Granulocyte colony-stimulating factor treatment plus dipeptidylpeptidase-IV inhibition augments myocardial regeneration in mice expressing cyclin D2 in adult cardiomyocytes

Abstract

Aims

Although pharmacological interventions that mobilize stem cells and enhance their homing to damaged tissue can limit adverse post-myocardial infarction (MI) remodelling, cardiomyocyte renewal with this approach is limited. While experimental cell cycle induction can promote cardiomyocyte renewal following MI, this process must compete with the more rapid processes of scar formation and adverse remodelling. The current study tested the hypothesis that the combination of enhanced stem cell mobilization/homing and cardiomyocyte cell cycle induction would result in increased myocardial renewal in injured hearts.

Methods and results

Myocardial infarction was induced by coronary artery ligation in adult MHC-cycD2 transgenic mice (which exhibit constitutive cardiomyocyte cell cycle activity) and their non-transgenic littermates. Mice were then treated with saline or with granulocyte colony-stimulating factor (G-CSF) plus the dipeptidylpeptidase-IV (DPP-IV) inhibitor Diprotin A (DipA) for 7 days. Infarct thickness and cardiomyocyte number/infarct/section were significantly improved in MHC-cycD2 mice with G-CSF plus DipA treatment when compared with MHC-cycD2 transgene expression or G-CSF plus DipA treatment alone. Echocardiographic analyses revealed that stem cell mobilization/homing and cardiomyocyte cell cycle activation had an additive effect on functional recovery.

Conclusion

These data strongly suggest that G-CSF plus DPP-IV inhibition, combined with cardiomyocyte cell cycle activation, leads to enhanced myocardial regeneration following MI. The data are also consistent with the notion that altering adverse post-injury remodelling renders the myocardium more permissive for cardiomyocyte repopulation.

Introduction

Many forms of cardiovascular disease result in either acute or progressive loss of cardiomyocytes, which in turn often results in adverse remodelling of the ventricle.¹ Therapeutic approaches aimed at altering this process are in development and include efforts to impact directly on post-injury remodelling,² efforts to salvage at-risk cardiomyocytes,³ and efforts to replace lost cardiomyocytes via cell cycle activation or transplantation of cardiomyocytes and/or cardiomyogenic stem cells.⁴

Direct engraftment or mobilization of marrow-derived stem cells has been employed to reduce adverse post-MI remodelling in animal models, largely through increasing the number of vascular cells and decreasing the loss of at-risk cardiomyocytes.⁵ These processes are thought to be mediated largely via the stromal-derived factor 1α (SDF-1α)/CXCR4 signalling pathway.^6,7 The homing properties of SDF-1α are known to be inactivated by N-terminal cleavage via the serine protease CD26/dipeptidypeptidase-IV (DPP-IV). Recent studies have demonstrated that a dual pharmacological therapy which combined stem cell mobilization with granulocyte colony-stimulating factor (G-CSF) and inhibition of SDF-1α degradation (via DipA-mediated inhibition of CD26/DPP-IV activity) enhanced stem cell recruitment to injured myocardium in an SDF1/CXCR4-dependent manner.⁸ Combinatorial G-CSF plus DipA treatment resulted in enhanced myocardial function, reduced infarct size, reduced adverse left ventricular (LV) wall thinning, reduced cardiomyocyte apoptosis, and increased border zone vascularization when compared with treatment with either DipA or G-CSF alone. However, this beneficial modulation of adverse remodelling did not lead to overt replacement of lost cardiomyocytes.

A number of genetic pathways have been identified which can be exploited to drive cardiomyocyte proliferation in normal or injured mammalian hearts.⁹ For example, expression of the D-type cyclins (obligate co-factors of cyclin-dependent kinase 4) under the regulation of the α-cardiac myosin heavy chain (MHC) promoter is sufficient to initiate cardiomyocyte cell cycle activity in adult hearts.^10,11 Interestingly, mice expressing cyclin D2 (MHC-cycD2 mice) exhibited sustained cardiomyocyte cell cycle activity following myocardial injury.¹⁰ Cardiomyocyte cell cycle induction in MHC-cycD2 mice resulted in repopulation of the infarcted myocardium, with a progressive reduction in infarct size and a concomitant improvement in LV function.^10,12 Nonetheless, significant scarring of the myocardium persisted in these animals, limiting the ultimate regenerative effect of transgene-induced cardiomyocyte cell cycle activity.

