The underlying mechanism(s) of improved left ventricular function (LV) due to mesenchymal stem cell (MSC) administration after myocardial infarction (MI) remains highly controversial. Myocardial regeneration and neovascularization, which leads to increased tissue perfusion, are proposed mechanisms. Here we demonstrate that delivery of MSCs 3 days after MI increased tissue perfusion in a manner that preceded improved LV function in a porcine model. MI was induced in pigs by 60-min occlusion of the left anterior descending coronary artery, followed by reperfusion. Pigs were assigned to receive intramyocardial injection of allogeneic MSCs (200 million, ∼15 injections) (n = 10), placebo (n = 6), or no intervention (n = 8). Resting myocardial blood flow (MBF) was serially assessed by first-pass perfusion magnetic resonance imaging (MRI) over an 8-wk period. Over the first week, resting MBF in the infarct area of MSC-treated pigs increased compared with placebo-injected and untreated animals [0.17 ± 0.03, 0.09 ± 0.01, and 0.08 ± 0.01, respectively, signal intensity ratio of MI to left ventricular blood pool (LVBP); P < 0.01 vs. placebo, P < 0.01 vs. nontreated]. In contrast, the signal intensity ratios of the three groups were indistinguishable at weeks 4 and 8. However, MSC-treated animals showed larger, more mature vessels and less apoptosis in the infarct zones and improved regional and global LV function at week 8. Together these findings suggest that an early increase in tissue perfusion precedes improvements in LV function and a reduction in apoptosis in MSC-treated hearts. Cardiac MRI-based measures of blood flow may be a useful tool to predict a successful myocardial regenerative process after MSC treatment.
- myocardial blood flow
- perfusion magnetic resonance image
- myocardial infarction
interest in stem cell-mediated myocardial repair has grown substantially, and there are rapidly accumulating clinical and preclinical data supporting this approach. While the administration of various preparations of adult stem cells holds great promise to improve failing myocardium (9, 31, 42), currently the mechanistic underpinnings of cardiac regenerative therapy are highly controversial.1
One of the general principles of successful cardiac regenerative therapy is that vascularization of the infarct segment of myocardium is a prerequisite, either as a primary or a facilatory mechanism. Neovascularization is implicated in amelioration of myocardial apoptosis on the one hand (18, 38) and participation in restoring myocardial stem cell niches on the other (5). In theory, successful myocardial regeneration and engraftment of transplanted cells is dependent on proper oxygen and blood supply (10, 27).
Mesenchymal stem cells (MSCs), which have substantial capacity to facilitate cardiac repair following acute myocardial infarction (MI), participate in vasculogenesis and angiogenesis by paracrine signaling, incorporation into newly formed vessels, and most likely transdifferentiation into vascular endothelium and smooth muscle (34, 38, 39, 43). A key issue that has not previously been addressed is the relationship between vasculogenesis and actual tissue perfusion.
Here we used cardiac MRI to serially measure tissue perfusion in pigs treated with allogeneic MSCs after MI. The aims of the study were 1) to assess tissue perfusion noninvasively and correlate its time course relative to restoration of cardiac function; 2) to histologically characterize myocardial vascularity after cardiac recovery; and 3) to evaluate the role of myocyte apoptosis. Together our findings suggest that early tissue perfusion may be an important precursor to myocardial recovery and may be of value in predicting a successful response to cellular transplantation.
Induction of Myocardial Infarction
All animal studies were approved by the Johns Hopkins University Institutional Animal Care and Use Committee and comply with the Guide for the Care and Use of Laboratory Animals (NIH Pub. No. 80-23, revised 1985). Female Yorkshire pigs (n = 24) were used for the study. Three animal groups were studied: pigs received unlabeled MSCs (n = 7), Feridex-labeled MSCs (n = 3), placebo (n = 6), or no injection (n = 8). Feridex-labeled and unlabeled MSCs were combined into one treatment group because of the evidence that Feridex labeling does not affect the biological activity of cells (13, 14).
MI was induced by occlusion of the mid-left anterior descending coronary artery (LAD) with an inflated coronary angioplasty balloon (35). After 60 min, occlusion of the vessel was terminated by deflating the balloon, removing it, and establishing reperfusion.
