Introduction

As the regenerative capacity of mature cardiomyocytes is minimal, infarcted myocardium undergoes reparative fibrosis and is replaced by scar tissue. These myocardial reparative mechanisms lead to diminished myocardial function. The subsequent development of congestive heart failure represents the leading cause of hospitalization in the industrialized world and is not sufficiently affected by current therapeutic approaches. Thus, cell-based therapies for myocardial infarction (MI) have been met with great enthusiasm. Various studies have demonstrated improved cardiac function in animal models of MI and have implicated the use of myoblasts, bone marrow or stem cells as a source of precursors in the regeneration of infarcted myocardium [7, 12, 20, 29, 43].

In addition, several small clinical studies have led to varying degrees of improvement in myocardial perfusion and contractile function [10, 28, 39, 40, 44].

However, recent landmark studies [3, 24, 26] have demonstrated that hematopoetic stem cells injected into infarcted myocardium overwhelmingly adopt mature hematopoetic states and that generation of cardiomyocytes occurs, if at all, through cell fusion on rare occasions. In the light of these findings, the addition of contractile elements by cell-based therapy appears increasingly unlikely as mechanism underlying the observed functional improvement. Nevertheless, cell TX may be associated with the release of growth factors eliciting cell proliferation and survival as well as accelerating neovascularization and subsequent healing processes [17, 19, 25, 35, 36, 40]. In addition, stem cells and bone marrow-derived angioblasts have been demonstrated to integrate into vascular structures and to promote angiogenesis [2, 5, 17, 22] and ex vivo expanded endothelial progenitor cells (EPC) have been shown to contribute to myocardial neovascularisation and have a beneficial impact on preservation of left ventricular (LV) function in a rat model of MI [14]. Furthermore, several studies have assessed the importance of growth factors and proinflammatory cytokines in the context of MI [4, 8, 31] and congestive heart failure [6].

In previous experiments [23], we were able to show that TX of human umbilical vein endothelial cells (HUVEC) triggered an infiltration with macrophages. The structural and functional improvement observed in this study appeared to be macrophage mediated. Furthermore, TUNEL staining revealed that a major proportion of HUVEC were apoptotic 1 day after injection suggesting that cell debris may also have contributed to the observed inflammatory response.

Here, we address the influence of inflammatory processes on cell-based therapy by investigating the effect of human fibroblasts, macrophages and microspheres (uniform polystyrene microspheres, 10 µm diameter) directly transplanted into infarcted myocardium four weeks after MI and their effect on infarct size, LV dimensions and LV function.

Materials and methods

Induction of myocardial infarction

Female adult Sprague–Dawley rats (200–250 g, 8–10 animals per group) were intubated under general anesthesia (1 ml/kg ketamine and 10 mg/kg xylasine, intraperitoneal) and positive pressure ventilation was maintained (room air supplemented with oxygen) using a rodent respirator. Hearts were exposed through a 2-cm left thoracotomy and MI was induced by suture occlusion of the left anterior descending artery (LAD) between the left atrium and the right pulmonary outflow tract (7/0 polyprolene, Ethicons). The muscle layer and skin incision were closed layer by layer with silk suture. Animal experiments were approved by local authorities and complied with German animal protection laws.

Cells

Fibroblasts were isolated from human dermis after removal of epidermis by dispase treatment over night at 4°C. Subsequently, the dermis was incubated with collagenase for 1 h at 37°C. Cells were washed and grown in DMEM (high glucose) with 10% fetal calf serum and 1% l-glutamin. In further groups, human monocytic Mono Mac 6 cells (MM6) and microspheres (FluoSpheres, uniform polystyrene microspheres, 10 µm diameter—comparable to diameter of transplanted cells, yellow green fluorescent (505/515) for blood flow determination—1.0 × 106 beads/ml, molecular probes) were used for TX. One day prior to TX, labeled MM6 cells and fibroblasts were incubated with BrdU (Zymed) as described previously [37]. Thus, in our experimental setting, BrdU was applied as a marker for the presence of transplanted cells.

