- Email: [email protected]
Stem cell therapy enhances electrical viability in myocardial infarction
Journal of Molecular and Cellular Cardiology 42 (2007) 304 – 314 www.elsevier.com/locate/yjmcc
Original article
Stem cell therapy enhances electrical viability in myocardial infarction William R. Mills a , Niladri Mal b , Matthew J. Kiedrowski b , Ryan Unger b , Farhad Forudi b , Zoran B. Popovic c , Marc S. Penn b,c,d , Kenneth R. Laurita a,d,⁎ a
The Heart and Vascular Research Center, MetroHealth Campus, Case Western Reserve University, Cleveland, OH 44109, USA b Department of Cell Biology, Cleveland, OH 44109, USA c Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, OH 44109, USA d The Center for Stem Cell and Regenerative Medicine, Cleveland, OH 44109, USA Received 1 May 2006; received in revised form 12 September 2006; accepted 20 September 2006 Available online 27 October 2006
Abstract Clinical studies suggest increased arrhythmia risk associated with cell therapy for myocardial infarction (MI); however, the underlying mechanisms are poorly understood. We hypothesize that the degree of electrical viability in the infarct and border zone associated with skeletal myoblast (SKMB) or mesenchymal stem cell (MSC) therapy will determine arrhythmia vulnerability in the whole heart. Within 24 h of LAD ligation in rats, 2 million intramyocardially injected SKMB (n = 6), intravenously infused MSC (n = 7), or saline (n = 7) was administered. One month after MI, cardiac function was determined and novel optical mapping techniques were used to assess electrical viability and arrhythmia inducibility. Shortening fraction was greater in rats receiving SKMB (17.8% ± 5.3%, p = 0.05) or MSC (17.6% ± 3.0%, p < 0.01) compared to MI alone (10.1% ± 2.2%). Arrhythmia inducibility score was significantly greater in SKMB (2.8 ± 0.2) compared to MI (1.4 ± 0.5, p = 0.05). Inducibility score for MSC (0.6 ± 0.4) was significantly lower than SKMB (p = 0.01) and tended to be lower than MI. Optical mapping revealed that MSC therapy preserved electrical viability and impulse propagation in the border zone, but SKMB did not. In addition, injected SKMBs were localized to discrete cell clusters where connexin expression was absent. In contrast, infused MSCs engrafted in a more homogeneous pattern and expressed connexin proteins. Even though both MSC and SKMB therapy improved cardiac function following MI in rat, SKMB therapy significantly increased arrhythmia inducibility while MSC therapy tended to lower inducibility. In addition, only MSC therapy was associated with enhanced electrical viability, diffuse engraftment, and connexin expression, which may explain the differences in arrhythmia inducibility. © 2006 Elsevier Inc. All rights reserved. Keywords: Arrhythmia; Electrophysiology; Heart failure; Skeletal myoblasts; Mesenchymal stem cells
1. Introduction Myocardial infarction (MI) is a leading cause of ventricular arrhythmias and sudden cardiac death. Cell therapy for MI using skeletal myoblasts (SKMB) or bone marrow-derived cells has been shown to improve cardiac function [1–6], perfusion [7], symptoms [8], and decrease infarct size [6]. While cell therapy for MI shows promise, some evidence suggests that such therapy is unable to improve electrical function and may even ⁎ Corresponding author. Rammelmamp 6th Floor, Case Western Reserve University, MetroHealth Campus, 2500 MetroHealth Drive, Cleveland, OH 44109, USA. Tel.: +1 216 778 7342; fax: +1 216 778 1261. E-mail address: [email protected] (K.R. Laurita). 0022-2828/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.yjmcc.2006.09.011
be proarrhythmic [9]. Early clinical and experimental studies of SKMB therapy have shown a concerning incidence of ventricular arrhythmias [10–12]. Consistent with this clinical observation, basic studies have shown that SKMB do not electrically couple with native myocardium in vivo [13] and when injected intramyocardially tend to cluster near injection sites [14,15], both of which may be the basis for arrhythmias reported in clinical studies. In contrast, recent studies using bone marrow derived cell therapy [5,12,16] have not reported a significant incidence of arrhythmias. Mesenchymal stem cells (MSC) can be delivered by intravenous infusion soon after MI and have been shown to form gap junctions with host cardiomyocytes in vivo [17,18]. Thus, MSC may enhance electrical viability of myocardium damaged by MI.
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Despite the rapidity in which cellular therapy for MI has progressed from the laboratory to the clinic, many fundamental questions regarding the electrophysiological and arrhythmic consequences are unresolved. Importantly, it is unknown whether cell therapy can enhance electrical viability of damaged myocardium and reduce the risk of post-MI arrhythmias. In this report, we utilize novel optical mapping techniques to determine the electrophysiological consequence of cell therapy in the intact heart, as commonly tested in clinical and experimental studies. In rat hearts with MI, SKMB were directly injected into the infarct and infarct border zone, and MSCs were delivered by intravenous infusion. We tested the hypothesis that the degree of electrical viability in the infarct and border zone associated with SKMB or MSC therapy will determine arrhythmia vulnerability in the whole heart. 2. Materials and methods 2.1. Experimental model This investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication No 85-23, Revised 1996) and was approved by the IACUC of both Case Western Reserve University and the Cleveland Clinic Foundation. To create a model of MI, we performed permanent ligation of the left anterior descending coronary artery (LAD) in rats as described previously [19]. Four groups of rat models were utilized in this study: normals (n = 5), control MI (n = 7), MI + SKMB (n = 6), and MI + MSC (n = 7). The methods used to create these groups are described below. 2.2. SKMB procedure SKMB were isolated and prepared as described previously [20]. SKMB were labeled with DiI (1,1′-dioctadecyl 3,3,3,3′tetra-methylind-carbocyanine percolate) before intramyocardial injection (see below). SKMB were injected into the lateral aspects of the infarct zone approximately 5 min after LAD ligation. A total of five injections (40 μL each containing ∼ 0.4 × 106 cells) were made, two each along the right and left borders of the infarct zone, and one near the apex of the heart. 2.3. MSC procedure Rat bone marrow was isolated by flushing the femurs with 0.6 ml DMEM (GIBCO, Invitrogen, Carlsbad, CA). Clumps of bone marrow were gently minced with an 18-gauge needle. Cells were separated by a Percoll density gradient and then centrifuged for 10 min at 260 G and washed with three changes of PBS with 100 U/ml penicillin/100 g/ml streptomycin (Invitrogen, Carlsbad, CA). The washed cells were then resuspended and plated in DMEM-LG (GIBCO, Invitrogen, Carlsbad, CA) with 10% FBS and 1% antibiotic and antimycotic (GIBCO, Invitrogen, Carlsbad, CA). The
305
cells were incubated at 37 °C. Non-adherent cells were removed by replacing the medium after 3 days. Cultures were refed every 3–4 days. When cultures became 70% confluent, adherent cells were detached following incubation with 0.05% trypsin and 2 mM EDTA (Invitrogen, Carlsbad, CA) for 5 min and subsequently passaged. MSC Cultures were then depleted of CD45+ and CD34+ cells simultaneously by negative selection using 10 μl each of primary PE-conjugated mouse anti-rat CD45 (BD Biosciences, San Diego, CA) and CD34 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) per 106 cells using the EasySep PE selection kit according to the manufacturer's instruction (Stem Cell technologies) to prevent non-specific selection of endothelial progenitor cells (EPC), monocytes, and macrophages. In preceding experiments, confluent cells were labeled with green fluorescent protein (GFP, see below), passaged and plated out at 1:4 to 1:5 dilutions until passage 11. In vitro differentiation assays confirmed that the GFP-tagged MSC had retained the potential for adipogenic, osteogenic, and chondrogenic differentiation. Two million labeled cells (GFP expressing MSC) harvested in 200 μl of PBS were infused via tail vein 24 h after MI. 2.4. GFP labeling of MSC We implemented a third-generation vesicular stomatitis virus glycoprotein (VSV-G) pseudotyped lentivirus expressing enhanced green fluorescent protein (EGFP) reporter gene and zeocin resistance gene driven by the cytomegalovirus (CMV) and the simian virus 40 (SV40) promoter respectively. The lentivirus was made using a four plasmid vector system. MSC were transduced twice for 8 h with purified lentivirus in the presence of 8 μg/ml of polybrene at a multiplicity of infection (MOI) of 30. The media were changed 72 h post-transfection and replaced with regular media containing zeocin (400 μg/ ml). Only cells that incorporated the viral genome, including the zeocin resistance gene, survived. Transfection of MSC with our protocol resulted in expression of EGFP 36 h later in 30% of the cells. FACS analysis to sort GFP expressing cells resulted in an enriched population with ∼ 90–95% of the cells being GFP+. MSC were then expanded using standard techniques. 2.5. DiI labeling of SKMB Because Dil does not spread from labeled to unlabeled cells, and it does not appear to have any adverse effects on living cells, we used Dil to trace transplanted SKMB and to obtain critical information about location, histology, and number of transplanted cells [14]. In short, we trypsinized, washed, and resuspended SKMB in serum-free culture media at a density of 1 × 106 /ml 5 μL (Molecular Probes, Carlsbad, CA). Cells were incubated for 20 min in a dark CO2 incubator at 37 °C. Cells were washed and resuspended in warm media three times. Before transplantation, the cells were pelleted and resuspended in warmed PBS to a concentration of 2 × 106 cells/200 μL.
