Bioenergetic and Functional Consequences of Cellular Therapy

Abstract

Rationale

The mechanism by which endogenous progenitor cells contribute to functional and beneficial effects in stem cell therapy remains unknown.

Objective

Utilizing a novel ³¹P magnetic resonance spectroscopy–2-dimensional chemical shift imaging method, this study examined the heterogeneity and bioenergetic consequences of postinfarction left ventricular (LV) remodeling and the mechanisms of endogenous progenitor cell contribution to the cellular therapy.

Methods and Results

Human embryonic stem cell–derived vascular cells (hESC-VCs) that stably express green fluorescent protein and firefly luciferase (GFP⁺/Luc⁺) were used for the transplantation. hESC-VCs may release various cytokines to promote angiogenesis, prosurvival, and antiapoptotic effects. Both in vitro and in vivo experiments demonstrated that hESC-VCs effectively inhibit myocyte apoptosis. In the mouse model, a fibrin patch–based cell delivery resulted in a significantly better cell engraftment rate that was accompanied by a better ejection fraction. In the swine model of ischemia-reperfusion, the patch-enhanced delivery of hESC-VCs resulted in alleviation of abnormalities including border zone myocardial perfusion, contractile dysfunction, and LV wall stress. These results were also accompanied by a pronounced recruitment of endogenous c-kit⁺ cells to the injury site. These improvements were directly associated with a remarkable improvement in myocardial energetics, as measured by a novel in vivo ³¹P magnetic resonance spectroscopy–2-dimensional chemical shift imaging technology.

Conclusions

The findings of this study demonstrate that a severely abnormal heterogeneity of myocardial bioenergetics in hearts with postinfarction LV remodeling can be alleviated by the hESC-VCs therapy. These findings suggest an important therapeutic target of peri-scar border zone and a promising therapeutic potential for using hESC-VCs together with the fibrin patch–based delivery system.

A transmural myocardial infarction, left ventricular (LV) remodeling with chamber dilation, and hypertrophy occur to compensate for the loss of contracting myocardium. Although stable LV remodeling may be achieved for a period of time, progressive myocardial dysfunction can develop and ultimately lead to overt congestive heart failure (CHF). The mechanisms that contribute to the transition from the compensated state to CHF remain unclear but may be related to progressive contractile dysfunction in the region of viable myocardium that surrounds the infarct (border zone, BZ).^1,2 We have recently demonstrated that BZ myocardium has a severely reduced energetic capacity, operates at a very low energetic state and is therefore more vulnerable to oxidative and other stresses.² We hypothesize that chronically elevated systolic wall stress in the BZ surrounding a myocardial infarct results in progressive abnormalities of oxidative phosphorylation and contractile dysfunction in this region and that in the absence of treatment, the energetic and contractile abnormalities of the BZ continue to expand radially to involve the entire left ventricle, thereby leading to global LV dysfunction and the development of CHF.^1,2

Currently, the available therapeutic options for heart failure due to transmural LV infarct are limited. Several exciting recent studies have shown that tissue specific stem cells may have the ability to generate cells of tissues from unrelated organs.^3–16 Whether this unexpected plasticity constitutes “transdifferentiation” or whether a small population of resident cardiac progenitor cells (CPC) persists in postnatal heart remains unknown. Although it is a consistent observation in the literature that cell transplantation improves LV contractile function,^{7,11–15,17,18} the cell engraftment rate a few weeks after the transplantation is usually very low, suggesting a trophic effect also contribute significantly to the functional beneficial effects. Namely, the cytokines released from the engrafted cells early after the transplantation benefit recipient hearts for a significantly longer term. We hypothesize that the beneficial effects of a fibrin patch–enhanced cell transplantation is associated with a significantly better engraftment and regeneration, as well as activation of recipient endogenous CPCs, which in turn increase neovascularization and myocardial protection. These effects initiate the improvement in BZ systolic contractile performance, perfusion, and consequently improved BZ wall stress, bioenergetics, and attenuated expansion of the BZ size. To evaluate this hypothesis we used a swine model of postinfarction LV remodeling with or without transplantation of human embryonic stem cell–derived vascular cells (hESC-VCs). The high energy phosphate content and PCr/ATP ratios in the BZ and RZ were measured using a novel method of ³¹P magnetic resonance spectroscopy in combination with 2-dimensional chemical shifting imaging (³¹P-2DCSI). To determine whether newly formed myocytes were derived from engrafted cells versus those originating in the host CPCs, a pulse-chase bromodeoxyuridine (BrdU) labeling technique¹⁹ was used.

Methods

An expanded Methods section is available in the Online Data Supplement.

All experiments were performed in accordance with the animal use guidelines of the University of Minnesota, and the experimental protocol was approved by the University of Minnesota Research Animal Resources Committee. The investigation conformed to the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health (NIH publication No. 85–23).

Generation of Vascular Cells From Human Embryonic Stem Cells

Vascular cells (VCs) were derived from the human embryonic stem cell (hESC) line H9 (Wicell, Madison, WI) and consisted of 2 cell types: endothelial cells (EC) and smooth muscle cells (SMC). Details of hESC culture and differentiation are described in our recent publications^20,21 and are briefly described in Online Data Supplement. The methods of cell characterization as well as in vitro cytoprotection experiments are also included in the Online Data Supplement.

