OUP user menu

Prostacyclin improves transcoronary myocardial delivery of adipose tissue-derived stromal cells

Rosalinda Madonna , Lucia Rinaldi , Cosmo Rossi , Yong-Jian Geng , Raffaele De Caterina
DOI: http://dx.doi.org/10.1093/eurheartj/ehl154 2054-2061 First published online: 24 July 2006


Aims Heart transplantation of adipose tissue-derived stromal cells (ADSCs) is under evaluation as a therapy for cardiac repair. Prostacyclin (PGI2), a vasodilator with additional effects on platelet aggregation and blood cell adhesion, exerts cardioprotection and might favour the myocardial delivery of ADSCs. We investigated the engraftment and influence on cardiac function of the transcoronary delivery of ADSCs and the effects of PGI2 compared with nitroglycerin (NTG) and adenosine (Ado) in isolated–perfused mouse hearts.

Methods and results Infusion of ADSCs at <1×106 cells/mL caused no significant changes in contractility and rhythm, whereas higher cell doses caused cardiac dysfunction. Perfusion with PGI2, NTG, and Ado concentration-dependently increased coronary flow (CF). Perfusion with PGI2, at variance from NTG and Ado, increased ADSC delivery and entrance into the myocardial interstitium without affecting ventricular or metabolic functions and CF (engrafted ADSCs, as percentage of control, at doses producing 50% of maximum vasodilation: PGI2: 220±12, P<0.001; NTG: 110±8, P=N.S.; Ado: 80±5, P=N.S.).

Conclusion PGI2 safely increases myocardial delivery of ADSCs, by mechanisms independent of its vasodilatory properties, with a potential for its use in cell therapy for cardiac repair.

  • Adipose tissue
  • Stem cells
  • Cell therapy
  • Prostacyclin
  • Cardiac repair
  • Coronary vasodilators


Transplantation of stem or progenitor cells is a promising strategy for cardiac repair, currently attracting increasing attention. Although a variety of stem and progenitor cells may be used for this purpose, so far only skeletal myoblasts and bone marrow-derived stromal cells have been tested in safety studies and clinical trials.16 The stromal vascular fraction freshly isolated from the adipose tissue has recently attracted much attention because of its angiogenic capacity to differentiate into mature endothelial cells and to form new blood vessels.7,8 Stromal stem cells freshly isolated from animal adipose tissues have also shown to transdifferentiate into contractile myogenic cells.9

One method for cell transplantation into the heart is the intracoronary delivery, which has advantages in less myocardial and graft injury compared with trans-epicardial or trans-endocardial methods.10 Intracoronary infusion is, however, limited in the possibility of delivering large number or large size of stem cells, such as myoblasts, for their potential for macro- and microembolization and poor rate of retention into the myocardial interstitium.

We reasoned that combining stem cell therapy with pharmacological treatments helping to decrease the risk of coronary thrombosis and embolization and increase cardioprotection might provide added benefits compared with stem cell therapy alone. Prostacyclin (PGI2), a microcirculatory vasodilator with additional effects on platelet aggregation,11,12 cell–cell interactions,13 cell permeability,14 and blood cell adhesion to injured endothelium15 or extracellular matrix,13 exerts cardioprotection in ischaemia/reperfusion and inflammation-related damage.16 Therefore, in the present study, using an ex vivo model of mouse heart Langendorff perfusion, we have investigated the efficiency of ADSC myocardial engrafting after intracoronary administration and its influence on myocardial function. Secondly, we have analysed the efficacy and safety of PGI2 in comparison with other vasodilators, i.e. nitroglycerin (NTG) and adenosine (Ado), in increasing transcoronary delivery of ADSCs.


Animal care

All studies were performed with the approval of the Institutional Ethics Committee for Animal Research. The investigations conformed to the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.


