Mesenchymal stem cells (MSC) exhibit tropism for sites of tissue damage as well as the tumor microenvironment. Many of the same inflammatory mediators that are secreted by wounds are found in the tumor microenvironment and are thought to be involved in attracting MSC to these sites. Cell migration is dependent on a multitude of signals ranging from growth factors to chemokines secreted by injured cells and/or respondent immune cells. MSC are likely to have chemotactic properties similar to other immune cells that respond to injury and sites of inflammation. Thus, the well-described model of leukocyte migration can serve as a reasonable example to facilitate the identification of factors involved in MSC migration.Understanding the factors involved in regulating MSC migration to tumors is essential to ultimately develop novel clinical strategies aimed at using MSC as vehicles to deliver antitumor proteins or suppress MSC migration to reduce tumor growth. For example, radiation enhances inflammatory signaling in the tumor microenvironment and may be used to potentiate site-specific MSC migration. Alternatively, restricting the migration of the MSC to the tumor microenvironment may prevent competent tumor-stroma formation, thereby hindering the growth of the tumor. In this review, we will discuss the role of inflammatory signaling in attracting MSC to tumors.
Brief description of the mesenchymal stem cell
Mesenchymal stem cells (MSC) are non-hematopoietic adult stem cells with multilineage potential. MSC are defined by plastic adherence, differentiation potential and cell surface marker expression.1 MSC, or MSC-like cells, have been isolated from nearly every organ or tissue in the body, making it challenging to characterize the MSC as a completely homogenous population. MSC contribute to the maintenance and regeneration of connective tissues and have the capacity to differentiate within osteoblasts, adipocytes, chondrocytes, myocytes and cardiomyocytes. MSC express markers including CD29, CD44, CD51, CD73 (SH3/4), CD105 (SH2), CD166 (ALCAM) and Stro-1, but the expression of specific combinations of markers appear microenvironment-dependent, suggesting a strong influence of tissue context on MSC phenotypes. In general, MSC appear to be a non-immunogenic population of cells; however, a few studies demonstrate immune-repressive functions of MSC through the induction of peripheral tolerance evident in autoimmune disorders such as multiple sclerosis.2
The use of bone marrow-derived MSC have been employed in support and engraftment of the transplantation of hematopoietic stem cells (HSCs) following high-dose chemotherapy in an effort to replenish the destroyed bone marrow cell population.3 Additionally, pre-clinical studies have explored the use of MSC in the reduction of graft-versus-host disease, for tissue repair; including cerebral injury,4 bone fracture,5 myocardial ischemia/infarction,6 muscular dystrophy,7 as well as tumor homing.
Migration of MSC to tumors is thought to be due to inflammatory signaling in a tumor resembling that of an unresolved wound.8 The innate tropism of MSC for tumors can be exploited for the delivery of antitumor agents to the tumor microenvironment. Gene-modified MSC expressing interferon-β have been used to significantly reduce tumor burden and in some cases extend survival in murine models of melanoma,9 lung,10 breast cancer11 and glioma.12 However, the mechanism and factors responsible for the targeted tropism of MSC to these wounded microenvironments remain to be fully elucidated. MSC are likely to have chemotactic properties similar to other immune cells that respond to injury and sites of inflammation. Thus, the well-described model of leukocyte migration can serve as a reasonable example to facilitate the identification of factors involved in MSC migration. Following a discussion of alterations in the ‘wounded’/tumor microenvironment that are shown to enhance MSC tumor-specific migration, we will introduce alternative, injury-induced cell migration models based on literature reviewing leukocytes and their progenitor cell line, the HSC. These alternative migration systems will provide rational for the tumor-specific MSC migration and will lead us into an overview of the current literature concerning MSC migration specifically. Ultimately, we will conclude with a discussion of the potential clinical applications of tumor-directed MSC migration.
