Thyroid hormone drives fetal cardiomyocyte maturation

Abstract

Tri-iodo-l-thyronine (T₃) suppresses the proliferation of near-term serum-stimulated fetal ovine cardiomyocytes in vitro. Thus, we hypothesized that T₃ is a major stimulant of cardiomyocyte maturation in vivo. We studied 3 groups of sheep fetuses on gestational days 125–130 (term ∼145 d): a T₃-infusion group, to mimic fetal term levels (plasma T₃ levels increased from ∼0.1 to ∼1.0 ng/ml; t_1/2∼24 h); a thyroidectomized group, to produce low thyroid hormone levels; and a vehicle-infusion group, to serve as intact controls. At 130 d of gestation, sections of left ventricular freewall were harvested, and the remaining myocardium was enzymatically dissociated. Proteins involved in cell cycle regulation (p21, cyclin D1), proliferation (ERK), and hypertrophy (mTOR) were measured in left ventricular tissue. Evidence that elevated T₃ augmented the maturation rate of cardiomyocytes included 14% increased width, 31% increase in binucleation, 39% reduction in proliferation, 150% reduction in cyclin D1 protein, and 500% increase in p21 protein. Increased expression of phospho-mTOR, ANP, and SERCA2a also suggests that T₃ promotes maturation and hypertrophy of fetal cardiomyocytes. Thyroidectomized fetuses had reduced cell cycle activity and binucleation. These findings support the hypothesis that T₃ is a prime driver of prenatal cardiomyocyte maturation.—Chattergoon, N. N., Giraud, G. D., Louey, S., Stork, P., Fowden, A. L., Thornburg, K. L. Thyroid hormone drives fetal cardiomyocyte maturation

Thyroid hormones (THs) are phylogenetically ancient and known to regulate diverse developmental processes among lower vertebrates, as exemplified by their fundamental role in the transformation of aquatic tadpoles into terrestrial frogs (1). In humans, abnormal TH levels in the fetus can lead to a range of complications, including decreased cardiac output, growth restriction, tachycardia, neuropathologies, and even fetal demise (2–4). Nevertheless, the influence of TH on cardiomyocyte development in mammals has been studied very little. This gap in knowledge remains a serious detriment to our understanding of potential mechanisms by which fetal hearts may be harmed when TH levels fall outside the normal range.

The primary form of TH secreted by the mammalian thyroid gland is 3,3′,5,5′tetra-iodo-l-thyronine (thyroxine or T₄; ref. 5). T₄ levels increase during gestation as the human fetal thyroid gland becomes active at around 15 wk, but fetal plasma levels are also determined by maternal T₄ transported across the placenta (6, 7). In the ovine fetus, T₄ levels in plasma gradually decrease toward term as T₄ is converted to the more active form of TH, 3,3′,5 tri-iodo-l-thyronine (T₃; ref. 8). The T₄ to T₃ conversion occurs through the activity of deiodinases whose expression is regulated by cortisol (9). As term approaches and the hypothalamic-pituitary-adrenal axis becomes activated (10), increased circulating cortisol concentrations stimulate an increase in expression of monodeiodinases, including the D2 form, which transforms T₄ to T₃. Thus, total T₃ levels increase from ∼0.3 to ∼1.5 nM in fetal sheep during the last weeks before delivery (8). This prepartum rise in T₃ promotes maturational processes in lung, liver, and kidney (11–13), and presumably in other organs as well.

The developing ovine myocardium is ideal for investigating the maturational role of THs because like human hearts—but unlike rodent hearts—most ovine cardiomyocytes go through their terminal maturation phase before birth (14–16). One sign of maturation in sheep and rat cardiomyocytes is the generation of a second nucleus and terminal differentiation, at which time cardiomyocytes permanently exit the cell cycle. However, the timing between rodents and large mammals is different. The myocyte population in the ovine myocardium is ∼80% binucleated at birth (16), but only ∼7% of cardiomyocytes in the rodent heart. By 2 wk of age, >95% of rodent cardiomyocytes have become binucleated (17).

The ovine heart grows primarily by increasing cardiomyocyte numbers (hyperplasia) over the first two-thirds of gestation, but in the last few weeks, cardiomyocyte volume expansion (hypertrophy) becomes increasingly important in myocardial growth (16, 18). In fetal sheep, the maturation of cardiomyocytes over the last third of gestation is characterized by binucleation, suppression of mitotic activity, and cardiomyocyte enlargement (16, 19). While the roles of angiotensin II, cortisol, and insulin-like growth factor 1 are known to regulate the hyperplastic response (15, 20, 21), the chemical signals that regulate these aforementioned “macromaturational” processes have not been discovered.

The features that characterize the maturational process in cardiomyocytes have been described in several experimental animals, including sheep (18, 22–24). In mammals, the maturing ventricular myocyte expands via physiological hypertrophy, packs the myocyte interior with contractile elements, establishes the t-tubule system to expedite excitation-contraction coupling, and switches to an internal Ca²⁺ cycling system through development of the internal Ca²⁺ storage compartment, the sarcoplasmic reticulum (SR) (14, 25, 26). Cardiac maturation is normally associated with increased expression of SR calcium ATPase (SERCA2a; ref. 27). In rats, SERCA2a is a known target of T₃ action (28).

Atrial natriuretic peptide (ANP) is made by ventricular and atrial cardiomyocytes in the fetal heart, and its expression increases as gestational age proceeds. In normal adults, ANP is not expressed in the cardiac ventricles but is reexpressed in response to stress (29). ANP appears to be important as a protector of excessive myocardial growth (30–33). Hypothyroid conditions in the ovine fetus decrease ANP production (34) and decrease cardiac output (2). Thus, ANP is another molecule that is associated with myocardial maturation.

