Molecular mechanisms of migration and homing of intravenously transplanted mesenchymal stem cells

Irina V. Kholodenko1*, Alina A. Konieva2, Roman V. Kholodenko3 and Konstantin N. Yarygin1 *Correspondence: irkhol@yandex.ru 1Orekhovich Institute of Biomedical Chemistry of the Russian Academy of Medical Sciences (Moscow), Russian Federation. 2North Ossetian State Medical Academy (Vladikavkaz), Russian Federation. 3M.M. Shemyakin–Yu.A. Ovchinnikov Institute of Bioorganic Chemistry of the Russian Academy of Sciences (Moscow), Russian Federation.


Introduction
At present stem cell-based technologies comprise the main trend in the emerging field of regenerative medicine, while mesenchymal stem cells (MSCs) isolated from different tissues are regarded as one of the most handy and versatile raw materials for preparation of cell therapy and tissue engineering products. Bone marrow-derived mesenchymal stem cells (BM-MSCs) are definitely the best characterized type of mesenchymal cells, although attention is gradually switching to alternative sources of MSCs, such as adipose tissue, placenta, umbilical cord/ Wharton's jelly, umbilical cord blood, and others.
Regardless of the source, cultured MSCs have common characteristics and behavioral traits. The ability to plastic adherence is widely used for their isolation from different tissues [1]. All MSCs have similar fibroblast-like morphology in vitro and are able to differentiate into cells of mesodermal lineage, namely, osteocytes, chondrocytes and adipocytes. MSCs from diverse sources may show certain differences in the expression of surface markers. However, in general, the phenotypes of these cells are very similar, and the typical mesenchymal stem cell surface markers are CD44 (receptor for hyaluronic acid), CD54 (ICAM-1; Intercellular Adhesion Molecule 1), CD90 (Thy-1), CD105 (endoglin), CD106 (VCAM-1; vascular cell adhesion molecule 1), HLA-I (MHC class I). Also MSCs do not express hematopoietic and endothelial cell markers, such as CD31, CD34, CD38, CD45, CD133, MHC class II [2,3].
Furthermore, MSCs have unique immunomodulatory properties [4]. In vitro co-cultured MSCs modulate differentiation of certain immune precursor cells and/ or maturation of differentiated immune cells [5]. The in vitro results are in accordance with immunomodulatory action seen both in animal experiments [6] and in clinical studies. Therefore, MSCs are being tested in clinical trials to reduce graft-versus-host disease, and for the treatment of autoimmune disorders such as multiple sclerosis [7,8].
An important property of MSCs is their capacity for migration and homing in or around the zones damaged by ischemia, inflammation, and trauma or tumor sites. Understandably, the efficacy and time course of cell invasion into the damaged tissues depends upon the route of transplantation. Many ways of MSCs administration have been tested in cell therapy experiments and clinical trials. Among those, local (eg., directly into ischemic myocardium or brain), intra-arterial or epidural routes of cell administration usually demand at least minor surgery and are often associated with major risks. Iv infusion is probably the most convenient way of MSCs delivery both in experiments and clinical trials. This method is minimally invasive and allows repeated injections of multiple doses of cells at certain time intervals. Moreover, it possibly imitates the likely physiological mechanism of bone marrow-derived stem cell participation in regeneration 2 doi: 10.7243/2050-1218-2-4 of different tissues [9].
It is now well documented that iv transplanted cells migrate in substantial quantities into the injured tissue [10]. It is noteworthy, though, that transplanted MSCs move not only to the areas of tissue damage, but to at least some stem cell niches including the subventricular zone and the hippocampal subgranular zone in the brain [11]. Here this very important migration route is just mentioned, but not systematically considered.
It is important to understand the mechanisms underlying the migration and homing of MSCs. This review presents recent data concerning the distribution of MSCs in the body after iv administration, as well as possible mechanisms of their migration and homing. It also includes some data related to other routes of MSCs transplantation, but critical for understanding the essential features of their relocation within intact and damaged tissues.