Inflammation-targeted homing in the tumor microenvironment
Inflammation is a cellular response that takes place under conditions of cellular injury and in sites of tissue wounding. Over two decades ago, Dvorak and co-workers described the tumor as an unhealed wound that produces a continuous source of inflammatory mediators (cytokines, chemokines and other potential chemoattractant molecules). Cancer progression has been correlated with an increase in inflammatory mediator gene expression, and this is thought to occur via disruption, damage and cellular turnover occurring in the tumor microenvironment. This constant production of inflammatory mediators perpetuates the maintenance and progression of the tumor environment and becomes a target for the MSC. Tumor-generated inflammatory mediators have a role in determining the conditions of the tumor microenvironment, as they regulate invasion, motility, extracellular matrix interaction through autocrine effects as well as coordinating cell movement through paracrine signaling.13 Tumor-produced and tumor-induced inflammatory chemokines are known to have an important role in leukocyte/macrophage infiltration into tumors.14 Based on this evidence, one can speculate that inflammation-induced chemokines participate in the directed migration of stem cells, such as MSC, to tumors and inflamed microenvironments. Previous studies have shown the importance of inflammation to the successful homing of systematically infused stem cells (HSC) to cardiac tissue,15 thus reinforcing the notion of inflammation and chemokine production in migration of MSC. In addition to the secreted chemotactic molecules secreted by the tumor and its surrounding stroma, the tumor cells themselves retain a chemotactic disparity amongst its cellular components. When fractionated, the tumor cell membrane possesses a superior chemotaxis-induction potential compared with the other cellular components such as cytosolic fractions, including organelles such as the nuclei, mitochondria, lysosomes, microsomes and ribosomes (Figure 1).
Hypoxia contribution to MSC homing
Many tumors exhibit hypoxia, a state of reduced oxygen that often parallels and perpetuates inflammation. A common feature between inflammation and hypoxic environments is the expression of pro-angiogenic molecules. The hypoxia-induced transcription factor HIF-1α activates the transcription of genes including vascular endothelial growth factor (VEGF), macrophage migration inhibitory factor, tumor necrosis factors (TNF-α), numerous proinflammatory cytokines and the activation of the transcription factor nuclear factor κB.16, 17 Nuclear factor κB is frequently activated in response to inflammatory mediators18 and has been shown to induce several chemokines (RANTES (CCL5), MIP-2 (CXCL2), MIP-1α (CCL3), monocyte chemoattractant protein 1 (MCP-1) (CCL2), interleukin-8 (CXCL8)) that are implicated in leukocyte migration.19, 20, 21, 22 Hypoxia in the tumor microenvironment is a cyclical event, and perpetuates the inflammatory response by ensuring a constant production of angiogenic and inflammatory mediators. Briefly, hypoxic conditions result in the generation of reactive oxygen species, which can increase DNA damage in neighboring cells. Tumor cells with more virulent mutations can then proliferate and invade neighboring cells/tissues resulting in tissue damage, which increases the demand for nutrients and oxygen, which continues to be short in supply.23 The reactive oxygen species induced in the microenvironment has also been demonstrated to increase secretion of inflammatory cytokines via nitric oxide (NO); NO has been linked to the regulation/induction of MIP-1α, MCP-1, macrophage inflammatory protein-related protein-1 and osteopontin.24, 25 Hypoxia has an important role in perpetuating the inflammatory process in tumors which results in the generation of chemokines that are involved in immune cell and likely MSC migration to tumors.
Archetype for MSC trafficking
The mechanism behind MSC migration is still in its infancy. However, factors involved in regulating migration of leukocytes have been studied extensively, and it is likely that many of the same factors are involved in regulating MSC migration. A list of receptors expressed on MSC that have been previously implicated in cell migration is shown in Table 1. Growth factor, cytokine/chemokine, adhesion molecules and toll-like receptors (TLRs) are expressed on MSC, as described by Chamberlain et al.,26 Ringe et al.,27 Allen et al.28 and Viola et al.29 Although the table does not encompass all receptors, it is a comprehensive list of those receptors that have been studied on other cell types, including leukocytes and HSC. The T cell, macrophage and dendritic cells are all considered relevant to the understanding of MSC homing because of their functional similarities in targeting inflamed/injured tissues. The HSC is considered because it is the precursor to the myeloid and lymphoid lineages in addition to being resident of the bone marrow where we also find the MSC.