Two of the animals that were not injected died before completion of the study protocol at 10 days and 4 wk after MI of cardiac causes. One MSC-treated pig died from cardiac rupture during follow-up 8 wk after MI. For the time course, five MSC and five control animals were analyzed. For the consecutive MRI-perfusion analysis six nontreated, six placebo-treated, and nine MSC-treated animals were evaluated at week 1.
Results from a subset of these animals regarding infarct size and global and regional cardiac function were published previously (3, 4). Myocardial perfusion data, all data from the placebo group, and MRI-tagging data from the Feridex-labeled MSC animals have not been published previously.
Stem Cell Harvest and Isolation
Male swine MSCs were obtained, isolated, and expanded as previously described (29, 34). Briefly, bone marrow was obtained from the iliac crest, and aspirates were passed through a density gradient to eliminate undesired cell types. Five to seven days after plating, hematopoietic and other nonadherent cells were washed away during medium changes. After expanding and culturing of the remaining purified MSC population, all utilized cells were harvested once they reached 80–90% confluence at passage 3. MSCs were placed in a cryopreservation solution consisting of 10% DMSO, 5% porcine serum albumin, and 85% Plasmalyte. The cells were stored in cryobags and frozen in a control rate freezer at −180°C until the day of implantation. Trypan blue staining was used to ensure the viability of all thawed MSC lots. Before use in this study, MSC viability was verified to be >85%. To ensure that the grafts were allogeneic, MSC donor animals were of a different strain (Durok-Landrace) from recipients (Yorkshire).
Two days after MI, animals received either intramyocardial injections of allogeneic porcine MSCs, Plasmalyte, or no injection, to serve as treatment, placebo, or control group, respectively. All the transplantation procedures were performed by one experienced operator (A. W. Heldman). The Stiletto Endocardial Direct Injection Catheter System (Boston Scientific, Natick, MA) was utilized per the manufacturer's protocol. In brief, arterial access was obtained by cutdown to the internal carotid artery. A left ventricular (LV) cineangiogram was performed with a 7-Fr pigtail catheter advanced through the 9-Fr LV sheath. The diastolic LV endocardial borders were traced manually in two roughly orthogonal fluoroscopic projections (see Fig. 1, A, B, and E), and areas of segmental hypokinesis and akinesis were marked (see Fig. 1E). The pigtail catheter was then exchanged for a 7-Fr steering guide to complete the guide-in-guide steerable delivery system for directional control during injection targeting. The Stiletto injection catheter consists of a central hypotube connected to a spring-loaded needle tip, which is placed within the steering guide. Radiopaque marker bands at the distal end of the LV sheath and steering guide assist in positioning the injection catheter for therapy delivery into the target area. Potential injection sites in the infarct (akinetic segments) and border zone (hypokinetic segments) were visualized in both projections to confirm endocardial contact in the target myocardial region. After deployment of the 26-gauge injection needle (see Fig. 1, C and D) Plasmalyte or MSCs were delivered to the endocardium in 0.5-ml aliquots (see Fig. 1E). A total of 2 × 108 MSCs were delivered into 12–17 injection sites.
Global and regional function.
MRI images were acquired with a 1.5-T MR scanner (CV/i, GE Medical Systems) at four time points after injection: 2 days, and 1, 4, and 8 wk. Global LV function was assessed with a steady-state free precession pulse sequence (37). A total of 8–10 contiguous short-axis slices were prescribed to cover the entire LV, from base to apex. Image parameters were the following: repetition time (TR)/echo time (TE) = 4.2 and 1.9 ms; flip angle = 45°; 256 × 160 matrix; 8-mm slice thickness/no gap; 125 kHz; 28-cm field of view (FOV); and 1 number of slices achieved. Cine images were analyzed with a custom research software package (Cine Tool, GE Medical Systems, Waukesha, WI). Endo- and epicardial borders of the LV were each defined in the end-diastolic and end-systolic frame in contiguous slices, and standard LV function parameters were calculated to evaluate resting LV function (see Fig. 4D).
