Abstract
The regenerative and reparative potential of mesenchymal stem cells (MSCs) make them attractive candidates for numerous cell-directed therapies. The variant degree of tissue repair by transplanted MSCs has been assessed in several published reports. There are many gaps in the knowledge of MSC biology and the underlying reasons for their disparate effectiveness in tissue repair. This review examines successful preclinical models of MSC-directed repair, particularly of myocardial repair, in an attempt to shed light into the events dictating MSC therapeutic efficacy. The reparative advantage of genetically altered MSCs will be described. This overview will elucidate possible molecular mechanisms that can influence MSC engraftment, differentiation, self-renewal, and ultimately increase wound repair.
Mesenchymal Stem Cell (MSC) Characterization
MSCs are a heterogeneous population of fibroblastoid cells isolated by their ability to form adherent colonies (5,43,59). There are variable methodologies for the initial isolation of the MSCs (5,12). Physical dissociation of the starting tissue may or may not be followed by immunodepletion of unwanted cells (79). MSCs are expanded in vitro in serum-containing media to attain a characteristic spindle-like shape (12). Although bone marrow (BM)-derived MSCs will be the focus of this review; reviews concerning MSCs isolated from adipose tissue (39), umbilical cord blood (27,83), and fetal tissues (49) are available. Regardless of the source, MSCs retain the ability to differentiate into distinct mesenchymal lineages, particularly adipogenic, chondrogenic, and osteogenic (48,66).
Table 1 contains a list of cell surface markers that characterize the antigenic phenotype of murine BM-derived MSCs; their categories and the appropriate references are included. To separate the MSCs from any hematopoietic contaminants of the BM, murine MSCs must not express the protein tyrosine phosphatase cluster of differentiation 45 (CD45-) or markers of hematopoietic lineages such as B-cells or T-cells [lineage negative (Lin-)]. MSCs must also express the stem cell antigen 1 (Sca1+) as well as the hyaluronic acid cell adhesion molecule, HCAM (CD44+).
Antigenic Phenotype of Murine Mesenchymal Stem Cells (MSCs)
CD, cluster of differentiation; Mac-1, macrophage-1 antigen; GR-1, granulocyte differentiation antigen-1; Flk-1: fetal liver kinase 1 (or vascular endothelial growth factor receptor-2); SCA-1, stem cell antigen-1; Oct-4, octamer binding protein 4.
A large amount of data is available which characterizes human MSCs (17,48). The minimal criteria to define human MSCs have been established by the Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy. Human MSCs must express CD105, CD73, and CD90, and lack expression of CD45, CD34, CD14 or CD11b, CD79α or CD19, and human leukocyte antigen (HLA)-DR surface molecules (20).
The above description of the isolation, epitope identification, and expansion of MSCs points to the fact that there is variability in the definition of MSCs. Variations in the biological activities of these cells are present between isolates and passages (19,25,34). The consensus reached to define human MSCs is addressing this point; however, no such consensus exists for the murine MSC field. Moreover, the description of the nature of MSCs has relied on the in vitro characterization of these cells due to the lack of a definitive antigen in vivo.
Do MSCs Exist in Vivo without Tissue Culture Manipulation?
Because MSCs are isolated by their adherence to tissue culture plates, the field debates the existence of MSCs in situ. In an attempt to ensure MSCs were not purely an in vitro phenomenon, the presence of stem cells within the neural crest of the vertebrate embryo that give rise to mesenchymal tissues have been documented. Analysis of a subset of these stem cells has demonstrated that they have mesenchymal properties also in vitro. These studies identify the source of MSCs in the developing embryo and several reviews on these are available (21,36). Signaling cascades involved in the in vivo generation of these neural crest mesenchymal progenitor cells have been identified (9,33).
In the adult, a hypothesis that equates MSCs to pericytes or vascular-associated mesenchymal cells has been proposed (14,60,64). The work of Simmons et al. demonstrated that bone marrow-derived cells isolated by virtue of their STRO-1 positivity had MSC qualities not depending on their adherence (22,60). The STRO-1+ cells were localized to blood vessel walls of human bone marrow sections, thus the perivascular hypothesis (56). Pericytes were detected in several human tissues (all known sources of MSCs) and when isolated gave rise to MSCs in vitro (14). Interestingly, these pericytes expressed several known MSC markers in their native in vivo state (14). Although proof of endogenous MSCs or MSC-like pericytes has been documented, their role in tissue repair is not understood. Among other things, it is still unclear whether these tissue resident cells migrate to sites of injury or solely secrete factors that contribute to wound healing (16).