306
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
2.6. Echocardiography Echocardiography was performed within 1 week of LAD ligation to confirm MI, and repeated 1 month later to assess ventricular function. Two-dimensional echocardiography was performed using a 15-MHz linear-array transducer interfaced with an Acuson Sequoia C256 (Siemens Medical Solutions, Malvern, PA, USA) as described previously [19,22]. Animals were lightly sedated with ketamine (50 mg/kg i.p.) for each echocardiogram. The presence of significant anterior MI was confirmed by anterior wall akinesis present in at least two adjacent segments on two-dimensional echocardiography [23]. As a measure of left ventricular function, the shortening fraction was calculated from M-mode recordings. Dimensions were measured between the anterior wall and the posterior wall from the short-axis view at the level of the papillary muscles. 2.7. Optical mapping Optical mapping was performed in Langendorff-perfused rat hearts 1 month after MI as described recently [24]. Normal (n = 5), control MI (n = 7), MI + MSC (n = 7), and MI + SKMB (n = 6) rats were injected with heparin 300 U i.p. 15 min before the heart was removed. Rats were then anesthetized with sodium pentobarbital 50 mg/kg i.p. After attainment of adequate anesthesia, the heart was immediately excised and arrested by direct aortic perfusion of cold cardioplegia solution (10 ml). Extraneous tissue was carefully dissected away from the heart, which was then placed in a beaker containing 20 ml oxygenated ice-cold cardioplegia and 180 μM of 4-[β-[2-(di-n-butylamino)6-naphthyl]vinyl]pyridium (Di-4-ANEPPS) for 20 min. The beaker was shaken gently every 60 s to ensure uniform staining of the entire epicardial surface. After completion of dyesuperfusion staining, the aorta was cannulated, and perfusion with 37 °C Tyrode's solution containing (mM) 121.7 NaCl, 25.0 NaHCO3, 2.74 MgSO4, 4.81 KCl, 5.0 dextrose, and 2.0 CaCl2 (pH 7.4) was started at ascending flow rates. Perfusion pressure in the cannula was monitored with a pressure transducer (Harvard Apparatus, Saint Laurant, Quebec, Canada), allowing for a ramp up to a target pressure of 70– 90 mm Hg over a 5-min period. After achievement of the target perfusion pressure, the cannulated heart was immersed in 37 °C Tyrode's solution in a specialized optical recording chamber for the duration of each experiment. ECG, perfusion pressure, and bath temperature were measured continuously during all experiments. The optical mapping system used has been described in detail previously [24,25]. Briefly, action potentials were optically recorded at a magnification of 2.1× from 256 sites within an 8.3 mm × 8.3 mm mapping field (0.52-mm interpixel resolution) on the anterior epicardial surface of the left and right ventricle. Fluorescence was excited with uniform light from a 270-W tungsten-halogen light source (filtered 514 + 20 nm) and transmitted to a 16 × 16-element photodiode array detector through a tandem-lens imaging system (emission filter > 610 nm). Photocurrent from each photodiode underwent current-to-voltage conversion, amplification, and bandpass
filtering (0.1–500 Hz) and was multiplexed and digitized (1000 samples/s per channel) with 12-bit precision. A CCD camera that is optically aligned with the photodiode array was used to obtain visible images of the optical mapping field relative to anatomical landmarks (e.g., infarct scar and border zone). 2.8. Experimental protocol After staining with Di-4 ANEPPS, all hearts were placed against the optical imaging window. CCD images of the epicardial surface of the heart were used to initially position and maintain the location of the heart relative to the mapping field. The transition between non-infarcted (i.e., pink) and infarcted (i.e., white) tissue was used to approximate the location of the infarct border zone. This ensured that non-infarcted, border zone, and infarcted tissue were as close as possible to the center of the mapping field for each heart tested (see Fig. 1A) [24]. Each heart was perfused with Tyrode's plus 20 mM of diacetylmonoxine (DAM) to eliminate motion artifact. All hearts were paced at a steady state cycle length of 150 ms using a silver unipolar electrode placed on the epicardial surface. Optical recordings were made during steady state pacing and during programmed stimulation. All preparations remained viable for the entirety of each optical mapping experiment. Coronary perfusion pressure was maintained at 80 ± 10 mm Hg for all experiments. At the conclusion of optical mapping, hearts were stained with triphenyltetrazolium chloride (TTC) using the method of Downey [31]. Fluorescence microscopy, using specialized di-4-ANEPPS fluorescence filters, was then performed on ventricular cross sections to ensure homogenous optical dye staining of non-infarcted and infarcted myocardium. 2.9. Data analysis Using previously described automated algorithms [27], activation times were assigned for each action potential by identifying the greatest positive change in fluorescence (maximum derivative) during each beat. All assigned times were verified visually by an experienced investigator. From local depolarization times, spatial contour maps (i.e., activation maps) of impulse propagation were calculated and projected onto the epicardial surface of the heart [28]. Impulse block was defined by the absence of an action potential at a recording site, other than from sites corresponding to the corners of the array, where lens vingetting significantly reduced fluorescence intensity independent of tissue viability. To assess electrical viability, conduction velocity and unnormalized optical action potential amplitude were measured in non-infarcted, border zone, and infarcted tissue. Conduction velocity was calculated in the infarct border zone during steady state pacing with respect to fiber orientation using an average of velocity vectors parallel to the direction of conduction, as described previously [25,28]. The unnormalized amplitude of optically recorded action potentials was determined by calculating the difference in fluorescence just before depolarization (during rest) and immediately after depolarization at the time corresponding to
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
the peak of the action potential as described previously in hearts with MI [24]. To determine arrhythmia vulnerability, programmed stimulation was performed with up to 3 premature stimuli, and an inducibility score was assigned to each heart. Inducibility was defined as the ability to provoke non-sustained or sustained (≥ 30 s) ventricular tachycardia or ventricular fibrillation. A score of 0 was assigned if the preparation was not inducible; 1 if inducible with 3 premature stimuli; 2 if inducible with 2 premature stimuli; and 3 if inducible with 1 premature stimulus. The higher the score, the easier it was to induce arrhythmias. The unpaired T test was used to test for statistical
307
significance, except for multiple comparisons, when ANOVA was performed followed by Newman–Keuls post hoc analysis. A p value <0.05 was considered statistically significant. Data are presented as mean ± SE unless otherwise noted. 2.10. Confocal immunofluorescence and histology Immunostaining was performed on all rat hearts following optical mapping as previously described [24]. In short, tissue was fixed in formalin and embedded in paraffin blocks. Antigen retrieval was performed using 10 mM sodium citrate buffer (pH
Fig. 1. Experimental preparation. (A) Photograph of the anterior surface of a Langendorff-perfused rat heart with MI within the optical mapping chamber. Note that the optical mapping area (black box) includes non-infarcted zone (NI), border zone (BZ), and infarct zones (IZ). Shown below are representative H&E images from each group. Tissues sections are mid-ventricular LV cross-sections.
308
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Fig. 2. SKMB (n = 6) and MSC (n = 7) therapy improved cardiac function to a similar extent compared to control MI (n = 7). Shown is a histogram illustrating shortening fractions measured by echocardiography 1 month after MI. Values shown are mean ± SD.
oven, de-paraffinized in xylene, and stained with hematoxylin and eosin (H&E). Photomicrographs of each section were acquired using an ArtixScan 4000tf slide scanner (Microtek, Carson, CA) in conjunction with the PathScan Enabler (Meyer Instruments, Houston, TX) at a resolution of 4000 dpi. Total infarct size was estimated from mid-ventricular LV crosssections using Image-Pro Plus version 5.1 image analysis software as we have previously shown [22]. Total infarct size was quantified by calculating the percent of the LV area with infarcted tissue compare to the total LV area. Total LV area and infarct area were calculated based on total epicardial and endocardial circumference and the epicardial and endocardial circumferential extent of the infarct, respectively. Two independent observers blinded to the study measured the amount of infarct in percent, and their values were then averaged. 3. Results
6.0) heat at 95 °C for 5 min. The buffer was replaced with fresh buffer and reheated for an additional 5 min and then cooled for approximately 20 min. The slides were then washed in deionized water three times for 2 min each. Specimens were then incubated with 1% normal blocking serum in PBS for 60 min to suppress non-specific binding of IgG. Slides were then incubated for 60 min with appropriate primary antibody. Optimal antibody concentration was determined by titration. Slides were then washed with phosphate-buffered saline (PBS) and then incubated for 45 min with corresponding secondary antibody conjugated with Alexa Fluor fluorochrome 488 (green) or 594 (red) diluted to 1.5 μg/ml in PBS with serum and incubated in a dark chamber. In double staining experiment, slides were incubated with another set of primary and secondary antibody in succession. After washing extensively with PBS, coverslips were mounted with aqueous mounting medium Vectashield with DAPI (Vector Laboratories, Burlingame, CA). Primary antibodies: rabbit anti-connexin-43 polyclonal IgG antibody (Santa Cruz Biotechnology, Inc.); rabbit anti-connexin 45 polyclonal IgG antibody (Santa Cruz Biotechnology, Inc.); goat polyclonal anti-connexin-40 IgG antibody (Santa Cruz Biotechnology, Inc.). Secondary antibodies: donkey anti-rabbit IgG Alexa Fluor 488 (Molecular Probes); donkey anti-rabbit IgG Alexa Fluor 594(Molecular Probes); donkey anti-goat IgG Alexa Fluor 488 antibody (Molecular Probes); donkey anti-goat IgG antibody Alexa Fluor 594 (Molecular Probes). An upright spectral laser scanning confocal microscope (Model TCS-SP; Leica Microsystems, Heidelberg, Germany) equipped with blue argon (for DAPI), green argon (for Alexa Fluor 488), and red krypton (for Alexa Fluor 594) laser was used for confocal analysis. Data were collected by sequential excitation to minimize “bleed-through”. Image processing, analysis and the extent of colocalization were evaluated using Leica Confocal software. Optical sectioning was averaged over four frames and the image size was set at 1024 × 1024 pixels. There were no digital adjustments made to the images. In three hearts from each group, mid-ventricular and apical segments that were paraffin-embedded were sectioned (5 μm) and applied to glass slides. Then, the slides were dried in an
In all infarcted rats, echocardiography confirmed the presence of anterior wall infarction and decreased cardiac function. Shown in Fig. 1 (bottom) are representative H&Estained sections used to qualitatively assess infarct size for each group with MI. In control MI, MI + SKMB, and MI + MSC, significant thinning of the injured wall and dilatation of the ventricular cavity is evident compared to a normal heart. The percent of infarcted LV was reduced for MI + SKMB (25% ± 12%, n = 3) and MI + MSC (30% ± 7%, n = 3) compared to control MI (37% ± 16%, n = 3); however, there was no statistical difference among all groups. As shown in Fig. 2, echocardiography demonstrated a significant improvement in shortening fraction in both MI + SKMB (17.8% ± 5.3%, p = 0.05) and MI + MSC (17.6% ± 3.0%, p < 0.01) groups compared to control MI (10.1% ± 2.2%).
Fig. 3. Skeletal myoblast therapy was associated with increased arrhythmia vulnerability. Arrhythmia vulnerability was determined by an arrhythmia inducibility score (see text for details) where the larger number indicates higher arrhythmia inducibility. SKMB (n = 6) was associated with increased vulnerability to ventricular tachycardia compared to normal hearts (n = 5), control MI (n = 7), and MI + MSC (n = 7). MSC therapy was associated with a nonsignificant decrease in arrhythmia vulnerability compared to control MI.
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
During programmed stimulation, ventricular tachycardia was induced in 57% of hearts with MI alone (control MI) compared to 0% in normal hearts. Importantly, arrhythmia inducibility in
309
MI + SKMB (100%) was greater than MI + MSC (29%). Similarly, arrhythmia inducibility score was significantly higher in the MI + SKMB group (2.8 ± 0.2) compared to both control
Fig. 4. MSC, but not SKMB, therapy was associated with enhanced electrical viability in the infarct and border zone as determined by greater action potential amplitude and evidence of impulse propagation (activation maps). Activation maps of impulse propagation and unnormalized optically recorded action potentials obtained from normal (A), control MI (B), MI + SKMB (C), and MI + MSC (D) hearts during optical mapping. The images in each panel show the area mapped in each heart, and the signals beside each activation map were recorded from non-infarcted (site 1), border zone (sites 2 and 3), and infarcted (site 4) tissue. Action potentials are plotted as absolute florescence. For all activation maps, each contour line represents 2 ms. The thick black lines indicate impulse block.
310
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Fig. 5. Enhanced electrical viability of the infarct border zone of rats treated with mesenchymal stem cells. Shown is a plot of action potential amplitude versus distance across non-infarcted, border zone, and infarcted tissue of normal (n = 5, circle), control MI (n = 7, square), MI + MSC (n = 7, triangle), and MI + SKMB (n = 6, diamond) hearts. The border zone occurs between −1 and +1 mm (corresponding to signals recorded from sites 2 and 3 in each panel of Fig. 4, shown here in parentheses above each point). Asterisks (*) denote statistical significance between MI + MSC, normal, and control MI groups.
