Cell distribution after intracoronary bone marrow stem cell delivery in damaged and undamaged myocardium: implications for clinical trials

Introduction Early randomized clinical trials of autologous bone marrow cardiac stem cell therapy have reported contradictory results highlighting the need for a better evaluation of protocol designs. This study was designed to quantify and compare whole body and heart cell distribution after intracoronary or peripheral intravenous injection of autologous bone marrow mononuclear cells in a porcine acute myocardial infarction model with late reperfusion. Methods Myocardial infarction was induced using balloon inflation in the left coronary artery in domestic pigs. At seven days post-myocardial infarction, 1 × 10(8) autologous bone marrow mononuclear cells were labeled with fluorescent marker and/or 99mTc radiotracer, and delivered using intracoronary or peripheral intravenous injection (leg vein). Results Scintigraphic analyses and Υ-emission radioactivity counting of harvested organs showed a significant cell fraction retained within the heart after intracoronary injection (6 ± 1.7% of injected radioactivity at 24 hours), whereas following peripheral intravenous cell injection, no cardiac homing was observed at 24 hours and cells were mainly detected within the lungs. Importantly, no difference was observed in the percentage of retained cells within the myocardium in the presence or absence of myocardial infarction. Histological evaluation did not show arterial occlusion in both animal groups and confirmed the presence of bone marrow mononuclear cells within the injected myocardium area. Conclusions Intravenous bone marrow mononuclear cell injection was ineffective to target myocardium. Myocardial cell distribution following intracoronary injection did not depend on myocardial infarction presence, a factor that could be useful for cardiac cell therapy in patients with chronic heart failure of non-ischemic origin or with ischemic myocardium without myocardial infarction.

Our aim was, with an experimental design of acute myocardial infarction (MI) and autologous cardiac cell therapy that closely match clinical trials [2,3] to analyze and quantify whole body and heart distribution of injected autologous bone marrow-mononuclear cells (BM-MNCs). We also evaluated the optimal route of delivery (intracoronary versus peripheral intravenous injection).

Animal procedures
Animal procedures were approved by the Institutional Committee for Use and Care of Laboratory Animals at the Veterinary School of Nantes and conform to the Guide for the Care and Use of Laboratory Animals (NIH Publication No.85-23, revised 1996). Forty-six domestic pigs (35 to 40 kg) were included in the study. For MI induction and cell injection, anesthesia was induced by intramuscular injection of ketamine (10 mg/kg)/xylazine (2 mg/kg) and maintained with mechanical tracheal venti la tion using inhaled isofl urane (2%). All animals received subcutaneously morphine (0.2 mg/kg) prior and after all procedures.

Experimental myocardial infarction
We used the MI experimental model initially described by Suzuki et al [6], with slight modifi cations. Five days prior MI induction, animals received β-blocker carazolol (20 μg/kg) twice per day. Propranolol (0.05 mg/kg) was injected intravenously just before MI induction procedure to prevent arrhythmia. After the placement of a 6F sheath in the femoral artery, an extra-back up 6F guiding catheter (Boston Scientifi c Scimed, Inc, Boston, USA) was placed under fl uoroscopy into the left main coronary artery. A 3 mm over the wire balloon catheter was placed either into the left anterior descending coronary artery distal to the fi rst diagonal branch or into the left circumfl ex. After balloon infl ation a coronary thrombus was induced by injections below the balloon of thrombin (900 UI, Sigma Aldrich Corporation, Saint Quentin Fallavier, France) and fi brinogen (5 mg, Sigma). Th e balloon was defl ated two to three minutes after injections. Coronary occlusion was checked by angiogram, and MI was confi rmed by ECG alteration and by plasma creatine kinase elevation 24 hours after MI induction (Refl otron, Roche Diagnostics, Meylan, France).

Preparation and cryopreservation of autologous bone marrow mononuclear cells (BM-MNCs)
Five days after MI, a total of 100 mL of bone marrow was aspirated into heparin-treated Bone Marrow Aspiration/ Intraosseous Infusion needle (15Ga, 10mm, Cardinal Health France 205 S.A.S, Châteaubriand, France) from the iliac crests and shoulder blades under general anaesthesia. Five points of puncture were performed at each site of 5 mL each. Th e duration of the procedure was 15 to 20 minutes. BM-MNCs were isolated using Ficoll gradient (Eurobio Les Ulis, France), and a total of 244 x 10 6 ± 33 x 10 6 BM-MNCs was obtained.
BM-MNCs were cryopreserved in foetal calf serum 10% DMSO. Cryo-preservation for two days avoided two general anaesthesias the same day in seriously disabled animals and decreased the number of red blood cells to be injected.

