The Influence of Naturally Occurring Differences in Birthweight on Ventricular Cardiomyocyte Number in Sheep.код для вставкиСкачать
THE ANATOMICAL RECORD 292:29–37 (2009) The Influence of Naturally Occurring Differences in Birthweight on Ventricular Cardiomyocyte Number in Sheep VICTORIA STACY, ROBERT DE MATTEO, NADINE BREW, FOULA SOZO, MEGAN E. PROBYN, RICHARD HARDING,* AND M. JANE BLACK Department of Anatomy and Developmental Biology, Monash University, Victoria, Australia ABSTRACT In most species including man, cardiomyocytes cease proliferating soon after birth when they become terminally differentiated. A reduced complement of cardiomyocytes in infancy may adversely impact on the function and adaptive capabilities of the heart in later life. Low birthweight is associated with an increased risk of heart disease in adults, but little is known about its effect on the number of cardiomyocytes. Using naturally occurring differences in birthweight, our aim was to determine the effect of birthweight on cardiomyocyte number in postnatal lambs. At 9 weeks after term birth, when the final number of cardiomyocytes is considered to be established, hearts were collected at necropsy from seven singleton and seven twin lambs. Hearts were perfusion-fixed, and tissue samples were systematically taken from the left ventricle plus intraventricular septum (LV1S) and the right ventricle (RV). The number of cardiomyocyte nuclei was estimated using an unbiased optical disector–fractionator stereological technique, and the total number of cardiomyocytes was determined. Weights of the total heart, LV1S and RV were significantly related to both birthweight and necropsy weight. In the LV1S but not the RV, cardiomyocyte number was significantly and directly related to heart tissue weight, birthweight, and necropsy weight. We conclude that the final number of cardiomyocytes in the LV1S is related to prenatal and early postnatal growth, and is proportionate to the weight of heart tissue. A low cardiomyocyte number in the LV1S following restricted fetal growth may contribute to the increased incidence of heart disease in adults born with low birthweight. Anat Rec, 292:29–37, 2009. Ó 2008 Wiley-Liss, Inc. Key words: fetus; newborn; heart; cardiomyocytes; growth Epidemiological studies have shown that low birthweight due to intrauterine growth restriction (IUGR) is associated with an increased risk of cardiovascular disease in adult life (Barker et al., 1993; Zhao et al., 2002; Lawlor et al., 2004). Previous experimental studies have shown that IUGR alters the development and function of major components of the cardiovascular system such as the kidneys (Manalich et al., 2000), vasculature (Goodfellow et al., 1998; Martin et al., 2000), and the activity of the sympathetic nervous system (Phillips and Barker, 1997). However, the effects of IUGR on the development of the heart have not been well defined, with Ó 2008 WILEY-LISS, INC. Grant sponsor: National Health and Medical Research Council of Australia. *Correspondence to: Richard Harding, Ph.D., D.Sc., Department of Anatomy and Developmental Biology, PO Box 13C, Monash University, VIC 3800, Australia. Fax: 1613-9905-2330. E-mail: email@example.com Received 7 February 2008; Accepted 7 July 2008 DOI 10.1002/ar.20789 Published online 24 October 2008 in Wiley InterScience (www. interscience.wiley.com). 30 STACY ET AL. little being known about the effects of IUGR on the generation of cardiomyocytes. The proliferation of cardiomyocytes is thought to cease soon after birth (Oparil et al., 1984; Li et al., 1996), after which postnatal growth of the heart muscle is predominantly due to cardiomyocyte hypertrophy and deposition of extracellular matrix. If cardiomyocyte number is reduced by IUGR, this could have important implications for cardiovascular health later in life. A reduced number of cardiomyocytes could contribute to cardiac dysfunction and susceptibility to heart disease, especially after catchup growth in body size, which is common in individuals subjected to restricted fetal growth (Eriksson and Forsen, 2002); catch-up in body growth (and even more so with obesity) would result in a reduced number of cardiomyocytes per kilogram of adult body weight. A recent study in our laboratory, using an experimental model of late gestational placental insufficiency to induce IUGR, has shown that the hearts of IUGR fetuses were smaller and structurally and functionally less mature than in non-IUGR fetuses (Bubb et al., 2007). Similar findings of cardiac muscle immaturity at birth following IUGR induced by placental restriction in sheep have recently been reported (Morrison et al., 2007). However, in neither of these studies was cardiomyocyte number determined. A study of neonatal rat pups born to mothers subjected to reduced protein intake during pregnancy has shown that there were fewer cardiomyocytes in the hearts of IUGR pups when compared with normally grown pups (Corstius et al., 2005). The growth-restricted pups also had significantly lighter hearts, a finding made in another study of rat pups, whose mothers were subjected to a restriction in global food intake (Desai et al., 2005). To date, little is known about the effects of differences in birthweight within the normal range on heart development. In particular, it is unknown whether individuals who have low