MI (1.4 ± 0.5, p = 0.05) and MI + MSC (0.6 ± 0.4, p = 0.01) groups (Fig. 3). Interestingly, the MI + MSC group tended to be less vulnerable to arrhythmias than control MI; however, this
difference did not reach statistical significance. These data suggest that SKMB therapy for MI increases vulnerability to ventricular arrhythmias whereas MSC therapy may reduce arrhythmia vulnerability. To determine electrical viability, optical action potentials were recorded simultaneously from non-infarcted, border zone, and infarcted tissue. Shown in Fig. 4 are representative examples of activation contour maps (constructed from activation times determined at each recording site) and unnormalized action potentials, plotted as absolute fluorescence, recorded from normal, control MI, MI + SKMB, and MI + MSC hearts. The image in each panel shows the location from which action potentials were recorded (right) corresponding to non-infarcted tissue (site 1), border zone (sites 2 and 3), and infarct zone (site 4). The activation map and action potentials shown in Fig. 4A were recorded from a normal heart and demonstrate uniform impulse propagation without evidence of block (thick black lines), in addition to uniform amplitude and normal action potential morphology. In this example, the conduction velocity between sites 2 and 4 is 0.55 m/s. Shown in Fig. 4B is a representative example of impulse propagation and unnormalized action potentials recorded from a control MI
Fig. 6. Representative SKMB (Panel A) and MSC therapy (Panel B) engraftment patterns and connexin expression (Cx40, top; Cx43, middle, Cx45, bottom). Each panel is a confocal immunofluorescent photomicrograph taken at 40× magnification (scale bar 40 μm) of ventricular cross sections from the MI + SKMB (Panel A) and MI + MSC (Panel B) group. Each image roughly corresponds to tissue between the sites labeled as 3 and 4 in each panel in Fig. 4, as indicated by the colocalization of DAPI (blue) and Dil (red).
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
311
Fig. 6 (continued).
heart. The activation map indicates regions of impulse block (thick black line) corresponding to the location of the border zone (between sites 2 and 3). The action potential recorded from non-infarcted tissue (1) was similar to that measured from normal hearts, whereas signals recorded from the border zone (2 and 3) demonstrated an abrupt decrease in amplitude, and a slower depolarization phase. No action potential activity was observed further into the infarct zone (site 4). In this example, the conduction velocity from sites 2 to 3 (i.e., in the border zone) was 0.31 m/s. For MI + SKMB (Fig. 4C), impulse propagation and unnormalized action potentials recorded from non-infarcted tissue (1) were similar to that measured from normal rats, whereas signals recorded from the border zone (2 and 3) were of low amplitude and had a slower depolarization phase. In this example, the conduction velocity was 0.21 m/s. No action potential activity was observed at site 4 (similar to control MI). Importantly, impulse block and significant conduction slowing was observed in the infarct border zone of MI + SKMB (0.21 ± 0.05 m/s) compared to control MI (0.33 ± 0.05 m/s; p = 0.02). Shown in Fig. 4D are impulse propagation and unnormalized action potentials recorded from a heart with MI + MSC. The activation map reveals impulse propagation beyond the border zone (sites 3 and 4) towards the center of the infarct and a smaller region of conduction block compared to MI + SKMB and control MI. The action potential recorded from non-infarcted tissue (1) was
similar to action potentials recorded in normal hearts. Action potentials recorded from the border zone (sites 2 and 3) were smaller than normal, but not as small as those recorded from the border zone in control MIs (Fig. 4B). In addition, unlike control MIs, an action potential was measured at site 4 (furthest point into the infarct zone) with an activation time that was delayed compared to that measured at site 3, indicating some degree of impulse propagation in the infarct zone (as also indicated in the activation map). In this example, the conduction velocity from sites 1 to 4 was (0.24 m/s). Overall, mean conduction velocity in the border zone for MI + MSC (0.24 ± 0.07) was slower but not significantly different from control MI (0.33 ± 0.05, p = NS). Enhanced electrical viability was observed only in hearts that received MSC therapy (4 of 7). Shown in Fig. 5 are average action potential amplitudes in non-infarcted tissue (site 1), border zone (sites 2 and 3), and infarct zone (site 4), recorded from normal hearts and hearts with MI, MI + MSC, and MI + SKMB. In each heart, action potential amplitude was normalized to that recorded from noninfarcted tissue (site 1). In normal hearts (circles), action potential amplitude decreased by an average of 16% ± 6.8% from sites 1 to 4. This decrease likely resulted from a slight fall off in excitation light intensity or lens vingetting. In hearts with control MI (squares), action potential amplitude furthest into the infarct zone (site 4) decreased by 82% ± 6.9%. In contrast, in hearts with MI + MSC (triangles), action potential amplitude
312
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
furthest into the infarct zone decreased by only 45% ± 20.8% and was significantly greater that that compared to control MI (p < 0.05). In MI + SKMB hearts (diamonds), action potential amplitude furthest into the infarct decreased by 86% ± 3.0% and was similar to hearts with control MI. These data suggest that MSC therapy is associated with enhanced electrical viability of the infarct and border zone, whereas SKMB is not. Shown in Fig. 6 are representative confocal immunofluorescent photomicrographs taken at 40× magnification (scale bar 40 μm) from ventricular cross sections in the, MI + SKMB (Fig. 6A) and MI + MSC (Fig. 6B) groups. In each panel, expression of Cx40 (first row), Cx43 (middle row), and Cx45 (bottom row) are shown. Each image corresponds to an area of tissue near sites 3 or 4 in each panel shown in Fig. 4. As demonstrated in Fig. 6A, intramyocardially injected SKMB (red) colocalized with DAPI (blue), indicating that DiI-labeled cells survived and tended to engraft heterogeneously in cell clusters. In addition, expression of Cx40 (top row, green), Cx43 (middle row, green), and Cx45 (bottom row, green) were absent where Dil-positive cells were located, indicating a lack of connexin expression associated with SKMB therapy. In contrast, intravenously infused MSCs (Fig. 6B, green) were homogeneously distributed. Interestingly, MSC engraftment was associated with the expression of all connexin proteins tested, as indicated by the colocalization of connexin protein (red) and MSCs (green) in the merged images. In summary, these data suggest that in contrast to SKMB therapy, the homogeneous engraftment and connexin expression associated with MSC therapy led to enhanced electrical viability and reduced arrhythmia inducibility. 4. Discussion Despite the incidence of ventricular arrhythmias observed in early clinical trials [10,11], there have been very few reports that have systematically evaluated the electrophysiological and arrhythmic consequence of cell therapy for MI in the whole heart. In this report, we have shown that (1) although SKMB therapy improves cardiac function, it increases arrhythmia vulnerability and does not enhance electrical viability in the infarct or border zone; and (2) MSC therapy improves cardiac function and enhances electrical viability in the infarct and border zone, which may explain why arrhythmia vulnerability was similar or, possibly, lower than control MI.
SKMB and cardiomyocytes. Moreover, Krucoff et al. [30] have shown that arrhythmia vulnerability is greater when cells are injected into the border zone compared to the central zone of an infarct. It is possible that DiI-labeled cells, which lacked the expression of connexin proteins, formed in clusters near the injection sites, producing electrical barriers to conduction, as we have previously shown when SKMB are injected into normal myocardium [14]. MI + MSC hearts were significantly less vulnerable to arrhythmias than MI + SKMB hearts, and trended towards being less vulnerable to arrhythmias than control MIs. The low arrhythmia vulnerability we observed in MI + MSC hearts is also consistent with the low incidence of sudden death and ventricular arrhythmias reported in studies of bone-marrow derived cell therapies [5,12,16]. Interestingly, MI + MSC hearts tended to exhibit conduction slowing in the border zone as did MI + SKMB (albeit not statistically significant). This finding is consistent with recent work by Chang et al. [26], who showed conduction slowing in cocultures of MSC and cardiomyocytes. Therefore, slow conduction in the border zone may not be a sole determinant of arrhythmia inducibility as measured in the whole heart. On the other hand, enhanced electrical viability (see next section) and the expression of connexin protein in the border zone of MI + MSC hearts may explain the observed reduction in arrhythmia vulnerability. Our study suggests that the delivery method of cell therapy (SKMB injection versus MSC IV infusion) may have affected arrhythmia vulnerability. For example, the clustering of directly injected SKMB may have formed a local zone of slow conduction [14], which is a substrate for reentrant excitation. In contrast, MSCs delivered by IV engrafted in a more homogeneous pattern (Fig. 6). Would SKMBs engraft in a more homogeneous pattern if delivered by IV infusion? Possibly, but it is unlikely that SKMB would engraft in the heart at all when delivered by IV infusion. Would MSCs delivered by direct intramyocardial injection result in an engraftment pattern similar to SKMB? Possibly; but unlike SKMBs, MSCs should express connexin proteins. Moreover, Fernandes et al. [12] demonstrated that MSCs delivered by direct injection also have a much lower rate of arrhythmia inducibility compared SKMB, suggesting that the cell type may be just as important as the method of delivery. Additional studies are required to determine the optimal delivery method for each cell type.
