche or, as in the case of cancer stem cells, to the organ where they may metastasize. SDF-1 is a potent chemoattractant for normal and malignant stem cells [5–10, 14, 25–31]. SDF-1 both chemoattracts cells and increases their motility [14, 15, 28, 29, 91]. This effect on cell motility [28, 29, 73] corresponded to changes in or rearrangements of cytoskeletal proteins, and cells exposed to SDF-1 display significant increases in the number and thickness of F-actin bundles. Moreover, stimulation of cells by SDF-1 also led to the formation of a leading edge in migrating tumor cells. By using time-lapse monitoring of the locomotion of individual cells, we were able to observe that SDF-1 increases the motility of several human tumor cells (e.g., breast cancer, rhabdomyosarcoma, and small lung cancer cell lines) [28, 29]. Tumor cells migrated extensively in the presence of SDF-1, which also increased the final displacement and the average velocity of cell displacement [28, 29].
Furthermore, SDF-1, just as in CXCR4+ stem cells, has been reported to induce a directional chemotaxis of several human solid tumor cell lines [23, 25, 28, 107, 138]. The stimulation of cell movement by SDF-1 (motility and directional migration) was inhibited in normal CD34+ hematopoietic cells as well as in several established cell lines after the PI-3K–AKT axis was blocked using Ly290042 or wortmanin and/or after inhibition of Gi signaling by preincubating cells with pertussis toxin [28, 96, 139]. This role of the PI-3K–AKT pathway in cell motility is further indicated by the observation that murine cells with a disruption of a single gene that encodes Pten (phosphatase and tensin homologue), a negative regulator of AKT phosphorylation/activation, shows enhanced phosphorylation of AKT, and as a result enhanced chemotaxis to SDF-1 [140]. SDF-1–mediated cell motility and chemotaxis, however, are affected by several other signaling events, probably involving MAPK p42/44 and phosphatases [109]. As reported recently, whereas the chemotaxis of normal HSCs was dependent on signaling through an atypical isoform of PKC, the chemotaxis of the Jurkat T-cell line was positively regulated by nonreceptor tyrosine kinase Pyk2 and negatively by p120RasGAP-associated p62 protein Dok-1 [102, 108]. As we previously discussed, the second, third, and C-terminal intracellular domains of CXCR4 are needed to transduce signals involved in SDF-1–mediated chemotaxis [111].
The SDF-1–CXCR4 Axis and Cell Adhesion
Circulating stem cells have to adhere to the endothelium in the organ where they will home or metastasize (Fig. 4). It is known that SDF-1 modulates adhesion of cells to fibrinogen, fibronectin, stroma, and endothelial cells [88, 141, 142], an effect that is mediated by the activation of various adhesion molecules (e.g., integrins) on the surface of target cells rather than by an increase in their de novo expression on the cell surface [88]. In accordance with this, SDF-1 has been reported to activate integrins LFA (lymphocyte function-associated antigen)-1, VLA (very late activation antigen)-4, and VLA-5 on human HSCs [88]. Lending further support to this evidence, we found that SDF-1 also activates IIb/ß3 integrin (CD41) on the surface of megakaryocytic cells [142]. Furthermore, SDF-1 induces firm adhesion and transendothelial migration of human CD34+ hematopoietic cells which are dependent on LFA-1/intracellular adhesion molecule (ICAM)-1 and VLA-4/vascular adhesion molecule (VCAM)-1 interactions. These interactions could be inhibited by pertussis toxin and cytochalasin D, indicating the involvement of Gi protein downstream signaling and the requirement of an intact cytoskeleton [88].
Interestingly, an alternative possibility was recently suggested postulating that adhesion molecules may trigger signals for both enhanced CXCR4 expression and increased function [99]. Activation of L-selectin on the surface of CXCR4+ cells upregulated both expression of and signaling from CXCR4. In light of these findings, we infer that interactions between CXCR4 and adhesion molecules are more complex than initially thought and seem in fact to constitute a two-way street. In the model proposed, CXCR4 signaling regulates the function of adhesion molecules on the cell surface, and, in turn, activation of certain adhesion molecules may enhance the signaling/expression of CXCR4 [99].
By means of similar mechanisms, SDF-1 may also increase the adhesion of tumor cells [75]. To support this notion, SDF-1 has been demonstrated, for example, to increase adhesion of neuroblastoma cells activated by TNF- and INF- HUVEC [25], prostate cancer cells to osteosarcoma layer and marrow-derived endothelium [143], and rhabdomyosarcoma cells to laminin and fibronectin [28, 29]. Furthermore, small lung cancer cells were reported to adhere to BM-derived stroma in response to SDF-1 in a VLA-4/VCAM-1–dependent manner [138].