The goal of the current study was to determine whether the beneficial impact of G-CSF plus DipA treatment on adverse post-injury remodelling would have an additive effect with the beneficial impact of cyclin D2-mediated cardiomyocyte cell cycle activity following MI. Accordingly, MHC-cycD2 mice and their non-transgenic (NON-TXG) littermates were subjected to permanent coronary artery occlusion, followed by treatment with either saline or G-CSF plus DipA. The combination of G-CSF plus DipA treatment and cardiomyocyte cell cycle activity reduced LV thinning, increased infarct cardiomyocyte content, and enhanced cardiac function, when compared with stem cell mobilization or cell cycle activity alone. These data suggest that limiting adverse post-injury remodelling renders the infarcted myocardium more permissive to cell cycle-mediated regenerative growth.

Methods

Mice

MHC-cycD2¹⁰ and MHC-nLAC¹³ mice utilize the mouse α-cardiac MHC promoter to target expression of a cDNA encoding cyclin D2 (nucleotide residues 267–1152) or a nuclear β-galactosidase reporter, respectively, to cardiomyocytes. All mice analysed at 7 days post-MI carried the MHC-nLAC reporter transgene, while 13% (5 of 37) of the mice analysed at 60 days post-MI carried the reporter transgene. All mice were in an inbred DBA/2J genetic background (>30 backcross generations). All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee. All analyses were initiated when mice reached 10–12 weeks of age.

Myocardial infarction

Myocardial infarction (MI) was performed in MHC-cycD2 mice and their NON-TXG littermates. Mice were intubated and ventilated with 2% isoflurane and supplemental oxygen. Via left thoracotomy, the left coronary artery was ligated at the inferior border of the left auricle as described previously.¹⁴

Experimental design

After MI, the mice were randomly assigned to the following groups: saline-treated NON-TXG (n = 10), G-CSF plus DipA-treated NON-TXG (100 μg/kg/QD i.p. and 70 mg/kg/BID, respectively; n = 7), saline-treated MHC-cycD2 (n = 10), and G-CSF plus DipA-treated MHC-cycD2 (n = 10). Pharmacological treatment was administered for 7 days post-MI. In some instances, the mice also carried the MHC-nLAC reporter transgene.

Echocardiography

Mice were lightly anaesthetized with 1.5% isoflurane until the heart rate (HR) stabilized at 400–500 b.p.m. Two-dimensional short-axis images were obtained with a high-resolution Micro-Ultrasound system (Vevo 770, VisualSonics Inc., Toronto, Canada) equipped with a 40 MHz mechanical scan probe. Heart rate, end-systolic diameter, end-diastolic diameter, stroke volume, fractional shortening, and ejection fraction (EF) were calculated with Vevo Analysis software (version 2.2.3) as described previously.¹⁵ Haemodynamic measurements and data analyses were performed independently by a blinded person using Vevo Analysis software (version 2.2.3).

Histology

Hearts were excised and fixed in 10% zinc formalin solution. Hearts were then embedded in paraffin and sectioned (10 µM thick) using standard protocols.¹⁶ Haematoxylin and eosin (H&E) and Masson's trichrome staining were performed according to the manufacturer's protocol (Sigma-Aldrich, St Louis, MO, USA). Infarct size was calculated as the average of four coronal sections sampled at 2 mm intervals from the apex to the base using the following formula developed by Pfeffer et al.¹⁷: infarct size = [coronal infarct perimeter (epicardial plus endocardial)/total coronal perimeter (epicardial plus endocardial)] × 100. Infarct wall thickness was measured in Masson's trichrome stained sections by taking the average length of five segments along evenly spaced radii from the centre of the LV through the infarcted free LV wall.¹⁸ Peri-infarct vessel content was quantified in Masson's trichrome stained sections using a stereological approach.¹⁹ Vascular structures (identified by the presence of inter-luminal blood cells) within 0.5 mm of the infarct border and which overlaid points on an 8 × 8 grid in micrographs (captured using a ×40 objective) were scored. Infarct vessel content was calculated as (the number of grid points overlying vessels/the total grid points) × 100.

