Author contributions: D.S.: conception and design, collection and assembly of data, final approval of manuscript; I.T.: conception and design, collection and assembly of data, final approval of manuscript; S.M.: conception and design, administrative support, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; N.N.: conception and design, collection and assembly of data, final approval of manuscript; N.H.: conception and design, collection and assembly of data; H. Tsuji: conception and design, collection and assembly of data, final approval of manuscript; H. Tsuruta: conception and design, collection and assembly of data, final approval of manuscript; K.S.: conception and design, collection and assembly of data, final approval of manuscript; Y.T.: provision of study material or patients, collection and assembly of data, final approval of manuscript; S.O.: financial support, administrative support, final approval of manuscript; A.U.: financial support, administrative support, final approval of manuscript. D.S. and I.T. contributed equally to this article.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS December 9, 2010.
The efficacy of transplantation of default human marrow-derived mesenchymal stem cells (MSCs) was modest. In this study, our challenge was to improve the efficacy of MSC transplantation in vivo by pretreatment of MSCs with pioglitazone. MSCs were cultured with or without medium containing 1 μM of pioglitazone before cardiomyogenic induction. After cardiomyogenic induction in vitro, cardiomyogenic transdifferentiation efficiency (CTE) was calculated by immunocytochemistry using anti-cardiac troponin-I antibody. For the in vivo experiments, myocardial infarction (MI) at the anterior left ventricle was made in nude rats. Two weeks after MI, MSCs pretreated with pioglitazone (p-BM; n = 30) or without pioglitazone (BM; n = 17) were injected, and then survived for 2 weeks. We compared left ventricular function by echocardiogram and immunohistochemistry to observe cardiomyogenic transdifferentiation in vivo. Pretreatment with pioglitazone significantly increased the CTE in vitro (1.9% ± 0.2% n = 47 vs. 39.5% ± 4.7% n = 13, p < .05). Transplantation of pioglitazone pretreated MSCs significantly improved change in left ventricular % fractional shortening (BM; −4.8% ± 2.1%, vs. p-BM; 5.2% ± 1.5%). Immunohistochemistry revealed significant improvement of cardiomyogenic transdifferentiation in p-BM in vivo (BM; 0% ± 0% n = 5, vs. p-BM; 0.077% ± 0.041% n = 5). Transplantation of pioglitazone-pretreated MSCs significantly improved cardiac function and can be a promising cardiac stem cell source to expect cardiomyogenesis. STEM CELLS 2011; 29:357–366
Cardiac stem cell therapy is a promising therapy for patients with severe heart failure; however, it is still less-developed. One of the critical problems is absence of ready-to-use stem cell sources. Embryonic stem cells  or induced pluripotent stem cells  have cardiomyogenic differentiation potential and are major candidates for future stem cell therapy, but, there are still several problems to overcome before clinical application, that is, neoplasm formation  and chromosomal stability . Human bone marrow-derived mesenchymal stem cells (BM-MSCs) have a cardiomyogenic transdifferentiation ability; however, the cardiomyogenic transdifferentiation efficiency (CTE) is extremely low . Residential cardiac precursor cells , having a mesenchymal phyenotype , have a higher CTE potential in comparison with BM-MSCs, but may not be so high in comparison with the other types of human mesenchymal cells, that is, umbilical cord blood-derived (UCB-MSC) , menstrual blood-derived (MMC) , placental chorionic plate-derived MSC (PCPCs) , and amniotic membrane-derived MSC (AMCs) . Moreover, these mesenchymal cells are expected to be used in an allograft manner and not as ready-to-use materials until the establishment of a stem cell bank system that will cover all major histocompatibility complexes to avoid rejection. Among these human mesenchymal types of cells, only BM-MSCs have been clinically utilized, having shown safety and modest efficacy; the efficacy was believed to be due to the paracrine effect of engrafted BM-MSCs on residual host cardiomyocytes  or neovascularization [13, 14]. However, if we were to find the method to increase the CTE of BM-MSCs and/or increase the paracrine effect of BM-MSCs, the efficacy of ongoing clinical BM-MSC-based therapy may be significantly advanced. As preferred culture methods to elicit the maximal response of the paracrine effect of human BM-MSCs were still undetermined, clinical studies might do not yet show a substantial paracrine effect of BM-MSCs.