MSC distribution after transplantation
One of the main issues relating to the clinical use of stem cells is the way to deliver them to the target tissue. As mentioned above, there are several routes of cell transplantation including the site-specific, directly into the site of injury or near it, and the systemic, into the bloodstream. Both of these methods have been tested in pre-clinical studies in animal models and in clinical practice.
A number of publications demonstrated that the sitespecific, or site-directed, delivery of MSCs resulted in effective engraftment of target tissues by transplanted cells, especially in the damage/inflammation conditions. Thus, in several studies effective functional recovery was observed after BM-MSCs injection directly into the ischemic myocardium [12] . Gojo S. et al., [13] inoculated mouse MSCs into the myocardium of healthy adult animals and found signs of neoangiogenesis near the injection site 7 days after transplantation. In this case, the donor cells were found in the walls of the newly formed blood vessels. They also differentiated into cardiomyocytes, endothelial cells and pericytes, indicating ability to engraft not only in damaged, but in healthy tissue, too.
Hofstetter et al., [14] injected rat MSCs into the spinal cord of rats suffering from paraplegia 1 week after injury, and found that the transplanted cells were concentrated at the core of injury site and stimulated the recovery of transected spinal cord. In our work [15], we injected human placental MSCs labeled with fluorescent microparticles into the mouse spinal cord immediately after injury. We observed that transplanted human cells at 30 days migrated into the mouse spinal cord, and these cells mostly accumulated at the injection site and on the dorsal side of the spinal cord. In addition, the labeled human cells migrated to the rostral and caudal directions from the injury site. A small part of the transplanted cells migrated deeper into the spinal cord tissue and spread within a few millimeters from the injury. Thus, at least a fraction of the injected human placental MSCs survived 1 month in the mouse body and migrated within the host spinal cord, promoting functional recovery of animals after spinal cord injury. Andrews et al., [16] injected human bone marrow MSCs in the ipsilateral hemisphere after one-sided middle cerebral artery occlusion in rats and discovered that at 30 days after transplantation viable human BM-MSCs spread out in the cortex of ipsilateral and contralateral hemispheres.
As follows from the above examples, the site-directed delivery of MSCs provides effective engraftment of area close to the injection site and allows transplanted cell migration to relatively distant locations. Apparently, it creates maximal accumulation of cells in the site of injury, probably contributing to the effectiveness of transplantation. However, it has significant drawbacks and, hence, is not always applicable in clinical practice. Firstly, the injection of MSCs directly into tissues other than skin and mucosal linings requires surgery, which is bothersome, associated with risks, and restricts the possibility of using multiple doses of cells at certain intervals of time. Secondly, local MSCs transplantation causes additional, albeit minor, tissue damage. Finally, most of the cells form dense clusters at the injection site producing undesirable consequences, such as calcification spots in myocardium after intramyocardial injection of MSCs isolated from bone marrow [17].
Cell administration into the bloodstream is more suitable for clinical use. Intra-arterial infusion may be quite effective with respect to attaining maximum cell concentration in the area supplied by a particular artery [18,19]. However, it also requires at least minor surgery and may end up with plugging of local capillaries with transplanted cells.
Iv administration is safer than the above mentioned transplantation routes and at the same time allows cell delivery to the majority of tissues. The main advantages of this method of MSC transplantation are the following: 1) availability and the absence of any contraindications, as opposed to site-directed transplantation; 2) the possibility to inject high doses of cells for a long time, for example, using a dropper, and the ability to infuse multiple doses at certain intervals; and 3) absence of the necessity of surgical intervention. However, this method of stem cell administration has one major drawback, which lies in the fact that not all the injected cells, but rather their smaller number, reach the site of injury.