There are three main participants in leukocyte trafficking: adhesion molecules, integrins and chemoattractants.30 The latter is the most important when considering the paracrine-mediated gradient trafficking and migration without cellular contact, such as what has been observed for MSC migration to sites of tumors and sites of injury.
Cytokine and chemokine receptors have an important role in leukocyte and likely MSC migration. A decrease in leukocyte migration has been observed in a mouse knockout model of the IL-8 receptors, CXCR1 and CXCR2.31 Similarly, neutrophil recruitment in ischemia-reperfusion models was inhibited by blocking the CXCR1 and CXCR2 receptors.32 The presence of both CXCR1 and CXCR227 on MSC suggests that they may have a similar function in MSC migration. Nearly every chemokine receptor has been found on the surface of MSC,3 while CCR2 and CCR3 are two receptors that may have a particularly important role in leukocyte and MSC trafficking. Macrophage trafficking has been shown to be mediated by CCR2, a chemokine receptor with ligands including MCP1, 2, 3 and 4 (MCP-1, 2, 3, 4, or CCL2, 8, 7, 13, respectively).30 Of note, blocking CCR3, the receptor for eotaxtin (CCL11), RANTES (CCL5), MCP2, MCP3 and MCP4 (CCL 8, 7, 13) has been effective in reducing trafficking in leukocytes. Yet another seven transmembrane receptor, CD97, an epidermal growth factor receptor appears necessary for the migration of neutrophils to sites of inflammation33 and has been implicated in the promotion of angiogenesis. Increased expression of CD97 in tumor cell lines, such as colorectal carcinoma, correlates with the increased migration potential of leukocytes.34 Of note, MSC express CD55, the ligand to CD97, suggesting a potential CD55–CD97 interaction thereby influencing both MSC and leukocyte migration.
Other than cytokines/chemokines, numerous bioactive molecules can also serve as leukocyte chemoattractants. Such molecules include lipids, (leukotrienes and prostaglandins), peptides such as chemerin and other elements of the extracellular matrix.35 The adhesion molecule, CD44, which is widely expressed on leukocytes and parenchymal cells interacts with components of the extracellular matrix known to be involved in inflammation including hyaluronic acid, collagen, laminin and fibronectin. The presence of CD44 is thought to mediate and enhance localized inflammation leading to the increased migration of leukocytes, it also serves as a receptor for growth factors, integrins, cytokines (osteopontin), glycosaminoglycans (hyaluronic acid), peptides (collagen), and is able to both induce inflammation and serve as a receptor mediator in the homing toward sites of inflammation.36, 37 Hyaluronan, which is a by-product of tissue repair, results in persistent inflammation38 and in addition to being critical to cell motility during development, the inhibition of hyaluronan synthesis in prostate tumors impairs growth and vascularization.39 Of importance, CD44, the receptor for hyaluronan, is expressed on MSC, and a recent publication suggests CD44 roles in the migration of MSC toward injured kidney tissue.40 In addition to CD44, hyaluronan by itself can interact with TLR2 and TLR4, thereby enhancing the inflammatory response.38 The participation and involvement of TLRs in MSC migration will be addressed to a further extent in a subsequent section.