In our previous data , we found that the CTE of MMCs had been increased dramatically by use of L-carnitine-containing medium, which is essential for mitochondrial free fatty acid metabolism. Cardiomyocytes require free fatty acid as a major energy source. Therefore, mitochondrial function may play an important role in expressing cardiomyocyte phenotype. In fact, the mitochondrial morphology of cardiomyocytes is quite different from its morphology in other organs . The mitochondrial proteins are synthesized not only by its own DNA but also by the DNA of the host nucleus, that is, perosixome proliferators-activated receptor-γ (PPAR-γ) and PPAR-γ coactivator-1α are known to activate mitochondrial gene transcription . From another point of view, we speculated that environmental stimulation may cause change in mitochondrial function that might affect the gene expression of the host nucleus DNA, which may also play a role in the CTE of MSCs.
In this study, we hypothesized that stimulation of PPAR-γ by pioglitazone may increase the CTE of human MSCs. We examined the effect of pioglitazone, as a PPAR-γ activator, on efficiency of cardiomyogenic transdifferentiation of human BM-MSCs, in vitro and in vivo.
MATERIALS AND METHODS
Isolation of Human Mesenchymal Cells
After informed consent was obtained, 15 ml of bone marrow aspirate was obtained from the iliac bone of a 41-year-old male as described previously . The collected sample was cultured in Dulbecco's modified Eagle's medium high-glucose supplemented with 10% human serum and established MSCs (BM-MSCs). BM-MSCs of 10–20 population doublings were used in the present study. We performed bone marrow aspiration procedure for six times and there was no difference among the BM-MSCs obtained from each procedure.
Cardiomyogenic Induction and Chemical Agents
The method of cardiomyogenic induction in vitro was described previously [5, 8–10]. In short, enhanced green fluorescent protein (EGFP)-labeled BM-MSCs were cocultured with murine cardiomyocytes (see detail in supporting information Material and Methods). In this system, the incidence of cell fusion was around 0.3% [8, 9, 18] and the evidence of cell fusion-independent cardiomyogenesis was extensively shown in the previous study [5, 8, 9, 11]. In our pilot study, we have confirmed that the incidence of cell fusion was not affected by pioglitazone (0.2%). Human cells were preincubated with medium containing 3 μmol/l of pioglitazone for 2 weeks before coculture and/or culture with the pioglitazone-containing medium after coculture. In another experiment, we administrated 1 μmol/l of GW9662, a specific PPAR-γ blocker. Pioglitazone (10 mM) and GW9662 (1 mM) stock solutions were freshly prepared with dimethyl sulfoxide (DMSO) (Sigma) solution and added to the culture media to give a final concentration. Evaluation of efficiency of cardiomyogenic transdifferentiation was described previously [8, 9, 11, 15]. In short, cocultivated BM-MSCs were enzymatically isolated, a smear sample was made, and then immunocytochemistry using mouse monoclonal antibody against anti-cardiac troponin-I (Trop-I, #4T21 Hytest, Euro, Finland, http://www.hytest.fi/) antibody was performed (described later). Isolated cells (spherical shape), in which Trop-I localized at the cytoplasm were considered as Trop-I-positive cells. The CTE was defined as the incidence of Trop-I/EGFP double-positive cells in EGFP-positive BM-MSCs.
Immunocytochemistry and Immunohistochemistry
A laser confocal microscope (FV1000, Olympus, Tokyo, Japan, http://www.olympus.co.jp/) was used. As described previously, samples were stained with Trop-I, with mouse monoclonal anti-human atrial natriuretic peptide (hANP) antibody (YLEM, MCV928), or with mouse monoclonal anti-sarcomeric α-actinin antibody (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com/) and rabbit polyclonal anti-connexin 43 (Cx43) antibody (Sigma) diluted 1:300 overnight at 4°C, then stained with tetramethylrhodamine isothiocyanate-conjugated anti-mouse IgG antibody (Sigma) and Cy5-conjugated anti-rabbit IgG antibody (Chemicon, Billerica, MA, http://www.millipore.com/) diluted 1:100, containing 4′-6-diamidino-2-phenylindole (Wako, Tokyo, Japan, http://www.wako-chem.co.jp/) at 1:300 for 30 minutes at 25°C–28°C.