Dual fl uorescence and radioactive labelling
Seven days after MI induction, BM-MNCs were quickly thawed and labeled immediately before injection using a green intracellular fl uorescent dye, Carboxy Fluorescein diacetate Succinimidyl Ester (CFSE, Molecular Probes Invitrogen, Cergy Pontoise, France) [7]. 99m Technetium ( 99m Tc) labeling was performed after CFSE labeling, using the Isolink TM kit according to the manufacturer's instructions (Mallinckrodt Medical B.V., Petten, the Netherlands). CFSE + BM-MNCs were incubated with Isolink TM / 99m Tc for 40 minutes at 37°C and pH = 6. Before injection, BM-MNCs were centrifugated and resuspended in 10 mL NaCl to obtain a fi nal solution containing 300-400 MBq 99m Tc-CFSE + BM-MNCs. Less than 5% cell mortality was observed in all experiments using trypan blue exclusion method after thawing, after radiotracer and fl uorescence labeling, and immediately before injection. 99m Tc cell leakage was evaluated before injection and after four hours incubation at 37°C in vitro. Every hour, supernatant was collected. 99m Tc activities were quantifi ed on both supernatant and BM-MNC pellet using a gamma counter. Radio-labeling effi cacy, defi ned as the percentage of 99m Tc cell activity in the fi nal cell solution, was 98 ± 5% (n = 3) after four hours and was stable during the four hour experiment, suggesting no 99m Tc cell leakage.

Autologous BM-MNC injection
At Day 7 after MI, a second catheterization was performed. A 3 mm over the wire balloon catheter was advanced to the site of MI induction. An angiogram was performed before completing the procedure. In the absence of coronary reperfusion, a recanalization was performed with the guide wire before its distal place ment. BM-MNCs were injected less than 30 minutes after preparation. Th e balloon was infl ated at low pressure (two bars) and 3.3 mL of labeled BM-MNCs were injected through the lumen. Th e balloon was defl ated two to three minutes later. Th e same procedure was repeated three times with a two minute interval between procedures. A total of 10 8 BM-MNCs were injected in 10 mL of NaCl at a fi nal concentration of 10 7 BM-MNCs/mL. For intravenous delivery, 10 8 BM-MNCs were injected via the superfi cial vein of the left front leg, using a bolus injection of 15 seconds.

Nuclear imaging and gamma counting
Planar scintigraphic images of 99m Tc-BM-MNC body distri bution were obtained 1hour and 24 hours after tracer injection, using a gamma camera fi tted with a low energy all purpose high resolution parallel hole collimator (DS7, Sopha Medical, GEMS, Buc, les Yvelines, France). Th e energy window was centred at 140 keV ± 20%. Using a dedicated image processing co mputer, a static image was recorded into a 64 × 64 matrix for 35 seconds. Th e camera was calibrated using a standard source measured with a gamma counter before each acquisition to convert image counts to activity in MBq. For each experiment a calibrated dose of technetium was used to quantify activity decay after 24 hours. Anesthetized animals were positioned in supine positions, at a constant distance of 30 cm from the collimator, and images were recorded in thoracic and abdominal projections. On each image, a region of interest was drawn around each organ to quantify counts within the drawn area. Organ activity in each region of interest was expressed as the percentage of the total initial injected activity (measured as the diff erence between the 99m Tc-BM-MNCs syringe content before and after injection) with correction for 99m Tc physical decay.
Animals were sacrifi ced, their organs were excised and imaged with the gamma camera, and activity was measured in weighted samples of each organ and used to quantify organ gamma counting radioactivity (ACAD activimeter, Lemer Pax, Nantes, France). All quantitative analyses were performed by two investigators blinded to animal treatment group.

Pathologic examination and immunohistology
Animals were sacrifi ced 1 hour or 24 hours after BM-MNC injection. Heart samples were cryopreserved and sliced into contiguous transverse 5 μm thick cryosections or were embedded in paraffi n and stained with Hematoxylin-Eosin-Safran (HES). Immunolabeling was performed to detect VonWillebrand Factor antibody (DakoCytomation SA. France, Trappes France). Cell nuclei were labeled using propidium iodide. Histology analyses were performed by two pathologists blinded to animal treatment group.

