OUP user menu

A quantitative, randomized study evaluating three methods of mesenchymal stem cell delivery following myocardial infarction

Toby Freyman, Glenn Polin, Hashim Osman, Jody Crary, MinMin Lu, Lan Cheng, Maria Palasis, Robert L. Wilensky
DOI: http://dx.doi.org/10.1093/eurheartj/ehi818 1114-1122 First published online: 1 March 2006


Aims Mesenchymal stem cells (MSCs), rare bone marrow-derived stem cell precursors of non-haematopoietic tissues, have shown promise in potentially repairing infarcted myocardium. These and similar cell types are being tested clinically, but understanding of delivery and subsequent biodistribution is lacking. This study was designed to quantitatively compare MSC engraftment rates after intravenous (IV), intracoronary (IC), or endocardial (EC) delivery in a porcine myocardial infarction (MI) model.

Methods and results Allogeneic, male MSCs were cultured from porcine bone marrow aspirates. Iridium nanoparticles were added during culturing and internalized by the MSCs. An MI was induced in female swine (27–40 kg in size) by prolonged balloon occlusion of the mid-left anterior descending artery. Animals (n=6 per group) were randomized to one of three delivery methods. Cellular engraftment was determined 14±3 days post-delivery by measuring ex-vivo the iridium nanoparticle concentration in the infarct. Confirmation of cellular engraftment utilized both DiI and fluorescence in situ hybridization (FISH) labelling techniques. During MSC infusion, no adverse events were noted. However, following IC infusion, half of the pigs exhibited decreased blood flow distal to the infusion site. At 14 days, the mean number of engrafted cells within the infarct zone was significantly greater (P≤0.01) following IC infusion than either EC injection or IV infusion and EC engraftment was greater than IV engraftment (P≤0.01). There was less systemic delivery to the lungs following [EC vs. IV (P=0.02), EC vs. IC (P=0.06)]. Both DiI and FISH labelling demonstrated the presence of engrafted male MSCs within the female infarcted tissue.

Conclusion IC and EC injection of MSCs post-MI resulted in increased engraftment within infarcted tissue when compared with IV infusion, and IC was more efficient than EC. However, IC delivery was also associated with a higher incidence of decreased coronary blood flow. EC delivery into acutely infarcted myocardial tissue was safe and well tolerated and was associated with decreased remote organ engraftment with compared with IC and IV deliveries.

  • Intracoronary
  • Intravenous
  • Endocardial
  • Bone marrow derived stem cell
  • Regeneration


Cardiomyocyte loss resulting from an acute myocardial infarction (MI) is a leading cause of death, congestive heart failure, and ventricular arrhythmias. As cardiomyocytes are unable to regenerate, replacement by fibrous tissue results, which may lead to dysfunctional myocardium. One approach to the treatment of permanent myocardial cell loss has been replacement therapy with either skeletal muscle or stem cells.

Mesenchymal stem cells (MSCs) are rare bone marrow-derived stem cell precursors of non-haematopoietic tissues, which have shown promise in potentially repopulating infarcted myocardium.1,2 MSCs have been used allogenically without immune suppression, are self-renewing, and can be transduced by a number of vectors while maintaining transgene expression after in vivo differentiation.1,3 Human MSCs can differentiate into a cardiomyocyte-like phenotype in adult mouse hearts, expressing β-myosin heavy chain, α-actinin, cardiac troponin I, and phospholamban at levels similar to host cardiomyocytes.4 In the ischaemic mouse heart, implantation of 5×104 to 5×105 bone marrow stem cells, containing MSCs, reduced infarction size and fibrosis.5 In the porcine MI model, direct injection of 6×107 MSCs into the infarct resulted in engraftment in all animals with expression of muscle-specific proteins, resulting in attenuation of contractile dysfunction and decreased wall thinning.6 In two human studies, infusion of autologous bone marrow-mononuclear cells after successful percutaneous coronary intervention for ST elevation MI resulted in an increase in left ventricular ejection fraction.7,8 More recent data have suggested that injection into ischaemic myocardium of 5×106 genetically engineered rat MSCs using retroviral transduction to overexpress the prosurvival gene Akt1 resulted in regeneration of 80–90% of lost myocardial volume and normalized left ventricular systolic and diastolic cardiac functions.9

