The oxygen-rich postnatal environment induces cardiomyocyte cell-cycle arrest through DNA damage response

Abstract

The mammalian heart has a remarkable regenerative capacity for a short period of time after birth, after which the majority of cardiomyocytes permanently exit cell cycle. We sought to determine the primary postnatal event that results in cardiomyocyte cell-cycle arrest. We hypothesized that transition to the oxygen-rich postnatal environment is the upstream signal that results in cell-cycle arrest of cardiomyocytes. Here, we show that reactive oxygen species (ROS), oxidative DNA damage, and DNA damage response (DDR) markers significantly increase in the heart during the first postnatal week. Intriguingly, postnatal hypoxemia, ROS scavenging, or inhibition of DDR all prolong the postnatal proliferative window of cardiomyocytes, whereas hyperoxemia and ROS generators shorten it. These findings uncover a protective mechanism that mediates cardiomyocyte cell-cycle arrest in exchange for utilization of oxygen-dependent aerobic metabolism. Reduction of mitochondrial-dependent oxidative stress should be an important component of cardiomyocyte proliferation-based therapeutic approaches.

Figure 1

Oxidation state, activity of mitochondrial respiration, oxidative stress and the activation of DNA damage response (DDR) correspond to cardiac regenerative capacity. (A) Fishes and mammalian fetuses are under low-oxygenated environment, whereas postnatal mammals are in well-oxygenated atmosphere. (B) qPCR analysis revealed post-natal increase in mitochondrial DNA (mtDNA) contents per gram of tissue (ventricles) until postnatal day 14 (P14). Relative mtDNA content in adult zebrafish was even smaller than that in P1 mouse. (C) TEM images of ventricles showed more mature cristae structure in P7 mouse heart comparing with P1 mouse heart and adult zebrafish heart (left). The number of mitochondrial cristae counted from SEM images increased in P7 mouse heart compared to P1 mouse heart (table, blue bars) and also to adult zebrafish heart (table, red bar). (D) HPLC detection of a superoxide probe dihydroethidium (DHE) revealed a significant increase in both 2-hydroxyethidium (EOH), a specific product for superoxide anion radical, and in ethidium (E), oxidized by other reactive oxygen species such as H₂O₂ (mainly) and ONOO from P1 to P7. (E) Imaging of ROS on cryosections with dihydrorhodamine 123 staining indicated linear increase in cardiomyocyte ROS level from P1 to P7 (arrows). (F) Immunostaining with oxidative DNA damage and DDR markers. A marker for oxidative base modification in DNA, 8-oxo-7,8-dihydroguanine (8-oxoG, left panels), and for activation of DDR, Ser1987 phosphorylated ATM (pATM, right panels) were not detected in cardiomyocyte nuclei at P1 (top panels, white arrows), whereas at P7 (middle panels) and at P14 (lower panels) both 8-oxoG and pATM showed nuclear localization (co-localized with Hoechst 33258, Ho) indicated by arrows. (G) Oxidative DNA damage and DDR were induced before P7 in cardiomyocyte. Left panels show representative images of 3D imaging of 8-oxoG staining and the number of 8-oxoG foci per cardiomyocyte. Right panels show western blot and quantification of pATM indicating upregulation of DDR. Cardiac troponin T (TnT) was used to normalize sample loading. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 2

