Adipose tissue-derived MSC (haMSC) are considered a promising stem cell type given the abundance of stem cells in this tissue, which has no donor limitation and is easily available by low-invasive methods [40]. Results in AMI treatment are modest, however, and the mechanisms involved are still not fully understood [40, 41]. Recent studies using two distinct cardiac stem cell (CSC) populations showed encouraging results in early clinical evaluation [1618].

Survival, engraftment, and persistence of transplanted cells or their progeny is extremely limited [42]. The frequently reported moderate improvement in cardiac function is thought to be produced by the liberation of paracrine factors that mediate survival, neovascularization, remodeling, and cell proliferation [43, 44]. Exosomes were also shown to mediate many MSC functions [45], later observed in CSC [46]. Analysis of the angiogenic potential of MSC-secreted factors (conditioned medium) for direct therapeutic use indicated a reduction in infarct size and conservation of systolic and diastolic cardiac output, which confirmed the value of these factors in counteracting AMI. Finally, other studies have also demonstrated the immunomodulatory properties of MSC that are related to their capacity to migrate to injury sites and/or neovascularize in ischemic areas, acting on different subsets of immune cells [47]; this property has been potentiated using genetically modified MSC [48].

Our initial studies of paMSC growth established appropriate conditions to obtain adequate cell doses for treatments. Proliferation studies showed distinct behavior of paMSC compared with human MSC grown at different oxygen concentrations. In low oxygen conditions, human MSC cultures showed significantly greater genetic stability and higher yields [36, 49]; this was related to elevated oxidative stress and DNA damage caused by growth at high oxygen tension, which helped to accelerate senescence [36, 50]. Results in the porcine model showed a similar paMSC growth profile at low and high oxygen tension, which suggests greater paMSC resistance to oxidative stress than with human MSC. Genetic stability of paMSC was confirmed in both oxygen conditions [51], and cells maintained multipotent differentiation capacity. Some differences were nonetheless observed in comparison with haMSC, which suggested intrinsic biological differences that could affect the therapeutic responses of paMSC vs human MSC.

Post-AMI stimuli activate CSC mediated by paracrine feedback between myocytes and the CSC. In response to stress, myocytes produce growth factors and cytokines for which CSC have receptors [25]. After demonstrating that CSC respond to growth factors secreted by adjacent myocytes, Urbanek et al. confirmed the efficiency of combined recombinant IGF-1/HGF treatment in mice and dogs [52]. CSC are activated in situ by local administration of IGF-1/HGF in a porcine heart infarction model [53], which improved ventricular function in pigs. Therefore, intracoronary administration of these factors was proposed as a strategy to reduce post-AMI cardiac remodeling and induce cardiac regeneration [25]. Finally, direct IGF-1 + HGF administration was recently evaluated in a porcine model of chronic myocardial infarction (MI), in which growth factor delivery reduced pathological hypertrophy, led to formation of new small cardiomyocytes, and increased capillarization [54].

Priming (preconditioning) of MSC or co-administration with growth factors is also used to augment therapeutic potential. An IGF-1 + HGF combination loaded in polylactic-co-glycolic acid microcarriers with haMSC enhanced engraftment of the transplanted haMSC cells and showed a 1.3-fold higher density of medium-sized blood vessels in the infarct border zone [29]. hMSC preconditioning with IGF-1 prior to transplant in infarcted rats increased engrafted cell survival in the ischemic heart, decreased myocardium cell apoptosis, and reduced inflammatory cytokines [55]. In most preconditioning strategies, IGF-1 is also proposed as a mediator [56].

IGF-1 and HGF are thus being evaluated in different modalities of cardiac repair [43, 57], although the mechanisms involved remain to be fully understood. Some studies showed that HGF treatment post-AMI attenuates systolic cardiac remodeling and cardiac dysfunction, with a cardioprotective effect; these effects were linked to angiogenic and anti-apoptotic mechanisms [58]. Evidence also implicates IGF-1 in vascular protection, which might be beneficial in chronic cardiac insufficiency [59] and in treatment of sepsis-associated cardiac dysfunction [60].