A number of studies have described the iv route of MSCs administration. Barbash et al., [20] iv injected rat MSCs to animals with myocardial infarction, and found that most of the cells penetrated into the lungs, and a small cell number engrafted into the heart, liver, and spleen. Only some of the MSCs retain their ability to homing into the site of damaged myocardium, however, much less than after intracardiac injection. Many studies after iv injection of cultured MSCs observed accumulation of donor cells in the lungs. At present, there is evidence that the MSCs iv injected into a healthy animals, mostly arrested in doi: 10.7243/2050-1218-2-4 the vascular system of the lungs [21]. It is likely that this phenomenon can be explained by the fact that MSCs expanded in culture are activated and express a great variety of adhesion molecules; moreover these cells have a relatively large size. However, Gao et al., [22] found that treatment of animals with vasodilating sodium nitroprusside reduced the number of cells penetrated into the lungs.
Despite the fact that MSCs can be arrested in the lungs, there is lot of evidence that showed that MSCs are able to home to the injured tissues after iv administration. Cultured rat and human MSCs migrate to sites of brain damage after cerebral ischemia, when iv transplanted into rats [23,24]. Wu et al., [25] transplanted rat MSCs iv for the treatment of cardiac allograft rejection in rats and found that they vigorously migrated into the site of allograft rejection, mainly differentiated into fibroblasts and a small number of transplanted cells differentiated into myocytes. MSCs have also been used for the treatment of lung injury in mice when administered iv. Ortiz et al., [26] found that mouse MSCs after transplantation into the mice exposed to bleomycin (model of pulmonary fibrosis) are homing to the lungs in response to injury, where they acquired epithelium-like phenotype, and reduced inflammation and collagen deposition. They also found a 23-fold increase in the engraftment level of donor cells compared to mice that were not exposed to bleomycin.
Analysis of distribution after iv MSCs administration was performed by different methods: 1) by fluorescently labeled proteins [27,28,29]; 2) by detection of human genes into animal recipients [30,31]; 3) by histological methods [32]; and 4) by immunohistochemistry [33,34]; 5) by fluorescence in situ hybridization [35]; and 6) by realtime PCR [26,27]. For example, Devine et al., [27] iv injected MSCs transduced with a green fluorescent protein (GFP) retroviral construct into the lethally irradiated baboons. At 9-21 months after MSCs infusion they evaluated for the presence of the GFP transgene in purified genomic DNA by sensitive real-time polymerase chain reaction (PCR). Moreover, two baboons were treated with autologous cells, and one animal was treated with allogeneic MSCs. In both cases, the distribution of transplanted cells was similar. Since it has been shown that most of the transplanted cells found in the gastrointestinal tract and the relatively high number of cells was also observed in the kidney, lung, liver, thymus and skin. The level of engraftment in these tissues was approximately 0.1% to 2.7% of injected cells. In the intact model (health animals) due to detection of GFPtransfected murine MSCs Deak et al., [29] demonstrated that most GFP-positive organs are the lungs, liver, kidneys, skin, and intestines as examined at 24 hours after MSCs transplantation.
However, the aforementioned methods are invasive and static. The development of non-invasive techniques such as magnetic resonance imaging (MRI) of MSCs labeled with superparamagnetic iron oxide nanoparticles (SPIO) [36], the combined single-photon emission CT (SPECT)/CT scanning [37] and tracking quantum dots [38], have improved our ability to investigate homing of MSCs and their fate and organ-specific accumulation. Cell tracking by MRI using SPIO is a method with the least risk to monitoring for the activity of stem cells in humans. Hsiao et al., [39] showed that MSCs successfully labeled Ferucarbotran (ionic SPIO, which is used in the clinic) without the aid of transfection agents, which does not affect cell viability, proliferation, changes in mitochondrial membrane potential, production of reactive oxygen species, or the ability to differentiate. Approximately 45.2% of the labeled MSCs can be detected with a single-level 3D gradient echo sequence and four repetitions using 1.5T MRI. Using SPECT/CT imaging in a model of acute myocardial infarction, Kraitchman et al., [37] demonstrated that the initial localization of BM-MSCs was observed in the lungs and the cells migrated to the organs that are not targets, such as the liver, kidney and spleen at 24 and 48 h after administration. Increase of the number of MSCs in the heart tissue that is affected by a heart attack, occurred simultaneously with a gradual decrease of the initial MSCs concentration in the lungs 24 hours after infusion, and MSCs persisted up to 7 days after injection. In addition, the labeling of BM-MSCs with bioconjugated quantum dots does not alter their self-renewing and differentiation potential toward chondrogenic, osteogenic and adipogenic lineages [40]. Moreover, it is very useful not only for tracking MSCs, but also in the study of behavioral changes in cells when MSCs are injected in combination chemicals, such as heparin [38]. When imaging the mice with acute liver failure Yukawa et al., [38] showed that almost all of the transplanted adipose-derived MSCs (AdMSCs) during 10 minutes accumulate in the lungs in the absence of heparin treatment. However, when heparin is used in combination with AdMSCs, accumulation of transplanted cells is found not only in the lungs, but also in the liver, and the number of accumulated cells increased up to 30% in the injured liver. Thus, studies that use different methods of tracking MSCs showed initial concentration of MSCs in the lungs after iv administration [22,29,37,38,41], then the majority of MSCs gradually migrated to sites of injury [29,37,38,41] or in liver, spleen, kidneys and bone marrow [22].