Main Clinical Applications
The National Library of Health provides information about ongoing clinical trials at http://clinicaltrials.gov (68). Currently, there are 73 clinical trials in which bone marrow-derived MSCs are to be administered as therapy for a variety of human conditions. None of these trials utilize genetically altered MSCs, a technology that might be of use in the future. Table 2 contains a list of these conditions, the number of trials engaged in the particular disease, and the method of administration of MSCs.
MSCs as Therapy, Current Clinical Trials, and Their Method of Administration (Clinicaltrials.gov)
Homing to Sites of Injury
The data included in Table 2 clearly demonstrate that the primary form of MSC transplantation, in 20 out of 30 conditions (67%), is intravenous. This form of administration takes advantage of the homing capacity of MSCs. Clinicians are relying on the ability of MSCs to travel through the vasculature and localize within injured tissue (18,53,54).
The effectiveness of homing ability varies between published preclinical reports. One explanation for this was delineated by Rombouts et al. by demonstrating that green fluorescent protein (GFP)-tagged murine MSCs lose their homing abilities in a syngeneic mouse model following prolonged in vitro expansion (51). Recently, several groups have shown that regulation of the CXC chemokine receptor 4 (CXCR4) and its ligand, stromalderived factor 1-α (SDF1-α), play an important role in the motility of human and rodent MSCs (55,74). Shi et al. demonstrated that incubation of MSCs in medium containing five cytokines [fms-related tyrosine kinase 3 (Flt-3) ligand, SCF, interleukin-6 (IL-6), hepatocyte growth factor (HGF), and IL-3] resulted in upregulation of both cell surface and intracellular CXCR4 (55). Tail vein injection of cytokine-cultured, and thus CXCR4-expressing, MSCs into sublethatlly irradiated nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice increased the homing to the bone marrow by sevenfold, albeit an increase in detection from 0.03% to 0.2% of the injected cells only 24 h posttransplant (55). This indicates that even in optimized conditions the efficacy of targeting is far less than 1%. The in vitro culture conditions in which the MSCs are expanded have an effect on the expression of CXCR4; 1-day exposure to hypoxia (1% oxygen tension) was sufficient to increase CXCR4 and CX3CR1 mRNA and protein levels enough to positively impact MSC migration and engraftment in a xenograft chick embryo model; this, however, was only an increase in engraftment from 0.3% to 0.9% (30).
There are other ligands and receptors involved in MSC migration besides the SDF1-α and CXCR4 axis. HGF and its receptor c-met increase the migration ability of MSCs in vitro (61) and have been implicated in the recruitment of MSCs to the site of injury (11). Human MSCs express the N-formyl peptide receptor (FPR), suggesting they are able to migrate to inflammatory sites where N-formylated peptides are present in the same way immune cells home to injury (72).
Migrating to the site of injury is only part of the challenge MSCs face for clinical application. MSCs must be able to traverse the 3-dimensional connective tissue. Type I collagen is the dominant extracellular matrix molecule found in mammalian tissues (62) and so MSCs must be capable of proteolytic degradation of this component to invade and extravasate into the tissue. The membrane-tethered 1 matrix metalloproteinase (MT1-MMP) has been recently shown to be involved in human MSC trafficking, in vitro and in vivo (38). MSCs were able to traverse type I collagen matrices 2D and 3D in vitro systems, and this activity was blocked completely by siRNA silencing of MT1-MMP. The chick chorioallantoic membrane (CAM) was used to confirm these results in vivo. Within 2 days of their placing on top of the membrane, hMSCs invaded the CAM surface and were detected in the underlying stroma; MT1-MMP silencing abrogated their invasive capabilities (38).
Understanding the genes/proteins involved in MSC homing, and their ability to remodel the collagenous extracellular matrix, will ultimately help increase their therapeutic efficiency. The CXCR4/SDF1-α signaling components seem to be important in the recruitment of MSCs. MT1-MMP seems involved in the degradation and penetration of type I collagen tissue barriers. Further work is needed to elucidate the regulation of the “homing” signature necessary for improved migration, invasion, and extravasation of MSCs into targeted tissues.
