The regulation of cardiomyocyte proliferation is important for heart development and function. Proliferation levels of mouse cardiomyocytes are high during early embryogenesis and start to decrease at midgestation. Many cardiomyocytes undergo mitosis without cytokinesis, resulting in binucleated cardiomyocytes during early postnatal stages, following which the cell cycle arrests irreversibly. It remains unknown how the proliferation pattern is regulated, and how the irreversible cell cycle arrest occurs. To clarify the mechanisms, fundamental information about cell cycle regulators in cardiomyocytes and cell cycle patterns during embryonic and postnatal stages is necessary. Here, we show that the expression, complex formation, and activity of main cyclins and cyclin-dependent kinases (CDKs) changed in a synchronous manner during embryonic and postnatal stages. These levels decreased from midgestation to birth, and then showed one wave in which the peak was around postnatal day 5. Detailed analysis of the complexes suggested that CDK activities were inhibited before the protein levels decreased. Analysis of DNA content distribution patterns in mono- and binucleated cardiomyocytes after birth revealed changes in cell cycle distribution patterns and the transition from mono- to binucleated cells. These analyses indicated that the wave of cell cycle regulator expression or activities during postnatal stages mainly produced binucleated cells from mononucleated cells. The data obtained should provide a basis for the analysis of cell cycle regulation in cardiomyocytes during embryonic and postnatal stages.
The heart begins to function in the early stages of embryonic development. The regulation of proliferation in cardiomyocytes is important during development because it is required for normal morphogenesis and determines the appropriate heart size, which is necessary for pumping appropriate volume of blood. In mammals, cell proliferation and hypertrophy in cardiomyocytes increase the heart size before and after birth. Proliferation levels of mouse cardiomyocytes are high during early embryogenesis and start to decrease around embryonic day 10–12 (E10–12) (Erokhina 1968; Toyoda et al. 2003). Many cardiomyocytes undergo mitosis without cytokinesis, resulting in binucleated cardiomyocytes during first the 2 weeks in postnatal stages (Soonpaa et al. 1996), following which the cell cycle arrests. Although the proliferation of a limited number of adult cardiomyocytes has been observed, the percentages are extremely low (normal mice 0.0006%; injured mice 0.0083%; injured human 0.015–0.08%; Soonpaa & Field 1997, 1998; Kajstura et al. 1998; Beltrami et al. 2001). The low levels prevent cardiac regeneration by the proliferation of pre-exiting cardiomyocytes. It remains unknown how the proliferation pattern is regulated, and how the irreversible cell cycle arrest occurs during postnatal stages. To elucidate the mechanisms, fundamental information about cell cycle regulators in cardiomyocytes and the cell cycle patterns during development is necessary.
The central cores of the cell cycle machinery are cyclins and cyclin-dependent kinases (CDKs). CDKs are activated by complex formation with cyclins, following which they phosphorylate complex-specific substrates. The sequential activation of CDKs advances the cell cycle. In mammals, the complexes are cyclin D-CDK4/6 (G1 phase), cyclin E-CDK2 (G1/S phase), cyclin A-CDK2/1 (S/G2 phase), and cyclin B-CDK1 (M phase).
In the present study, we describe the detailed expression patterns of these main cyclins and CDKs, and the activity patterns of CDKs during embryonic and postnatal stages. In addition, we also present the cell cycle distribution patterns of both mono- and binucleated cardiomyocytes after birth.
Materials and methods
Mice with the C3H/HeJ Jcl (Clea Japan, Tokyo, Japan) genetic background were used. The presence of a vaginal plug was regarded as E0.5. Mice were injected intraperitoneally with 50 µg/g body weight EdU (5-ethynl-2′-deoxyuridine; Invitrogen) for detection of DNA synthesis (Salic & Mitchison 2008) 2 h before euthanasia.