All reagents were from Sigma (Sigma, St Louis, MO, USA), unless otherwise specified. A stock solution of Ado (MP Biomedical, Irvine, CA, USA) was prepared in distilled water; a stock solution of PGI2 was prepared in absolute ethanol. NTG was purchased from AstraZeneca (Basiglio, MI, USA) as alcoholic solution. Serial dilutions of vasodilators to desired concentrations were made in distilled water. At dilutions achieved, a preliminary subset of experiments from BalbC mice had demonstrated no effect of ethanol in some of the solutions. All concentrations tested in the experiments included drug concentrations achieved in blood during intravenous infusions.1719

Preparations of cells for the intracoronary administration

Peri-epidydimal adipose tissues from BalbC and Rosa26 LacZ+ mice (Charles River Laboratories, Margate, Kent, UK) was harvested, and adipose stromal cells (ADSCs) isolated using a modification of published methods.20 In brief, adipose tissue was mechanically minced and digested with collagenase. After adipocyte removal, the vascular stromal fraction was plated (at 10 000 cells/cm2 density) in DMEM medium supplemented with penicillin (100 U/mL), streptomycin sulfate (100 µg/mL), and 10% fetal bovine serum. After 24 h, non-adherent cells were removed, and adherent ADSCs passaged until 80% confluence and used before passage 4. Before perfusion, ADSCs were incubated with 50 µmol/L bisbenzimide at 37°C for 15 min. Cells were resuspended in a modified Krebs–Henseleit buffer (Krebs), containing 120 mM NaCl, 25 mM NaHCO3, 4.7 mM KCl, 1.2 mM KH2PO4, 2.5 mM CaCl2, 1.2 mM Mg2SO4, 15 mM glucose, and 2 mM pyruvate, and then filtered through a 75 µm nylon filter immediately before perfusion.

Heart preparation and perfusion

We used a standard Langendorff isolated-perfusion system for adult BalbC mice (20–30 g) of either sex. Details of the procedure are reported in the Supplementary material online.

Experimental protocol

All hearts were equilibrated for 10 min before each experiments. Four types of heart perfusions were performed sequentially, with mice in each group selected at random: (i) untreated hearts perfused with Krebs (n=24); (ii) hearts perfused with ADSCs (n=24); (iii) hearts perfused with vasodilators (PGI2 or NTG or Ado) alone (n=18 for each drug); (iv) hearts pretreated with vasodilators for 10 min and then perfused with ADSCs + vasodilators for 30 min (n=18) in order to have a minimum of three replicates for each single point of each dose–response curve. Therefore, all three vasodilators were tested in groups 3 and 4.

Baseline data for CF, HR, and developed pressures were sampled at the end of equilibration. Functional assessment of hearts after cell transplantation was done using increasing concentration of cells (105–0.5×106–106–2×106–4×106 cell/mL) at different passages (from passage 0 to passage 3). Concentration–response curves were constructed for each drug infusing increasing concentration of vasodilator (PGI2: 10−8–10−6 mol/L; NTG: 10−7–10−5 mol/L; Ado: 10−6–10−4 mol/L) into the perfusion system. All experiments aimed at evaluating the effects of vasodilators on transcoronary delivery of ADSCs were done using primary cultures (passage 0) of ADSCs at 105 cell/mL concentration. Data were recorded at the end of the 30 min infusion period.

Assay for lactate dehydrogenase and cardiac Troponin I

Evaluations of the metabolic damage after the reperfusion procedure were done on the effluent samples measuring markers of myocardial necrosis, such as the concentration of cardiac specific troponin-I (cTnI) and the activity of lactate dehydrogenase (LDH), using commercially available kit (Life Diagnostics, Inc., West Chester, PA, USA). The total release of cTnI and LDH was calculated as the cumulative efflux during 30 min of reperfusion and expressed as percentage of Krebs-perfused value.

Histological evaluation of bisbenzimide-labelled ADSCs

Evaluations of the number and distribution of ADSCs present in the cardiac tissue after the reperfusion procedures were performed counting blue-fluorescent, bisbenzimide-positive cells. Myocardial tissue specimens were frozen in OCT (Miles, Elkhart, IN, USA) immediately after procedure and stored at −80°C. Five micrometre sections were obtained from each specimen, then observed and photographed at 5× and 10× magnification under a fluorescence microscope. An average of 20 fields in each section were examined by two independent observers and photographed.