Analysis of Cytokines Released From hESC-VCs

ECs or SMCs (0.4 million) were harvested from their normal growth media, washed 3 times using PBS, and then cultured in T75 flasks with 8 mL basal medium. The cell cultures were then subjected to either normoxic or hypoxic conditions for 48 hours. Supernatants were then harvested and analyzed by ELISA array (Human C series Angiogenesis Array, RayBiotech, Norcross, GA). Basal media exposed to identical conditions was used as blank controls to remove any background signals.

In Vivo Mouse Study

Details of the mouse model of myocardial infarction (MI) procedure were as reported previously.²² Briefly, 12-week-old, immunodeficient NOD/SCID/γc^–/– (NSG) mice (Jackson Laboratory, Bar Harbor, ME) were used in the current study. A left thoracotomy was performed to expose the heart and the left anterior descending artery was ligated with a 6.0 surgical silk suture. Fifteen minutes after ligation, the surviving animals were randomly assigned to the following groups to receive respective interventions: saline (MI, n=8); intramyocardial injection of hESC-VCs (MI+C, n=8; 0.125 million each of hESC-ECs and SMCs); fibrin patch without cells (MI+P, n=7), or hESC-VCs (0.125 million each of hESC-ECs and SMCs) seeded in fibrin patch (MI+P+C, n=9). Eight mice were used to experience identical open-chest surgery but without MI (SHAM, n=8). These 5 groups (40 mice) were followed up for 28 days. Additional 30 mice from groups of MI, MI+P, and MI+P+C were used and euthanized at day 2 (total 15, n=5 per group) and day 7 (total 15, n=5 per group) for histological evaluations. Another 6 sham mice were used for FACS analysis. Ten mice died during the surgery of myocardial infarction or during the follow-up. The total number of mice used for this study is 86. More detailed procedures for the animal model are included in the Online Data Supplement.

In Vivo Swine Study

The swine model for myocardial ischemia/reperfusion (I/R) and fibrin patch–based cell transplantation were as previously described in detail.^15,21,23 Briefly, female Yorkshire swine (≈15 kg at 55 days old) were anesthetized with inhaled 2% isoflurane, intubated, and ventilated with a respirator and supplemental oxygen. During surgery, the animal body temperature, ECG, arterial blood pressure, and oxygen saturation were monitored. A left thoracotomy was performed in the 4^th intercostals space by layers, and the root of the 1^st and 2^nd diagonal coronary arteries from the left anterior descending coronary artery were occluded for 60 minutes and followed by reperfusion. The animals were randomly assigned to the groups receiving the respective treatment as follows: sham control group (sham, n=6), animals were exposed to identical open-chest surgery but without I/R or any treatment; I/R+saline (MI, n=7); I/R+fibrin patch without cells (MI+P, n=6); I/R+hESC-VCs (hESC-ECs and hESC-SMCs, 2 million each) seeded in fibrin patch (MI+P+C, n=7). All these 4 groups of animals were followed up for 28 days (n=26). Additional 21 animals of MI, MI+P and MI+P+C groups were used and euthanized at day 3 (n=9, 3 each group), and day 7 (n=12, 4 each group) after surgery for histological evaluations. Five animals died of lethal arrhythmia during surgery or early in the follow-up. The total number of animals used in this study is 52. More detailed methods for surgery and resuscitation, cardiac MRI, and histology are also included in Online Data Supplement.

Fibrin Patch–Based Cell Delivery

The methods of fibrin patch–based cell delivery were as previously described in detail.^21,24 ES-derived vascular cells (endothelial and smooth muscle cells, 2 million each) were harvested freshly from cell culture immediately before transplantation, resuspended in 1 mL total volume of fibrinogen solution (25 mg/mL). This solution was then coinjected with catalytic thrombin solution (75 NIH units/mL) supplemented with 4 mL CaCl₂ and 2 mmol/L ε-aminocaproic acid onto the epicardium of the infarcted area. A plastic holder was placed on top of the heart to serve as a mold for fibrin patch. Usually within 1 minute, the mixture would be solidified, resulting in a circular fibrin patch with approximately 3 cm in diameter and 2 to 3 mm in thickness.

In Vivo Myocardial Energetic Mapping

In vivo myocardial energetic mapping was achieved using ³¹P MR spectroscopy with 2-dimensional chemical shift imaging (³¹P-2DCSI). MR measurements were performed on a 65 cm-bore 9.4-T magnet interfaced with a Vnmrj console (Varian, CA).²⁵ Radiofrequency transmission and magnetic resonance spectroscopy (MRS) signal detection were performed with a 28-mm-diameter, double-tuned (¹H and ³¹P) surface coil. The proton signal from water was used to adjust the position of the animal in the magnet so that the coil was at the magnetic isocenter, and to shim the magnetic field. ³¹P MR spectra were acquired with adiabatic half passage pulse to minimize flip angle variation due to B₁ inhomogeneity from the surface coil. ³¹P-2DCSI uses 10×8 phase encoding steps to cover a field of view of 5×4 cm², resulting in a spatial resolution of 0.5×0.5 cm². ³¹P MR acquisition was gated according to both the cardiac and respiratory cycle, with an average repetition time (TR) of 2.7 seconds. Each phase encoding step utilized 12 repetitive scans, resulting in a total data acquisition time of 43 minutes for ³¹P-2DCSI. The raw data were subject to Fourier series windowing reconstructions²⁶ on a home-built Matlab program, and the peaks of PCr and ATPγ from each voxel were integrated. The ratio of PCr to ATPγ (PCr/ATP) was then corrected based on 2 global spectra of TR=2.7 seconds and TR=12 seconds, to compensate for the partial saturation effects. A schematic view of ³¹P-2DCSI sequence and the results of phantom studies are shown in Online Figure I.