Systemic MSC infusion has been shown to be safe in oncology trials10 and their ability to home to injured tissues may result in localized therapeutic benefits. As such, MSCs could become an important therapeutic approach to increase angiogenesis or improve left ventricular function following MI. However, little is known regarding the optimal delivery strategy for stem cells. Considering that the cardiology field has moved on to clinical trials of stem cell therapy, maximizing the benefits, or even showing proof-of-principle, will depend on efficient delivery. Accordingly, the current study was designed to quantitatively compare engraftment rates of MSCs delivered by three different delivery approaches: intravenous (IV), intracoronary (IC), and endocardial (EC) following an MI in a relevant large animal model.


MSC preparation

Allogeneic, male MSCs were cultured from porcine bone marrow aspirates using a standardized protocol (Osiris Therapeutics, Inc., Baltimore, MD, USA).11 Briefly, mononuclear cells were isolated from anti-coagulated, porcine bone marrow aspirates by density-gradient centrifugation. The resulting mononuclear cells were then transferred to culture flasks. Non-adherent cells were removed by subsequent media changes and MSCs were culture expanded (∼1:4 at each passage) out to passage 3. In the final 24 h of culture, iridium nanoparticles, 300 nm in diameter (CL-300G02-1; BioPhysics Assay Labs Inc., Worcester, MA, USA), were added to the culture medium at a concentration of 20 µL/mL. Labelled cells were subsequently harvested, cryopreserved, and stored (<−150°C) as individual doses.

In order to determine the average amount of label internalized per cell, cryopreserved cells were thawed and resuspended in SanSaLine (R-7000, BioPhysics Assay Labs Inc.). Triplicate samples of known cell number ranging from 2.5 to 500×104 cells per sample (covering the anticipated engrafted cell numbers in tissues) were prepared. Measured values were averaged and the relationship between iridium content and cell number was determined by linear regression. This equation was then used to calculate cell number in tissue samples from experimental animals.

Prior to in vivo administration, doses of 50×106 viable cells were prepared for each animal. Frozen aliquots were thawed in a water bath at 37°C and diluted 1:1 with saline (PlasmaLyte-A, Baxter, Deerfield, IL, USA). No further manipulations were performed when cells were prepared for IV infusion, except to check cell concentration and viability using Trypan Blue exclusion and a cytometer. To prepare cells for IC infusion or EC injection, thawed cells were centrifuged and resuspended in saline to obtain the prescribed cell concentration. For all approaches, doses were used only if cell viability exceeded 70%.

Infarct procedure

MIs were induced in female Yorkshire domestic swine (27–40 kg in size) by prolonged balloon occlusion of the mid-left anterior descending artery (LAD). A standard angioplasty balloon sized 1.1:1 to the arterial diameter was used to occlude the LAD at the second diagonal branch for 60 min. Unpublished data from our laboratory using MRI 3 days post-infarction have demonstrated that this approach results in an infarction affecting 20–30% of the left ventricular wall mass. The infarct is permanent as noted by MRI and gross examination 12 weeks later. The level of anti-coagulation was determined by monitoring the activated clotting time (ACT) and heparin was administered to maintain an ACT of >300 s. Oximetery was used for monitoring arterial O2 saturation and continuous electrocardiographic monitoring was performed to assess for arrhythmias.

Evaluation of extracellular iridium clearance

Pilot studies were performed to evaluate the clearance of free iridium label in infarcted myocardium (Figure 1). A volume of iridium label, equivalent to that contained in one dose of cells, was mixed with 14 mL of saline. The label was delivered by IC infusion (discussed subsequently) into five pigs following infarction. Two pigs were sacrificed immediately and three at 14 days. The infarcted myocardium was isolated and processed for iridium label analysis and iridium particles were quantified using a neutron activation technique (BioPhysics Assay Labs Inc.). This analysis provided values, in disintegrations per minute, proportional to the amount of iridium contained in each sample. Average label content in the infarcted myocardium was calculated for both time points.