To assess regional cardiac function, tagging MRI images were acquired with an ECG-gated, segmented K-space, fast gradient recalled echo pulse sequence with spatial modulation of magnetization to generate a grid tag pattern as previously described (4). Briefly, images were obtained at the same location as the cine-MRI images, and image parameters were as follows: TR/TE = 6.7 and 3.2 ms; flip angle = 12°; 256 × 160 matrix; views per second: 4; 8-mm slice thickness/no gap; 31.25 kHz; 28-cm FOV; 1 NSA; and 6 pixels tagging space. Tagged images were quantitatively analyzed with a custom software package (Diagnosoft HARP, Diagnosoft, Palo Alto, CA). Regional strain magnitude was determined from the 24 radially displaced segments for each short-axis section covering the entire LV and averaged among slices for each region and each time point, generating a strain map for each point over the cardiac cycle. The peak systolic circumferential strain (Ecc) was determined from the strain map for each point.
Myocardial perfusion imaging and delayed enhancement.
First-pass perfusion imaging was performed continuously for ∼1.5 min at rest immediately after an intravenous bolus injection of Gd-DTPA (0.1 mmol/kg, 5 ml/s; Magnevist, Berlex, Wayne, NJ) with an ECG-gated interleaved saturation recovery gradient echoplanar imaging pulse sequence (EFGRET-ET) (37). An entire short-axis stack was acquired every 2–4 heartbeats. Imaging parameters were as follows: TR/TE = 7.2 and 1.8 ms; flip angle = 20°; 128 × 128 matrix; 8-mm slice thickness/no gap; bandwidth 125 kHz; 28-cm FOV; and 0.5–1 NSA.
After completion of first-pass image acquisition, a second bolus of Gd-DTPA (0.1 mmol/kg) was injected. Delayed enhancement (DE) images were acquired 15 min later with an ECG-gated, breath-hold, interleaved, inversion recovery, fast gradient echo pulse sequence. DE-MRI images were acquired at the same location as the short-axis cine-images. Imaging parameters were TR/TE/inversion time (TI) = 7.3, 3.3, and ≈200 ms; flip angle = 25°; 256 × 196–160 matrix; 8-mm slice thickness/no gap; 31.2 kHz; 28-cm FOV; and 2 NSA. Inversion recovery time was adjusted as need to null the normal myocardium (37).
For perfusion analysis region of interest signal intensity curves were generated with Cine Tool (GE Medical Systems) by manually segmenting the endocardial and epicardial borders of the images. The segmented myocardium enclosed by the endo- and epicardial borders was then subdivided into eight segments of equal circumferential extent along the myocardial centerline dividing the two contours. The segments were adjusted to start at the anterior junction of the left and right ventricles (Fig. 2). Signal intensity-time curves were obtained by averaging the signal intensity in each myocardial segment and for every one of the 50 acquired images per slice (Fig. 2).
The segments of myocardial infarction were verified by the hyperenhancement in the DE images for each slice (Fig. 2). The upslope of the signal intensity was calculated in the myocardium, and the perfusion parameters were corrected either by the upslope of contrast agent in the ventricular cavity or by areas of the remote noninfarct myocardium. In addition, we calculated the areas under the signal intensity curve as described by Klocke et al. (20) and normalized the values to the signal of the LV blood pool (LVBP).
The initial value of the tissue impulse residue represents the mean blood flow in milliliters per minute per gram. Although the upslope parameter and the area under the curve are considered semiquantitative measurements of myocardial perfusion, both perfusion parameters are demonstrated to correlate closely with microsphere flow ratios.
At 8 wk, samples of infarct and noninfarct myocardium were obtained and stored in Streck's tissue fixative (Streck Laboratories, Omaha, NE) for histological analysis.
Immunohistochemical studies were carried out on 4-μm-thick paraffin sections. For antigen retrieval, the sections were steamed with target retrieval solution (TRS; Dako, Carpenteria, CA) for 20 min before blocking with avidin and biotin blocking solutions (Dako) and 10% normal goat serum. The density and size of blood vessels were evaluated by staining with antibody against von Willebrand factor (vWF) (Biocare Medical, Walnut Creek, CA). After washing, the samples were incubated with biotinylated goat anti-mouse IgG (E0433, Dako) at 1:400, and streptavidin peroxidase (Vectastain, Vector Laboratories, Burlingame, CA) was applied. The peroxidase activity was developed 5 min with 0.066% diaminobenzidine (DAB) + 0.01% H2O2 + 2.5% NiSO4. The slides were counterstained with hematoxylin and coverslipped.
