Timing of cardiomyocyte growth, maturation, and attrition in perinatal sheep

Abstract

Studies in altricial rodents attribute dramatic changes in perinatal cardiomyocyte growth, maturation, and attrition to stimuli associated with birth. Our purpose was to determine whether birth is a critical trigger controlling perinatal cardiomyocyte growth, maturation and attrition in a precocial large mammal, sheep (Ovis aries). Hearts from 0-61 d postnatal lambs were dissected or enzymatically dissociated. Cardiomyocytes were measured by micromorphometry, cell cycle activity assessed by immunohistochemistry, and nuclear number counted after DNA staining. Integration of this new data with published fetal data from our laboratory demonstrate that a newly appreciated >30% decrease in myocyte number occurred in the last 10 d of gestation (P < 0.0005) concomitant with an increase in cleaved poly (ADP-ribose) polymerase 1 (P < 0.05), indicative of apoptosis. Bisegmental linear regressions show that most changes in myocyte growth kinetics occur before birth (median = 15.2 d; P < 0.05). Right ventricular but not left ventricular cell number increases in the neonate, by 68% between birth and 60 d postnatal (P = 0.028). We conclude that in sheep few developmental changes in cardiomyocytes result from birth, excepting the different postnatal degrees of free wall hypertrophy between the ventricles. Furthermore, myocyte number is reduced in both ventricles immediately before term, but proliferation increases myocyte number in the neonatal right ventricle.

Keywords: apoptosis; cell number; fetus; neonate; proliferation.

Figure 1.

A) The increase in heart and body weight slows somewhat near birth (n = 114). B) LV and RV free wall weights are similar in fetal sheep. After birth both free walls continue to increase in weight, but the LV does so more rapidly than the RV (combined n = 39; P = 0.0075). C) The ratio of each free wall to total heart weight diverged rapidly at birth (n = 39). Regression (solid) lines and 95% CI (dashed) best fit to each data set are shown; bars show the 95% CI of the breakpoint. Fetal data from Jonker et al. (3). Regression data are included in the text and Tables 1 and 2.

Figure 2.

A, B) Cardiomyocyte length was best described by bisegmental regression, with increased growth occurring in the neonatal segment (n = 79 fetuses and neonates were studied, of which n = 75 LV and n = 74 RV had binucleated myocytes, and n = 15 had quadrinucleated myocytes in sufficient quantities to measure; at least 100 myocytes were measured per ventricle of each animal). C, D) Cardiomyocyte width was best described by bisegmental regression only in the LV; linear regression best describes myocyte width in the RV. Shown are the best-fit linear regression lines (solid) of each data set and 95% CI (dashed); bars show 95% CI of the breakpoints. Fetal data from Jonker et al. (3). Regression data are included in Tables 1 and 2. bi, binucleate; mono, mononucleate.

Figure 3.

LV (A) and RV (B) cell volumes were calculated and were best fit by bisegmental regression that showed increased growth in the neonatal segment (n = 79 fetuses and neonates were studied, of which only n = 75 LV and n = 74 RV had binucleated myocytes. LV and RV quadrinucleated myocytes were only measured in n = 15 of the neonates. At least 100 myocytes were measured per ventricle of each animal). C) The cardiomyocytes cell volume per nucleus increased rapidly in the postnatal segment. D) The slope of the relationship of binucleated and mononucleated myocyte volume between ventricles is not significantly different from 2 for the LV (n = 75) or RV (n = 74), indicating that cell volume is similarly regulated as a function of nuclei number in mononucleated and binucleated myocytes. Fetal data from Jonker et al. (3). Regression (solid) lines and 95% CI (dashed) best fit to each data set are shown; bars show 95% CI of the breakpoints. Regression data are included in Tables 1 and 2. bi, binucleate; mono, mononucleate.

Figure 4.

In the LV (A) and RV (B), the percent of cardiomyocytes with 1 nucleus declines rapidly before birth (n = 79 animals, at least 300 myocytes per ventricle of each animal were counted). Most of these myocytes become binucleated. A small but growing proportion of neonatal cardiomyocytes have 4 or more nuclei. In the LV (C) and RV (D), the proportion of myocytes in the cell cycle is low at birth (n = 79 animals, at least 500 myocytes per ventricle of each animal were counted). Insets: Cell cycle activity normalized to the proportion of myocytes that are mononucleated. Fetal data from Jonker et al. (3). Regression lines (solid) and 95% CI (dashed; best fit to each data set) are shown; bars show quadrinucleate 95% CI of the breakpoints. Regression data are included Tables 1, 2, and 3. mono, mononucleate.

Figure 5.

LV cardiomyocytes from sheep 92 dPC (A), 1 d after birth (B), and 45 d after birth (C). Cells imaged for myosin (red) and DNA (cyan). Scale bar, 50 μm (applies to all panels).

Figure 6.

LV (A) and RV (B) total free wall myocyte numbers in the fetus and neonate. Bisegmental linear regressions were fitted to the previously published fetal data (n = 63 fetuses) from Jonker et al. (3). Separately, linear regressions were fitted to the neonatal data (n = 16 neonates). An increase in RV but not the LV cell number was found during the neonatal period (P = 0.028). Regression lines (solid) and 95% CI (dashed; best fit to each data set) are shown; bars show 95% CI of the breakpoints. Regression data are included in Tables 1 and 2. LV (C) and RV (D) myocyte numbers were compared at 10 d intervals before term and after birth. Comparison of adjacent ages (mean 134 ± 1 dPC, n = 6; 140 ± 1 dPC, n = 13; 4 ± 2 d postnatal, n = 5) determined that total myocyte number decreased immediately prior to birth (by 34% in the LV and 38% in the RV), but not at birth. sd of total number is shown. Quadrinucleate myocytes contribute less than the thickness of the line to any group. E) LV and (F) RV abundance of cleaved to whole PARP-1 was compared in early gestation and at 10 d intervals before term and after birth (mean 94 ± 1 dPC, n = 6; 133 ± 3 dPC, n = 7; 139 ± 2 dPC, n = 7; 4 ± 3 d postnatal, n = 7). Error bars show standard error of the mean. G) A representative PARP-1 Western blot. Blot image is inverted. Ladder overlay is contrast-enhanced white light image.