Main

In recent years, cell-based therapy has emerged as a potential new strategy for organ repair. The optimal source of cells for repairing damaged tissue is a topic of intense research. Bone marrow mononuclear cells (BMNCs) have been the cell source most frequently used for the angiogenic cell therapies applied for cardiovascular disease.1, 2, 3, 4, 5 Two randomized controlled trials in Europe, however, failed to prove any functional benefit of intracoronary applied BMNCs in patients with acute myocardial infarction (MI).4, 5 The results of these trials suggest that the effects of BMNC implantation may be insufficient to improve cardiac function. A new cell therapy using other cell types, a therapy more effective than BMNC implantation, is expected to be established soon.

Our group previously demonstrated the important contributions of the stem/progenitor cells among the BMNCs in improving limb ischemia.6 There has been considerable interest in the development of a new therapy with a non-hematopoietic subset of adult bone marrow (BM) stem/progenitor cells referred to as the mesenchymal stem cells or multipotent stromal cells (MSCs) in cardiac repair or regeneration.7, 8, 9, 10, 11 MSCs have been shown to protect against ischemic injury through both direct prevention of cell death and through the stimulation of angiogenesis.7, 11 In experiments comparing cultured human MSCs with BM hematopoietic stem/progenitor cells, Iso et al7 found that the former expressed higher mRNA levels for angiogenic factors such as vascular endothelial growth factor (VEGF) and adrenomedullin. Encouragingly, translational studies on small and large models toward clinical implication have shown that MSC administration in the acute phase of MI improves cardiac function.8, 9, 12, 13, 14 It remains undetermined, however, whether MSC implantation is effective in a chronic MI or heart failure model.

As the selection of cell types advances, methods and devices for cell delivery will also have to be developed for clinical cell therapies. The methods used for cell implantation have included direct injection into myocardium after thoracotomy,14 transendocardial injection via an injection catheter,8 and intracoronary injection via a balloon catheter.9 The first of these methods, via thoracotomy, poses an increased risk of morbidity and mortality, as it requires highly aggressive intervention and general anesthesia. The second method, transendocardial injection with a catheter, also has flaws, including risks such as cardiac tamponade and ventricular arrhythmia. As hopeful alternatives, new methods of cell transplantation have recently been reported.15, 16, 17, 18 We previously demonstrated the safety and efficacy of retrograde injection of BMNCs via the coronary vein into the myocardium through a single balloon infusion catheter in a swine MI model.15

Our group conducted this study for two purposes: first, to investigate whether the low invasive delivery of MSCs via the coronary vein restores cardiac function in chronic ischemic cardiomyopathy model; second, to explore the mechanism behind the improvement, if any improvement is found.

MATERIALS AND METHODS

Isolation of Swine MSCs

This study was conducted according to the Guide for the Care and Use of Laboratory Animals published by the United States National Institutes of Health. The experimental protocol was approved by the Animal Care and Use Committee of Showa University School of Medicine.

Swine BM cells were harvested from the femurs of the swine as the MI model was being created (Figure 1a). These cells were resuspended in MSC culture medium (10% fetal bovine serum in Dulbecco's modified Eagle's medium containing 4.5 g/l glucose) with antibiotic/antimycotic supplements (Invitrogen, Carlsbad, CA, USA), and the cultures were maintained at 37°C in a humidified atmosphere containing 95% air and 5% CO2. When the cultures reached subconfluence, the cells were harvested with 0.25% trypsin and 1 mM EDTA, then one-half of the harvested cells were replaced. After one series of passages, the attached MSCs were devoid of hematopoietic cells. After two series of passages, the cells were harvested again with 0.25% trypsin and 1 mM EDTA, washed with phosphate buffer saline, labeled with Celltracker Orange (Invitrogen), and resuspended in 10 ml RPMI Medium 1640 (Invitrogen).

Figure 1
figure 1

Experimental protocol and procedure. (a) Schedule and scheme for the experimental protocol. (b) Left coronary angiography and coronary venography (left), and X-ray image at balloon inflation with cell infusion (right). MI, myocardial infarction. BM, bone marrow. MSC, multipotent stromal cell.