The SDF-1–CXCR4 Axis and Cell Secretion
Circulating stem cells have to release factors (e.g., MMPs) that will enable them to cross the endothelial barrier and home to the organ (Fig. 4). Evidence has accumulated that, under certain circumstances, normal and malignant stem cells may express and secrete several proteins such as growth factors, cytokines, or proteolytic enzymes. These factors allow stem cells to interact with their surrounding environment as well as play some role in autocrine/paracrine regulation of the stem cell pool in niches [144–146]. Interactions of CXCR4 with SDF-1 lead to the activation of NF-B, which regulates cell secretion. It has been shown that normal and tumor cells secrete more MMPs (e.g., MMP-2 and MMP-9), nitric oxide, and certain angiopoietic factors such as VEGF after stimulation by SDF-1 [89, 90, 92, 142]. These factors/enzymes secreted by target cells are involved in migration of HSCs and tumor cells across vascular basement membranes (e.g., MMPs), in tumor vasculogenesis, and in crosstalk between CXCR4+ cells and endothelium (e.g., VEGF, HGF). These interactions are crucial both for trafficking of normal stem cells and for metastasis of cancer stem cells. Interestingly, both normal early stem/progenitor cells as well as cancer cells secrete several factors that promote angiogenesis. Thus, SDF-1 plays an important role in angiogenesis both by directly chemoattracting CXCR4+ endothelial TCSCs as well as by stimulating cells to secrete other angiopoietic factors (e.g., VEGF) [147, 148]. Similar mechanisms may operate during tumor vasculogenesis, in which SDF-1 has been reported to be upregulated in tumor and tumor-surrounding tissues under steady state conditions as well as after radiochemotherapy [24, 29]. Thus, in response to SDF-1 stimulation, cancer cells may also secrete several proangiopoietic factors [149] and metalloproteinases [150, 151].
The SDF-1–CXCR4 Axis on Cell Proliferation and Survival
Finally, normal stem cells or cancer stem cells have to survive/expand in the new distant tissue location (Fig. 4). Several chemokines have been reported to positively [40, 93] or negatively [152] modulate cell growth. It has been suggested that the ratio between activated p38 stress kinase (negative regulator) and the PI-3K–AKT axis (positive regulator) as a result of chemokine receptor activation determines whether a particular chemokine will support or inhibit cell survival [153]. Nevertheless, the direct effect of SDF-1 on cell proliferation and survival remains controversial. For example, in contrast to several other chemokines such as MIP-1 and PF-4 that have been reported to inhibit the proliferation of early hematopoietic cells [154, 155], SDF-1 has been found to stimulate the proliferation and survival of hematopoietic cells under certain experimental conditions [21, 156]. As we previously discussed, SDF-1 has been reported to be an autocrine survival factor for purified CD34+ CD38+ BMMNCs and its prosurvival effect is PI-3K–AKT–axis dependent [94].
Nevertheless, when we evaluated the effect of SDF-1 on normal human CD34+ stem/progenitor cells, lineage-expanded erythroblasts, megakaryoblasts, and myeloid cells, we found that SDF-1 did not affect cell survival or proliferation in any of them [90, 155]. Similarly, we did not observe an effect of SDF-1 on proliferation/survival in 26 different established human T and B lymphoid and myeloid cell lines [90]. However, in several of these cell lines as well as in normal human CD34+ cells and megakaryoblasts, SDF-1 induced phosphorylation of MAPK p42/44 and serine-threonine kinase AKT, which are widely accepted as being associated with cell survival/proliferation.
Interestingly, SDF-1 was found to stimulate the proliferation and survival of astrocytes and some tumor cell lines [93, 157]. The proliferation effect of SDF-1 on astrocytes was MAPK p42/44-dependent and sensitive to inhibition by PD98059, an inhibitor of MEK kinase which activates MAPK p42/44 [93]. Some tumor cells (e.g., prostate cancer) were found to secrete SDF-1, and an autocrine SDF-1–CXCR4 interaction had been suggested to regulate the proliferation/survival of these cells [24, 143]. Moreover, SDF-1 was found to be a survival factor for glioma and glioblastoma cells, an effect that has been correlated with prolonged activation of AKT and MAPK p42/44 [158].