birthweight due to mild, naturally occurring prenatal growth restriction, such as that induced by twinning, have altered development of heart muscle. In the present study, we have used sheep as, in this species, the timing of cardiomyocyte differentiation during late gestation and early postnatal life is similar to that in humans (Oparil et al., 1984; Burrell et al., 2003). Our aim was to determine the effects of naturally occurring differences in birthweight, as an index of fetal nutrition and growth, on the number of cardiomyocytes in the hearts of postnatal lambs; our hypothesis was that low birthweight would be associated with a decreased number of cardiomyocytes. In order to increase the range of birthweights, we studied the hearts of singleton and twin lambs. We analyzed the heart muscle at 9 weeks after birth, as at this age, ovine cardiomyocytes are thought to be terminally differentiated and binuclear (Burrell et al., 2003). As postnatal nutrition, growth, and arterial pressure could also influence cardiomyocyte differentiation, we have measured postnatal growth rates and arterial pressure in these animals. We have also compared data between males and females, as gender differences in cardiac development may exist. MATERIALS AND METHODS All experimental procedures were approved by the Monash University Animal Welfare Committee. In order to measure cardiomyocyte numbers, we used 14 lambs from 14 date-mated Border Leicester 3 Merino ewes; seven carried singletons (three males, four females), and seven carried twins. Only one lamb (randomly chosen) from each twin pair was studied (two males, five females). The ewes delivered their lambs vaginally at term. After birth, lambs were raised with their mothers and were weighed daily; at 5–6 weeks after birth, they were separated from their mothers and provided with lucerne chaff and free access to water. At 3 weeks of age, the lambs underwent surgical implantation of a catheter into the femoral artery for later measurement of arterial pressure and heart rate. At 58 6 1 days after birth, the lambs were placed prone in a sling, while awake, for recording of arterial pressure for 2 hours on 2 consecutive days. Necropsy At 64 6 1 days after birth, the lambs were humanely killed by an overdose of sodium pentobarbital (325 mg/ mL i.v.), and body and organ weights recorded. Hearts were removed intact and weighed. A catheter was placed in the aortic arch and heparin sodium (to prevent clotting), papaverine hydrochloride (to maximally dilate the vasculature), and potassium chloride (to relax the cardiomyocytes) were administered via the catheter. Hearts were perfused via the aortic arch with saline to clear the coronary vasculature of blood and then perfusion-fixed with 4% formaldehyde. The fixed hearts were stored in 10% buffered formalin. Sampling of Ventricular Muscle The fixed hearts were trimmed of connective tissue and weighed; the atria were removed and both ventricles weighed. The ventricles were cut into a series of transverse ( 4 mm) slices. A central slice was used for the measurement of ventricular wall thickness (see later). In each slice, the right ventricle (RV) was separated from the left ventricle plus interventricular septum (LV1S). The LV1S were kept together and treated as a unit as the septum resembles the LV both functionally and structurally. Each slice of the LV1S was cut into four pieces, and the pieces were laid out from largest to smallest. Using a smooth-fractionator approach, five to seven pieces were sampled; these were further sectioned using a slicing device and again, using a smooth fractionator approach (Gundersen, 2002), eight to 10 pieces of LV1S were sampled. The sampled pieces of LV1S were placed in 70% ethanol in preparation for tissue processing. A similar sampling procedure was repeated for the RV, resulting in the generation of eight to 10 pieces of RV. Measurement of Ventricular Wall Thickness A transverse slice of the ventricles, as described earlier, from the mid region of the heart was used to measure ventricular wall thickness. A digital image was taken of the slice and thicknesses of the LV wall, and septum and RV wall were measured (Image ProPlus Version 6.0 for WindowsTM, Media Cybernetics, Silver Spring). FETAL GROWTH AND CARDIOMYOCYTE NUMBER Tissue Processing The sampled LV1S and RV tissue was embedded in glycol-methacrylate (Technovit 7100 resin, Heraeus Kulzer, Germany), and the blocks were exhaustively sectioned at 20 mm. Every thirtieth section for LV1S and every twentieth section for RV (commencing at a randomly selected number between 1 and 10) were collected and mounted onto a glass slide and stained with hematoxylin in a 1,000-Watt microwave oven (50% power for 2.5 min). Estimation of the Number of Cardiomyoctye Nuclei An optical disector–fractionator approach was used to estimate the total number of cardiomyocyte nuclei in the LV1S and the RV (Corstius et al., 2005). Counting was performed using a computer-assisted stereologic package (CAST 2002; Olympus, Albertsland, Denmark). An unbiased counting frame was superimposed over the sections using a 1003 objective lens. A systematic uniform random sample of fields was obtained on a 2,000 3 2,000 mm grid for the LV1S and a 1,500 3 1,500 mm grid for the RV; nuclei were counted in a ‘‘disector’’ (area of 329.6 mm2) within a depth of 10 mm. To