Full size image

Chronic MI Model and Cell Implantation

In all, 19 male domestic swine, aged 2–4 months and weighing about 25 kg, were used. After receiving an intramuscular injection of ketamine (20 mg/kg) and atropine (60 μg/kg), each animal was intubated, ventilated, and maintained with a mixture of halothane (2%) and oxygen. Electrocardiograph monitoring was performed continuously, and a coronary angiography was carried out after inserting an introducer from the right cervical artery. Next, the guidewire was inserted into the left anterior descending coronary artery and sterilized beads were placed in the mid left anterior descending coronary artery by microcatheter (Figure 1a and b) according to a method previously described.15 Peripheral blood was collected before and 1 and 4 days after MI. Serum creatine kinase (CK) levels were determined.

At 4 weeks after the induction of MI by permanent coronary occlusion with the beads, the 13 surviving animals were anesthetized again by the same procedure. After coronary angiography and left ventriculography (LVG), a balloon catheter (2.5 mm in diameter) was advanced through the coronary sinus into the anterior intraventricular vein and positioned just beside the occlusive site of the left anterior descending coronary artery (Figure 1a and c). The swine were randomly assigned to either an MSC infusion group (MSC group: n=8) or medium infusion group (control group: n=5). In the MSC group, MSCs (1.02±0.34 × 107 cells) labeled with the red fluorescent dye were suspended in 10 ml culture medium and infused retrogradely through the balloon catheter from the anterior intraventricular vein over a 10-min period of balloon inflation. The same volume of medium was infused in the control group by the same procedure.

LVG was performed before and at 4 weeks after the cell implantation to determine cardiac function as previously described.19 Later, on the completion of follow-up coronary angiography and LVG at 4 weeks after the treatment, the animals were killed and their hearts were harvested. Samples were collected from the infarcted, border, and remote areas in the treated hearts. The tissue was prepared as previously described.19

Flow Cytometric Analysis of the Cultured Cells

The characteristics of the cultured MSCs were confirmed by flow cytometric analyses using purified or directly conjugated antibodies against SWC3a (BD PharMingen, San Diego, CA, USA), CD44 (BD PharMingen), SLA-class I (BD PharMingen), CD29 (BD PharMingen), CD44H (FITC-labeled, BD PharMingen), SLA-DQ (BD PharMingen), CD31 (BD PharMingen), and CD90 (PE-labeled; BD PharMingen). The cells were detached and stained for 30 min at 4°C with primary antibodies and immunofluorescent secondary antibodies. After washing, the cells were analyzed on a Cytomics FC500 analyzer (Beckman Coulter, Fullerton, CA, USA).

Immunostaining and Morphometric Analysis

Immunohistochemistry and the morphometric analysis were performed as previously described.19 The primary antibodies were anti-von Willebrand factor (vWF) for endothelial cells at a 1:400 dilution (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-α-smooth muscle actin (SMA) for small arteries at a 1:300 dilution (Sigma, Saint Louis, MO, USA) and phospho-signal transducer and activator of transcription 3 (phospho-STAT3, Cell Signaling Technology, MA, USA) at a 1:50 dilution.

Images of the Mallory-stained sections were recorded, digitized, and analyzed with Win ROOF image analysis software (Mitani, Fukui, Japan) to measure the fibrotic area. The measurements were taken in 10 sections for each animal. To evaluate neovascularization, the numbers of α-SMA- and vWF-positive vessels in the border areas were manually counted in a randomly selected high-power field ( × 400). The measurements were taken in 10 areas for each animal. The investigator performing this analysis was blinded to the treatment.

Immunofluorescence

Before staining, frozen sections were fixed for 10 min in chloroform at room temperature. The sections were incubated with the first antibody: either anti-vWF antibody (Santa Cruz Biotechnology) at a 1:300 dilution for 60 min at room temperature, or anti-troponin T antibody (NeoMarkers, Fremont, CA, USA) at a 1:100 dilution for 60 min at room temperature, or anti-VEGF antibody (Santa Cruz Biotechnology) at a 1:800 dilution for 24 h at 4°C. After incubation with the first antibody, the sections were incubated with the second antibody: Alexa 488-labeled IgG (green color; Molecular Probes, Eugene, OR, USA; 1:400 dilution) for 1 h. Nuclear staining was performed by DAPI.