Statistical analysis
Data were expressed as mean ± SD. Th e Kruskal-Wallis test was used for comparison of independent samples and Spearman's test for correlation coeffi cient. A value of P <0.05 was considered statistically signifi cant.

Procedural results
Myocardial infarction was induced in 37/46 pigs ( Figure 1A). MI induction was complicated by death related to ventricular fi brillation during the procedure in eight cases (21.6%), and immediately after at the wake up time in four other cases (10.8%). Total death rate was 32% during and immediately after MI induction, probably due to extended MI [8].
Th e 25 surviving animals were randomly assigned to three diff erent groups: fi ve were used as controls (MI without BM-MNC injection), 16 received intracoronary BM-MNCs, four received intravenous BM-MNCs. Two deaths were observed during balloon infl ation for BM-MNC intracoronary injection (one from a ventricular fi brillation and one from a second MI, probably related to the injection site location being within a non-infarcted area). In the 9/47 animals without MI induction, seven received intracoronary BM-MNC injection and two were used as radioactivity controls (intracoronary injection of Isolink TM / 99m Tc only).

Myocardial infarction model
MI was induced by complete obstruction of the left coronary artery using balloon infl ation and injection of thrombin and fi brinogen (n = 23; left anterior descending artery: n = 15; left circumfl ex artery: n = 8) ( Figure 1B). Transient ST-segment elevation and isolated ventricular extrasystoles occurred during each coronary balloon occlusion, but abnormal ECG profi le (ST-segment elevation) persisted only after injections of thrombin and fi brinogen. Plasma creatine kinase levels increased 24 hours after coronary occlusion (1,222 ± 604 UI/l before coronary occlusion, and 4060 ± 1961 UI/l at 24 hours, P <0.01), confi rming MI induction. Macroscopic examination of the heart performed seven to eight days after coronary occlusion showed extensive transmural antero-inferior MI encompassing 41.25 ± 5% of left anterior ventricle area (n = 23), with a thinning of the injured wall. Statistical analysis did not show any diff erence in MI size between animal groups, nor between animals with left anterior descending artery or left circumfl ex coronary occlusion, nor between animals with or without spontaneous reperfusion at the time of cell injection (not shown). Histopathologic examination confi rmed MI, with associated infl ammatory reaction and loss of cardiomyocyte nuclei ( Figure 1C). Coronary angiography at seven days after MI showed spontaneous reperfusion in 18/23 pigs. Histological analyses showed an adherent residual thrombus within the coronary artery without total occlusion of the vessel lumen. Coronary occlusion in the fi ve out of 23 remaining pigs required reperfusion before BM-MNC injection. Neither in pigs with spontaneous reperfusion nor in those with mechanical reperfusion, angiographic control showed abnomalities of blood fl ow distribution when compared to basal angiographic data obtained before MI induction.

Biodistribution and 99m Tc-BM-MNC cardiac engraftment after MI and intracoronary injection
Scintigraphic images were used to track 99m Tc-BM-MNCs after intracoronary injection. Radioactivity in the kidneys and the bladder due to free 99m Tc clearance, was observed 15 minutes after injection (Figure 2A). At one hour after injection, whole pig body images revealed intense regional accumulation of radioactivity in the heart ( Figure 2B). At 24 hours a persistence of heart and lung radioactivity was observed, with a decrease in signal intensity ( Figure 2C).
Radioactivity was expressed as the fraction of measured radioactivity, decay-corrected and divided by the injected radioactivity amount. In the fi rst set of experiments, radioactivity was quantifi ed one hour after 99m Tc-BM-MNC intracoronary injection both in regions of interest drawn on the planar images and in isolated organs ( Figure 3A). Th ere was no signifi cant diff erence between both quantifi cation methods (n = 6, R = 0.94 for heart, and R = 0.84 for lungs). Th erefore results from either method were used in following experiments. As controls, two pigs without MI but with intracoronary injection of Isolink TM / 99m Tc only were used to quantify free Isolink TM / 99m Tc fi xation in each organ. At one hour radioactivity rate was 11.5 ± 2.1% in the heart, 3.5 ± 2.1% in the lungs and 14.5 ± 2.1% in the liver ( Figure 3B), and at 24 hours radioactivity rate was null (not shown). Consistent with the in vivo scintigraphic images (Figure 2A), after 99m Tc-BM-MNC intracoronary injection high levels of radioactivity were measured in the heart (34.8 ± 9.9% of injected radioactivity) and in the lungs (32.6 ± 13.9%) (n = 6, Figure 3B). Small quantities of radioactivity were observed in other organs (2 ± 1.4% in the liver, 1.7 ± 0.6% in the kidneys). Radioactivity was also quantifi ed in the infarcted site and in other cardiac tissues (right ventricle, healthy left ventricle and auricles, Figure 3C). We observed an accumulation of radioactivity in the infarcted area, Almost no radioactivity was detected remote from injection site (right ventricle, healthy left ventricle, auricles). 24 hours after BM-MNC injection, radioactivity persisted with lower intensity in the heart (6 ± 1.7%) and in the lungs (11 ± 2.6%) (n = 3, Figure 4B). Importantly, there was no correlation between the infarction size and the heart radioactivity rate at one hour and 24 hours (not shown).