4.1. Arrhythmia vulnerability 4.2. Electrical viability Strikingly, we report that 100% of MI + SKMB hearts had inducible ventricular arrhythmias. This uniform inducibility observed in the MI + SKMB group is consistent with that from Fernandes et al. [12] and a recent study, in which we also reported 100% inducibility in rats with chronic MI that received SKMB 2 months after MI [21]. The increase in arrhythmia vulnerability observed in MI + SKMB hearts may be related to reduced conduction velocity in the infarct border zone compared to control MI. These results are consistent with those reported by Abraham et al. [29], who showed slowing of conduction and inducible sustained reentry in cocultures of
To determine electrical viability associated with cell therapy, we measured optical action potential amplitude and impulse propagation within the infarct border zone. In MI + SKMB hearts, we observed a fall off in action potential amplitude at the border zone and no evidence of impulse propagation into the infarct zone, a pattern that was undistinguishable from control MI. This finding is consistent with the work of Reinecke et al. [13], who showed that SKMB do not electrically integrate into the myocardium. In contrast, the border zone of MI + MSC hearts was electrically more viable than control MI, as
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
evidenced by action potential and impulse propagation deeper into the infarct zone. The optical action potential amplitude, however, must be interpreted with caution because it depends on several factors such as dye staining and excitation light, in addition to the number of electrically viable cells in a given area. To address these factors, we have shown previously that the voltage-sensitive dye superfusion technique homogenously stains non-infarcted and infarcted myocardium [24]. In addition, we showed that over a relatively short distance (∼ 6 mm) action potential amplitudes were very similar in normal hearts (Fig. 5, circles), ensuring uniform excitation light intensity and dye staining. However, 95% of Di-4-ANEPPS fluorescence is measured from a depth of < 0.5 mm [32]. Therefore, our previous study [24] and the current study demonstrate that optical action potential amplitude is a reasonable approximation of relative electrical viability within the outermost epicardial layer. The increase in electrical viability may have resulted from the homogenous distribution of viable MSCs throughout the border zone and the expression of connexin proteins (Fig. 6B). Valiunas et al. [17] and Potapova et al. [18] have shown that low frequency gap junction formation between MSC and cardiomyocytes can occur. It is possible that MSC therapy may have provided passive coupling between cardiomyocytes, as has been shown with fibroblasts [33]. Therefore, enhanced electrical coupling between MSC and native cardiomyocytes may have, at least in part, led to the increase in electrical viability we observed. As a result, the electrical size of the infarct might decrease and reduce the path length for reentrant excitation and thus make arrhythmias less likely. Further studies are needed to test this possibility. The presence of all three connexin proteins may reflect MSCs at various stages of differentiation. It is well known that the developing mammalian heart expresses varying levels of connexins during embryological development, and the expression of connexins in the present study may be analogous to the ubiquity of these connexins during development [34–36]. It is possible that some of our results may be caused by a paracrine effect. For example, MSC therapy has been previously associated with increased angiogenic factors and a down-regulation of proapoptotic proteins [37]. We did not observe cardiogenesis based on differentiation of injected cells. Thus, it seems likely that the improved function (Fig. 2) observed in response to cell therapy was due to paracrine effects that might be the result of similar or distinct pathways between MSC and SKMB. 4.3. Clinical implications SKMB therapy was associated with improved cardiac function but did not enhance electrical viability. Moreover, SKMB therapy was associated with a significant increase in vulnerability to ventricular arrhythmias and slow conduction in the infarct border zone compared to control MI. These findings may explain the reported high incidence of arrhythmias observed in early clinical trials of SKMB therapy. In contrast, therapy with intravenous MSC led to improved cardiac
313
function, enhanced electrical viability of the infarct and border zone, and was not associated with an increase in vulnerability to ventricular arrhythmias. Overall, these data suggest that the electrophysiological and arrhythmic consequence of cell therapy for MI depends on changes in electrical viability of the infarct and border zone of MI as determined by the cell type and delivery method. If these data are corroborated by future basic and clinical studies, MSC could emerge as a valuable therapy for MI-associated sudden cardiac death in patients. Acknowledgments This study was supported by NIH grants HL68877 (KRL) and HL74400 (MSP), and grants from The Wilson Foundation (MSP), Shalom Foundation (MSP), Ohio Valley Affiliate of the American Heart Association (KRL), and the Biological Research Technology Transfer Fund of the State of Ohio (MSP and KRL). This work was performed during Dr. Mills' tenure as the Michael Bilitch Fellow in Cardiac Pacing and Electrophysiology of the NASPE-Heart Rhythm Society. References [1] Taylor DA, Atkins BZ, Hungspreugs P, Jones TR, Reedy MC, Hutcheson KA, et al. Regenerating functional myocardium: improved performance after skeletal myoblast transplantation. Nat Med 1998;4:929–33. [2] Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, et al. Bone marrow cells regenerate infarcted myocardium. Nature 2001;410: 701–5. [3] Yau TM, Tomita S, Weisel RD, Jia ZQ, Tumiati LC, Mickle DAG, et al. Beneficial effect of autologous cell transplantation on infarcted heart function: comparison between bone marrow stromal cells and heart cells. Ann Thorac Surg 2003;75:169–76. [4] Menasche P, Hagege AA, Scorsin M, Pouzet B, Desnos M, Duboc D, et al. Myoblast transplantation for heart failure. Lancet 2001;357:279–80. [5] Assmus B, Schachinger V, Teupe C, Britten M, Lehmann R, Dobert N, et al. Transplantation of progenitor cells and regeneration enhancement in acute myocardial infarction (TOPCARE-AMI). Circulation 2002;106: 3009–17. [6] Strauer BE, Brehm M, Zeus T, Kostering M, Hernandez A, Sorg RV, et al. Repair of infarcted myocardium by autologous intracoronary mononuclear bone marrow cell transplantation in humans. Circulation 2002;106: 1913–8. [7] Stamm C, Westphal B, Kleine HD, Petzsch M, Kittner C, Klinge H, et al. Autologous bone-marrow stem-cell transplantation for myocardial regeneration. Lancet 2003;361:45–6. [8] Tse HF, Kwong YL, Chan JK, Lo G, Ho CL, Lau CP. Angiogenesis in ischaemic myocardium by intramyocardial autologous bone marrow mononuclear cell implantation. Lancet 2003;361:47–9. [9] Makkar RR, Lill M, Chen PS. Stem cell therapy for myocardial repair: is it arrhythmogenic? J Am Coll Cardiol 2003;42:2070–2. [10] Menasche P, Hagege AA, Vilquin JT, Desnos M, Abergel E, Pouzet B, et al. Autologous skeletal myoblast transplantation for severe postinfarction left ventricular dysfunction. J Am Coll Cardiol 2003;41:1078–83. [11] Smits PC, van Geuns RJ, Poldermans D, Bountioukos M, Onderwater EE, Lee CH, et al. Catheter-based intramyocardial injection of autologous skeletal myoblasts as a primary treatment of ischemic heart failure: clinical experience with six-month follow-up. J Am Coll Cardiol 2003;42:2063–9. [12] Fernandes S, Amirault JC, Lande G, Nguyen JM, Forest V, Bignolais O, et al. Autologous myoblast transplantation after myocardial infarction increases the inducibility of ventricular arrhythmias. Cardiovasc Res 2006;69:348–58.
314
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
[13] Reinecke H, MacDonald GH, Hauschka SD, Murry CE. Electromechanical coupling between skeletal and cardiac muscle. Implications for infarct repair. J Cell Biol 2000;149:731–40. [14] Fouts KH, Fernandes B, Mal N, Liu JQ, Laurita KR. Electrophysiological consequence of skeletal myoblast transplantation in normal and infarcted canine myocardium. Heart Rhythm 2006;3:452–61. [15] Thompson RB, Emani SM, Davis BH, van den Bos EJ, Morimoto Y, Craig D, Glower D, et al. Comparison of intracardiac cell transplantation: autologous skeletal myoblasts versus bone marrow cells. Circulation 2003;108(Suppl 1):II264–71. [16] Wollert KC, Meyer GP, Lotz J, Ringes-Lichtenberg S, Lippolt P, Breidenbach C, et al. Intracoronary autologous bone-marrow cell transfer after myocardial infarction: the BOOST randomised controlled clinical trial. Lancet 2004;364:141–8. [17] Valiunas V, Doronin S, Valiuniene L, Potapova I, Zuckerman J, Walcott B, et al. Human mesenchymal stem cells make cardiac connexins and form functional gap junctions. J Physiol 2004;555:617–26. [18] Potapova I, Plotnikov A, Lu Z, Danilo Jr P, Valiunas V, Qu J, et al. Human mesenchymal stem cells as a gene delivery system to create cardiac pacemakers. Circ Res 2004;94:952–9. [19] Askari AT, Unzek S, Popovic ZB, Goldman CK, Forudi F, Kiedrowski M, et al. Effect of stromal-cell-derived factor 1 on stemcell homing and tissue regeneration in ischaemic cardiomyopathy. Lancet 2003;362:697–703. [20] Jain M, DerSimonian H, Brenner DA, Ngoy S, Teller P, Edge AS, et al. Cell therapy attenuates deleterious ventricular remodeling and improves cardiac performance after myocardial infarction. Circulation 2001;103: 1920–7. [21] Deglurkar I, Mal N, Mills W, Popovic ZB, Kiedrowski MJ, Forudi F, McCarthy PM, Blackstone EH, Laurita KR, Penn MS. Mechanical and electrical effects of cell therapy are independent. Hum Gene Ther in press. [22] Askari A, Unzek S, Goldman CK, Ellis SG, Thomas JD, DiCorleto PE, et al. Cellular, but not direct, adenoviral delivery of vascular endothelial growth factor results in improved left ventricular function and neovascularization in dilated ischemic cardiomyopathy. J Am Coll Cardiol 2004;43: 1908–14. [23] Morgan EE, Faulx MD, McElfresh TA, Kung TA, Zawaneh MS, Stanley WC, et al. Validation of echocardiographic methods for assessing left ventricular dysfunction in rats with myocardial infarction. Am J Physiol Heart Circ Physiol 2004;287:H2049–53. [24] Mills WR, Mal N, Forudi F, Popovic ZB, Penn MS, Laurita KR. Optical mapping of late myocardial infarction in rats. Am J Physiol Heart Circ Physiol 2006;290:H1298–306.
[25] Laurita KR, Chuck ET, Yang T, Dong WQ, Kuryshev YA, Brittenham GM, Rosenbaum DS, Brown AM. Optical mapping reveals conduction slowing and impulse block in iron-overload cardiomyopathy. J Lab Clin Med 2003;142:83–9. [26] Chang M, Emokpae R, Zhang Y, Tung L, Marban E, Abraham RA. Coculture of mesenchymal stem cells and neonatal rat ventricular myocytes produces an arrhythmic substrate. Heart Rhythm 2005;S48 [Abstract]. [27] Laurita KR, Girouard SD, Rosenbaum DS. Modulation of ventricular repolarization by a premature stimulus. Role of epicardial dispersion of repolarization kinetics demonstrated by optical mapping of the intact guinea pig heart. Circ Res 1996;79:493–503. [28] Eloff BC, Lerner DL, Yamada KA, Schuessler RB, Saffitz JE, Rosenbaum DS. High resolution optical mapping reveals conduction slowing in connexin43 deficient mice. Cardiovasc Res 2001;51:681–90. [29] Abraham MR, Henrikson CA, Tung L, Chang MG, Aon M, Xue T, et al. Antiarrhythmic engineering of skeletal myoblasts for cardiac transplantation. Circ Res 2005;97:159–67. [30] Krucoff MW, Crater S, Taylor DA, Soliman AM, Morimoto Y. Cell location may be a primary determinant of safety after myoblast transplantation into the infarcted heart. JACC 2004;43:15. [31] Downey JM, Measuring infarct size by the tetrazolium method. Available at: http://www.usouthal.edu/ishr/help/ttc. 2005. [32] Girouard SD, Laurita KR, Rosenbaum DS. Unique properties of cardiac action potentials recorded with voltage-sensitive dyes. J Cardiovasc Electrophysiol 1996;7:1024–38. [33] Gaudesius G, Miragoli M, Thomas SP, Rohr S. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003;93:421–8. [34] Delorme B, Dahl E, Jarry-Guichard T, Briand J-P, Willecke K, Gros D, Théveniau-Ruissy M. Expression pattern of connexin gene products at the early developmental stages of the mouse cardiovascular system. Circ Res 1997;81:423–37. [35] Van Kempen MJA, Vermeulen JLM, Moorman AFM, Gros D, Paul DL, Lamers WH. Developmental changes of connexin40 and connexin43 mRNA distribution patterns in the rat heart. Cardiovasc Res 1996;32: 886–900. [36] Alcoléa S, Théveniau-Ruissy M, Jarry-Guichard T, Marics I, Tzouanacou E, Chauvin JP, et al. Downregulation of connexin 45 gene products during mouse heart development. Circ Res 1999;84:1365–79. [37] Tang YL, Zhao Q, Qin X, Shen L, Cheng L, Ge J, et al. Paracrine action enhances the effects of autologous mesenchymal stem cell transplantation on vascular regeneration in rat model of myocardial infarction. Ann Thorac Surg 2005;80:229–36.