Enzyme-Linked Immunosorbent Assay
Angiogenic humoral factors (angiogenin, angiotensin-2, epidermal growth factor [EGF], basic fibroblast growth factor [bFGF], heparin-binding EGF-like growth factor [HB-EGF], hepatocyte growth factor [HGF], leptin, platelet-derived growth factor-BB [PDGF-BB], phosphatidylinositol-glycan biosynthesis class F protein [PIGF], vascular endothelial growth factor [VEGF]) in culture medium supernatant (cultured with 10% serum containing medium for 7 days) were measured by enzyme-linked immunosorbent assay (ELISA). The assay was performed with Quantibody Human Angiogenesis Array I kit (RayBiotec, Inc., Norcross, GA, http://www.raybiotech.com/) and was conducted according to manufacturer recommended protocol.
Transplantation of Pioglitazone-Pretreated Marrow-Derived MSCs in Myocardial Infarction Model In Vivo
Myocardial infarction (MI) was induced in the open chests of anesthetized female F344 nude rats (Clea Japan, Inc.; 6 weeks of age) as described previously . Two weeks after MI, 1–2 × 106 of EGFP-labeled BM-MSCs were injected into the myocardium at the border zone of the MI. Two weeks after the first operation, rats with MI were randomized in a blind study of the following groups: the sham-operated group (sham), the control MI group (MI), the MI+plain BM-MSCs-transplanted group (BM), and the MI+pioglitazone-pretreated MSCs-transplanted group (p-BM). After cellular transplantation, pioglitazone (2.5 mg/kg/day) was orally administered in some of the experiments (po). Randomization occurred immediately before echocardiogram. Immediately before cell transplantation, two-dimensional and M-mode echocardiographic (8.5 MHz linear transducer; EnVisor C, Philips Medical Systems, Andover, MA, http://www.healthcare.philips.com/) images were obtained to assess left ventricular end-diastolic dimension (LVEDd) and left ventricular end-systolic dimension (LVESd) at the midpapillary muscle level by a single-blinded observer. Two weeks after the transplantation, a similar echocardiogram was performed again. Left ventricular % fractional shortening (% FS), thickness of anterior wall, and thickness of posterior wall were calculated from 5–6 traces and averaged. Left ventricular pressure, brain natriuretic peptide (BNP), body weight, and heart weight (wet) were measured as previously described. Tissue samples were obtained by slicing along the short-axis of the left ventricle, for every 1 mm of depth. After masson trichrom staining, the area of fibrosis was digitized from each slice, and then the % fibrosis volume in the left ventricular myocardium was calculated, as previously described . Immunohistochemical analysis was performed using anti-rat CD34 antibody (1:200; R&D Systems, Minneapolis, MN, http://www.rndsystems.com/; AF4117) to evaluate vascular density. Then, biotinylated goat immunoglobulins (Dako, Carpinteria, CA, http://www.dako.com/; E0466) were used as a second antibody, next, strept ABC complex/Horseradish peroxidase (Dako; K0377), and finally, 3,3′-diminobenzine substrate (Wako; K3183500) was used. The images were digitized and the % brown pixel area of the capillary vessels was counted in the peri-infarct normal zone (Nz) and the center of the MI zone (MI) using a light microscope at ×10 magnification. The area in five high-power fields was calculated and averaged.
All data are shown as mean value ± SE. The difference between mean values was determined with one-way analysis of variance (ANOVA) test or one-way repeated measures ANOVA test and Bonferroni post hoc test. Statistical significance was set at p < .05.