The number and size of blood vessels visible per high-power field (HPF) were determined from randomly chosen areas of the endomyocardial rim. Five areas from each slide were counted, starting at the endocardial surface of the endocardium and ending in the infarct scar tissue. Vessel size was determined as previously described by Schuster et al. (32). Briefly, small vessels were defined as one or two endothelial cells spanning the vessel circumference. Vessels with five or more cells spanning the endothelium were considered large. A medium-sized vessel spanned three or four endothelium cells. The vessel size was evaluated by two independent investigators (A. J. Boyle and K. H. Schuleri) blinded to the animal. Both results showed a high correlation (r = 0.985, P < 0.01) Vessel morphology was further investigated with Masson's trichrome staining.
In situ detection of DNA fragmentation (TUNEL staining).
To detect the DNA fragmentation in situ, fluorescein-based nick-end labeling was performed with Roche chemicals and protocol (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's description. Briefly, the nuclei were stripped of proteins by incubation with 20 μg/ml proteinase K for 10 min after deparaffinization. The cardiomyocytes were analyzed on a Zeiss Axiovert 200 microscope with 510-Meta confocal laser scanning module and were counted as TUNEL positive as previously described (30). Apoptotic nuclei of cardiomyocytes were identified by double immunofluorescence labeling with α-sarcomeric actin. Caspase-3 (R&D Systems Minneapolis, MN) staining to evaluate apoptosis was performed as described above, and the slide analysis was performed in the same way as the TUNEL staining. All histological apoptosis stains were evaluated by blinded readers (TUNEL by K. H. Schuleri and caspase-3 by M. R. Gutman).
All analyses were blinded to the randomization. All values are expressed as means ± SE. Differences between groups were compared with t-tests. Differences in the perfusion parameters were compared within the groups with repeated-measures analysis of variance (ANOVA) and between the groups with two-way ANOVA with interaction term. A P level <0.05 was considered statistically significant.
Myocardial Blood Flow
MRI imaging was used to monitor regional tissue perfusion in infarct myocardium. Perfusion in the infarct zone was similar between control animals and MSC-treated animals 3 days after injection (Fig. 3). Over the first week, resting MBF in the infarct area of MSC-treated pigs increased compared with untreated animals (Fig. 3). Untreated animals showed a slow improvement of resting MBF over time (Fig. 3).
In contrast, MBF increased early in MSC-treated animals, peaking 1 wk after stem cell injection (0.13 ± 0.007 and 0.08 ± 0.002, respectively, signal intensity ratio of MI to LVBP; P < 0.05 vs. nontreated; Fig. 3). The increased flow 1 wk after intramyocardial MSC delivery was evident as well compared with placebo injections (Figs. 4A and 5). There was no significant difference between heart rates during follow-up or between groups at any time point. (Table 1). We applied all currently used semiquantitative correction methods for MBF (8, 26, 40) in additional animals, and tissue perfusion in the infarct in MSC-treated animals was always increased relative to control and placebo animals at week 1 (P < 0.01 and P < 0.001; Fig. 5). Importantly, the heart rate was similar at this time point [96 ± 5, 100 ± 8, and 93 ± 7 beats/min for MSC, control, and placebo animals, respectively; P = not significant (NS)]. Four and eight weeks after injection, MBF was similar between the control group and the MSC group (P = NS; Figs. 3 and 4A). Taken together, week 1 was the only significantly different time point within the MSC group and between both groups throughout the follow-up time of the study.
LV Function and Infarct Size
The early increase in resting MBF with MSC treatment was followed by subsequent improvement in regional contractility in the injected infarct territories (Fig. 4B) and increase in LV ejection fraction (LVEF) in these animals (Fig. 4D) evaluated by MRI. LVEF increased from 30.5 ± 2.1% at week 1 to 39.4 ± 2.6% at week 8 in the treatment group (P < 0.05 vs. control and MSC week 1). In contrast, global cardiac function remained unchanged in the untreated group (31.4 ± 0.8% to 27.2 ± 4.5%, week 1 and week 8, respectively; P = NS) and placebo group (34.7 ± 1.5% to 30.2 ± 0.8%, week 1 and week 8, respectively; P = NS). In terms of regional function, endocardial peak Ecc in the infarct area of MSC-treated animals improved from −7.45 ± 0.52 at week 1 to −13.22 ± 0.57 at week 8 (P < 0.05 vs. MSC week 1) but remained unchanged in the control (−5.49 ± 0.59 and −5.24 ± 0.62, week 1 and week 8, respectively; P = NS) and the placebo (−7.85 ± 0.88 and −4.45 ± 1.52, week 1 and week 8, respectively, P = NS) group (Fig. 4B).