Similar to the leukocytes, the migratory machinery of HSC are also well characterized in regards to defined receptors/ligands required for migration and should be considered in the evaluation of potential candidates for the elucidation of MSC migration. In fact, one of the most widely recognized receptor/ligand pairs for HSC trafficking is CXCR4/CXCL12 (SDF1); in addition, CXCR4/SDF1 appears important in dictating migration of several tumor cell lines to metastatic sites.41, 42 However, in contrast to the key role of CXCR4/SDF1 in HSC migration, a recent publication by Ip and co-workers showed blocking of the CXCR4 receptor had no impact on MSC migration, suggesting key differences between the migration signals of these two stem cells.43 This controversial receptor indicates that receptor function differs between cell types and enhances the importance of examining receptors found mutually expressed on MSC- and HSC-like CXCR4, CCR544 and VEGFR.45
MSC trafficking—finding the tumor
The past year has revealed a surge of publications attempting to define the homing properties of MSC. Due to the more comprehensive command of knowledge in the field of leukocyte migration, those chemokines and corresponding receptors chosen for evaluation on MSC is based on prior data illustrated in leukocyte models. However, as noted above, the migratory competence of these receptors clearly varies: CXCR4/CXCL12 are crucial in bone marrow retention and homing of HSC where as in MSC, it appears that CXCR4/CXCL12 do not posses the same migration importance unless enforced expression of CXCR4 is employed.43, 46, 47 Bearing in mind that each of the following receptors and their respective ligands (chemokines, cytokines, growth factors, peptides small biomolecules) have not been implicated as a single primary mechanism for the migration of MSC; however, altogether may function in an additive manner. Thus, the coalescence of known migratory mechanisms is vital to the understanding of the complete migratory disposition of the MSC in relationship to cancer biology.
Preconditioning the MSC
Activation of MSC with proinflammatory cytokines (that is, TNF-α) prior to reinfusion has been demonstrated to increase MSC in vivo migratory and adhesion capacity through the increased expression of receptors. Existing data suggest that the cytokines IL-1β and TNF-α activate adherence properties of the MSC including the upregulation of the VCAM-1-VLA-4 adhesion pathway.48 Additional receptors known to be upregulated by TNF-α priming include CCR3 and CCR4; these findings correlate with the observation of increased in vitro migration to RANTES (a CCR3 ligand) and macrophage-derived cytokine (MDC (CCL22)—a CCR4 ligand).3
Several growth factors have been shown to induce MSC migration. Insulin-like growth factor 1 (IGF-1) increases the expression of chemokine receptors on the MSC, thereby enhancing migration. Li and co-workers demonstrated that IGF-1-induced upregulation of CXCR4 expression on MSC increased the migratory capacity of the cells toward an in vitro SDF-1 gradient through a phosphoinositide-3 kinase-dependent pathway without altering the proliferation status of the MSC.49 The inability for MSC migration to occur in vivo through a CXCL12/CXCR4 mechanism was mentioned previously; accordingly, IGF-1 is capable of stimulating the expression of other chemokine receptors such as CCR5—the RANTES (CCL5) receptor.50 Other growth factors including basic fibroblast growth factor and VEGF are associated with angiogenesis and are secreted under hypoxemic stress. MSC demonstrate an increased migratory propensity in the presence of basic fibroblast growth factor through a phosphoinositide-3 kinase/AKT pathway downstream of the basic fibroblast growth factor receptor on the MSC.51
Chemokine/cytokine secretion and the respective receptors found on MSC
Monocyte-chemoattractant protein 1 (CCL2), a chemokine secreted by tumor cells, was shown to be a potent chemoattractant for MSC migration toward breast carcinomas.52 However, prior studies show conflicting data on the migratory response to CCL2. The inconsistent results may be attributed to differences between primary cells and passaged cell lines.27, 53 Additionally, Dwyer and co-workers examined the expression of CCL2 in breast-tumor explants revealing the cytokine presence not only in the tumor cells, but also surprisingly showing that the majority of the CCL2 detected was from the stromal fibroblast population.27 This finding further enforces the importance of the tumor microenvironment participation in MSC migration as well as the tumor cells themselves.
The cytokines, VEGFα and platelet-derived growth factor αβ (PDGFαβ) both harness chemoattractant properties; Ball and co-workers showed that VEGFα was able to stimulate migration through the PDGF receptor, confirming the intricacies involved in the induction of signaling pathways.54 In support of this, data from our group demonstrate the combined potential of PDGF and VEGF acting as chemoattractants inducing MSC migration in an in vitro matrigel migration assay; this combination of growth factors is more potent than either growth factor alone (Figure 1).