Immunomodulatory Properties
The majority of the current clinical trials (44%) are utilizing MSCs for their immunomodulatory properties. In conditions like graft versus host disease, Crohn's disease, primary Sjogren's syndrome, organ transplantation/rejection, systemic sclerosis, type 1 diabetes, systemic lupus erythematosus, multiple sclerosis, neuroblastoma, and nonmalignant red blood cell disorders, MSCs are being transplanted as treatment by themselves or as adjunct therapy. The biology behind the effects of MSCs on the immune system has been explored in several experimental models. A wide range of immune-suppressant factors have been shown to be secreted by MSCs; information on these and their function can be found in several recent reviews (57,65,73,80) and will not be addressed in this review.
Myocardial Infarction (MI) Therapy
The second most common application for MSCs in the clinic is treatment for ventricular dysfunction and myocardial repair (70). Although 10 current trials are under way utilizing MSCs in the setting of MI, very little clinical data have been gathered. A private company, Osiris Therapeutics, recently published clinical results demonstrating that MSC transplantation proved a safe therapy with increased benefits (increased left ventricular ejection fraction and reversed remodeling) compared to the placebo-treated group (26). Although these data are exciting, it is important to note that this study had a relatively small sample size (n = 53), a follow-up time of only 6 months, and is the first to show an overall statistical improvement in heart function in all MSC-treated patients (global symptom score p = 0.027). Because the available clinical data are scarce, many investigators are trying to dissect the necessary events, and the molecular mechanism that drive them, for MSCs to positively affect the heart following MI.
Possible Mechanisms of MSC-Mediated Repair
Figure 1 shows the possible results of the transplanted cells within the site of injury. Regardless of whether the mechanism of repair reflects their differentiation (direct contribution) or paracrine effects, the MSCs must be able to successfully reach the wound, survive, and expand within it. This section will focus on the molecular mechanisms thought to be involved in these events by looking at the effects of genetically altered MSCs in preclinical models of myocardial repair.
Possible effects of MSCs on injured tissue.
Cardiogenic Differentiation of MSCs Within Infarcted Myocardium
As early as 1999, researchers noticed that MSCs could be differentiated in vitro (by 5-azacytidine treatment) into beating, myocyte enhancer factor 2A (MEF-2A)- and MEF-2D-expressing cardiomyocytes (40). More recent work has demonstrated that these cells express functional adrenergic and muscarinic (β1, β2, M1, and M2) receptors after 5-azacytidine treatment in vitro (24). Instead of using this potentially toxic inhibitor of DNA methylation to derive cardiomyocytes from MSCs, Behfar et al. developed a “cardiogenic cocktail” (6). This mix of trophic factors include transforming growth factor-β1 (TGF-β1), bone morphogenetic protein 2 (BMP2), insulin-like growth factor (IGF1), fibroblast growth factor 4 (FGF4), IL-6, leukemia inhibitory factor (LIF), α-thrombin, vascular endothelial growth factor A (VEGF-A), tumor necrosis factor-α (TNF-α), and retinoic acid. The cocktail was able to drive the nuclear expression of cardiac NK2 transcription factor related, locus 5 [Nkx2.5] and MEF-2C, providing insight into the molecular mechanisms and signaling pathways that govern MSC cardiogenic differentiation (6). This in vitro phenomenon suggested MSCs could be a novel therapy used to regenerate damaged cardiac tissue and it prompted investigators to look for the in vivo differentiation of MSCs into cardiomyocytes following their administration.
GFP-tagged Lin- c-kit+ MSCs injected into the peri-infarct area following coronary artery ligation were fully incorporated into the myocardium and seemed to give rise to new, regenerated tissue after only 9 days (46). These results seem unique as several studies of this nature revealed that the level of in vivo cardiogenic differentiation of MSCs is very low (32,53) or completely lacking (13). Instead of MSC-derived regeneration of the heart, researchers have shown a fusion event of MSCs with injured cardiomyocytes (4).
Despite the low levels of direct contribution of the MSCs to the myocardium, MSCs have a positive effect in the function and remodeling of injured myocardium (31,52), albeit to varying degrees. It is widely accepted that the most probable effect of the MSCs in the site of injury is to supply the wound with soluble factors that enhance the repair process.
Secreted Factors Involved in Cardiac Reperfusion by MSCs
An ischemic environment is a direct consequence of coronary artery occlusion (69). The severity of the occlusion may lead to myocardial death and necrosis (78). Neovascularization nourishes the wounded tissue with the necessary nutrients and oxygen supply to prevent further tissue damage and increase repair (58). Several reports have demonstrated that MSCs induce VEGF expression, and therefore, neovascularization in ischemic myocardium (63,75,77,81). In vitro analysis of MSC-conditioned media (MSC-CM) demonstrated that MSCs themselves secrete a variety of angiogenic factors, among them VEGF and placental growth factor (PlGF); and that these are upregulated in response to hypoxic culture conditions (35). Together, these studies suggest that MSCs may play a role on the reperfusion of the infarcted myocardium.