In addition, infarct size decreased with MSC treatment from 16.1 ± 1.1% LV mass at week 1 to 12.4 ± 1.0% LV mass at week 8 in the treatment group (P < 0.05 vs. control and MSC week 1). However, the infarct size remained unchanged in the control and slightly increased in the placebo groups (Fig. 4E).
Vessel Density and Apoptosis
Tissue perfusion and overall vessel density, assessed by staining for vWF, was equal in both groups at 8 wk (20.1 ± 1.51 and 22.6 ± 3.571/HPF, MSC vs. control; P = NS). However, the distribution of the vessel sizes showed important differences. While untreated animals primarily exhibited small vessels in the infarct zone (8.8 ± 0.84 and 15.0 ± 2.53/HPF, MSC vs. control; P < 0.05), the MSC-treated animals had larger-sized vessels and enhanced smooth muscle coating (4.1 ± 0.54 and 2.2 ± 0.55/HPF, MSC vs. control; P < 0.05) (Fig. 6 and Table 2). The overall vessel density and breakdown of vessel sizes were alike in the nonischemic areas of treated and untreated animals (Table 2).
The degree of myocyte apoptosis as assessed by TUNEL and caspase-3 staining was less in the endomyocardial rim of the infarct area of MSC-treated pigs versus untreated animals at 8 wk (control vs. MSC, 3.3 ± 0.06 and 1.29 ± 0.40 TUNEL-positive myocytes/mm2, P < 0.05; Fig. 7). In the nonischemic remote areas, the apoptotic index was similar in both groups, thus highlighting a preferential impact of MSC treatment to reduce apoptosis in peri-infarct areas. The fluorescence TUNEL results match the evaluation from the caspase-3 immunohistochemistry (Fig. 7).
Green fluorescent protein-labeled MSCs were detected in the vessel wall of injected animals a week after cell delivery, while placebo animals showed no green fluorescence (Fig. 8).
The major new finding reported here is that transplantation of allogeneic MSCs in regions of MI produces an early increase in tissue perfusion 1 wk after administration. Recovery in cardiac performance and reduction in infarct size occur typically 4–8 wk after therapy, at a time when the tissue exhibits more mature blood vessels and a lower degree of myocyte apoptosis (3). Together these findings suggest that an early effect on tissue perfusion represents an important feature of successful subsequent cardiac regeneration, and this increase offers mechanistic insights that could be useful as a predictive tool to assess improving cardiac function.
Neovascularization is proposed as a dominant effect for improved cardiac function in response to cellular therapeutics. Kinnaird et al. (19) first reported a beneficial effect of MSCs on collateral remodeling and perfusion in a mouse model of hindlimb ischemia. With regard to the heart, there are reports in rats (19, 38, 43), pigs (39), and dogs (34) that implanted MSCs differentiate into endothelial cells and enhance vascular density. Indeed, we too have noted evidence that MSCs participate in neovascularization (3), and we show the evidence of physical incorporation of MSCs into the epithelium of vessels in our porcine model. However, previous studies on vessel morphology were not accompanied by actual measurements of tissue perfusion. Here we used MRI to serially measure tissue perfusion, and accordingly our present data offer novel insights into the time course of improved blood flow in the postinfarct heart.
The only previous MSC study that correlated resting MBF with increased vascular density was conducted in rats and showed increased tissue perfusion after 8 wk of autologous MSC therapy (38). The present data on vessel numbers are in agreement with other investigators who did not observe increased vessel density at 10–15 wk after MSC injection in cryoinfarct and LAD ligation infarct rat models (17, 21). Moreover, Jaquet et al. (17) showed similar results compared with our group for vessel density in infarct and noninfarct myocardium.