The distribution of ADSCs in the cardiac tissue after the reperfusion procedure was evaluated by indirect immunofluorescence for the endothelial cell marker platelet-endothelial cell adhesion molecule-1 (PECAM-1). Details of this procedure are reported in the Supplementary material online.

Histological evaluation of β-galactosidase expression

β-Galactosidase (β-Gal) expression in ADSCs harvested from Rosa26 LacZ+ mice seeded in eight-well chamber slides was first confirmed with β-Gal staining in vitro and then assessed in tissue sections after reperfusion. Each heart after perfusion was divided in two parts, one for the measurement of β-Gal activity and the other for β-Gal staining. After Langendorff perfusions, specimens for the assay of β-Gal activity were immediately frozen in liquid nitrogen, whereas specimens for β-Gal staining were frozen in an embedding medium, directly after cell or control Krebs perfusion and sectioned frontally, transversely, and sideways on a cryostat (Leica, Leitz, Wetzlar, Germany). Sections were fixed with 0.2% glutaraldehyde for 10 min, rinsed with PBS twice, and incubated with X-gal working solution (5 mmol/L K3[Fe(CN)6], 5 mmol/L K4[Fe(CN)6], 2 mmol/L MgCl2, 0.5 mg/Ml 5-bromo-chloro-3-indolyl-β-D-galactoside) at 37°C for 24 h. Slices were mounted with aqueous mounting medium (Biomedica, Foster City, CA, USA) and visualized by light microscopy.

Assay for β-Gal activity

Assay for β-Gal activity was performed using standard methods to assess the amount of transplanted cells existing in the heart (see Supplementary material online).

Assay for NO production

The nitrite accumulation in the coronary effluents, taken as an index of NO production, was determined using the Griess reagent (1% sulfanilic acid and 0.1% N-[1-naphtyl] ethylenediamine–HCl in 5% phosphoric acid), as reported previously.21

Statistical analyses

All results are presented as means±SD. In selected cases, means and SD are expressed as percentage of the control group means and SD. Values of concentrations producing 50% of the maximum responses (EC50) were derived plotting the concentration of drug from the concentration–response curve on the abscissa as a function of the effect measured on the ordinate and defined as the drug concentration at which CF was half-maximal. Differences in dose–response at individual drug concentrations were analysed by Student's t-test for two-group comparisons and by analysis of variance (ANOVA) for multiple group comparisons after checking for normality of distributions. The existence of individual differences, in case of significant F values at ANOVA, was tested by Scheffé's multiple contrasts. For all tests, statistical significance was set at P<0.05.


Perfused ADSCs enter myocardial interstitium

After stabilization, isolated hearts were perfused intracoronarily for 30 min with increasing concentrations of LacZ+ ADSCs (at 0, 105, 0.5×106, 106, 2×106, 4×106 cell/mL). In addition to carrying the LacZ gene, the nuclei of ADSCs were labelled with the membrane-permeable DNA-binding dye bisbenzimide, at sublethal concentrations. The estimate of ADSC retention rate in the hearts after perfusion, evaluated by both β-Gal activity and count of fluorescent cells, indicated increasing number of ADSCs retained within the myocardium after perfusion with increasing concentration of ADSCs (Figure 1; Supplementary material online, Figure S1). To localize the cardiac distribution of ADSCs, the hearts subjected to LacZ+/bisbenzimide-labelled ADSC transfusions were collected at the end of the 30 min perfusion, sectioned, and stained for β-Gal with X-gal and for the endothelial cell marker CD31 (PECAM-1). We observed cells positively stained in blue for β-Gal activity and for bisbenzimide (blue fluorescence) in the myocardium throughout the distribution territories of both the left and the right coronary arteries (RCA) and throughout all cardiac layers, where they appeared to permeate the myocardial interstitium (Figure 1B and C). No β-Gal-positive or bisbenzimide-positive cells were found in the control hearts without cell transfusions. High-power view at microscopy revealed the presence of β-Gal-positive cells with blue fluorescent nuclei in the myocardial interstitium. ADSCs were clearly distinguishable from PECAM-1-positive endothelial cells and appeared adjacent to blood vessels of the microcirculation.

Figure 1

(A) Retention rates, expressed as β-Gal activity or fluorescent cell count of LacZ+/bisbenzimide-labelled ADSCs in mice hearts after cell perfusion. Increasing concentrations of LacZ+/bisbenzimide-labelled ADSCs (0–105–0.5×106–106–2×106–4×106 cells/mL, at passage 0) were transfused into the coronary circulation of isolated–perfused mouse hearts. Sections from frozen hearts were observed under fluorescence microscope, and blue-fluorescent engrafted cells were counted by two independent observers. An average of 20 fields in each section were examined and photographed. Control hearts showed no positive staining or β-Gal activity assay (n=3 for each treatment group). See Supplementary material online for a colour version of this figure. (B and C) Histological findings of LacZ+/bisbenzimide-labelled ADSCs after intracoronary delivery. (B) In situ staining for β-Gal shows postitive (dark blue) ADSCs disseminated throughout the LV free wall 30 min after transfusion of 105 ADSCs/mL. Inset: β-Gal staining of LacZ+ ADSCs before perfusion. (C) Immunofluorescence staining showing blue-fluorescent, bisbenzimide-labelled ADSCs aligned with microcirculatory blood vessels (green fluorescence, after staining with a fluoresceine-labelled antibody against the endothelial marker CD-31). ADSCs with blue fluorescent nuclei here do not appear stained with the anti-CD31 antibody. Scale bar: 25 µm for (B and C). (D, E and F) Concentration-dependent decline (expressed as percentage of Krebs-perfused control) of HR, LV end-diastolic pressure (LVEDP), and LV end-systolic pressure (LVESP) in isolated–perfused mice hearts submitted to ADSC perfusion. Values are mean±SD; n=3 for each treatment group; *P<0.05 vs. untreated controls.

Cell-perfused hearts retain a normal function up to a critical concentration of transfused ADSCs

Cardiac function was evaluated to assess the maximum number per volume unit of perfused ADSCs compatible with a normal cardiac performance. We did not observe, at the selected time points, any functional reduction in cardiac performance and any release of the cardiac-specific injury marker troponin I from the hearts, after perfusion with freshly isolated, non-passaged ADSCs (passage 0) at <1×106 cells/mL, compared with Krebs-treated hearts in terms of HR, −dp/dtmax, +dp/dtmax, LVESP, and LVEDP (Tables 1 and 2; Figure 1D, E, and F).

View this table:
Table 1

Baseline functional data after cell perfusion

View this table:
Table 2

Cumulative release of cTnI after 30 min perfusion with increasing concentration of ADSCs

At higher cell numbers (≥1×106 cells/mL) or higher cell passage, we observed cardiac dysfunction and significant cardiac damage. This concentration-dependent change in heart function indicated the cell number of 0.5×106 cells/mL as the maximum possibly useful ADSC concentration for cardiac ADSC perfusion in mice.

Under control conditions, heart rate (HR) remained constant, and the contraction–relaxation cycle or heart rhythm were regular during and at the end of the 30 min perfusions (Supplementary material online, Figure S2A). Mean LVESP and LVEDP were 15 and 110 mmHg, respectively. Transfusion of freshly isolated ADSCs at passage 0, at numbers up to 0.5×106 cells/mL, did not cause any major changes in LV pressure, HR, or heart rhythm (Supplementary material online, Figure S2B). Control-perfused hearts and hearts perfused with non-passaged cells had similar values of LVESP and LVEDP during the perfusion time (Supplementary material online, Figure S2B). With increasing cell numbers, however, there were gradual reductions in HR, maximum +dp/dt and maximum −dp/dt, LVESP, and LVEDP (Supplementary material online, Figure S2C; Table 1). We also observed that at higher passages, ADSCs triggered changes in heart function at lower concentrations compared with non-passaged cells (Supplementary material online, Figure S2D). Thus, from this standpoint, non-passaged, undifferentiated ADSCs appeared to be superior to differentiated ADSCs for transcoronary myocardial delivery.

