Mesenchymal stem cells for vascular regeneration

■ Role of shear stress on MSCs

Under physiological conditions, the endo thelium in arteries experiences shear stress within the range of 10 to 20 dynes/cm³ [42,43]. The shear stress that results from blood flow appears to modulate EC function by activating downstream signaling pathways. To study the effect of shear stress on cell behavior, both in vitro and in vivo systems can been used. The advantage of the in vitro system is the ability to control the magnitude and direction of shear stress, whereas in vivo models are a step closer to translating in vitro findings to clinically relevant diseases [44]. Fluid shear stress is commonly modeled in vitro by laminar flow using a parallel-plate chamber as a 2D culture model, or in the case of interstitial flow, by a porous 3D matrix.

To date, studies of shear stress on MSC migration, proliferation and vascular differentiation remains limited in number. Table 3 summarizes the results from published findings related to shear stress effects on MSC phenotype and behavior. In one study using STRO-1-expressing rat bone marrow stromal cells, 36-h shear stress (∼14 dyn/cm²) stimulation by pulsatile flow increased the number of smooth muscle myosin heavy chain-expressing cells, although SMA protein levels was not significantly different when compared with the static control group [45]. In another study, shear stress applied to murine C3H/10T1/2 embryonic mesenchymal progenitor cells led to induced expression of endothelial markers such as CD31, vWF and VE-cadherin at the mRNA and protein level after 12 h of steady shear stress levels of 15 dyn/cm² [46]. Shear stress could also stimulate endothelial function, including the uptake of acetylated low-density lipoproteins and formation of capillary-like structures on matrigel.

■ Role of mechanical stretch on MSCs

Physiologically relevant levels of strain in blood vessels range from 5 to 30% strain and 30 to 90 cycles per min [37]. In 2D culture, strain is often studied by uniaxial or equiaxial models in which cells are cultured on elastic substrates such as poly(dimethylsiloxane) (PDMS), which can be repetitively stretched to controllable strain magnitudes and frequencies. For 3D stretch models mimicking vascular environments, equiaxial stretch conditions can be simulated by embedding cells in a tubular matrix and then stretching the construct from the luminal side.

The effect of strain is well-characterized in vascular cell types, but not for MSCs [47]. As summarized in Table 3, we have previously shown that for human bone marrow-derived CD105- and CD166-expressing MSCs, 5% cyclic equiaxial strain at 60 cycles per min for 24 h downregulates SMA and SM22α, whereas cyclic uniaxial strain transiently increases the expression of these markers [48]. However, in another study in which rat bone marrow-derived adherent cells were exposed to 10% cyclic uniaxial strain at 60 cycles per min for 7 days, the expression of SMA and h1-calponin was induced [49]. The seemingly opposing results may be attributed to differences in stimulation time, species-specific response or stem cell phenotype. In the former report, immunophenotypically characterized human cells were studied for a period of 24 h at 5% equiaxial strain, whereas the latter report utilized rat adherent cells without immunophenotypic characterization and examined the effect of 10% cyclic uniaxial strain for 7 days. Consequently, it is unclear which specific factors may be attributed to these conflicting results.

In addition to global gene-expression changes, uniaxial strain induces cell realignment in the perpendicular direction, relative to the direction of stretch. To assess whether uniaxial strain could promote MSC differentiation if cell orientation was restricted, we cultured human bone marrow-derived CD105- and CD166-expressing MSCs on microgrooves in the directions parallel or perpendicular to 5% uniaxial strain. Notably, global changes in MSCs were observed only when MSCs were kept aligned in the direction of uniaxial strain [50]. DNA microarray analysis showed an upregulation of h1-calponin expression, alterations in cell signaling and increases in cell proliferation. Together, these studies suggest that physiological levels of stretch can modulate MSC gene expression, cell alignment, proliferation and phenotype towards those that are consistent with smooth muscle characteristics.

■ Role of multimodal mechanical stimulation on MSCs

To mimic the dynamic physiological environment, multimodalities of mechanical stimulation have been studied in well-controlled in vitro environments. O'Cearbhaill et al. cultured bone marrow-derived CD73- and CD105-expressing human MSCs in the presence of pulsatile pressure (40−120 mmHg), radial distention of 5% and a shear stress of 10 dyn/cm² at a frequency of 60 cycles per min for up to 24 h [51]. Histological analysis revealed that the majority of MSCs responded to the mechanical stimuli by aligning within 20° to the direction of flow and by adopting a compact cell size, as is characteristic of ECs. However, gene expression and immunofluorescence analysis of both the stimulated and static groups reported that the cells did not express EC marker vWF factor on either the mRNA or protein level. Instead, the mechanically stimulated cells exhibited greater levels of smooth muscle-associated markers SMA and calponin in comparison to static samples. These results implicate the potential of MSCs as a cell source for the generation of SMCs, although the morphology and orientation appeared to resemble that of ECs.

