OUP user menu

Stem cell-related cardiac gene expression early after murine myocardial infarction

Susanne Vandervelde, Marja J.A. van Luyn, Mark H. Rozenbaum, Arjen H. Petersen, Rene A. Tio, Martin C. Harmsen
DOI: http://dx.doi.org/10.1016/j.cardiores.2006.11.030 783-793 First published online: 1 March 2007


Objective: Clinical experimental stem cell therapy after myocardial infarction appears feasible, but its use has preceded the understanding of the working mechanism. The ischemic recipient cardiac environment is determinative for the attraction and subsequent fate of stem cells. Here, we studied expression levels of genes that are anticipated to be essential for adequate stem cell-based cardiac repair at various time-points during the 1 month period following myocardial infarction (MI).

Methods: Gene expression in the hearts of mice that underwent MI by permanent or transient (30 min) ligation of the coronary artery was monitored using quantitative RT-PCR analysis of mRNA isolated from whole heart sections as well as from specific, laser micro-dissected, regions of sections. Protein expression was performed by immunohistochemical stainings and Western blot analysis.

Results: Many inflammatory genes were highly expressed for at least 1 week after MI. The expression of pro-angiogenic genes such as bFGF, VEGF-A and VEGF-R2 changed only marginally post-MI. Markers used to test stem cell gene expression remained unchanged post-MI with the exception of G-CSF and GM-CSF, which are genes that are also known to enhance the inflammatory response. Analysis of micro-dissected regions revealed that SDF-1, SCF (both stem cell attractants) and VEGF-R2 (involved in angiogenesis) gene expression was slightly decreased especially in the infarcted region.

Conclusion: Genes that are generally considered to participate in stem cell-related processes and angiogenesis were not upregulated after MI, whereas the inflammatory gene expression dominated. Modulation of this imbalance might be of value for stem cell-mediated therapy.

  • Myocardial infarction
  • Gene expression
  • Stem cells
  • Angiogenesis
  • Cytokines
  • Growth factors

1. Introduction

The first experimental clinical applications of stem cells for regenerative therapy after myocardial infarction appear safe and seem to improve cardiac function [1–3]. However, the detailed working mechanism underlying this improved function remains unclear.

The contribution of bone marrow-derived stem cells (BMSCs) to cardiac homeostasis and repair seems marginal [4,5] and endogenous BMSCs are not stimulated adequately after myocardial infarction [6]. Moreover, transplanted stem cells may rapidly die after intramyocardial transplantation [7] and may not increase cardiac function, although they seem to be favourably involved in infarct remodelling [8].

Thus, exploration of the niche that needs to be regenerated is indispensable to optimize future stem cell therapy. Recent investigation of the post-infarcted murine heart on a cellular level revealed a great influx of inflammatory cells mainly during the first week post-MI and a decrease of small vessels in the infarcted region immediately post-MI [9]. However, exploration of the niche on a molecular level has not been investigated thoroughly yet to determine the level of expression of stem cell-related cytokines and growth factors at various time-points during the first month after MI.

Therefore, we analysed the gene expression profile of several factors after myocardial infarction in mice, i.e. cyto/chemokines and growth factors, and their cognate receptors, divided into three functional groups: inflammatory factors, angiogenic factors and stem cell factors. This classification is based on the inflammatory process that occurs immediately post-MI and the processes that are required for adequate regeneration i.e. angiogenesis, stem cell mobilization and engraftment [9]. Without doubt some of these factors may serve more promiscuous roles and, therefore, belong to more than one functional group.

Here, we show gene expression dynamics in whole heart cross sections and in selectively microdissected regions of sections of the two most widely used murine models of acute myocardial infarction: permanent ligation (“ligation”) and ligation followed by reperfusion of the left anterior descending (LAD) coronary artery (“reperfusion”). Furthermore, gene expression is correlated to protein expression by immunohistochemical stainings and Western blot analysis of some of the investigated genes. This study provides fundamental molecular knowledge of the murine ischemic cardiac environment, which might support the development of novel therapeutic modalities aimed at optimising the ischemic cardiac environment for stem cell-mediated cardiac regeneration.