Different vasodilators differentially affect CF in control- and ADSC‐perfused hearts

PGI2, NTG, and Ado all caused concentration-dependent increases in CF (vasodilation) in isolated hearts perfused at constant pressure (Supplementary material online, Figure S3). Maximal response to each drug did not vary significantly among vasodilators. For each drug used, we established values of drug concentrations producing 50% of the maximum increase in CF (EC50 values, Table 3) in order to subsequently compare the specific properties of the drugs in favouring ADCS myocardial delivery independent of their overall vasodilatory properties. Hearts transfused with ADSCs at 105 cells/mL showed no major changes in CF (Figure 2) compared with Krebs-perfused hearts. CF in the presence of PGI2 increased up to 172±10% of baseline (Krebs-perfused hearts, Supplementary material online, Figure S3A) at 10−6 mol/L concentration, and this maximal effect was non-significantly different from the value of 160±9% of baseline in ADSC-perfused hearts (Figure 2A). However, the maximal CF induced by NTG at 10−5 mol/L was significantly reduced from 142±43% (Supplementary material online, Figure S3B) to 107±16% in ADSC-perfused hearts (Figure 2B). Likewise, 10−4 mol/L Ado resulted in maximal CF of 144±5% of baseline (Supplementary material online, Figure S3C), whereas this effect was reduced to 112±14% in ADSC-perfused hearts (Figure 2C).

Figure 2

Changes in CF (expressed as percentage of Krebs-perfused control) in isolated mouse hearts after perfusion with ADSCs in the presence or absence of (A) PGI2, (B) NTG, and (C) Ado. Values are means±SD; n=3 for each treatment group; *P<0.05 vs. untreated controls.

View this table:
Table 3

EC50values of various vasodilators and corresponding CFs in the perfused hearts

Different vasodilators differently affect nitrite levels but not cardiac injury markers in Krebs and ADSC-perfused hearts

We assayed nitrite accumulation and cardiac injury markers in the coronary effluents to evaluate possible differential effects of the various drugs used. As shown in Table 4, PGI2 (10−8–10−6 mol/L) and NTG (10−7–10−5 mol/L) but not Ado (10−6–10−4 mol/L) increased nitrite production in a concentration-dependent manner compared with Krebs-perfused hearts. The concomitant perfusion with ADSCs at 105 cells/mL did not significantly modify nitrite levels in hearts treated with the various vasodilators.

View this table:
Table 4

Cumulative release of cTnI and nitrate production after 30 min ADSC perfusion in the presence or absence of various vasodilators

There was negligible release of cardiac-specific (troponin I) or non-specific (LDH) injury markers from the hearts under basal conditions (Krebs-perfused hearts) and throughout the perfusion with ADSCs at 105 cells/mL, in the presence or absence of vasodilators. The total cTnI release during the 30 min of perfusion in different groups is shown in Table 4, indicating the absence of significant cardiac damage at drug concentrations used in the presence or in the absence of ADSC administration.

Prostacyclin increases retention rate of ADSCs

Perfused hearts were evaluated for the number of bisbenzimide-positive ADSCs before and after perfusion with vasodilators, observing the OCT-embedded sections under a fluorescence microscope at low-power (10×) magnification. Heart perfusions with NTG (10−7–10−5 mol/L) increased the number of ADSCs appearing in the myocardial interstitium non-significantly (absolute numbers of engrafted cells/field: basal, 80±8; NTG 10−7 mol/L, 83±9; NTG 10−6 mol/L, 88±2; NTG 10−5 mol/L, 144±5) (Figure 3B), whereas perfusion with PGI2 (10−8–10−6 mol/L) strongly increased it in a concentration-dependent manner without affecting ventricular and metabolic functions and without impairment of CF compared with PGI2 alone (absolute numbers of engrafted cells/field: basal, 80±8; PGI2 10−7 mol/L, 128±9; PGI2 10−6 mol/L, 168±5; PGI2 10−5 mol/L, 224±16) (Figure 3A). Ado (10−6–10−4 mol/L) caused no major changes in the number of engrafted ADSCs compared with ADSC-perfused hearts without Ado (absolute number of engrafted cells/field: basal, 80±8; Ado 10−7 mol/L, 88±12; Ado 10−6 mol/L, 64±3; Ado 10−5 mol/L, 85±2) (Figure 3C). Figure 4 shows distribution and number of engrafted ADSCs in the hearts after perfusion with vasodilators at EC50 values, clearly demonstrating that bisbenzimide-labelled cells with blue fluorescent nuclei were extremely more abundant in the hearts perfused with PGI2 than in those with the other two vasodilators, at the concentrations that caused similar elevation of CF.