Mechanisms of migration and homing
Homing is the ability of cells to migrate into the "right" place, "home" organ and/or tissue (stem cell niche), or in the damaged area. Homing to specific organs is a multistep process that includes cell attachment and rolling in the vessel lumen, adhesion and extravasation across the vascular endothelium and migration through the tissue stroma. Migratory ability of the cells is due to biochemical signals from the "right" area, reception system and the ability of cells to chemotaxis. Based on the foregoing, it can be concluded that the mesenchymal stem cells, regardless of the mode of administration, have a high doi: 10.7243/2050-1218-2-4 potential for migration and homing to sites of injury [15,19,20]. However, in order to improve this ability and, as a consequence, increase the therapeutic efficacy of these cells it is necessary to study in detail the mechanisms of migration and homing, and possibility to regulate these processes.
There is much evidence to support the theory that MSCs are homing to different tissues, especially to injured or inflamed tissues. This process involves the migration of cells in the blood stream (probably by rolling), transendothelial migration through the vascular walls and the invasion into the target tissue. The mechanism by which MSCs directed to the tissues and migrate across the endothelium, is not yet fully understood, but it is likely that the damaged tissue expresses specific receptors or ligands to facilitate traffic, adhesion and infiltration of MSCs to the site of injury, as well as with the involvement of leukocytes to sites of inflammation. Chemokine receptors and their ligands are essential components involved in the migration of leukocytes to sites of inflammation, and it was recently shown that MSCs also express some of these molecules. In addition, it is known that some adhesion molecules and integrins involved in leukocyte rolling and transendothelial migration, are also expressed by MSCs. Moreover, for invasion into the targeted tissues MSCs use certain extracellular proteases similar to immune cells, as well as metastatic cancer cells, which use different proteases to penetrate the tissue extracellular matrix, such as matrix metalloproteinases (MMPs), serine, cysteine and aspartate proteases [42][43][44]. Data about the involvement of certain molecules in the MSC migration/homing summarized in Table 1.

Chemokines and their receptors
As discussed earlier, MSCs have the ability to migrate into the tissue from the circulation, possibly in response to signals that are induced by injury. Although the mechanisms by which MSCs attracted in tissue and transmigrate through vascular endothelial cell layer are not yet fully understood, it is possible that this process involved chemokines and their receptors, similar to hematopoietic stem cells and immune cells [45].
Chemokines is a large family of proteins that regulate many functions, such as chemotaxis. While the majority of chemokines are presented at secreted form and consist of 67-127 amino acid residues, some chemokines are membrane-bound (eg., CXCL16 and CX3CL1). Chemokines are classified into four groups according to the spatial location of the first two cysteine residues near the NH2end: CXC chemokines, CC chemokines, C chemokines and CX3C chemokines. Chemokine receptors are classified as receptors associated with G-proteins for CXC, CC, C or CX3C chemokines [46].