2. Materials and methods

2.1. Animals

Nine to 12 week old C57BL/6JolaHsd male mice (22–25 g) were obtained from Harlan Nederland (Horst, the Netherlands). Mice were housed individually prior to and after surgery. All procedures performed on mice were approved by the local committee for care and use of laboratory animals, and were performed according to strict governmental and international guidelines on animal experimentation. The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).

2.2. Myocardial infarction

Mice were anesthetized, orally intubated and mechanically ventilated (Hugo Sacks Elektronik, March-Hugstetten, Germany) with isoflurane (2.5%), N2O (2%), O2 (2%) and fixed into position with their left side up. Thirty minutes prior to surgery, mice received analgesic drug temgesic (0.03 mg/kg) intramuscularly. A left thoracotomy was performed via the third intercostal space, and muscles and pericardium were carefully dissected. The LAD coronary artery was localized using a dissection microscope, and ligated with a 6-0 non-absorbable Prolene suture just proximal to the bifurcation of the LAD. For the reperfusion model, the suture was removed after 30 min of occlusion, for the permanent ligation model, the suture remained in place. Sham operated mice underwent the same procedure except for arterial occlusion. The intercostal space was closed with a 6-0 non-absorbable Prolene suture and the skin with a 5-0 absorbable Safil suture. Mice received 100% oxygen until wakening after which the endotracheal tube was withdrawn. Animals were given 100% oxygen via nasal cone until full recovery of consciousness. During recovery mice were kept warm with a heat lamp.

At 6 and 24 h, 2, 7 and 28 days post-MI mice were killed. The hearts were dissected and snap-frozen in liquid nitrogen and stored at −80 °C (n=3 per time point per model).

2.3. Histological stainings

Hearts fixed in liquid nitrogen were cut transversally in the middle third between base and apex in sections of 5 μm. Masson Trichrome (collagen blue, myocardial cells red) was performed according to standard protocol [10] on cryo sections of 6–8 μm, fixed first in 4% formaldehyde for 1 h room temperature followed by Bouin fixative at 60 °C for 5 min. To explore protein expression, cryosections (5 μm) were stained immunohistochemically for TGF-β1 (polyclonal goat anti-human TGF-β1, Santa Cruz Biotechnology, Santa Cruz, USA), VEGF-A (polyclonal rabbit anti-human VEGF-A-20, Santa Cruz Biotechnology, Santa Cruz, USA) EPO receptor (polyclonal rabbit anti-mouse EpoR M-20, Santa Cruz Biotechnology, Santa Cruz, USA), and VEGF-Receptor 2 (polyclonal rabbit anti-mouse FLK-1 (C-1158) Santa Cruz Biotechnology). Endogenous biotin and avidin were blocked with avidin/biotin blocking kit according to the manufacturer's instructions (Vector Laboratories, Burlingame, USA). Endogenous peroxidase was blocked by incubation with 0.1% H2O2 in PBS for 10 min. Colour development in immunoperoxidase stainings was performed with 3-amino-9-ethyl-carbazole (AEC; Sigma, Steinheim, Germany) and sections were counterstained using Mayer's hematoxilin (Fluka Chemie, Buchs, Switzerland).

2.4. mRNA isolation and cDNA synthesis

Total RNA was isolated from heart slices (15 sections of 5 μm thickness) with the Absolutely RNA® Microprep Kit according to manufacturer's instructions (Stratagene, La Jolla, CA, USA), including DNase treatment. All RNA preparations were free from genomic DNA, as confirmed by qPCR carried out on total isolated RNA i.e. without the reverse transcription step. RNA yield and quality was determined by measuring the absorbance at 230, 260, and 280 nm with a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies, Montchanin, DE, USA), and only samples with 260/280 ratios between 1.8 and 2.2 were used for further analysis. Equal amounts (100 ng) of total RNA from all samples were taken for cDNA synthesis using M-MuLV Reverse Transcriptase in the presence of random hexamers, according to manufacturer's instructions (First strand cDNA Synthesis Kit, Fermentas Life Sciences, St Leon-Rot, Germany). For qRT-PCR, TaqMan “assay by demand” primer/probe sets specific for mouse β2-microglobulin served as internal standard control (Applied Biosystems, Foster City, CA).