Matsumo et al. reported that the adenoviral overexpression of VEGF by MSCs (VEGF-MSCs) increases capillary density (approximately twofold) in the VEGF-MSC-treated hearts 28 days post-MI. In this study, the MSCs were followed by β-galactosidase expression and the number of dual α-smooth muscle actin- and LacZ-positive cells was increased approximately fourfold in the treatment group. These authors suggest that the transdifferentiation of MSCs to endothelial progeny led to a significant increase in cardiac function and a significant decrease in infarct size in the VEGF-MSC-treated group (42).
Erythropoietin (Epo) is involved in neovascularization of damaged tissue (28,71) and thus increased vascular density was observed in the infarcted hearts treated with Epo-overexpressing MSCs (13). Without data to indicate endothelial transdifferentiation of MSCs, the authors conclude the MSCs were accessory to the newly formed blood vessels. The secretion of trophic factors, especially Epo, was what ultimately improved cardiac function post-MI.
Increased Myocardial Survival in Response to the MSC Secretome
The MSC-secreted factors seem to have more than just proangiogenic effects; MSC treatment reduces the apoptotic index within the peri-infarct area (37,68). In vitro, cultured cardiomyocytes have improved survival in hypoxic conditions when treated with MSC-CM (31) or in coculture experiments (15,68,76). An important prosurvival factor, Akt, has been shown to be expressed by MSCs in response to hypoxia (68). Overexpression of Akt in MSCs (Akt-MSCs) increased their survival compared to control but, importantly, injection of their MSC-CM to the myocardium following MI had a significantly positive effect on the apoptotic index of the cardiomyocytes (44).
Microarray analysis performed on the CM obtained from Akt-MSCs documented the presence of secreted frizzled-related protein 2 (sFRP2) (44). Rat cardiomyocytes in vitro, as well as infarcted mice cardiomyocytes in vivo, had decreased apoptosis when treated with CM from the Akt-MSCs. The prosurvival effect on the cardiomyocytes was due to sFRP2 since CM from “Akt-MSCs minus sFRP2” (knocked down sFRP2 expression by siRNA in Akt-MSCs) did not have the same prosurvival effect in either setting. The decrease in cardiomyocyte apoptosis post-MI due to Akt-MSC-CM led to decreased infarct size (44). It is important to note that the authors looked at infarct size 72 h posttreatment and that they did not assess any functional cardiac parameters.
When heme-oxygenase 1 (HO1) was overexpresseed by MSCs (HO1-MSCs), similar effects as the ones described above were observed (67). This group saw significant decrease in the in vitro apoptotic index of the HO1-MSCs compared to the control MSCs. In vivo, HO1-MSC-treated infarcted hearts had increased capillary density but this was attributed to the increase of VEGF by the HO1-MSCs. Transgene expression led to a significant decrease in the apoptotic index of cardiomyocytes in the treatment hearts. HO1-MSC treatment following MI led to a smaller infarct size, decreased ventricular remodeling, and increased cardiac function 28 days post-MI. There was no mention of engraftment or differentiation by the MSCs in this report. The authors conclude that the most important role of the transgene was its antiapoptotic and antioxidant effects on the cardiomyocytes (67).
Enhancing MSC Survival in the Wound
Retroviral expression of the prosurvival gene Akt-1 decreased the apoptotic index of MSCs in vitro (41). The authors transplanted the Akt-MSCs, or GFP-MSCs as a control, directly onto the rat myocardium following MI. Although the transgene expression decreased the apoptosis of the MSCs in vivo, this analysis was performed 24 h after the injection. No long-term effect on MSC survival was established. The degree of the incorporation of the MSCs into the tissue was not discussed despite observation of MSC-derived cardiomyocytes. Improved cardiac function due to Akt-MSC treatment was determined at 2 weeks post-MI, albeit using an ex vivo Langendorff model (41).
Jiang et al. tested the in vivo efficacy of MSCs transduced with a retroviral vector to express both Akt and angiopoietin-1 (Ang-1) (32). The animals in this study underwent permanent coronary occlusion and were treated with an intramyocardial injection of MSCs. Although the authors claim transgene expression afforded the MSC with increased engraftment, these data were not presented in their publication. Expression of Akt and Ang-1 by MSCs increased blood vessel density within the peri-infarct area and led to a statistical increase in heart function (fractional shortening and ejection fraction) compared to the control media treatment group.