Western blotting analysis and in vitro kinase assay for CDKs
Ventricles were lysed with a lysis buffer (50 mmol/L HEPES [hydroxyethyl piperazineethanesulfonic acid], 150 mmol/L NaCl, 1 mmol/L ethylenediaminetetraacetic acid [EDTA], 2.5 mmol/L EGTA [ethylene glycol tetraacetic acid], 10% glycerol). Immunoprecipitation and Western blotting analysis were performed as described previously (Shirato et al. 2009 and Toyoda, 2000 #194, respectively). Antibodies used for immunoprecipitation and Western blotting analysis are shown in Table 1. Immunoprecipitates were suspended with 100 µL kinase buffer (50 mmol/L HEPES, 10 mmol/L MgCl2, 2.5 mmol/L EGTA, 0.2 mmol/L dithiothreitol [DTT], 1 µg histone H1, 50 µmol/l adenosine triphosphate [ATP], 0.2 µBq [γ-32P] ATP; Institute of Isotopes, Budapest, Hungary), and incubated at 30°C for 30 min. Reactants were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and phosphorylated proteins were detected by autoradiography. The intensity of the bands in Western blotting analysis or in vitro kinase assay was quantified using Image J, and the signal intensity was normalized using each glyceraldehyde 3-phosphate dehydrogenase (GAPDH) level.
Table 1. Antibodies
IF, immunofluorescence; IP, immunoprecipitation; WB, Western blotting.
SP4; Lab vision Co.
07-687; Millipore Co.
sc-751; Santa Cruz Biotechnology Inc.
CY-A1; Sigma Chemical Co.
GNS1; Thermo Co.
V152; Cell Signaling Technology
sc-260; Santa Cruz Biotechnology Inc.
sc-163; Santa Cruz Biotechnology Inc.
55/Cdk2; BD Transduction Lab.
sc-54; Santa Cruz Biotechnology Inc.
sc-32233; Santa Cruz Biotechnology Inc.
5c5; Sigma Chemical Co.
Dissociation of cardiomyocytes and measurement of DNA content by microphotometry
Dissociation of cardiomyocytes was performed as described previously (Shioya 2007) using smaller cannulae. Dissociated cells were fixed with 4% paraformaldehyde at 4°C for 24 h. The cells were washed with phosphate-buffered saline (PBS) and then smeared on slide glasses. EdU incorporation was detected as described previously (Salic & Mitchison 2008). Cardiomyocytes were stained with an antibody against sarcomeric actin (Table 1), as described in histological methods previously (Motoyama et al. 1997). DNA was stained with 4́6́-diamidino-2-phenylindole dihydrochloride (DAPI). After staining, fluorescence images were photographed, and then DNA content per nucleus and a percentage of EdU positive cells were analyzed. DNA content was measured with Cell Cycle Application Module of MetaMorph software (Molecular Device). Only sarcomeric actin positive cells were analyzed as cardiomyocytes. Mono- and binucleated cells could be easily distinguished, and were examined independently. A total of approximately 500 cardiomyocyte nuclei were analyzed at each stage. NIH3T3 cells in the growth state were also examined as a control experiment.
Expression and activity patterns of cell cycle regulators during cardiac development
The expression levels of the main cyclins (cyclin D1, cyclin E, cyclin A and cyclin B1) and CDKs (CDK4, CDK2 and CDK1) were investigated using Western blot analysis (Figs 1, 2). The expression levels of all proteins changed in a synchronous manner during embryonic and postnatal stages. The levels were high during early embryonic stages and gradually decreased from embryonic day 12 (E12) or E16. The minimum levels occurred on postnatal day 0 (birth data, P0) or P3, then the levels increased again, peaked around P5, and decreased again. All levels were extremely low after P14. Multiple bands of cyclin D1, cyclin B1, CDK2 and CDK1 might be derived from modified forms, such as phosphorylation.