Original article
Stem cell therapy enhances electrical viability in myocardial infarction William R. Mills a , Niladri Mal b , Matthew J. Kiedrowski b , Ryan Unger b , Farhad Forudi b , Zoran B. Popovic c , Marc S. Penn b,c,d , Kenneth R. Laurita a,d,⁎ a
The Heart and Vascular Research Center, MetroHealth Campus, Case Western Reserve University, Cleveland, OH 44109, USA b Department of Cell Biology, Cleveland, OH 44109, USA c Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, OH 44109, USA d The Center for Stem Cell and Regenerative Medicine, Cleveland, OH 44109, USA Received 1 May 2006; received in revised form 12 September 2006; accepted 20 September 2006 Available online 27 October 2006
Abstract Clinical studies suggest increased arrhythmia risk associated with cell therapy for myocardial infarction (MI); however, the underlying mechanisms are poorly understood. We hypothesize that the degree of electrical viability in the infarct and border zone associated with skeletal myoblast (SKMB) or mesenchymal stem cell (MSC) therapy will determine arrhythmia vulnerability in the whole heart. Within 24 h of LAD ligation in rats, 2 million intramyocardially injected SKMB (n = 6), intravenously infused MSC (n = 7), or saline (n = 7) was administered. One month after MI, cardiac function was determined and novel optical mapping techniques were used to assess electrical viability and arrhythmia inducibility. Shortening fraction was greater in rats receiving SKMB (17.8% ± 5.3%, p = 0.05) or MSC (17.6% ± 3.0%, p < 0.01) compared to MI alone (10.1% ± 2.2%). Arrhythmia inducibility score was significantly greater in SKMB (2.8 ± 0.2) compared to MI (1.4 ± 0.5, p = 0.05). Inducibility score for MSC (0.6 ± 0.4) was significantly lower than SKMB (p = 0.01) and tended to be lower than MI. Optical mapping revealed that MSC therapy preserved electrical viability and impulse propagation in the border zone, but SKMB did not. In addition, injected SKMBs were localized to discrete cell clusters where connexin expression was absent. In contrast, infused MSCs engrafted in a more homogeneous pattern and expressed connexin proteins. Even though both MSC and SKMB therapy improved cardiac function following MI in rat, SKMB therapy significantly increased arrhythmia inducibility while MSC therapy tended to lower inducibility. In addition, only MSC therapy was associated with enhanced electrical viability, diffuse engraftment, and connexin expression, which may explain the differences in arrhythmia inducibility. © 2006 Elsevier Inc. All rights reserved. Keywords: Arrhythmia; Electrophysiology; Heart failure; Skeletal myoblasts; Mesenchymal stem cells
1. Introduction Myocardial infarction (MI) is a leading cause of ventricular arrhythmias and sudden cardiac death. Cell therapy for MI using skeletal myoblasts (SKMB) or bone marrow-derived cells has been shown to improve cardiac function [1–6], perfusion [7], symptoms [8], and decrease infarct size [6]. While cell therapy for MI shows promise, some evidence suggests that such therapy is unable to improve electrical function and may even ⁎ Corresponding author. Rammelmamp 6th Floor, Case Western Reserve University, MetroHealth Campus, 2500 MetroHealth Drive, Cleveland, OH 44109, USA. Tel.: +1 216 778 7342; fax: +1 216 778 1261. E-mail address: [email protected] (K.R. Laurita). 0022-2828/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.yjmcc.2006.09.011
be proarrhythmic [9]. Early clinical and experimental studies of SKMB therapy have shown a concerning incidence of ventricular arrhythmias [10–12]. Consistent with this clinical observation, basic studies have shown that SKMB do not electrically couple with native myocardium in vivo [13] and when injected intramyocardially tend to cluster near injection sites [14,15], both of which may be the basis for arrhythmias reported in clinical studies. In contrast, recent studies using bone marrow derived cell therapy [5,12,16] have not reported a significant incidence of arrhythmias. Mesenchymal stem cells (MSC) can be delivered by intravenous infusion soon after MI and have been shown to form gap junctions with host cardiomyocytes in vivo [17,18]. Thus, MSC may enhance electrical viability of myocardium damaged by MI.
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Despite the rapidity in which cellular therapy for MI has progressed from the laboratory to the clinic, many fundamental questions regarding the electrophysiological and arrhythmic consequences are unresolved. Importantly, it is unknown whether cell therapy can enhance electrical viability of damaged myocardium and reduce the risk of post-MI arrhythmias. In this report, we utilize novel optical mapping techniques to determine the electrophysiological consequence of cell therapy in the intact heart, as commonly tested in clinical and experimental studies. In rat hearts with MI, SKMB were directly injected into the infarct and infarct border zone, and MSCs were delivered by intravenous infusion. We tested the hypothesis that the degree of electrical viability in the infarct and border zone associated with SKMB or MSC therapy will determine arrhythmia vulnerability in the whole heart. 2. Materials and methods 2.1. Experimental model This investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication No 85-23, Revised 1996) and was approved by the IACUC of both Case Western Reserve University and the Cleveland Clinic Foundation. To create a model of MI, we performed permanent ligation of the left anterior descending coronary artery (LAD) in rats as described previously [19]. Four groups of rat models were utilized in this study: normals (n = 5), control MI (n = 7), MI + SKMB (n = 6), and MI + MSC (n = 7). The methods used to create these groups are described below. 2.2. SKMB procedure SKMB were isolated and prepared as described previously [20]. SKMB were labeled with DiI (1,1′-dioctadecyl 3,3,3,3′tetra-methylind-carbocyanine percolate) before intramyocardial injection (see below). SKMB were injected into the lateral aspects of the infarct zone approximately 5 min after LAD ligation. A total of five injections (40 μL each containing ∼ 0.4 × 106 cells) were made, two each along the right and left borders of the infarct zone, and one near the apex of the heart. 2.3. MSC procedure Rat bone marrow was isolated by flushing the femurs with 0.6 ml DMEM (GIBCO, Invitrogen, Carlsbad, CA). Clumps of bone marrow were gently minced with an 18-gauge needle. Cells were separated by a Percoll density gradient and then centrifuged for 10 min at 260 G and washed with three changes of PBS with 100 U/ml penicillin/100 g/ml streptomycin (Invitrogen, Carlsbad, CA). The washed cells were then resuspended and plated in DMEM-LG (GIBCO, Invitrogen, Carlsbad, CA) with 10% FBS and 1% antibiotic and antimycotic (GIBCO, Invitrogen, Carlsbad, CA). The
305
cells were incubated at 37 °C. Non-adherent cells were removed by replacing the medium after 3 days. Cultures were refed every 3–4 days. When cultures became 70% confluent, adherent cells were detached following incubation with 0.05% trypsin and 2 mM EDTA (Invitrogen, Carlsbad, CA) for 5 min and subsequently passaged. MSC Cultures were then depleted of CD45+ and CD34+ cells simultaneously by negative selection using 10 μl each of primary PE-conjugated mouse anti-rat CD45 (BD Biosciences, San Diego, CA) and CD34 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) per 106 cells using the EasySep PE selection kit according to the manufacturer's instruction (Stem Cell technologies) to prevent non-specific selection of endothelial progenitor cells (EPC), monocytes, and macrophages. In preceding experiments, confluent cells were labeled with green fluorescent protein (GFP, see below), passaged and plated out at 1:4 to 1:5 dilutions until passage 11. In vitro differentiation assays confirmed that the GFP-tagged MSC had retained the potential for adipogenic, osteogenic, and chondrogenic differentiation. Two million labeled cells (GFP expressing MSC) harvested in 200 μl of PBS were infused via tail vein 24 h after MI. 2.4. GFP labeling of MSC We implemented a third-generation vesicular stomatitis virus glycoprotein (VSV-G) pseudotyped lentivirus expressing enhanced green fluorescent protein (EGFP) reporter gene and zeocin resistance gene driven by the cytomegalovirus (CMV) and the simian virus 40 (SV40) promoter respectively. The lentivirus was made using a four plasmid vector system. MSC were transduced twice for 8 h with purified lentivirus in the presence of 8 μg/ml of polybrene at a multiplicity of infection (MOI) of 30. The media were changed 72 h post-transfection and replaced with regular media containing zeocin (400 μg/ ml). Only cells that incorporated the viral genome, including the zeocin resistance gene, survived. Transfection of MSC with our protocol resulted in expression of EGFP 36 h later in 30% of the cells. FACS analysis to sort GFP expressing cells resulted in an enriched population with ∼ 90–95% of the cells being GFP+. MSC were then expanded using standard techniques. 2.5. DiI labeling of SKMB Because Dil does not spread from labeled to unlabeled cells, and it does not appear to have any adverse effects on living cells, we used Dil to trace transplanted SKMB and to obtain critical information about location, histology, and number of transplanted cells [14]. In short, we trypsinized, washed, and resuspended SKMB in serum-free culture media at a density of 1 × 106 /ml 5 μL (Molecular Probes, Carlsbad, CA). Cells were incubated for 20 min in a dark CO2 incubator at 37 °C. Cells were washed and resuspended in warm media three times. Before transplantation, the cells were pelleted and resuspended in warmed PBS to a concentration of 2 × 106 cells/200 μL.