Figure 3

Retention rates (expressed as percentage of ADSCs-transfused control) of ADSCs (injected at 0.5×106 cells/mL) in the hearts after perfusion in the presence or absence of (A) PGI2, (B) NTG, and (C) Ado. Values are means±SD; n=3 for each treatment group; *P<0.05 vs. untreated control. **P<0.05 vs. ADSCs-perfused controls. See Supplementary material online for a colour version of this figure.

Figure 4

Histological findings of ADSCs in the myocardial interstitium of hearts treated with EC50 of the various vasodilators; 105 cells/mL of blue fluorescent, bisbenzimide-labelled ADSCs were transfused into the coronary circulation in combination with vasodilators (PGI2, NTG, and Ado). Five micrometre sections were observed under a fluorescence microscope; ADCSs, found throughout the territories of both the right and the left coronary arteries and in all cardiac layers, were then counted by two independent observers. An average of 20 fields in each section were examined and photographed. Control hearts showed no positive staining (magnification 10×). n=3 for each treatment group. Scale bar: 25 µm.


Increasing the efficacy of stem cell delivery and of their entry into the myocardial interstitium is likely key to better outcomes in cardiac repair. However, coronary perfusion with a large number of stem cells may increase the risk of microembolism, which can cause ischaemic injury to the myocardium. Understanding the relationship between stem cell perfusion rates and the myocardial response can help design proper in vivo approaches to deliver stem cells into diseased myocardial tissue. Strategies to increase myocardial stem cell delivery to the myocardium with intracoronary administration include the use of higher numbers of transfused cells, higher infusion pressure, longer incubation time, or the use of drugs increasing vascular permeability.22,23 The present study was therefore aimed at clarifying the dose–response relationship of ADSC intracoronary administration with regard to myocardial damage and the dose–response relationship in terms of safety and efficacy in increasing the delivery of ADSCs by combinining ADSC administration and pharmacological therapies with known vasodilators.

We have shown here that the intracoronary administration of ADSCs up to a certain number per unit volume can provide safe and effective cell delivery and that the adjunctive use of PGI2 may further increase the efficiency of such ADSC delivery into the myocardial interstitium. Because our data on the β-Gal activity assay prove the viability of delivered ADSCs and therefore suggest engraftment of these cells in the myocardial interstitium, intracoronary ADSC administration for myocardial cell therapy in cardiovascular disease appears feasible also in vivo.

The possibility of using ADSCs as a reservoir of adult multipotent stem cells has been recently reported.7,8,24 Several research teams have recently initiated clinical trials in humans, delivering bone marrow cells into the heart at sites of infarction.4 Another approach for stem cell delivery is through intracoronary injection.5,6 However, crucial questions here remain as to how many cells should be used, how fast the injection should be performed, and how immediately safe is such an approach for the myocardium. These issues are likely variable according to the cell type used. In this study, we addressed some of these issues using ADSCs as a source of stem cells for intracoronary administration.