Statistics and Data Analysis

Statistical analyses were performed using Sigmastat version 3.5 (San Jose, CA). All data are expressed as a mean±SD. Data were analyzed with 1-way analysis of variance for repeated measures. A value of P<0.05 was considered significant. When a significant result was found, individual comparisons were made using the method of Scheffé.

Results

Derivation of Endothelial and Smooth Muscle Cells From hESC

Vascular differentiation of hESCs was initiated when cocultured with M2–10B4 stromal cells, giving rise to a population of CD34⁺CD31⁺ cells as previously described (Figure 1).^20,21,27 From these CD34⁺CD31⁺ cells, both endothelial cells (hESC-ECs) and smooth muscle cells (hESC-SMCs) were derived by culture under appropriate conditions.²⁰ The 2 cell populations assumed typical phenotypes and expressed cell-specific markers corresponding to ECs and SMCs, respectively, and formed tube-like structures in vitro. The cells with endothelial and smooth muscle phenotype that have not yet formed tube-like structures were used for transplantation into the infarcted hearts.

Figure 1. Characterization of human embryonic stem cell–derived vascular cells (hESC-VCs).

A, Flow cytometric analysis of hESC-derived endothelial cells (ECs) for CD31, CD34, and CD146. Red line shows the samples stained with indicated antibodies, with blue line showing the isotype control. B and D, Representative immunostaining images of hESC-ECs and smooth muscle cells (SMCs), respectively. The ECs expressed endothelial specific markers such as CD31, von Willebrand factor (vWF), and VE-cadherin, whereas the SMCs expressed smooth muscle actin (SMA), SM22, and Calponin. C, hESC-VCs formed capillary tube-like structures when cultured in matrigel.

Antiapoptotic Effects of hESC-VCs

To investigate the protective effects of hESC-VCs in vitro, cardiac myocytes (HL-1) were cultured in normal growth medium, serum-free medium, or hESC-VC-conditioned media for 24 hours. The conditioned media from hESC-VCs demonstrated protective effects on HL-1 cells in vitro as a result of reduced apoptosis (Online Figure II).

The antiapoptotic capabilities of hESC-VCs in vivo were tested on both mouse and swine models of myocardial infarction (Figure 2). At either day 2 (mouse) or day 3 (swine) after surgery, animals were euthanized and the infarct center of LV myocardium was subject to terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining, where apoptotic cells would be visualized as red nuclei. Compared with both MI and MI+P groups, the MI+P+C group of both mouse and swine models showed a significant reduction in apoptosis, indicating that hESC-VCs were able to protect the nearby myocardium from infarction.

Figure 2. hESC-VCs reduce cardiomyocyte apoptosis in vivo.

A through C, Representative TUNEL staining of the infarct center in mouse (A and B) and swine hearts (C), respectively. Panels a1–a4 (MI mouse) and b1– b4 (MI+P+C mouse) illustrate the individual channels and merged images of TUNEL staining with higher magnification. Panels C1–C3 illustrate individual channels and merged images of TUNEL staining from MI, MI+P and MI+P+C pig hearts, respectively. Animals were euthanized at day 2 (mouse) or day 3 (swine) after surgery. Colors shown are: cardiomyocytes (cTnI, green), apoptotic cells (TUNEL, red), nuclei (DAPI, blue), and merged TUNEL and DAPI-positive cells (apoptotic nuclei, pink). D, Quantification of TUNEL staining on both mouse and pig hearts in terms of percentage of TUNEL⁺ cell number relative to total cells number (TUNEL⁺%). TUNEL quantification was based on 5 fields (infarct center) per heart and 3 hearts per group. CMs indicates cardiomyocytes. The MI+P+C group showed a significant reduction in TUNEL⁺% in myocardium of both mouse and pig hearts. *P<0.05 versus MI; #P<0.05 versus MI+P.

To confirm that the hESC-derived VCs were capable of cytoprotection via paracrine effects, we investigated the cytokines released from hESC-VCs using an ELISA array; the results are summarized in Online Table I. A total of 19 cytokines were detected in the supernatants of the cell cultures, including those that promote angiogenesis, cell proliferation, inhibition of apoptosis, and the induction of cell migration.^28,29

Enhanced Cell Delivery Using Fibrin Patch

Using the myocardial infarction mouse model and hESC-VCs that were genetically modified to stably express green fluorescent protein (GFP) and firefly luciferase (GFP⁺/Luc⁺), we compared the efficacy of different cell delivery methods. The same dose of hESC-VCs was transplanted into infarcted mouse hearts using either a fibrin patch or intramyocardial injection. The temporal change of cell retention was examined using in vivo bioluminescent imaging, which detected luciferase expressing hESC-VCs. Interestingly, a significantly higher cell retention rate was observed in the fibrin patch–based delivery method (Figure 3B, P<0.05). The engraftment rate of fibrin patch–based hESC-VCs transplantation in swine hearts was evaluated based on immunostaining against GFP (Online Data Supplement). The fibrin patch–based delivery resulted in an engraftment rate of 2.6±0.6% at week 4 after transplantation, based on an initial cell number of 4 million.

Figure 3. Fibrin-patch delivery improves the efficacy of cellular therapy for myocardial repair.