Quantification of cardiomyocytes

Cardiomyocyte number/mm² was quantified planimetrically in the infarct area by automated cell counting of at least 50 high-power images per group using ImageJ software. The cardiomyocyte number/infarct/section was calculated by multiplying the cardiomyocyte infarct number/mm² × average infarct wall thickness (mm) × average infarct coronal perimeter (mm). Cardiomyocyte minimal fibre diameter was quantified in high-power images using ImageJ software (at least 50 infarct border cardiomyocytes per sample).

Cardiomyocyte DNA synthesis assays

Cardiomyocyte DNA synthesis was monitored in experimental animals carrying the MHC-nLAC reporter transgene. Cardiomyocyte nuclei in MHC-nLAC mice are readily identified in tissue sections by incubation with the chromogenic β-galactosidase substrate X-GAL (5-bromo-4-chloro-3-indolyl-β-d-galactoside) which gives rise to a blue reaction product. For tritiated thymidine (³H-Thy) incorporation,²⁰ mice received a single injection of radiolabelled nucleotide (200 µCi i.p. at 20 Ci/mM, Perkin Elmer, Boston, MA, USA) and were sacrificed 4 h later. Hearts were harvested, immersion fixed in 50 mM cacodylic acid/1% paraformaldehyde, cryoprotected in 30% sucrose, embedded, and sectioned at 10 µM using standard histological techniques.¹⁶ Sections were reacted with X-GAL (1 mg/mL in 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 2 mM magnesium chloride in phosphate-buffered saline). The sections were counter-stained with DAPI, and autoradiographic emulsion was applied and processed as described previously.²⁰ Cardiomyocyte DNA synthesis, identified by the co-localization of blue nuclear β-galactosidase activity and silver grains, was scored at the infarct border zone (defined as myocardial tissue within 0.5 mm of the fibrous scar tissue) and within the infarct LV wall.

Statistics

Results were expressed as mean ± SEM. Multiple group comparisons were performed by one-way analysis of variance (ANOVA) followed by the Bonferroni procedure for comparison of means. Comparisons between two groups were performed using the unpaired Student's t-test. All tests were two-sided. Statistics were calculated using the data analysis tool from Microsoft Office Excel 2007. Data were considered statistically significant at P< 0.05.

Results

Granulocyte colony-stimulating factor plus DipA treatment and cardiomyocyte cell cycle induction have an additive beneficial effect on post-myocardial infarction wall thinning

To determine whether G-CSF plus DipA treatment and cardiomyocyte cell cycle activity have an impact on post-infarct remodelling, MHC-cycD2 mice and their NON-TXG littermates were subjected to permanent coronary artery ligation followed by 7 days of treatment with saline or with G-CSF plus DipA. The animals were sequestered for an additional 53 days, and then hearts were harvested and subjected to histological analyses. Prominent infarcts were observed in saline-treated NON-TXG mice at 60 days post-MI (Figure 1 and Table 1A). In agreement with previous observations, G-CSF plus DipA treatment and cardiomyocyte cell cycle activation each resulted in a significant decrease in infarct size. Combined G-CSF plus DipA treatment and cyclin D2 expression did not have an additive effect on infarct size reduction. However, stem cell mobilization/homing and cell cycle activation did have an additive effect on the lessening of post-MI wall thinning. Left ventricular infarct wall thinning was very pronounced in saline-treated NON-TXG mice at 60 days post-MI (Figure 2 and Table 1B). Left ventricular infarct wall thickness in mice with G-CSF plus DipA treatment was increased to 1.8 ± 0.14× vs. saline-treated NON-TXG mice. Saline-treated MHC-cycD2 mice exhibited a moderate (1.4 ± 0.10×) increase in LV infarct wall thickness. Granulocyte colony-stimulating factor plus DipA-treated MHC-cycD2 mice exhibited an LV infarct wall thickness that was increased to 2.4 ± 0.17× vs. saline-treated NON-TXG animals.

Figure 1.

Infarct size at 60 days post-myocardial infarctionMI. (A) Representative images of haematoxylin and eosin-stained transverse sections. Bar = 500 µm. (B) Infarct size (%, mean ± SEM) for: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice. *P< 0.05 vs. saline/NON-TXG.

Table 1.

Impact of stem cell mobilization and/or cardiomyocyte cell cycle activation on left ventricular remodelling and cardiomyocyte renewal

Parameter	Saline/NON-TXG	G-CSF + DipA/NON-TXG	Saline/MHC-cycD2	G-CSF + DipA MHC-cycD2
A. Infarct size (%)	36.9 ± 3.06 (n = 10 mice)	25.5 ± 2.20* (n = 7 mice)	27.9 ± 1.95* (n = 10 mice)	25.0 ± 3.02* (n = 10 mice)
B. Left ventricle infarct wall thickness (mm)	0.15 ± 0.010 (n = 10 mice)	0.28 ± 0.021*,† (n = 7 mice)	0.21 ± 0.015* (n = 10 mice)	0.36 ± 0.025*,Ψ,† (n = 10 mice)
C. Peri-infarct vessel content (%)	7.2 ± 1.20 (n = 6 mice)	14.2 ± 0.75*,† (n = 6 mice)	8.0 ± 0.72 (n = 6 mice)	20.2 ± 2.93*,† (n = 6 mice)
D. Cardiomyocyte 3H-thymidine-labelling index (%; n = 4 mice)	0.008 ± 0.0041 (n = 35 706 nuclei)	0.024 ± 0.0020 (n = 49 113 nuclei)	1.58 ± 0.328*,Ψ (n = 23 684 nuclei)	1.81 ± 0.534*,Ψ (n = 33 226 nuclei)
E. Cardiomyocyte number/infarct/section	121 ± 7.9 (n = 10 mice)	446 ± 23.7* (n = 7 mice)	569 ± 27.7*,Ψ (n = 10 mice)	1130 ± 168.9*,Ψ,† (n = 10 mice)
F. Minimal cardiomyocyte diameter (μm)‡	19.1 ± 1.07 (n = 10 mice)	19.1 ± 0.62 (n = 7 mice)	20.5 ± 0.32 (n = 10 mice)	20.3 ± 0.60 (n = 10 mice)
G. Cardiomyocyte number/infarct/section (7 days)	141 ± 13.4 (n = 10 mice)	343 ± 72.2* (n = 7 mice)	273 ± 11.7 (n = 10 mice)	369 ± 36.4* (n = 10 mice)
H. Minimal cardiomyocyte diameter (7 days; μm)	18.7 ± 0.30 (n = 10 mice)	17.8 ± 0.56 (n = 7 mice)	19.5 ± 0.44 (n = 10 mice)	18.4 ± 1.44 (n = 10 mice)

All values are mean ± SEM. All assays were performed at 60 days post-injury except sections D, G, and H, which were at 7 days post-injury.

*P < 0.05 vs. saline/NON-TXG.

^†P< 0.05 vs. saline/MHC-cycD2.

^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG.

^‡Minimal fibre diameters were not significantly different among the groups at 60 days post-MI; P = 0.961 for NON-TXG vs. NON-TXG + G-CSF + DipA, P = 0.272 for NON-TXG vs. CycD2 + saline, P = 0.372 for NON-TXG vs. CycD2 + G-CSF + DipA, P = 0.078 for NON-TXG + G-CSF + DipA vs. CycD2 + saline, P = 0.209 for NON-TXG + G-CSF + DipA vs. CycD2 + saline, and P = 0.800 for CycD2 + saline vs. CycD2 + G-CSF + DipA.

Figure 2.

Left ventricular infarct wall thickness at 60 days post-myocardial infarction. (A) Representative photomicrographs of Masson's trichrome-stained sections. Bar = 500 µm. (B) Infarct thickness (mm, mean ± SEM) for: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice.*P < 0.05 vs. saline/NON-TXG; ^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG; ^†P< 0.05 vs. saline/MHC-cycD2.