Postnatal cardiomyocyte cell cycle arrest is dependent on environmental oxygen concentration. (A) Neonates were exposed to hyperoxic (100% O₂) or mildly hypoxic (15% O₂) environment from perinatal stage for 4 days (hyperoxia) or 7 days (hypoxia). (B) Oxidative DNA damage indicated by nuclear signal with anti-8-oxoG antibody in cardiomyocyte was increased in hyperoxic hearts (left panels), and decreased in hypoxic hearts (right panels). (C) Heart weight versus body weight (HW/BW) ratio showed no statistically significant difference in mice exposed to hyperoxia, whereas significantly increased in mice exposed to hypoxia. (D) Cardiomyocyte cell size did not show significant difference in hyperoxia, and was decreased in hypoxia. (E) Co-immunostaining with anti-phospho-histone H3 Ser10 (pH3) and anti-cardiac Troponin T (TnT) antibodies showed drastic decrease in cardiomyocyte mitosis in hyperoxia-exposed hearts, and in contrast, significant increase in cardiomyocyte mitosis in hypoxia-exposed hearts. Cardiomyocyte apoptosis did not increase in neither treatments as shown by TUNEL staining. (F) Co-immunostaining with anti-Aurora B and anti-sarcomeric actinin antibodies showed decreased cytokinesis in hyperoxic hearts, whereas mild increase in hypoxic hearts. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 3

Injection of ROS generator diquat induced accelerated cardiomyocyte cell cycle arrest. (A) 5mg/kg of diquat was injected subcutaneously for 3 days after birth and hearts were harvested at P3. (B) Immunostaining showed nuclear accumulation of 8-oxoG in cardiomyocytes in diquat injected neonates. (C) HW/BW ratio has no statistically significant difference. (D) WGA staining showed significantly increased cardiomyocyte cell size in hearts of diquat injected neonates. (E) Co-immunostaining with anti-pH3 and anti-TnT antibodies showed drastic decrease in cardiomyocyte mitosis in diquat injected hearts. TUNEL assay showed no significant increase in cardiomyocyte apoptosis. (F) Diquat injection resulted in decreased cardiomyocyte cytokinesis as shown by co-immunostaining with anti-Aurora B, anti-TnT antibodies. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 4

Injection of a scavenger against ROS suppresses post-natal cardiomyocyte cell cycle arrest. (A) N-acetyl-cysteine (NAC) was administered for 21 days after birth. (B) ROS level in cardiomyocytes significantly decreased as shown by dihydrorhodamine 123 staining in NAC treated neonates. (C) Reduced 8-oxoG staining was seen in NAC injected heart, demonstrating reduced oxidative DNA damage in cardiomyocyte at P7. (D) NAC injection suppressed activation of DNA damage response pathway. Co-immunostaining with anti-pATM and anti-alpha actinin antibodies showed that NAC treatment reduces nuclear phospho-ATM at P7 compared with control (arrows in left panels). (E) HW/BW ratio in control and NAC treated hearts at P14 showed no significant difference. (F) Cell size quantification using WGA staining showed significantly decreased cell size in NAC treated hearts. (G) Total number of cardiomyocyte was significantly increased in NAC treated heart at P21. (H) Co-immunostaining with anti-pH3 and anti-TnT antibodies showed increased cardiomyocyte mitosis in NAC treated hearts. Images shown are hearts from PBS-injected control or NAC-injected neonates at P14. (I) NAC injection induced cardiomyocyte cytokinesis indicated by Aurora B and TnT double positive cardiomyocytes. (J) Percentage of binucleated cardiomyocytes was significantly decreased and mononucleation was siginificantly increased in NAC treated hearts at P14. Immunocytochemistry on isolated myocyte with anti-Cx43 antibody and Hoechst 33258 (Ho) nuclear specific dye were used to visualize boundary and nuclei of each cardiomyocyte, respectively. (K) Apoptotic cell death visualized with TUNEL assay was not increased in NAC-treated heart at neither P7 nor P14. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 5