Figure 1

Outline and timelines of studies.

MSC delivery

Animals were randomized to the delivery method by a blinded draw during the infarction procedure. By design, any animal dying within 24 h of the infarct was removed from and replaced in the study. Deaths <24 h after cell delivery were thought to be secondary to the effects of the infarct. A total of six animals per group were included in this study. The following procedures were used.

IV infusion

Fifteen minutes after MI completion, the MSC solution was infused into an ear vein catheter at 2–3 mL/min.

IC infusion

Fifteen minutes after completion of the infarct, an over-the-wire angioplasty balloon, sized 1:1 to the arterial diameter, was inflated at the site of occlusion. Two millilitres of MSC suspension was infused during a 2-min balloon inflation via the guide-wire lumen. The balloon was then deflated, coronary artery blood flow restored, and the myocardium was perfused for 2 min. The cycle was then repeated for a total of six to seven times to deliver the total dose of 14 mL.12 Following the last infusion, an angiogram was acquired to assess blood flow in the target artery.

EC injection

Fifteen minutes after completion of the infarct, the animal was rotated ∼30° along the long axis. Left ventriculography at 30° right anterior oblique and 60° left anterior oblique (LAO) angles was performed to visualize the extent and degree of infarction and to visualize the anterior and septal walls (the areas of infarction). Using maps drawn from the ventriculogram, selective targeting of injections in the infarct area (distal 1/2 of the septal/anterior wall) was performed. An EC injection catheter (Stiletto™, Boston Scientific Corporation, Natick, MA, USA) was prepared and loaded with the cell suspension. Targeted injection was performed using a 7 Fr steering catheter within a 9 Fr left ventricular sheath placed in the left ventricle. Using the LAO view, 13 injections of 0.2 mL each were made to the infarct and border zone areas. The pattern and location of injections were recorded on a mylar sheet to insure that the entire infarct area was injected with MSCs.

Determination of cellular engraftment

Cell engraftment was evaluated 14±3 days post-delivery. Following euthanasia (Euthasol 0.3 mL/kg), the heart was removed and the left coronary tree perfused with 200 mL SanSaLine solution (BioPhysics Assay Labs Inc.) for 30 min at physiological pressure (100 mmHg). The peripheral organs were examined for potential injury or toxicity and liver and lung samples (∼10 g) were removed for evaluation of systemic delivery of MSCs. Following removal of the right ventricular free wall and both atria, the remaining left ventricle was sectioned in a bread-loaf fashion into cross-sections ∼1-cm thick and the sections weighed. The infarcted area was removed including 1 cm of border zone tissue at each end. The infarct sections, a section of remote, non-infarcted myocardium, and lung and liver sections were placed in a desiccator (Garden Master Dehydrator, American Harvest) at 65°F for >24 h. The dried sections were sent to BioPhysics Assay Labs Inc. for analysis of tissue iridium content. The extent of engrafted MSCs was calculated from Figure 2.

Figure 2

The relationship between iridium content and cell number was determined by analyzing samples with known numbers of cells prepared in vitro. Linear-regression analysis returned an equation which describes this relationship. Iridium counts returned from tissue samples can be converted to cell number using this equation. Error bars are obscured by the data markers.

Confirmation of cellular delivery and engraftment

Cellular engraftment after delivery was confirmed by two independent methods: MSCs labelled with DiI and FISH staining of male MSCs delivered to female recipients. For DiI labelled MSCs, hearts were obtained either immediately or 7 days after MI (Figure 1). Myocardial sections from the infarct and border zones were stored at −80°C in OCT. Serial 8-µm frozen sections were cut and placed on two sets of the slides. Sections were fixed with 4% PF for 10 min and washed with ddH2O. One set of slides was mounted with Vectashield mounting medium (Vector) for DiI. The adjacent sections were stained for haematoxylin and eosin. For FISH, sections of paraffin-embedded myocardial tissue in both the infarct and border zone was used to evaluate the presence of cells containing the Y-chromosome. DNA in situ hybridization for the human Y-chromosome was performed with CEP Y satellite III (Vysis), a Y chromosome repeat marker that is labelled with fluorescein. Cell nuclei were counterstained with Hoechst 33258 nuclear dye.

Determination of ventricular injury

At necropsy, the myocardium was evaluated for the presence and extent of MI and additional injury related to MSC delivery. Reference was made to the presence of LAD occlusion as well as EC injury tracts related to EC delivery. The extent of myocardial fibrosis was quantified using a 0–3 score with 0=no myocardial fibrosis, 1=patchy fibrosis <25% of the infarcted myocardium, 2=patchy fibrosis >25% but <50% of the infarcted myocardium, and 3=>50% of the infarcted myocardium.

Statistical analysis

The primary endpoint of the study was cellular engraftment within the infarct zone. Results are presented as the mean±standard deviation, unless otherwise noted. Statistical comparisons between groups were performed using the Ryan–Einot–Gabriel–Welsch multiple range test. Comparison of the fibrous score was performed using a T test, assuming unequal variance. All tests were two-tailed and a P-value <0.05 was considered significant.


Pilot data on iridium and MSCs

Conversion to cell number. Regression analysis showed that iridium particle content and cell number were linearly related (Figure 2). The lower limit of detection was 30 000 cells, representing 0.06% of the administered dose of 50×106 cells.

Label clearance. Fourteen days after IC delivery, 80% of the label was cleared with 20% of the iridium label still present within the myocardium, indicating acute myocardial retention or uptake by phagocytic cells of some iridium particles (Figure 3).

Figure 3

Iridium label was partially cleared from infarcted myocardium during the 14 days following IC infusion. During this time period, iridium content was reduced by 80%. Iridium content represents the mean±standard error.

Procedural results

In total, 22 animals underwent the infarct procedure. During the procedure, two animals had intractable ventricular fibrillation and could not be resuscitated. These animals were not randomized into a delivery group. Two animals, both randomized to the IC delivery group, died within 24 h of the infarct and were replaced with additional animals. One animal died of a cardiac arrest, presumably on the basis of an arrhythmia and the other animal died of a dissected iliac artery caused by insertion of the arterial sheath. During MSC infusion, no adverse events were noted in any group. Following IC MSC infusion, however, three of six pigs exhibited decreased blood flow (TIMI grades 1–2),13 distal to the infusion site (Figure 4). No animals exhibited any adverse effects and all had normal recovery, surviving to scheduled sacrifice.

Figure 4

No-reflow following MSC infusion post-infarct. (A) Initial angiogram obtained prior to MI and (B) angiogram obtained immediately after infarction. Arrow denotes location of occluding balloon. Open arrow denotes normal distal arterial blood flow. (C) No-reflow 15 min following MSC infusion. Blood flow abruptly stops distal from infusion site (arrow).

Infarct size

All hearts exhibited signs of infarction with 6/18 demonstrating myocardial thinning, 13/18 hearts showing fibrous tissue in-growth, and 7/18 hearts showing haemorrhagic degradation of the myocardium (Figure 5). The mean LV mass and LV-to-body mass ratio were similar for all groups and the mass of the infarcted tissue and percentage of LV infarcted were consistent across groups (Table 1).

Figure 5

Representative images showing MI 14 days post-injury and cell delivery. (A) Following sectioning of the left ventricle into bread-loafed sections, ∼1-cm-thick gross evaluation revealed that the infarct is transmural in extent involving the distal septal wall and anteriolateral wall (pale areas). This animal had mesenchymal cells delivered by IC infusion. (B) Magnified view of second slice from apex in (A). Slight fibrosis is noted in two areas (asterisk). (C) Mid-level, infarcted septal-anterior wall from EC injection group with slight thinning and fibrosis is noted. (D) Mid-level slice from another heart in the EC group showing the infarct area (pale). The infarct was not thinned and there was no fibrosis noted. (E) Outline of typical extent of myocardial injury 14 days after infarction (dotted line). Arrow denotes location of occluding balloon. (F) Photomicrograph of area of infarction. Note the large number of neutrophils and inflammatory cells and the relative absence of viable myocardium. Scale bars are in centimetres, except in (E) and (F) in which line represents 100 µm.