Previously, several groups have studied the expression of chemokine receptors on human MSCs derived from bone marrow. Wynn et al., [56] evaluated the expression of CXCR4 on hMSCs isolated from bone marrow (hBM-MSCs), and showed that the receptor is present on the cell surface membrane of less than 1% of the cells; although they found, that 83% -98% of the hBM-MSCs expressed intracellular CXCR4 at high level. Other studies have shown functional expression of CXCR4 at the surface of BM-MSCs [57]; expression of CCR1, CCR7, CXCR4, CXCR6, and CX3CR1 was shown on individual cells (2-25%) [52]. Honczarenko M. et al., [58] performed flow cytometric analysis of the expression of CC chemokine receptors (except for CCR10), CXC receptors and CX3CR1. They found, that 43%-70% of the hBM-MSCs expressed functional (as determined by chemotaxis assay) CCR1, CCR7, CCR9, CXCR4, CXCR5 and CXCR6. Another group showed expression of CCR2, CCR8, CXCR1, CXCR2, and CXCR3 in BM-MSCs by RT-PCR and immunohistochemistry [59]. Ponte et al., [60] demonstrated that hBM-MSCs expressed CCR2, CCR3, CCR4, and CXCR4, and found that TNF increased the expression of CCR2, CCR3 and CCR4, but not CXCR4.
Despite contradictory data obtained in different studies concerning expression of several chemokine receptors, it is obvious that MSCs express a great variety of these receptors. Such contradictions may be due to several factors: the heterogeneity of ex vivo expanded MSCs, which is a feature of these cells, the differences in the methods of MSCs isolation and cultivation and the number of passages that have undergone cells before analysis. In this context, the number of passages is particularly important in the analysis of chemokine receptor expression. Thus, ex vivo cultured MSCs, which were initially positive for CCR1, CCR7, CCR9, CXCR4, CXCR5, and CXCR6, at passage 12-16 loss of their surface chemokine receptor expression, which was accompanied by a lack of chemotactic response to chemokines, and by further decline in the expression of adhesion molecules, including ICAM-1, ICAM-2 and VCAM-1 [58]. In addition to the long-term cultivation of MSCs and to cultural conditions, changes in the expression pattern of surface receptors dependent on confluency of cells, site of isolation and the environment in the process of incubation (normoxic or hypoxic conditions) [61].
Furthermore, the fact that MSCs derived from different sources, express distinct set of chemokine receptors may indicate that they have a different ability to migrate into various tissues. For example, similar to the white blood cells, MSCs expressing CCR9 can use this receptor for homing to the intestine, and CCR1-expressing MSCs can use the receptor for homing into inflamed tissue of the joints in rheumatoid arthritis, or in the brain in multiple sclerosis.
The most studied of chemokine/receptor pair in cell migration and homing are CXCL12, or SDF-1 (stromal cell-derived factor-1) and its receptor CXCR4. It is known that CXCL12 is a major chemotactic factor that promotes hematopoietic cell homing into the bone marrow at the development process, maintenance them in the adult bone marrow, as well as migration of mature lymphocytes towards sites of inflammation and injury [62]. In addition, CXCL12 and its receptor CXCR4 are involved in the migration of primordial germ cells [63], as well as in the attracting endothelial progenitor cells into ischemic tissue [64].
However, in the MSCs migration and homing this chemokine and its receptor play a less significant role, except when the expression of CXCR4 was stimulated by various factors. For example, it was found that IGF-1 is not only itself induced migration of MSCs, but also initiated the migratory activity of MSCs in response to SDF-1α through upregulation of CXCR4 [65]. Interestingly, the BM-MSCs after stimulation with TNF-α also showed increased susceptibility to SDF-1α, despite on the invariably of SDF-1α surface binding and CXCR4 expression. Moreover, shortterm treatment of BM-MSCs with cytokine cocktail, which includes Flt-3 ligand, stem cell factor (SCF), IL-6, hepatocyte growth factor (HGF) and IL-3, induced a significant upregulation of surface and intracellular CXCR4 [66].
