Mesenchymal Stromal Cell Migration: Possibilities to Improve Cellular Therapy

Introduction

M esenchymal stromal cells (MSC) were originally identified by Friedenstein et al. as the bone marrow (BM) stromal cells supporting hematopoietic stem and progenitor cells [1]. MSC represent a very heterogeneous cell population. Some of these cells have a high proliferative potential, generating colonies when plated in tissue culture at a low density. These colonies are the so-called colony forming unit-fibroblast (CFU-F) [2]. Another hallmark of MSC is their ability to differentiate into several mesenchymal lineages [3 –5].

MSC are a minor fraction in BM; the frequency in human BM has been estimated at 0.001%–0.01% of the total nucleated cells [3]. MSC frequency seems to decline with age, from 1/10,000 nucleated BM cells in a newborn to about 1/1,000,000 nucleated marrow cells in an 80-year-old person [3]. Besides BM, also other tissues contain MSC. They can be obtained from almost every postnatal organ [6,7], including adipose tissue [8,9], the periosteum [10], brain [11], liver [6], skeletal muscle [11], hair follicles [12], peripheral blood [13], umbilical cord blood [14], and Wharton's Jelly [15] as well as fetal tissues [16 –18].

It has been reported that a very small number of MSC consistently circulate in the blood. This circulating pool was found increased under hypoxic conditions [19], in case of major injury [20,21], or large skin burns [22], suggesting that MSC can be mobilized and can migrate to places where they are needed. However, contrasting reports exist on the presence of MSC in (mobilized) peripheral blood and the success rate of identifying any circulating MSC is low (reviewed in Ref. [23]).

MSC have shown to possess the ability to migrate to sites of inflammation and injury, for example, in an animal model of cerebral ischemia [24]. In a model of multiple organ failure, Chapel et al. showed that tagged MSC homed to numerous tissues with localization correlating to the severity and site of injury [25]. The homing efficiency of infused MSC has been reported to be greatly influenced by the variety of protocols currently used for isolation and culture expansion of MSC. Studies by Rombouts and Ploemacher [26] demonstrated that primary BM-derived MSC were able to effectively home in irradiated mice, whereas cultured MSC had lost this homing capacity already after 24 h of culture. Further, culturing of MSC has been associated with a decrease in expression of adhesion molecules, the loss of chemokine receptors, and a subsequent lack of chemotactic response [27,28]. Since migration of MSC toward the site where they are required is crucial for most clinical applications, in this review we aim to highlight the current understanding of MSC migration and to identify further opportunities to improve the clinical benefits after MSC transplantation via their homing/migratory capacity.

Characterization and Phenotype(s) of MSC

Although in situ characterization of MSC populations is still in its infancy, the International Society of Cellular Therapy (ISCT) has postulated a definition for ex vivo expanded MSC to be used in clinical studies, based on 3 main characteristics: (1) their adhesion to plastic; (2) their expression of a specific set of membrane molecules CD105, CD90, and CD73 together with lack of expression of the hematopoietic markers CD14, CD34, and CD45 and HLA-DR; and (3) trilineage differentiation into osteoblasts, adipocytes, and chondrocytes in vitro [29]. Recently, a new set of MSC markers has been identified (CD140b, CD146, CD271, CD340, and CD349) [30 –33] that enable prospective isolation of MSC and enrichment of CFU-F from human BM. Based on expression of CD271 and CD146, we and others identified 3 subpopulations of MSC, co-expressing classical MSC markers CD105 and CD90 to be present in BM: CD271⁺/CD146⁺, CD271⁺/CD146⁻, and CD271⁻/CD146⁺ populations, further referred to as double positive, CD271 single positive (sp), and CD146 sp populations, respectively. Only the double-positive population and the CD271 sp population highly enrich for clonogenic MSC [34,35]. Tormin et al. recently demonstrated that the double-positive cells are mainly located in the vascular niche and the marrow space, whereas CD271 sp cells reside in the endosteal space [35].

It is currently unknown whether these subsets have different biologic properties, neither it is known whether culture expansion of MSC from BM mononuclear cells selects for outgrowth of a certain population. For clinical applications, MSC are cultured from BM mononuclear cells without any selection. Therefore, this review focuses on MSC that are characterized according to the ISCT criteria [29].