A, Bioluminescent imaging time course of mice receiving 0.25 million hESC-VCs (Luc⁺, GFP⁺) either via fibrin patch or intramyocardial injection. B, Quantification of bioluminescent data from A. *P<0.05 versus injection method at the same time. C, Ejection fraction analysis at week 4 after surgery measured using echocardiography. *P<0.05 versus MI; #P<0.05 versus MI+P; ‡P<0.05 versus MI+C. D, The hESC-VC patch was extensively revascularized at week 1 after transplantation but negative for cardiac regeneration, evidenced by costaining of smooth muscle actin (SMA) and alpha-sarcomeric actin (αSA). Figure inset shows von Willebrand factor (vWF) and SMA double staining from the adjacent section, indicating the maturation of these vessels. E, Representative Immunostaining of human-specific CD31 (hCD31) from pig heart (MI+P+C) at 4 weeks after transplantation. Shown in the images are 2 vessels positive for hCD31 (red arrows, inset f). In the same field, there are another 4 vessels (black arrows) from the host that were not positive in hCD31 staining (serve as negative controls). Figure insets e (e1: hCD31 channel; e2: DAPI channel) illustrate transplanted endothelial cells (arrowheads) that have not yet integrated into vascular structure. Figure insets f (f1: hCD31 channel; f2: DAPI channel) illustrate one vessel generated from transplanted hCD31⁺ cells (red arrows).

At 4 weeks after surgery, cardiac function was assessed using echocardiography (Figure 3C). The hESC-VC-patch– transplanted group (MI+P+C) showed remarkably better contractile function (P<0.05 versus MI+C and MI+P), supporting the effectiveness of the fibrin patch cell delivery system verses direct injection.

Myocardial Protection From hESC-VC Patch Transplantation

Groups of mice with MI, MI+P, and MI+P+C, were studied and euthanized at 1 week after surgery to examine the mechanistic basis of myocardial repair from hESC-VC patch transplantation. Left anterior descending coronary artery ligation resulted in a severe transmural scar in the LV myocardium, which was characterized by the loss of myocytes. In response to hESC-VC-patch transplantation, nearby myocardium was protected, resulting in a significant increase in thickness of the LV wall (Online Figure II). The week 1 cardiomyocytes were negative for GFP or BrdU colocalization (Online Figure II), indicating a mechanism of myocardial protection which spared myocytes that would otherwise have undergone apoptosis. These findings, together with the anti-apoptosis and cytokine release results, demonstrate a cardiac protection effect from the hESC-VC patch transplantation. The VC patch area was also extensively revascularized as evidenced by vessel staining in Figure 3D. Similar results were also observed in swine hearts (Figure 7F2). This reestablished circulatory system inside the fibrin patch could provide nutrition to support the long-term engraftment of transplanted hESC-VCs, further strengthening the efficacy of cardiac protection. To examine the origin of these reestablished vessels, human-specific CD31 antibody was used to evaluate the histology of these sections. The results indicate that majority of the well-resolved vessel structures are human-specific CD31 negative, only a few of the vessels are human-specific CD31 positive, such as shown in the Figure 3E. These data suggest that fibrin patch–enhanced delivery of hESC-VCs induced a significant increase of neovascularization, the majority of which are secondary to angiogenesis and only small portion of the new vessels structure are result of vasculogenesis.

Figure 7. Contrast-enhanced first-pass MR perfusion imaging confirms improvements in infarct zone (IZ) and border zone (BZ) myocardial perfusion after hESC-VCs transplantation.

A and B, At week 4 after surgery, quantitative analysis of myocardial blood flow (MBF) was performed from contrast-enhanced first-pass MR perfusion imaging.⁹ Regions of interest (ROIs) corresponding to remote zone (RZ, green), border zone (BZ, yellow), infarct zone (IZ, red), and LV chamber (blue) were drawn from delayed enhancement MRI (A) and transferred to perfusion images to compute the temporal change of image intensity within corresponding regions (B1 through B2). Panels C1–C3 illustrate the time dependence of relative MRI signal intensities from regions of LV chamber (C1, blue), remote zone (C2, green) and infarct zone (C3, red), respectively. A Fermi model was used to generate the regional MBF using the LV curve (C1) as an input function. D, Summarized quantitative analysis of MBFs determined using an automatic program CIMRA. E1 through E2, Vascular staining of infarct zone myocardium from MI and MI+P+C hearts. E3, hESC-VC patch transplantation increased the density of arterioles of swine hearts. F1 and F2, hESC-VC patch transplantation induced extensive neovascularization of patch area that is not observed in hearts with cell-free patch transplantation (B1, MI+P). F3, MI+P+C hearts showed significantly higher vascular densities (number of vWF⁺ vessels per mm²) in both IZ and BZ as compared with MI group. *P<0.05 versus MI.

Mobilization of Endogenous CPCs

hESC-VC-patch transplantation also induced recruitment of c-kit⁺ progenitor cells. Representative c-kit staining of the hESC-VC-patch area is shown in Figure 4A, where ≈25 c-kit⁺ cells were recruited. Quantification of c-kit⁺ cells revealed that MI+P+C hearts had a significantly greater number of c-kit⁺ cells as compared with the MI group (Figure 4A2). Interestingly, the c-kit⁺ cell density in the hESC-VC-patch area was ≈40 fold higher than that of the MI hearts, further supporting the hypothesis of c-kit⁺ cell recruitment by the hESC-VCs. The proliferation activity of these c-kit⁺ cells was assessed using a 7-day BrdU incorporation protocol as well as Ki67 staining (Figure 4B and 4C). Quantification of BrdU⁺/c-kit⁺ and Ki67⁺/c-kit⁺ cells suggested that the hESC-VC-patch transplantation was also associated with elevated proliferation activity of endogenous c-kit⁺ cells.

Figure 4. hESC-VC patch transplantation activates c-kit⁺ endogenous progenitor cells.

A1, c-kit⁺ cells (green) being recruited into hESC-VC patch at week 1 after surgery. A2, Significantly higher numbers of c-kit⁺ cells were seen in the hESC-VC patch–transplanted hearts, with highest density of cells remaining in the fibrin patch. B1 through B3, BrdU incorporation (arrows) suggests increased cell cycling activities of the c-kit⁺ cells in MI+P+C hearts. C1 through C2, Costaining of c-kit (green) and Ki67 (red) further confirmed the increased c-kit⁺ cell proliferation in MI+P+C hearts. D, Representative flow cytometric analysis of a myocyte-depleted single-cell preparation from sham mouse heart; 13.8% of c-kit⁺ cell population is double-positive for CD45 (shown in blue). Quantification based on 6 sham mouse hearts yielded a percentage of 12.7±1.8 for CD45⁺/c-kit⁺ cells within c-kit⁺ cell population, which is consistent with the result from histology. All histological quantification was performed based on 5 slides per heart and 5 hearts per group. An average of >100 c-kit⁺ cells were evaluated for each heart.

Assessment of Endogenous CPCs

To evaluate the contribution of hematopoietic progenitors to the cardiac c-kit⁺ cell pool, additional in vivo and in vitro experiments were performed to examine the fraction of c-kit⁺ cells that were also CD 45⁺. Flow cytometry analysis of myocyte-depleted single-cell preparations based on 6 sham mouse hearts yielded a percentage of 12.7±1.8% for CD45⁺/c-kit⁺ cells within c-kit⁺ cell pool (Figure 4D), which is consistent with previous reports.³⁰ This result was further confirmed by histological analysis (costaining of CD45 and c-kit, data not shown) performed on 5 MI+P+C mouse hearts at 1 week after MI. Over 100 c-kit⁺ cells were evaluated for each heart. The fraction of CD45⁺/c-kit⁺ cells within the c-kit⁺ population was estimated to be 10±2%.

In Vivo Swine Study

hESC-VCs was then used, with a swine model of I/R (Figure 6A and 6B) and fibrin patch–based transplantation of hESCVCs that was also associated with substantial cell engraftment, extensive neovascularization, and recruitment of endogenous c-kit⁺ cells (Online Data Supplement).

Figure 6. Fibrin patch–based transplantation of hESC-VCs alleviates remodeling in swine hearts after damage from I/R.

A and B, Schematic and actual view, respectively, of the fibrin patch–based cell transplantation system. C, The level of cardiac hypertrophy is attenuated in the MI+P+C group, as measured by the ratio of LV weight over body weight (LV/BW). D through F, Hematoxylin and eosin staining of the border zone (BZ) myocardium showing a decrease in cell size and an increase in nuclei density in the treated group. G, Ejection fraction; H, thickening fraction; and I, end-systolic LV wall stress of the infarct region in groups at week 4 after transplantation. *P<0.05 versus MI hearts; #P<0.05 versus MI+P hearts.

Myocyte Turnover Level in Response to Cell Transplantation

Recently, there is strong evidence supporting the concept that there is a certain level of myocyte turnover rate in normal heart and that this rate is increased in response to myocardial injury.^31–34 To evaluate the cardiomyocyte cycling activity and its relationship with the endogenous cardiac progenitors, 8 swine (n=4 for each group of MI and MI+P+C) were subjected to 7-day BrdU incorporation protocol (starting from the day of myocardial infarction) and euthanized at either week 1 (time point BrdU is discontinued) or week 4 (21 days after the BrdU had been discontinued). The percentage of BrdU-labeled myocytes was quantified and summarized in Figure 5. The cell-treated hearts (MI+P+C) showed significantly higher percentage of BrdU⁺ myocytes than nontreated hearts (MI), indicating an elevated myocardial regeneration level in response to cell transplantation.

Figure 5. BrdU incorporation in cardiomyocytes.

A1 through A3, Mice, and B1 through B4, swine, were subject to 7-day BrdU incorporation protocol and euthanized at either week 1 (A1, B1, and B3) or week 4 (A2, A3, B2, and B4) after surgery. Costaining of myocyte (green) and BrdU (red) indicate significant increased activity of myocyte turnover rate of MI+P+C hearts at week 4 (arrows in A3 and C1 through C4), which is not observed in MI hearts.

hESC-VC Patch Improved LV Contractile Function and Wall Stress

At 4 weeks after surgery, the LV contractile performance and wall stress were examined. The LV chamber function, in terms of ejection fraction (Figure 6G), was significantly improved in the hESC-VC patch–transplanted swine (MI+P+C, P<0.05 versus MI and MI+P). In addition, significant improvement of the regional myocardial systolic thickening fraction in both peri-infarct border (BZ) and infarct zones (IZ), was also observed in hESC-VC-patch–transplanted hearts but not in the fibrin patch only group (Figure 6H). Similarly, the systolic LV wall stress was significantly reduced in hearts receiving the hESC-VC patch treatment (Figure 6I).

hESC-VC Patch Attenuated Myocardial Hypertrophy

At 4 weeks after surgery, animals were euthanized, and cardiac hypertrophy, in terms of left ventricle weight over body weight (LV/BW), was examined. The severe cardiac hypertrophy secondary to myocardial infarction was attenuated in hESC-VC patch–treated hearts (Figure 6C). The attenuation of cardiac hypertrophy was further supported by hematoxylin and eosin staining of BZ myocytes, which showed the smallest myocyte cross-sectional area in MI+P+C group (Figure 6D through 6F).

