Intramyocardial injection of bone marrow mononuclear cells (MNCs) with hyaluronan (HA) hydrogel is beneficial to the ischemic heart in a rat model of myocardial infarction (MI). However, the therapeutic efficacy and safety must be addressed in large animals before moving onto a clinical trial. Therefore, the effect of combined treatment on MI was investigated in pigs. Coronary artery ligation was performed in minipigs to induce MI followed by an intramyocardial injection of normal saline (n = 7), HA (n = 7), normal saline with 1 × 108 freshly isolated MNCs (n = 8), or HA with 1 × 108 MNCs (HA-MNC; n = 7), with a sham-operated group serving as a control (n = 7). The response of each experimental group was estimated by echocardiography, ventricular catheterization, and histological analysis. Although injection of HA or MNCs slightly elevated left ventricular ejection fraction, the combined HA-MNC injection showed a significant increase in left ventricular ejection fraction, contractility, infarct size, and neovascularization. Importantly, injection of MNCs with HA also promoted MNC retention and MNC differentiation into vascular lineage cells in pigs. Therefore, this study not only provides evidence but also raises the possibility of using a combined HA-MNC injection as a promising therapy for heart repair.
- myocardial infarction
- cell therapy
- large animal model
acute myocardial infarction (MI) has severe morbidity and high mortality rates. Although early revascularization by thrombolytic therapy and/or coronary intervention has shown cardiac benefits, the ischemic attack still causes permanent damage with disease progression, ultimately leading to heart failure (22, 26). To overcome this obstacle, cell therapy has emerged as a new approach to repair the injured heart. Among the many different types of cells used, bone marrow-derived cells are particularly attractive owing to their unique advantage of being derived from an autologous source, which avoids immunogenicity (2, 10, 20). Furthermore, bone marrow cells are easily isolated and are available for immediate use without long-term cultivation, making their use clinically practical.
Previous studies (12, 13) in animals have shown that the injection of bone marrow cells improves cardiac performance after MI through paracrine factor secretion such as growth factors and cytokines to prevent cardiomyocyte apoptosis and promote angiogenesis. Despite these advantages, in clinical trials, the therapeutic efficacy in the infarcted human heart remains unclear and controversial (1, 10, 11, 16, 19, 20, 24, 25). The main controversy is the low survival and retention rate of the injected cells (8). Therefore, a material-based cell delivery system has been designed to overcome these problems.
Hyaluronan (HA), also referred to as hyaluronic acid, is composed of nonsulfated glycosaminoglycan and is widely distributed in the extracellular matrix of tissues. During the embryonic stage, HA is temporarily produced in the heart, and without HA, the formation of cardiac vasculature becomes abnormal (3). In a previous study (4), we demonstrated that HA enhances the therapeutic efficacy of bone marrow mononuclear cells (MNCs), which, in turn, reduces cardiomyocyte apoptosis and inflammatory cell infiltration at the early stage of MI. Furthermore, intramyocardial injection of HA-MNCs restores heart function by increasing cell retention, promoting neovascularization, and decreasing scar formation after MI in rats (4). However, although this combined treatment results in elevated cardiac performance and a significant improvement in left ventricular (LV) function in the rat MI model, the cardiovascular characteristics of rats are quite different from those of humans. For example, the resting heart rate of rats is five times faster, its myocardium displays a very short action potential that lacks a plateau phase, and the predominant myosin isoform is different between rats and humans (7). On the other hand, large animals such as pigs have blood vessel diameters, anatomic structure, disease progression, and a remodeling pattern similar to humans and are therefore commonly used to study cardiovascular diseases (5, 17).
In the present study, we tested the hypothesis that an intramyocardial injection of autogenic bone marrow MNCs with HA increases MNC retention and therapeutic efficacy to attenuate pathological remodeling and enhance cardiac function in a pig model of MI, toward the goal for future translation in cardiac regenerative medicine.