Granulocyte colony-stimulating factor plus DipA treatment increases vascularization

Our previous data showed that G-CSF plus DPP-IV inhibition with DipA resulted in an increased number of angiogenic CD34⁺/c-kit⁺, CD34⁺/CXCR4⁺, and CD34⁺/Flk1⁺ progenitor cells in the ischaemic myocardium, which was associated with increased vessel density 6 days post-MI.⁸

In order to analyse the long-term effect of this treatment, blood vessel content was quantified within the infarcted myocardium near the border zone at 60 days post-MI (Figure 3 and Table 1C). Blood vessels comprised ∼7% of the area of the infarct near the border zone in saline-treated NON-TXG mice. Granulocyte colony-stimulating factor plus DipA treatment of NON-TXG mice increased vessel content to 2 ± 0.08× vs. saline-treatment mice. Cardiomyocyte cell cycle activation alone did not result in an overt increase in vessel content at 60 days post-MI in saline-treated MHC-cycD2 mice. Vessel content was increased to 2.8 ± 0.34× in G-CSF plus DipA-treated MHC-cycD2 mice when compared with saline-treated NON-TXG mice. Although vessel content in G-CSF plus DipA-treated MHC-cycD2 mice was greater than in G-CSF plus DipA-treated NON-TXG animals, the difference did not reach statistical significance (P = 0.094).

Figure 3.

Peri-infarct vessel density at 60 days post-myocardial infarction. (A) Representative photomicrographs of Masson's trichrome-stained sections. Bar = 20 µm. (B) Infarct vessel area (%, mean ± SEM) for: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice. *P < 0.05 vs. saline/NON-TXG; ^†P< 0.05 vs. saline/MHC-cycD2.

Cyclin D2 expression results in similar levels of cardiomyocyte cell cycle induction in saline vs. granulocyte colony-stimulating factor plus DipA-treated mice

Cardiomyocyte DNA synthesis was monitored as a surrogate readout for proliferation. These experiments utilized the MHC-nLAC reporter transgene, which directs expression of a nuclear-localized β-galactosidase reporter exclusively to cardiomyocytes.¹³ MHC-nLAC and MHC-cycD2 mice were intercrossed, and offspring carrying either the MHC-nLAC transgene alone or both the MHC-nLAC and MHC-cycD2 transgenes were sequestered. At 12 weeks of age, the mice were subjected to permanent coronary artery occlusion, followed by saline or G-CSF plus DipA treatment as described above. On the seventh day (i.e. the last day of treatment), the mice received a single injection of tritiated thymidine and 4 h later hearts were harvested, and then sectioned, reacted with X-GAL, and processed for autoradiography (Figure 4A and B, and Table 1D). Cardiomyocyte DNA synthesis is evidenced by the presence of silver grains over X-GAL-positive nuclei. Very few surviving cardiomyocytes at the infarct border zone or within the infarct itself exhibited cell cycle activity in saline-treated or G-CSF plus DipA-treated NON-TXG mice, although DNA synthesis in non-cardiomyocytes was readily detected (arrowheads identify ³H-thymidine-labelled non-cardiomyocyte nuclei in the micrographs). In contrast, cardiomyocyte cell cycle activity was increased 200 ± 29.4× in mice carrying the MHC-cycD2 transgene (arrows identify ³H-thymidine-labelled cardiomyocyte nuclei). Granulocyte colony-stimulating factor plus DipA treatment did not significantly enhance cardiomyocyte cell cycle activity in these animals. Bromodeoxyuridine incorporation analyses confirmed sustained cardiomyocyte DNA synthesis at 60 days post-MI in the MHC-cycD2 mice, consistent with our previous studies using ³H-thymidine incorporation.¹⁰

Figure 4.

Infarct and infarct border zone cardiomyocyte ³H-Thy-labelling at 7 days post-myocardial infarction. (A) Representative photomicrographs of autoradiograms of X-GAL-stained sections. Arrows indicate cardiomyocyte DNA synthesis (as evidenced by silver grains over blue, β-galactosidase-positive nuclei); arrowheads indicate non-cardiomyocyte DNA synthesis (as evidenced by silver grains over non-blue nuclei). Bar = 20 µm in all panels. (B) Cardiomyocyte ³H-Thy-labelling index (%, mean ± SEM) for: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice. *P < 0.05 vs. saline/NON-TXG; ^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG.

