Vascular endothelial growth factor (VEGF) as a key therapeutic trophic factor in bone marrow mesenchymal stem cell-mediated cardiac repair

INTRODUCTION

Clinical trials of mesenchymal stem cell (MSC) therapies are currently being conducted for treating a broad spectrum of diseases [1]. Remarkably, although the multi-lineage differentiation potential (stemness) of MSCs was originally thought to mediate their therapeutic attributes, it has now become clear that the secretion of multiple growth factors and cytokines (trophic action) by MSCs is primarily responsible for many of the observed therapeutic benefits [2; 3]. Consistent with this notion, studies have attributed the cardiac repair function of MSCs to their trophic actions independent of their differentiation potentials [4; 5; 6]. We have in particular demonstrated that while MSCs injected into the hind limb muscle are trapped in the local musculature with no detectable migration to the heart, the administered MSCs exert a prominent cardiac repair function in a hamster heart failure model, and this therapeutic effect can be reproduced with a cell-free strategy based on intramuscular injections of MSC-conditioned medium [7]. These studies together illustrate that the trophic mediators secreted by MSCs improve cardiac function by a combination of multiple mechanisms such as attenuating tissue injury, inhibiting fibrotic remodeling, promoting angiogenesis, mobilizing host tissue stem cells, and reducing inflammation, although the bio-active factors remain largely undefined.

While clinical trials of MSCs have demonstrated no significant adverse side-effects even with multiple cell administrations [1; 8], long-term safety of MSC therapy remains to be evaluated. Notably, MSC transformation has been reported in prolonged ex vivo expansion [9; 10; 11], and over-expression of telomerase or inactivation of the Wnt pathway has been shown to provoke tumor formation from MSCs [12; 13]. MSC abnormalities have also been detected in patients with multiple myeloma [14; 15], suggesting that MSCs may be sensitive to precancerous or cancerous state of the host. Since MSC-derived trophic factors appear to underlie the cardiovascular therapeutic effects, they may be explored for development of trophic factor-based cell-free therapeutic strategies, thus eliminating the safety concern associated with cell transformation. Toward this end, we have begun to characterize the trophic actions of human and porcine bone marrow MSCs in relation to their therapeutic efficacy [7; 16; 17]. We report here that cardiac repair mediated by the trophic actions of MSCs can be critically influenced by the ability of MSCs to express and secrete adequate levels of vascular endothelial growth factor (VEGF). This finding bears significance to development of MSC-based therapeutic regimens for cardiac regeneration.

METHODS AND MATERIALS

RESULTS

Preclinical and clinical studies of myocardial stem cell therapy often use invasive cell delivery approaches such as intramyocardial injection or intracoronary infusion, which can introduce harmful scar tissue, arrhythmia, calcification, or microinfarction in the heart [20; 21; 22; 23]. Aiming at developing a safer cell delivery regimen and taking advantage of the trophic actions of MSCs, we recently demonstrated a prominent cardiac repair mechanism achieved by injections of pMSCs or pMSC-conditioned medium into the hind limb muscle of TO2 cardiomyopathic hamsters [7; 18]. Intramuscularly injected pMSCs and pMSC-conditioned medium each significantly improved ventricular function one month post therapy. The functional improvement is associated with mobilization of bone marrow progenitor cells, increased myocardial capillary density, renewed cardiomyocyte populations, attenuated apoptosis, and decreased fibrosis [7].

We were particularly encouraged by the finding that intramuscular injections of pMSC-conditioned medium could achieve significant cardiac repair as this strategy could be further optimized for a trophic factor-based cell-free therapeutic regimen, eliminating the concern of potential MSC transformation after cell administration [10; 11; 12; 13]. Since we noted species-related differences in the growth and trophic actions of MSCs during our characterizations of hMSCs and pMSCs [16], we first examined whether hMSCs delivered to the hind limb muscle of TO2 cardiomyopathic hamsters could also repair the failing hamster heart as demonstrated with pMSCs [7]. Figure 1A shows that hMSCs and pMSCs both significantly increased LVEF by ~30% four weeks after the cell therapy. We then compared the therapeutic potency of hMSC- and pMSC-conditioned medium (designated as hMSC CM and pMSC CM, respectively). Notably, although the injected hMSCs were therapeutically effective, hMSC CM failed to improve ventricular function of the failing hamster heart (Fig. 1B), suggesting that the hMSC-conditioned medium might not contain an adequate level of therapeutic trophic factors.

To further understand the difference in the trophic actions of hMSCs and pMSCs, we used qRT-PCR to quantify expression of growth factor and cytokine genes. Figure 2 shows that among the seven genes analyzed pMSCs significantly expressed higher levels of insulin-like growth factor 1 (IGF1), monocyte chemotactic protein 1 (MCP1), and vascular endothelial growth factor-A (VEGF) than hMSCs. Since VEGF has been used in clinical trials for heart disease [24; 25], the gene was singled out for further analysis. We further performed VEGF ELISA to validate the qRT-PCR data. Figure 3 shows that pMSC CM contained ~5 fold higher VEGF concentration than hMSC CM (225 ± 17 vs. 40 ± 5 pg/ml; P<0.001), which is in agreement with the qRT-PCR data presented in Figure 2.

To determine whether the decreased VEGF output might be associated with the observed therapeutic deficit of hMSC CM, we overexpressed VEGF in hMSCs using a recombinant adenoviral vector as described previously [17; 26], and prepared conditioned medium (designated as VEGF hMSC CM) from the VEGF-expressing hMSCs. Figure 3 reveals that infection of hMSCs by the VEGF adenoviral vector caused a pronounced increase in VEGF concentration of the conditioned medium (10,118 ± 210 pg/ml). Injections of VEGF hMSC CM were subsequently found to completely restore the cardiac therapeutic potency of the hMSC-conditioned medium (Fig. 1B). Thus, although hMSCs exhibit a lower VEGF output compared to pMSCs, this deficit can be readily remedied by VEGF overexpression. These findings thus define VEGF as a key therapeutic trophic factor in MSC-mediated cardiac repair.

DISCUSSION

Despite recent recognition that the trophic actions of MSCs underlie their therapeutic effects in various preclinical and clinical studies [2; 3; 4], the trophic factors responsible for tissue functional improvement after MSC therapy remain largely undefined. While our study explored the feasibility of using a cell-free therapeutic strategy for cardiac repair based on MSC-conditioned medium, which can eliminate the concern of potential MSC transformation [9; 11; 12; 13], it further demonstrated VEGF as a key therapeutic trophic factor in MSC-mediated cardiac repair. This demonstration is consistent with a recent report showing that MSC-derived VEGF plays a critical role in attenuating acute kidney injury [27].

Development of trophic factor-based cell-free therapeutic regimens for tissue repair can be facilitated through analysis of the trophic actions of MSCs as demonstrated here. Using a cocktail of well-defined trophic factors for this purpose is not only pharmaceutically sound but also strategically advantageous since the trophic factor cocktail can be precisely modified to treat different degenerative conditions. Our comparative study of pMSC CM, hMSC CM, and VEGF hMSC CM suggests that a cocktail with a VEGF dosage greater than ~200 pg/ml is sufficient to elicit significant cardiac repair for the failing hamster heart in conjunction with other yet undefined MSC trophic factors. This VEGF dosage amounts to ~2 ng VEGF/kg body weight per intramuscular injection, which is three orders of magnitude lower than the typical VEGF dosages used in single growth factor therapeutic trials (>10 μg/kg body weight per injection) [28]. Interestingly, we have found that VEGF when administered alone (in the absence of MSC trophic factors) could also rescue the failing hamster heart at higher dosages (0.1–1 μg/kg body weight per intramuscular injection) [29]. Despite repeated intramuscular injections, side effects such as tissue edema, angioma formation, and inflammation, which have been associated with VEGF administration [30; 31], were not observed in our study. A prominent feature of the intramuscular VEGF or MSC delivery approach is that the local muscular bed produces additional beneficial soluble mediators in response to trophic factor signaling [7; 29]. Activation of the host muscle trophic factor network thus further amplifies the trophic actions of MSCs, culminating in an effective cardiac repair mechanism.

MSCs are thought to have a perivascular origin, and can be derived from multiple human organs including bone marrow, skeletal muscle, placenta, pancreas, and adipose tissue [32]. It is presently unknown whether the therapeutic potency of MSCs may be affected by the tissue source. In transitioning toward the clinical use of hMSCs, additional variables such as differences in donor gender, age, and life style can potentially complicate the logistic aspect of MSC therapy. For instance, limited information indicates that female stem cells may possess a more pronounced regenerative potential than male stem cells [33], which is in line with the finding that female patients typically exhibit certain cardioprotective phenomenon from acute myocardial infarction and better outcome after the incidence compared to male patients [34]. Although these gender influences are thought to be mediated through differential sex hormone receptor signaling, a recent study shows that female rodent MSCs produce a higher level of VEGF than male rodent MSCs in response to hypoxia [35]. Our demonstration here that VEGF hMSC CM, but not hMSC CM, improved cardiac function further highlights the critical role of VEGF in mediating the therapeutic function of MSCs. We suggest that the cardiac therapeutic potency of a given donor MSC may correlate with the cell’s ability to express and secrete an adequate level of VEGF, which can be readily assessed by qRT-PCR and ELISA as shown here.

The cardiovascular effects of VEGF appear to stem from its multiple biological attributes. VEGF is well known for its ability to mobilize bone marrow progenitor cells, and these recruited progenitor cells can participate in myogenesis and angiogenesis [36]. In addition to being an angiogenic growth factor, VEGF also possesses a nonangiogenic role important for the cardiovascular system [37]. VEGF can promote differentiation of stem cells into cardiomyocytes and endothelial cells [38; 39], and boost the function of MSCs and endothelial progenitor cells [40; 41]. We have demonstrated that MSCs engineered to overexpress VEGF exhibit more pronounced trophic factor outputs [17]. These trophic factors acting in concert can mediate cardiac repair through their progenitor cell-mobilizing, angiogenic, cytoprotective, myogenic, anti-fibrotic, and anti-inflammatory properties [7; 16]. Interestingly, an MSC-derived anti-inflammatory factor designated as TSG-6 has been found to mediate cardiac repair after myocardial infarction [6]. Given that tissue damage and degeneration is complex in nature, multiple MSC trophic factors acting in synergy are expected to be involved in tissue healing. Identification of these trophic factors can eventually lead to the development of cell-free trophic factor cocktails suitable for various clinical applications.