allow for inconsistencies in cut surfaces and loss of material, the upper and lower 5 mm of the section were used as a ‘‘guard" area. Nuclei were counted when they came into clear focus within the volume of the disector; however, if any part of a nucleus touched the ‘‘forbidden lines’’ of the grid that nucleus was not counted. The cardiomyocyte nuclei were easily distinguishable from other cells (e.g., endothelial cells and fibroblasts) by their elongated oval shape, light purple staining, and visible chromatin and prominent nucleoli. The total number of cardiomyocytes in the individual ventricles was determined by the following equation: Ncardiomyocytes5 Q2 3 1/f1 3 1/f2 3 1/f3 3 1/f4 3 1/f5, where Q2 is the number of cardiomyocyte nuclei counted, f1 and f2 are the initial fixed tissue sampling fractions, f3 is the third level of sampling where every thirtieth section was chosen from a random start, f4 is the fraction of the section area actually counted (LV: 329.6/(2,000 3 2,000); RV: 329.6/(1,500 3 1,500), and f5 is the fraction of the depth of the section (10/20) counted. The total number of cardiomyocytes was subsequently determined by adjusting for the proportion of mononucleates/binucleates (see later). Proportion of Binuclearity To determine the proportion of binucleated cells to mononucleated cells, we used two additional singleton lambs. At 9 weeks of age, these lambs were humanely killed by an overdose of sodium pentobarbital (325 mg/ mL i.v). The freshly dissected hearts were attached to a Langendorff apparatus via an intra-aortic cannula and perfused with 0-Ca21 physiological saline at 358C for 8 min to clear remaining blood. To digest the ventricles, saline-containing collagenase (Worthington CL-2, 200 U/ mL21) was infused for 25 min. To relax the cardiomyocytes, hearts were then perfused for 20 min with Hepesbuffered high K1 solution: 117 mM KCl, 36 mM NaCl, 1 mM MgSO4, 60 mM Hepes, 8 mM ATP-Na, 50 mM 31 EGTA; pH adjusted to 7.0 with KOH (Burrell et al., 2003). The LVs and RVs were separated and placed in Hepes-buffered high K1 solution and gently agitated for 5 min. The isolated cells were filtered through gauze, and cardiomyocytes from the LV and RV were separately fixed in 4% paraformaldehyde in phosphate-buffered saline (pH 5 7.4) (Bubb et al., 2007). Samples of fixed LV and RV cardiomyocytes were smeared onto slides and stained with hematoxylin and eosin. Cardiomyocytes were analyzed under a light microscope (4003 magnification); all cells that fell entirely within a field of view were analyzed. Six hundred cells were counted, and the proportions of mononucleated and binucleated cells determined (Bubb et al., 2007). Cardiomyocyte Proliferation To confirm that cardiomyocytes had ceased proliferating at 9 weeks of age, we performed immunohistochemistry on paraffin-embedded tissue sections from the LVs and RVs using an antibody against the Ki67 protein (1:100, mouse anti-human monoclonal antibody; DakoCytomation, Glostrup, Denmark). The Ki67 protein is expressed during all active phases of the cell cycle but is absent in resting cells and is therefore used as a marker of cell proliferation (Gerdes et al., 1984). Tissue sections were boiled for 20 min in a microwave oven (on high power) in an antigen retrieval solution (0.01 M sodium citrate, pH 6.0) and an EnVision 1 Dual Link SystemHRP (DAB1) immunohistochemistry kit (DakoCytomation, CA) was used to visualize positively stained nuclei. Specificity of immunostaining for Ki67 was confirmed by omission of the primary antibody. A section from the heart of a normal sheep fetus at 133 days of gestation ( 14 days before term) was included as a positive control, as many cardiomyocytes are not terminally differentiated at this age. Sections were counterstained with hematoxylin, and the proportion of cardiomyocytes undergoing cell proliferation was qualitatively analyzed using a light microscope. Analysis of Collagen Content of Cardiac Muscle To assess the collagen content of the cardiac muscle, we stained 5-mm sections with Masson’s Trichrome stain. Using image analysis software (Image ProPlus; Version 6.0 for Windows), we analyzed eight sections per ventricle and nine fields per section. In each field, we measured the proportion of the area occupied by collagen (stained blue) in the LVs and RVs. Data Analysis Data are presented as mean 6 SEM. Data were analyzed using a two-way ANOVA, using male vs. female and singleton vs. twin as factors. Associations between cardiomyocyte number and birthweight, necropsy weight, and heart weight were tested using linear regression analysis. Differences between data from the LV1S and RV (cardiomyocyte number and density) were analyzed by paired t test. Significance was taken at P < 0.05. 