Mitogen-activated protein kinase (MAPK) and phosphoinositol-3 kinase (PI3K)-related cascades are important in the regulation of growth and maturation of the heart, including the hypertrophic growth of the postnatal heart (35–37). However, roles of T₃ in regulating specific branches of the MAPK and PI3K cascades involved in cardiomyocyte growth and maturation, including the extracellular signal-related kinase (ERK) and the mammalian target of rapamycin (mTOR) protein, are not known.

In clinical studies, low maternal plasma TH levels during pregnancy are also associated with neuropathologies in offspring and increased rates of fetal demise (38–40). Conversely, fetal hyperthyroxinemia leads to fetal complications that include growth restriction, tachycardia, and suppression of the hypothalamic-pituitary-thyroid axis (3). Thus, it appears that adverse outcomes occur with TH levels outside a normal concentration window. Whether abnormal concentrations of TH affect the pattern of fetal heart development remains unknown.

To understand the role of T₃ in regulating the maturation of fetal cardiomyocytes, we hypothesized that T₃ is the primary driver of cardiomyocyte maturation in fetal sheep in vivo. We studied 3 groups of fetuses with widely different circulating T₃ levels to characterize the normal macro features of cardiomyocyte maturation under these conditions as judged by 3 criteria: stimulation of binucleation, suppression of proliferation, and stimulation of cardiomyocyte enlargement among binucleated myocytes. We tested whether elevated T₃ would activate target signaling proteins, ERK and mTOR, and alter cell cycle regulatory genes, including the cell cycle stimulant protein cyclin D1, the cell cycle inhibitor protein p21, and the antiproliferative peptide ANP. T₃ was infused into the first group of preterm fetuses; a second group was thyroidectomized to produce lower than normal T₃ levels; and a third group served as normal intact vehicle-infused controls.

MATERIALS AND METHODS

Animals

Sheep (Ovis aries, mixed western breed) were studied in accordance with the Institutional Animal Care and Use Committee at Oregon Health and Science University. The fetal sheep model was used rather than a rodent model because the size and growth rate of the heart is similar to that of the human; large numbers of cardiomyocytes can be dissociated from a single heart from midgestation onward; ovine fetal cardiomyocytes go through terminal differentiation and become binucleated during fetal life; and fetal hemodynamic factors, including heart rate and blood pressure, can be monitored continuously in the unanesthetized preparation from midgestation onward.

Surgery

Twelve time-mated ewes, carrying twin fetuses, underwent sterile surgery at 120 ± 1 d gestational age (dGA; term ∼145 d). Surgeries were performed on both of the twin fetuses. Polyvinyl catheters placed in the fetal carotid artery and jugular vein were used for continuous hemodynamic recordings, infusion of T₃ or vehicle (i.v.), and blood sampling; a catheter attached to the fetal skin was used to measure amniotic fluid pressure (15). A subset of fetuses was thyroidectomized at the time of surgery prior to implanting catheters (41). Ewes were allowed 4 d recovery from surgery before they were moved to stanchions for the duration of the experiment. Infusions started 5 d after surgery.

Infusion studies

Fetuses (n=8/group) were assigned randomly to one of three groups: group 1, nonthyroidectomized (intact) T₃-infusion (high T₃) group; group 2, thyroidectomized (TX; low T₃) group; or group 3, intact vehicle-infusion (normal T₃) group. Each twin of a pair was treated similarly to negate any communication of treatments between the two fetuses. In group 1, T₃ (Sigma, St. Louis, MO, USA) was infused at a rate of 54 μg/d i.v., for a period of 5 d (from 125 to 130 dGA). A target concentration of T₃ in plasma of 1.0 ng/ml was chosen for the infusion group because it represents the circulating prepartum concentration (∼2 wk later than the age we are studying; ref. 8). Other investigators have used a similar regimen in fetal sheep (8, 42). In blood, >99% of total T₄ and T₃ is transported bound to thyroxine-binding globulin and other plasma proteins. Only the free unbound form of the hormone is available to stimulate receptors within individual cells. TH actions occur through their nuclear receptors TRα and TRβ (43–46) and through separate nongenomic mechanisms (47–49). TH receptors have affinities for T₃ that are some 12 times higher than for T₄ (48), thus explaining the greater potency of T₃ vs. T₄. Hence, we chose to infuse T₃ rather than T₄ directly into the sheep fetus in order to better control the target T₃ concentration over the course of the study. Control animals were infused with a vehicle solution (54 μl 1M NaOH/d in lactated Ringers solution, 38.4 ml/d i.v.) at the same rate and duration as for T₃. TX fetuses were infused with sterile lactated Ringers solution as vehicle. Fetal arterial blood was collected before the study and then daily to determine fetal pH, blood gases, and hematocrit (Radiometer ABL 720; Radiometer A/S, Copenhagen, Denmark; values corrected to 39°C); plasma samples were used to measure free and total T₃.

Hemodynamic measurements

Fetal aortic pressure, right atrial pressure, amniotic pressure, and heart rate were continuously monitored and recorded throughout the study. Pressures were measured with Abbott Transpac pressure transducers (Abbott, Abbott Park, IL, USA) and a calibrated computerized recording system (Powerlab, ADInstruments, Colorado Springs, CO, USA; Apple, Cupertino, CA, USA). The system was calibrated against a mercury manometer and rezeroed for drift before each measurement. All fetal intravascular pressures were referred to amniotic fluid pressure as 0. Zeroes were checked every morning, after which the data were taken over the next 60-min period, averaged, and recorded for later analysis. Heart rates were derived from arterial pressure measurements.