In addition to stimulating differentiation, multimodalities of mechanical strain and shear could also affect matrix deposition by MSCs. When ovine bone marrow-derived CD45- and CD14-negative MSCs were cultured on poly(glycolic acid) (PGA) and poly(l-lactic acid) (PLLA) scaffolds in the presence of both cyclic stretch (60 cycles per min) and shear stress (∼1 dyne/cm²) for 3 weeks, a 75% higher collagen content was observed compared with cells with stretch, shear or static control groups [52]. These results suggested that that combined stretch and shear could stimulate MSCs to remodel the matrix microenvironment, which has applications for accelerating tissue formation. These results together demonstrate the effect of shear stress and strain in modulating stem cell phenotype and behavior.

■ Role of matrix rigidity

Physiological matrix rigidity, a bulk property of the material that is characterized by the Young's modulus, varies within three orders of magnitude from soft tissues of the brain to hard tissues of the bone [65]. In particular, matrix rigidity plays an important role in cardiovascular applications, as cardiovascular cells respond to changes in matrix stiffness due to injury or disease such as MI or atherosclerosis. To utilize MSCs as a potential cell source for cardiovascular treatments, it is therefore important to understand the effects of matrix rigidity on MSC function and behavior as well.

The ability to sense matrix rigidity relies on mechanosensing structures in the cell such as integrins and focal adhesions. Integrins are transmembrane glycoproteins that serve as mediators of communication between the intracellular and extracellular environment from the inside-out or outside-in [66]. Structurally, they consist of an α- and β-subunit that assembles together in different combinations to form heterodimers. In human MSCs, α1 through α6 and β1 through β4 are reportedly expressed [67,68]. Binding of the extracellular domain of integrin mediates cell–cell and cell–extracellular matrix (ECM) adhesion, whereas binding of the cytoplasmic domain to intracellular ligands can activate signaling pathways or modulate cytoskeletal assembly. Related to integrins are focal adhesions, a complex of cytoskeletal proteins and enzymes that can bind the cytoplasmic domain of integrins or to actin [69]. Together, integrins and focal adhesions mediate the transduction of external cues that regulate cell morphology, migration, proliferation and differentiation.

To study the role of matrix rigidity, Pelham and Wang developed a method to vary the stiffness by changing the concentration of acrylamide and bis-acrylamide crosslinker used to form the matrix, while maintaining the same chemical composition [70]. This model enabled well-controlled studies of the effects of rigidity independent of chemical composition. Their studies, which assessed the role of substrate rigidity on fibroblast migration, led to striking observations that rigidity alone could modulate cell morphology and spreading area. In particular, they found that cells on softer substrates had lower spreading areas and faster migration rates, in comparison to cells on more rigid substrates [71]. Besides controlling cell spreading, matrix rigidity can also modulate the directionality of migration, as cells tend to migrate from soft (14 kPa) to hard (30 kPa) substrates [72].

Besides modifying the behavior of differentiated cells, rigidity also has profound effects on the fate of MSCs. Engler et al. demonstrated that human bone marrow-derived MSCs differentiated on polyacrylamide substrates towards osteogenic, muscular and neuronal lineages according to the physiological rigidity of respective tissues [73]. In particular, soft matrices that mimic brain tissue stimulated neurogenesis, whereas rigid matrices that resemble the stiffness of bone induced osteogenesis. These findings were supported by transcriptional profiling data showing upregulation of neurogenic markers on soft gels (0.1−1 kPa), myogenic markers on intermediate gels (11 kPa) and osteogenic markers on stiff gels (34 kPa). Lineage specification by matrix rigidity appeared to be dependent on non-muscle myosin II, as demonstrated by the loss of lineage specificity when cells were exposed to blebbistatin, an inhibitor of non-muscle myosin II. However, further research is necessary to understand the role of matrix rigidity in modulating MSC morphology, migration and vascular lineage specification.