Quantitative mRNA Expression Analysis for Basic Fibroblast Growth Factor (bFGF) and VEGF

An RNA extraction, reverse transcription-polymerase chain reaction (RT-PCR) analysis and quantification of the PCR products were performed as previously described.19 The PCR primers were as follows: bFGF, forward 5′-TCAAAGGAGTGTGTGCGAAC-3′ and reverse 5′-CAGGGCCACATACCAACTG-3′; VEGF, forward 5′-ATGCGGATCAAACCTCACC-3′ and reverse 5′-ATCTGGTTCCCGAAACGCTG-3′; GAPDH, forward 5′-TCACCATCTTCCAGGAGCGA-3′ and reverse 5′-CACAATGCCGAAGTGGTCGT-3′. PCR reactions were carried out as follows: for bFGF, 35 cycles of denaturation for 60 s at 94°C, annealing for 45 s at 59°C, and extension for 45 s at 72°C; for VEGF, 27 cycles of denaturation for 30 s at 94°C, annealing for 1 min at 60°C, and extension for 2 min at 68°C, followed by a final extension for 7 min at 68°C; for GAPDH, 30 cycles of denaturation for 1 min at 94°C, annealing for 90 s at 55°C, and extension for 2 min at 72°C.

The sizes and quantities of the PCR products were detected by an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA) according to the manufacturer's instructions. In brief, 9 ml of gel-dye mix was fed into each of 16 wells and then forced into a micro channel network. Each of two buffer wells was also filled with 9 ml of the gel-dye mix, and sample and ladder wells were filled with 5 ml of marker mix followed by 1 ml of DNA sizing ladder and sample, respectively. The prepared microchip was vortexed and placed into the bioanalyzer for analysis. All experiments were performed using Agilent biosizing software (Version A.02.12).

Immunoblotting

The immunoblotting was performed by a method previously described,20 with minor modifications. A total of 20 μg of protein was separated by SDS-PAGE. The protein was transferred from gel to a clear blot P membrane sheet (Atto, Tokyo, Japan) by horizontal electrophoresis at 108 mA for 90 min at room temperature. Immunostaining was carried out with anti-VEGF (Santa Cruz Biotechnology) and bFGF (Millipore, CA, USA) antibodies.

In Vitro Angiogenesis Assay

Tubule formation experiments were conducted using an Angiogenesis kit (Kurabo, Osaka, Japan), according to the manufacturer's instructions. Briefly, human umbilical vein endothelial cells (HUVECs) and human fibroblasts were admixed and seeded into 24-well plates. Human MSCs (5 × 104 cells/well) were applied and co-cultured for the experiments. After 10 days of culture, the HUVECs were fixed with 70% ethanol at 4°C and immunostained with an anti-human CD31 antibody using BCIP/NBT as a substrate for the secondary antibody. Five fields per well were selected for digital photography under a microscope (Olympus, Tokyo, Japan), and the areas of the tubule-like structures were measured quantitatively using angiogenesis image analyzer software (Kurabo).

Statistical Analysis

All values are expressed as means±s.e.m. The means were compared by the unpaired Student's t-test. The paired t-test was used for comparison of temporal changes in the ejection fraction (EF). Comparisons among the three groups were conducted by one-way analysis of variance followed by Scheffé's multiple comparison test. A P-value of <0.05 was considered statistically significant.

RESULTS

Phenotype of Swine MSCs

Bone marrow was aspirated from the swine femurs as the MI model was being created (Figure 1a). The mononuclear cell layer from the BM was cultured, and the MSCs adhering to the culture plastic after 24–48 h were isolated. The cultured swine MSCs were phenotypically similar to human MSCs. They had a fibroblastic spindle-like shape typical of mesenchymal cells and were positive for CD44, SLA class I, CD29, and CD90, and negative for CD31 (Figure 2).

Figure 2
figure 2

Phenotype of swine MSCs. (a) Phase contrast microscopic images of MSCs (A, low magnification; B, high magnification). (b) Flow cytometric analysis of swine MSCs. MSCs are positive for SWC3a, CD44, SLA class I, CD29, CD44H, and CD90, and negative for SLA-DQ, CD31, and SLA-DR.