Plasma hormone concentrations

Daily plasma samples were collected from each fetus and stored at −20°C until analysis. Plasma total T₃ and free T₃ were measured using respective Coat-a-Count radioimmunoassays (Siemens Healthcare Diagnostics, Inc., Los Angeles, CA, USA). The lower limits of detection were 0.07 ng/ml for total T₃ and 0.2 pg/ml for free T₃. The interassay coefficients of variation were 8% for total T₃ and 7.8% for free T₃.

Cardiac myocyte isolation

At the end of the study (130±1 dGA), ewes were euthanized by intravenous injection of a commercial solution of sodium pentobarbital (SomnaSol, ∼80 mg/kg; Butler Schein Animal Health, Dublin, OH, USA). A bolus dose of heparin (10,000 U) followed by 10 ml of saturated potassium chloride (KCl) was injected into the umbilical vein to arrest the fetal heart in diastole. Prior to dissociation of the heart to isolate cardiomyocytes, a 0.5-cm² section of the left ventricle (midwall, full thickness) was removed and flash-frozen for molecular analyses. The cut edges of the hole in the ventricular wall were lightly cauterized before the myocyte dissociation procedure so that dissociation solutions would not leak.

Hearts were enzymatically dissociated as described previously by our laboratory (50). Briefly, hearts were perfused retrogradely with 3 gassed (95% O₂ and 5% CO₂ at 39°C) solutions: Tyrode's buffer for 5–10 min (140 mM NaCl, 5 mM KCl, 1 mM MgCl₂·6H₂O, 10 mM glucose, and 10 mM HEPES; pH adjusted to 7.35 with NaOH; no calcium added; components from Sigma); a collagenase/protease solution for 5–10 min consisting of type II collagenase (160 U/ml; Worthington Biochemical Corp., Lakewood, NJ, USA) and type XIV protease (0.78 U/ml; Sigma,) in Tyrode's buffer; and a calcium-free Kraftbrühe (KB) solution (74 mM glutamic acid, 30 mM KCl, 30 mM KH₂PO₄, 20 mM taurine, 3 mM MgSO₄, 0.5 mM EGTA, 10 mM HEPES, and 10 mM glucose; pH adjusted to 7.37 using KOH; components from Sigma). Following the perfusion, the left ventricular (LV) and right ventricular (RV) free walls were dissected from the heart and placed in separate tubes containing KB solution. The tissue was gently swirled to release the myocytes. The cell slurry rested at room temperature for 30 min. Myocytes were fixed with an equal volume of 2% paraformaldehyde for measurement of maturational state (binucleation), cell cycle activity, and myocyte size.

Myocyte size

These measurements were performed as described previously in our laboratory (15). Cardiac myocyte lengths and maximal widths were measured from photomicrographs of methylene blue-stained wet mounts at ×400 view (Zeiss Axiophot; Bartels and Stout, Bellevue, WA, USA). Cardiac myocyte lengths and widths were measured using calibrated imaging software (Image-Pro Plus 6.2 for PC; Media Cybernetics, Bethesda, MD, USA). Myocyte volume was calculated as described previously by correcting from the volume of a cylinder with a height equal to the length of the myocyte and a diameter equal to the myocyte width (14). At least 50 mononucleated and binucleated myocytes were measured per ventricle per fetus. For all analyses of dissociated cardiac myocytes, cells were selected by nonrepeating, random sampling across the microscope slide.

Binucleation

Methylene blue staining allowed the number of nuclei per cardiac myocyte to be determined. At least 300 myocytes per ventricle per animal were counted to determine percentage binucleation, separately of the cells measured for size.

Cell cycle activity

Cell cycle activity was estimated by determining the fraction of cardiomyocytes that were positive for Ki-67 staining using methods previously described by our laboratory (16). The Ki-67 protein is present in all phases of the cell cycle; it disappears in cells in G₀. Isolated fixed myocytes (50,000) were dried and postfixed onto slides (SuperFrost; Fisher, Pittsburgh, PA, USA) by immersion in acetone for 30 min at 4°C. Cells were permeabilized for antigen retrieval in sodium citrate (0.01 M, pH 6.0) for 6 min at 85°C. Nonspecific staining and endogenous peroxidase activities were blocked with 0.3% hydrogen peroxide. Slides were incubated overnight at 4°C with the Ki-67 antibody (1:200 dilution, mouse monoclonal MIB-1; DakoCytomation, Carpinteria, CA, USA). Samples were incubated with the biotinylated secondary antibody (1:200; Vectastain ABC Kit, Mouse IgG; VectorLabs, Burlingame, CA, USA). Positive nuclei were stained with 3,3′-diaminobenzidine chromagen (DAB) in substrate buffer (DakoCytomation) for 1–5 min, and cells were counterstained with 0.1% methylene blue. A minimum of 500 myocytes were counted per ventricle per fetus to determine the number of Ki-67-positive mononucleated myocytes (51). Results are expressed as the percentage of Ki-67-positive mononucleated myocytes ÷ total number of mononucleated myocytes. Phospho-histone 3 (pH 3, 1:200 dilution; Cell Signaling, Danvers, MA, USA) was used to compare the percentage of cells in mitosis at the end of the infusion period to those that are in the cell cycle. Histone 3 phosphorylation is tightly correlated with chromosome condensation in mitotic prophase. The staining protocol and analysis is the same as that described for Ki-67.

Western blot

Protein extracted from frozen LV tissue [lysis solution: 5 mM Tris-HCl, 5 mM EGTA, 5 mM EDTA, 0.06% SDS (RIPA buffer; Upstate, Temecula, CA, USA); protease inhibitor Complete Mini EDTA-free tablet (Roche, Indianapolis, IN, USA); and inhibitors against alkaline, serine/threonine, and tyrosine phosphatases (Sigma)] was quantified by BCA assay (Pierce, Rockford, IL, USA). Cytosolic protein from each ventricle (20 μg) was separated by SDS-PAGE and transferred onto reinforced nitrocellulose membranes (Optitran BA-S 83; Whatman-GE, Piscataway, NJ, USA) before incubation with primary antibodies (listed below). Protein was visualized by chemiluminescence (SuperSignal; Pierce) and digitally quantified (ImageJ 1.43; U.S. National Institutes of Health, Bethesda, MD, USA) (50).

Antibodies for Western blot analysis

The following antibodies were obtained from Cell Signaling. Phospho-p44/42 MAPK (ERK1/2) rabbit polyclonal antibody detects endogenous levels of p44 and p42 MAP kinase (ERK1 and ERK2) when phosphorylated either individually or dually at Thr202 and Tyr204. Phospho-AKT (Ser473; D9E) XP rabbit monoclonal antibody detects endogenous levels of AKT only when phosphorylated at Ser473. Phospho-p70 S6 kinase (Ser371) rabbit monoclonal antibody detects endogenous levels of p70 S6 kinase (p70^S6K) only when phosphorylated at Ser371. AKT rabbit polyclonal antibody detects endogenous levels of total AKT1, AKT2, and AKT3 proteins. Rabbit p70^S6K antibody detects endogenous levels of total p70^S6K protein. Rabbit polyclonal Phospho-mTOR (Ser2448) antibody detects endogenous levels of mTOR only when phosphorylated at Ser2448, and rabbit polyclonal total mTOR detects endogenous mTOR. Mouse monoclonal p21 Waf1/Cip1 (DCS60) antibody detects endogenous levels of total p21 protein. The antibody does not cross-react with other cdk inhibitors and recognizes the amino-terminal portion of p21. Rabbit monoclonal α-tubulin detects endogenous levels of total α-tubulin protein, and does not cross-react with recombinant β-tubulin. Mouse monoclonal cyclin D1 recognizes full-length peptide (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Horseradish peroxidase-linked secondary antibodies (goat anti-rabbit IgG and horse anti-mouse IgG) were obtained from Cell Signaling.

Quantitative polymerase chain reaction (PCR) assays for hypertrophic genes

The following primers were used for PCR.

ANP: forward 5′-CCCGTGTATGGCTCTGTGTC-3′, reverse 5′-GAGGGCACAGCCTCATCTTC-3′ (195 bp).

SERCA2a: forward 5′-TGACAGGTGTACCCACATTC-3′, reverse 5′-AAGTTGGCAGAGTCCTCAAG-3′ (186 bp).

β-Actin: forward 5′-CTGCACCACCAACTGCTTAG-3′, reverse 5′-CAGTGGATGCAGGGATGATG-3′ (150 bp).

RNA was extracted from 50 mg of frozen fetal heart tissue using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized from 1 μg RNA from each sample using the GeneAmp RNA PCR kit (Applied Biosystems, Foster City, CA, USA). The size of each PCR product was verified on a 1% agarose gel with a 1 kb+ Ladder (Invitrogen). In subsequent runs, melting curve analysis (where temperature is raised from 65 to 95°C at a rate of 0.1°C/s) was used to verify correct production of product. PCR reactions were performed using ABI Power SYBR Green (Applied Biosystems) in the Stratagene MX-3005P PCR instrument (Stratagene, La Jolla, CA, USA). Optimal cycling conditions were determined for each primer pair. Quantification values were generated from samples with only a single product that melts at the expected temperature. For each unknown sample, the concentration of the amplicon was calculated based on the equation derived from the standard curve. These concentrations were normalized to the concentration of the housekeeping gene, β-actin, for the correlating sample. β-Actin was chosen for normalization after determining that it was not affected by changing experimental conditions.

Statistics

The in vivo effect of T₃ on fetal hemodynamics was analyzed by 2-way analysis of variance (ANOVA) with post hoc Bonferroni's multiple comparison test against d 0 values. Myocyte parameters (cell sizes, cell cycle activity, and binucleation), protein expression, and immunohistochemistry (Ki-67 and pH3) from the 3 experimental groups were compared by 1-way ANOVA with post hoc Tukey's multiple comparison test. Western blot densitometry was quantified over the linear range using NIH ImageJ 1.4. Phosphoproteins analyzed by Western blot were normalized to their respective total unphosphorylated protein equivalent. Remaining protein analysis was normalized to α-tubulin, which served as the loading control. While the same myocyte analysis was performed on both ventricles, data from the right ventricle, except for cell size, are not shown, as the trends were the same as for the left ventricle. No tissue was collected for molecular studies from right ventricles. All statistics were performed using GraphPad Prism 4.03 for Windows (GraphPad, San Diego, CA, USA). Values of P < 0.05 were accepted as significant. Graphs display all three groups of animals per respective measurement. Data are expressed as means ± se.

RESULTS

Hemodynamic measurements

Fetal blood-gas values were unchanged among groups over the course of the study (Table 1). Mean fetal arterial pressure was elevated significantly but only on d 5 in the T₃-infused group when compared to control animals (Fig. 1A). Fetal arterial pressure was not changed across the 5 experimental days in any group. However, pressures were slightly higher in the high-T₃ group than in TX fetuses. Fetal heart rates were unchanged in control fetuses or in fetuses receiving T₃, though a decrease in heart rate was found in TX fetuses compared to control fetuses across the 5-d experimental period (Fig. 1B).

Table 1.

Hemodynamic measurements with reference to thyroid status

Measurement	Thyroid status
	TX	Control	T3 infusion
Arterial pressure (mmHg)	37.1 ± 0.7+++	40.1 ± 1.0	47.6 ± 1.2***
Heart rate (bpm)	146.2 ± 5.4+	162.9 ± 3.3	163.2 ± 3.3
Blood gas
Po2 (mmHg)	23.5	20.9	22.8
Pco2 (mmHg)	50.1	52.1	48.6
pH	7.34	7.33	7.35

Measurements on final day of study. Data are means ± se; n = 8/group. Significant differences between groups are indicated.