306
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
2.6. Echocardiography Echocardiography was performed within 1 week of LAD ligation to confirm MI, and repeated 1 month later to assess ventricular function. Two-dimensional echocardiography was performed using a 15-MHz linear-array transducer interfaced with an Acuson Sequoia C256 (Siemens Medical Solutions, Malvern, PA, USA) as described previously [19,22]. Animals were lightly sedated with ketamine (50 mg/kg i.p.) for each echocardiogram. The presence of significant anterior MI was confirmed by anterior wall akinesis present in at least two adjacent segments on two-dimensional echocardiography [23]. As a measure of left ventricular function, the shortening fraction was calculated from M-mode recordings. Dimensions were measured between the anterior wall and the posterior wall from the short-axis view at the level of the papillary muscles. 2.7. Optical mapping Optical mapping was performed in Langendorff-perfused rat hearts 1 month after MI as described recently [24]. Normal (n = 5), control MI (n = 7), MI + MSC (n = 7), and MI + SKMB (n = 6) rats were injected with heparin 300 U i.p. 15 min before the heart was removed. Rats were then anesthetized with sodium pentobarbital 50 mg/kg i.p. After attainment of adequate anesthesia, the heart was immediately excised and arrested by direct aortic perfusion of cold cardioplegia solution (10 ml). Extraneous tissue was carefully dissected away from the heart, which was then placed in a beaker containing 20 ml oxygenated ice-cold cardioplegia and 180 μM of 4-[β-[2-(di-n-butylamino)6-naphthyl]vinyl]pyridium (Di-4-ANEPPS) for 20 min. The beaker was shaken gently every 60 s to ensure uniform staining of the entire epicardial surface. After completion of dyesuperfusion staining, the aorta was cannulated, and perfusion with 37 °C Tyrode's solution containing (mM) 121.7 NaCl, 25.0 NaHCO3, 2.74 MgSO4, 4.81 KCl, 5.0 dextrose, and 2.0 CaCl2 (pH 7.4) was started at ascending flow rates. Perfusion pressure in the cannula was monitored with a pressure transducer (Harvard Apparatus, Saint Laurant, Quebec, Canada), allowing for a ramp up to a target pressure of 70– 90 mm Hg over a 5-min period. After achievement of the target perfusion pressure, the cannulated heart was immersed in 37 °C Tyrode's solution in a specialized optical recording chamber for the duration of each experiment. ECG, perfusion pressure, and bath temperature were measured continuously during all experiments. The optical mapping system used has been described in detail previously [24,25]. Briefly, action potentials were optically recorded at a magnification of 2.1× from 256 sites within an 8.3 mm × 8.3 mm mapping field (0.52-mm interpixel resolution) on the anterior epicardial surface of the left and right ventricle. Fluorescence was excited with uniform light from a 270-W tungsten-halogen light source (filtered 514 + 20 nm) and transmitted to a 16 × 16-element photodiode array detector through a tandem-lens imaging system (emission filter > 610 nm). Photocurrent from each photodiode underwent current-to-voltage conversion, amplification, and bandpass
filtering (0.1–500 Hz) and was multiplexed and digitized (1000 samples/s per channel) with 12-bit precision. A CCD camera that is optically aligned with the photodiode array was used to obtain visible images of the optical mapping field relative to anatomical landmarks (e.g., infarct scar and border zone). 2.8. Experimental protocol After staining with Di-4 ANEPPS, all hearts were placed against the optical imaging window. CCD images of the epicardial surface of the heart were used to initially position and maintain the location of the heart relative to the mapping field. The transition between non-infarcted (i.e., pink) and infarcted (i.e., white) tissue was used to approximate the location of the infarct border zone. This ensured that non-infarcted, border zone, and infarcted tissue were as close as possible to the center of the mapping field for each heart tested (see Fig. 1A) [24]. Each heart was perfused with Tyrode's plus 20 mM of diacetylmonoxine (DAM) to eliminate motion artifact. All hearts were paced at a steady state cycle length of 150 ms using a silver unipolar electrode placed on the epicardial surface. Optical recordings were made during steady state pacing and during programmed stimulation. All preparations remained viable for the entirety of each optical mapping experiment. Coronary perfusion pressure was maintained at 80 ± 10 mm Hg for all experiments. At the conclusion of optical mapping, hearts were stained with triphenyltetrazolium chloride (TTC) using the method of Downey [31]. Fluorescence microscopy, using specialized di-4-ANEPPS fluorescence filters, was then performed on ventricular cross sections to ensure homogenous optical dye staining of non-infarcted and infarcted myocardium. 2.9. Data analysis Using previously described automated algorithms [27], activation times were assigned for each action potential by identifying the greatest positive change in fluorescence (maximum derivative) during each beat. All assigned times were verified visually by an experienced investigator. From local depolarization times, spatial contour maps (i.e., activation maps) of impulse propagation were calculated and projected onto the epicardial surface of the heart [28]. Impulse block was defined by the absence of an action potential at a recording site, other than from sites corresponding to the corners of the array, where lens vingetting significantly reduced fluorescence intensity independent of tissue viability. To assess electrical viability, conduction velocity and unnormalized optical action potential amplitude were measured in non-infarcted, border zone, and infarcted tissue. Conduction velocity was calculated in the infarct border zone during steady state pacing with respect to fiber orientation using an average of velocity vectors parallel to the direction of conduction, as described previously [25,28]. The unnormalized amplitude of optically recorded action potentials was determined by calculating the difference in fluorescence just before depolarization (during rest) and immediately after depolarization at the time corresponding to
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
the peak of the action potential as described previously in hearts with MI [24]. To determine arrhythmia vulnerability, programmed stimulation was performed with up to 3 premature stimuli, and an inducibility score was assigned to each heart. Inducibility was defined as the ability to provoke non-sustained or sustained (≥ 30 s) ventricular tachycardia or ventricular fibrillation. A score of 0 was assigned if the preparation was not inducible; 1 if inducible with 3 premature stimuli; 2 if inducible with 2 premature stimuli; and 3 if inducible with 1 premature stimulus. The higher the score, the easier it was to induce arrhythmias. The unpaired T test was used to test for statistical
307
significance, except for multiple comparisons, when ANOVA was performed followed by Newman–Keuls post hoc analysis. A p value <0.05 was considered statistically significant. Data are presented as mean ± SE unless otherwise noted. 2.10. Confocal immunofluorescence and histology Immunostaining was performed on all rat hearts following optical mapping as previously described [24]. In short, tissue was fixed in formalin and embedded in paraffin blocks. Antigen retrieval was performed using 10 mM sodium citrate buffer (pH
Fig. 1. Experimental preparation. (A) Photograph of the anterior surface of a Langendorff-perfused rat heart with MI within the optical mapping chamber. Note that the optical mapping area (black box) includes non-infarcted zone (NI), border zone (BZ), and infarct zones (IZ). Shown below are representative H&E images from each group. Tissues sections are mid-ventricular LV cross-sections.
308
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Fig. 2. SKMB (n = 6) and MSC (n = 7) therapy improved cardiac function to a similar extent compared to control MI (n = 7). Shown is a histogram illustrating shortening fractions measured by echocardiography 1 month after MI. Values shown are mean ± SD.
oven, de-paraffinized in xylene, and stained with hematoxylin and eosin (H&E). Photomicrographs of each section were acquired using an ArtixScan 4000tf slide scanner (Microtek, Carson, CA) in conjunction with the PathScan Enabler (Meyer Instruments, Houston, TX) at a resolution of 4000 dpi. Total infarct size was estimated from mid-ventricular LV crosssections using Image-Pro Plus version 5.1 image analysis software as we have previously shown [22]. Total infarct size was quantified by calculating the percent of the LV area with infarcted tissue compare to the total LV area. Total LV area and infarct area were calculated based on total epicardial and endocardial circumference and the epicardial and endocardial circumferential extent of the infarct, respectively. Two independent observers blinded to the study measured the amount of infarct in percent, and their values were then averaged. 3. Results
6.0) heat at 95 °C for 5 min. The buffer was replaced with fresh buffer and reheated for an additional 5 min and then cooled for approximately 20 min. The slides were then washed in deionized water three times for 2 min each. Specimens were then incubated with 1% normal blocking serum in PBS for 60 min to suppress non-specific binding of IgG. Slides were then incubated for 60 min with appropriate primary antibody. Optimal antibody concentration was determined by titration. Slides were then washed with phosphate-buffered saline (PBS) and then incubated for 45 min with corresponding secondary antibody conjugated with Alexa Fluor fluorochrome 488 (green) or 594 (red) diluted to 1.5 μg/ml in PBS with serum and incubated in a dark chamber. In double staining experiment, slides were incubated with another set of primary and secondary antibody in succession. After washing extensively with PBS, coverslips were mounted with aqueous mounting medium Vectashield with DAPI (Vector Laboratories, Burlingame, CA). Primary antibodies: rabbit anti-connexin-43 polyclonal IgG antibody (Santa Cruz Biotechnology, Inc.); rabbit anti-connexin 45 polyclonal IgG antibody (Santa Cruz Biotechnology, Inc.); goat polyclonal anti-connexin-40 IgG antibody (Santa Cruz Biotechnology, Inc.). Secondary antibodies: donkey anti-rabbit IgG Alexa Fluor 488 (Molecular Probes); donkey anti-rabbit IgG Alexa Fluor 594(Molecular Probes); donkey anti-goat IgG Alexa Fluor 488 antibody (Molecular Probes); donkey anti-goat IgG antibody Alexa Fluor 594 (Molecular Probes). An upright spectral laser scanning confocal microscope (Model TCS-SP; Leica Microsystems, Heidelberg, Germany) equipped with blue argon (for DAPI), green argon (for Alexa Fluor 488), and red krypton (for Alexa Fluor 594) laser was used for confocal analysis. Data were collected by sequential excitation to minimize “bleed-through”. Image processing, analysis and the extent of colocalization were evaluated using Leica Confocal software. Optical sectioning was averaged over four frames and the image size was set at 1024 × 1024 pixels. There were no digital adjustments made to the images. In three hearts from each group, mid-ventricular and apical segments that were paraffin-embedded were sectioned (5 μm) and applied to glass slides. Then, the slides were dried in an
In all infarcted rats, echocardiography confirmed the presence of anterior wall infarction and decreased cardiac function. Shown in Fig. 1 (bottom) are representative H&Estained sections used to qualitatively assess infarct size for each group with MI. In control MI, MI + SKMB, and MI + MSC, significant thinning of the injured wall and dilatation of the ventricular cavity is evident compared to a normal heart. The percent of infarcted LV was reduced for MI + SKMB (25% ± 12%, n = 3) and MI + MSC (30% ± 7%, n = 3) compared to control MI (37% ± 16%, n = 3); however, there was no statistical difference among all groups. As shown in Fig. 2, echocardiography demonstrated a significant improvement in shortening fraction in both MI + SKMB (17.8% ± 5.3%, p = 0.05) and MI + MSC (17.6% ± 3.0%, p < 0.01) groups compared to control MI (10.1% ± 2.2%).
Fig. 3. Skeletal myoblast therapy was associated with increased arrhythmia vulnerability. Arrhythmia vulnerability was determined by an arrhythmia inducibility score (see text for details) where the larger number indicates higher arrhythmia inducibility. SKMB (n = 6) was associated with increased vulnerability to ventricular tachycardia compared to normal hearts (n = 5), control MI (n = 7), and MI + MSC (n = 7). MSC therapy was associated with a nonsignificant decrease in arrhythmia vulnerability compared to control MI.
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
During programmed stimulation, ventricular tachycardia was induced in 57% of hearts with MI alone (control MI) compared to 0% in normal hearts. Importantly, arrhythmia inducibility in
309
MI + SKMB (100%) was greater than MI + MSC (29%). Similarly, arrhythmia inducibility score was significantly higher in the MI + SKMB group (2.8 ± 0.2) compared to both control
Fig. 4. MSC, but not SKMB, therapy was associated with enhanced electrical viability in the infarct and border zone as determined by greater action potential amplitude and evidence of impulse propagation (activation maps). Activation maps of impulse propagation and unnormalized optically recorded action potentials obtained from normal (A), control MI (B), MI + SKMB (C), and MI + MSC (D) hearts during optical mapping. The images in each panel show the area mapped in each heart, and the signals beside each activation map were recorded from non-infarcted (site 1), border zone (sites 2 and 3), and infarcted (site 4) tissue. Action potentials are plotted as absolute florescence. For all activation maps, each contour line represents 2 ms. The thick black lines indicate impulse block.
310
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
Fig. 5. Enhanced electrical viability of the infarct border zone of rats treated with mesenchymal stem cells. Shown is a plot of action potential amplitude versus distance across non-infarcted, border zone, and infarcted tissue of normal (n = 5, circle), control MI (n = 7, square), MI + MSC (n = 7, triangle), and MI + SKMB (n = 6, diamond) hearts. The border zone occurs between −1 and +1 mm (corresponding to signals recorded from sites 2 and 3 in each panel of Fig. 4, shown here in parentheses above each point). Asterisks (*) denote statistical significance between MI + MSC, normal, and control MI groups.