In addition, we observed more mature vessel 8 wk after MSC transplantation in the infarct area, a finding that could be reflective of angiogenesis due to MSCs in the infarct area. In terms of preventing adverse post-MI remodeling, it has been postulated that promoting vessel maturation may have an impact beyond the obvious benefit of oxygen and blood supply (7).
Several technical features regarding the use of MRI to measure tissue perfusion warrant mention. In this regard, several different approaches have been used in human and animal studies. We focused on tissue signal intensity change over time (slope) in response to intravenous gadolinium injection because of its correlation with microsphere measurements (41). We also examined the area under the signal intensity curve because both parameters are used in clinical studies as semiquantitative assessments of tissue perfusion (1, 23, 28, 33). The results from both of these approaches supported a similar conclusion. In terms of signal corrections, we also used two different approaches. The upslope parameter was corrected to the gadolinium signal intensity both in the LVBP (8, 33) and in the remote noninfarct myocardium (26, 40) as previously described. Both approaches control for hemodynamic variance during the tracking of the gadolinium bolus while the images are acquired. Again, both correction methods showed similar results.
It is attractive to speculate that the early increase in resting MBF may be due to paracrine effects. MSCs can secrete factors such as vascular endothelial growth factor, stromal cell-derived factor, basic fibroblast growth factor, hepatocyte growth factor, and insulin-like growth factor that could protect or salvage endangered ischemic myocardium (15). Moreover, following the mounting evidence that the heart is not a postmitotic organ capable of regenerating parenchymal cells, implanted MSCs might induce the intrinsic cardiac stem cell pool to home to the site of injection and improve cardiac repair via paracrine signals (5). The beneficial effect of increased perfusion in patients with ischemic cardiomyopathy has been demonstrated and is associated with improved prognosis (12, 24). These data do not exclude the possibility that MSCs differentiate into cardiac myocytes, and to the extent that they do, the present findings suggest that augmented tissue perfusion likely due to neovascularization occurs in advance of significant cardiomyogenesis.
Cardiac MRI measurement of resting MBF is a new modality that has been validated by comparison to single-photon emission computed tomography (SPECT) (11, 22) and PET (16, 25, 33). We focused on resting MBF because increased vessel density should be reflected in increased tissue perfusion at rest. However, the absence of any difference in perfusion after MSC therapy over the long term may be due to the resting state. The more mature vessels observed in MSC-treated animals may allow for sustained improvement in perfusion at exercise or under pharmacological stress, which may account for the reduced apoptosis observed in the endocardial rim. From a practical standpoint this increase in resting MBF, a measurement obtainable from a standard gadolinium-enhanced cardiac MRI image, was among the earliest observable features of the hearts responding to MSC therapy. Accordingly, this could have substantial value in predicting responsiveness to cellular therapeutics, and future work will be directed to addressing this hypothesis.
In summary, whether increased tissue perfusion accounts for cardiac recovery following MSC cardiomyoplasty is controversial. Here we demonstrate with MRI technology that intramyocardial injection of allogeneic MSCs transiently increases resting MBF. This is followed by enhanced maturation of blood vessels, reduction in apoptosis in the endomyocardial rim, less infarct scarring, and improved ventricular function. Our data suggest that sustained resting MBF is not necessary for the improvement in LV function in MSC-treated hearts. However, it seems to be a critically important feature of the myocardial regenerative process. We conclude that an early improvement in resting MBF leading to an increase in mature vessels, reduced apoptosis, and improvement in LV function is a main characteristic of intramyocardial MSC transplantation.
This work was supported by the Johns Hopkins University School of Medicine Institute for Cell Engineering (ICE), National Institutes of Health Grants U54-HL-081028 (Specialized Center for Cell Therapeutics), R21-HL-72185, and RO1-NIA-AG-025017, and the Donald W. Reynolds foundation. J. M. Hare has received research grant funding from Osiris Therapeutics.
The authors thank Virginia Eneboe and Jeff Brawn for excellent technical assistance in performing the animal studies. The authors also thank Norman J. Barker for his expertise and helpful suggestions obtaining pathology and histology images.
↵1 This paper was presented at the 9th Cardiovascular-Kidney Interactions in Health and Disease Meeting at Amelia Island Plantation, Florida, on May 26–29, 2006.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society