Song et al., [67] found that tumor conditioned medium induced up-regulation of CXCR4 on BM-MSCs, so that the cells acquire capacity to increased migration toward tumor cells. However, AMD3100 (a specific inhibitor of CXCR4) is not completely abolished migration of BM-MSCs to tumor cells PC3 (human prostate cancer cell line), MCF-7 (breast cancer cell line) and RIF-1 (fibrosarcoma), which suggests that there must be other factors involved in the specific migration of BM-MSCs to tumors. Interestingly, the CXCR4 inhibitor had no effect on the migration of BM-MSCs toward human prostate cancer cell line DU145, showing that different tumors may use different mechanisms to doi: 10.7243/2050-1218-2-4 attract BM-MSCs. In addition, tumor conditioned medium induced the production of SDF-1α by human BM-MSCs [68]. Gao et al., [69] show that SDF-1 acts as a signaling molecule and not necessarily as a chemotactic factor to facilitate hMSC migration towards tumor environment. It is possible that SDF-1 acts in conjunction with other factors, including cytokines such as IL-8 and IL-6, to provide hMSC migration in response to different stimuli. It is possible to assume that the expression levels and distribution patterns of CXCR4 may be regulated by various cytokines and growth factors. It is known that at the damage of various tissues, such as myocardial infarction [70], ischemic stroke [71] and others, in the site of injury produced a wide range of cytokines and growth factors that can simultaneously serve as chemoattractants for MSCs, and stimulate these cells to expression of different chemokine receptors, particular CXCR4. The same can be said for solid tumors, which are often seen as unhealed wound [72], which is a continuous producer of inflammatory mediators (cytokines, chemokines and other potential chemoattractant molecules).

Cytokines, growth factors, and receptors
In addition to chemokines and their receptors, different cytokines and growth factors play an important role in the MSC homing. It is known that during the inflammatory process damaged tissues release a large number of cytokines, including IL-6, PDFG-BB and IGF-1 [73]. Thus, a number of studies have shown the active role of these cytokines in the migration of BM-MSCs [74]. Moreover, PDGF-BB shows greater attracting potential, than the other cytokines. Other inflammatory cytokines, such as transforming growth factor (TGF)-β1, interleukin (IL)-1β and tumor necrosis factor (TNF)-α, stimulate the production of matrix metalloproteinases (MMPs) by the MSCs, leading to powerful stimulation of chemotactic migration through the extracellular matrix [75]. Human adipose-derived MSCs (AdMSCs) migrate in response to various growth factors and cytokines, including platelet-derived growth factor (PDGF)-AB, TGF-β1, TNF-α, and SDF-1α [76]. Spaeth et al., [73] demonstrated that the combination of PDGF and VEGF acts as chemoattractants, inducing migration of MSCs in vitro; this combination of growth factors is more powerful than each growth factor separately. These results show that increasing of MSC homing can be achieved by modulating their response to different growth factors and cytokines, which may allow improving their therapeutic potential.
Toll-like receptors (TLR) is an essential component of the innate immune response. In humans, there were found 11 TLRs, each of which recognizes a wide range of conservative molecules known as pathogen-associated molecular patterns. Depending on the stimulation TLR signaling may regulate the expression of both CC and CXC chemokines by activation of nuclear factor κB [77]. Dimers of TLR1/2 and 2/6 recognize lipopeptides and peptidoglycans; TLR3 recognizes double-stranded RNA; TLR4 recognizes lipopolysaccharides; TLR5 recognizes the extracellular matrix molecules; TLR7 and 8 recognize pathogen-associated molecular patterns (PAMPs) that are expressed on infectious agents; TLR9 recognizes unmethylated CpG sequences in DNA molecules [78]. TLR1-6 have been identified on primary human MSCs [79]. Waterman et al., [80] shown that stimulation of TLR enhanced MSC migration function; the most powerful induction of migration occurs when stimulated TLR3 and TLR4, and their inhibition reduced the migration ability of MSCs by more than 50%. Nurmenniemi et al., [81] showed constitutive expression of TLR9 in human BM-MSCs, and stimulation of the cells with TLR9 agonist, CpG oligonucleotides, increased the expression of MMP-13 and enhanced the ability of these cells to invasion in vitro.