Full size image

Cardiac Function in Chronic MI after the MSC Implantation

We measured the serum CK levels to assess acute myocardial damage after the sterilized beads were implanted in the swine LADs. The levels of CK peaked at 1 day after the infarction, and the peak levels were similar in the two groups (3338.6±538.5 mg/dl in the control group, 3447.7±561.7 mg/dl in the MSC group).

At 4 weeks after the MI, the swine were treated with the MSCs (1.02±0.34 × 107 cells, n=8) or vehicle (n=5) via coronary vein infusion (Figure 1c). Coronary venography confirmed that the vessels were free of complications such as vein rupture or embolization during the procedure. To determine cardiac function, LVG was performed immediately before and at 4 weeks after the treatment. The EF assessed by the LVG did not significantly differ between the two groups at baseline. The EF significantly decreased in the control group over time (P<0.05), whereas the cardiac function was preserved in the MSC group (Figure 3a). The difference in the EF (δ-EF) between the baseline and follow-up was significantly smaller in the controls than in the MSC animals (P<0.01, Figure 3b). There were no statistically significant differences between the two groups in the EF or end-diastolic volume at follow-up (Figure 3c).

Figure 3
figure 3

Cardiac function assessed by left ventriculography. (a) Time course changes in ejection fraction (EF) immediately before and at 4 weeks after treatment. Left, control group (n=5); right, MSC infusion group (n=8). (b) Differences between pre-treatment EF and EF after 4 weeks (δ-EF). (c) Endodiastolic volumes at 4 weeks after treatment. No significant differences between the two groups were observed. Data are means±s.e.m.*P<0.05; ‡P<0.01.

Full size image

These findings from the LVG studies indicate that retrograde delivery of the MSCs via the coronary vein suppresses the progression of systolic LV dysfunction in chronic MI.

Effects of the Cell Infusion on Cardiac Fibrosis and Neovascularization

We conducted histopathological studies to address the effects of the MSC administration on cardiac fibrosis and neovascularization. Azan-Mallory-stained sections of hearts from the MSC-treated swine showed somewhat less fibrosis than those from the controls, but the difference was not significant in quantitative analysis (Figure 4a). The numbers of vWF- and α-SMA-positive vessels were assessed in the border areas (Figure 4b). The number of vWF-positive vessels was significantly higher in the MSC-treated swine hearts than in the control group (P<0.0005). Likewise, α-SMA-positive vessel counts were also significantly higher in the MSC group than in the control group (P<0.0005).

Figure 4
figure 4

Cardiac fibrosis and neovascularization in hearts after cell infusion. (a) Left, blue staining indicates fibrosis in the Azan-Mallory-stained sections. Right, the morphometric analysis shows no significant differences in the fibrotic area between the two groups. (b) Left, neovessels are evaluated in the von Willebrand factor (vWF)- and the α-smooth muscle actin (α-SMA)-stained sections. Right, the numbers of vWF- and α-SMA-positive vessels in the border areas. Data are means±s.e.m. †P<0.0005.

Full size image

The effects of the MSC infusion were limited in terms of modification of cardiac fibrosis in the chronic post-infarcted myocardium, while the cell administration significantly promoted angio/arteriogenesis in the border zones. Thus, the favorable effects of the MSCs on cardiac function appeared to derive, at least in part, from enhanced neovascularization.

In a previous report from another group,21 freshly isolated BM cell implantation induced a calcified lesion formation in the infarcted myocardium of rat. Noting this, we carried von Kossa staining to detect abnormal calcium deposition in the frozen sections. This staining, however, revealed no calcified lesions (data not shown).

Engraftment and Differentiation of the Implanted MSCs

We assessed MSC engraftment and differentiation by immunofluorescence. MSCs marked by red fluorescence dye were sparsely observed in an interstitial area of the border area at 4 weeks after the infusion, but we could not detect any cells differentiated to cardiomyocytes (Figure 5a and b). A few MSCs expressed vWF and were incorporated into the vessel structure (Figure 5c).