Given the promising results with direct HGF and/or IGF-1 administration [25, 54], several attempts have been made to engineer MSC to vehiculate IGF-1 or HGF expression. Kouroupi et al. manipulated neural stem/precursor cells (NPC) to overexpress IGF-1; using live-imaging techniques, they reported that IGF-1 transduction enhanced the motility and tissue penetration of grafted NPC [61], although no significant in vivo improvement was demonstrated [62]. Human MSC and paMSC were transduced with lentiviral vectors to overexpress IGF-1 [63]; overexpression of this gene improved induction by 5-azacytidine and promoted limited cardiomyocyte-like differentiation [63]. Experience is broader for genetic manipulation of MSC to overexpress HGF. Early work with rat BM-MSC showed decreased infarcted scar area and increased angiogenesis in HGF-MSC-treated animals [64]. In the porcine model, paMSC (alone or vascular endothelial growth factor (VEGF)/HGF-transfected) improved cardiac function and perfusion, probably by increasing angiogenesis and reducing fibrosis; MSC + HGF was superior to MSC + VEGF [65].

We used the porcine model to explore the synergistic effect of combined, sustained administration of paMSC modified individually to overexpress IGF-1 and HFG, labeled with fluorescent markers. This approach would allow later adjustment of the balance between growth factor supply by altering the ratio of the two paMSC populations. We generated and validated optimized lentiviral vectors and transduced paMSC, followed by purification, by which we obtained enriched paMSC-IGF-1-GFP and paMSC-HGF-Cherry cells that were evaluated in vitro and in vivo. Results showed that the cell doses used in the animals caused no toxicity or short-term safety problems.

We found improvement in cardiac function (increases in LVEF, cardiac output, and stroke volume, and reduction in heart rate and infarction area) in all groups from day 15 (T3), although these changes were not statistically significant in any case and were similar to non-paMSC-treated infarcted controls (group I). Although intragroup variability was marked, it did not appear to be the main reason for these results. Overall results for treatment group III pigs (paMSC IGF/HGF) were in fact poorer than those for group II (paMSC-GFP; paMSC bearing the empty vector). At termination of the in vivo experiment, the heart apex and IVS showed fibrotic areas during macroscopic assessment and sampling, and evaluation of infarction degree by Masson’s trichrome staining showed a larger proportion of affected areas in group III (see Additional file 10: Figure S8), and these pigs had larger infarctions than in other groups.

Although paMSC could not be identified in all samples, immunohistochemistry and molecular assays confirmed live paMSC in some group II and III tissues at 1 month (T4) follow-up, which suggests that the functional results are not due to elimination of paMSC-IGF- 1/HGF. Comparable results have been reported after autologous porcine BM-MSC implant for treatment of aortic injury [66].

Hematoxylin/eosin staining was used to evaluate the degree of resolution of cardiac damage after each treatment. The control group (I) had the lowest degree of damage, mainly in Heart 2 and IVS samples, but pericarditis was more severe compared with other groups and lesions were visible in all sections. By contrast, group III Heart 2 and IVS sections were the most affected by fibrosis and inflammation, followed by those of group II. Angiogenesis was also more evident in sections from group III than the other groups. We also found a correlation between fibrosis and neovascularization, especially in groups II and III.

Several factors could contribute to the lack of significant differences. Kren et al. showed accelerated healing and repair kinetics in young pig models of AMI reperfusion like ours [67], which would limit the length of the experimental window. In addition, high intragroup variability and a reduced number of animals per group are study limitations, especially in the case of negative results.

The time of paMSC transplant also influences effectiveness; the optimal range varies from the time of AMI to 1 week later, although a specific suitable time point remains to be established [68]. Beneficial effects have been reported following in vivo transfer of modified MSC or recombinant factors from 1 week to 1 month post-infarct induction. In addition, recent studies have shown that a large proportion of the injected cells are lost from the myocardium within the first few minutes post-injection and not more than 0.1–15 % are retained.

Although positive therapeutic results are reported for IGF-1- and HGF-expressing MSC individually, we conclude that co-administration of paMSC that overexpress IGF-1 and HGF does not appear to have a synergistic effect or promote effective cardiac repair. This could be caused by interference from high local levels of either factor. The positive correlation of enhanced neovascularization and fibrosis in a number of paMSC-IGF-1/HGF-treated pigs suggests that sustained exposure to high HGF + IGF-1 levels promotes both beneficial and deleterious effects, with no regenerative advantage. In any case, global results strongly suggest that delivery of growth factor by implantation of biodegradable microparticles in the affected area is superior to the transplant of paMSC-mod cells that probably are not efficiently retained. The consequence should be a much less efficient local supply of growth factor, with reduced increment in a therapeutic index. In addition, in case of a significant therapeutic effect using modified MSC, it will be mandatory to eliminate the possibility that non-retained MSC could find other propitious niches and favor other pathological conditions [69, 70].