MATERIALS AND METHODS
All animal protocols were approved by the Animal Care and Use Committee of National Cheng Kung University (Tainan, Taiwan). Sexually mature Lanyu minipigs (∼5 mo old, body weight: 24 ± 1.4 kg) were acquired from the National Taitung Animal Propagation Station. All animals were starved overnight before surgery. Anesthesia was used on all animals before intubation and baseline echocardiography. Animals received intramuscular injections of Zoletil (12.5 mg/kg, Virbac) and Rompun [xylazine (0.2 ml/kg), Bayer Healthcare] before further induction with isoflurane (1.5–2%, Baxter Healthcare). For the sustained administration of saline or anesthetic, a venous indwelling catheter was maintained in an ear vein during surgery. After surgery, an antibiotic (Ampolin, YSP) to prevent infection and an analgesic to alleviate pain (Keto, YSP) were given.
Induction of MI and HA and MNC administration.
A total of 36 Lanyu minipigs were used. All animals were randomly divided into the following groups: sham operated (open-chest surgery without coronary artery ligation, n = 7), MI (open-chest surgery with coronary artery ligation to induce MI, n = 7), MI with 1% HA (preparation in normal saline, Sigma-Aldrich, St. Louis, MO; 2 ml HA was injected into the heart 10 min after open-chest surgery with coronary artery ligation, n = 7), MI with MNCs (2 ml normal saline + 1 × 108 MNCs injected into the heart 10 min after open-chest surgery with coronary artery ligation, n = 8), and MI with 1% HA plus MNCs (2 ml HA + 1 × 108 MNCs injected into the heart 10 min after open-chest surgery with coronary artery ligation, n = 7). MI surgery was performed by permanent occlusion at the midleft anterior descending artery. All treatments were equally distributed to the infarct region (∼40 delivery sites, 50 μl/site, 26-gauge needle) 10 min after artery occlusion, and no animal died during the surgical period. HA powder (1,630 kDa, Sigma-Aldrich) was dissolved in PBS at 4°C for 24 h to form 1% HA hydrogel. All treatments were randomized and given to the animals by one person who was blinded to the treatment.
Bone marrow MNC isolation, purification, and DiI fluorescent dye labeling.
Pig bone marrow MNC isolation, purification, and DiI labeling were performed according to a previous study (15). Bone marrow was aspirated before coronary artery ligation, and three 5-ml syringes were attached on the bone cavity to aspirate the liquid bone marrow (total 15 ml).
Heart function was evaluated with echocardiography (Vivid 7 with a 3.5-MHz probe, GE Healthcare, Horten, Norway) before surgery and 1 day, 1 mo, and 2 mo after coronary artery ligation. Both LV end-diastolic diameter (LVEDD) and LV end-systolic diameter (LVESD) were measured in the long-axis view. LV ejection fraction (LVEF) in the long-axis view was calculated as follows: (LVEDV − LVESV)/LVEDV × 100%, where LVEDV is LV end-diastolic volume and LVESV is LV end-systolic volume. LVEDV was calculated as 7.0 × LVEDD3/(2.4 + LVEDD), and LVESV was calculated as 7.0 × LVESD3/(2.4 + LVESD).
Hemodynamics were measured by direct ventricular catheterization after echocardiography 2 mo after coronary artery ligation. In brief, 5.0-Fr. pressure-volume sensing catheters (Millar Instruments, Houston, TX) were used to collect the data. Before recording, 10 ml of blood were taken from the right jugular vein and collected in an EDTA blood collection tube to calibrate the electrical conductivity for volume conversion. LV pressure-volume loops were recorded after the signal became stable. To change preload, we created an incision in the upper abdomen and transiently compressed the inferior vena cava. Before the animal was euthanized, 10 ml of 25% saline were injected into the right jugular vein to confirm the conductance. All data were collected and recorded digitally and analyzed with PVAN3.2 software (Millar Instruments).
Infarct size, length, and area.
Two months after the infarction and all other measurements, animals were euthanized by an overdose with intravenous pentobarbital (85–150 mg/kg). The heart was harvested and then cut into five transverse slices from apex to base. Images of these slices were then captured by a digital camera (P500, Nikon), and the white area was considered as necrotic tissue. The ratio of the white necrotic area to the total area as well as the infarcted midline length to the total midline length were measured with ImageJ (National Institutes of Health), as previously described by Takagawa et al. (23).