On the other hand, several other tumor cell lines did not respond to stimulation by SDF-1 by proliferation or increased survival, making it very likely that SDF-1 is not directly involved in growth of HSCs, TCSCs, or cancer stem cells expressing CXCR4. It is possible, however, that SDF-1’s effect on cell proliferation/survival is an indirect one. By increasing cell adhesion, SDF-1 may increase cell survival of cells through signaling from adhesion contact molecules or by directing their migration may expose cells to other prosurvival factors that are coexpressed with SDF-1 in tissues. It is well known that signals generated from adhesion molecules prevent cells from undergoing apoptosis in a mechanism described as "anoikis" [159]. Lending support to this idea is the fact that hematopoietic cells from SDF-1 transgenic mice display both enhanced survival in vitro in response to growth factor withdrawal and enhanced myelopoiesis in vivo [156]. Finally, it is possible that SDF-1 increases cell proliferation by enhancing secretion of other autocrine growth factors.
Implications for the Future
The SDF-1–CXCR4 axis has emerged as an important regulator of trafficking of normal and malignant stem cells, which means that it is potentially a target for various therapeutic interventions. Modulation of this axis may be important for (a) enhancing homing of stem cells for tissue/organ regeneration, (b) mobilization of normal hematopoietic and nonhematopoietic stem cells into PB, and (c) inhibiting metastasis of CXCR4+ cancer stem cells.
Enhancement of stem cell homing could be accomplished by employing several recently identified priming agents of the SDF-1–CXCR4 axis, such as C3a, des-Arg C3a, or hyaluronic acid, before transplantation to prime the SDF-1–dependent homing of HSPCs into BM [11, 73–75, 160]. Such pharmacological modulation of CXCR4 signaling may also play an important role in chemoattraction of CXCR4+ TCSCs and regeneration of damaged organs and tissues by these cells. Based on this, we expect, for example, that in the future, homing and engraftment efficiency of HSCs after transplantation can be significantly improved by the use of drugs containing molecules that prime/increase the responsiveness of HSCs to an SDF-1 gradient. This strategy would be especially important in cord blood transplantation, in which the number of HSCs is limited, especially for reconstituting an adult recipient.
On the other hand, several inhibitors of the CXCR4–SDF-1 axis could be used as new drugs to increase the mobilization of HSCs and nonhematopoietic stem cells into PB. Small-molecular inhibitors of CXCR4 such as T140 and AMD 3,100, modified recombinant SDF-1, blocking antibodies against CXCR4 or molecules perturbing incorporation of CXCR4 into membrane lipid formation (e.g., statins) are examples of such potential compounds [12, 13, 79].
All of these blocking strategies may also have value as potent new antimetastatic drugs for CXCR4+ cancer cells, but there needs to be a careful assessment of the possible side effects resulting from such therapy that might occur during trafficking and development of hemato/lymphopoietic stem cells, maturation of lymphocytes, and/or platelet formation. Encouragingly, the first clinical trials in HIV-infected patients using blocking agents (to prevent binding of HIV to CXCR4) have proven pharmacologically safe [161]. Strategies based on utilization of antisense oligodeoxynucleotides or double-stranded RNA-mediated interference (RNAi) could also have potential as downregulators of the expression of CXCR4 in target cells [162]. Similarly, downregulation of HIF-1 by an siRNA strategy may lead to downregulation of SDF-1 in various tissues and inhibit spreading of CXCR4+ tumor cells [163]. Recently, a small molecular inhibitor of transcriptional coactivation of HIF-1, called chetomin, has been identified. Interestingly, systemic administration of chetomin inhibited hypoxia-inducible transcription within tumors and inhibited tumor growth in mice [164].
In this review, we have focused on a role of the SDF-1–CXCR4 axis in regulation of cell trafficking. This process, however, is regulated by several other regulatory axes that together are responsible for the efficacy and tissue specificity of this process. The most important motomorphogens that are involved, beside SDF-1, in regulating cell trafficking are hepatocyte growth factor/scatter factor [58], vascular endothelial growth factor [132], basic fibroblast growth factor [133, 165], kit ligand [134], and leukemia inhibitory factor [166]. We are on the beginning of a long journey to understand how all of these factors orchestrate stem cell migration.
|
CONCLUSION |
|---|
|
TopAbstractIntroductionConclusionReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopAbstractIntroductionConclusionReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionConclusionReferences |
|---|
B promotes breast cancer cell migration and metastasis by inducing the expression of the chemokine receptor CXCR4. J Biol Chem 2003;278:21631–21638.