Clinical Application of MSC

The immune-suppressive capacity and the regenerative potential of MSC have raised clinical interest in these cells. Although BM-derived MSC are currently the most frequently used source for a wide range of therapeutic applications (reviewed in Refs. [36,37]) also adipose tissue-derived MSC are regarded a safe and suitable cell type for therapy [38 –40]. Until September 2010, no significant adverse events have been reported in these trials [41]. Transplanted MSC have been applied in bone tissue engineering strategies to reduce clinical symptoms of osteogenesis imperfecta [42] and large bone defects [43], in regenerative treatments to further stimulate repair of pancreatic islets [44], and the infarcted myocardium [21,45,46]. Further, MSC have been used in several small trials for immunomodulatory treatments of autoimmune diseases, including Crohns disease [47,48] and diabetes mellitus [49]. However, feasibility for these applications has not been demonstrated yet.

The application of MSC in hematopoietic stem cell transplantation (SCT) is far ahead of other indications, reviewed in Refs. [36,37,50]. Currently, the 2 main potential applications of MSC in SCT are (1) prevention and/or treatment of graft-versus-host disease (GvHD) and (2) enhancement of engraftment.

(1) Severe acute GvHD after allogeneic SCT is associated with high morbidity and mortality [51], particularly in corticosteroid resistant patients. Because of the immune modulatory effect of MSC in vivo and in vitro, use of ex vivo expanded MSC can be applied for treatment of GvHD. Le Blanc et al. were the first to show rapid improvement of GvHD after treatment with BM-derived MSC [52]. After this initial report, further studies confirmed that infusion of ex vivo expanded adipose or BM-derived MSC can alleviate GvHD. Le Blanc et al. also demonstrated that so-called third-party mesenchymal stem cells (derived from unrelated HLA-mismatched donors) were as effective as HLA-identical or haplo-identical cells to treat GvHD [53]. This finding has practical implications and suggests that third-party cells can be prepared and stored frozen to be used for GvHD therapy. However, response rates reported so far differ substantially, ranging from 15% to 83% [39,53 –55].

(2) MSC orchestrate the BM microenvironment and produce multiple growth factors [56 –58] and a large quantity of various Wnt proteins [34,59], which are most likely involved in hematopoietic support. Since MSC provide support for primitive hematopoietic progenitor cells in vivo, it was postulated that they might enhance engraftment after SCT. Indeed, simultaneous intravenous injection of donor MSC and HSC was found to accelerate recovery of hematopoiesis after myeloablative therapy in animal models [60,61]. This effect is further enhanced by intrabone injections of MSC, suggesting that homing of MSC to the BM is relevant [62,63]. Koc et al. were the first to show rapid hematopoietic engraftment after co-infusion of autologous peripheral blood stem cells and autologous expanded BM-derived MSC in patients [64]. Thereafter, MSC-derived from the HSC donor have been investigated for their ability to enhance hematopoietic recovery in allogeneic SCT [65 –69], and co-transplantation of third-party MSC was studied as well [70]. The beneficial effects on hematopoietic recovery in these studies were highly variable and they only indicate that MSC could enhance hematopoietic engraftment. To establish this indication, larger randomized trials are required. This is especially of interest for patients vulnerable to engraftment failure, for example, after cord blood transplantation.

A future clinical application may include MSC as cellular delivery system for the local production of anticancer therapeutics in tumors (reviewed in Ref. [38]). As tumors resemble damaged tissue by releasing a range of cytokines, MSC were found to preferentially migrate to a variety of tumors. This capacity was adopted to deliver anti-tumor drugs in a preclinical setting. Studeny et al. showed that MSC transduced with interleukin (IL)-1β successfully migrate to tumors and the secreted IL-1β contributes to loss of tumor mass [71,72]. Local irradiation increases MSC migration to tumors in a dose-dependent manner [73,74], which is mediated by the increased release of transforming growth factor-β1, vascular endothelial growth factor, and platelet-derived growth factor (PDGF)-BB by the tumor cells and upregulation of CCR2 on MSC.