View this table:
Table 1

Comparison of injury between delivery groups

GroupTotal LV mass (g)LV mass/body mass (g/kg)Infarct mass (g)Percentage of LV infarcted
IC (n=6)110.9±16.82.9±0.227.9±4.226.3±4.1
EC (n=6)111.1±17.42.8±0.228.6±6.625.8±4.9
IV (n=6)109.9±17.82.7±0.528.7±6.726.0±2.5
  • All values are mean±standard deviation.

Gross evaluation of the myocardium demonstrated no myocardial injury other than infarction. No chronic LAD occlusions were noted. No perforations or injury tracts, attributable to EC delivery, were observed. Of animals receiving MSCs by the IC route, four of six had grade 3 fibrosis in the infarcted area. Three of the four animals had decreased coronary blood flow after cellular infusion. The mean fibrosis grade after IC delivery was 2.33±1.03, after IV delivery 1.50±1.29, and after EC delivery 1.33±1.21 (IC vs. IV P=0.32, IC vs. EC P=0.32).

Engraftment of MSCs

At 14 days, the mean number of retained iridium particles within the infarct zone was significantly greater following IC infusion than either EC injection or IV infusion (IC vs. EC P=0.01, IC vs. IV P=0.0008, EC vs. IV P=0.003, Table 2). Normalization of the results to infarct mass did not alter the differences (Table 2). For IC infusion and EC injection, 14-day retention represented 6 and 3% of the administered dose, respectively. For the IV infusion group, none of the infarcts contained a measurable number of cells. In none of the delivery groups were cells detected in remote, non-infarcted myocardial samples.

View this table:
Table 2

Engraftment of MSCs in tissues varies by delivery technique

Mesenchymal stem cell engraftment 14 days after delivery
Infarct zone (cells)Infarct zone (cells/g)Liver (cells/g)Lung (cells/g)
IC2 864 000±983 000106 000*± 43 0001000±100011 000±2000
EC1 393 000±618 00051 000**±24 000700±7004000±3000***
IVNone detectedNone detected9±2013 000±2000
  • *IC vs. EC (P=0.01), IC vs. IV (P=0.0008).

  • **EC vs. IV (P=0.003).

  • ***EC vs. IC (P=0.06), EC vs. IV (P=0.02).

Confirmation of engraftment was observed at both 7 and 14 days after delivery. Histological evaluation of infarcted and border zone myocardium at 7 days demonstrated that the majority of engrafted cells was found in areas with some viable myocardium. Infarction resulted in large areas of dead myocardium (Figure 5) with few viable myofibrils and an abundance of neutrophils. Within the infarcted zone, no MSCs could be located except in association with viable myofibrils. Engrafted cells were often observed near arterioles (Figure 6). There was no histological evidence of macrophage uptake of iridium nanoparticles. Using FISH, MSCs of male origin were observed engrafted in both infarct and border zones of female hearts following IC and EC deliveries (Figure 7).

Figure 6

Two examples of engraftment of MSCs labelled with DiI 7 days after MI and IC infusion. Note the location of engrafted cells near an arteriole containing MSCs (top). Engrafted cells within area of viable myocardium (bottom). Red cells denote DiI positive engrafted cells and green cells denote native myocardial cells. Original magnification 40×.

Figure 7

Engraftment of male MSCs in female hearts 14 days following MI and either IC infusion or EC injection. (A) Female porcine myocardium acting as negative control. (B) Positive control: porcine testis showing presence of Y-chromosome (arrow). (C) MSC demonstrating a Y-chromosome (arrows) within the infarct border zone following IC infusion. (D) MSC with Y-chromosome with the infarct zone following EC injection (arrow).

Analysis of representative samples of liver and lung demonstrated increased iridium signalling in these organs (Table 2). There was reduced engraftment in the lung following delivery by EC injection when compared with IC (P=0.06) or IV (P=0.02) infusion, respectively (Table 2). When considering the entire mass of the lungs (∼1 kg), this accounted for >20% of the total dose administered following IC or IV infusion techniques.


Cellular cardiomyoplasty has been conceptualized as an approach to treat acute cell loss resulting from MI, repopulate areas of chronic cellular loss in the setting of chronic ischaemic heart disease, and improve left ventricular function in patients with either ischaemic or non-ischaemic cardiomyopathy. These disease states may require different delivery strategies with regard to timing of and approach to delivery, delivered cell type, anticipated myocardial response (angiogenesis vs. cardiomyocyte replacement), and clinical endpoints (i.e. symptom relief vs. reduction in mortality). Hitherto, there have been no studies comparing the various approaches of stem cell delivery and engraftment in either large animals or humans. Therefore, this study was designed to compare quantitatively the three most common MSC delivery approaches following infarction in a large animal model with cellular engraftment as the endpoint. The procedure was performed in a manner that is consistent with potential clinical approaches to MSC delivery, i.e. in a fully anti-coagulated animal and a patent infarct related artery following a large MI. The results show that using the parameters outlined earlier, (i) IC delivery of MSCs was associated with a significantly greater number of engrafted MSCs when compared with EC or IV delivery, (ii) IC delivery was also associated with decreased distal arterial blood flow and greater myocardial injury, potentially obviating the advantage of such a delivery approach, (iii) no measurable directional homing of MSCs to areas of myocardial injury was noted after IV delivery, and (iv) local EC delivery appeared more site specific with less systemic engraftment than the other delivery methods.

The optimal cell delivery technique is that which provides the most therapeutic benefit, i.e. recovery of myocardial or vascular function. However, the relationship between cell retention, engraftment, timing of delivery, and subsequent therapeutic benefit remains unknown. This study was designed to evaluate cell delivery techniques solely on short-term cell retention or engraftment, as this endpoint provides specific information regarding cell delivery parameters. This study was not designed to evaluate the differential functional response of modified myocardium after one of the delivery approaches, as these studies require further insight into the optimal delivery techniques.

Cells were labelled with internalized iridium nanoparticles rather than more traditional methods that utilize stains or genetic manipulations, in order to permit a quantitative comparison between delivery methods with regard to cellular retention. This technique was selected, as it allows for assessment of label retention throughout the entire infarct and border zones. As such, this method represents an improvement over cell distribution assessments based on a limited number of histology slides or techniques using β-galactosides or other reporter genes in which the assessment of distribution and engraftment is limited to the surface of ‘bread-loafed’ sections.14 Histological assessment enables evaluation of cell viability, morphology, and context within the tissue; however, those techniques sample only a small fraction of the targeted infarct zone (1000 high power fields samples represent <0.1% of a 30 g infarct). This may lead to high variability when comparing delivery methods which yield heterogeneous cell distribution (i.e. multiple injections).

Pilot results demonstrated that the internalized iridium technique was sensitive enough to detect 0.06% of the total dose or as few as 30 000 cells, a 10-fold lower than the anticipated cell engraftment at 14 days. It cannot be confirmed that the entire measured iridium label in vivo was associated with viable, engrafted cells. Acutely retained cells that later die will release iridium nanoparticles. If not cleared, these particles will be counted as engrafted cells and result in overestimation of engrafted cell number. However, residual extracellular label still results from acutely retained cells, so the values reported at 14 days (Table 2) may be considered a composite of acutely retained and engrafted cells, although the presence of macrophages containing iridium nanoparticles after MSC death cannot be excluded. Therefore, this technique enabled analysis of all infarct tissue and was capable of detecting small numbers of engrafted cells, but did not provide confirmation that only viable, engrafted cells were counted. However, the quantitative comparison of delivery methods remains valid and FISH and DiI labelling confirmed that viable, transplanted cells were present in the targeted tissue and were present at 7 and 14 days.