^***P < 0.001 vs. control;

⁺P < 0.05,

⁺⁺⁺P < 0.001 vs. T₃.

Figure 1.

Hemodynamic measurements with changes in thyroid status. A) Mean aortic pressure did not differ between control (solid squares) and TX (open circles), but pressure was significantly increased on d 5 in T₃-infused fetuses (open triangles). Mean aortic pressure was depressed in TX compared to T₃-infused fetuses throughout the study period. B) Heart rate was decreased significantly in the TX group. C) Total T₃ levels were unchanged in control fetuses, while T₃ infusion increased concentrations to those found near term. T₃ was undetectable in TX fetuses. Data are means ± se; n = 8/group. ***P < 0.001 vs. control; ⁺P < 0.05, ⁺⁺⁺P < 0.001 vs. TX.

TH concentrations

As expected, the plasma concentrations of total T₃ in control fetuses and in the pre-T₃-infusion animals were similar at ∼0.1 ± 0.1 ng/ml. The target concentration of T₃ in plasma of 1.0 ng/ml was approximated within the first day of T₃ infusion, reached 1.1 ± 0.3 ng/ml by d 2, and was maintained for the duration of the study. Free T₃ levels reached 2.2 ± 0.8 pg/ml in the T₃-infusion group and 0.1 ± 0.1 pg/ml in the control group; total and free T₃ levels in the TX group were not detectable throughout the study period.

Organ weights

Unadjusted organ weights (g), as well as organ weights adjusted for body weight (g/kg), are summarized for heart, brain, liver, combined kidney, combined adrenal, and combined thyroid gland (control and T₃ groups only) in Table 2. Unadjusted heart weight was unchanged by thyroid status, but heart weight to body weight ratio was significantly lower in TX fetuses. Kidney weights (g and g/kg body weight) were increased in TX fetuses compared to control fetuses and compared to T₃-infused fetuses. No effect of thyroid status was found on brain, liver, adrenal glands, or thyroid gland weights.

Table 2.

Organ weights

Organ	Thyroid status
	TX		Control		T3 infusion
	Raw (g)	Ratio (g/kg)	Raw (g)	Ratio (g/kg)	Raw (g)	Ratio (g/kg)
Body (kg)	3.5 ± 0.1		3.2 ± 0.2		3.3 ± 0.2
Heart	21.1 ± 1.2	5.9 ± 0.2*,+	21.6 ± 0.9	6.8 ± 0.3	22.3 ± 1.0	6.8 ± 0.3
Brain	41.4 ± 1.3	11.7 ± 0.1+	40.1 ± 1.5	12.6 ± 0.5	42.4 ± 1.4	13 ± 0.5
Liver	74.5 ± 7.2	20.8 ± 1.3	58.8 ± 0.0	18.2 ± 0.6	65.7 ± 4.3	20 ± 1
Kidney	22.6 ± 1.6*	6.4 ± 0.3**,++	17.3 ± 1.4	5.3 ± 0.2	17.7 ± 1	5.4 ± 0.2
Adrenal	0.3 ± 0.02	0.1 ± 0.01	0.3 ± 0.03	0.1 ± 0.01	0.3 ± 0.03	0.1 ± 0.01
Thyroid	NA	NA	0.6 ± 0.05	0.2 ± 0.02	0.7 ± 0.05	0.2 ± 0.01

Data are means ± se; n = 8/group. Ratio, organ to fetal body weight; NA, tissue not available. Significant differences between groups are indicated.

^*P < 0.05,

^**P < 0.01 vs. control;

⁺P < 0.05,

⁺⁺P < 0.01 vs. T₃.

Effect of T₃ on maturational state of fetal cardiomyocytes (LV only; similar trends in RV)

One index of cardiomyocyte maturation, cardiomyocyte binucleation, was related to the concentration of circulating T₃ levels in the fetal groups (Fig. 2). In the left ventricle, 21.4 ± 4.1% of cardiomyocytes from TX animals were binucleated compared to 42.6 ± 1.1% in vehicle-infused animals (P<0.001) and 56.8 ± 2.4% in T₃ infused fetuses (P<0.001). Binucleation was increased significantly in the left ventricles of T₃-infused fetuses compared to vehicle-infused fetuses (P<0.01). Cardiomyocyte binucleation in right ventricles showed a similar trend (TX, 20.6±8.6%; control, 45.8±9%; and T₃, 55.8±8.5%).

Figure 2.

Increased circulating T₃ promotes terminal differentiation of fetal sheep cardiomyocytes. Elevated T₃ concentration increased the portion of LV myocytes that were binucleated. Data are means ± se; n = 8/group. **P < 0.01, ***P < 0.001 vs. control; ⁺⁺⁺P < 0.001 vs. T₃.

Proliferative markers in cardiomyocytes (LV only reported; similar trends in RV)

The percentage of mononucleated myocytes that stained positive for Ki-67 (Fig. 3A) was lower in fetuses exposed to elevated levels of T₃ compared to control fetuses (6.1±0.7 vs. 11.4±1.1%, P<0.01) and even lower in fetuses with depressed T₃ levels (3.1±0.6%, P<0.001 vs. control; Fig. 3A). The proportion of mononucleated myocytes positive for Ki-67 did not differ between T₃ and TX fetuses. The percentage of phospho-histone 3-positive (Fig. 3B) mononucleated myocytes followed the same trend. In T₃-infused fetuses, 0.8 ± 0.2% of mononucleated cardiomyocytes were phospho-histone 3 positive vs. 1.2 ± 0.1% in controls (P<0.05), and in thyroidectomized fetuses 0.4 ± 0.2% were positive (P<0.01). Consistent with this change, p21 protein levels were increased and cyclin D1 protein decreased in hearts from T₃-infused fetuses compared to control and TX (P<0.01; Fig. 4). However, no difference was found in protein levels for either p21 or cyclin D1 in TX hearts compared to control.