MI (1.4 ± 0.5, p = 0.05) and MI + MSC (0.6 ± 0.4, p = 0.01) groups (Fig. 3). Interestingly, the MI + MSC group tended to be less vulnerable to arrhythmias than control MI; however, this
difference did not reach statistical significance. These data suggest that SKMB therapy for MI increases vulnerability to ventricular arrhythmias whereas MSC therapy may reduce arrhythmia vulnerability. To determine electrical viability, optical action potentials were recorded simultaneously from non-infarcted, border zone, and infarcted tissue. Shown in Fig. 4 are representative examples of activation contour maps (constructed from activation times determined at each recording site) and unnormalized action potentials, plotted as absolute fluorescence, recorded from normal, control MI, MI + SKMB, and MI + MSC hearts. The image in each panel shows the location from which action potentials were recorded (right) corresponding to non-infarcted tissue (site 1), border zone (sites 2 and 3), and infarct zone (site 4). The activation map and action potentials shown in Fig. 4A were recorded from a normal heart and demonstrate uniform impulse propagation without evidence of block (thick black lines), in addition to uniform amplitude and normal action potential morphology. In this example, the conduction velocity between sites 2 and 4 is 0.55 m/s. Shown in Fig. 4B is a representative example of impulse propagation and unnormalized action potentials recorded from a control MI
Fig. 6. Representative SKMB (Panel A) and MSC therapy (Panel B) engraftment patterns and connexin expression (Cx40, top; Cx43, middle, Cx45, bottom). Each panel is a confocal immunofluorescent photomicrograph taken at 40× magnification (scale bar 40 μm) of ventricular cross sections from the MI + SKMB (Panel A) and MI + MSC (Panel B) group. Each image roughly corresponds to tissue between the sites labeled as 3 and 4 in each panel in Fig. 4, as indicated by the colocalization of DAPI (blue) and Dil (red).
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
311
Fig. 6 (continued).
heart. The activation map indicates regions of impulse block (thick black line) corresponding to the location of the border zone (between sites 2 and 3). The action potential recorded from non-infarcted tissue (1) was similar to that measured from normal hearts, whereas signals recorded from the border zone (2 and 3) demonstrated an abrupt decrease in amplitude, and a slower depolarization phase. No action potential activity was observed further into the infarct zone (site 4). In this example, the conduction velocity from sites 2 to 3 (i.e., in the border zone) was 0.31 m/s. For MI + SKMB (Fig. 4C), impulse propagation and unnormalized action potentials recorded from non-infarcted tissue (1) were similar to that measured from normal rats, whereas signals recorded from the border zone (2 and 3) were of low amplitude and had a slower depolarization phase. In this example, the conduction velocity was 0.21 m/s. No action potential activity was observed at site 4 (similar to control MI). Importantly, impulse block and significant conduction slowing was observed in the infarct border zone of MI + SKMB (0.21 ± 0.05 m/s) compared to control MI (0.33 ± 0.05 m/s; p = 0.02). Shown in Fig. 4D are impulse propagation and unnormalized action potentials recorded from a heart with MI + MSC. The activation map reveals impulse propagation beyond the border zone (sites 3 and 4) towards the center of the infarct and a smaller region of conduction block compared to MI + SKMB and control MI. The action potential recorded from non-infarcted tissue (1) was
similar to action potentials recorded in normal hearts. Action potentials recorded from the border zone (sites 2 and 3) were smaller than normal, but not as small as those recorded from the border zone in control MIs (Fig. 4B). In addition, unlike control MIs, an action potential was measured at site 4 (furthest point into the infarct zone) with an activation time that was delayed compared to that measured at site 3, indicating some degree of impulse propagation in the infarct zone (as also indicated in the activation map). In this example, the conduction velocity from sites 1 to 4 was (0.24 m/s). Overall, mean conduction velocity in the border zone for MI + MSC (0.24 ± 0.07) was slower but not significantly different from control MI (0.33 ± 0.05, p = NS). Enhanced electrical viability was observed only in hearts that received MSC therapy (4 of 7). Shown in Fig. 5 are average action potential amplitudes in non-infarcted tissue (site 1), border zone (sites 2 and 3), and infarct zone (site 4), recorded from normal hearts and hearts with MI, MI + MSC, and MI + SKMB. In each heart, action potential amplitude was normalized to that recorded from noninfarcted tissue (site 1). In normal hearts (circles), action potential amplitude decreased by an average of 16% ± 6.8% from sites 1 to 4. This decrease likely resulted from a slight fall off in excitation light intensity or lens vingetting. In hearts with control MI (squares), action potential amplitude furthest into the infarct zone (site 4) decreased by 82% ± 6.9%. In contrast, in hearts with MI + MSC (triangles), action potential amplitude
312
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
furthest into the infarct zone decreased by only 45% ± 20.8% and was significantly greater that that compared to control MI (p < 0.05). In MI + SKMB hearts (diamonds), action potential amplitude furthest into the infarct decreased by 86% ± 3.0% and was similar to hearts with control MI. These data suggest that MSC therapy is associated with enhanced electrical viability of the infarct and border zone, whereas SKMB is not. Shown in Fig. 6 are representative confocal immunofluorescent photomicrographs taken at 40× magnification (scale bar 40 μm) from ventricular cross sections in the, MI + SKMB (Fig. 6A) and MI + MSC (Fig. 6B) groups. In each panel, expression of Cx40 (first row), Cx43 (middle row), and Cx45 (bottom row) are shown. Each image corresponds to an area of tissue near sites 3 or 4 in each panel shown in Fig. 4. As demonstrated in Fig. 6A, intramyocardially injected SKMB (red) colocalized with DAPI (blue), indicating that DiI-labeled cells survived and tended to engraft heterogeneously in cell clusters. In addition, expression of Cx40 (top row, green), Cx43 (middle row, green), and Cx45 (bottom row, green) were absent where Dil-positive cells were located, indicating a lack of connexin expression associated with SKMB therapy. In contrast, intravenously infused MSCs (Fig. 6B, green) were homogeneously distributed. Interestingly, MSC engraftment was associated with the expression of all connexin proteins tested, as indicated by the colocalization of connexin protein (red) and MSCs (green) in the merged images. In summary, these data suggest that in contrast to SKMB therapy, the homogeneous engraftment and connexin expression associated with MSC therapy led to enhanced electrical viability and reduced arrhythmia inducibility. 4. Discussion Despite the incidence of ventricular arrhythmias observed in early clinical trials [10,11], there have been very few reports that have systematically evaluated the electrophysiological and arrhythmic consequence of cell therapy for MI in the whole heart. In this report, we have shown that (1) although SKMB therapy improves cardiac function, it increases arrhythmia vulnerability and does not enhance electrical viability in the infarct or border zone; and (2) MSC therapy improves cardiac function and enhances electrical viability in the infarct and border zone, which may explain why arrhythmia vulnerability was similar or, possibly, lower than control MI.
SKMB and cardiomyocytes. Moreover, Krucoff et al. [30] have shown that arrhythmia vulnerability is greater when cells are injected into the border zone compared to the central zone of an infarct. It is possible that DiI-labeled cells, which lacked the expression of connexin proteins, formed in clusters near the injection sites, producing electrical barriers to conduction, as we have previously shown when SKMB are injected into normal myocardium [14]. MI + MSC hearts were significantly less vulnerable to arrhythmias than MI + SKMB hearts, and trended towards being less vulnerable to arrhythmias than control MIs. The low arrhythmia vulnerability we observed in MI + MSC hearts is also consistent with the low incidence of sudden death and ventricular arrhythmias reported in studies of bone-marrow derived cell therapies [5,12,16]. Interestingly, MI + MSC hearts tended to exhibit conduction slowing in the border zone as did MI + SKMB (albeit not statistically significant). This finding is consistent with recent work by Chang et al. [26], who showed conduction slowing in cocultures of MSC and cardiomyocytes. Therefore, slow conduction in the border zone may not be a sole determinant of arrhythmia inducibility as measured in the whole heart. On the other hand, enhanced electrical viability (see next section) and the expression of connexin protein in the border zone of MI + MSC hearts may explain the observed reduction in arrhythmia vulnerability. Our study suggests that the delivery method of cell therapy (SKMB injection versus MSC IV infusion) may have affected arrhythmia vulnerability. For example, the clustering of directly injected SKMB may have formed a local zone of slow conduction [14], which is a substrate for reentrant excitation. In contrast, MSCs delivered by IV engrafted in a more homogeneous pattern (Fig. 6). Would SKMBs engraft in a more homogeneous pattern if delivered by IV infusion? Possibly, but it is unlikely that SKMB would engraft in the heart at all when delivered by IV infusion. Would MSCs delivered by direct intramyocardial injection result in an engraftment pattern similar to SKMB? Possibly; but unlike SKMBs, MSCs should express connexin proteins. Moreover, Fernandes et al. [12] demonstrated that MSCs delivered by direct injection also have a much lower rate of arrhythmia inducibility compared SKMB, suggesting that the cell type may be just as important as the method of delivery. Additional studies are required to determine the optimal delivery method for each cell type.