Myocardial Perfusion and Vascular Density

Utilizing gadolinium-based delayed enhancement and first-pass perfusion MRI,³⁵ regional myocardial perfusion was assessed noninvasively. The infarct was visualized using a delayed-enhancement MRI (Figure 7A), where regions of interest corresponding to IZ, BZ, and RZ were drawn and applied to first-pass perfusion images (Figure 7B) to measure the time-dependent change of regional signal intensities. Fermi model fitting of the regional signal intensity curves was used to quantify myocardial blood flows (Figure 7C), which are summarized in Figure 7D. The myocardial blood flow rate was significantly lower in the IZ and BZ of MI hearts. In response to the hESC-VC patch transplantation, the myocardial blood flow significantly improved in both IZ and BZ (P<0.05 versus MI and MI+P). The structural basis of the observed improvement of myocardial perfusion was examined by immunostaining for arteriolar and vascular density, respectively (Figure 7E and 7F and Online Data Supplement). Consistent with the perfusion data, I/R severely damaged the IZ vessel structure, which was partially ameliorated in hESC-VC patch–treated hearts (Figure 7E and 7F).

hESC-VC Patch Improved BZ Myocardial Energetics

In vivo mapping of myocardial energetics was achieved using a novel method of ³¹P MRS in combination with 2-dimensional chemical shift imaging (2DCSI). Superior spatial localization of 2DCSI sequences was demonstrated in vitro using phosphate phantoms (Online Figure I). A typical in vivo myocardial energetic map of an infarcted swine heart is shown in Figure 8A, where the ³¹P spectra corresponding to different regions of the IZ, BZ, and RZ are color-coded and highlighted in Figure 8B. Prominent abnormalities of myocardial energetics were detected in both IZ (depletion of high-energy phosphates) and BZ (reduction of PCr/ATP). The depletion of high energy phosphates in IZ is consistent with the fact that the scar area consists of fibrotic tissue and no high-energy phosphates. The in vivo bioenergetic mapping was measured in the three groups of MI, MI+P+C, and sham (n=4 each). Myocardial infarction resulted in severe reduction of PCr/ATP ratio in the BZ myocardium. The abnormality was significantly alleviated in response to hESC-VC patch transplantation (Figure 8E).

Figure 8. Bioenergetic consequences of border zone myocardium secondary to hESC-VCs transplantation.

A and B, Using ³¹P magnetic resonance spectroscopy in combination with 2-dimensional chemical shift imaging (³¹P-2DCSI) technique, high-energy phosphates levels of phosphocreatine (PCr) and adenosine triphosphate (ATP) from different regions of remote zone (RZ, black), border zone (BZ, blue), and infarct zone (IZ, red) could be measured at the same time. Myocardial infarction resulted in regional abnormalities of myocardial energetics that is characterized by absence of PCr or ATP in IZ and significantly reduced PCr/ATP ratio in BZ. C, Representative spectrum from a phantom containing fresh LV blood, which demonstrates prominent 2,3-diphosphoglycerate (2,3-DPG) resonance peaks, with the ATP resonance peaks under the noise level. D, Representative spectrum from an in vivo heart with postinfarction LV remodeling with similar signal intensity of 2,3-DPG and prominent high-energy phosphates peaks. C and D demonstrate that within the s/n ratio of this NMR experimental setting, the ATP signal contribution from the erythrocytes to the phosphorous spectrum is negligible. E, hESC-VC patch transplantation significantly increased the PCr/ATP ratio in the BZ myocardium when compared with MI group, *P<0.05. F, Plotting the individual BZ wall stress levels against their corresponding PCr/ATP ratios establishes a strong correlation (P<0.01), suggesting a relationship between improvement of bioenergetic levels and alleviation of wall stress.

Discussion

We have developed a novel in vivo 2DCSI MRS method, which demonstrates a therapeutic target of energetically vulnerable BZ in hearts with postinfarction LV remodeling (Figure 8). The new findings of the present study demonstrate that the fibrin patch–based delivery of hESC-VCs results in the improvements of myocardial bioenergetics and perfusion (Figures 7 and 8) at BZ of hearts with postinfarction LV remodeling. These functional improvements are accompanied by significant increase of BZ vascular density, particularly the resistant vessels (Figure 7E3), and recruitment of endogenous CPCs for cardiac repair (Figures 4 and 5).