■ Role of matrix patterning

In addition to matrix rigidity, matrix patterning can also direct cell organization and differentiation [63]. Therefore, to mimic physiologically ordered vascular tissues, micropatterning and nanopatterning tools can enable well-controlled studies of restricted cell alignment. Soft lithography is a convenient method to generate micropatterns for biological applications. The desired pattern can be generated by photolithography onto a silicon wafer, which can be transferred to an elastomeric material such as PDMS [74]. Cells grown directly on the PDMS stamp would be subjected to the topographical features. Alternatively, the patterns on the PDMS stamp can be transferred onto other substrates by microfluidic patterning or microcontact printing.

Using microfludic patterning of collagen I in parallel microchannels, we have previously shown that vascular SMCs can align according to the direction of the channels [75]. Furthermore, the proliferation rate of SMCs decreased on substrates with channels that were 20 and 30 μm wide, compared with nonpatterned substrates. In addition to SMCs, skeletal myoblasts could also align according to topographically patterned PDMS and form globally aligned myotubes in the direction of the microchannels [76]. Myotube length was significantly longer on micropatterned substrates than that on nonpatterned substrates, suggesting a potential role of micropatterning as a tool for generating physiologically aligned cellular constructs for tissue-engineering applications.

While the effects of micropatterning on vascular and muscle cells have been studied, the effect of micropatterning on MSCs remains to be elucidated. McBeath et al. employed micropatterned islands of fibronectin to probe the role of cell shape in mediating human bone marrow-derived MSCs towards adipogenic and osteogenic differentiation [77]. Notably, adipogenesis was associated with small (1024 μm²) islands in which the cells remained rounded in morphology, whereas large (10,000 μm²) islands preferentially induced osteogenic differentiation and an adherent morphology. It was further demonstrated that actin assembly and Rho GTPase mediate the differentiation of MSCs towards adipogenic and osteogenic lineages.

In addition to microscale features, nanoscale features can also restrict cellular alignment, behavior and differentiation. Nanofibrous matrices can be prepared by electrospinning, a process that generates nanoscale fibers formed by high-voltage electrostatic field [78]. In the electrospinning process, a thin stream of electrically charged polymer solution travels to a collector plate as it evaporates, forming fibrous polymer strands [78]. When a rotating drum is used as the collector plate, parallel-aligned fibrous networks can be generated [79]. Alternatively, randomly oriented fibers can be stretched to generate aligned fibers.

We have utilized electrospinning technology to generate parallel-aligned PLLA substrates to create aligned skeletal myoblasts, neurite outgrowth and MSCs [76,80,81]. In particular, we have shown that CD105- and CD166-expressing human bone marrow-derived MSCs cultured on aligned PLLA nanofibers with diameters similar to that of collagen fibrils (500−1000nm) remain viable and align in the direction of the nanofibers in vascular tissue-engineering applications [81]. Moreover, the cellular organization of MSCs was similar to that of SMCs on native arteries, suggesting that the nanofibrous substrates could mimic the structure native collagen matrix upon which SMCs are grown, which can be used to construct tissue-engineered vascular grafts.

■ Vascular grafts

In 2005, almost 500,000 coronary artery bypass procedures were performed [1]. Despite the improvement in overall quality of life by bypass procedures, there is still a growing need to engineer off-the-shelf vascular graft replacements to eliminate the need for donor vessels, which have a 35% 10-year failure rate. Synthetic acellular vascular grafts, which can be made of various biomaterials such as poly(ethylene terephthalate) and polyurethane [82,83], have been restricted to sizes larger than 5 mm in diameter owing to the heightened risk of thrombosis and occlusion associated with smaller graft diameter sizes. As a result, engineering living blood vessels is a promising alternative.

Blood vessels consist of three cell layers, namely an endothelial lining in the innermost layer, vascular SMCs in the medial layer and a fibroblast adventitial outermost layer. Models of living vascular grafts have been developed to contain cells alone, matrix alone, or the combination of cells, soluble factors and matrix [81,84–88], and the grafts may consist of multiple layers or single layer of cells and/or matrix (Figure 2). For example, Weinberg et al. showed that a piece of blood vessel can be constructed with ECs as inner lining and SMCs in collagen gel, supported by a Dacron mesh [84]. L'Heureux et al. demonstrated the feasibility of generating trilaminar vascular grafts without the use of a scaffold. The authors grew sheets of SMCs and wrapped them around a tubular support to produce the medial layer. Then, in a similar manner, they wrapped a sheet of fibroblasts around the SMC layer to generate the adventitial layer, before ECs were finally seeded in the lumen [86]. The same approach has recently been used to construct a bilayered vascular grafts containing ECs and fibroblasts [89]. Niklason et al. demonstrated the feasibility of engineering vessels by culturing ECs and SMCs in a tubular PGA porous scaffold in a bioreactor [87]. The engineered vessels were mechanically robust and could respond to pharmacological agents like native vessels. When implanted in vivo, the engineered arteries remained patent for up to 24 days.