4.1. Arrhythmia vulnerability 4.2. Electrical viability Strikingly, we report that 100% of MI + SKMB hearts had inducible ventricular arrhythmias. This uniform inducibility observed in the MI + SKMB group is consistent with that from Fernandes et al. [12] and a recent study, in which we also reported 100% inducibility in rats with chronic MI that received SKMB 2 months after MI [21]. The increase in arrhythmia vulnerability observed in MI + SKMB hearts may be related to reduced conduction velocity in the infarct border zone compared to control MI. These results are consistent with those reported by Abraham et al. [29], who showed slowing of conduction and inducible sustained reentry in cocultures of
To determine electrical viability associated with cell therapy, we measured optical action potential amplitude and impulse propagation within the infarct border zone. In MI + SKMB hearts, we observed a fall off in action potential amplitude at the border zone and no evidence of impulse propagation into the infarct zone, a pattern that was undistinguishable from control MI. This finding is consistent with the work of Reinecke et al. [13], who showed that SKMB do not electrically integrate into the myocardium. In contrast, the border zone of MI + MSC hearts was electrically more viable than control MI, as
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
evidenced by action potential and impulse propagation deeper into the infarct zone. The optical action potential amplitude, however, must be interpreted with caution because it depends on several factors such as dye staining and excitation light, in addition to the number of electrically viable cells in a given area. To address these factors, we have shown previously that the voltage-sensitive dye superfusion technique homogenously stains non-infarcted and infarcted myocardium [24]. In addition, we showed that over a relatively short distance (∼ 6 mm) action potential amplitudes were very similar in normal hearts (Fig. 5, circles), ensuring uniform excitation light intensity and dye staining. However, 95% of Di-4-ANEPPS fluorescence is measured from a depth of < 0.5 mm [32]. Therefore, our previous study [24] and the current study demonstrate that optical action potential amplitude is a reasonable approximation of relative electrical viability within the outermost epicardial layer. The increase in electrical viability may have resulted from the homogenous distribution of viable MSCs throughout the border zone and the expression of connexin proteins (Fig. 6B). Valiunas et al. [17] and Potapova et al. [18] have shown that low frequency gap junction formation between MSC and cardiomyocytes can occur. It is possible that MSC therapy may have provided passive coupling between cardiomyocytes, as has been shown with fibroblasts [33]. Therefore, enhanced electrical coupling between MSC and native cardiomyocytes may have, at least in part, led to the increase in electrical viability we observed. As a result, the electrical size of the infarct might decrease and reduce the path length for reentrant excitation and thus make arrhythmias less likely. Further studies are needed to test this possibility. The presence of all three connexin proteins may reflect MSCs at various stages of differentiation. It is well known that the developing mammalian heart expresses varying levels of connexins during embryological development, and the expression of connexins in the present study may be analogous to the ubiquity of these connexins during development [34–36]. It is possible that some of our results may be caused by a paracrine effect. For example, MSC therapy has been previously associated with increased angiogenic factors and a down-regulation of proapoptotic proteins [37]. We did not observe cardiogenesis based on differentiation of injected cells. Thus, it seems likely that the improved function (Fig. 2) observed in response to cell therapy was due to paracrine effects that might be the result of similar or distinct pathways between MSC and SKMB. 4.3. Clinical implications SKMB therapy was associated with improved cardiac function but did not enhance electrical viability. Moreover, SKMB therapy was associated with a significant increase in vulnerability to ventricular arrhythmias and slow conduction in the infarct border zone compared to control MI. These findings may explain the reported high incidence of arrhythmias observed in early clinical trials of SKMB therapy. In contrast, therapy with intravenous MSC led to improved cardiac
313
function, enhanced electrical viability of the infarct and border zone, and was not associated with an increase in vulnerability to ventricular arrhythmias. Overall, these data suggest that the electrophysiological and arrhythmic consequence of cell therapy for MI depends on changes in electrical viability of the infarct and border zone of MI as determined by the cell type and delivery method. If these data are corroborated by future basic and clinical studies, MSC could emerge as a valuable therapy for MI-associated sudden cardiac death in patients. Acknowledgments This study was supported by NIH grants HL68877 (KRL) and HL74400 (MSP), and grants from The Wilson Foundation (MSP), Shalom Foundation (MSP), Ohio Valley Affiliate of the American Heart Association (KRL), and the Biological Research Technology Transfer Fund of the State of Ohio (MSP and KRL). This work was performed during Dr. Mills' tenure as the Michael Bilitch Fellow in Cardiac Pacing and Electrophysiology of the NASPE-Heart Rhythm Society. References [1] Taylor DA, Atkins BZ, Hungspreugs P, Jones TR, Reedy MC, Hutcheson KA, et al. Regenerating functional myocardium: improved performance after skeletal myoblast transplantation. Nat Med 1998;4:929–33. [2] Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, et al. Bone marrow cells regenerate infarcted myocardium. Nature 2001;410: 701–5. [3] Yau TM, Tomita S, Weisel RD, Jia ZQ, Tumiati LC, Mickle DAG, et al. Beneficial effect of autologous cell transplantation on infarcted heart function: comparison between bone marrow stromal cells and heart cells. Ann Thorac Surg 2003;75:169–76. [4] Menasche P, Hagege AA, Scorsin M, Pouzet B, Desnos M, Duboc D, et al. Myoblast transplantation for heart failure. Lancet 2001;357:279–80. [5] Assmus B, Schachinger V, Teupe C, Britten M, Lehmann R, Dobert N, et al. Transplantation of progenitor cells and regeneration enhancement in acute myocardial infarction (TOPCARE-AMI). Circulation 2002;106: 3009–17. [6] Strauer BE, Brehm M, Zeus T, Kostering M, Hernandez A, Sorg RV, et al. Repair of infarcted myocardium by autologous intracoronary mononuclear bone marrow cell transplantation in humans. Circulation 2002;106: 1913–8. [7] Stamm C, Westphal B, Kleine HD, Petzsch M, Kittner C, Klinge H, et al. Autologous bone-marrow stem-cell transplantation for myocardial regeneration. Lancet 2003;361:45–6. [8] Tse HF, Kwong YL, Chan JK, Lo G, Ho CL, Lau CP. Angiogenesis in ischaemic myocardium by intramyocardial autologous bone marrow mononuclear cell implantation. Lancet 2003;361:47–9. [9] Makkar RR, Lill M, Chen PS. Stem cell therapy for myocardial repair: is it arrhythmogenic? J Am Coll Cardiol 2003;42:2070–2. [10] Menasche P, Hagege AA, Vilquin JT, Desnos M, Abergel E, Pouzet B, et al. Autologous skeletal myoblast transplantation for severe postinfarction left ventricular dysfunction. J Am Coll Cardiol 2003;41:1078–83. [11] Smits PC, van Geuns RJ, Poldermans D, Bountioukos M, Onderwater EE, Lee CH, et al. Catheter-based intramyocardial injection of autologous skeletal myoblasts as a primary treatment of ischemic heart failure: clinical experience with six-month follow-up. J Am Coll Cardiol 2003;42:2063–9. [12] Fernandes S, Amirault JC, Lande G, Nguyen JM, Forest V, Bignolais O, et al. Autologous myoblast transplantation after myocardial infarction increases the inducibility of ventricular arrhythmias. Cardiovasc Res 2006;69:348–58.
314
W.R. Mills et al. / Journal of Molecular and Cellular Cardiology 42 (2007) 304–314
[13] Reinecke H, MacDonald GH, Hauschka SD, Murry CE. Electromechanical coupling between skeletal and cardiac muscle. Implications for infarct repair. J Cell Biol 2000;149:731–40. [14] Fouts KH, Fernandes B, Mal N, Liu JQ, Laurita KR. Electrophysiological consequence of skeletal myoblast transplantation in normal and infarcted canine myocardium. Heart Rhythm 2006;3:452–61. [15] Thompson RB, Emani SM, Davis BH, van den Bos EJ, Morimoto Y, Craig D, Glower D, et al. Comparison of intracardiac cell transplantation: autologous skeletal myoblasts versus bone marrow cells. Circulation 2003;108(Suppl 1):II264–71. [16] Wollert KC, Meyer GP, Lotz J, Ringes-Lichtenberg S, Lippolt P, Breidenbach C, et al. Intracoronary autologous bone-marrow cell transfer after myocardial infarction: the BOOST randomised controlled clinical trial. Lancet 2004;364:141–8. [17] Valiunas V, Doronin S, Valiuniene L, Potapova I, Zuckerman J, Walcott B, et al. Human mesenchymal stem cells make cardiac connexins and form functional gap junctions. J Physiol 2004;555:617–26. [18] Potapova I, Plotnikov A, Lu Z, Danilo Jr P, Valiunas V, Qu J, et al. Human mesenchymal stem cells as a gene delivery system to create cardiac pacemakers. Circ Res 2004;94:952–9. [19] Askari AT, Unzek S, Popovic ZB, Goldman CK, Forudi F, Kiedrowski M, et al. Effect of stromal-cell-derived factor 1 on stemcell homing and tissue regeneration in ischaemic cardiomyopathy. Lancet 2003;362:697–703. [20] Jain M, DerSimonian H, Brenner DA, Ngoy S, Teller P, Edge AS, et al. Cell therapy attenuates deleterious ventricular remodeling and improves cardiac performance after myocardial infarction. Circulation 2001;103: 1920–7. [21] Deglurkar I, Mal N, Mills W, Popovic ZB, Kiedrowski MJ, Forudi F, McCarthy PM, Blackstone EH, Laurita KR, Penn MS. Mechanical and electrical effects of cell therapy are independent. Hum Gene Ther in press. [22] Askari A, Unzek S, Goldman CK, Ellis SG, Thomas JD, DiCorleto PE, et al. Cellular, but not direct, adenoviral delivery of vascular endothelial growth factor results in improved left ventricular function and neovascularization in dilated ischemic cardiomyopathy. J Am Coll Cardiol 2004;43: 1908–14. [23] Morgan EE, Faulx MD, McElfresh TA, Kung TA, Zawaneh MS, Stanley WC, et al. Validation of echocardiographic methods for assessing left ventricular dysfunction in rats with myocardial infarction. Am J Physiol Heart Circ Physiol 2004;287:H2049–53. [24] Mills WR, Mal N, Forudi F, Popovic ZB, Penn MS, Laurita KR. Optical mapping of late myocardial infarction in rats. Am J Physiol Heart Circ Physiol 2006;290:H1298–306.
[25] Laurita KR, Chuck ET, Yang T, Dong WQ, Kuryshev YA, Brittenham GM, Rosenbaum DS, Brown AM. Optical mapping reveals conduction slowing and impulse block in iron-overload cardiomyopathy. J Lab Clin Med 2003;142:83–9. [26] Chang M, Emokpae R, Zhang Y, Tung L, Marban E, Abraham RA. Coculture of mesenchymal stem cells and neonatal rat ventricular myocytes produces an arrhythmic substrate. Heart Rhythm 2005;S48 [Abstract]. [27] Laurita KR, Girouard SD, Rosenbaum DS. Modulation of ventricular repolarization by a premature stimulus. Role of epicardial dispersion of repolarization kinetics demonstrated by optical mapping of the intact guinea pig heart. Circ Res 1996;79:493–503. [28] Eloff BC, Lerner DL, Yamada KA, Schuessler RB, Saffitz JE, Rosenbaum DS. High resolution optical mapping reveals conduction slowing in connexin43 deficient mice. Cardiovasc Res 2001;51:681–90. [29] Abraham MR, Henrikson CA, Tung L, Chang MG, Aon M, Xue T, et al. Antiarrhythmic engineering of skeletal myoblasts for cardiac transplantation. Circ Res 2005;97:159–67. [30] Krucoff MW, Crater S, Taylor DA, Soliman AM, Morimoto Y. Cell location may be a primary determinant of safety after myoblast transplantation into the infarcted heart. JACC 2004;43:15. [31] Downey JM, Measuring infarct size by the tetrazolium method. Available at: http://www.usouthal.edu/ishr/help/ttc. 2005. [32] Girouard SD, Laurita KR, Rosenbaum DS. Unique properties of cardiac action potentials recorded with voltage-sensitive dyes. J Cardiovasc Electrophysiol 1996;7:1024–38. [33] Gaudesius G, Miragoli M, Thomas SP, Rohr S. Coupling of cardiac electrical activity over extended distances by fibroblasts of cardiac origin. Circ Res 2003;93:421–8. [34] Delorme B, Dahl E, Jarry-Guichard T, Briand J-P, Willecke K, Gros D, Théveniau-Ruissy M. Expression pattern of connexin gene products at the early developmental stages of the mouse cardiovascular system. Circ Res 1997;81:423–37. [35] Van Kempen MJA, Vermeulen JLM, Moorman AFM, Gros D, Paul DL, Lamers WH. Developmental changes of connexin40 and connexin43 mRNA distribution patterns in the rat heart. Cardiovasc Res 1996;32: 886–900. [36] Alcoléa S, Théveniau-Ruissy M, Jarry-Guichard T, Marics I, Tzouanacou E, Chauvin JP, et al. Downregulation of connexin 45 gene products during mouse heart development. Circ Res 1999;84:1365–79. [37] Tang YL, Zhao Q, Qin X, Shen L, Cheng L, Ge J, et al. Paracrine action enhances the effects of autologous mesenchymal stem cell transplantation on vascular regeneration in rat model of myocardial infarction. Ann Thorac Surg 2005;80:229–36.