Pretreatment with Pioglitazone Increased Efficiency of Cardiomyogenic Transdifferentiation via PPAR-γ Receptor Activation
Administration of pioglitazone did not cause any significant change in morphology of BM-MSCs (Fig. 1A, 1B), whereas improved CTE in vitro was observed. In this study, treatment with pioglitazone (pio) dramatically increased the incidence of beating BM-MSCs. Immunocytochemistry revealed a dramatic increase in incidence of cardiac troponin-I (Trop-I)/EGFP-double positive BM-MSCs in the pioglitazone-administrated experiment (Fig. 1F–1H), whereas the double-positive BM-MSCs were rare in the control (CNT) experiment (Fig. 1C–1E). The staining pattern of Trop-I showed a clear striation pattern in the differentiated BM-MSCs (Fig. 1I, 1J). EGFP and Trop-I-staining appeared alternately in a striated manner, suggesting that Trop-I was expressed in the cytoplasm of EGFP-positive cells. The transdifferentiated pioglitazone-treated BM-MSCs showed diffuse dot-like staining of hANP around the nuclei (Fig. 1K, 1L), clear striation pattern of sarcomeric α-actinin, and a diffuse dot-like staining pattern of Cx43 at the margin of the cells, whereas these did not show in the control BM-MSC experiment (supporting information Fig. 1). Similar pioglitazone-induced improvement in CTE was also observed in PCPCs and AMCs (data not shown).
To clarify the target of the pioglitazone, pioglitazone was administered only before the coculture or only after the coculture. Smear preparations of enzymatically isolated cells were stained with Trop-I (Fig. 1M) and CTE was calculated (Fig. 1N). Administration of pioglitazone after the start of coculture did not affect the CTE; on the other hand, pretreatment with pioglitazone significantly increased CTE (Fig. 1M). This suggests that pioglitazone modified the character of BM-MSCs to be able to cause higher CTE.
The effect of pioglitazone is known to be mediated by two nuclear receptors, PPAR-γ and retinoid X receptor . In the present study, therefore, we use GW9662 as a specific PPAR-γ blocker to know which receptor is the essential one for this pioglitazone effect. The result showing that GW9662 completely blocked pioglitazone-induced increase in CTE in BM-MSCs (supporting information Fig. 2), suggests this effect is caused by a PPAR-γ-dependent pathway. PPAR-γ is the nuclear receptor, which regulates gene expressions and which might affect pioglitazone-induced increases in CTE. Therefore, gene chip analysis was performed to compare the expression pattern between control and pioglitazone-treated BM-MSCs. Consistently, upregulated genes and downregulated genes by the administration of pioglitazone in BM-MSCs are shown in supporting information Table 1 and supporting information Table 2, respectively.
Effect of PPAR-γ-Activated BM-MSC Transplantation on Cardiac Function In Vivo
The BM-MSCs were transplanted into the hearts of nude rats with chronic MI, in vivo, and the effect on cardiac function was examined. Before the cellular transplantation (2 weeks), there was no difference in the % FS (Fig. 2B) at the baseline. The difference in % FS (Δ% FS) at 2 weeks after transplantation is shown in Figure 2C. Oral administration of pioglitazone did not affect Δ% FS of MI model (CNT vs. CNT+po). Default BM-MSC transplantation also did not cause a significant effect (CNT vs. BM). On the other hand, pretreatment with pioglitazone significantly improved the efficacy of BM-MSC transplantation (BM vs. p-BM, BM+po vs. p-BM+po). Oral administration of pioglitaone, also significantly improved % FS in p-BM group (p-BM vs. p-BM+po). There was no difference in ΔLVEDd in any group (Fig. 2D) and significant decrease in ΔLVESd of p-BM group compared with BM group (Fig. 2E). Thus, p-BM transplantation improved systolic function. It is also notable that the efficacy of default BM-MSC transplantation was significantly improved by oral administration of pioglitaone, suggesting that the effect of pioglitazone on BM-MSCs, to some extent, can be expected even after the transplantation. There was no difference in body weight (Fig. 3A), in serum BNP concentration (Fig. 3B), or other hemodynamic parameters (Fig. 3C–3F).
Heart weights were significantly increased by MI (Sham vs. CNT; Fig. 4A). The MI-induced increase in heart weight was blocked by orally administrated pioglitazone (CNT vs. CNT+po), and pioglitazone-pretreated BM-MSC transplantation (BM vs. p-BM). Left ventricular (LV) systolic pressure was significantly improved by pioglitazone-pretreated BM-MSCs (BM vs. p-BM; Fig. 4B). Orally administrated pioglitazone alone improved LV systolic pressure of CNT groups, whereas it did not improve echocardiographic parameters, suggesting that pioglitazone-induced water retention  might have played a role in this LV pressure to some extent.