Inflammatory chemokine receptor expression on MSC is influenced by microenvironmental conditions. For example, the CC-, but not CXC-, chemokine receptors have been shown to be upregulated by TNF-α.3 This upregulation of certain chemokine receptors in response to cellular signals may have a role in tissue-specific homing.3, 27, 55 Additionally, the cytokines produced by the MSC themselves may not have a direct role in MSC migration, but may increase adhesion molecule expression in preparation for contact with the site of inflammation/injury as a mechanism of self-conditioning. In vitro co-culture experiments demonstrate that tumor-secreted factors influence expression of chemokines and chemokine receptors in MSC (Figure 2). Furthermore, different tumor cell lines produce different patterns of gene expression in MSC (Figure 3). Further investigation of these differentially expressed factors may shed light on mechanisms of MSC migration to tumors.
Additional receptors implicated in MSC migration
Toll-like receptors are a vital component of the innate immune response. There are 11 TLR found in humans, each recognize a conserved yet broad range of molecules known as pathogen-associated molecular patterns. Depending on the stimulation, downstream TLR signaling can regulate the expression of both CC and CXC chemokines via nuclear factor κB activation.56 Motif recognition varies between receptors: the dimers, TLR1/2 and 2/6 recognize lipopeptides and peptidoglycans; TLR3 recognizes double-stranded RNA; TLR4 recognizes lipopolysaccharide; TLR5 recognizes extracellular matrix molecules; TLR7 and 8 recognize synthetic antiviral compounds; TLR9 recognizes unmethylated CpG DNA.57 TLR signaling results in the activation of several pathways, MAPK, MyD88, c-Jun N-terminal kinase and inhibitor of κB kinase leading to the activation of the downstream activation of transcription factors nuclear factor κB and AP-1, which ultimately lead to the transcription of proinflammatory chemokines and cytokines.57 TLR1–6 have been identified on primary human MSC by both reverse transcription-PCR and flow cytometry.58 Tomchuck and co-workers reported that TLR stimulation enhanced the migratory function of MSC; the most potent migratory induction occurred with the stimulation of TLR3. Likewise, the inhibition of TLR3 signaling through a neutralizing antibody decreased MSC migration capacity by over 50%.
Manipulation of MSC migration
Augmentation of MSC migration via genetic manipulation of MSC
With the mechanistic understanding behind MSC migration being slowly deciphered, the potential for enhancing MSC-targeted therapies appears promising. The genetic manipulation of MSC to overexpress target receptors should enhance their migration to site-specific locations. The introduction of exogenous DNA into MSC, as reviewed by Damme and co-workers, enables enforced expression/secretion of a desired therapeutic factor into the targeted environment.59 One could imagine the potential in the overexpression of TLR3 and CCR2, the receptor for CCL2 (MCP-1) on MSC may improve their migration efficiency to specific tumor cells. Such a specific and directed approach will prove promising to the field of gene therapy for the treatment of cancers, allowing the cellular targeting of specific tumors, inflammatory diseases and other tissue injuries including myocardial infarction and ischemic cerebral damage.
The diverse collection of receptors present on MSC suggest the formation of a cascade of competitive events that enable a hierarchy of chemoattractants that are responsible for a step-wise chemotaxis to the tumor microenvironment. Conflicting chemotactic signals will lead to a biased migration to previously encountered chemoattractants in leukocyte-trafficking models.60 This heterologous signaling pathway activation based on the variation of surface receptors expression may justify both the preconditioning of MSC as discussed previously and the manipulation of surface receptor expression to enhance chemotaxis.