Currently little is known regarding the optimal delivery parameters (e.g. MSC concentration), so the statistically greater number of engrafted MSCs following IC administration compared with EC should be viewed within the limitations of the study design. As IC was associated, in half of the deliveries, with decreased blood flow and increased myocardial damage, the dose used (50×106 cells) may be close to maximal because any increase in the concentration may result in a greater likelihood of ‘no reflow.’ With regard to IV delivery, several groups have suggested that cardiac recruitment of stem cells requires both myocardial injury and expression of stromal-derived factor 1α (sdf-1).15,16 Sdf-1 upregulation occurs immediately following myocardial injury,15 although forced expression using adenoviral gene delivery may improve homing.16 Thus, in order to improve cardiac implantation after IV delivery, simultaneous direct injection of sdf-1 may be necessary. The optimal dose of MSCs delivered by an EC approach is also unknown but potentially could be increased by altering parameters such as number of injections, volume per injection, and/or cell concentration of the injectate, resulting in increased engraftment. Ultimately, EC delivery may prove more efficacious in improving delivery efficacy and reducing the systemic effect of a more systemic delivery. We did not evaluate the retrograde coronary sinus method of MSC delivery. This is another potential approach to access the myocardium and has demonstrated high delivery rates. Engraftment efficacy, however, has not been tested.

Decreased blood flow during a percutaneous coronary intervention (called no-reflow) is well known to interventional cardiologists and is defined as decreased distal blood flow and reduced tissue perfusion in the absence of a flow-limiting coronary lesion.17 It is often observed following percutaneous coronary intervention for MI or thrombotic lesions. The cause is incompletely understood but is thought to result from a combination of tissue oedema, neutrophil plugging of microvasculature,18 and microvascular spasm.19 Hence, it is not unanticipated that MSCs delivered directly into the culprit coronary artery following an MI resulted in no-reflow, as MSCs are of a size (∼10–20 µm) that could cause capillary plugging (Figure 8). This angiographic observation is important because no-reflow has been associated with increased infarct size and expansion, greater degree of scar thinning, and an increased incidence of long-term cardiac complications, potentially related to adverse left ventricular remodelling.20,21 In the current study, increased fibrosis and thinning of the infarcted myocardium were noted in those animals that suffered no-reflow.

Figure 8

Photomicrographs obtained from adjacent sections of myocardium demonstrating DiI labelled MSCs within a coronary arteriole. Note ‘plugging’ of the arteriole in the haematoxylin and eosin stained section. The DiI stained section confirms that the majority of cells is MSCs.

In a clinical setting, in which a percutaneous coronary intervention is performed to improve myocardial function, development of no-reflow following IC cellular delivery may mitigate the potential improvement in function afforded by MSC delivery. Previous studies evaluating the efficacy of bone marrow MSCs delivery by IC infusion have yielded conflicting results. Chen et al.22 did not comment on a potential reduction in blood flow after infusion of 6 mL of MSC solution (8–10×109 cells in total), infused a mean of 18.4 days after infarction and revascularization. No data were presented on the incidence of CPK release following cell delivery, which, if present, indicates that distal embolization and microvascular plugging may have resulted from IC delivery. Conversely, Vuillet et al.23 observed ST changes and microinfarcts following IC delivery of mesenchymal stromal cells in dogs. Clinically, Kang et al.24 noted that a mild increase in CPK MB after cell infusion, although coronary flow reserve measurements were not reduced. It is possible that the reduced blood flow observed in the current study may have resulted from the short time period between infarction and MSC infusion. However, the finding of reduced blood flow by angiography as well as immunohistochemical evidence of microvascular plugging should alert the clinician to a potential limitation of IC infusion. This finding should be substantiated in subsequent studies, with functional endpoints, and attempts should be made to identify the cause and potential approaches to reduce no-reflow associated with IC delivery of cells.