32 STACY ET AL. TABLE 1. Body and heart weights in 9-week-old lambs: data are presented comparing singletons and twins, comparing males and females and for all animals combined Single (n 5 7) Birth weight (kg) Autopsy weight (kg) Heart weight (g) Heart weight/autopsy weight (g/kg) LV1S weight (g) LV1S/autopsy weight (g/kg) RV weight (g) RV/autopsy weight (g/kg) LV1S wall thickness (mm) RV wall thickness (mm) 4.79 18.89 83.1 4.4 52.3 2.8 16.9 0.9 10.5 5.4 6 6 6 6 6 6 6 6 6 6 0.23 0.88 4.7 0.3 2.7 0.1 1.1 0.1 0.4 0.3 Twin (n 5 7) Male (n 5 5) 3.67 13.58 57.0 4.2 35.6 2.6 11.3 0.8 8.7 4.1 4.06 16.80 71.8 4.3 44.9 2.7 14.3 0.8 9.2 4.7 6 6 6 6 6 6 6 6 6 6 0.31* 1.10* 4.4* 0.2 2.4* 0.1 0.9* 0.1 0.2* 0.3* 6 6 6 6 6 6 6 6 6 6 0.57 2.15 10.2 0.3 6.4 0.8 2.3 0.1 0.3 0.4 Female (n 5 9) 4.32 15.92 69.1 4.4 43.4 2.7 14.0 0.9 9.8 4.7 6 6 6 6 6 6 6 6 6 6 0.26 1.21 5.8 0.2 3.5 0.1 1.2 0.1 0.4 0.4 All (n 5 14) 4.23 16.24 70.0 4.3 43.9 2.7 14.1 0.9 9.6 4.7 6 0.24 6 1.00 6 4.8 6 0.1 6 2.9# 6 0.1# 6 1.0 6 0.03 6 0.3# 60.3 Abbreviations: LV1S, left ventricle plus septum; RV, right ventricle. Values shown are the means 6 SEM. *Indicates twins significantly different from singles (P < 0.05). Indicates significant difference between LV1S and RV (P < 0.05). # RESULTS Fetal and Postnatal Growth All lambs were born at term (146 6 0 days, range, 144–148 days). The mean birthweight was 4.23 6 0.24 kg (Table 1), with individual weights ranging from 2.8– 5.6 kg. As a group, twins were 20% lighter at birth than singletons (3.67 6 0.31 kg vs. 4.79 6 0.23 kg; P < 0.01); there was no difference between males and females. Growth rates after birth are shown in Fig. 1. On average, body weights increased 3.8-fold between birth and 9 weeks. Twins had a significantly lower absolute growth rate (g/day) during the first 9 weeks after birth compared with singletons (153.6 6 15.8 g/day vs. 227.4 6 8.1 g/day, P < 0.01); however, there was no significant difference between the weekly percentage weight gain in the two groups. There was no difference in postnatal growth rates between males and females (Fig. 1). During the experimental period of birth to 9 weeks of age, there was no evidence of catch-up growth in the lower birthweight lambs. The absolute weights of both the LV1S (twins, 35.6 6 2.4 g vs. singletons, 52.3 6 2.7 g; P < 0.001) and RV (twins 11.3 6 0.9 g vs. singletons 16.9 6 1.1 g; P < 0.001) were significantly lower in twins than singletons; when these values were adjusted for body weight, there was no difference between groups. There were no differences in weights of the LV1S or RV between males and females (Table 1). Ventricular Wall Thickness The walls of LV1S were approximately twice as thick as those of the RV (Table 1). Compared with singletons, twins had significantly thinner LV (twins, 8.7 6 0.2 mm vs. singletons, 10.5 6 0.4 mm; P < 0.001) and RV (twins, 4.1 6 0.3 mm vs. singletons, 5.4 6 0.3mm; P < 0.001) walls. Ventricular wall thicknesses (both LV1S and RV) were significantly and positively correlated with birthweight in all lambs (LV1S, r2 5 0.337, P 5 0.037; RV, r2 5 0.470, P 5 0.010). There were no differences in ventricular wall thicknesses between males and females (Table 1). Arterial Pressure At 8 weeks, the mean arterial pressure in all lambs averaged 75 6 4 mmHg, the systolic pressure was 96 6 4 mmHg, and the diastolic pressure was 65 6 5 mmHg. There were no significant differences between twins and singletons or between males and females in mean arterial pressure, systolic pressure, or diastolic pressure. There were no significant correlations between arterial pressure (mean, diastolic or systolic) and birthweight or body weight at 8 weeks. Heart Weights At 9 weeks after birth, the mean heart weight in all lambs was 70.0 6 4.8 g (4.3 6 0.1 g/kg body weight). In twins, the heart weight was 32% lower than in singletons (57.0 6 4.4 g vs. 83.1 6 4.7 g; P < 0.001), but when adjusted for body weight there was no significant difference between groups (Table 1). There was no difference between males and females in either absolute or relative heart weights. Among all animals, the weight of LV1S was 62.7% of the total heart weight, and the RV was 20.1%. In relation to body weight (all animals combined), the weight of the RV was significantly lower than that of the LV1S (Table 1). Number of Cardiomyocytes in the LV1S In the LV1S, the mean number of cardiomyocytes in the 14 lambs was 31.8 3 108 6 2.61 3 108; when adjusted for tissue weight, this was 7.31 3 107 6 0.48 3 107 cardiomyocytes/g (Table 2). The absolute number of cardiomyocytes in LV1S was significantly correlated with birthweight (r2 5 0.578, P 5 0.002), body weight at necropsy (r2 5 0.414, P 5 0.013), and heart weight (r2 5 0.511, P 5 0.004) (Fig. 2). There were significantly more cardiomyocytes in the LV1S in singletons than in twins, but their density was not different (Table 2). In the LV1S, there was a trend for males to have fewer cardiomyocytes than females (males, 27.4 3 108 6 3.81 3 108 cells vs. females, 34.3 3 108 6 3.48 3 108 cells; P 5 0.086); however, their density was not different (Table 2). Number of Cardiomyocytes in the RV In the RV, the mean number of cardiomyocytes for all animals was 8.13 3 108 6 0.50 3 108; the number of cardiomyocytes per gram of RV tissue was 6.03 3 107 6 0.45 3 107 (Table 2). The number of cardiomyocytes in the RV was not significantly correlated with birthweight (r2 5 0.113, P 5 0.239) or necropsy weight (r2 5 0.058, FETAL GROWTH AND CARDIOMYOCYTE NUMBER 33 Fig. 1. Postnatal growth rates between birth and necropsy at 9 weeks. The upper panels compare singleton lambs (filled circles, n 5 7) and twin lambs (open circles, n 5 7): (A) shows the increase in body weight and (B) shows the mean percent weight gain per week. The lower panels show the data for males (filled triangles, n 5 5) and females (open triangles, n 5 9): (C) shows the increase in body weight and (D) shows the mean percent weight gain per week. Data points show the mean 6 SEM. *indicates that twins are significantly different from singletons (P < 0.05). P 5 0.406), but there was a trend for a correlation with heart weight (r2 5 0.271, P 5 0.056). There was no difference in cardiomyocyte number between twins and singletons; however, there was a trend for a higher density of cardiomyocytes in the twins compared with the singletons (P 5 0.055). In the RV, there was no difference for either cardiomyocyte number or density between males and females. The total number of cardiomyocytes was significantly lower in the RV than the LV1S (P < 0.001), and there was a trend for a lower density of cells in the RV compared with the LV1S (P 5 0.062). 