Large doses of MSC, ranging from 0.4×10⁶/kg to 10×10⁶/kg body weight [53,69], are currently required for successful clinical application. Although transplantation of these large doses of MSC in clinical trials seems to be beneficial to the patients, the engraftment potential of MSC in vivo is limited. For example, after co-transplantation of HSC and MSC, 100% donor chimerism was observed for hematopoietic cells in blood and BM, whereas the MSC remain of host origin [75]. This does not exclude the possibility that transient MSC engraftment may exert favorable effects through the secretion of cytokines or other paracrine factors, which engage and recruit recipient cells in productive tissue repair [37]. However, the observation that intrabone injections of the MSC [62,63] further enhance hematopoietic recovery suggests that most therapeutic applications can be improved if more MSC reach the target site. In murine models, MSC seem to preferentially home to damaged tissue [76], although the observed frequency of engraftment is highly variable depending on the conditioning regimen and the route of administration, but the overall consensus points to limited homing efficiency (0.00023%–0.00030%) of expanded MSC (reviewed in Ref. [77]). In vivo imaging of MSC after infusion will be helpful to determine the homing efficiency and to provide insights in the process of MSC migration a clinical setting. In animal models, real time tracking of MSC after infusion is often dependent on fluorescent dyes and proteins [74,78] or on reporter constructs that integrate into the genome [79,80]. These methods usually depend on optical tracking; however, due to low light penetration through the body, these strategies cannot be applied in a clinical setting [81]. For tracking in of MSC in patients, nonintegrating dyes that can be detected by real-time nonoptical techniques are required. Three techniques meet these criteria: magnetic resonance imaging (MRI), single photon emission computed tomography, and positron emission tomography. Out of these techniques, detection of superparamagnetic iron oxide (SPIO)-labeled cells by MRI is the most sensitive, but also the most expensive [81]. Tracking MSC using these 3 methods has only been studied in animal models [81], and only one clinical trial using MRI to trace SPIO-labeled dendritic cells has been approved so far [82]. The techniques mentioned above and other options to track MSC in vivo have recently been reviewed extensively by Reagan and Kaplan [81].

Chemokines, Cytokines, and Growth Factors Related with MSC Migration

It is assumed that common mechanisms of cell migration also apply to MSC migration (reviewed in Ref. [77]). Studies on leukocyte [83,84] and HSC [85,86] migration have provided insight into common mechanisms of migration. Chemokines, cytokines, and growth factors released upon injury provide migratory cues for cells. They induce upregulation of selectins and activation of integrins on the cell surface, enabling cells to interact with the endothelium. Cells subsequently adhere and transmigrate across the endothelial layer into tissues.

MSC express a wide variety of chemokine and growth factor receptors: among others, CXCR4 [27,87,88], PDGF receptors alpha and beta [89], and the hepatocyte growth factor (HGF) receptor cMet [90]. In vitro migration studies have demonstrated that several chemokines and growth factors are chemotactic stimuli for MSC, including stromal cell-derived factor-1α (SDF-1α) [28,91], PDGF [92,93], HGF [90], monocyte chemoattractant protein1 [94], and basic fibroblast growth factor [95]. These stimuli induce migration of MSC derived from various adult and fetal tissues [89,96 –98]. The optimal chemotactic stimulus varies between MSC derived from multiple origins, and this may be caused by tissue-dependent imprinted characteristics [96]. MSC were reported to migrate across endothelial cell monolayers [99,100] and through the underlying extracellular matrix [101,102], which are pivotal capacities since in most clinical trials MSC are administered intravenously. However, in multiple studies it has been shown that only a small fraction of MSC shows strong in vitro migratory characteristics [76,89,96]. To understand why only a small proportion of all culture-expanded MSC are able to migrate, the characteristics of migratory MSC have to be studied. These data will be important to explore strategies to improve directed migration of MSC.

In search of differences between migratory and nonmigratory MSC, we have previously investigated the functionality of the molecular machinery involved in cell migration. Upon stimulation with SDF-1α or PDGF-BB, the majority of all expanded MSC were able to rearrange the actin cytoskeleton, to polarize and to increase phosphorylation of the focal adhesion adapter protein paxillin, which is required for focal adhesion turn over and ultimately migration. Thus, these observations could not explain the small subset of migratory MSC [96]. The MSC that migrated toward SDF-1α maintained their migratory capacity in a secondary migration experiment, indicating that these cells are intrinsically different from nonmigratory cells. However, we could not distinguish migratory and nonmigratory fractions by cell surface marker such as integrins, adhesion molecules, or chemokine and growth factor receptors. Migratory MSC contained relatively less cells in S and G2/M phase of the cell cycle, but this cannot explain the observed difference in migration [96]. In search of genes that regulate MSC migration, 12 genes were found to be differentially expressed between migratory and nonmigratory expanded fetal BM-derived MSC in a micro-array screen. This list did not include cell surface markers. The early response transcription factors Nur77 and Nurr1 were upregulated in migratory MSC [103]. While in our studies we screened on migratory and nonmigratory fractions, Lee et al. looked into differences between low-passage and low-density cultures versus MSC from expanding, near-confluent cultures. Six surface markers were found preferentially expressed on early passage MSC in low confluency cultures: podocalyxin-like protein PODXL, CD49f, CD49d, c-Met, CXCR4, and CX3CR1 [104]. PODXL and CD49f were previously associated with MSC migration. Sorting for PODXL^hi/CD49f^hi cells resulted in selection of early MSC progenitors with increased homing to the heart in a murine myocardial infarction model [104].