Whilst the use of ECs and SMCs for generating vascular conduits would require patient-specific cell isolation, MSCs are a favorable alternative cell source because of their ability to be delivered allogenically as an off-the-shelf therapy. Studies to date provide encouraging results that suggest that MSC-seeded grafts may be a feasible therapeutic option as a vascular conduit (Table 4). In one study, green fluorescence protein (GFP)-labeled rat bone marrow-derived MSCs that express CD90 and CD73 were cultured on polyurethane vascular prosthesis and assessed for their ability to differentiate into vascular lineages in vivo [90]. The cells were found to be organized in a multicellular structure that mimics that of native arteries 2 weeks after implantation in mice. The cells co-expressed GFP and SMC markers, suggesting that the in vivo setting induced smooth muscle protein maturation. MSC-seeded grafts were also more highly endothelialized than acellular grafts to prevent thrombosis. However, the electrophysiological profile of MSC-differentiated SMCs appeared to be different from that of mature SMCs. These results suggested that only partial differentiation of MSCs into SMCs were achieved by in vivo implantation.

We have recently shown that CD105- and CD166-expressing human bone marrow-derived MSCs that were cultured on aligned electrospun nanofibrous PLLA conduits can promote efficient in vivo infiltration of vascular cells and ECM remodeling [81]. In vivo transplantation of the MSC-seeded or acellular conduits as interpositional grafts in rat common carotid arteries showed that the cell-seeded grafts enabled long-term patency with well-formed layers of endothelial and smooth muscle layers. Both the cell-seeded and acellular grafts were shown to have extensive collagen deposition, but only cell-seeded grafts developed an elastic lamina adjacent to the lumen, implicating that the ECM and cellular components of MSC-seeded grafts closely resembled those in native vessels. Interestingly, only few MSCs could be found in the grafts after explantation at 7 days after transplantation, suggesting that the long-term retention or differentiation of MSCs may not be necessary to confer the therapeutic benefits of anti-thrombogenesis and vascular remodeling.

The patency of MSC-seeded grafts may be related to the anti-thrombogenic properties conferred by MSCs [81]. Our in vitro studies demonstrated that platelet adhesion and aggregation was notably higher on SMCs than on MSCs or ECs, indicating that MSCs are comparable to ECs in their ability to confer antiplatelet adhesion properties. The mechanism of resisting platelet aggregation appears to be related to MSC cell surface heparin sulfate proteoglycans. The disruption of heparin sulfate proteoglycans by heparinase treatment led to increased platelet adhesion. This study demonstrated an important role of heparin sulfate proteoglycans on MSCs in resisting platelet adhesion.

Recent studies have utilized genetic modification of MSCs to better mimic the function of vascular cells. Kanki-Horimoto et al. transduced rat bone marrow-derived MSCs with an adenovirus harboring endothelial nitric oxide synthase (NOS), which is normally released by the endothelium to prevent thrombogenesis, to better mimic ECs [91]. While culturing the modified or naive MSCs onto small-caliber expanded polytetrafluoroethylene vascular prostheses, they monitored NOS activity based on the conversion of radiolabeled arginine to citrulline and found that the citrulline levels were increased in modified MSCs, in comparison to the naive cells. Furthermore, NOS activity could be blocked by the NOS inhibitor, N(G)-nitro-l-arginine methyl ester. This study highlights a promising technology for developing grafts with improved patency rates.

Tissue-engineered vascular grafts have also been shown to be mechanically durable. One study assessed the mechanical properties and therapeutic potential of a tubular scaffold consisting of a compact polyurethane layer sandwiched by porous PLGA on either side [92]. The burst strength of vascular grafts ranged from 160 to 183 kPa, which was similar to the burst pressure in veins (∼223 kPa) [86]. In addition, the compliance was 0.57−53.80%/100mmHg, which was higher than that of conventional Dacron grafts. This tubular scaffold was next seeded with canine bone marrow stromal cells purified by Ficoll gradient centrifugation and implanted for up to 24 weeks in the canine abdominal aorta. Histological analysis of the vessels after explantation demonstrated full patency and differentiation towards EC and SMC lineages. These results suggest that this trilaminar scaffold favors mechanical strength and patency. Together, these studies demonstrate the therapeutic potential of MSCs and polymer scaffolds for developing vascular conduits that confer mechanical strength, patency and vascular differentiation potential.