2.5. Laser dissection microscopy

Heart sections of 5 μm of time points t=24 h, 7 and 28 days (n=3 per time point) were mounted on Palm slides (P.A.L.M. Microlaser Technology, Bernried, Germany) and were stained for 1 min with RNAse-free Mayer's hematoxilin (Sigma) and air-dried. Equal surface areas (2 mm2) of tissue were laser dissected from infarcted hearts i.e. infarcted area, border zone and spared myocardium. The border zone was defined as a rim of approximately 0.5 mm of tissue bordering the infarction, while the spared myocardium was defined as unaffected tissue of the LV, both septum and free LV wall (Fig. 1). Dissected tissue was collected in RNA-se free lysisbuffer (Absolutely RNA® Microprep Kit, Stratagene). RNA was prepared according to the protocol used for whole sections, as described above. cDNA synthesis from total RNA of microdissected samples were reverse transcribed with Sensiscript Reverse Transcriptase in the presence of random hexamer primers, according to manufacturer's instructions (Sensiscript® Reverse Transcriptase, Qiagen Benelux B.V., Netherlands). For qRT-PCR, a selection of TaqMan “assay by demand” primer/probe sets were used: β2-microglobulin (β2M), Stromal Derived Factor-1 (SDF-1), Stem Cell Factor (SCF), Transforming Growth Factor-β1 (TGF-β1), basic Fibroblast Growth Factor (bFGF), Vascular Endothelial Growth Factor-Receptor2 (VEGF-R2) and CXCL2 (Applied Biosystems, Foster City, CA). CXCL2 (KC, Gro-α) was chosen as the murine Interleukin-8 (IL-8) homologue, but will be referred to as IL-8 in this article.

Fig. 1

Laser dissected areas. Hematoxilin stained cross-sections of a heart 1 day after reperfusion (original magnification 100×). The heart is dissected in infarcted area (i), borderzone (B) and spared myocardium (S) by laser-dissection-microscopy. The borderzone is defined as the rim of approximately 0.5 mm of myocardium bordering the infarct area.

2.6. Quantitative RT-PCR

Quantitative RT-PCR, according to the “Taqman” principle, was performed in 384-well microtitre plates in a final reaction volume of 10 μl, composed of 5 μl TaqMan universal PCR Master Mix, 0.5 μl primer/probe mix (Applied Biosystems), 3.5 μl Milli Q and 1 μl cDNA, using an ABIPRISM7900HT System (Applied Biosystems). Amplifications were performed starting with a 2 min AmpErase UNG activation step at 50 °C, followed by a 10 min Amplitaq Gold Enzyme Activation step at 95 °C, followed by 45 cycles of denaturation at 95 °C for 15 s and combined primer annealing/extension at 60 °C for 1 min. Fluorescence increase of 6-FAM (6-Carboxyfluorescein) was automatically measured during PCR. Cycle thresholds (CT) for the individual reactions were determined using ABI Prism SDS 2.0 data processing software (Applied Biosystems). All cDNA samples were amplified in triplicate and normalized against a triplicate of β2M in the same plate.

2.7. Western blot analysis

Protein samples were obtained from 5 μm cryo sections lysed in a buffer containing 20 mM Tris–HCl, 5.0 mM EDTA, 2 mM EGTA, 100 mM NaCl, 0.05% SDS, 0.05% NP-40 and Proteinase inhibitor cocktail (Sigma, Zwijndrecht, the Netherlands). 20 μg of protein sample, was loaded on a sodium dodecyl sulfate–15% polyacrylamide gel. Proteins were electrophoretically transferred onto nitrocellulose membranes (Protron, Schleicher and Schnell Biosciences, Dassel, Germany). Nonspecific protein binding was blocked using PBS with 0.1% Tween 20 and 5% nonfat dry milk (Campina, Eindhoven, The Netherlands). After incubation with polyclonal rabbit anti-human VEGF-A-20 (Santa Cruz Biotechnology, Santa Cruz, USA) 1:200 for 1 h, the membranes were incubated with peroxide conjungated secondary antibodies 1:3000 for 1 h and detection was done using Supersignal West Dura Extended Duration Substrate (ECL, Pierce Chemical Company, Rockford, USA).