Myocardial infarction and BM-MNC cardiac engraftment
To investigate the role of MI on BM-MNC homing and engraftment in the heart, 99m Tc-BM-MNC were intracoronary injected in pigs with MI and in pigs without MI. In both animal groups, ECG alteration (ST-segment elevation) was observed during intracoronary BM-MNC injection. STsegment elevation stopped immediately after the procedure, suggesting that no micro-infarction was induced by injection itself favouring cardiac homing. Importantly, no signifi cant diff erence in heart and lung BM-MNC distribution was observed between both groups, with a high radioactivity ratio in the heart (34.8 ± 9.9% in MI group versus 51.2 ± 17.2% in no MI group, Figure 3B).

Biodistribution of BM-MNCs and delivery route
Th e effi cacy of intracoronary administration versus systemic intravenous administration was evaluated in pigs with MI. After intravenous cell delivery at 1 hour and 24 hours, radioactivity was detected mainly in the lungs both on scintigraphic images (n = 4, Figure 4A) and by radioactivity quantifi cation, in marked contrast with intracoronary injection ( Figure 4B). After intravenous injection, cardiac radioactivity was almost null (0.16 ± 0.23%) as compared to pigs after intracoronary injection (34.8 ± 9.9%). After 24 hours, no cardiac radioactivity was observed in intravenously injected pigs, suggesting an absence of cardiac homing at 24 hours.

Histological analyses
For histological analyses, BM-MNCs were labeled with a fl uorescent marker (CFSE) prior to injection and animals were sacrifi ced one hour or 24 hours after injection. As noticed by others [9] microscopic fl uorescent artefacts were observed in the myocardium of control animals with MI and without BM-MNC injection, characterized by fl uorescent leukocytes and autofl uorescence within the infarcted tissue. However, these artefacts could be distinguished from BM-MNCs by their small size and their limited location within necrotic areas [10] ( Figure 5A). BM-MNCs in non necrotic areas were detected in each animal that received CFSE+ BM-MNCs ( Figure 5B-5C), and were not observed in any control animals (with MI and without BM-MNC injection, Figure 5D). CFSE+ BM-MNCs were also observed in heart tissue 24 hours after intracoronary injection in each animal that received CFSE+ BM-MNCs ( Figure 5J). Th e presence of BM-MNCs was confi rmed in HES histological analyses ( Figure 5E-5F). BM-MNCs appeared as clusters of mononuclear cells, characterized by round    No BM-MNCs were observed remote from the injection site (auricle, right ventricle, non infarcted myocardium).
Tissue sections were stained for endothelial cells and capillaries using anti-VonWillebrand antibody ( Figure 5G). Neither thrombus nor arterial occlusion was observed in both groups. BM-MNCs were incorporated within myocardial tissue and were not observed within the lumen neither into wall blood vessel, and no cells were observed with double staining CFSE+ and endothelial marker. Double labeling for macrophages using anti-MAC-1 antibody was negative (not shown), suggesting that transplanted BM-MNCs had not been phagocyted.
We also investigated extracardiac BM-MNC biodistribution by histological analysis. CFSE+ BM-MNCs were disseminated in whole lungs ( Figure 5H). Sections of other vital organs (liver, spleen, kidneys and thymus) demonstrated no signifi cant cell location, only cell debris ( Figure 5I). Importantly, the presence of BM-MNCs was confi rmed one hour ( Figure 5K-5L) and 24 hours (not shown) after intracoronary injection in every pig without MI and again, neither thrombus nor arterial occlusion was observed. Th ese data were in accordance with analyses of scintigraphic data and radioactive organ quantifi cation ( Figure 3B).
Th e absence of cardiac homing after BM-MNC intravenous injection at 24 hours as observed on scintigraphic images was also confi rmed by histological analyses, showing an absence of fl uorescence in myocardium inside and outside MI area (not shown). In contrast, CFSE+ BM-MNCs were numerous and dispersed in whole lungs when injected intravenously (not shown).