Collagen deposition and wheat germ agglutinin staining.
Heart tissues were collected from regions remote from the infarct. All tissues were fixed with 4% paraformaldehyde at 4°C for at least 24 h before being embedded in paraffin. Masson's trichrome staining was performed to reveal the collagen content. Five randomly selected images from the remote region were used to calculate the collagen content with Axiovision software (AxioVision, Zeiss). Wheat germ agglutinin (WGA; WGA-Alexa 488, Molecular Probes, Eugene, OR) was used for cardiomyocyte membrane staining remote from the infarct. All sections were observed and photographed under a Zeiss microscope. Five hundred cardiomyocytes from five randomly selected images of the remote region were used to calculate the average diameter with Axiovision software (AxioVision, Zeiss).
Pig hearts were harvested and cut into five equal-width slices from the apex to the base (papillary muscle) using a geometric structure as shown in Fig. 3A. Nine pieces of tissue were then collected from slices 2–4 (each slice provided three pieces of tissue: one from the peri-infarct region, one from the infarct region, and one from the remote region). After the tissues were sectioned, deparaffined, and rehydrated, each section was boiled in sodium citrate buffer (pH 6, Sigma-Aldrich) for 10 min to retrieve the antigen. Sections were incubated with anti-smooth muscle 22α (1:200, Abcam, Cambridge, UK), anti-isolectin IB4 (1:100, Invitrogen), and anti-troponin I (1:200, DSHB, Iowa, IA) antibodies at 4°C overnight. After being washed three times, sections were incubated with secondary antibody, which was either Alexa fluor 488 or 568. Nuclei were stained with 4′,6-diamidino-2-phenylindole (Sigma-Aldrich), mounted, and observed under a fluorescence microscope (Zeiss). Capillary and arteriole densities were calculated by averaging five randomly selected images at the peri-infarct region (×200 magnification). MNC retention was calculated by counting the number of DiI-positive cells with nuclei in six randomly selected images at the peri-infarct and infarct regions (×200 magnification).
All in vivo measurements are presented as means ± SE. Statistical significance was calculated using one-way or two-way ANOVA. P values of <0.05 were considered statistically significant.
Combined HA-MNC injection elevates cardiac function, thickens the interventricular septum, and prevents ventricular remodeling after MI.
Heart function was measured by echocardiography to confirm that MI was successfully induced in minipigs. The cardiac performance in pigs from different experimental groups was evaluated at 1 and 2 mo post-MI. Pigs from the sham group showed no significant change in LVEF and interventricular septum thickness at diastole and systole between 1 and 2 mo post-MI. However, LVEF was decreased in 2.82% of the MI group, 1.47% of the HA-only group, and 0.55% of the MNC-only group, whereas it increased 3.3% in the HA-MNC treatment group between 1 and 2 mo post-MI (P < 0.001 vs. MI, MI + HA, and MI + MNC groups; Fig. 1). Systolic and diastolic interventricular septum thicknesses also increased in the combined HA-MNC treatment group (systole: 0.41 ± 0.01 cm and diastole: 0.37 ± 0.01 cm in the MI group, systole: 0.41 ± 0.01 cm and diastole: 0.37 ± 0.02 cm in the MI + HA group, systole: 0.40 ± 0.02 cm and diastole: 0.36 ± 0.01 cm in the MI + MNC group, and systole: 0.47 ± 0.03 cm and diastole: 0.42 ± 0.03 cm in the MI + HA-MNC group). These results reveal that HA-MNC treatment promotes cardiac function and prevents ventricular remodeling (P < 0.05 vs. the MI group, P < 0.05 vs. the MI + HA group, and P < 0.01 vs. the MI + MNC group; Fig. 1) at 2 mo post-MI.
Combined HA-MNC injection improves hemodynamics after MI.
To further assess whether injection of HA alone, MNC alone, or HA along with MNC affects heart function by improving cardiac contractility compared with the MI group, we used direct ventricular catheterization to measure cardiac hemodynamics 2 mo post-MI. The results showed that both systolic and diastolic function were elevated in the HA-MNC group (Fig. 2). Moreover, both LVEDP and LVEDV were reduced in the HA-MNC group compared with the MI group. In addition, the maximal chamber elasticity was lower in the MI and MNC-only groups but higher in the HA-only and HA-MNC groups, suggesting higher contractility of the injured heart after HA and HA-MNC injection (Fig. 2). These data indicate that HA-MNC injection prevents ventricular dilation post-MI.