2.8. Statistical analyses

To determine differences in gene expression CT values were normalized against β2M expression (internal control housekeeping gene) levels by employing the δCT method [11]. Fold difference in gene expression was equated relative to absolute gene expression levels in control mice (Fig. 3). Methodologically, large increases in gene expression can be expected from, and was seen for genes that were not expressed, or very lowly expressed in controls (on/off phenomenon). This does not necessarily imply a larger absolute quantity of transcripts. In this study we chose to present changes in gene expression after MI, rather than displaying absolute quantities [11].

Fig. 3

Absolute gene expression levels in controls. The absolute levels of basal gene expression in control hearts, expressed as CT value, differed considerably for the various genes. A low CT value represents a high gene expression. SDF-1, Flk-1 and VEGF-A were constitutively highly expressed. G-CSF was not expressed in control hearts and therefore the CT value was set on 40. Fold difference of gene expression was equated relatively to these absolute gene expression levels in control mouse hearts. B2M served as an internal valuation standard.

Differences in expression levels between experimental and control mice were expressed as fold difference of expression, calculated as 2−δδCT[11]. The linear range of amplification was determined for the primer sets by using tenfold serially diluted cDNAs. In all cases data confirmed that it was valid to employ the 2−δδCT for processing our qRT-PCR data.

Data are expressed as mean±S.E.M. Data were analyzed using statistical software (GraphPad Prism, version 3.00, GraphPad Software Inc.). Inter-rater agreement was evaluated by paired samples t-testing between controls and MI and between the reperfusion and ligation model. The significance of differences in the findings was evaluated by a one-way ANOVA followed by a two-tailed unpaired Student's t-test. A difference of P<0.05 was considered statistically significant.

3. Results

3.1. Morphology and histology

Infarct size, expressed as the percentage infarction of LV surface was 26.5±9.3% for the reperfusion model and 33.9±17.4% for the ligation model, values which were statistically not significantly different at all investigated time points. After permanent ligation, infarcts were transmural and showed more progressive wall thinning and LV lumen dilatation than after (non-transmural) reperfused infarcts (Fig. 2A). In both models, massive numbers of neutrophils and macrophages infiltrated the infarct in the first week post-MI (Fig. 2B,C). Previously, we reported a detailed comparison of remodelling, cellular influx and vascularization between both models [9].

Fig. 2

The morphology of the heart after myocardial infarction. A) Masson trichrome stained cross section of the heart at 28 days post-MI induced by permanent ligation (right) or ligation of 30 min followed by reperfusion (left). There is more pronounced LV wall thinning and lumen dilatation after ligation, while in both models massive collagen deposition (blue) was observed (original magnification 12.5×). B/C) Details of hematoxilin/eosin stained sections of the infarcted area at 1 day (B) and 7 days (C) post-MI of the reperfusion (left) and ligation model (right) (original magnification 200×). At 1 day post-MI in both models remnants of cardiomyocyte are visible. In reperfused MI there is a great infiltration with neutrophils (specific staining not shown), which is less in ligated MI. At 7 days post-MI cardiomyocyte remnants are cleared after reperfusion, where after ligation few remnants are still present. In both models there is a great influx of inflammatory cells, which is predominantly comprised of macrophages.

3.2. Quantitative gene expression analysis on whole heart cross-sections

The absolute expression level in healthy non-infarcted controls of the analyzed genes (Fig. 3), expressed as CT value, ranged from CT of 25 (which means high expression) for SDF-1 and Vascular Endothelial Growth Factor-A (VEGF-A) to undetectable (CT set on 40, very low or no expression) for Granulocyte-Colony Stimulating Factor (G-CSF), while β2M had a CT value of 22. The household gene β2M (CT of 22) served as the internal standard to calculate δCT values, since it was constitutively and highly expressed at a remarkably constant level (both in healthy controls and after MI (CT=22.34±0.65 for all animals). We did not observe gene expression differences between sham-operated and non-operated controls (P=0.41), indicating that surgery itself, at least in our hands, had no influence on cardiac gene expression.