34 STACY ET AL. TABLE 2. Cardiomyocyte number and density in 9-week-old lambs: data are presented comparing singletons and twins, comparing males and females and for all animals combined Single (n 5 7) 8 LV1S (10 ) LV1S/LV1S weight (107 #/g) RV (108) RV/RV weight (107 #/g) LV1S 1 RV (108) 36.6 6.99 8.93 5.29 45.5 6 6 6 6 6 3.67 0.61 0.83 0.32 4.36 Twin (n 5 7) 27.0 7.63 7.33 6.80 34.4 6 6 6 6 6 3.11* 0.82 0.48 0.79 2.79* Male (n 5 5) 27.4 6.23 7.69 5.74 35.1 6 6 6 6 6 3.81 0.54 0.78 0.87 4.54 Female (n 5 9) 34.3 7.91 8.38 6.20 42.7 6 6 6 6 6 3.48 0.63 0.69 0.58 3.73 All (n 5 14) 31.8 7.31 8.13 6.03 40.0 6 6 6 6 6 2.61# 0.48 0.50 0.45 2.88 Abbreviations: LV1S, left ventricle plus septum; RV, right ventricle. Values shown are the means 6 SEM. *Indicates twins significantly different from singles (P < 0.05). # Indicates difference between LV1S and RV (P < 0.05). Number of Cardiomyocytes in Both Ventricles In the LV1S and RV combined, the average total number of cardiomyocytes (from all hearts examined) was 40.0 3 108 6 2.88 3 108. Twins had significantly fewer ventricular cardiomyocytes than singletons (twins, 34.4 3 108 6 2.79 3 108 cells vs. singletons, 45.5 3 108 6 4.36 3 108 cells; P 5 0.031). There was a trend for males to have fewer ventricular cardiomyocytes than females (males, 35.1 3 108 6 4.54 3 108 cells vs. females, 42.7 3 108 6 3.73 3 108 cells; P 5 0.069) (Table 2). When the data for all lambs were pooled, there were positive, significant associations between total ventricular cardiomyocyte number and birthweight (r2 5 0.560, P 5 0.002), necropsy weight (r2 5 0.391, P 5 0.017), and heart weight (r2 5 0.546, P 5 0.003). Proportion of Binucleated Cardiomyocytes In the LV, 99% of the cells were binucleated in both lambs studied, whereas in the RV 93% of the cells were binucleated and the remaining 7% were mononucleated. Cardiomyocyte Proliferation There were no Ki67-positive cells in the LVs or RVs from all lambs studied, confirming that cardiomyocytes had ceased proliferating by 9 weeks of age. In contrast, there was evidence of cell proliferation in fetal heart tissue at 133 days of gestation (Fig. 3). Ventricular Collagen Content The area of collagen in the RV (1.03% 6 0.12%) was significantly greater than in the LV (0.70% 6 0.06%, P < 0.05) (Fig. 4). There was no significant relation between the amount of collagen in the ventricles and either birthweight or necropsy weight. Similarly, there were no differences between singletons and twins, nor between males and females. DISCUSSION The major finding of this study is that the number of cardiomyocytes in the LV1S, at a time of development when cardiomyocyte proliferation has ceased, is directly related to body weight and heart tissue weight. This was found to be true across a wide range of birthweights within the normal range, including both singleton and twin offspring of both genders. The same was not true for the number of cardiomyocytes in the RV. Our findings have important implications for cardiac structure and function in adult life, especially in low birthweight (small for gestational age) individuals following catch-up growth or exposure to cardiovascular challenges. To examine the effects of a wide range of birthweights on cardiomyocyte development, we included twin and singleton lambs in our study, as well as male and female lambs. Twins are normally 20%–30% lighter than singletons at birth, in both sheep (Ross et al., 2005; Bloomfield et al., 2007; De Matteo et al., 2008) and humans (de Geus et al., 2001), presumably because of a mild restriction of fetal growth imposed by a smaller placenta. We found that twin lambs, as a group, were not different from singleton lambs in that heart weight and cardiomyocyte number were directly related to body weight, and cardiomyocyte data from singletons and twins showed the same relation to body and heart weight (i.e., data fell on the same line of identity). This suggests that across the normal range of birthweights for singletons or twins, the final number of cardiomyocytes is determined primarily by the size of the heart, which in turn is related to fetal and early postnatal growth. Hence, larger individuals born at term will likely begin life with a greater complement of cardiomyocytes than those who are smaller at term birth and in early infancy. Conversely, individuals who are small for gestational age, as a result of a chronic reduction in fetal nutrient or oxygen availability, are likely to have a reduced final complement of cardiomyocytes. In the present study, we chose to examine the hearts of lambs at 9 weeks after birth, as by this age, it was expected that the cardiomyocytes