Cell transplantation

TX was performed 4 weeks after MI. The rats were anesthetized and the hearts were exposed through thoracotomy as described above. The areas for injection were chosen by visual identification based on surface scarring and wall motion akinesis. Cells and microspheres were re-suspended in 100 µl of phosphate buffer saline (PBS) and injected into marginal zones of the MI by syringe injection (for 1-min injection time) at three distinct but adjacent sites. Animals were divided into four groups: the control group (n = 10) animals received only culture medium (RPMI, 100 µl), the second group (n = 8) received human MM6 cells (1 × 106), the third group (n = 8) received human fibroblasts (1 × 106) and the fourth group (n = 8) received microspheres (1 × 106).

Echocardiography

Four weeks after MI (prior to transplantation) and 2 months after TX, rats were anesthetized and two-dimensional and M-mode measurements were performed with a 15-MHz linear phased-array probe connected to a SONOS 7500 echocardiography unit (Philips). The animals were placed in the supine-lateral position and excessive pressure on the thorax was avoided. Parasternal long-axis and short-axis views of the left ventricle (LV) were obtained, ensuring that the mitral and aortic valves and apex were well visualized. Area fraction and wall area were determined by planimetry of end-diastolic and systolic volumes in parasternal short axis. Measurements of LV end-diastolic and end-systolic dimensions were obtained in M-mode at mid-papillary level from more than three beats and FS was calculated as FS(%) = (LVIDd − LVIDs)/LVIDd) × 100, where LVID is LV internal diameter, s is systole and d is diastole. Areas of maximum measurable infarct size were chosen, diastole is defined as maximum measurable area, systole is defined as minimum measurable area.

Langendorff perfusion and assessment of infarct area

Two months after TX, the rats were anesthetized and hearts removed. Heart function was analyzed applying a Langendorff apparatus with filtered Krebs–Henseleit buffer at a pressure of 65 mmHg equilibrated with 5% CO2 and 95% O2. A latex balloon was placed into LV to measure pressure and heart rate, registered with a Watanabe Mark VII plotter. After 30 min of stabilization, coronary flow was assessed by timed collection. Preload was increased from 5 to 10, 15 and 20 mmHg. Systolic pressure was recorded at the different preloads and developed pressure was calculated. Subsequently isoproterenol (70 nmol/l) was diluted in Krebs–Henseleit buffer and infused continually for 5 min. LV pressure development was measured during catecholamine infusion. The scar size of LV-free wall was assessed by computed planimetry of digitized images (Diskus software, Hilgers) taken from hearts fixed in distension (30 mmHg) with 10% formalin and cut into 5 µm slices.

Identification of the cells

Infarct size was assessed by computer-assisted planimetry (Diskus software, Hilgers) measuring scar dimension. Scar area is expressed as percentage of total ventricle area derived from four different sections.

For cell identification, slides (5 µm) were stained with anti-BrdU kit (Zymed). Universal quick kit and alkaline phosphatase substrate kit (Vector Laboratories) were used with anti-CD31 antibody (Santa Cruz Biotechnology) for vessel identification. The nuclei undergoing apoptosis were stained with MEBSTAIN apoptosis kit II (MBL). Accustain trichrome stain (Masson, Sigma) was used to determine collagen content of infarcted regions. The stained areas were measured by computer-assisted planimetry (Diskus software, Hilgers). Macrophages and Neutrophiles were stained with α-naphthyl acetate esterase (Sigma) and naphthol as-d chloroacetate (Sigma), respectively. Lipofuscin was stained with saturated solution of sudan black (Sigma).

Data analysis

Data represent mean ± SD. Statistical analysis was performed with Prism 4 software (Graph Pad) using unpaired Student t test and 1-way ANOVA followed by Newman–Keuls test where applicable. Differences with P < 0.05 were considered significant.

Results

Four weeks after ligation of the left coronary artery, fibroblasts (n = 8), macrophages (n = 8), medium (n = 10) and microspheres (n = 8) were injected into infarcted myocardium and adjacent border zones. Transplanted cells were previously labeled with BrdU to allow further tracking.

Histology involved (1) tracking of transplanted cells by immunohistochemistry, measurement of (2) apoptosis, (3) monocyte infiltration, (4) tissue inflammation/lipofuscin staining, (5) vascular density, (6) collagen content. Myocardial function was evaluated by 2D echocardiography prior to injection and 2 months later as well as by isolated heart studies 2 months after TX (retrograde perfusion according to Langendorff).