In general, inflammatory genes showed the highest expression levels within 48 h post-MI in both models and declined (rapidly) thereafter. In the reperfusion model (Fig. 4), the pro-inflammatory genes Tumor Necrosis Factor-α (TNFα), IL-1β and IL-6 peaked at 24 h post-MI (Fig. 4A). IL-6 expression had the highest fold difference of the inflammatory group with a peak of 420-fold difference at 24 h post-reperfusion. Although the pro-inflammatory genes showed a similar course of expression, TNF-α was relatively lowly expressed in this group with a peak of 6-fold difference and was already back at control levels at 7 days post-MI whereas IL-1β and IL-6 expression was still increased. IL-6 gene expression was still slightly elevated at 28 days post-MI after reperfusion. In the permanent ligation model (Fig. 5), gene expression of TNF-α at 24 h and 2 days was lower than in the reperfusion model (P<0.05). Also the peak of IL-1β gene expression was 24 h later after ligation(rep vs. lig 24 h P<0.001). Moreover, after ligation IL-6 gene expression had already peaked at 6 h and its expression remained higher at 2 days post-MI than after reperfusion (rep vs. lig at 6 h, 24 h and 2 days P<0.01).

Fig. 5

Gene expression after ligation MI. Fold difference in gene expression in whole heart cross-sections after ligation of inflammatory genes (A), angiogenic genes (B) and stem-cell related genes (C) at 6 and 24 h and 2, 7 and 28 days. * indicates a statistically significant difference compared to control hearts. Gene expression patterns after ligation is relatively similar to reperfused MI (Fig. 4). Also after ligation, highest fold difference was in the inflammatory group and for G-CSF and GM-CSF.

Fig. 4

Gene expression after reperfusion MI. Fold difference in gene expression in whole heart cross-sections after reperfused MI of inflammatory genes (A), angiogenic genes (B) and stem-cell related genes (C) at 6 and 24 h and 2, 7 and 28 days. * indicates a statistically significant difference compared to control hearts. Fold differences were highest in the inflammatory gene group and for G-CSF and GM-CSF in the stem cell-related gene group. In the angiogenic and stem cell-related group, there were no or only small differences compared to control hearts, except for IL-8, G-CSF and GM-CSF, which are genes that also relate to the inflammatory response.

After reperfusion, the expression of IL-10, a potent anti-inflammatory cytokine and inhibitor of the expression of pro-inflammatory cytokines TNF-α, IL-1β and IL-6, was already increased by 4-fold at 6 h post-MI and peaked 2 days post-MI with a 10-fold difference compared to controls. IL-10 expression remained slightly elevated until 28 days post-MI after reperfusion. After permanent ligation however, IL-10 expression peaked at 7 days instead of 2 days post-MI (rep vs. lig at 7 days P<0.001) and was back at control level at 28 days post-MI.

The inflammatory cytokines Monocyte Chemoattractant Protein-1 (MCP-1, CCL2) and Macrophage Inflammatory Protein-1α (MIP-1α, CCL3), two potent chemo-attractants for monocytes and neutrophils respectively, both peaked at 2 days post-MI (50-fold and 32-fold respectively). Although MCP-1 was back at control levels after 7 days, MIP-1α remained slightly elevated throughout 28 days post-reperfusion. After permanent ligation, the expression dynamics of MCP-1 was similar to reperfusion, but gene expression of MIP-1α was more constantly elevated from 6 h until 7 days post MI after ligation, without a peak in expression at 24 h or 2 days (rep vs. lig at 6 h, 24 h, 2 days and 7 days P<0.05).

Of the representative angiogenic genes investigated (Figs. 4B and 5B), the only factor that was highly expressed in both models was CXCL-2 (KC), the murine homologue of human IL-8, which also exerts inflammatory properties as a neutrophil attractant and activator. Nevertheless, IL-8 peaked with 27.5-fold and 29.6-fold difference in reperfusion and ligation respectively at 6 h (P>0.05) and gradually declined thereafter to control levels at 7 days post-MI similarly in both models. bFGF, a potent angiogenic factor [12], was only slightly increased at 24 h and 2 days in the reperfusion model. In the ligation model, there was also a slight increase in gene expression at 6 h post-MI (lig vs. rep P<0.01). VEGF-A, another powerful angiogenic factor [13–15], showed virtually no changes in expression, except for a transient small downregulation (−2.8 fold) at 7 days in both models and after ligation a persistent slightly negative expression 28 days (rep vs. lig at 28 days P<0.01). VEGF-R2 (Flk-1), the VEGF receptor enrolled in neo-vascularization, was slightly downregulated at 2 and 7 days in the reperfusion model (2.2-fold and 2.3-fold). In the permanent ligation model, this downregulation was significantly lower and extended to 28 days (rep vs. lig at 28 days P<0.01). However, as mentioned earlier VEGF-A and VEGF-R2 genes were both constitutively expressed at high levels in control hearts (Fig. 3). Therefore their absolute mRNA quantities may still be high despite the small downregulation in gene expression post-MI.