would be terminally differentiated and binuclear, and thus would have ceased dividing (Smolich et al., 1989; Burrell et al., 2003). Our use of Ki67 immunohistochemistry confirmed the absence of cardiomyocyte proliferation at 9 weeks, whereas marked proliferation was observed before birth. Interestingly, when we examined enzymatically isolated cardiomyocytes, there were negligible mononuclear cardiomyocytes (1%) in the LV1S, whereas in the RV, 7% of the cardiomyocytes were mononuclear, suggesting that these cardiomyocytes were still capable of division. Therefore, we can confidently conclude that the total number of cardiomyocytes we counted in the LV1S (the dominant chamber of the heart after birth) represents the final complement of cardiomyocytes; in contrast, in the RV the final number of cardiomyocytes may increase further if the mononucleated cells continue to divide. The difference in the proportion of mononucleated/ binucleated cells between the LV and RV may relate to the hemodynamic transition at birth, whereby the LV becomes the dominant pumping chamber; it is thus con- FETAL GROWTH AND CARDIOMYOCYTE NUMBER 35 Fig. 3. High-power (bar 5 10 mm) images demonstrating Ki67-positive cardiomyocytes (brown staining), which are indicative of cells that are proliferating. Note the absence of staining in heart tissue from 9week-old lambs (A), and the positive staining in heart tissue from a sheep fetus at 133 days of gestation (C). Staining was absent in the negative control sections in which the primary antibody was omitted (B, 9-week-old heart; D, fetal heart). Fig. 2. Relationship between cardiomyocyte number in the left ventricle plus septum (LV1S) and (A) birthweight, (B) necropsy weight at 9 weeks, and (C) heart weight. Data for singleton lambs (n 5 7) are shown by filled circles; data for twin lambs (n 5 7) are shown by open circles. ceivable that cardiomyocytes in the LV will have accelerated maturation compared with cardiomyocytes in the RV. The number of cardiomyocytes in the RV was considerably less than in the LV1S, and there was a trend for reduced density in the RV as well. Although we have not estimated cardiomyocyte number at birth, the differences between ventricles are again likely due to the considerable remodeling (including apoptosis) within the RV soon after birth (Smolich et al., 1989; Kajstura et al., 1995; Burrell et al., 2003), following the hemodynamic transition at birth. In the fetus, both ventricles effectively operate in concert, because of the presence of the ductus arteriosus that shunts blood from the pulmonary artery to the aorta. After birth, when the ductus arteriosus closes and the right heart perfuses only the low resistance, low pressure pulmonary circulation, there is marked apoptosis in the RV, resulting in the loss of cardiomyocytes (Kajstura et al., 1995; Burrell et al., 2003). This could account for the greatly reduced number of cardiomyocytes in the RV compared with the LV1S, and the lack of a significant relationship between the number of RV cardiomyocytes, with both birthweight and body weight at 9 weeks of age. The number of cardiomyocytes that we measured in the LV1S at 9 weeks (3.2 3 109) is similar to the number obtained in a previous study (Burrell et al., 2003) of lambs at 4 and 6 weeks after birth ( 2.0 3 109). The difference between the absolute values obtained between studies is likely due to differences in postnatal ages (a small proportion of cardiomyocytes may still be dividing 36 STACY ET AL. Fig. 4. Representative images showing collagen staining (blue) in the left ventricle (A) and right ventricle (B) of lambs at 9 weeks of age. Bar 5 10 mm. at 6 weeks) or alternatively, differences in methodologies or breed of sheep. In the earlier study (Burrell et al., 2003) cardiomyocyte number was derived indirectly from mean cardiomyocyte volume and LV mass, whereas we calculated the number using an unbiased optical disector–fractionator technique. As we studied lambs at 9 weeks after birth, it is likely that both prenatal and postnatal nutrition and growth contributed to the differences in both heart weight and cardiomyocyte number. Between birth and 9 weeks of age, the animals grew in proportion to their body weight, with no evidence of catch-up growth in animals born with low birthweight; that is, for individual animals postnatal growth rates (% increase per day) were likely to have been similar to fetal growth rates. Differences in the birthweights of our animals were not due to differences in gestation length, as all were born at term, and presumably were a result of differences in fetal growth resulting from either genetic factors or differences in fetal nutrition and/or oxygenation. As we found that arterial pressure was unrelated to body weight, differences in cardiomyocyte number between animals of different body weights could not be attributed to arterial pressure differences. Heart sparing, as a result of preferential perfusion of the myocardium, has often been described in growth-restricted human fetuses (al-Ghazali et al., 1989; Rizzo and Arduini, 1991; Baschat et al., 1997). In the present study, the growth of the heart was proportional to body weight, with no evidence of disproportionately large hearts in the smaller animals. Although redistribution of cardiac output favoring the heart has previously been described