Tracking of transplanted cells (BrdU labeling) and microspheres

Two months after injection, microspheres could be identified through their autofluorescent signal in clusters surrounded by fibrous tissue (Fig. 1). In the group of fibroblast transplantation, only rare and disorderly BrdU signals, surrounded by fibrous tissue, could be detected in host myocardium. In comparison, no BrdU labeled cells could be identified in the macrophage transplanted group. A possible explanation for this phenomenon might be the macrophages’ ability to migrate.

Fig. 1
figure 1

Transplanted cell population in infarcted areas at 2 months. Immunohistochemical BrdU staining of infarcted areas 2 months after cell or microsphere transplantation shows only rare BrdU signals (see arrows) in areas transplanted with fibroblasts. Microspheres were detected by characteristic round shape. Objective 40×, scale bars 25 µm. Inset in left upper corner shows a positive control for BrdU

Full size image

Apoptosis

Fig. 2a shows TUNEL staining demonstrating significant differences between groups after medium and macrophage TX compared to groups transplanted with fibroblasts and microspheres. Significantly fewer apoptotic cells were found in infarcted areas after fibroblast (232.2 ± 19.1 cells/mm2) and microsphere (233.2 ± 16.8 cells/mm2) injection compared to the number of apoptotic cells after macrophage (328.6 ± 37.4 cells/mm2) and medium (338.7 ± 16.5 cells/mm2; P < 0.05) injection.

Fig. 2
figure 2

Apoptotic cells in infarcted areas. Note the increased number of apoptotic cells in medium and macrophage treated hearts (a green TUNEL pos. nuclei, objective 40×, scale bar 25 µm, *P < 0.05 vs. control and MM6, insets show blue DAPI staining for all nuclei). Vessels per mm2 (b red CD 31 positive vessels, objective 20×, scale bar 50 µm) and collagen content (c blue, scale bar 100 µm) did not differ significantly

Full size image

Monocyte infiltration

As an indicator of inflammation we assessed monocyte infiltration in all groups by staining for α-naphthyl acetate esterase. Monocyte infiltration was significantly more pronounced in fibroblast and microsphere-injected hearts (fibroblasts 94.7 ± 7.1 cells/mm2, microspheres 82.2 ± 3.0 cells/mm2) compared to macrophage und medium-injected hearts (macrophages 56.0 ± 9.9 cells/mm2, medium 46.4 ±  9.0 cells/mm2, P < 0.05; Fig. 3a).

Fig. 3
figure 3

Numbers of neutrophiles (PMNs, red), monocytes (yellowish-brown) and lipofuscin positive cells in infarcted areas. Monocytes and lipofuscin positive cells are significantly increased in fibroblast and microsphere treated hearts (*P < 0.05 vs. control and MM6, objective 40×, scale bar 25 µm)

Full size image

Tissue inflammation/lipofuscin staining

For further assessment of inflammation myocardium was stained with lipofuscin as an established cumulative marker of inflammation [35] particularly within infarcted regions. Lipofuscin is a granular pigment commonly associated with aging but found to be accumulating in large quantities in infarcted myocardium, following increased local inflammation [41, 42]. We could show that after TX in the groups treated with microspheres and fibroblasts the content of lipofuscin positive cells is significantly higher in infarcted areas (fibroblasts 94.94 ± 10.8 cells/mm2, microspheres 97.20 ± 14.34 cells/mm2) compared to macrophage und medium-injected hearts (macrophages 36.12 ± 8.42 cells/mm2, medium 45.23 ± 7.66 cells/mm2, P < 0.05; Fig. 3b).

Vascular density

Inflammation and angiogenesis are frequently regarded as closely related processes with neovascularization being the prominent vascular response to chronic inflammation. To examine whether injection of macrophages, fibroblasts and microspheres might affect angiogenesis, vascular density was assessed by staining and quantification of CD 31 positive vessels in the area of infarction. Neither cell nor microsphere injection resulted in a significantly increased number of vessels at 2 months after TX as compared to medium treated control hearts (control 347.6 ± 44.6 vessels/mm2, microspheres 355.6 ± 34.8 vessels/mm2, fibroblasts 342.8 ± 28.7 vessels/mm2, macrophages 336.4 ± 14.5 vessels/mm2; Fig. 2b). Furthermore, there were no BrdU positive cells detected in association with vessel walls, suggesting that injected cells were not incorporated into newly forming vessels (data not shown).