Regarding the stem cell factor group (Figs. 4C and 5C), SDF-1 and SCF are potent chemoattractants for stem cells [16]. Moreover, SDF-1 exerts regenerative capacity when overexpressed in the heart [17]. SDF-1 gene expression was unchanged in the reperfusion model compared to uninfarcted control hearts and was even slightly downregulated in the ligation model (2.3 fold) at 28 days post-MI (lig vs. rep P>0.05). Furthermore, gene expression of SCF did virtually not change compared to control hearts, except for small downregulations at 7 days in both models and also at 28 days (by 2.0 fold) after permanent ligation. C-kit and CXCR4 are the receptors for SCF and SDF-1 respectively. Gene expression of these receptors also only minimally changed post-MI in both models, with fold differences up to maximally 2.5 times (for CXCR-4 expression at 2 days after ligation).

G-CSF and GM-CSF are both potent mobilizers of stem cells, but are also involved in maturation of progenitor cells into granulocytes and macrophages. G-CSF and GM-CSF were very highly upregulated at 6 h post-MI (232-fold and 42-fold respectively after reperfusion), but their expression returned to uninfarcted control levels at 7 days post-reperfusion and at 28 days post-ligation. Although G-CSF expression was similar in both models, GM-CSF expression was higher at 2 and 7 days after ligation compared to reperfusion (P<0.05).

Erythropoietin-Receptor (EPO-R), the receptor of EPO present on BMSCs and involved in vasculogenesis and myogenesis [18–20], was continuously expressed at control level post-MI in both models. Both IGF-1α and HGF, cardioprotective factors and potential mitogens for cardiomyocytes [21,22], showed only modest increase predominantly at 2 and 7 days post-MI. IGF-1α was slightly upregulated after reperfusion on 2 days (rep vs. lig P<0.05) and on 7 days in both models (3-fold difference). The upregulation of HGF was small and comparable in both models. TGF-β1, promotor of myogenic differentiation of BMSCs [23], but also known as anti-inflammatory protein and stimulator of matrix deposition [24], had a slight (3-fold) increase in expression at 24 h and 2 days in both models. Gene expression of TGF-β1 was still slightly increased 7 days post-ligation compared to reperfusion (P>0.05).

3.3. Local gene expression pattern analysis

In order to determine the contribution of the specific regions of the myocardium in gene expression post-MI, we investigated the levels of gene expression in amounts of tissue, extracted by laser dissection microscopy from the infarcted area, the border zone and the spared myocardium of the LV wall (Fig. 1). We determined the expression levels of IL-8, SDF-1, SCF, TGF-β1, bFGF and Flk1. Except for IL-8, these stem cell-related and angiogenic genes did not, or only slightly changed in whole heart cross sections compared to control hearts. We were interested in regional differences of gene expression in the separate regions that could have been veiled in whole heart cross sections.

In general, there were often significant differences in gene expression levels between the dissected regions (Fig. 6). Conversely, the level of gene expression in the spared myocardium was always comparable to controls except for IL-8 gene expression at day 1 in both models and SDF-1 gene expression on day 1 and 28 in the ligation model. IL-8 was the only exception on whole heart cross sections that showed highly upregulated gene expression primarily in the first 2 days post-MI, which is reflected in the IL-8 gene expression in separate regions at 1 day post-MI. Remarkably, the increased IL-8 expression in both models was located predominantly in the infarcted area and the border zone, and only to a far lesser degree in the spared myocardium.

