Author + information
- Received November 2, 2004
- Revision received February 3, 2005
- Accepted February 8, 2005
- Published online November 1, 2005.
- ↵⁎Reprint requests and correspondence:
Dr. Robert A. Kloner, The Heart Institute, Good Samaritan Hospital, University of Southern California, 1225 Wilshire Boulevard, Los Angeles, California 90017-2395.
Objectives We investigated whether granulocyte colony-stimulating factor (G-CSF) and stem cell factor (SCF) could promote myocardial regeneration after coronary artery occlusion and improve left ventricular (LV) function.
Background Cytokine-induced mobilization of bone marrow stem cells in the heart may represent a promising strategy for replacing infarcted myocardium.
Methods Sprague-Dawley rats were subjected to permanent coronary occlusion. A treated group (n = 19) received G-CSF (100 μg/kg) and SCF (25 μg/kg) subcutaneously, starting 2 h after surgery and continuing daily for an additional 4 days. Control rats (n = 21) received sterile water. The peripheral blood content in hematopoietic progenitor cells was analyzed.
Results At eight weeks, LV angiograms (rest and dobutamine stress) and histologic analysis were performed. At rest, LV ejection fraction (LVEF) was 0.45 in controls and 0.52 in treated hearts (p = 0.16). For any infarct size, LVEF was greater in the treated group (p = 0.045). Under dobutamine stress, treated animals had smaller LV end-diastolic and -systolic volumes (0.37 ± 0.04 ml and 0.16 ± 0.03 ml) versus control animals (0.51 ± 0.05 ml and 0.26 ± 0.04 ml; p = 0.026 and 0.048) with a 7% improvement in ejection fraction. Scar thickness was 1.1 ± 0.1 mm in treated hearts and 1.0 ± 0.1 mm in controls (p = 0.36). Scar morphology was similar in both groups without obvious new muscle in the scar.
Conclusions Because we did not find evidence of new muscle cells in the infarct area, our conclusion is that G-CSF and SCF enhanced the LV functional reserve of the heart without replacing scar tissue.
Bone marrow stem cell mobilization induced by cytokines may represent a promising strategy for repairing infarcted myocardium. Several growth factors, including interleukin (IL)-3, IL-8, granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and stem cell factor (SCF) have been reported to mobilize hematopoietic precursor cells from the bone marrow (1,2) and stimulate endothelial cell migration and proliferation (3).
Granulocyte colony-stimulating factor, which is currently used therapeutically for the treatment of tumors, is one of the most studied in the setting of myocardial infarction (MI). Granulocyte colony-stimulating factor stimulates the proliferation and differentiation of precursor cells committed to the neutrophil/granulocyte cell type (4). Granulocyte colony-stimulating factor in combination with SCF was also found to mobilize primitive bone marrow cells in splenectomized mice (5). Stem cell factor stimulates early pluripotent and committed stem cells to form colony-forming units and can act synergistically with other growth factors, such as IL-1, IL-3, IL-6, GM-CSF, G-CSF, and erythropoietin to increase the production of hematopoietic cells (6,7). Several studies have shown that G-CSF improves cardiac function and left ventricular (LV) remodeling after MI (5,8–10). However, negative results of recent studies have dampened researchers’ enthusiasm and questioned the capacity of growth factors to induce regeneration and neovascularization of myocardial tissue (11–15). At present, the situation is controversial. Therefore, more research is necessary to provide conclusive answers.
In this study, we investigated whether a combination of the two hematopoietic growth factors, G-CSF and SCF, would result in regeneration of myocardial tissue and improvement of LV function after permanent coronary artery occlusion in rats.
Fifty-three Sprague-Dawley female rats (200 to 250 g) were used in this study. All the rats were maintained in accordance with the policies and guidelines of the position of the American Heart Association on research animal use (16) and the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, National Research Council (Department of Health, Education, and Welfare Publication No. 85-230). All animal procedures were approved by the Institutional Care and Use Committee at Good Samaritan Hospital.
The rats were anesthetized with an intraperitoneal injection of 75 mg/kg ketamine and 5 mg/kg xylazine, shaved, endotracheally intubated, and ventilated with a rodent respirator. Under sterile conditions, a left thoracotomy in the fourth intercostal space was performed after injection of the nerve blocker bupivicane (0.1 mg/kg) into the third through fifth intercostal spaces. The pericardium was removed and the proximal left coronary artery was ligated with a 4-O silk suture, causing permanent coronary artery occlusion. The chest was then closed and the rats were weaned from the respirator and extubated. The animals received buprenex (0.02 mg/kg) as required for analgesia.
The rats were randomly divided into two groups: treated and control. The treated group received a daily dose of 100 μg/kg of recombinant methionyl human G-CSF (Neupogen [filgrastim], Amgen Inc., Thousand Oaks, California) and 25 μg/kg of recombinant rat SCF (Peprotech Inc., Rocky Hill, New Jersey), subcutaneously, starting 2 h after coronary occlusion and continuing for the following four days (5-day treatment). The control group received an equivalent volume of sterile water.
Evaluation of hematopoietic progenitors in circulation
Blood samples were collected and assayed before and after the initiation of growth factor treatment. The blood was collected in anticoagulant citrate dextran (Sigma-Aldrich, St. Louis, Missouri) and the progenitor content in the blood was measured using a methylcellulose-based progenitor assay (17). Whole blood cells were assayed in 1% methylcellulose medium containing recombinant cytokines according to the manufacturer’s recommended guidelines (Methocult GF R3774, Stem Cell Technologies, Vancouver, Canada). The samples were plated at half-log dilutions from 1 × 104to 3 × 105with three plates at each cell concentration. The colonies were counted on days 10 to 14 of culture, with the dilution containing ∼30 well-separated colonies used to calculate the progenitor content.
Physiologic assessment of LV function
Left ventricular function was assessed eight weeks after induction of ischemia. The rats were anesthetized and ventilated as described earlier. Fluid-filled catheters were inserted into the carotid artery and jugular vein. Heart rate and arterial pressures were measured through a pressure transducer connected to the catheter in the carotid artery (ADInstruments, Colorado Springs, Colorado). In order to assess heart function with stress, dobutamine was given as an infusion at a dose of 10 μg/kg/min. Angiograms (with the rat in a lateral position) were performed before (rest) and during the infusion of dobutamine (stress) by using a mini C-arm fluoroscopy system (XiScan C-arm Imaging system, Xite Inc., East Windsor, Connecticut). The LV was visualized by an injection of non-ionic contrast (1.5 ml) via the jugular vein. The contrast circulates to the right ventricle, lung, and then LV, where images of LV end-diastolic and end-systolic volumes are obtained as previously described (18,19). The angiograms were recorded on a VHS videotape and blindly analyzed to measure LV systolic/diastolic volumes and ejection fractions.
Regional myocardial blood flow and capillary assessment
Regional myocardial blood flow (RMBF) was measured using 1 × 106radioactive microspheres labeled with 141Ce (Perkin-Elmer Life Sciences, Boston, Massachusetts). After 10 min of stabilization following the angiograms, the chest was opened, 0.5 ml of sucrose solution containing the microspheres was injected into the LV, and simultaneously, a reference blood sample was withdrawn from the carotid artery at a rate of 0.37 ml/min.
In order to identify perfused blood vessels in the infarcted area, 0.8 ml of 50% Unisperse blue pigment suspension (Ciba-Geigy, Hawthorne, New York) was injected via the jugular vein. Seconds later, 2 ml of potassium chloride was injected into the same vein to stop the heart in diastole. The rats were deeply anesthetized at the time of the injection.
The hearts were excised and fixed in 10% formalin at a constant intraventricular pressure (11 cm H2O) for 30 min. Postmortem LV volumes were measured by weighing the empty LV, filling it with distilled water, and weighing it again. Three measurements were made and averaged.
The hearts were sliced from apex to base in four transverse sections, which were weighed and photographed. One slice from the middle area was processed for histology; it was embedded in paraffin and sections were stained with both hematoxylin and eosin and picrosirius red. The sections stained with picrosirius red (collagenous scar stains red and normal tissue stains yellow) were used to calculate infarct size (IS). Various parameters were determined by computerized planimetry (SigmaScan Pro, Chicago, Illinois): areas and circumferences of ischemic and normally perfused regions, LV cavity area, thickness of the infarcted and non-infarcted walls (average of three measurements), and the thinnest and thickest regions of the scar.
Measurement of RMBF
Samples from infarcted and non-infarcted areas were dissected from the other two sections of the heart and weighed. The radioactivity of these tissue samples, together with the radioactivity of the reference blood sample, was counted in a gamma counter, and RMBF was calculated (20).
Capillary density was evaluated by microscopic examination of the slide stained with picrosirius red. The number of capillaries per high-power field (400×), recognized as round or tubular structures containing blue pigment, were counted in the scar area, in the left and right border zone adjacent to the scar, and in the non-infarcted area. Results were expressed as number of capillaries per mm2.
All data are presented as mean ± SEM and were analyzed using SAS software (Cary, North Carolina). Student ttests were performed to compare discrete values between treated and control groups. The relationship between the size of the infarct and either the ejection fraction or the end-diastolic/systolic volumes was analyzed using analysis of covariance (ANCOVA). Values of p < 0.05 were considered statistically significant.
Ten of the 53 rats used in the study died during the day of surgery after coronary occlusion. Forty-three rats survived surgery and myocardial infarction. Three rats were excluded because the hearts did not demonstrate an infarct when sacrificed at eight weeks. Data are presented for 40 animals: 19 treated and 21 control rats.
Evaluation of progenitor mobilization following growth factor treatment
A separate group of rats was used to assess the mobilization of hematopoietic progenitors into the periphery following growth factor treatment. In these experiments the progenitor content was evaluated in the whole blood cell fraction and the analysis showed a significantly higher level of progenitors in the blood following growth factor treatment (9.5 × 105leukocytes) compared with the control group (3.65 × 105leukocytes) (p = 0.0062; ttest). These data demonstrate that the growth factor treatment regimen used in these studies resulted in the mobilization of hematopoietic progenitors into the blood of treated animals.
Heart rate and mean arterial pressure (MAP) were measured at eight weeks after myocardial infarction. There were no significant differences between the control and the treated groups at any of the time points studied: baseline, before dobutamine infusion (rest), during dobutamine infusion (stress), or before RMBF (after opening the chest) (Table 1).
At rest, the angiographic variables LV end-diastolic and LV end-systolic volumes between the two groups did not differ significantly (Table 1).
Under dobutamine stress, treated hearts manifested improvement in these parameters when compared with control hearts. Left ventricular end-diastolic volume was 0.51 ± 0.05 ml in the control hearts versus 0.37 ± 0.04 ml in the treated hearts during dobutamine stress (p = 0.026) (Fig. 1).Left ventricular end-systolic volume was 0.26 ± 0.04 ml in the control hearts versus 0.16 ± 0.03 ml in the treated hearts during dobutamine (p = 0.048) (Fig. 1).
There was also a trend (although non-significant) toward higher LV ejection fraction (EF) in the animals that received the growth factors compared with control animals, both at rest (0.52 ± 0.04 vs. 0.45 ± 0.03, p = 0.161) and with dobutamine stress (0.58 ± 0.04 vs. 0.54 ± 0.04, p = 0.450). The increase was 13% and 7%, respectively.
Relationship among IS, EF, and LV dimensions
To verify a possible direct relationship between IS and either EF or LV end-diastolic/systolic volumes, individual values of IS in both control and treated groups were analyzed and plotted against the correspondent EF (Fig. 2)or LV end-diastolic and -systolic volumes (Fig. 3).The results showed that, for any infarct size, EF at rest was greater in the treated group (p = 0.045 by ANCOVA) (Fig. 2). Ejection fraction under stress was not significantly different between the two groups. The data also indicated that, under dobutamine stress, for any infarct size, treated rats had smaller LV end-diastolic and -systolic volumes (p = 0.017 and p = 0.016, by ANCOVA) (Fig. 3).
RMBF and capillary density
In both groups, treated and control, RMBF in the ischemic area (1.21 ± 0.20 ml/min/g and 1.35 ± 0.22 ml/min/g, respectively) was lower than that in the non-ischemic area (2.09 ± 0.32 ml/min/g and 2.48 ± 0.34 ml/min/g, respectively), with the RMBF in the scar being 58% (control) and 55% (treated) of the RMBF in the non-ischemic tissue. However, RMBF values in each area were similar in the two groups (Table 1).
Capillary density, as a measure of neovascularization, was not significantly different between the rats receiving the growth factors versus control rats (Table 1).
Histology and postmortem volumes
Histologic analysis showed similar mean IS in the two groups either when infarct was expressed as a percentage of area or when expressed as a percentage of circumference (Table 1).
Other measurements relative to the scar, such as overall thickness and the thinnest and the thickest regions, were not different between the two groups (Table 1). Scar thickness was 1.0 ± 0.1 mm in control hearts and 1.1 ± 0.1 mm in treated hearts (p = 0.358).
The thickness of the non-infarcted LV area instead tended to be lower in the rats that received growth factor therapy (1.8 ± 0.1 mm vs. 2.1 ± 0.1 mm).
There were no qualitative differences in histologic appearance of the scars between groups. Specifically, the infarct scars were primarily composed of collagen, fibroblasts, and a few residual cardiomyocytes along the endocardial edge and sometimes the epicardial edge. There were no large new patches of cardiac muscle as we had previously observed with implantation of neonatal or fetal cardiomyocytes (Fig. 4).
Left ventricular cavity areas measured by histology and LV postmortem volumes in the pressure-fixed hearts were similar in treated and control animals (Table 1).
In this study we investigated whether G-CSF and SCF in combination could promote myocardial regeneration and improve LV function after coronary artery occlusion in rats. Our findings show that the beneficial effect of G-CSF + SCF treatment relies on a mechanism that is different from regeneration of myocardium in the infarct (suggested by some investigators to be due to stem cell mobilization). Under stress, growth factor therapy reduced LV diastolic and systolic volumes, with a trend to improved EF without evidence of new muscle. For any given infarct size, EF was improved in the G-CSF and SCF group. Scar morphology, size, and thickness were indeed similar in both groups, without obvious new cardiac muscle growth within the scar. Because we did not find evidence of new cardiac muscle cells in the infarct area, our conclusion is that the combination of G-CSF and SCF enhanced the LV functional reserve of the heart without replacing scar tissue.
The two growth factors G-CSF and SCF are well known to promote hematopoiesis (4,21,22). In addition, several studies have shown that G-CSF, either by itself or in combination with SCF, has beneficial effects in MI (5,8–10). Orlic et al. (5) showed that bone marrow cells, mobilized by administration of G-CSF and SCF to splenectomized mice, home to the infarcted area, differentiate into cardiomyocytes, endothelial, and smooth muscle cells; promote myocardial repair; and reduce mortality by forming vascular structures and improving anatomic remodeling and LV function. Kocher et al. (8) demonstrated that injection of bone marrow-derived endothelial progenitor cells mobilized by recombinant human G-CSF contribute to neovascularization of ischemic myocardium in rats and improve cardiac function. Minatoguchi et al. (10) suggested that the improvement in LV function that they observed in a rabbit model of coronary occlusion/reperfusion after G-CSF treatment could be a consequence of the regeneration of myocardial tissue and acceleration of the healing of postinfarction wounds induced by G-CSF.
However, despite these positive and encouraging results, there is still controversy surrounding the efficiency of growth factor therapy compared to cell transplantation for the therapy of MI. Can these factors really stimulate mobilization of bone marrow stem cells and differentiation of these cells into cardiomyocytes?
Recent reports have demonstrated a limited ability of some growth factors to restore cardiac function in infarcted hearts of different animal models (11–15). Terrovitis et al. (11) evaluated the effect of GM-CSF administration in a pig model of MI. They did not see a decrease in infarct size or an improvement in LV remodeling. Maekawa et al. (12) investigated the influence of GM-CSF (different from G-CSF that we used) induction on post-MI remodeling after occlusion of the coronary artery in Wistar rats. They demonstrated that administration of romurtide, a GM-CSF inducer, delayed collagen production and promoted both monocyte recruitment to the infarcted heart and infarct expansion, which resulted in aggravation of LV remodeling. Sakakibara et al. (13) showed that administration of G-CSF to mice after MI was not as effective as bone marrow cell transplantation for infarct repair. Yano et al. (14) did not observe any significant differences in infarct sizes and in LV systolic and diastolic properties following administration of G-CSF to rats after MI. Chachques et al. (15) demonstrated that administration of vascular endothelial growth factor to sheep three weeks after MI induced angiogenesis but did not improve LV function or remodeling.
The combination of G-CSF and SCF mobilized hematopoietic progenitors into the circulation in our rat model of MI. However, this mobilization did not induce either regeneration of myocardium in the infarcted area or angiogenesis. In previous studies involving cell transplantation, we have shown that it is possible to replace scar with muscle (18,19,23,24). We have demonstrated that fetal and neonatal cardiomyocytes injected directly into the heart muscle survive the infarct, differentiate, develop a muscular supply, thicken the infarct wall with reduction of cavity area and LV dilation, and improve EF.
In the present study, the thickness of the scar, the LV cavity area, and the infarct size were similar in both growth factor-treated and control rats. If there was recruitment of cells to the infarcted tissue, it should have resulted in an increase in wall thickness of the scar without a corresponding increase in the circumference of the scar. We calculated and expressed IS either as a percentage of LV area or as a percentage of the LV circumference, and in both cases we did not find any difference between the two groups. Also, in contrast to our previous studies, which injected fetal or neonatal cardiomyocytes into the scar in which clear-cut patches of new muscle could be identified, we did not identify new muscle in any of the scars of treated animals in the present study. The number of capillaries and the blood flow in the scar tissue were also similar between the rats receiving the growth factors versus control rats, indicating equivalent vascularization.
Administration of growth factors, however, elicited some beneficial effects. Ejection fraction under resting conditions when analyzed in relation to infarct size was significantly higher in the treated group. There was a trend toward improvement in EF by 15% at rest and by 7% under stress conditions in treated rats compared with untreated rats. Under dobutamine stress, the treated group had smaller LV end-diastolic and LV end-systolic volumes. This was also true when correlation analysis was conducted between volumes and infarct size. These data suggest reduced ventricular dilation of the heart under stress following growth factor administration. Evaluation of the blood revealed that there was an increase in hematopoietic progenitor cells after five days of treatment with G-CSF + SCF. Thus, there was some mobilization of bone marrow stem cells into the circulation. However, we did not see evidence for myocardial regeneration in the scar, which has been suggested to occur with mobilization of stem cells. There was similar scar morphology in both groups without recruitment of cells to the infarct area. Our conclusion is that the combination of G-CSF and SCF had some benefits in MI through an action on residual myocardium that is as yet unknown. Although one possibility is that the infarct stimulated homing of stem cells to the injured area, we did not specifically study this concept. However, it is important to observe that homing of stem cells is a time-dependent process, as Murry et al. have demonstrated (25). It is therefore possible that in our study, early after the treatment with G-CSF and SCF and for a relatively short period of time, there was some recruitment of stem cells to the heart. The stem cells may have only been able to survive a brief period of time in the hostile and hypoxic environment of the infarct, or perhaps they migrated to other tissues. Rather than developing new heart muscle, the stem cells may have released humoral factors that favorably affected surrounding viable myocardium (that is, a paracrine effect). This could explain why we did not see new muscle in the scar of treated rats after eight weeks from the treatment, but did observe an improvement in cardiac functional parameters.
It is possible that these beneficial effects might be due to some of the growth factor-related signaling systems. It is known that G-CSF activates a variety of intracellular signaling cascades such as JAK/STAT, Ras-Raf-MAP kinase, and Src family kinase pathways (26). Given the fact that the JAK/STAT pathway plays a critical role in the cardiovascular system, it might be that the beneficial effects shown in our study are a consequence of the activation of the JAK/STAT signaling pathway associated with a cytokine receptor. In a recent preliminary report, Harada et al. (27) have shown that G-CSF protects the myocardium from apoptotic cell death by activation of the JAK/STAT pathway and induction of the expression of anti-apoptotic molecules in the heart. This resulted in prevention of cardiac remodeling after MI. At present, the concept that G-CSF may have worked through this mechanism in our study is speculative; however, it would be an interesting path to explore in future experiments.
The authors thank the Kenneth T. and Eileen L. Norris Foundation.
This study was funded in part by NHLBI grant 1RO1-HLO73709.
- Abbreviations and Acronyms
- ejection fraction
- granulocyte colony-stimulating factor
- granulocyte-macrophage colony-stimulating factor
- infarct size
- left ventricular/ventricle
- mean arterial pressure
- myocardial infarction
- regional myocardial blood flow
- stem cell factor
- Received November 2, 2004.
- Revision received February 3, 2005.
- Accepted February 8, 2005.
- American College of Cardiology Foundation
- Laterveer L.,
- Lindley I.J.,
- Hamilton M.S.,
- Willemze R.,
- Fibbe W.E.
- Welte K.,
- Bonilla M.A.,
- Gillio A.P.,
- et al.
- Orlic D.,
- Kajstura J.,
- Chimenti S.,
- et al.
- Dawn B.,
- Guo Y.,
- Rezazadeh A.,
- et al.
- Minatoguchi S.,
- Takemura G.,
- Chen X.-H.,
- et al.
- Maekawa Y.,
- Anzai T.,
- Yoshikawa T.,
- et al.
- Sakakibara Y.,
- Nakajima H.,
- Yoshimoto M.,
- et al.
- Yano T.,
- Miura T.,
- Miki T.,
- et al.
- Eaves C.J.
- Muller-Ehmsen J.,
- Peterson K.L.,
- Kedes L.,
- et al.
- Metcalf D.
- Avalos B.R.
- Harada M.,
- Yingjie Q.,
- Takano H.,
- et al.