in the presence of acute fetal hypoxia in sheep (Sheldon et al., 1979; Smolich et al., 1989), increases in relative heart weight were not detected in IUGR lambs following placental embolization (Bubb et al., 2007) or carunclectomy (Morrison et al., 2007) in which hypoxia would be chronic. Similarly, there were no effects on relative heart weight at 8 weeks of age (Manalich et al., 2000) or at 2 years (Louey et al., 2000) following IUGR induced by placental embolization. In the present study, as the hearts were analyzed at 9 weeks of age, we cannot exclude the possibility that there may have been accelerated cardiac growth in the smaller fetuses in utero, which subsequently normalized after birth. There have been a number of experimental studies linking elevations in fetal cortisol levels with altered cardiomyocyte growth; however, the findings are not consistent; studies have shown increased cardiomyocyte proliferation (Giraud et al., 2006), cardiomyocyte hypertrophy (Martin et al., 2000), or no effect (Phillips and Barker, 1997). In this regard, it is important to note that we have previously shown that plasma cortisol levels are not elevated as a result of the spontaneous growth restriction associated with twinning (De Matteo et al., 2008), and so potentially confounding effects of elevations in fetal cortisol levels do not apply in the present study. In animal models, an increase in fibrosis has often been described in the RV when compared with the LV in the adult heart (Weber and Brilla, 1991). Interestingly, by 9 weeks of age in the present study, there was already an increased level of fibrosis in the RV compared with the LV, which is likely a result of replacement fibrosis following cardiomyocyte apoptosis accompanying the postnatal RV remodeling. Implications for Adult Cardiovascular Function and Disease Although there is convincing epidemiological data linking low birthweight with an increased incidence of heart disease later in life (Eriksson and Forsen, 2002), very few studies have investigated the effects of IUGR and low birthweight on the structure of the heart. In two recent sheep studies (one from our laboratory), it has been shown that IUGR, as a result of chronic placental insufficiency induced by either placental embolization (Bubb et al., 2007) or carunclectomy (Morrison et al., 2007), leads to immaturity of the cardiac muscle in late gestation. In both studies the proportion of undifferentiated, mononucleated cardiomyocytes was significantly greater in the IUGR hearts compared with the hearts of normally grown fetuses; this implies delayed differentiation and/or increased proliferation of cardiomyocytes in these IUGR hearts. What effect the immaturity of the myocardium has on the final number of cardiomyocytes in these IUGR hearts is yet to be determined. As cardiomyocytes, in general, cease proliferating soon after birth, our present findings suggest that infants of low birthweight may have a reduced capacity for physiological hypertrophy of the heart later in life, especially if their adaptive capabilities are challenged; this could occur, for example, as a result of catch-up growth, induction of obesity, or development of hypertensive LV hypertrophy. In support of this concept, maternal protein restriction in rats leads to fewer cardiomyocytes in the heart at birth (Corstius et al., 2005), and this is associated with an increased deposition of interstitial fibrosis within the myocardium by early adulthood (Lim et al., 2006). In addition, the hearts of IUGR rat offspring exposed to hypoxia in utero are more vulnerable to ischemia and reperfusion injury in adulthood (Li et al., 2003). CONCLUSIONS We have shown that the number of cardiomyocytes in the LV, at a time when cardiomyocyte differentiation has essentially ceased, is related to birthweight and heart weight. That is, individuals with a low birthweight, FETAL GROWTH AND CARDIOMYOCYTE NUMBER regardless of whether they are singleton or multiple births, are likely to begin life with fewer LV cardiomyocytes than individuals born heavier. As cardiomyocytes cease proliferating soon after birth, cardiomyocyte number is unlikely to be affected even if catch-up growth occurs after infancy. A lower cardiomyocyte number following restricted fetal growth, and hence low birthweight and low neonatal weight may compromise the postnatal adaptive capabilities and function of the heart, especially if challenged in adulthood. It is possible that a reduced number of cardiomyocytes in the LV of individuals born with lower birthweights may contribute to the association between low birthweight and increased risk of heart disease in adulthood, especially in individuals who catch up in body weight or who become obese later in life. ACKNOWLEDGMENTS The authors gratefully acknowledge the help of N. Wreford, A. Satragno, N. Blasch, M. Tare, and S. Connell. LITERATURE CITED al-Ghazali W, Chita SK, Chapman MG, Allan LD. 1989. Evidence of redistribution of cardiac output in asymmetrical growth retardation. Br J Obstet Gynaecol 96:697–704. Barker DJ, Gluckman PD, Godfrey KM, Harding JE, Owens JA, Robinson JS. 1993. Fetal nutrition and cardiovascular disease in adult life. Lancet 341:938–941. Baschat AA, Gembruch U, Reiss I, Gortner L, Diedrich K. 1997. Demonstration of fetal coronary blood flow by Doppler ultrasound in relation to arterial