The heart section was stained with masson-trichrome (Fig. 4C) and the fibrosis volume was measured and averaged (Fig. 4D). The % fibrosis volume was significantly decreased by p-BM transplantation compared with BM.
Mechanism for Recovery in Cardiac Functions
ELISA analysis revealed default BM-MSCs significantly secrete angiogenesis-related molecules, that is, angiotensin-II, bFGF, HB-EGF, HGF, leptin, PIGF, and VEGF in the culture medium, and they were not changed by pioglitazone administration (Fig. 5). Concordant with this result (Fig. 6G), at the MI area, BM transplantation significantly increased the vessel density at the center of the MI area (CNT vs. BM), but there was no statistical significance between BM versus p-BM. On the other hand, orally administrated pioglitazone significantly increased the vessel density at the peri-MI normal area (Nz, CNT vs. CNT+po), and additionally increased the vessel density at the center of the MI area of p-BM groups (Nz, p-BM vs. –BM+po). This angiogenic effect was also observed independent of BM-MSC transplantation, suggesting that the effect was partially caused by pioglitazone-induced VEGF-expression in vascular smooth muscle .
Immunohistochemical analysis was performed to observe the fate of EGFP-labeled BM-MSCs in situ. In the BM group and the BM+po group, there were many EGFP-positive cells in the MI area; however, these cells were enucleated and there were no EGFP/Trop-I double-positive cardiomyocytes (Fig. 7A). In the p-BM group, there were a lot of EGFP-positive cells in the MI area (Fig. 7B), and sometimes we observed EGFP-negative/Trop-I positive cells adjacent to the EGFP-positive cells (Fig. 7C), which may represent p-BM-induced survival of host cardiomyocytes in the MI zone, which may be due to an augmented paracrine effect of p-BM. Surprisingly, in the p-BM+po group, the survival rate of EGFP/Trop-I double-positive cells are significantly increased (Fig. 7I) and there were many EGFP/Trop-I double-positive cells at the border zone of the MI (Fig. 7D, arrows) that showed a clear striation staining pattern of Trop-I (Fig. 7E). Many EGFP/sarcomeric a-actinin double-positive cardiomyocytes were also observed (Fig. 7F) that showed a clear striation staining pattern (Fig. 7G). Furthermore, there was band-like staining of Cx43 expression at the margin of the EGFP-positive cardiomyocytes to the host cardiomyocytes (Fig. 7G), suggesting tight electrical coupling. The α-actinin and EGFP staining were observed alternately in a striated manner (*), suggesting that α-actinin is expressed in the EGFP-positive cells (Fig. 7H). In the p-BM+po group, cardiomyocytes were enzymatically isolated and EGFP-positive rod-shaped cardiomyocytes (striation was observed by phase-contrast image, supporting information Fig. 3A, 3B) and EGFP-negative rods were selected; then, the fluorescence in situ hybridization experiment was performed. The EGFP-negative rods were negative for human-specific Alu and positive for Rat-X chromosome (supporting information Fig. 3C–3F). On the other hand, EGFP-positive rods were positive for human-specific Alu, whereas negative for Rat-X (supporting information Fig. 3G–3J), suggesting that EGFP-positive cardiomyocytes were derived from human BM-MSCs not by cell fusion but by cardiomyogenic transdifferentiation.
In this study, pioglitazone dramatically increased the CTE of human BM-MSCs via PPAR-γ activation in vivo and in vitro. Pioglitazone-pretreated BM-MSC transplantation significantly improved impaired cardiac function in the MI model in vivo and significantly reduced % volume of the MI. Orally administrated pioglitazone significantly improved cardiac function further when BM-MSCs were transplanted, and significantly improved the survival of BM-MSC-derived cardiomyocytes in vivo.