Integrins and adhesion molecules
One of the most important stages of cell homing, including MSCs, is their movement in the bloodstream and transmigration through the vascular endothelium. This multi-step process includes several types of adhesion molecules, cytokines and proteases [82]. It is assumed that MSCs, moving through the blood vessels, use the same mechanism of rolling, as a white blood cells and HSCs. In order to leave the blood stream, mature lymphocytes and hematopoietic progenitor cells undergo a series of coordinated adhesion steps, which are mainly mediated by selectins and their ligands [83]. Then the cells roll along the vascular endothelium and at some point encounter chemokines, which ultimately activate integrins, leading to arrest of the cells in a certain place of vasculature and their subsequent transendothelial migration.
Rüster B. et al., [84] demonstrated that hBM-MSCs interact with endothelial cells, including rolling with rapid extension of podia and activation of selectin-dependent and integrin-dependent binding. The authors contend that although hBM-MSCs do not express P-selectin glycoprotein ligand-1 (PSGL-1) and an alternative ligand CD24, they interact with endothelial cells in the P-selectindependent manner. At the same time, in contrast to HSCs, human MSCs do not express or express small amounts of E-and L-selectins, and their contribution to traffic MSC is not significant when compared with the P-selectin. That would suggest that MSCs express a new carbohydrate ligand, which can serve counterligand for P-selectin expressed on endothelium. Also these data suggest that hMSCs, like white blood cells, rolled on endothelial cells at the first stage of their recruitment [84].
The majority of leukocyte adhesion molecules, such as β2 integrin-related molecules, are not expressed on the surface of MSCs [85]. However, MSCs express very late antigen-4 (VLA-4, α4β1 integrin) and vascular cell adhesion molecule-1 (VCAM-1) [84]. It was shown that β1 integrins are important for the intramyocardial traffic of MSCs [86] and, as mentioned above, are crucial in the rolling and adhesion of human MSCs [84]. Moreover, an anti-VCAM-1 antibody reduces the adhesion of rat MSCs to endothelial cells of microvessels [87]. In addition, it was found that VCAM-1 and VLA-4 are key receptors for transendothelial migration of human MSCs. Because transmigration of MSCs is not blocked completely by anti-VCAM-1 and anti-VLA-4 antibodies, it is possible that this process may involve other integrin α-chain.
Vascular endothelium forms a major barrier to the transmigration of molecules and cells from the blood into the surrounding tissue. It has been shown that endothelial cells actively modulate leukocyte diapedesis [90]. To date, there is very little data concerning the effects of vascular endothelium on MSC transmigration. In various areas of vasculature endothelium is functionally different, and, therefore, varies its barrier function [91]. It is known that various endothelial phenotypes reflect functional differences. For example, variations in the number of intercellular tight junctions exist at various points throughout the vasculature and are the cause of differences in endothelium permeability [91]. Steingen et al., [92] showed that endothelial phenotype strictly influences on transmigration of human bone marrow MSCs in vitro. Coronary artery endothelium enables the most rapid integration of MSCs in vitro. In contrast, MSC transmigration through the venous endothelium was most effective over time. On the other hand, the adhesion of rat MSCs to endothelial cells of the aorta and cardiac microvascular endothelial cells did not differ significantly [87]. In summary, the above data demonstrate that the rolling, transmigration and invasion efficiency of MSCs into different organs depend on the patterns of adhesion molecules and integrins expressed by MSCs, and may vary depending on the type of surrounding vessels.