Fig. 6

Gene expression pattern analysis in micro-dissected regions of myocardium. Equal amounts of tissue from infarcted area (white bars), border zone (grey bars) and spared myocardium (black bars) were collected using laser dissection microscopy and gene expression was measured at 1, 7 and 28 days post-MI. Graphs on the right side represent the reperfusion model, and graphs on the left side the permanent ligation model. * indicates a significant difference compared to control non-infarcted hearts.

Although SDF-1 expression in whole cross sections was similar to control levels, with the exception of a small downregulation in the ligation model 28 days post-MI, the measurement of SDF-1 gene expression in the separate regions of the heart elucidated that there was a significant downregulation of SDF-1 located in the infarcted area at 1 and 28 days in both models (P<0.05).

For SCF, the strongest downregulation of gene expression was again localized in the infarcted area, and not in the border zone nor in the spared myocardium in both models at all time points. The highest gene expression of TGF-β1 was predominantly observed in the border zone in both models. bFGF expression in the separate regions was ambiguous. The upregulation at 1 day post-MI of bFGF mainly originated from the infarcted region in both models, whereas at 7 days the highest expression was observed in the borderzone. In contrast, at 28 days bFGF gene expression was downregulated in the infarcted region after reperfusion, whereas in the borderzone and in the spared myocardium as well as in all regions after ligation gene expression of bFGF was similar to control levels.

Similar to SCF in both models, VEGF-R2 was mostly downregulated in the infarct zone only (7 days and 28 days post-MI). The downregulation of VEGF-R2 in the infarcted area after reperfusion 28 days post-MI was not reflected in whole heart cross-sections, which we could not explain. At last, we would like to note that adding the folds gene expression of the separate regions after laser dissection does not always exactly equals fold difference of whole hearts. We investigated equal amount of dissected tissue from the infarctzone, borderzone and spared myocardium. The limitation of this procedure is that fold differences of the individual regions do not add up to the fold difference found for whole heart sections, because these regions are not equally distributed in the whole heart sections.

3.4. Protein expression

To assess whether gene expression correlated with the expression of the encoded protein, we performed immunohistochemical stainings for TGF-β, VEGF-A, EPO-R and VEGF-R2 and did Western blot analysis of VEGF-a (representative micrographs in Fig. 7, see figures in supplementary data for more micrographs of the infarcted area, border zone and spared myocardium of both MI models on 1, 7 and 28 days post-MI). Although we could not exactly quantify protein expression in immunohistochemical stainings, the spatio-temporal pattern of protein expression of TGF-β and VEGF-R2 was similar to localised gene expression, determined in micro-dissected areas, at the same time (Fig. 7). The immunohistochemical staining of VEGF-A showed a reduction of VEGF-A in the infarcted area, which explains the decrease in VEGF-A at 7 days post-reperfusion and 7 and 28 days post-ligation. Furthermore, Western blot analyses for the expression of VEGF-A, showed a expression pattern similar to the gene expression for VEGF-A, i.e. a reduced expression of VEGF-A at 7 day post-reperfusion and similarly at day 28 post-ligation. The decreased expression on day 7 post-ligation is not observed, which we could not explain. The high level of VEGF-A expression in the sham operated mice indicated that overall post-MI VEGF expression was reduced, which was comparable to the decrease observed by qRT-PCR.

Fig. 7

Protein expression analysis. Representative micrographs of immunohistological stainings of TGF-β (A), VEGF-R2 (B) and VEGF-A (C). Original magnification 200×. Red staining indicates protein expression. Here we show protein expression parallels spatial and temporal patterns of gene expression, presented at the right side. The protein TGF-β is mostly expressed in the border zone (A). VEGF-R2 expression is reduced primarily in the infarcted area. The decrease in VEGF-A expression originates from the infarcted area. Western blot analysis of VEGF-A (C) indicated VEGF-a protein expression is decreased post-MI.

4. Discussion

In this study we show the gene expression profile of representative stem cell-related, angiogenic and inflammatory genes up to 1 month post-MI. Early after MI the expression of inflammatory genes predominated. The expression of pro-angiogenic genes such as bFGF, VEGF-A and VEGF-R2 changed only marginally post-MI. Stem cell gene expression remained unchanged post-MI with the exception of high gene expression of G-CSF and GM-CSF, which are genes that are also known to enhance the inflammatory response. Local analysis of micro-dissected regions revealed that the infarct was the primary location of the decline in gene expression of SDF-1, SCF and VEGF-R2. Furthermore, gene expression in the spared myocardial region virtually did not change.