Collagen content as determined by specific collagen staining (Accustain trichrome) did not differ among the four groups (Fig. 2c). Likewise, infarct size was similar and independent of the respective groups’ treatment (data not shown).

Cardiac function

To assess cardiac function we performed 2D echocardiography before MI as well as prior to TX (4 weeks after MI induction) and 2 months after TX. Four weeks after MI, reduced FS was observed in all groups.

However, after fibroblast transplantation/microsphere injection assessment of FS revealed a significant improvement in contrast to injection of macro- phages or medium alone (FS microspheres 38.9 ±  4.6%; FS fibroblasts 36.84 ± 6.05%, FS macrophages 29.16 ± 8.7%, FS medium 27.2 ± 7.2%; P < 0.05, Fig. 4a).

Fig. 4
figure 4

Echocardiographic results and isolated heart studies. Fractional shortening in percent (a) after transplantation revealed significant differences towards improved left ventricular function after transplantation of microspheres and fibroblasts. Isolated heart studies: Significantly higher base line pressures (b *P < 0.05) and significantly higher contractile response to isoproterenol application (c) after transplantation of fibroblasts and microspheres

Full size image

For further quantification of LV function, we performed isolated heart studies with retrograde perfusion according to Langendorff. Left ventricular developed pressure (LVDP) was significantly higher in hearts treated with fibroblasts and microspheres 2 months after injection (LVDP fibroblasts 129 ± 32.9 mmHg, LVDP microspheres 119.2 ± 24.1 mmHg) compared to medium and macrophage injected hearts (LVDP medium 67 ± 22.6 mmHg, LVDP macrophages 75.9 ±  24.8 mmHg, P < 0.05; Fig. 4b).

While medium-treated hearts remained unaffected by isoproterenol stimulation (LVDP medium 77 ±  30.4 mmHg), microsphere and fibroblast injection resulted in a significant response to catecholamine challenge with isoproterenol (LVDP microspheres 152.4 ± 28.3 mmHg, LVDP fibroblasts 159.4 ±  13.1 mmHg, P < 0.05, Fig. 4c). Macrophage-injected hearts also responded to catecholamine challenge albeit to a lesser extent than the former (LVDP macrophages 100.6 ± 13.9 mmHg, P < 0.05; Fig. 4c). Coronary flow rates did not differ between the groups.

Discussion

Functional improvement following cell TX has been attributed to various mechanisms. TX of ex vivo expanded EPCs augmented neovascularization in a model of hindlimb ischemia [13]. Reduced LV dilatation and preserved cardiac function was also demonstrated in a rat model of MI by TX of EPC and attributed to improved neovascularization [14, 33]. Subsequent studies have supported the notion of neovascularization as the main mechanism underlying improved LV function after EPC TX [9, 21, 32, 33].

It has also been proposed, that transplanted mesenchymal stem cells and other cell types like embryonic stem cells [22] are able to express muscle proteins after integration in host myocardium [43] and repopulate scar areas [22]. Menard et al. [22] explained an improvement after cell TX by a provision of new contractile elements. Other studies showed survival of transplanted cells in host myocardium and development of gap junctions between transplanted and host cells as a sign of electromechanical coupling [30, 34]. Orlic et al. [27] also reported, that hematopoetic stem cells, directly injected into infarcted myocardium, can differentiate into cardiac myocytes and lead to improvement of LV function in mice. However, in conflict with these findings, rather than differentiating into cardiac muscle or vasculature, hematopoetic stem cells injected after MI have been shown to adopt mature hematopoetic fates [3, 24] without contractile function. Furthermore, transplanted allogenic mesenchymal stem cells improved global LV function only transiently [5]. These transient beneficial effects suggest a possible early paracrine effect of transplanted cells. Indeed, in several studies [11, 15, 47] bone marrow-derived cells differentiated into pericytes, fibroblasts and leukocytes producing growth factors and chemokines thus inducing collateralization and vascular growth in a mouse model of hind limb ischemia.