Discussion
In a porcine model of acute MI with late reperfusion and autologous cardiac cell therapy, we quantifi ed biodistribution using a radiolabeling technique. Most of the BM-MNCs delivered by the intracoronary route were distributed within the heart and lungs. Th e presence of MI did not modify BM-MNC cardiac homing and engraftment after intracoronary injection. Finally,   target area. Intracoronary delivery off ers the advantages of a non-surgical method that can be performed percutaneously during angioplasty for acute MI. In the present study we showed that intracoronary BM-MNC injection can allow cell engraftment into myocardium tissue, whereas peripheral systemic intravenous adminis tration was not effi cient at 24 hours. Th is suggests that the cardiac niche eff ect favoured by MI was not suffi cient to attract intravenously injected BM-MNCs within 24 hours, thus limiting the effi cacy of systemic intravenous injection. Chemoattractant factors secreted by the infarcted heart might be too diluted in the body of large animals like pigs to attract BM-MNCs. If this is the case, similar results may be expected in humans. Another hypothesis is that the ability to trap intravenously injected BM-MNCs within the heart is limited, since only 4% to 5% of cardiac output is dedicated to supplying the coronary arteries [11]). Finally, cardiac engraftment following systemic intravenous BM-MNC delivery might be limited by lung entrapment prior to coronary artery access. Numerous studies have been performed in rodent [12] and large animal models [13,14], using intravenous injection of mesenchymal stem cells, showing mesenchymal stem cell engraftment after 24 hours. Several hypotheses can be discussed to explain these discrepancies with our results. First, our protocol closely matched clinical trial protocols, in which a seven-day delay between acute MI and cell injection seems to be the best time for effi cacy following intracoronary injection [3,15]. However, this time may be too long after intravenous cell injection, in regards to local infl ammatory response and niche eff ect. In most studies mesenchymal stem cells were injected 1 to 72 hours after myocardial infarction [12][13][14]. Second, BM-MNCs were labeled with 99m Technetium, a radioelement that has a short half life and may not have been detected at 24 hours. However, we did not observe any fl uorescent cell within the heart 24 hours after intravenous injection, corroborating our radioactivity data. Finally, mesenchymal stem cells may have a better capability for cardiac homing than BM-MNCs, that contain a very small stem cell number. Interestingly, a recent study comparing intra-aortic, intravenous, and intramyocardial delivery of mesen chymal stem cells in rats observed 5% cell survival at 48 hours after intravenous delivery, mostly in the lungs [15].
Th e radioactive labeling using Isolink TM kit did not discriminate cell type, and was not species nor cell specifi c. In a pilot study we evaluated 99m Tc-hexa-methylpropylen-amine-oxime (HMPAO) labeling, often used in humans. However, 99m Tc-HMPAO labeling of BM-MNCs was totally unsuccessful (not shown). We then switched to the radioactive linker Isolink TM , which allowed us to inject a large quantity of 99m Tc-labeled cells, signifi cantly higher than in other studies [16,17]. Although, similar to previous studies [11,18,19], the viability of Isolink TM / 99m Tclabeled BM-MNCs was not altered, the adverse eff ects of labeling on the migratory and functional abilities of BM-MNCs cannot be entirely excluded. Th e majority of cardiac BM-MNC homing studies were performed in rodents, and only a few studies established the fraction of transplanted cells retained within the myocardium using direct radioactive labeling of the cells. In our study, radioactivity quantifi cation in each organ showed 34.8 ± 9.9% total radioactivity in the heart with an accumulation within the injection site, one hour after intracoronary BM-MNC injection, and 6.0 ± 1.7% at 24 hours. Importantly, histology results were in accordance with scintigraphic imaging data confi rming the presence of numerous BM-MNCs at one hour and 24 hours after intracoronary injection. As in our study, Hou et al [8], using a pig model of reperfused MI with intracoronary cell delivery and radioactive cell quantifi cation, observed a largely right-sided distribution of BM-MNC. However, only 2.6 ± 0.3% of BM-MNCs were detected in the heart, with some BM-MNCs being distributed to the right ventricle. Importantly, the