Next, we examined the patterns of complexes and activities of cyclin-CDK complexes. Cyclin E-CDK2, cyclin A-CDK1/2, and cyclin B1-CDK1 were immunoprecipitated using anti-cyclin E, cyclin A, and cyclin B1 antibodies, respectively (Fig. 3A). Cyclin A binds to two partners, CDK2 and CDK1. CDK activity was investigated with an in vitro kinase assay using histone H1 after the immunoprecipitation (Fig. 3B). In all cases, the patterns of the complexes were similar to those of the expression of cyclins and CDKs shown in Figure 1 (Fig. 3A). CDK2 immunoprecipitated with cyclin E or cyclin A is mainly the higher mobility form in which Thr160 is phosphorylated (Gu et al. 1992). The activity patterns were also similar and showed a transient peak around P3 or P5 (Fig. 3B).
It is important to determine whether the CDK activities decreased simply due to a decrease in the levels of the CDK protein bound to cyclins, especially before cell cycle arrest. In addition, the patterns of proteins and activities after birth were slightly different among experiments, probably due to individual differences. For example, the peaks of CDK activity immunoprecipitated with cyclin E were at P3 or P5. Therefore, we repeated the experiments for Cyclin E-CDK2 and cyclin A-CDK1/2 complexes multiple times in postnatal stages (Fig. 4). Data using average levels showed that CDK activities revealed different patterns from those of CDK protein levels. In the case of immunoprecipitates with cyclin E (CycE-IP), the activity levels of CDK2 decreased drastically at P7, earlier than the stage when the CDK2 protein levels decreased (P14). In fact, statistical analysis using Tukey's multiple comparison tests after one-way anova showed that the activity level at P5 and protein level at P10 were significantly higher than all later levels, and not significantly different from those of earlier stages. In the case of CycA-IP, CDK activity drastically and significantly decreased at P7, while the protein levels of CDK2 drastically decreased at P14. Although the protein level of CDK1 showed a decrease at P7, it decreased more drastically at P14. These data suggested that the CDK activities were inhibited by other factors after P7.
These data showed that the levels of expression, complex formation, and activation of main cyclins and CDKs were high and gradually decreased during embryonic stages, and showed tentative minimum levels around P0, following which a transient increase forming peaks around P5 was observed. These activities decreased most likely due to the inhibition of activity but not a decrease in CDK proteins.
Binucleation and DNA content distribution patterns of cardiomyocytes during postnatal development
Many cardiomyocytes become binucleated during the postnatal stages (Soonpaa et al. 1996). We analyzed the DNA content distribution patterns during these stages in order to determine the cell cycle distribution patterns and how the cell cycle changes during transition from mononucleated cells to binucleated cells. Cells whose nuclear DNA contents are 2C, 2C–4C, and 4C are in the G1, S, and G2/M phases in the cell cycle, respectively. To measure nuclear DNA content, flow cytometry is generally used. In the cases of binucleated cells, isolation of nuclei from cells is necessary to measure DNA content per nucleus by using flow cytometry. However, we can not distinguish nuclei of binucleated cells from those of mononucleated cells after isolation. Moreover, it is difficult to measure the DNA content of cells in the M phase, especially from prometaphase to anaphase, because the chromosomes would be scattered after isolation from the cells. Therefore, we smeared dissociated cardiomyocytes (without isolation of nuclei) on slide glasses, stained the DNA in nuclei with DAPI and then measured the DNA content in each nucleus with microphotometry. Using this method, we were able to distinguish mono- and binucleated cardiomyocytes, measure the DNA content of the M phase cells and analyze the differences in the cell cycle distribution patterns between the two types of cardiomyocytes in all cell cycle phases. Similar methods have been used to measure DNA content for cell cycle analysis (Hitomi & Stacey 1999; Sa & Stacey 2004). We first compared the result from our method using microphotometry with that from flow cytometry using NIH3T3 cells to confirm the validity of our method (Fig. 5), and similar results were obtained with both methods.