The Mechanisms of Improvement in CTE by PPAR-γ Activation
Pioglitazone, a major antidiabetic agent, enhances adipocyte differentiation via PPAR-γ activation  and secretion of adiponectin , and improves insulin sensitivity in patients with type II diabetes mellitus . The endogenous agonist for PPAR-γ receptor is 15-deoxy-Δ(12,14)-prostaglandin J2 (15d-PGJ2), an end product of arachidonic acid and prostaglandin D2. As, PPAR-γ abundantly distributes in adipose tissue, the main target has been believed to be adipose tissue; however, PPAR-γ is known to be expressed in mesenchymal cells , and PPAR-γ activation changed the phenotype of mesenchymal cells during cardiomyogenesis. In the present study, we concluded that pioglitazone-induced increase in transdifferentiation was caused by PPAR-γ, because it was completely blocked by GW9662, a specific PPAR-γ blocker. The favorable effect on CTE was observed by rosiglitazone (1 μmol/l, 37.0% ± 2.9% n = 20), also suggesting the class effect.
It is notable that neither administration of pioglitazone alone (without cocultivation) nor administration of pioglitazone after the onset of cocultivation caused cardiomyogenic transdifferentiation. These data suggest that pioglitazone did not directly induce the cardiomyogenic transdifferentiation or augment the effect of cardiomyogenic transdifferentiating humoral factors derived from murine cardiomyocytes. In our pilot study, pretreatment with pioglitazone for 2 days before the start of cocultivation did not cause any improvement in efficiency, suggesting the slow effect of pioglitazone. From these observations, we concluded that pretreatment with pioglitazone changed the cellular biology of the MSC and we, therefore, called this state a “potentiation of cells,” in the present study, but precise molecular mechanisms remain elucidated.
Furthermore, mechanisms of cardiomyogenic “transdifferentiation” are also unclear. Our previously reported AMCs , UCB-MSCs , MMCs, and endometrial gland-derived mesenchymal cells (EMCs)  might be referred to as “potentiated cells.” Despite this potentiated cells express almost all cardiomyocytes-specific genes, that is, Nkx 2.5, GATA-4, cardiac troponin-I. at the default state [8–10], however, they did not spontaneously transdifferentiate into cardiomyocytes. Therefore, we speculated that some epigenetic stimulation or cell-to-cell communication may play an important role in causing cardiomyogenic transdifferentiation of the MSCs.
The Effect of Orally Administrated Pioglitazone in Cardiac Functions of MI Model
The orally administrated pioglitazone alone significantly reduced MI-induced cardiac hypertrophy and LV systolic pressure, and did not affect other hemodynamic parameters. Therefore, the effect on LV pressure may be caused in part by pioglitazone-induced fluid retention . Despite increased vascular density at the peri-MI by pioglitazone administration, % fibrosis volume was not changed; therefore, this angiogenic effect could not help reduce % fibrosis volume, in the present study. Shiomi et al.  clearly showed that the oral administration of pioglitazone significantly improved left ventricular systolic functions in the acute MI model; however, in the chronic MI model , the effect of pioglitazone could not be observed. Thus, we concluded that orally administrated pioglitazone alone did not cause any improvement of ventricular systolic functions in the chronic MI model and the beneficial effect was observed only when the BM-MSCs were transplanted, suggesting that the effects are mediated by BM-MSCs.
The Effect of Pioglitazone-Pretreated Human Marrow-Derived MSCs on Cardiac Functions In Vivo
In this study, BM-MSCs slightly improved Δ% FS, but there was no statistical significance. On the other hand, pretreatment with pioglitazone of BM-MSCs significantly improved Δ% FS, ΔLVESd, LV systolic pressure, % fibrosis volume, and normalized post MI-induced hypertrophy of heart weight. As the number of surviving BM-MSC-derived cardiomyocytes was rare, the improvement in cardiac function may not be due to newly generated cardiomyocytes. Pioglitazone pretreatment did not affect secretion of angiogenic factors in vitro and vessel density in vivo; thus, the effect of the pioglizatone on cardiac function may not be caused by the angiogenesis. We speculated that the potentiated BM-MSCs might increase secretion of unspecified factors, which may be contributing antiapoptotic action to the host myocardium  or suppressing post MI left ventricular remodeling.
Our data showed that the effect of “human” BM-MSCs at the default state on cardiac function was low. The discrepancy from a previous experimental report  performed by “nonhuman” BM-MSCs might be due to low angiogenic effect or low paracrine effect of “human cells.” From another aspect, our data of default human BM-MSCs can explain the poor efficacy of BM-MSC in the clinical reports . Such poor efficacy might be due to low angiogenic and/or paracrine effect of “human” BM-MSCs. Therefore, if we potentiate BM-MSCs by means of pioglitazone, we may be able to expand the efficacy of BM-MSCs in the clinical study.