Scavenging reactive oxygen species in mitochondria suppresses post-natal cardiomyocyte cell cycle arrest. (A) Mitochondrial-targeted catalase (mCAT) was over-expressed in cardiomyocyte by flipping floxed-stop cassette between GAPDH promoter and mCAT by crossing with mice harboring Myh6 promoter driven Cre recombinase. Myh6-Cre; floxed-stop mCAT mice are hereinafter called mCAT mice. Litter-mate Myh6-Cre mice were used as control. (B) The level of ROS in cardiomyocyte decreased in mCAT heart compared with control. (C) mCAT significantly suppressed DNA damage shown by 3D imaging and the quantification of 8-oxoG foci in cardiomyocyte (D) HW/BW ratio in control and mCAT mouse hearts at P14 and in the adult was not changed. (F) WGA staining indicated slightly decreased cardiomyocyte cell size in mCAT mouse heart, but not statistically significant level. (G) Increased cardiomyocyte mitosis in mCAT mouse hearts at p14 indicated by co-immunostaining with anti-pH3 and anti-troponin T antibodies. (H) Increase in cardiomyocyte cytokinesis shown by immunostaining with anti-Aurora B antibody. (I) Total number of cardiomyocyte was significantly increased at P14 in mCAT mouse heart. (J) Number of binucleated cardiomyocytes was significantly decreased and monucleated cardiomyocytes was significantly increased in mCAT hearts (K) TUNEL positive apoptotic cell death was within normal range in mCAT heart. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 6

Scavenging ROS extends postnatal cardiac regeneration window. (A) Animals were treated with NAC from P0 and subjected to ischemia reperfusion (IR) surgery at P21. Viability was investigated 24 hours after injury to evaluate the consistency of the IR surgery, and then cardiac function was examined by echocardiography at 1 week and 1 month after injury. (B) TTC staining showed no significant difference in the size of injury between control and NAC-pretreated hearts. (C) HW/BW ratio at 1 week or 1 month after IR surgery between control and NAC-pretreated mice was not changed. (D) WGA staining showed a decrease in cell size in NAC treated hearts 1 month after IR injury compared to control hearts, indicating the NAC treatment did not induce cardiomyocyte hypertrophy. (E) Left ventricular systolic function quantified by ejection fraction (EF) and shortening fraction (SF) (n=6 per group) before IR surgery (left), 1 week after surgery (middle) and 1 month after surgery (right) showed more functional recovery in NAC-pretreated hearts compared with control hearts. (F) Histological analysis with Masson trichrome staining showed reduced fibrotic scar formation in NAC treated hearts. (G) Cardiomyocyte mitosis was increased in NAC-treated mice as indicated by anti-pH3 and anti-TnT co-immunostaining at 3 days after IR injury. (H) Immunostaining with anti-BrdU and anti-Actinin antibodies showed increased cardiomyocyte BrdU incorporation NAC-pretreated heart. Error bars indicate SEM. *p < 0.05; **p < 0.01.

Figure 7

Inhibition of an effector of DNA damage response, Wee1 kinase, suppresses postnatal cardiomyocyte cell cycle arrest. (A) DNA damage response pathway activates Wee1 kinase that inhibits cell cycle regulator Cyclin/CDKs, resulting in cell cycle arrest. (B) Immunostaining with anti-Wee1 and anti-TnT antibodies showed no expression of Wee1 protein in cardiomyocytes at P1 (arrows), and then nuclear localized Wee1 protein at P7 and P14 (arrows). (C) Daily injection of Wee1 kinase inhibitor MK-1775 was performed from birth until P14. (D) Wee1 inhibitor injection resulted in slight, but not statistically significant, decrease in HW/BW ratio compared with control. (E) Inhibition of Wee1 induced mitosis in cardiomyocyte at P14 shown by pH3/TnT immunostaining. (F) Quantification of cardiomyocyte cell size with WGA staining showed decreased cardiomyocyte size in Wee1 inhibitor treated hearts. (G) AuroraB/TnT immunostaining showed increased cardiomyocyte cytokinesis in Wee1 inhibitor treated hearts. (H) No increase in apoptotic cell death in cardiomyocyte indicated by TUNEL assay. (I) Total number of cardiomyocyte was increased at P7 and P14 in Wee1 inhibitor treated hearts. (J) Number of binucleated cardiomyocytes was reduced, and mononucleated cardiomyocyte was increased in Wee1 inhibitor treated heart. Error bars indicate SEM. *p < 0.05; **p < 0.01.