Figure 3.

A) Cell cycle activity, measured by Ki-67, decreased in both T₃-infused and TX fetuses. A1) Image at ×400. B) Mitotic activity of myocytes, measured by phospho-histone 3 expression, was decreased in both T₃-infused and TX fetuses. B1) Image at ×400. Arrows indicate positively labeled cardiomyocytes. Data are means ± se; n = 8/group. *P < 0.05; **P < 0.01 vs. control.

Figure 4.

A) Immunoblots of cell cycle proteins in cardiomyocytes from T₃-infused and TX fetuses compared to controls. B) Cell cycle suppressant p21 was markedly elevated in T₃-infused compared to control and TX fetuses. C) Expression of cell cycle promoter cyclin D1 is diminished in T₃-infused compared to control and TX fetuses. Data are means ± se; n = 8/group. **P < 0.05 vs. control; ⁺⁺P < 0.01 vs. T₃.

Cardiomyocyte size

LV data are shown in Table 3. Cell measurements in the RV were similar to LV measurements and are not included in the table. No differences were found in the lengths of mononucleated or binucleated cardiomyocytes among the three groups (Table 3). However, the widths of both mononucleated and binucleated LV myocytes from T₃-treated fetuses were larger than those from control fetuses (P<0.05 and P<0.001, respectively; Table 3). T₃-infused fetuses also had wider mononucleated RV myocytes compared to control (12.9±0.3 vs. 11.5±0.1 μm, P<0.05), while thyroidectomy had no effect on mononucleated RV cardiomyocyte width. Binucleated cardiomyocytes from both ventricles of T₃-infused and TX fetuses were wider than control myocytes (Fig. 5). Binucleated LV myocytes were also wider in T₃-infused fetuses compared to TX fetuses (P<0.05; Table 3).

Table 3.

Myocyte measurements with reference to thyroid status

Measurement	Thyroid status
	TX	Control	T3 infusion
Mononucleated LV myocytes
Ki67+ (%)	3.1 ± 0.6***	11.4 ± 1.1	6.1 ± 0.7**
Phospho-histone 3+ (%)	0.4 ± 0.2**	1.2 ± 0.1	0.8 ± 0.2*
Dimensions
Length (μm)	73.0 ± 0.7	70.7 ± 0.9	72.3 ± 1.2
Width (μm)	10 ± 0.2*	9.9 ± 0.2	10.8 ± 0.2*
Volume (μm3)	5685 ± 309*	4608 ± 284	5725 ± 279*
Binucleated LV myocytes
Percentage	21.4 ± 4.1***,+++	42.6 ± 1.1	56.8 ± 2.4**
Dimensions
Length (μm)	91.3 ± 1.0	88.5 ± 1.6	91.1 ± 1.4
Width (μm)	12.6 ± 0.3*,+	11.6 ± 0.2	13.7 ± 0.3***
Volume (μm3)	10,000 ± 635***	8052 ± 31	11,330 ± 507*

Data are means ± se; n = 8/group. Data from left ventricle are shown; trends are the same in the right ventricle. Significant differences between groups are indicated.

^*P < 0.05,

^**P < 0.01,

^***P < 0.001 vs. control;

⁺P < 0.05,

⁺⁺⁺P < 0.001 vs. T₃.

Figure 5.

T₃ promotes cardiomyocyte hypertrophy. Binucleated cardiomyocyte width was increased in left (A) and right (B) ventricles from fetuses that were either deficient or had elevated circulating T₃. Both elevated T₃ levels and T₃ deprivation led to widening of ventricular binucleated myocytes. Data are means ± se; n = 8/group. *P < 0.05, ***P < 0.001 vs. control; ⁺P < 0.05 vs. T₃.

Molecular markers of cardiomyocyte maturation and hypertrophy

Figures 6 and 7 show expression levels of specific molecular markers in 3 fetal groups. In the high-T₃ group, the expression level of ANP was increased significantly compared to control (P<0.05; Fig. 6A), while no difference was found between TX and control fetuses. SERCA2a expression was related to the concentration of circulating T₃ across the groups. Compared to control hearts, SERCA2a expression was increased markedly in T₃ fetuses (P<0.001; Fig. 6B) and decreased in TX fetuses (P<0.05 vs. control; P<0.01 vs. T₃). Figure 7 shows that the phosphorylated form of ERK and mTOR protein increased with increasing levels of T₃ (P<0.05 vs. control, P<0.01 TX vs. T₃) with no significant differences between TX and control fetuses. No significant difference was found among the treatment groups for pAkt and p-p70^S6K (data not shown).

Figure 6.

T₃ promotes molecular markers of cardiomyocyte maturation. A) Expression of ANP was increased in the left ventricles of fetuses receiving exogenous T₃, but thyroidectomy did not alter ventricular ANP. B) SERCA2a expression levels were lower in TX fetal hearts and higher in T₃-infused hearts compared to controls. Data are means ± se; n = 8/group. *P < 0.05 vs. control; ⁺⁺⁺P < 0.001 vs. T₃.

Figure 7.

T₃ activates ERK and mTOR. Immunoblots showed elevated levels of phosphorylated ERK (A) and mTOR (B) in fetal hearts exposed to elevated levels T₃, but phosphoprotein levels were unchanged in hearts from TX fetuses. Data are means ± se; n = 8/group. *P < 0.05 vs. control; ⁺⁺P < 0.01 vs. T₃.