Strategies to Enhance Migration and Homing of MSC

Various approaches to modify MSC or to enhance expression of surface markers of MSC have been explored to enhance MSC migration. All strategies listed in this paragraph are summarized in Table 1. Lentiviral overexpression of Nur77 or Nurr1, which were found upregulated in migratory MSC, resulted in enhanced migration of MSC toward SDF-1α compared to control-transduced MSC [103]. Another successful approach to modify MSC has been demonstrated by Sackstein and colleagues. They show that modification of CD44 with an E-selectin binding motif increased specific homing of MSC to the BM [105].

As the CXCR4-SDF-1α axis is dominant in trafficking of cells to BM and functional CXCR4 expression is required for MSC to migrate toward SDF-1α [88], many groups focused on modulating CXCR4 expression to enhance MSC migration.

Transfection or transduction of MSC with CXCR4 resulted in increased migration toward SDF-1α in vitro [91,106,107] and toward infarcted myocardium in rats, whereas hardly any migration was observed to normal myocardium [107].

Cytokine pretreatment has been shown to enhance homing of HSC [108] and this approach has been explored in MSC as well. Short-term stimulation of MSC with Flt-3 ligand, SCF, IL-6, HGF, and IL-3 increased surface expression of CXCR4. Correspondingly, in vitro migration and long-term engraftment in mice were increased [109]. Preincubation of MSC with TNF-α, but not interferon-γ, also resulted in upregulation of chemokine receptors [110,111] and increased chemokine-mediated migration [89,110,111]. TNF-α did not affect growth factor-induced chemotaxis [89]. Pro-inflammatory cytokines also increased the production of matrix metalloproteinases (MMPs) in MSC, thereby increasing the ability of stimulated MSC to migrate through the extracellular matrix. SDF-1α pretreatment did not influence MMP expression or cell invasion [101]. Expression of Nur77 and Nurr1 is increased upon cytokine stimulation in various cell types [112], including MSC [103]; thus, these genes may be part of the mechanism that increases expression of homing receptors in MSC upon cytokine treatment.

As oxygen levels in body tissues are lower than those under standard cell culture conditions [113], Hung et al. explored the hypothesis that culturing MSC under hypoxic conditions increases homing capacity [114]. Exposure of MSC to hypoxic conditions as short as 1 day increased expression of homing receptors CX3CR1 [114], CXCR4 [114], and CXCR7 [115] and activates AKT and c-MET-signaling [116]. When transplanted into chick embryos or in a murine ischemic hind limb model, homing of hypoxia exposed cells was increased compared to cells cultured at normoxia [114,116]. Thus, short-term exposure to hypoxia may provide an opportunity to enhance MSC homing.

Culturing MSC for Clinical Application

As homing capacity and expression of homing receptors decreases upon culture expansion [26,104], manufacturing of MSC for clinical application will be a trade-off between obtaining sufficient cell number for transplantation and maintaining migratory and other designated MSC properties.

It has been shown that harvesting MSC at 50%–70% confluency favors MSC migration; higher confluency potential is associated with an increase in migration [96]. Increased culture confluence was also found to inhibit transendothelial migration in MSC by increasing the production of a natural MMP inhibitor, TIMP-3 [102].

Several lines of evidence point out that plating density is also an important factor in expanding MSC while maintaining clonogenicity, differentiation potential, and a normal karyotype. Low seeding densities favored long-term proliferation of MSC. Plating at 200 cells/cm² generated an optimal number of rat MSC [117]. In addition, low plating densities result in better preservation of clonogenic rat and human MSC compared with higher densities [117,118]. Adipogenic differentiation of human MSC is favored during short-term expansion, whereas chondrocyte yield was increased after prolonged cell culture [118].