Alteration of MSC migration via changes in external environment
As discussed previously, inflammatory mediators have been shown to increase targeted trafficking. Inflammation induction as an element of targeted treatment enables a broader scope of secreted inflammatory molecules that may influence migration. This global enhancement of inflammatory chemokines, cytokines, along with tissue damage by-products including, lipids (leukotriene), glycosaminoglycans (hyaluran), enzymes, free radicals, complement and fibrinopeptides will exacerbate the migratory response seen in MSC. Mice with irradiated tumors, as compared with unirradiated tumors, show an increase in MSC migration. Klopp and co-workers recently demonstrated this in an elegant experiment upon which bilateral hind leg tumor implants were used. The tumor in the left hind leg remained an internal control, whereas the right hind leg received local irradiation treatment after which intravenous injection of MSC revealed a higher number of MSC localizing to the irradiated tumor (Figure 4). Similar experiments using HSC have shown an increased migratory propensity post-irradiation treatment of glioma cells. The attraction of HSC to irradiated glioma cells was attributed to an increase in stress signaling that induced hypoxia-induced transcription factor 1α transcriptional activity dependent on a functional TGF-β signaling cascade to induce CXCL12 promoter activity.61
Low-dose irradiation promotes VEGF release that is dependent on matrix metalloproteinase 9 expression, which is also upregulated under stress responses such as radiation treatment.62 The angiogenic factor, VEGF, is secreted by both the tumor cells and the tumor stroma, and is a known migratory inducer of both HSC and MSC.63, 64 In the presence of irradiated tumor cell-conditioned media, MSC increase their expression of VEGFa, VEGFc and PDGFb. The expression of these growth factors may be an autocrine feedback loop mechanism to enhance receptor expression, or to counter balance an increased apoptotic mRNA expression seen in these cells.
Targeted migration of MSC to tumor sites will have a significant impact on the field of antitumor therapy. MSC exhibit an intrinsic homing property enabling them to direct migration to sites of inflammation. The exploitation of this process will be a valuable asset to directed therapy. Their capability to express exogenous gene products, genetic stability and allogeneic properties make MSC excellent delivery vehicles for antitumor therapy, previously demonstrated not only in tumor models but also for other diseases such as graft-versus-host, multiple sclerosis and arthritis.9, 65, 66
The induced expression of receptors critical to the migratory competence of MSC to tumors will allow an increased number of MSC to reach the target location. The increase in migratory efficiency will improve the therapeutic value of the overall system.
There have been many advances in determining the factors involved in the migration of MSC to ‘wounded’/inflammatory/tumor environments; however, their full potential as therapeutic-vehicle candidates can only be utilized when the mechanistic understanding behind their migration is elucidated. Many different receptors have been implicated in the homing of MSC: (1) the broad activation of growth factor receptors that activate further chemokine receptor expression like CXC and CC receptors; (2) the activation of TLR that also target downstream expression of CXC and CC receptors; (3) the activation of adhesion molecules and (4) integrins that may or may not be implicated in the direct role of paracrine cell movement. The key players implicated in MSC migration to date include the chemokines MCP-1 (CCL2),52 CXCL8,27 RANTES (CCL5);3 LL-37,58 integrinβ1,43 receptors CD44,40 CCR2,3 CCR3,3 and the receptor tyrosine kinases for the following growth factors, IGF-1,3 PDGF-bb,67 HGF,68 and VEGF.54
The ability to extract from previously described cell-migration models in the elucidation of MSC homing properties is an arduous task. Many common receptors have been identified on MSC; however, multiple ligands and co-receptors have the ability to alter the downstream signaling pathway through coupling, crosstalk or inhibitory mechanisms thereby rendering an alternative mechanistic feature for a common ligand/receptor pair depending on the cell-lineage. The future of targeted therapy using MSC will depend on the exploitation of these previous ligand/receptor interactions for the enhancement of the existing intrinsic migratory/homing propensity.
This work was supported in part by grants from the National Cancer Institute (CA-1094551 and CA-116199 for FCM, CA-55164, CA-16672, and CA-49639 for MA) and by the Paul and Mary Haas Chair in Genetics (MA). ES, AK and FCM are supported in part by grants from the Susan G Komen Breast Cancer Foundation.
About this article
Journal of The Royal Society Interface (2019)