Figure 5
figure 5

MSC engraftment and differentiation. (a) Engraftment of MSCs implanted in the border area. Red fluorescence indicates MSCs. Blue fluorescence indicates DAPI nuclear staining. (b) Lack of cardiomyocyte differentiation of the engrafted MSCs. Implanted Red fluorescence indicates MSCs. Green fluorescence indicates myocardium (TnT=troponin T). Blue fluorescence indicates DAPI nuclear staining. (c) Endothelial cell differentiation of the engrafted MSCs. Red fluorescence indicates MSCs. Green fluorescence indicates endothelial cells (vWF). Yellow (red and green) fluorescence indicates MSC differentiation. Blue fluorescence indicates DAPI nuclear staining. Bar=100 μm.

Full size image

The effect of the engrafted MSCs in promoting neovascularization seemed to be conferred partly through endothelial differentiation. The extent of the differentiation, however, was far too low to account for the significant increases in the vessel numbers.

Enhanced Expression of Angiogenic Factors after the MSC Implantation

Noting the reduced extent of cell incorporation to neovessels, we examined whether factors related to angiogenesis contributed to the increased neovascularization in the treated myocardium.

We determined the gene expression levels of bFGF and VEGF relative to GAPDH by quantitative RT-PCR analysis. The bFGF and VEGF levels were significantly higher in the MSC-treated hearts than in the control hearts (P<0.05, respectively; Figure 6a). Immunoblotting analysis confirmed the increased protein expressions of both bFGF and VEGF in the MSC group compared with the controls (Figure 6b). Of note, the MSC infusion enhanced the transcriptional levels not only in the border and infarcted areas, but also in the remote areas with little to no MSC engraftment (Figure 6a). In response to the MSC infusion, resident cardiac cells in the remote areas appeared to produce these angiogenic cytokines.

Figure 6
figure 6

Expressions of bFGF and VEGF in the hearts after cell infusion. (a) Representative images of electrophoretic gels after PCR amplification of FGF and VEGF mRNA in the border areas (upper). Quantitative analysis of bFGF mRNA expression relative to GAPDH mRNA in each area of the heart (middle). Quantitative analysis of VEGF mRNA expression relative to GAPDH mRNA in each area of the heart (bottom). Data are means±s.e.m., n=3 in each area. *P<0.05. **P<0.05 vs infarcted area in the MSC group. B, border area. I, infarcted area. R, remote area. (b) Immunoblotting analysis for bFGF and VEGF protein expressions in the border area. bFGF, basic fibroblast growth factor. VEGF, vascular endothelial growth factor.

Full size image

Immunohistochemical staining revealed that the MSCs engrafted in the border zones produced VEGF (Figure 7a). MSCs have been shown to secrete IL-6 and leukemia inhibitory factor.7, 22 The transcriptional factor STAT3 is activated by IL-6 family members, and its activation promotes capillary growth in adult heart via acceleration of VEGF transcription.23 We found that the area surrounded by the engrafted MSCs contained clusters of cells positive for phosphorylated +-STAT3 in sequential sections of the swine infarcted myocardium (Figure 7b).

Figure 7
figure 7

Expressions of VEGF and pro-angiogenic signal in the MSC-treated hearts. (a) VEGF production from the engrafted MSCs. Red fluorescence indicates MSCs. Green fluorescence indicates VEGF. Blue fluorescence indicates DAPI nuclear staining. Merged image (yellow fluorescence) shows VEGF expression in the MSCs. Bar=20 μm. (b) STAT3 activation in the resident cardiac cells. Left, MSC engraftment (red fluorescence and white arrow). Blue fluorescence indicates DAPI nuclear staining. Right, clusters of cardiac cells with activated STAT3. Brown indicates positive staining for phospho-STAT3. Bar=50 μm. STAT3, signal transducer and activator of transcription 3.

Full size image

Thus, VEGF might be produced not only the MSCs, but also by resident cardiac cells with STAT3 activated by the MSC-secreted factors.

In Vitro Angiogenesis Assay

We examined the culture system by an angiogenesis assay. HUVECs were immunostained with an anti-CD31 antibody, and the colored areas were quantified as capillary growth. Co-culture with human MSCs (5 × 104 cells) prominently induced the capillary network formation (Figure 8). The capillary area was significantly larger in the MSCs than in the control medium or the control medium supplemented with a high concentration (10 ng/ml) of VEGF (P<0.0001, respectively).