Besides MSCs, endothelial progenitor cells (EPCs) can be used to cover the luminal surface of vascular grafts and prevent thrombus formation [93–98]. However, this approach may only be used autologously owing to the potential immune responses and the limited expansion capability of human ECs and EPCs. Alternatively, if appropriate adhesion or receptor ligands are available on the vascular graft, EPCs in the circulating blood can attach and cover the luminal surface [99,100]. Modification of vascular grafts to recruit EPCs or MSCs from circulating blood will be an attractive approach to facilitate the endothelialization of the vascular graft.

■ MSCs for treatment of MI

Besides generating vascular conduits, MSCs have been utilized for treatment of MI, which results when blood supply to the heart is blocked, leading to cardiomyocyte cell death. The pathological remodeling process that results includes scar tissue formation and thinning of the free wall, which contribute to the impairment of cardiac performance [101–103]. To repair MI tissue, angiogenesis is initiated to restore blood supply to the tissue. The process of angiogenesis begins when supporting cells known as pericytes are first removed from the branching vessel. Next, the endothelial basal lamina and ECM degrade, and are replaced by a new matrix that attracts EC migration and proliferation. The ECs then form a tube-like structure that recruits pericytes and SMCs to the abluminal surface, forming a new vascular branch [104].

Angiogenesis is a highly organized process that is mediated by soluble chemical cues. For example, VEGF induces vascular hyperperme-ability, endothelial migration and proliferation. FGF mediates endothelial migration, proliferation and tube formation, and TGF-β plays an important role in matrix remodeling and vessel restabilization [104]. Since these soluble factors play an important role in vascular remodeling and angiogenesis, they have become widely studied as potential agents to modify MSC behavior and function for vascular repair.

Various cell types have been used to repair MI tissue, including fetal cardiomyocytes [105], skeletal myoblasts [106], embryonic stem cells [107] and MSCs [108]. The modes of delivery into ischemic tissue include systemic or local routes including local injection, systemic injection or the placement of a local patch. Systemic delivery into the circulation can be advantageous because of minimal invasiveness, but the efficiency of directed homing may be limited. On the other hand, local delivery to the ischemic tissue is preferable to ensure cell transplantation to desired tissue locations, although accessing the myocardium may be surgically invasive. Cells can also be delivered concomitantly with polymers, which provide structural support and mediate cell survival and proliferation. Furthermore, the polymers can be adapted for delivering other factors, including cytokines, plasmids, viruses and drugs [109].

For the repair of MI, numerous studies report a therapeutic effect of delivering MSCs to improve cardiac repair by stimulating neovascularization and vascular differentiation (Table 4 & Figure 3). For example, the delivery of density-gradient purified bone marrow-derived canine MSCs for treatment of chronic myocardial ischemia led to significant improvement in ejection fraction and vascular density in the cell-treated group, and MSCs were also found to express EC and SMC markers [110]. Schuleri et al. delivered density-gradient purified bone marrow-derived porcine MSCs in allogenic models of MI and monitored myocardial blood flow by first-pass perfusion MRI over 8 weeks [111]. Their results demonstrated a significant improvement in myocardial blood flow over the first week in MSC-treated animals, compared with placebo-treated animals and untreated animals, but this improvement was not sustained at 4 and 8 weeks. On the other hand, global cardiac function was improved at week 8. This finding suggests that early tissue perfusion appears to precede improvement in cardiac function. To address the therapeutic potential of MSCs derived from patients with ischemic heart disease, CD105- and CD106-expressing bone marrow derived-MSCs from the diseased patients were delivered to immunocompromised mouse models of MI [112]. MRI assessment and immunohistochemical analysis showed that, after 14 days, the ejection fraction was significantly higher in the group receiving cell treatment. In addition, the MSC-treated group had increased vascularity and reduced thinning of the infarct scar, along with the generation of MSC-derived cells that expressed endothelial marker vWF. The therapeutic benefits of improved cardiac function and cardiomyocyte survival after MSC treatment has been confirmed by others as well [113–115].