Comparison of Mesenchymal Cells with Other Organs
In contrast to other mesenchymal cells, BM-MSCs can be used as an autograft and have been transplanted into a clinical patient ; therefore, BM-MSCs are ready to use and are still an important candidate as a cardiac stem cell source. On the other hand, CTE of BM-MSCs at the default state was far less than other mesenchymal cells, that is, MMC , AMC , UCB-MSC . In this study, pretreatment with pioglitazone dramatically improved the CTE of BM-MSCs in vitro. We suspected that the difference in CTE of mesenchymal cells obtained from various organs might be due to different degrees of stimulation with intrinsic 15d-PGJ2 of each mesenchymal cell at the default state. But in our preliminary observation, pretreatment with GW9662 did not suppress the CTE of MMCs and EMCs; therefore, the precise mechanism of different CTE of mesenchymal cells at the default state is still unclear.
From the in vivo experiment, we feel the number of surviving BM-MSC-derived cardiomyocytes is not enough for restoring impaired cardiac function, at this time. But taking into account the extremely low CTE in the previous article [5, 13], it is a great advance that such a significant number of human BM-MSCs can be transdifferentiated into cardiomyocytes in vivo by use of pioglitazone. Our experimental model might provide us with a first clue as to the adequate BM-MSC-transplantation methods that can be expected to show dramatic cardiomyogenesis in vivo. Further experimentation should be done.
From another point of view, the beneficial effect of pioglitazone on cardiac function might be caused by augmentation of a paracrine effect of BM-MSCs. In this study, precise mechanisms for a paracrine effect were not yet determined, as it may be caused by unknown molecules and/or may be caused by a cell-mediated mechanism (i.e., immune cells). The precise molecular or cellular mechanism should be elucidated in the near future.
For the in vitro experiment, pioglitazone pretreatment before cardiomyogenic induction was necessary and was shown to be a sufficient condition for dramatic improvement of cardiomyogenic transdifferentiation. In contrast, not only pretreatment with pioglitazone but also additional oral administration of pioglitazone was essential to observe a significant number of surviving BM-MSC-derived cardiomyocytes in vivo. The apparently different results between in vitro and in vivo may be explained by a significant loss in the number of transplanted BM-MSCs that might have occurred in the in vivo condition, that is, apoptosis. and that might have been improved by orally administered pioglitazone. The orally administered pioglitazone-induced angiogenic effect at peri-MI normal zone and/or antiapoptotic effect  might be essential for the survival of transdifferentiated cardiomyocytes. Improvement of survival of BM-MSCs in situ might also improve the effect of transplantation of BM-MSCs in this study.
In this study, BM-MSCs were obtained from a single donor. However, the same effect of pioglitazone was observed in BM-MSCs from neonate and PCPCs , AMCs  in our pilot study, suggesting that the effect of pioglitazone can be applied generally to human mesenchymal cells.
The fluorescent intensity of EGFP significantly decreases around 2 weeks after the transfection ; therefore, we examined cardiac function and performed the histological experiment at 2 weeks after the transplantation, in the present study. The durability of the preferable effect of pioglitazone-treated BM-MSC transplantation for more than 2 weeks was unclear.
Not only improvement of cardiomyogenic transdiffereniation efficiency but also improvement of paracrine action of BM-MSCs was elicited by pretreatment with pioglitazone. Pioglitazone is a popular medicine, and clinical application of transplantation of BM-MSCs for hearts has already been done; therefore, our proposed method is a ready-to-use method. Cardiac stem cell therapy by autologus BM-MSC transplantation can improve the efficacy and expected cardiomyogenic transdifferentiation by pioglitazone treatment. Pioglitaozne-pretreated BM-MSC transplantation can be a promising cardiac stem cell therapy and may dramatically advance the modest efficacy of stem cell therapy at the present time.
The research was partially supported by a grant from the Ministry of Education, Science and Culture, Japan. A part of this work was undertaken at the Keio Integrated Medical Research Center.