Recently, we were able to demonstrate that human umbilical vein cell (HUVEC) treated hearts showed higher coronary flow rates and increased LVDP at baseline as well as under volume strain and catecholamine stress. Based on these findings [23], we hypothesized that observed structural and functional improvement are partially mediated by macrophages, because monocytes have juxta- or paracrine effects on positive inotropic receptors and stimulate angiogenic receptors [18, 46]. Therefore, it is tempting to speculate that inflammation triggered by transplantation has the potential to contribute to and modify the remodeling process and consequently improve cardiac function. Indeed, inflammatory chemokines have been suggested to modulate cardiac dysfunction following MI [1, 8, 45].

It has been further suggested that higher collagen content in infarcted myocardium could contribute to mechanic stability and improved function of the left ventricle [16]. In this context, a catheter-based transendocardial delivery of SDF-1α in a porcine model of MI resulted in impaired LV function associated with a significant loss of collagen in the peri-infarct area [16].

The presented study compared the effect of microsphere and cell injection on myocardial remodeling and contractile function. Our results demonstrate that microspheres persist at injection sites for at least 2 months. Without affecting infarct size, microsphere injection results in improved LV function and a tendency towards reversal of post infarct remodeling. The underlying mechanism to this unexpected benefit derived from injection of supposedly biologically passive microspheres appears to lie in an augmented inflammatory response. Indeed, the numbers of macrophages and lipofuscine positive cells found in infarct regions were significantly increased after microsphere injection. This concept is supported by a study from Heusch et al. [38] employing a porcine infarction model involving pre-treatment by microembolization with microspheres. In this model, an inflammatory reaction was demonstrated in combination with increased TNF-α expression and decreased infarct size.

The effect of microsphere injection into infarcted myocardium was compared to transplantation of human macrophages and fibroblasts. Rational for the selection of these cell groups was the expected stimulation of inflammatory responses (macrophages) and an augmented connective tissue response (fibroblasts), respectively.

Similar to microspheres, the injection of fibroblasts resulted in functional improvement detectable 2 months after TX, despite collagen content and neovascularisation remaining unaffected. While only few labeled fibroblasts could be demonstrated in the target area, an increased inflammatory response was detected, again similar to the effects observed after microsphere injection.

The transplantation of macrophages resulted in findings comparable to medium-injected hearts. After 2 months, cells were no longer present in the infarct zones and no inflammatory response could be detected at this time point.

While the association between inert microspheres, inflammatory response and improved contractility remains the key finding, further issues are worth addressing. Our findings demonstrate that an improvement of cardiac function after cell TX is not necessarily combined with higher rates of neovascularization. The density of CD 31 positive vessels was not different between the four groups. In accordance, coronary flow rates did not differ in isolated heart studies.

Furthermore, our results show no significant differences in amounts of collagen positive staining, suggesting that collagen content did not contribute to the observed improvement of LV function in the presented model.

In our study, improvement of myocardial function appears to be associated with lower amounts of apoptotic cells in infarcted regions. Statistical analysis, however, did not demonstrate quantity of apoptotic cells to be an independent predictor of improved LV function. Interpretation of these findings is limited since measurement and further differentiation of apoptosis, especially in non-infarcted myocardium, were not part of the original study protocol.

A further limitation of the presented study certainly lies in the fact that transplanted cells were not autologous (fibroblasts and macrophages). Presumably this will have led to an enhanced immune response and might in part explain the low amount of BrdU labeled cells observed 2 months after TX.

In conclusion, this study demonstrates that inflammatory processes induced by injection of unlikely candidates such as fibroblasts and even microspheres 1 month after MI are associated with improved myocardial function. Since coronary interventions, coronary surgery and especially application of autologous progenitor cells or fetal stem cells may also trigger or sustain an inflammatory response in patients so treated, these mechanisms of inflammation warrant further investigation.

Furthermore, standard pharmacologic therapy following MI including statin and ACE inhibitor therapy is known to modulate inflammatory processes. A better understanding of these multifaceted pro- and anti-inflammatory interactions might lead to improved therapeutic options for patients affected by MI.