Myocyte Turnover Rate and Activity of Endogenous CPCs in Response to Cell Transplantation

Although the quantitative levels remain to be controversial, there is strong evidence to support the concept that heart is not a postmitotic organ.^31–34 These studies demonstrate that a certain myocyte turnover rate exists in normal adult heart, and this cell proliferation is increased in response to myocardial injury.^31–34 To examine the short term activation, mobilization, and differentiation of endogenous CPCs, additional mouse study was completed. The additional mouse study yielded a density of 0.8±0.3 c-kit⁺ cells/mm² for MI+P+C mice at day 2 after surgery, which is not significantly different from that of sham mice (0.6±0.2 c-kit⁺ cells/mm²). CPCs increased significantly at day 7 after acute myocardial infarction and increased further more at day 14.^22,36,37

To determine whether the newly formed cells were derived from host CPCs, the pulse-chase BrdU labeling technique was used. hESC-VC transplantation was accompanied by significant recruitment of c-kit⁺ cells (Figure 4). We previously reported that resident Sca-1⁺/CD31^– CPCs exist in the adult mouse heart.²² Similarly, it has been reported that the numbers of c-kit⁺ CPCs are increased in hypertrophied and cardiomyopathic hearts.^38–40 In the present study, cardiomyocyte cycling activity was identified by an increased number of endogenous c-kit⁺ CPCs³¹ that were committed to myogenesis. Recent reports suggest that this reparative mechanism of CPCs exists throughout life in normal hearts.^32–34 The data from the present study suggest that without interventions, this rate of regeneration is very modest and not sufficient to prevent severe structural remodeling and a decreased ejection fraction after infarct (Figures 2 and 6). In response to hESC-VC transplantation, the number of myocardial c-kit⁺ cells was significantly increased with a significant portion costaining positive for BrdU and Ki67 (Figure 4), suggesting that cardiomyocyte replacement occurs in hearts receiving this cell treatment.

The increased myocyte cycling activity in response to patch-enhanced cell transplantation was directly evidenced by BrdU incorporation of myocytes (Figure 5). The hESCVC–transplanted hearts have a significantly higher percentage of BrdU⁺ myocytes than MI hearts, suggesting an elevated myocardial regeneration level in response to cell therapy.

Cytokine-Associated Myocardial Protection and Neovascularization

Findings from both in vitro and in vivo experiments demonstrate that hESC-VCs release cytokines that can prevent myocyte apoptosis (Figure 2, Online Figure II, and Online Table I). These secreted factors probably contribute to the beneficial antiapoptotic effects (eg, FGF, IGF), neovascularization promotion (eg, VEGF, FGF; Figure 7E and 7F), and mobilization of endogenous CPCs (eg, IL, IGF, EGF; Figure 4). Interestingly, the secretion of these factors was significantly increased under hypoxic conditions and was only observed in hESC-EC but not hESC-SMC (Online Table I), suggesting that EC may play an important role from the endocrine perspective. A remarkable protection of myocytes was observed by evaluating the apoptosis frequency in both in vitro and in vivo experiments. These findings are consistent with previous studies using other progenitor cell types.^14,15,41 It is a rather consistent finding in the literature that (1) only a very small percentage of transplanted cells show long-term engraftment in recipient myocardium,^14,42,43 and (2) an even smaller fraction of the engrafted cells appear transdifferentiate into cardiomyocytes or vascular cells.^14,42,43 These findings have led to the concept that early postimplantation paracrine interactions between the transplanted cells and native cardiomyocytes and vascular cells (and possibly cardiac and vascular progenitor cells) are the basis of much of the benefit observed after cell transplantation initiated quite early after an ischemia-reperfusion (I/R) event.^44,45 Indeed, there are many reports that cell transplantation performed shortly after an I/R event is associated with decreased early apoptosis of injured cardiomyocytes. This sparing of native cardiomyocytes that would have otherwise died after the initial I/R insult presumably reduced infarct size and this accounted for the decreased subsequent LV remodeling and dysfunction observed in the treated animals.^14,15

Myocardial Perfusion

hESC-VCs transplantation was accompanied by significant improvements in myocardial perfusion in both the infarct and border zones in addition to a significant increase in vascular density (Figure 6), suggesting that vessels associated with hESC-VC transplantation and regeneration are functional. It is possible that because of increased BZ wall stress, flow reserve in this region is inadequate to support energy demands during periods of increased cardiac work. Hence, to support LV contraction at basal and high cardiac work states, an increased vascularity may be required in the BZ. Myocardial perfusion by MRI provides direct evidence that regional perfusion is significantly improved by cell transplantation (Figure 6) and is supported by the structural evidence of increased vascular density. The histological staining methods applied in Figure 3B of the present study only pick up the arterioles, which are the resistant vessels that control the autoregulation. These are the smallest muscular vessels (50≈150 μm) that regulate the myocardial perfusion. Each arteriole supports capillaries where the exchanges of oxygen and carbon substrates occur. We have previously reported that hypertrophied and remodeled LV is associated with the subendocardial ischemia during the increased cardiac work states.^46,47 Therefore, the increase of the density of the resistant vessels as illustrated in Figure 7E3 is the important structural basis of the improved BZ myocardial blood flow rate measured by cardiac MRI (Figure 7D). The improvement of myocardial flow probably contributes to the reduction of apoptosis and improvement in LV contractile performance that observed in the present study.

Contractile Functional and LV Wall Stress Improvements in Relation to hESC-VC–Induced Structural Changes

We examined whether hESC-VC transplantation improves the LV overall structural and functional characteristics. The transplantation resulted in significant improvements of LV chamber and regional contractile function and reduction of myocardial wall stress, which in turn was accompanied by a significant reduction in myocardial hypertrophy (Figure 6). Recently, Suzuki et al³¹ reported that cell transplantation into stable, chronically hibernating swine myocardium resulted in a functional improvement in the hibernating region, accom panied by myocardial regeneration resulting from activation of resident c-kit⁺ cells in the myocardium. In the present study, the structural improvements of increased vascular density and reduction in hypertrophy were accompanied by remarkable functional improvements of myocardial perfusion (Figure 7), regional systolic thickening fraction, LV chamber function (Figure 6), and bioenergetics (Figure 8).