We investigated two murine models of acute myocardial infarction (i.e. ligation and reperfusion) in which we observed similar spatio-temporal gene transcription profiles, in spite of the considerable histological differences between ligation and reperfusion shown here and in previous studies [9]. These differences include an increased left ventricular dilatation and more progressive wall thinning after ligation, whereas a higher and faster inflammatory cell influx and increased neovascularisation was characteristic after reperfusion [9]. Our current results show that the augmented and faster inflammatory response after reperfusion in comparison to ligation, coincided with a higher early expression of TNF-α and an earlier peak in gene expression of IL-1β, MIP-1α and IL-10 after reperfusion. Also, the prolonged presence of neutrophils after permanent ligation coincided with a prolonged presence of MIP-1α. This may indicate a role for MIP-1α in neutrophil attraction after MI. The upregulation of anti-inflammatory genes IL-10 and TGF-β was longer after ligation compared to reperfusion, which was concomitant with lower number of macrophages up to 1 week post-MI and the slower clearance of cardiomyocyte remnants. After ligation the expression of the genes for VEGF-A and its receptor VEGF-R2 were strongly downregulated compared to reperfusion throughout the 28 days analysis period. This coincides with reduced neovascularization in the ligation model, which we already observed in previous investigations [9]. In conclusion, it appears that the distinct pathohistological differences between ligation and reperfusion can at least be partly explained by the differences in gene expression levels. The strong inflammatory gene upregelation in the first week post-MI seems to precede the adverse remodelling process, which starts directly after MI and is already significantly different 4 days post-ligation [9].

Here, we aimed to study and compare gene expression at the transcriptional level of factors that are involved in processes that occur after MI such as inflammation, angiogenesis, remodelling and stem cell-driven tissue repair. Our approach of quantitative RT-PCR allowed for the simultaneous analysis a large array of representative genes using the same sensitive detection technique. In fact, this allowed us to investigate the network of selected genes involved in the post-MI processes in a precise quantitative way. We further confirmed the gene expression data by protein expression using immunohistochemical stainings and Western blot analysis, which implied that the protein expression of the investigated factors did parallel the gene expression profile in time, although we could not exactly quantify this.

Concluding from our gene expression data we predict that the administration of combinations of angiogenic and stem cell-related factor might benefit the post-MI repair process. There are studies that report the beneficial effect of (re)introducing one or two factors in stem cell-mediated cardiac repair. Askari et al. showed a higher number of invading BMSC in the heart that was accompanied by an increased angiogenic response within the MI after transplantion in situ of cardiac fibroblasts that overexpress SDF-1α and G-CSF therapy [17]. Others show that the administration of G-CSF and SCF post-MI resulted in attenuated left ventricular remodelling, improved angiogenesis and new myocytes [25,26]. In our study we show a slightly decreased gene expression of SDF-1 and SCF primarily located within the infarcted region. We have found an increase in G-CSF gene expression, although limited to predominantly the first 2 days post-MI. These studies show that introduction of one or two factors over a longer period can positively influence the remodelling process post-MI, most likely because of the initial shortage of these factors post-MI, as we show here.

The gene expression analysis of laser dissected areas revealed that the infarcted region and the borderzone are mostly responsible for gene expression changes. In current investigation, we cannot exactly indicate the cell type responsible for gene expression changes. Neutrophil influx peaked at 24 h post-MI, macrophages at 4 days post-MI and lymfocytes were only sporadically present [9]. Genes that are highest expressed at 6 h post-MI (such as G-CSF, GM-CSF and IL-8) probably precede inflammatory cell influx, but genes that peak at 2 days such as MCP-1 and MIP-1α may well be expressed by infiltrated inflammatory cells.

To conclude, this study gives an overview of the local dynamics of various stem cell-related cytokines and growth factors early after MI that forms the molecular fundaments for further development of stem cell-mediated repair of the post-schemic heart.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cardiores.2006.11.030.


  • Time for primary review 34 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
View Abstract