and venous flow velocity waveforms and perinatal outcome—the ‘heart-sparing effect’. Ultrasound Obstet Gynecol 9:162–172. Bloomfield FH, Oliver MH, Harding JE. 2007. Effects of twinning, birth size, and postnatal growth on glucose tolerance and hypothalamic-pituitary-adrenal function in postpubertal sheep. Am J Physiol Endocrinol Metab 292:E231–E237. Bubb KJ, Cock ML, Black MJ, Dodic M, Boon WM, Parkington HC, Harding R, Tare M. 2007. Intrauterine growth restriction delays cardiomyocyte maturation and alters coronary artery function in the fetal sheep. J Physiol 578:871–881. Burrell JH, Boyn AM, Kumarasamy V, Hsieh A, Head SI, Lumbers ER. 2003. Growth and maturation of cardiac myocytes in fetal sheep in the second half of gestation. Anat Rec A Discov Mol Cell Evol Biol 274:952–961. Corstius HB, Zimanyi MA, Maka N, Herath T, Thomas W, van der Laarse A, Wreford NG, Black MJ. 2005. Effect of intrauterine growth restriction on the number of cardiomyocytes in rat hearts. Pediatr Res 57:796–800. de Geus EJ, Posthuma D, Ijzerman RG, Boomsma DI. 2001. Comparing blood pressure of twins and their singleton siblings: being a twin does not affect adult blood pressure. Twin Res 4:385–391. De Matteo R, Stacy V, Probyn M, Desai M, Ross M, Harding R. 2008. The perinatal development of arterial pressure in sheep: effects of low birth weight due to twinning. Reprod Sci 15:66–74. Desai M, Gayle D, Babu J, Ross MG. 2005. Permanent reduction in heart and kidney organ growth in offspring of undernourished rat dams. Am J Obstet Gynecol 193:1224–1232. Eriksson JG, Forsen TJ. 2002. Childhood growth and coronary heart disease in later life. Ann Med 34:157–161. Gerdes J, Lemke H, Baisch H, Wacker HH, Schwab U, Stein H. 1984. Cell cycle analysis of a cell proliferation-associated human 37 nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol 133:1710–1715. Giraud GD, Louey S, Jonker S, Schultz J, Thornburg KL. 2006. Cortisol stimulates cell cycle activity in the cardiomyocyte of the sheep fetus. Endocrinology 147:3643–3649. Goodfellow J, Bellamy MF, Gorman ST, Brownlee M, Ramsey MW, Lewis MJ, Davies DP, Henderson AH. 1998. Endothelial function is impaired in fit young adults of low birth weight. Cardiovasc Res 40:600–606. Gundersen HJ. 2002. The smooth fractionator. J Microsc 207:191– 210. Kajstura J, Mansukhani M, Cheng W, Reiss K, Krajewski S, Reed JC, Quaini F, Sonnenblick EH, Anversa P. 1995. Programmed cell death and expression of the protooncogene bcl-2 in myocytes during postnatal maturation of the heart. Exp Cell Res 219:110–121. Lawlor DA, Davey Smith G, Ebrahim S. 2004. Birth weight is inversely associated with coronary heart disease in post-menopausal women: findings from the British women’s heart and health study. J Epidemiol Community Health 58:120–125. Li F, Wang X, Capasso JM, Gerdes AM. 1996. Rapid transition of cardiac myocytes from hyperplasia to hypertrophy during postnatal development. J Mol Cell Cardiol 28:1737–1746. Li G, Xiao Y, Estrella JL, Ducsay CA, Gilbert RD, Zhang L. 2003. Effect of fetal hypoxia on heart susceptibility to ischemia and reperfusion injury in the adult rat. J Soc Gynecol Investig 10:265–274. Lim K, Zimanyi MA, Black MJ. 2006. Effect of maternal protein restriction in rats on cardiac fibrosis and capillarization in adulthood. Pediatr Res 60:83–87. Louey S, Cock ML, Stevenson KM, Harding R. 2000. Placental insufficiency and fetal growth restriction lead to postnatal hypotension and altered postnatal growth in sheep. Pediatr Res 48:808–814. Manalich R, Reyes L, Herrera M, Melendi C, Fundora I. 2000. Relationship between weight at birth and the number and size of renal glomeruli in humans: a histomorphometric study. Kidney Int 58:770–773. Martin H, Hu J, Gennser G, Norman M. 2000. Impaired endothelial function and increased carotid stiffness in 9-year-old children with low birthweight. Circulation 102:2739–2744. Morrison JL, Botting KJ, Dyer JL, Williams SJ, Thornburg KL, McMillen IC. 2007. Restriction of placental function alters heart development in the sheep fetus. Am J Physiol Regul Integr Comp Physiol 293:R306–R313. Oparil S, Bishop SP, Clubb FJ, Jr. 1984. Myocardial cell hypertrophy or hyperplasia. Hypertension 6:III38–III43. Phillips DI, Barker DJ. 1997. Association between low birthweight and high resting pulse in adult life: is the sympathetic nervous system involved in programming the insulin resistance syndrome? Diabet Med 14:673–677. Rizzo G, Arduini D. 1991. Fetal cardiac function in intrauterine growth retardation. Am J Obstet Gynecol 165:876–882. Ross MG, Desai M, Guerra C, Wang S. 2005. Programmed syndrome of hypernatremic hypertension in ovine twin lambs. Am J Obstet Gynecol 192:1196–1204. Sheldon RE, Peeters LL, Jones MD, Jr, Makowski EL, Meschia G. 1979. Redistribution of cardiac output and oxygen delivery in the hypoxemic fetal lamb. Am J Obstet Gynecol 135:1071–1078. Smolich JJ, Walker AM, Campbell GR, Adamson TM. 1989. Left and right ventricular myocardial morphometry in fetal, neonatal, and adult sheep. Am J Physiol 257:H1–H9. Weber KT, Brilla CG. 1991. Pathological hypertrophy and cardiac interstitium. Fibrosis and renin-angiotensin-aldosterone system. Circulation 83:1849–1865. Zhao M, Shu XO, Jin F, Yang G, Li HL, Liu DK, Wen W, Gao YT, Zheng W. 2002. Birthweight, childhood growth and hypertension in adulthood. Int J Epidemiol 31:1043–1051.