Combined HA-MNC injection reduces collagen accumulation and pathological hypertrophy in areas remote from the infarct.
Morphometric analysis of hearts from the MI group showed severe scar formation at the infarct. In contrast, all treated groups showed a smaller scar size and scar length, and an increased scar thickness, especially in the HA/MNC group seems to have synergistic effect (scar size: P < 0.001 vs. the MI group, P < 0.05 vs. the MI + HA group, and P < 0.05 vs. the MI + MNC group; scar length: P < 0.001 vs. the MI group, P < 0.01 vs. the MI + HA group, and P < 0.01 vs. the MI + MNC group; and scar thickness: P < 0.001 vs. the MI group, P < 0.05 vs. the MI + HA group, and P < 0.05 vs. the MI + MNC group; Fig. 3). To further investigate the therapeutic effects of HA alone, MNC alone, and combined HA-MNC treatment, we evaluated the degree of pathological remodeling using Masson's trichrome staining and WGA membrane staining in the remote region. We found that the accumulation of interstitial fibrosis and diameters of cardiomyocytes in the remote region were smaller in all treated groups, especially in the HA-MNC group (Fig. 4). Together, these results suggest a global effect of HA-MNC injection on myocardial salvage and successful prevention of ventricular remodeling.
Combined HA-MNC injection increases implanted cell retention and neovascularization.
To test whether injection with HA increased MNC retention in pig hearts, we labeled MNCs with DiI before injection and quantified DiI-positive cell numbers at the border and infarct regions. To rule out false positive signals, only DiI and 4′,6-diamidino-2-phenylindole double-positive cells were counted. The results showed an approximately twofold higher DiI-positive MNC retention in the HA-MNC group compared with the MNC-alone group at 2 mo post-MI (Fig. 5).
To determine whether neovascularization was promoted because cell retention increased, we performed immunohistochemical staining to identify endothelial cells (isolectin) and smooth muscle cells (smooth muscle 22α) and then quantified capillary and arteriole densities at the peri-infarct region. The results revealed that both capillary and arteriole densities were significantly higher in the HA-MNC group compared with those in the MI-only, HA-only, and MNC-only treatment groups (Fig. 6), indicating that HA-MNC treatment enhances not only transplanted cell retention but also neovascularization.
In this study, we tested the therapeutic efficacy of HA and bone marrow MNCs for post-MI cardiac repair in a large animal model using minipigs. We found that injection of HA alone or MNCs alone modestly improved cardiac function and reduced infarct size after MI. However, when HA and MNCs were combined together for injection, the therapeutic effects were significantly enhanced, showing better recovery of systolic and diastolic function and better preservation of myocardial structure at 2 mo post-MI. We propose that the injected HA plays an important role in creating a favorable microenvironment, facilitating the injected autologous MNCs to repair the injury as well as recruiting cells to the infarct to participate in the formation of new blood vessels. These data suggest that combined HA-MNC therapy is an effective approach to repairing the injured heart. The enhancement in both structural and functional recovery implies that this treatment can be applied clinically for myocardial repair and regeneration.
Previous studies (18, 27) in rodents have demonstrated that bone marrow cell therapy, regardless of the use of the whole bone marrow MNCs or only one specified population, may have better cardiac function compared with MI. Some earlier clinical studies (1, 11, 20) have also demonstrated that autologous bone marrow MNC therapy is safe and improves post-MI cardiac function in patients, although another study (10) failed to show such benefits. Interestingly, recent clinical trials (10, 19, 24) in the United States have shown no significant improvement in LV function. Possible causes of these controversial results may include the rapid loss of delivered cells through pumping from the delivery sites and the low cell survival under hypoxic and inflammatory conditions. Therefore, to satisfy the various needs in cardiac cell therapy, tissue engineering has been proposed as a potential solution, such as the use of injectable biomaterials along with cell therapy (9, 14). Although injection along with a biomaterial may increase cell retention after intramyocardial injection, it also requires a long period of time for the material to degrade or integrate with the host, which can illicit an enhanced immune or inflammatory response in the heart (15). In contrast, HA, a highly biocompatible material already widely used in clinics, is degraded quite rapidly in the body, as it has been shown that the half-life of HA is from 12 h to 3 days in tissue (6). Thus, injection of HA may not stimulate an immune response in the heart (4).