Myocardial Energy Metabolism

Previously, it has been impossible to examine heterogeneity of myocardial bioenergetics between different regions of the in vivo heart. The novel, spatially localized, ³¹P-2DCSI MRS experiments in the present study demonstrate that bioenergetic changes in hearts with postinfarction LV remodeling are remarkably heterogonous (Figure 8). The BZ of hearts with postinfarction LV remodeling is exposed to overstretching that may result in repetitive ischemia.^47–49 The spatially localized ³¹P-2DCSI experiments demonstrate that the improvements in myocardial bioenergetics are most pronounced in BZ of hearts with postinfarction LV remodeling (Figure 8). Bioenergetic improvements are accompanied by structural evidence for improvement in vascular density (Figure 7E and 7F), reduction of hypertrophy (Figure 6), and a reduction of apoptosis in hearts with postinfarction LV remodeling (Figure 2). The structural improvements are accompanied by beneficial outcomes in perfusion (Figure 7), contractile function and LV wall stresses (Figure 6). The data from the present study suggests a beneficial feedback cycle that is triggered by the sparing of BZ myocytes from apoptosis and myocardial regeneration from endogenous CPCs in response to hESC-VC transplantation. When we plotted BZ wall stress against BZ PCr/ATP for each heart, a linear relationship was evident (Figure 8D), supporting the concept that cellular therapy induced a beneficial feedback cycle. The remarkably higher wall stress and worsened energetics in the BZ, which was alleviated by the cellular therapy (Figures 6 and 8), is in agreement with the concept that hESC-VC was effective therapeutic target for protecting the BZ myocardium of hearts with postinfarction LV remodeling.

In summary, the findings of the present study demonstrate that the fibrin patch–based enhanced delivery of hESC-VCs resulted in a significant improvement of LV chamber function, regional systolic thickening fraction, and a reduction of wall stress in hearts with postinfarction LV remodeling. These contractile performance improvements were accompanied by significant reductions in myocardial apoptosis, hypertrophy, and vascular rarefaction. These structural benefits were accompanied by functional improvements in both myocardial perfusion and bioenergetics. In addition, this fibrin patch–based enhanced delivery of hESC-VC resulted in a pronounced mobilization of endogenous cardiac progenitor cells into the myocardial injury site, suggesting a novel mechanism of hESC-VC–induced endogenous CPC mobilization that contributes significantly to the aforementioned beneficial effects.

Novelty and Significance.

What Is Known?

• Cellular therapy has been shown to promote myocardial repair after injury. However, the mechanisms underlying the beneficial effects remain unclear.

• Hearts with postinfarction left ventricular (LV) remodeling are associated with myocardial bioenergetic inefficiency. The severity of this bioenergetic abnormality is linearly related to the severity of LV dysfunction.

What New Information Does This Article Contribute?

• The fibrin patch method has shown significantly better efficacy for stem cell delivery in vivo.

• A novel in vivo magnetic resonance spectroscopy–2-dimensional chemical shift imaging (MRS-2DCSI) method was developed, which demonstrates a therapeutic target of energetically vulnerable infarct BZ in hearts with postinfarction LV remodeling.

• The fibrin patch–based delivery of human embryonic stem cell– derived vascular cells (hESC-VCs) results in improvements of myocardial bioenergetics and perfusion at BZ of hearts with postinfarction LV remodeling. This is accompanied by pronounced activation and mobilization of endogenous cardiac progenitor cells.

Both experimental and clinical studies have demonstrated of the benefits of cellular therapy for treatment of heart failure. However, the mechanisms underlying beneficial effects of cell therapy remain largely unknown. Using both rodent and swine models of MI, we examined the therapeutic efficacy of hESC-derived endothelial cells and smooth muscle cells. The cells were delivered using a novel fibrin patch method, which showed superior efficacy. The cell therapy attenuated negative LV remodeling and improved BZ myocardial contractile function. This was accompanied by an improvement in myocardial perfusion and an increase in the density of resistance vessels. These beneficial effects were associated with activation and mobilization of endogenous cardiac progenitor cells at BZ. Additionally, we developed a novel ³¹P MRS-2DCSI method and report for the first time 2D mapping of high-energy phosphates in the in vivo heart. These findings demonstrate a bioenergetic heterogeneity of hearts during postinfarction LV remodeling, with the most severe abnormalities in the BZ myocardium. These findings demonstrate that combined use of the fibrin patch with hESC-derived vascular cells provides an effective means to therapeutically target the BZ myocardium of hearts with severe postinfarction LV dysfunction.

Non-standard Abbreviations and Acronyms

BrdU

bromodeoxyuridine

BZ

border zone

CHF

congestive heart failure

CPC

cardiac progenitor cell

CSI

chemical shift imaging

EC

endothelial cell

GFP

green fluorescent protein

hESC

human embryonic stem cell

I/R

ischemia reperfusion

IZ

infarct zone

Luc

firefly luciferase

LV

left ventricle

MRS

magnetic resonance spectroscopy

ROI

region of interest

RZ

remote zone

SMC

smooth muscle cell

TR

repetition time

TUNEL

terminal deoxynucleotidyl transferase dUTP nick-end labeling

VC

vascular cell