Some reports raise concerns over the incidence of MSC differentiation into cardiovascular cell types as artifacts of cell fusion [116,117]. One study utilizing Cre/lox recombination to detect cell fusion events of murine bone marrow-derived MSCs showed that the cells fused in vitro with neural progenitors [116]. In the same study, in vivo bone marrow transplantation demonstrated that the cells fused with hepatocytes, Purkinje neurons and cardiac muscle, but no evidence of differentiation could be detected. However, other evidence points to differentiation as a true phenomenon. For example, Kajstura et al. infused c-kit-expressing bone marrow-derived cells from male transgenic mice that expressed enhanced GFP (eGFP) into female recipient hearts [118]. Using both immunofluorescence detection of eGFP and fluorescence in situ hybridization for X and Y chromosomes, the authors detected newly formed cardiac and vascular cells that had one set of the X or Y chromosome as well as expression of eGFP. This study suggests that the cells differentiated rather than fused with host cells. Based on these studies, the relative contribution of cell fusion and differentiation is unclear.

Preservation of endogenous ECM and intercellular junctions also appears to be involved in promoting therapeutic benefits after MSC delivery. Miyahara et al. derived rat MSCs from adherent adipose tissue for transplantation as a patch onto the epicardial surface of infarcted hearts [119]. Their results showed that the sheets engrafted and grew in thickness, containing MSC-derived vascular cells, undifferentiated cells and few cardiomyocytes. The cell sheets also stimulated the formation of vascular structures originating from the host tissue, suggesting that MSCs could stimulate angiogenesis by acting through a paracrine mechanism. Another study comparing the therapeutic effect of rat MSCs derived from bone marrow by Percoll gradient centrifugation in either dissociated or sheet fragment form for repair of the infarcted heart reports significant enhancement of vascular density and graft/host cell connection for the cell sheet-treated animals [120]. The MSC sheet fragments were found to be aligned along the interstices of the injection sites, whereas very few cells were retained in the dissociated cell group.

The mechanism by which MSCs improve neovascularization after MI is unclear. Increasing reports provide evidence that MSCs act through a paracrine mechanism, in addition to differentiating into vascular lineages that directly contribute to vascular function. MSCs have been shown to secrete angiogenic factors such as VEGF and bFGF that may stimulate angiogenesis in ischemic tissues [121]. Furthermore, Tang et al. showed evidence that MSC-treated infarcts had higher levels of stromal cell-derived factor-1α, a stem cell homing factor, as well as reduced levels of proapoptotic protein Bax, when compared with rats receiving media treatment [122]. Therefore, it is likely that MSCs are involved in inducing angiogenesis and promoting survival cues in ischemic tissues.

In addition to cells alone, MSCs delivered in conjunction with polymers have resulted in therapeutic improvement. In one recent study, a cardiac patch was formed by embedding human bone marrow-derived CD44-negative MSCs in collagen I matrix [123]. Transplantation of the construct as an epicardial patch for immunocompetent rats after MI led to the reduction of left ventricular interior diameter at systole, thickening of the anterior wall and improvement in cardiac function as assessed by fractional shortening. Interestingly, the cells were not detectable at 4 weeks after transplantation, suggesting that sustained cell retention was not necessary for cardiac improvement. A similar study of transplanting an epicardial patch consisting of multilayered rat bone marrow-derived MSCs in a decellularized pericardial scaffold also led to improved heart function and the formation of MSC-derived neovasculature [124]. We have also delivered bio-compatible matrices to the infarcted myocardium as acellular matrices or in conjunction with cells [64,125]. We demonstrated that local injections of collagen I, Matrigel and fibrin into the infarcted rat myocardium induced significantly higher capillary density, in comparison to saline control treatment, and could stimulate cell infiltration [64]. Whether or not the induction may be related to matrix-mediated activation of angiogenic factors or prosurvival cues needs to be further examined. Together, these studies illustrate the therapeutic potential of MSCs for increasing neovascularization and vascular differentiation, as well as improving cardiac function.

Another area of research interest is in the improvement of MSC viability after transplantation into the ischemic myocardium. Owing to the low survival rate of transplanted cells, Mangi et al. showed that the overexpression of Akt1 could improve MSC survival in vitro and in vivo [126]. Using rat CD34-negative bone marrow-derived MSCs that were retrovirally transduced with Akt1 containing a GFP reporter construct, the authors found that in vivo delivery of these modified cells maintained both higher cell retention and higher viability at 24 h after cell delivery into the infarcted heart, in comparison to cells modified with GFP reporter construct alone. In addition, the cells improved cardiac function and were found to express cardiac phenotypes. Related approaches to enhance MSC survival in the ischemic heart by either hypoxic preconditioning or by genetic modification of bcl2 or angiogenin expression have also resulted in higher cell retention in vivo [127–129]. These studies highlight the utility of genetic modification or preconditioning to improve transplant viability.