We first administered ADSCs at different concentrations or at different plating passages into isolated–perfused Langendorff hearts, a widely used system to study myocardial physiology and pathophysiology. We found that unmanipulated, non-expanded β-Gal-positive ADSCs can be fairly well delivered into the myocardial interstitium and, here, remain viable, with little risk of functionally evident coronary embolism and minimal reduction in global cardiac performance. As a further index of the efficiency of cell delivery into the myocardium, we also used bisbenzimide-labelled, blue-fluorescent cell count. We observed that although the number of β-Gal-expressing cells in the myocardial interstitium steeply increased proportionally to the concentration of cells in the perfusion solution, the rate of increase in the number of bisbenzimide-labelled ADSCs was lower, reaching a plateau at numbers of 4×106 cells/mL in the perfusion solution. Minor differences in the computation of cell-grafting efficiency are accounted for by differences of measurement systems, as the assay of β-Gal activity measures the total enzyme activity of the tracer into the heart, which also is likely deriving from diffusion from non-engrafted cells on the epicardial surface as a consequence of some spilling from the perfusion apparatus, although fluorescent cell count assesses only cells penetrated in the myocardial interstitium. Irrespective, however, of these minor differences, too high numbers of perfused cells per unit volume (≥1 million ADSCs/mL) exert harmful effects on the hearts, as documented by a gradual decline in functional performance and the eventual stop of heart beat. This occurred more frequently when higher passage cells were transfused. The mechanism by which high-passage cells more easily trigger heart dysfunction is not clear. This could be a consequence of cell differentiation or positive selection of mature cell phenotypes, such as fibroblasts or smooth muscle cells. In fact, it is known that differentiated cells have a lower nucleus/cytoplasm ratio than undifferentiated cells (also confirmed in our own observations). Confirmation of such findings in in vivo systems and studies of mechanisms involved would be now warranted on this point.

In studying the retention rate of ADSCs in the hearts after perfusion with various vasodilators, acting either mostly on the epicardial vessels (NTG) or on resistance vessels and the microcirculation (Ado and PGI2), we found that there is no increase in ADSC delivery and entrance into the myocardial interstitium by Ado and a possibly moderate increase of ADSC delivery and entrance after NTG (not significant in our experiments). Both drugs significantly increased CF when given alone, whereas the addition of sufficiently high number of cells prevented such vasodilation, likely through the occurrence of microcirculatory plugging. On the contrary, the addition of ADSCs to PGI2 strongly increased ADSC delivery and entrance in a concentration-dependent manner without affecting ventricular and metabolic functions and the changes in CF induced by the drug. This peculiar effect of PGI2, not shared by other vasodilators such as NTG and Ado, should be attributed to additional properties of PGI2, beyond vasodilation, which likely minimize the extent or consequences of macro- and microembolization occurring as a consequence of the administration of excessive ADCS numbers. PGI2 has been indeed shown to exert effects on platelet aggregation,11,12 cell–cell interactions,13 cell permeability14 and blood cell adhesion to injured endothelium,15 as well as cardioprotection in ischaemia/reperfusion damage.16 It is also theoretically possible that PGI2 acts indirectly stimulating some secretory activity by the infused ADSCs. A thorough investigation on these issues is, however, beyond the scope of this article.

Our ex vivo model of isolated–perfused Langendorff mouse heart is free from influences of systemic haemodynamic changes as well as of blood constituents, including blood cells, hormones, coagulation factors, and proinflammatory cytokines. Although a molecular understanding of the mechanisms underlying this peculiar effects of PGI2 are beyond the scope of the present investigation, it is likely that they depend on direct influences of the drug on cell–cell interaction and cell adhesion to injured endothelium.

Whatever the underlying mechanisms, this study demonstrates the feasibility and safety of ADSC intracoronary delivery as a potential new resource for cardiac repair. Perfusion with proper concentrations (<10−6 cells/mL) of these cells seems to be safe. The addition of PGI2 increases the delivery of ADSCs into the heart without compromising safety. The use of non-passaged cells also appears better in terms of engraftment compared with the use of expanded cells. The possibility of using PGI2 in combination with cell therapies should be confirmed in further studies and explored both in terms of underlying mechanisms and as to clinical implications. In vivo experimental studies are warranted to confirm these findings in more physiological situations where coronary arteries are perfused with blood containing already its own cells compared with our ex vivo system.

Supplementary material

Supplementary material is available at European Heart Journal online.


This work was supported by the Italian Ministry of Research through an ‘ex-60%’ grant, a grant to the Center of Excellence on Aging of the University of Chieti, a grant of the Consorzio Italiano Ricerche Cardiovascolari (all to R.D.C.), grants to Y.-J.G. from the National Institutes of Health (R01HL59249 and R01HL69509), Department of Defense T5 programme, and Texas Higher Education Board ARP/ATP/TDT programme. We thank Dr James T. Willerson, University of Texas Health Science Center at Houston and Texas Heart Institute, for proofreading and commenting on this manuscript.

Conflict of interest: none declared.


View Abstract