Retroviral overexpression of Epo by MSCs increased the survival both in vitro and in vivo in a murine subcutaneous implantation model of Matrigel-embedded MSCs (13). Although assessment of the Epo-MSC- and control MSC-treated hearts 7 and 14 days post-MI did not demonstrate an increase in MSC engraftment or in the MSC-derived cardiac progeny, there was improved cardiac function related with Epo-MSC treatment.
Improving MSC Self-Renewal
Whether MSCs work by direct regeneration or through paracrine modulation, MSCs are unable to have a significant effect on the repair process unless a critical mass is attained within the wound (i.e., enhanced engraftment). The levels of engraftment of MSCs in preclinical models of myocardial repair are low (30,53,55). Few clinical data on the engraftment of MSCs are available that address long-term engraftment. In one example, fewer than 1% donor cells were detected 4–6 weeks postinfusion on osteogenesis imperfecta patients (29). Recently, our group demonstrated that retroviral overexpression of sFRP2 by MSCs (sFRP2-MSCs) increased the in vitro proliferation (2) and survival of MSCs by modulation of both the BMP and Wnt signaling cascades (3). This phenomenon provided the MSCs with a survival advantage also observed in vivo. Although sFRP2-MSCs showed long-term engraftment within infarcted myocardium approximately 2.5-fold greater than a GFP-MSC control 30 days post-MI (2), the observed degree of engraftment was still modest. The amount of engraftment observed was much greater in another model of tissue repair (2), suggesting engraftment also depends on the wound context. Achieving a critical mass of MSCs by improving their survival and self-renewal capacity should increase their numbers in the damaged tissue.
Conclusion
The use of genetically altered MSCs in the context of clinical myocardial repair may be possible in the near future. The safety and effectiveness of genetically altered stem cell therapy have been closely monitored, particularly in the hematopoietic stem cell (HSC) field. The clinical trials in this field aim to improve the effect of CD34+ cell treatment for a variety of diseases (i.e., gliobastoma multiforme, severe combined immunodeficiencies, and breast cancer) with overexpression of certain proteins {O-6-methylguanine-DNA methyltransferase (MGMT[P140K]), adenosine deaminase, and multidrug resistance (MDR1)} (70). These proteins may increase the survival/engraftment of the CD34+ cells as is the case for MGMT[P140K] (50) and MDR1 (45), or provide a missing enzyme to the patient (adenosine deaminase) (8). Information gained from these trials includes the transduction efficiency utilizing different types of vectors including retroviral (1) and lentiviral (23), to try to establish a cutoff for clinical effectiveness. The safety of these vectors has been assessed whereby integration sites showed modest dysregulation of surrounding genes (10). However, potential risks of transduced stem cell transplant still remain.
Although there is a long history of trials utilizing genetically altered CD34+ cells, this technology remains relatively unexploited and its application in the MSC field is tentative at best. An alternative approach to the transduction of MSCs prior to administration is suggested by a recent report. An overall improvement in the repair of chronic myocardial infarction with what was termed “guided cardiopoiesis” of the MSCs was demonstrated (7). In this report the authors directed hMSCs into cardiopoiesis in vitro with the use of a cardiogenic cocktail prior to in vivo administration in a murine model of myocardial infarction. Not only were there were no deleterious effects, the “guided” MSCs improved heart function more compared to control (7). In vitro priming might be an avenue to consider in the clinic.
This review described the mechanisms by which MSCs are thought to increase repair by looking closely at their effects on myocardial repair. The involvement of several genes on the ability of the MSCs to home to the site of injury, engraft within the myocardium, give rise to new myocytes/fuse with existing cells and/or increase the vascular content within the wound were discussed. These genes and their documented effects are depicted in Figure 2; the overall division is that of autocrine effects (affecting the MSCs themselves) or paracrine effects (affecting the wounded/infracted tissue). Although much more work needs to be done to fully understand MSC biology and their reparative capacities, the goal remains to find a gene or set of genes that could have both autocrine and paracrine effects so as to increase MSC-directed myocardial repair.
Lessons from genetically altered MSCs.
Footnotes
Acknowledgments
This work was supported, in whole or in part, by National Institutes of Health Grant R01-HL088424, Veterans Affairs merit award (P.P.Y.), and American Heart Association Grant 09PRE2010035 (M.P.A.).