Enhanced cardiac regeneration in cyclin D2 mice treated with granulocyte colony-stimulating factor plus DipA

Given the marked increase in cardiomyocyte cell cycle activity detected in mice carrying the MHC-cycD2 transgene, and given the salutary impact of G-CSF plus DipA treatment on post-infarct remodelling, cardiomyocyte number/infarct/section was quantified in all four groups at 60 days post-MI (Figure 5 and Table 1E). Few cardiomyocytes were observed in the infarcts of saline-treated NON-TXG mice, with the surviving cardiomyocytes localized predominantly along the endocardial surface. Cardiomyocyte number/infarct/section in G-CSF plus DipA-treated NON-TXG mice increased to 3.6 ± 0.20× when compared with saline-treated NON-TXG mice. An even higher increase in cardiomyocyte number (4.7 ± 0.23x) was observed in the infarcts of saline-treated MHC-cycD2 mice at 60 days post-MI, consistent with the increased cardiomyocyte cell cycle activity observed in these animals. Finally, cardiomyocyte number/infarct/section in MHC-cycD2 mice receiving G-CSF plus DipA was increased to 9.3 ± 1.39× vs. NON-TXG mice receiving saline. Since cyclin D2 has been implicated in hypertrophic cardiomyocyte growth,²¹ minimal fibre diameters of cardiomyocytes were analysed. Minimal fibre diameters were not significantly different among the groups at 60 days post-MI (Table 1F).

Figure 5.

Cardiomyocyte content at 60 days post-myocardial infarction. (A) Representative photomicrographs of Masson's trichrome-stained sections. Bar = 50 µm. (B) Cardiomyocyte number/infarct/section at 60 days post-myocardial infarction: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice. *P < 0.05 vs. saline/NON-TXG; ^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG; ^†P< 0.05 vs. saline/MHC-cycD2.

Improved heart function in cyclin D2 mice treated with granulocyte colony-stimulating factor plus dipeptidylpeptidase-IV inhibition

Since the combination of stem cell mobilization/homing and cardiomyocyte cell cycle activity resulted in an additive beneficial impact on myocardial remodelling, vessel density, and particularly infarct cardiomyocyte number, we anticipated that G-CSF plus DipA-treated MHC-cycD2 mice might exhibit superior cardiac function. Two-dimensional echocardiography (Figure 6; see also Table 2 for additional haemodynamic parameters) revealed that EF was markedly compromised in saline-treated NON-TXG mice at 60 days post-MI. Ejection fraction was rescued to a similar degree in G-CSF plus DipA-treated NON-TXG mice and in saline-treated MHC-cycD2 transgenic mice. The combination of G-CSF plus DipA treatment and cardiomyocyte cell cycle activation had an additive effect in rescuing cardiac function following MI; EF increased to 2.3 ± 0.17× vs. saline-treated NON-TXG mice.

Figure 6.

Cardiac function at 60 days post-myocardial infarction. (A) Representative echocardiograms. (B) Ejection fraction (EF%, mean ± SEM) for: (1) saline-treated NON-TXG mice, (2) G-CSF plus DipA-treated NON-TXG mice, (3) saline-treated MHC-cycD2 mice, and (4) G-CSF plus DipA-treated MHC-cycD2 mice. *P < 0.05 vs. saline/NON-TXG; ^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG; ^†P< 0.05 vs. saline/MHC-cycD2. Ejection fraction in non-infarcted, untreated NON-TXG adult mice was 87 ± 1.1.

Table 2.

Impact of stem cell mobilization and/or cardiomyocyte cell cycle activation on cardiac function

Hemodynamic parameters	Saline/NON-TXG	G-CSF + DipA/NON-TXG	Saline/MHC-cycD2	G-CSF + DipA MHC-cycD2
HR (b.p.m.)	461 ± 6 (n = 10 mice)	446 ± 13 (n = 7 mice)	454 ± 15 (n = 10 mice)	447 ± 17 (n = 10 mice)
ESD (mm)	4.9 ± 0.25	3.6 ± 0.21*	4.1 ± 0.19*	3.6 ± 0.24*
EDD (mm)	5.4 ± 0.21	4.3 ± 0.24*	4.9 ± 0.21	4.6 ± 0.21*
SV (μL)	27.1 ± 3.0	31.6 ± 3.9	33.0 ± 3.25	41.7 ± 3.19*
FS (%)	9 ± 1.3	16 ± 0.8*	15 ± 1.3*	22 ± 2.1*,Ψ,†
EF (%)	19 ± 2.8	34 ± 1.5*	32 ± 2.5*	44 ± 3.4*,Ψ,†

All values are mean ± SEM. Haemodynamic parameters analysed by echo 60 days after myocardial infarction. HR, heart rate; ESD, end-systolic diameter; EDD, end-diastolic diameter; SV, stroke volume; FS, fractional shortening; EF, ejection fraction.