Proteases
MSC invasion into the targeted tissue requires not only their transmigration through the endothelial barrier, but also through the basement membrane and extracellular matrix. It is known that leukocyte transmigration is due to the binding of α4β1 integrin on the surface of white blood cells with VCAM-1 on the surface of vascular endothelium in the site of inflammation, which then induces the expression of matrix metalloproteinase MMP-2, which facilitates invasion of the subendothelial matrix [93]. Matrix metalloproteinases (MMPs) are a family of endoproteinases that can cleave all of the extracellular matrix components and, thus, contribute to cell migration through the matrix. This system is regulated by the interaction of MMP with four tissue-specific inhibitors of MMPs (tissue inhibitor of metalloproteinase, TIMP). The balance of MMP/TIMP plays an important role in angiogenesis, proliferation, apoptosis and invasion. High expression of MMP associated with low levels of TIMP secretion promotes a higher rate of metastasis or tumor growth [75,94].
Son B.R. et al., [95] demonstrated that BM-MSCs and umbilical cord blood MSCs (UCB-MSCs) express membrane type of MMP-1 and MMP-2, which cleaves the main component of basement membrane collagen IV. In addition, De Becker et al., [82] showed that MMP-2, but not MMP-9, which is not expressed by these cells, plays a role in the MSC migration through the bone marrow endothelium. Steingen et al., [92] confirmed that MSCs do not express MMP-9 at the protein level and showed that in the myocardium active MMP localized precisely in areas of MSC transmigration and invasion, showing that during homing in cardiac tissue MSCs secrete MMP-2 to transmigrate through the basement membrane and penetrate the surrounding extracellular matrix.
Tondreau et al., [74] using qRT-PCR and protein analysis demonstrated constitutive expression of MMP-2, TIMP-1 and TIMP-2, but not MMP-9, and a very low level of TIMP-4 by BM-MSCs. The authors also found that the mRNA levels of MMP/TIMP in response to PDGF-BB, IL-6, IGF-I and SDF-1α significantly varied. Although MMP-2 and MMP-13 production is not always increased significantly in response to these chemokines/cytokines, variable levels of TIMP-1 and TIMP-2 were observed, confirming their participation in the migration process. Thus, the migration of BM-MSCs in the presence of PDGF-BB was the result of decreasing production of TIMP balanced by increasing levels of MMP-2. High migration of MSCs observed in the presence of IL-6 is due to the increasing level of MMP-13 without modulating of TIMP. At the same time adding SDF-1α did not affect the production of MMP/TIMP and correlated with the absence of migration. On the other hand, Song and Li [67] showed that normally BM-MSCs did not express MMP-2. However, MMP-2 expression is induced by treatment of BM-MSCs with tumor conditioned medium as early as 2 h, and after 24 hours its amount decreases, which indicates that the expression of MMP-2 is strictly regulated, possibly through tissue inhibitor of matrix metalloproteinase (TIMP)-1 and TIMP-2 at various stages of homing/engraftment BM-MSC The results presented in this chapter clearly indicate that MSCs similar to HSCs use matrix metalloproteinases, such as MMP-2, for transendothelial migration and invasion in the targeted tissue.

Conclusions
Despite the fact that the process of cell homing, particularly white blood cells and hematopoietic progenitor cells has been studied in detail, it is not possible to extrapolate the available knowledge on the migration and homing of mesenchymal stem cells. As can be seen from the presented review, the milestone events in homing of different cell types are the same, but the molecular mechanisms of these stages are different. Moreover, MSCs derived from different sources express different pattern of chemokines and their receptors, which may indicate the possibility of a preferred migration of different MSCs into various tissues. In addition, ex vivo culturing conditions of MSCs and number of passages can have a significant effect on the expression of chemokines/receptors as well as matrix metalloproteinases, which in turn affects the ability of the cells to migration and homing.
Thus, in order to create the most complete picture of the mechanisms of MSC migration and homing require more detailed studies aimed at investigating the various molecular signaling pathways involved in these processes. A more complete understanding of the mechanisms of MSC homing and migration will help effectively apply these cells in clinical practice and regenerative medicine when administered systemically. The possibility to modulate the migration ability of these cells, for example, by inflammatory cytokines, opens broad prospects for the development of directed migration of greater number of iv injected cells to the injury site.