To assist MSC expansion, addition of epidermal growth factor (EGF) [119,120] or antibodies against type 1 interferon receptors (IFNAR1) [121] to culture media have been explored. Short-term expansion in the presence of EGF results in increased proliferation and maintains 25% more CFU-F compared to control-treated cells [119], suggesting that EGF acts as a mitogen and survival factor for clonogenic MSC. IFNAR1 reversibly controls the quiescence of MSC. At initial stages of expansion, IFNAR addition increases CFU-F outgrowth and over 2-month culture period it leads to faster and greater amplification of MSC without losing differentiation potential [121].

At present, serum-free medium is not available for the generation of clinical grade MSC. Alternatives to fetal calf/bovine serum are human AB serum and/or platelet lysates [122,123], and reviewed in Ref. [124]. To reduce risk of contamination with animal or human infections diseases, it is important to develop chemically defined media supplemented with recombinant growth factors that allow derivation and efficient expansion of MSC. Such media were established for embryonic stem several years ago [125]. To date, it has not been studied whether the above-described culture procedures influence the migratory capacities of MSC.

A suitable chemically defined medium should meet a few criteria. First, it has to allow efficient generation of MSC from adipose tissue or 10–20 mL of BM aspirate. Second, it has to enable optimal expansion of the obtained MSC. Finally, the capacities to differentiate, to modulate the immune response, and to migrate and specifically home have to be preserved. Although the in vitro migratory capacity of MSC is not a release criterion for the clinic at the moment [126,127], its inclusion may be considered.

Summary and Concluding Remarks

MSC are applied in an expanding number of clinical trials. Although administration of a large dose of MSC to patients with a wide variety of diseases seems to benefit clinical outcome, lack of understanding of MSC homing to injured tissues is restricting efficacy of the therapies. As several routes of delivering MSC explored in murine models may be unfavorable to apply in a clinical setting, such as delivering cells into the heart, bone, or brain, in future studies one should focus on improving migration of intravenously administrated MSC.

Optimizing the cell culture protocol is a major future challenge. Generating a chemically defined serum-free medium that allows efficient expansion of MSC is important to reduce the risk of infectious diseases. This medium should be designed not only for efficient expansion but also to preserve migratory capacities and classical MSC hallmarks such as differentiation and immune modulation. Addition of growth factors or antibodies to direct the appropriate signaling pathways may be helpful to achieve these goals.

Further characterization of native MSC subsets and their biologic properties will help to develop optimal culture protocols per therapy, as different applications utilize other MSC properties, like immune modulation or regenerative properties. Conventional MSC expansion protocols may select for specific MSC subsets; thereby, several MSC capacities may be lost.

Several studies point out that cytokine pretreatment and/or exposure to hypoxia increases MSC migration in vitro and in animal models. A clinical-grade cytokine cocktail should enhance expression of homing receptors and increase expression of genes that were found to stimulate MSC migration. Increased homing and engraftment of MSC may then result in lower doses of transplanted cells, thereby decreasing the culture expansion time, costs, availability, and the risk of transformations during culture expansion.

It is still under debate whether the few engrafted MSC are all the progeny of one or multiple MSC clones that may have a better migratory capacity than other cells. Recently, a cellular barcoding tool was introduced to enable clonal tracking in the hematopoietic system [128]. This retroviral system inserts a small unique sequence or barcode in the genome of each individual hit cell. Since all daughter cells will carry the same barcode, the progeny of all labeled cells can be traced [128]. This method provides a potential tool to study the fate of transplanted MSC, which can contribute to better clinical applications. First of all, it will enable to measure how heterogenous MSC cultures actually are, because the number of clones that contribute to the culture can be enumerated. It can also help to determine preferential outgrowth of specific MSC clones, which may possess enhanced migratory properties. Second, upon transplantation it can be used to illustrate how many MSC clones are capable of sustained engraftment in targeted tissue. If an approach to enhance MSC migration is successful, it may be expected that more unique barcodes, thus, more MSC clones, can be recovered from the targeted tissue. Finally, if MSC are injected or loaded onto a transplantable scaffold, the barcoding tool can be used to determine whether MSC indeed migrate into adjacent or more distant tissues.

In conclusion, with respect to all future MSC-based cellular therapies, it will be crucial to develop GMP-grade next-generation MSC that have enhanced in vivo homing capacity. This will lead to more straightforward and tailored MSC transplantation procedures for disease-specific applications.