*P < 0.05 vs. saline/NON-TXG;

^ΨP< 0.05 vs. G-CSF plus DipA/NON-TXG;

^†P< 0.05 vs. saline/MHC-cycD2.

Discussion

The data presented here demonstrate that a strategy combining stem cell mobilization/homing with cell cycle activation in adult cardiomyocytes has a beneficial effect on myocardial regeneration. Of note, cardiac function and cardiomyocyte number/infarct/section were significantly increased, while adverse thinning of the infarcted LV free wall was less, in animals with both G-CSF plus DipA treatment and cell cycle activation when compared with animals with G-CSF plus DipA or cell cycle activation alone. Infarct border zone vessel density was enhanced predominately by G-CSF plus DipA treatment, while cardiomyocyte cell cycle activity was enhanced almost exclusively by expression of the MHC-cycD2 transgene. Collectively, these data suggest that modulating adverse remodelling in combination with inducing cardiomyocyte renewal results in better post-injury cardiac structure and function when compared with either intervention alone.

The additive effect of G-CSF plus DipA treatment and cardiomyocyte cell cycle induction on LV free wall thickness and cardiomyocyte number/infarct/section is of interest. Previous studies demonstrated that G-CSF plus DPP-IV inhibition limited adverse thinning of the infarcted LV free wall.⁸ The increased cardiomyocyte number/infarct/section at 7 days post-MI (Table 1G), coupled with the absence of an overt change in cardiomyocyte cell cycle activity (Table 1D), supports the notion that stem cell mobilization/homing results in marked cardioprotection; moreover, the bulk of this effect (>75%) is completed by 7 days post-MI. Although the precise mechanism for this observation is not fully understood, our previous data showed that G-CSF plus DipA treatment stimulated SDF-1/CXCR4-mediated recruitment of CD34⁺/c-kit⁺, CD34⁺/Sca-1⁺, CD34⁺/CXCR4⁺, and CD34⁺/Flk1⁺ stem cells, which was associated with reduced cardiomyocyte apoptosis and increased vessel density at the infarct border zone, consistent with early myocardial salvage.⁸ This view is further supported by previous reports showing that transplantation and mobilization of CD34⁺ stem cells contribute to enhanced neovascularization and reduced apoptosis after MI.^22,23 Moreover, G-CSF treatment intrinsically can reduce cardiomyocyte apoptosis following injury.²⁴ It should also be noted that G-CSF plus DipA-treated NON-TXG mice exhibited a slight enhancement of cardiomyocyte DNA synthesis when compared with saline-treated mice (Table 1D). A similar effect on peri-infarct cardiomyocyte cell cycle activity was previously reported following pravastatin-induced mobilization of CD133⁺ and c-Kit⁺ progenitors.²⁵

Previous studies have demonstrated that cardiomyocyte cell cycle induction in MHC-cycD2 mice resulted in increased cardiomyocyte content following injury.^10,26 The progressive increase in cardiomyocyte number/infarct/section observed in the saline-treated MHC-cycD2 mice in the current study (Table 1E and G) is consistent with these previous results. The additive effect that stem cell mobilization had on infarct wall thickness and increased infarct cardiomyocyte survival, in combination with the progressive increase in cardiomyocyte numbers due to cyclin D2-induced cell cycle activity, increased infarct cardiomyocyte content in the G-CSF plus DipA-treated MHC-cycD2 mice to 9.3 ± 1.39× vs. saline-treated NON-TXG animals. This marked increase in infarct cardiomyocyte content likely underlies the observed additive effect of stem cell mobilization and cell cycle induction on cardiac function.