One of the limitations of this study is that the pig heart is too big to quantify the total number of injected cells directly. Therefore, cell retention was quantified by collecting peri-infarct and infarct tissues from three cross-sections of the heart to estimate the total number. Interestingly, consistent with our previous finding in rats, we found that HA also facilitated the retention of MNCs at the injection site (4). Although the lifespan of HA may be short after injection, the retention rate of MNCs in the HA-MNC group showed an approximately twofold increase compared with that in the MNC-alone group, resulting in significant increases of angiogenesis and arteriogenesis in the HA-MNC group. Importantly, we also observed that most of the vessel cells at the peri-infarct area were DiI negative, suggesting an endogenous origin of the cells. These results indicate that the increase of vascularization by HA-MNC treatment may come from not only neovascularization but also preservation of the preexisting vessels.
Moreover, we found an elevated LVEF (5% increase after coronary artery ligation) and infarct thickness (30% thicker than the MI-only group), increased systolic and diastolic function (dP/dt: 1,616 mmHg/s and −dP/dt: −1,713 mmHg/s in the HA-MNC group vs. dP/dt: 1,159 mmHg/s and −dP/dt: −1,177 mmHg/s in the MI-only group), reduced scar size (11% less than the MI-only group), and retarded continuous remodeling specifically in the HA-MNC group but not in the HA- or MNC-only groups. The scar reduction effect has also been observed in other disease models using HA hydrogel as a treatment; for example, collagen formation is reduced in the injured spinal cord after HA treatment (21). This study therefore provides a potential solution for treating acute MI, particularly if delivered through a more clinically practical approach, such as by transendocardial delivery of the therapy. However, further work should be conducted before moving into clinical trials, such as investigations of the influence of HA hydrogel degradation, the optimization of timing of treatment (e.g., acute MI or chronic MI), the function of the neovessels, and the delivery route for hydrogel and cells (e.g., via ultrasound-guided injection or catheter); the potential risks (e.g., embolism) of the treatment should also be determined to develop an optimum treatment for MI.
In conclusion, our study demonstrates cardiac therapeutic effects of combined HA-MNC injection, such as promoting MNC cell retention and neovascularization in a large animal model of MI. Also, we show that the combined treatment improves cardiac performance by reducing scar size and collagen accumulation and by recruiting endogenous cells to the injury site for new blood vessel formation. Therefore, we postulate that HA-MNC therapy can be translated into a clinical treatment for MI.
This work was funded by Department of Health Executive Yuan Grants DOH102-TD-PB-111-TM019 and DOH102-TD-PB-111-TM020, National Science Council Grants 100-2314-B-006-046 and 101-2325-B-006-013, and the Academia Sinica Translational Medicine Program.
P. C. H. Hsieh received research support from Celgene Cellular Therapeutics and Meridigen Biotech.
Author contributions: C.-H.C., S.-S.W., and P.H. conception and design of research; C.-H.C. and M.-Y.C. performed experiments; C.-H.C., M.-Y.C., and P.H. analyzed data; C.-H.C. and P.H. interpreted results of experiments; C.-H.C. prepared figures; C.-H.C. drafted manuscript; C.-H.C., S.-S.W., and P.H. edited and revised manuscript; P.H. approved final version of manuscript.
The authors are grateful for support from the Taitung Animal Propagation Station and the National Cheng Kung University Animal Center for assistance with the pig experiments. The authors also thank Ching-Pin Chen, Chia-Wei Chang, and Ting-Yu Chu for help with pig surgery, Susan Lin for echocardiography, and Erika Wei and Prof. Iain C. Bruce for editing the manuscript.
- Copyright © 2014 the American Physiological Society