DISCUSSION

We sought to determine the degree to which varying plasma levels of T₃ affect the maturation of fetal ovine cardiomyocytes in vivo because we suspected the importance of T₃ in regulating myocardial growth. In one group of fetuses, we increased the plasma levels of T₃ by 10-fold for 5 d. Fetuses in this group were 125–130 dGA, ∼2 wk before they would experience the normal-term T₃ surge. The concentration range achieved in this group was similar to that reached naturally in fetuses at term during the normal TH surge. In a second group of fetuses, we removed the fetal thyroid gland at a comparable gestational age, and T₃ levels dropped to immeasurable levels within 4 d of surgery. A third group of intact fetuses received a vehicle infusion, had normal T₃ levels, and served as a control group to both the treated groups. We compared indices of cardiomyocyte maturation among groups.

Cardiomyocyte maturation in a high-TH environment

We hypothesized that elevated levels of T₃ in the fetal sheep would stimulate the maturation of the myocardium and would mimic the normal changes that ordinarily begin some 15 d later in gestation. We compared 3 features of cardiomyocyte maturation as a primary test of the hypothesis: increased cardiomyocyte binucleation, suppression of cardiomyocyte proliferation, and enlargement of cardiomyocytes. All three of our criteria for maturation changed in the directions the hypothesis predicted. The left ventricles of fetuses whose plasma T₃ levels were elevated had a higher fraction of binucleated cardiomyocytes in the myocardium of 56.8 ± 2.4 vs. 42.6 ± 1.1% in age-matched controls; a cardiomyocyte cell cycle activity index that diminished by 46%, as shown by decreased levels of Ki-67; and an increase by some 14% in the width of cardiomyocytes that contained 2 nuclei. Cardiomyocytes from the right ventricle exhibited similar trends. The two cell cycle regulatory proteins that were analyzed also changed as predicted: protein levels of the cell cycle promoter cyclin D1 were depressed by 150%, and the cell cycle suppressor protein p21was elevated 5-fold.

We posit that cardiomyocytes from the high-T₃ animals enlarged not because of mechanical stimulation in the form of wall stress but because cardiomyocytes were stimulated by the hypertrophic actions of T₃. While the hormonal effect may have been bolstered slightly by the elevated mean arterial pressure of ∼2.0 mmHg on d 5 in the T₃ infusion animals compared to the other groups, the load stimulus would likely have been small. Thus, it is likely that the hypertrophic response that we observed in the first group over 5 d of load was due to a direct effect of T₃ on cardiomyocytes, as shown in adult cardiomyocytes (52–54). The expression of ANP was increased in response to elevated levels of T₃. The peptide is expressed normally in all four chambers during fetal life but increases with maturation and/or with elevated loading conditions. Thus, the elevated levels in the ventricles may represent elevated maturation as well as increased loading, since mean arterial pressure was slightly elevated on the last day of T₃ infusion.

In the immature mononucleated myocyte, the sarcoplasmic ATPase SERCA2a removes Ca²⁺ from the cytoplasm during the diastolic period of each beat. As cardiomyocytes become enlarged during the maturation process in utero, they develop an extensive SR in concert with a mature t-tubule system (55, 56). Gradually, the source of activator calcium ion switches from an extracellular fluid to the internal SR. However, as the myocytes enlarge during the maturation process, new SERCA2a protein is manufactured and installed on the internal SR membrane to rapidly remove Ca²⁺ from the cytoplasm during the diastolic phase of the cardiac cycle. This study took advantage of the fact that SERCA2a expression is T₃ sensitive (23), as it normally rises toward term. In these experiments, SERCA2a expression was increased in the myocardium of fetuses exposed to elevated T₃ and was suppressed in hearts not exposed to circulating T₃.

Cardiomyocyte maturation in a low-TH environment

In thyroidectomized fetuses, plasma levels of T₃ decreased to values that were undetectable by the assay. In this group, the percentage of cardiomyocytes with 2 nuclei (21.4±4.1%) was significantly lower than in both vehicle (42.6±1.1%) and T₃ (56.8±2.4%) groups. This finding is in keeping with our maturation hypothesis and suggests that hearts growing in a low-TH environment had suppressed maturation. The hearts from thyroidectomized animals had fewer myocytes that labeled positive for Ki-67 or phospho-histone 3 than did control hearts, indicating that cardiomyocyte proliferation was depressed. We had reasoned that lower than normal levels of T₃ would remove the T₃-induced brake on proliferation and would allow it to increase. This was not the case. The explanation for this finding is uncertain. However, THs are known to be required for basal cell cycle activity in several cell types, including adult neural stem cells and hematopoietic progenitors (57, 58). Thus, one explanation for the suppression of proliferation in the absence of TH is that fetal cardiomyocytes require a minimal level of TH to maintain normal mitotic activity. Note that unlike hearts exposed to high T₃ levels, depressed mitotic activity in the absence of TH occurred without apparent changes in the expression of the cyclin D1 or p21 proteins. This finding indicates that the suppression of mitotic activity in the absence of TH involves a signaling process different from the one used to suppress mitotic activity when T₃ levels are elevated.

In the studies reported here, the widths of fetal binucleated cardiomyocytes increased in two very different hormonal environments, high plasma levels of T₃ and low T₃ levels. The potential explanation requires speculation. As noted above, the hearts from thyroidectomized fetuses had the lowest percentage of cardiomyocytes that were proliferating and that were binucleated. Thus, it is likely that the myocardium in a thyroid deficient fetus would be underendowed. If thyroidectomy led to a thinning of the ventricular free wall, its wall stress would have increased.