Figure 8
figure 8

In vitro angiogenesis assay. Comparison of %CD31 areas among the growth medium (100%), the growth medium supplemented with VEGF 10 ng/ml, and the MSC co-culture. Data are means±s.e.m., n=3–4. #P<0.0001 vs MSC.

Full size image

DISCUSSION

Various reports have demonstrated a potential relationship between cardiac angiogenesis and cardiac function.24, 25, 26 Disruption of the balance between myocyte growth and angiogenesis leads to contractile dysfunction and heart failure. VEGF plays a critical role in this process.

In an earlier study by our group, BMNCs implanted via the coronary vein improved cardiac function in the acute and subacute phase of MI.15 MSCs have been shown to be a robust source of angiogenic factors compared with hematopoietic cells in BM.7, 27 The low-invasive local delivery of MSCs preserved cardiac function in the swine chronic MI model in our present study. In comparisons against our controls, swine hearts implanted with MSCs exhibited increased expression of VEGF and bFGF and larger numbers of vessels in vWF and α-SMA staining. The enhanced neovascularization induced by the increased VEGF may help to prevent the progression of cardiac dysfunction in the chronic MI model.

Immunohistochemical staining revealed that the elevated VEGF levels were at least partly due to VEGF release from the engrafted MSCs. MSCs significantly promoted capillary network formation in our endothelial tube formation assay in vitro. The results of this in vitro study may support the hypothesis that factors secreted by the MSCs in vivo are responsible for enhanced neovascularization and preserved contractile function. Several previous studies support the paracrine hypothesis for MSC-mediated angiogenesis and cardioprotection.7, 11, 27, 28 Moreover, our RT-PCR analysis showed significant increases in the VEGF levels in the remote areas wholly devoid of MSCs. The MSC-treated heart contained clusters of resident cells positive for phospho-STAT3, the regulator of VEGF production. Human MSCs expressed higher mRNA levels of both VEGF and IL-6 family members than the human BM CD133+ progenitors, and secreted those factors in vitro.7 On this basis, we speculate that MSC-secreted factors, such as VEGF and IL-6, promote neovascularization both directly through endothelial proliferation and indirectly through activation of STAT3 in the resident cells.

Our previous study demonstrated that the safety and feasibility of the cell delivery via the coronary vein in a preclinical trial with the pig models.15 For the reasons explained below, we also selected coronary vein infusion as the route for MSC delivery into the myocardium in the present study. We did not, however, compare it with the other delivery methods (eg, intracoronary infusion and intramyocardial injection) in our hands. Intracoronary infusion of MSCs may cause distal embolization and consequent no-reflow phenomena. Freyman et al29 reported that intracoronary MSC infusion impaired coronary flow distal to the infusion site in half of their treated pigs. Similar to this report, the intracoronary administration showed elevated cardiac enzymes and micro infarctions in dogs with normal coronary arteries.30 Hoshino et al18 compared three methods for the injection of gelatin hydrogel microspheres, namely, antegrade injection, retrograde injection, and direct injection. The greatest number of microspheres was retained in the target area after the antegrade injection, but this method induced microinfarctions and failed to show any angiogenic effect. An increased coronary flow reserve was observed in the animals receiving retrograde injection and direct injection therapy, thus confirming the superior efficacy of these methods compared with antegrade injection. Importantly, retrograde injection was much less invasive than direct intramyocardial injection.

Coronary vein infusion of MSCs is a promising therapeutic strategy for the treatment and prevention of heart failure. Yet, the cell implantation failed to reduce cardiac fibrosis or achieve reverse remodeling in this study. Further studies in large animals to explore the methods to enhance or modify MSC function, such as gene transfer11 and pharmacological preconditioning,31 will be necessary.

Taken together, our results show that MSC implantation via the coronary vein is feasible and prevents cardiac contractile dysfunction in ischemic cardiomyopathy of pigs. The mechanism behind this benefit appears to be partly attributable to increased local expression of angiogenic factors such as VEGF and enhanced neovascularization in response to the cell implantation.