Although IV delivery of cells has been promulgated as a potential simple delivery strategy, insofar that circulating bone marrow cells may differentiate into cardiac myocytes after MI data evaluating the concept of MSCs homing to the myocardium is contradictory. Our results support previous studies showing minimal uptake when using a systemic approach.25,26 Kuramochi et al.21 demonstrated only a 0.04±0.02% rate of new cardiomyocytes from circulating bone marrow cells 1 month after infarction, a level lower than our detection limit. Barbash et al. demonstrated that local arterial infusion into the rat left ventricular cavity was associated with increased myocardial migration and colonization of myocardial MSCs when compared with an IV approach. These investigators also demonstrated increased uptake by lung, hepatic, and splenic tissue after IV delivery.26 Other investigators have reported homing of MSCs to recently infarcted myocardium.1,27 Kraitchman et al.27 demonstrated focal myocardial uptake of MSCs persisting for 7 days following IV delivery in a canine model. Increased uptake by the lungs, liver, spine, and spleen were noted. Studies to understand the link between functional improvement and MSC homing as well as studies to understand whether there is an effect of cell labels on homing will help to settle this controversy. Nonetheless, the effects of remote organ engraftment are unknown and may be benign; however, engraftment may result in uncontrolled cellular growth, differentiation, or malignant transformation in remote organs. During necropsy, we did not observe abnormal growths and are unaware of any such reports. However, given the functional plasticity of such cells, the potential effects of such remote delivery should be considered when designing clinical studies.

The low engraftment rates observed after all delivery approaches is not unexpected, as it is generally accepted that 1–5% of delivered cells actually engraft within the infarct zone. The low observed cellular retention rates may relate to transport of cells out of the myocardium via venous effluent, leading to entrapment within the lung, liver, or spleen or apoptosis of MSCs within the infarct milieu.28 In vitro testing suggested that the labelling technique used did not affect MSC proliferation and it is not thought that the low engraftment rates are related to the labelling technique. Other groups have also successfully incorporated particles, of this size range, into cells2931 and MSCs have been observed to engraft specifically in recently infarcted myocardium following arterial delivery.26

There are potential limitations in this study. The optimal time point for MSC delivery is unknown. We chose to deliver immediately after infarct revascularization, as this approach is the most clinically relevant scenario. However, a longer lag time between infarction and cell infusion/injection may be optimal. Given the potential role of sdf-1 in improving cardiac recruitment,16 the results obtained after a longer lag time may differ. The delineation of infarct size was grossly performed in unstained tissue. Tri-phenol-tetrazolium-chloride (TTC), the ‘gold standard’, assessment of infarct size could not be performed as TTC staining effects iridium measurements. It is possible that resection of infarcted tissue was not precise, thereby effecting engraftment results. However, the number of engrafted cells per gram of tissue should not be grossly altered. Finally, we did not assess MSC phenotypic changes or functional effects of the MSCs, so the effects of the differential engraftment rates are unknown.


IC and EC injection of MSCs post-MI resulted in increased engraftment, at the 14-day follow-up time within infarcted tissue, when compared with IV infusion. IC delivery was more efficient than EC delivery, although it was also associated with decreased coronary blood flow. EC delivery into acutely infarcted myocardial tissue was safe and well tolerated and was associated with decreased remote organ engraftment when compared with IC and IV deliveries. Given the decreased coronary blood flow observed after IC delivery, local EC delivery may be a preferable method for MSC delivery.


The assistance of Adolf Dossman and Oleg Shleyfer is gratefully acknowledged. In addition, the authors are grateful for the assistance provided by Sharron McCulloch and Timm Varney of Osiris Therapeutics.

Conflict of interest: T.F. and M.P. are employees of Boston Scientific Corp., manufacturers of the Stilletto™ catheter; R.L.W. is the recipient of grant funding from Boston Scientific Corp. and is a member of a Scientific Advisory Board of Boston Scientific Corp.


View Abstract