experimental model in this study was xenogeneic, with human BM-MNCs being injected into pigs. Human cells may lack the correct membrane receptors to home into the pig myocardium or may undergo acute lysis due to a xenogeneic reaction. In a recent study, using a similar model of reperfused MI in pigs with autologous BM-MNC transplantation [20] 6.5% of autologous transplanted BM-MNCs were detected in the heart four days after injection, a result close to ours at 24 hours. Finally, cardiac homing cell rate at one hour might be overestimated by including a possible leakage of radioactive label from the cells, although this appears unlikely as no radioactive cell leakage was observed in vitro in the next four hours following radioactive labeling.
Although phase I-II clinical trials have been completed using coronary delivery of BM cells, only one study used coronary delivery of 99m Tc-labeled BM-MNCs (in a single patient), and observed intense cardiac cell engraftment [17]. Th e study of Hofmann et al [11] showed that in fi ve patients with MI, only 1.3 to 2.6% of 18 F-FDG-labeled BM-MNCs were detected in the infarcted myocardium one hour after intracoronary BM-MNC injection. However, the number of injected BM-MNCs was 30-fold higher than in our study, so the absolute number of retained BM-MNCs within the heart was 39 to 78 x 10 6 BM-MNCs, a number close to ours (34.8 ± 9.9 x 10 6 retained BM-MNCs). Intracoronary injection of a very high cell number may saturate the binding sites for cardiac cell homing, raising questions about the ideal number of cells to be injected.
After intracoronary injection, we observed a similar biodistribution of BM-MNCs in animals with or without MI, despite the large size of the infarcted area (45% of the left ventricle). Several studies have shown that acute myocardial infarction is followed by an acute local infl ammatory reaction involving upregulation of chemokines receptors and adhesion molecules, thereby facilitating adhesion and infi ltration of cells involved in tissue repair, including stem cells [21][22][23]. Th e dynamic capability of BM-MNCs to migrate and the niche eff ect are central in regenerative medicine [21][22][23]. In our model, to carry out intracoronary injection, blood fl ow was stopped three times for at least two minutes to prevent backfl ow and prolong contact time between BM-MNCs and myocardium. When BM-MNCs were injected without prior MI induction, each maneuver for intracoronary cell injection resulted in ST-segment elevation, as already described for repeated angioplasty balloon infl ations [24,25]. Although we did not evaluate the intensity of the ischemia induced by balloon infl ation, this suggests that the injection technique created local downstream ischemic and preconditioning eff ects [25], thus rendering the local microenvironment more receptive to cell homing [24]. A recent study in a pig model of myocardial infarction and intracoronary injection of autologous BM cells, balloon occlusion was found ineff ective to increase cell homing [24]. In another study in a similar model, single-bolus delivery was as eff ective as three balloonocclusion deliveries [26]. Importantly, both studies were performed only in pigs with myocardial infarction and not in healthy pigs, suggesting that the cardiac niche eff ect favoured by MI was suffi cient to attract injected BM-MNCs without any further eff ect of balloon occlusion. Th e fact that the presence and the size of MI did not infl uence cell engraftment could be useful for cardiac cell therapy in patients with chronic heart failure of non-ischemic origin or with ischemic myocardium without myocardial infarction. Two clinical trials have been published with patients who have had a myocardial infarction at least three months before BM-MNC coronary injection [1,27,28]. A clinical benefi t was observed in both studies, suggesting that BM-MNCs homed to the myocardium despite the absence of acute myocardial infarction. Th e hypothesis of cardiac homing in the absence of acute myocardial infarction was recently confi rmed in two Phase I cardiac cell therapy clinical trials on patients with chronic ischemic cardiomyopathy and receiving an intracoronary injection of radiolabeled CD133+ or CD34+ cells [16,29].

Conclusions
Myocardial cell distribution following intracoronary injection did not depend on MI presence. Given these results, further eff orts to analyze the mechanisms of adhesion and transmigration through the vascular wall will be rewarding.