According to the Laplace relationship (59):

$$S_w=P_w/2\times r/h$$

where S_w is wall stress, P_w is transmural pressure, r is meridional radius of curvature, and h is average free wall thickness. From this relationship, one can see that, with a constant transmural pressure and radius, wall stress will increase if wall thickness decreases. If wall stress did increase, cardiomyocyte hypertrophy would surely follow (14, 60). Wall thickness was not measured in these hearts, so this speculation awaits further experimental evidence. We have noted in past experiments that binucleated cardiomyocytes are more likely to enlarge than mononucleated cells would when wall stresses are elevated (14), but in these experiments both mono- and binucleated cells were larger following T₃ deprivation and/or supplementation.

Signaling markers of maturation under differing T₃ conditions

TH is known to stimulate cardiomyocyte hypertrophy in adult humans and in neonatal rodent models (61–63). Activation of PI3K, Akt, and mTOR is known to result in hypertrophy in neonatal rat hearts following T₃ or T₄ administration both in vitro and in vivo (63, 64). The MAP kinase cascade and ERKs have been implicated in many aspects of cardiac growth, including hyperplastic growth during the proliferative stage of heart development and hypertrophic growth that characterizes the growth of terminally differentiated cardiomyocytes (65–67). In this study, T₃ was shown to stimulate phosphorylation of ERK and mTOR in the myocardium. Coupled with increases in myocyte size, these data suggest that T₃ promotes hypertrophic growth in fetal cardiomyocytes through recognized signaling pathways.

As discussed above, p21Cip1/Waf1 is a member of a family of transcriptionally regulated cell-cycle inhibitors whose loss is often associated with increased proliferation and cancer (68). The relationship of ERK activation and p21 expression has not been examined in fetal cardiac cells. The Ras/ERK signals that are required for proliferation have been implicated in cell cycle arrest via p21 (69) and cardiomyocyte hypertrophy (65, 70, 71). Therefore, Ras/ERK signals can have both proliferative and antiproliferative actions (72), even though their proproliferative actions are better described and better known. In some cell types, the ERK dependency for both opposing pathways can be explained by the differential requirement for transient vs. sustained ERK activation (73). In other cell types, sequestration of ERK signaling streams via scaffold proteins may allow for the concurrent activation of distinct ERK signaling cascades within the same cells that achieve distinct outcomes (74–76). However, it is not known whether elevations in p-ERK are related to p21 regulation in ovine fetal cardiomyocytes.

ANP and SERCA 2a are sensitive to T₃ levels

ANP is a well-accepted marker of cardiomyocyte hypertrophy in adult myocardium (56). The relative expression of the ANP gene in the heart was increased by T₃ but unchanged in TX fetuses. As mentioned above, the increase was most likely a normal response to cardiomyocyte maturation. We have recently shown that cardiomyocyte proliferation is suppressed following the generation of intracellular cGMP in response to the binding of the transmembrane guanylyl cyclase, ANP receptor (NPRA; ref. 33). Thus, the elevated levels of ANP in hearts exposed to high levels of T₃ would have suppressed the proliferation of cardiomyocytes along with the actions of T₃. The fact that ANP did not increase in the low-TH environment further suggests that the cardiomyocytes in those hearts remained in an immature state.

Clinical application

Studies in animals and humans associate low maternal and fetal TH levels with suppressed neurodevelopment in offspring (39, 77). The fact that both hyper- and hypothyroid conditions lead to detrimental fetal outcomes suggests that an optimal range of TH levels must support the appropriate growth and maturation of the myocardium. If the maturation process is completed before the appropriate number of cardiomyocytes is generated, the heart could suffer a deficit in cardiomyocyte endowment at birth. That the ovine fetal heart can be underendowed has been demonstrated in placental insufficiency studies (78, 79) and experimental reduction in systolic pressure (80). Hearts with fewer than normal numbers of myocytes are expected to be more vulnerable to heart failure or other detrimental outcomes in adult life (81, 82). Data from experiments in rats suggest that cardiomyocyte numbers cannot recover during postnatal life following episodes of hypoxic stress (83, 84) or vitamin D deficiency in the womb (85).

Maternal and fetal TH levels

In normal pregnancies, TH molecules in the fetal plasma during the last trimester come from both the fetal thyroid gland and the mother. In humans, but not in sheep, elevated maternal plasma levels of TH lead to elevated fetal TH levels (86, 87). Thus, maternal hyperthyroidism leads to fetal hyperthyroidism. The monocarboxylate transporter, which transports T₃ and T₄, has been identified in human placentas (88) and appears to be the primary placental transporter for THs. We studied the effects of doses of T₃ that led to plasma concentrations normally found in human and sheep fetuses in a later stage of their development. Maternal levels of circulating total T₄ (∼110 ng/ml) and circulating total T₃ (∼2.0 ng/ml) are elevated during normal pregnancy (89, 90). Normal human umbilical artery levels of T₄ (90.5±0.02 ng/ml) and T₃ (1.9±0.96 ng/ml) at birth are similar to fetal sheep levels (8, 91).

CONCLUSIONS

The present study suggests that the final maturation stages of the myocardium are regulated by conserved hormonal systems that are known to be phylogenetically ancient. When preterm cardiomyocytes were exposed to term levels of circulating T₃ for several days, they developed phenotypic features of maturity ordinarily seen in cardiomyocytes some 2 wk later. On the one hand, our data are consistent with the conclusion that T₃ serves as the primary driver of fetal cardiomyocyte maturation and powerfully reduces the generative capacity of the cardiomyocyte population near the time of birth. On the other hand, when circulating T₃ levels are severely reduced, several features of cardiomyocyte growth and maturation suffered. These studies show for the first time that the normal development of the near-term fetal heart requires T₃ concentrations to be maintained within a defined range. Over the past decade, it has become clear that gestational adversities may reduce heart cell numbers for life and increase vulnerability for chronic disease. Consequently, human hearts exposed to elevated T₃ concentrations before birth might be compromised for life if they respond to T₃ in a similar fashion to sheep.