Several other mechanisms may have contributed to the benefits which G-CSF plus DipA treatment imparted on cell cycle-induced cardiomyocyte renewal. For example, G-CSF plus DipA treatment had a profound impact on vessel density at the infarct border zone, which was still apparent at 60 days post-MI in G-CSF plus DipA-treated mice. Previous studies have shown that marrow-derived cells support sprouting and proliferation of endogenous endothelial cells by the secretion of angiogenic growth factors.^27,28 If perfusion was rate-limiting for myocardial regeneration induced by the MHC-cycD2 transgene, enhanced peri-infarct vessel content might contribute to the increase in infarct cardiomyocyte content and, consequently, the improvement in cardiac function, observed in transgenic mice with G-CFS plus DipA treatment.

The impact on adverse post-MI remodelling may also have benefited cell cycle-induced myocardial repopulation. For example, it is essential to maintain the extracellular matrix following transmural infarction in order to prevent dilation or rupture of the damaged area; consequently deposition of extracelluar matrix occurs quite rapidly following MI.^29,30 Cell cycle-dependent cardiomyocyte renewal in the MHC-cycD2 mice occurs at a low rate, requiring an estimated 39 days for the surviving intra-infarct cardiomyocyte population to double (assuming that cardiomyocytes which synthesize DNA divide within 24 h). It is possible that G-CSF plus DipA treatment fundamentally alters the kinetics and/or the composition of healing infarcts, rendering the microenvironment more permissive for cardiomyocyte repopulation. Such a model is consistent with previous observations with the MHC-cycD2 mice. For example, significant scarring persisted in MHC-cycD2 mice for as long as 180 days after permanent coronary artery ligation, where the rate of cardiomyocyte drop-out (and consequently, the rate of reparative extracellular matrix deposition) is quite high.¹² In contrast, cyclin D2-mediated cell cycle induction was sufficient to largely block myocardial scaring in a genetic model of atrial fibrosis resulting from transforming growth factor β-induced cardiomyocyte apoptosis,^26,31 where the time course of cardiomyocyte drop-out and myocardial fibrosis is much slower than that in acute MI. Thus, alterations in the kinetics and content of adverse post-injury remodelling impacted by G-CSF plus DipA treatment might have facilitated cell cycle-induced myocardial repopulation.

Limitations of the study

Although a cardiomyocyte-restricted MHC promoter (which is expressed in differentiated cardiomyocytes) was used to target cyclin D2 expression, the experimental design of the current study does not completely discriminate between enhanced proliferation of pre-existing cardiomyocytes and enhanced proliferation of de novo stem cell-derived cardiomyocytes. However, the following observations favour cardiomyocyte proliferation over de novo stem cell-derived cardiomyocytes as the main contributor for the observed regeneration. First, the ³H-thymidine-labelling index of genetically tagged cardiomyocytes (MHC-nLac) was 200 ± 29.4× increased in transgenic mice, consistent with a high degree of proliferation in differentiated cardiomyocytes. Second, minimal fibre diameters of border zone cardiomyocytes were not significantly different among the groups at 7- and 60 days post-MI (Table 1F and H), arguing against a special pool of small, stem cell-derived cardiomyocytes as has been evoked by others.³² Although these observations are most consistent with a mechanism entailing proliferation of pre-existing cardiomyocytes, definitive proof will require the use of a conditional, cardiomyocyte-restricted cyclin D2 transgene, or alternatively the use of a conditional genetic fate-mapping reporter.³³ However, the origin of the proliferating cardiomyocytes does not alter the main conclusion of the study, namely that lessening adverse post-MI remodelling via stem cell mobilization and increased stem cell homing has an additive effect on myocardial repopulation resulting from enhanced cardiomyocyte proliferation. Finally, a previous study using knock-out animals implicated cyclin D2 in hypertrophic cardiomyocyte growth.²¹ Although hypertrophic growth could also impact cardiac function, minimal fibre diameter measurements have failed to identify cardiomyocyte hypertrophy in normal or infarcted MHC-cycD2 mice when compared with NON-TXG controls (Table 1F and H, see also Pasumarthi et al.¹⁰).

Funding

This work was supported by the National Institutes of Health (grant numbers HL083126 and HL085098); and the German Research Foundation (grant number DFG ZA 575/1-1).

Conflict of interest: The Ludwig-Maximilians-University is the holder of the patents ‘Use of G-CSF for Treating Ischemia’ (EP 03 02 4526.0 and US 60/514,474) and ‘Remedies for Ischemia’ (EP2007/003272 and US 60/792,943).