Next, we examined postnatal cardiomyocytes (Fig. 6). We also examined EdU incorporation to detect DNA synthesis. Figure 6A represents dissociated cardiomyocytes and examples of DNA content at P0, P7 and P14. The morphology and size changed, and binucleation appeared to advance. The distribution patterns of cells with various DNA contents are shown in Figure 6B. Almost the same number of nuclei was investigated at each stage. Using the data, the percentages of nuclei with various levels of DNA content among all nuclei including both mono- and binucleated cells are shown in Figure 7A. The percentages of mono- and binucleated cardiomyocytes are also shown in Figure 7B. These data clearly show the transition from mono- to binucleated cardiomyocytes.
Almost all of the cardiomyocytes were mononucleated (98.7%; Fig. 7B) and the DNA content was mainly 2C (the G1 phase) at P0 (Fig. 7A). Nuclei with other DNA contents were very rare (Fig. 7A), indicating that the cell cycle activity was very low at P0. Subsequently, the percentages of nuclei with 2C in mononucleated cells decreased, while those with 4C (the G2/M phase) increased and showed a peak at P7. On the other hand, the percentages of nuclei with 2C in binucleated cells increased (Fig. 7A). The percentage of EdU positive cells peaked at P3 (Fig. 7C). No positive binucleated cells were observed.
These data suggest that more than 80% of the mononucleated cardiomyocytes progressed through their cell cycle and yielded binucleated cells. The cell cycle of the remaining mononucleated cells was arrested at the G1 phase. The majority of the binucleated cardiomyocytes most likely did not enter the cell cycle again, and their cell cycle was arrested at the G1 phase because nuclei with 2C–4C (the S phase) or 4C were rare and there were no cells positive for EdU.
Our analysis showed that the expression, complex formation, and activity of the main cyclins and CDKs changed in a synchronous manner during embryonic and postnatal stages (E12-adult, Figs 1-3). These levels decreased from E12 or E16 to P0, and then exhibited one wave in which the peak was around P5. Detailed analysis of the complexes suggested that CDK activities were inhibited before the protein levels decreased (Fig. 4). Analysis of the cell cycle distribution patterns and binucleation (Figs 6, 7) indicated that the wave mainly produced binucleated cells from mononucleated cells.
Even though the data presented here are basically consistent with previous data showing patterns of proliferation, binucleation, and expression of cyclin D1, D3 and CDK4 (Erokhina 1968; Cluzeaut & Maurer-Schultze 1986; Soonpaa et al. 1996), we performed more detailed analysis. For example, we analyzed the expression of almost all of the main cyclins and CDKs, their complexes, and the activities of cyclin E-CDK2, cyclin A-CDK1/2 and cyclin B1-CDK1. Moreover, the DNA content distribution patterns of both mono- and binucleated cardiomyocytes were examined at many stages.
We showed that proliferation activity reached the minimum point once around P0 (Figs 1-4, 6, 7). The proliferation pattern changes drastically between before and after birth. Before birth, the number of cardiomyocytes increases by cell cycle progression with cytokinesis, while after birth the number does not change markedly but more than 80% of the cells become binucleated by cell cycle progression without cytokinesis. P0 is most likely the changing point, and the downregulation of cell cycle progression might be necessary for transfer of the cell cycle system from the embryonic type to the postnatal type. In addition, the expression and activities of the main cyclins and CDKs are downregulated after P5, which causes cell cycle arrest. The mechanism, which regulates the downregulation, and the significance, are interesting. Our data suggest that CDK activities were inhibited (Fig. 4), however, the inhibitor is unknown. Genetic manipulation is a powerful method with which to investigate these issues, and the data presented here should provide an important basis for such studies.
The authors wish to thank Dr Takao Shioya of Saga University for kindly teaching us the method for the dissociation of cardiomyocytes. This work was partially supported by a research grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan.