Unexpected maturation of PI3K and MAPK-ERK signaling in fetal ovine cardiomyocytes

Abstract

In the first two-thirds of gestation, ovine fetal cardiomyocytes undergo mitosis to increase cardiac mass and accommodate fetal growth. Thereafter, some myocytes continue to proliferate while others mature and terminally differentiate into binucleated cells. At term (145 days gestational age; dGA) about 60% of cardiomyocytes become binucleated and exit the cell cycle under hormonal control. Rising thyroid hormone (T₃) levels near term (135 dGA) inhibit proliferation and stimulate maturation. However, the degree to which intracellular signaling patterns change with age in response to T₃ is unknown. We hypothesized that in vitro activation of ERK, Akt, and p70^S6K by two regulators of cardiomyocyte cell cycle activity, T₃ and insulin like growth factor-1 (IGF-1), would be similar in cardiomyocytes at gestational ages 100 and 135 dGA. IGF-1 and T₃ each independently stimulated phosphorylation of ERK, Akt, and p70^S6K in cells at both ages. In the younger mononucleated myocytes, the phosphorylation of ERK and Akt was reduced in the presence of IGF-1 and T_3. However, the same hormone combination led to a dramatic twofold increase in the phosphorylation of these signaling proteins in the 135 dGA cardiomyocytes—even in cells that were not proliferating. In the older cells, both mono- and binucleated cells were affected. In conclusion, fetal ovine cardiomyocytes undergo profound maturation-related changes in signaling in response to T₃ and IGF-1, but not to either factor alone. Differences in age-related response are likely to be related to milestones in fetal cardiac development as the myocardium prepares for ex utero life.

the number of myocytes in the heart is set during the perinatal period (23). At birth time, some 60–70% of ovine cardiomyocytes have gained a second nucleus and no longer divide (23) through a process known as terminal differentiation (10, 37). In fetal sheep terminal differentiation begins at about 110 days gestational age (dGA) and continues to birth (145 dGA) (4, 23). Before 110 days, mononucleated fetal cardiomyocytes divide rapidly to build the cellular endowment of the growing myocardium (23). The process of ovine “terminal” cardiomyocyte maturation is thus predominantly prenatal and is thought to be similar to human heart development (1, 40). This precocious growth pattern differs from that of the altricial rat and mouse heart in which terminal differentiation occurs after birth (22, 41, 42).

The number of cardiomyocytes contained within the heart at birth can vary according to prenatal hormonal and mechanical conditions (1, 6, 15, 21, 23, 28, 31, 44, 48, 51). Among the many regulators, insulin-like growth factor 1 (IGF-1) (44) powerfully stimulates proliferation rates of ovine cardiomyocytes and is the most powerful factor in promoting robust growth in the heart. In opposition to IGF-1, thyroid hormone is a powerful antiproliferation hormone that becomes increasingly effective as gestation proceeds (7). Thus the balanced actions of these two hormones determine the cardiomyocyte endowment at birth.

The ovine thyroid gland begins to secrete thyroxine (T₄) around 50 days gestational age (18, 47) and its plasma concentration rises over the first two-thirds of gestation. However, the affinity of T₄ for the alpha (TRα1) and beta (TRβ1) thyroid hormone receptors in fetal myocardium and other organs (24, 26, 29), is some 10–15 times less than it is for 3,3′,5-tri-iodo-l-thyronine (T₃) (9). Fetal T₃ levels increase some 10-fold as T₄ is converted to T₃ under the influence of ever increasing deiodinase activity in peripheral tissues during the last few weeks of gestation (12, 13, 35).

We previously reported that physiological concentrations of T₃ are powerful suppressants of serum-stimulated proliferation among isolated cardiomyocytes from mid- and late-gestation (7, 8). The inhibitory effect is associated with increases in the cell cycle suppressant protein, p21, and decreased expression of the cell cycle protein, cyclin D1. T₃ was found to drive phenotypic and physiological maturation of cardiomyocytes including augmenting Ca²⁺ transporter expression and maturation of the atrial natriuretic peptide system in vivo (6). These findings made it clear that slowing of cardiomyocyte proliferation during late gestation depends largely on the stimulatory influences of IGF-1 that gradually become antagonized as T₃ concentrations increase in the myocardial environment near the end of gestation.

There is evidence that cardiomyocyte sensitivities to stimulatory and inhibitory actions of hormones are continuously changing as cardiomyocytes mature over the course of gestation. In late-gestation ovine cardiomyocytes, the pro-proliferative action of IGF-1 requires activation of both the mitogen activated protein kinase (MAPK) pathway via the extracellular signal related kinase (ERK) cascade and the phosphoinositol 3 kinase (PI3K) pathway, including Akt (44). The basic signaling relationships in the ovine cardiomyocyte are shown in Fig. 1. The binding of IGF-1 to IGF1R leads to activation of both PI3K/Akt pathway (left) and the MAPK/ERK pathway (center). The downstream target, p70S6 kinase (p70^S6K), can be independently activated via either the MAPK/ERK or the PI3K/Akt/mammalian target of rapamycin complex 1 (mTORC1) pathway (39, 49). T₃, which binds both of the two thyroid hormone receptors (TR), TRα and TRβ, stimulates many pathways in common with IGF-1 including activation of both the PI3K pathway and the MAPK pathway. T₃ has also been shown to interact with Ras (14, 16, 17) but its actions in the immature cardiomyocyte are not fully understood (indicated by the black dashed arrow in Fig. 1).

Fig. 1.

A hypothetical signaling mechanism for T₃ and IGF-1 in fetal sheep cardiomyocytes. T₃, 3,3′,5-tri-iodo-l-thyronine; IGF-1, Insulin-like growth factor-1; IGFR, IGF receptor; IRS, insulin receptor substrate; MEK, mitogen-activated protein kinase kinase; ERK, extracellular signal-related kinase; PI3K, phosphoinositol-3-kinase; Akt; protein kinase B; mTORC1, mammalian target of rapamycin complex 1; TR, thyroid hormone receptor; AP-1, activator protein 1; Elk-1, ETS domain-containing protein.

Because IGF-1 and T₃ are key hormones that regulate the growth and maturation of the fetal myocardium in late gestation and because both hormones are known to signal through the ERK and Akt pathways, we sought to determine whether there were important maturation-related changes in their signaling as gestation proceeds. We tested the null hypothesis that maturational activation of ERK and Akt pathways under the influences of combined T₃ and IGF-1 would be similar between cardiomyocyte populations isolated from fetuses of 100 dGA and 135 dGA.

MATERIALS AND METHODS

Animals.

Animals studies were reviewed and approved by the Institutional Animal Care and Use Committee at Oregon Health and Science University (Portland, OR). Primary cultures of cardiomyocytes were obtained from noninstrumented, control fetal sheep (Ovis aries; mixed Western breed) at 101 ± 3 dGA (n = 7 for BrdU and Western blots; n = 4 for immunohistochemistry) and 135 ± 1 dGA (n = 7 for BrdU and Western blots; n = 4 for immunohistochemistry) where term is ∼145 dGA. Each n represents a separate fetus, not the average of replicates. The fetal cardiac myocytes that comprise the myocardium at ∼100 dGA are phenotypically homogeneous and are all mononucleated while myocytes in the 135 dGA heart are ∼40% mononucleated (23).

Materials.

3,3′,5-tri-iodo-l-thyronine (T₃; thyroid hormone), 5-bromo-2′-deoxyuridine (BrdU), insulin-transferrin-sodium selenite (ITSS), antibiotic-antimycotic solution (penicillin-streptomycin-amphotericin B, PSA), Type XIV protease, and laminin were obtained from Sigma-Aldrich (St. Louis, MO). Type II collagenase was obtained from Worthington Biochemicals (Lakewood, NJ). PI3K inhibitor LY294002 (LY), the MEK inhibitor U0126 (U0), and the mTORC1 (referred to as mTOR hereafter) inhibitor rapamycin (R) were from EMD Millipore (Billerica, MA). IGF-1 circulates largely bound to a family of at least 6 binding proteins in serum with only a small fraction existing in the “free” form. To determine the actions of free IGF-1 unassociated with binding protein, we used an IGF-1 analog Long R3 IGF-1 (referred to as IGF-1 hereafter, GroPep, Australia), which binds binding proteins with much lower affinity (200–1,000 fold) than endogenous IGF-1 (11, 44).

Antibodies for Western blot analyses.

The following antibodies were obtained from Cell Signaling (Danvers, MA.). 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. Akt rabbit polyclonal antibody detects endogenous levels of total Akt1, Akt2, and Akt3 proteins. Phospho-p70 S6 kinase (Thr389) rabbit monoclonal antibody detects endogenous levels of p70 S6 kinase only when phosphorylated at Thr389. This antibody also detects p85 S6 kinase when phosphorylated at the analogous site (Thr412). Total p70 S6 kinase antibody detects endogenous levels of the protein. This antibody also recognizes p85 S6 kinase. Rabbit monoclonal α-tubulin detects endogenous levels of total α-tubulin protein, and does not cross-react with recombinant β-tubulin. Mouse monoclonal ERK2 antibody recognizes COOH terminus of MAP kinase p42 (Santa Cruz Biotech, Santa Cruz. CA). Horseradish peroxidase linked secondary antibodies (goat anti-rabbit IgG and horse anti-mouse IgG) were obtained from Cell Signaling.

Cardiac myocyte isolation.

Ewes were euthanized by intravenous injection of a commercial solution of sodium pentobarbital (SomnaSol, ∼80 mg/kg, Butler Schein Animal Health, Dublin, OH). Fetuses received a bolus dose of heparin [5,000 U (100 dGA) or 10,000 U (135 dGA), Baxter, Deerfield, IL], followed by 5 ml (100 dGA) or 10 ml (135 dGA) of saturated potassium chloride (KCl) into the umbilical vein to arrest the heart in diastole. Fetuses were weighed and the heart excised, trimmed in a consistent manner. The ascending aorta and main pulmonary artery were trimmed at the level of the bifurcation for each heart, blotted, and weighed. Hearts were enzymatically dissociated as previously described by our laboratory (20). Briefly, hearts were perfused in a retrograde manner with gassed solutions (95% O₂ and 5% CO₂, 39°C); Tyrode buffer for 5 min [(no calcium added); 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂·6H₂O, 10 mM glucose, 10 mM HEPES; pH adjusted to 7.35 with NaOH] until the vessels were clear of blood; ∼2 min (for 100 dGA heart) or ∼8 min (for 135 dGA hearts) with 160 U/ml Type II collagenase and 0.78 U/ml Type XIV protease in Tyrode buffer to digest the tissue; last, ∼5 min with a high-potassium, 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, 10 mM glucose; pH adjusted to 7.37 using KOH). The left and right ventricular (LV, RV) free walls were dissected from the heart and placed into separate tubes containing 20 ml KB solution. The tissue was gently agitated to release cardiomyocytes. We achieve a cell purity of about 95% cardiomyocytes by our isolation methods; cardiomyocyte concentrations are further enriched by two preplating steps to remove fibroblasts before culturing the myocytes. Myocyte identity was confirmed by staining for myosin heavy chain α/β (See BrdU uptake analysis).

Cardiomyocyte cultures.

Cardiomyocytes were cultured as previously described by our laboratory (7, 31). The freshly isolated slurry of myocytes rested for 30–60 min at room temperature before centrifugation and resuspension in sterile serum media (10% fetal bovine serum, FBS, Invitrogen, Carlsbad, CA); DMEM low glucose (5.56 mM d-glucose, 4 mM l-glutamine, 1 mM sodium pyruvate, 5.33 mM KCl, 0.4 mM glycine, pH 7.4, Invitrogen, no. 11885-084 supplemented with 10 ml/l of ITSS, and 10 ml/l PSA). All cultures were performed at 39°C, 95% air, 5% CO₂. Cells were preplated twice to remove nonmyocyte cells (2 h each time). Myocytes were then seeded onto sterile 22 × 22 mm glass coverslips or 6-well plates as described in the following section. These were coated with laminin (4 mg/ml) for at least 4 h and aspirated just before plating at a density of 1.5 × 10⁵ cells per coverslip and 5 × 10⁵ cells (100 dGA) or 1 × 10⁶ (135 dGA) cells per well. Cardiomyocytes were incubated in serum media for 24 h. Cells were then incubated in serum-free media (SF) for 48 h, and the media changed again to fresh SF media for 24 h before treatment; treatment commenced on culture day 5 for all experiments.

Culture conditions and treatments.

Using the PI3K inhibitor (LY), the MEK inhibitor (U0), and the mTOR inhibitor (R) we determined the extent to which MAPK and PI3K constituent members participates in the downstream effects of T₃ in fetal cardiomyocyte growth suppression. Cells from each age (7, 8) were incubated under these conditions: 1) SF, 2) IGF-1 (1 μg/ml) in SF, 3) SF + IGF-1 + T₃ (0.37 nM), 4) SF + IGF-1 + T₃ (1.5 nM), 5) SF + IGF-1 + T₃ (1.5 nM) + U0 (10 μM), 6) SF + IGF-1 + T₃ (1.5 nM) + LY (10 μM), and 7) SF + IGF-1 + T₃ (1.5 nM) + R (10 nM). Cells used for the analysis of BrdU uptake were cultured on sterile 22 × 22 mm coverslips and the above conditions cocultured with BrdU (10 μM) for 48 h. Cells for signaling studies were cultured on either 6-well plates (Western blot analysis) or onto 22 × 22 mm coverslips (immunohistochemistry). We chose to treat cells in media lacking serum so that we would have the IGF-1 proliferative stimulus exclusively. The concentration of IGF-1 used in this study was derived from previous work by us (44) where a vigorous dose response was performed in 135 dGA cells and confirmed again for this study. In addition, we have measured the dose effect for the younger cells and while a much larger proportion of cells are able to proliferate, it is otherwise similar to the 135 day cardiomyocytes. Concentrations of T₃ represent the spectrum of total T₃ levels in the last one-third of gestation with 0.37 nM approximating the normal circulating levels near 100 dGA and 1.5 nM the concentration that is found prior to birth (35, 36, 50). Concentrations of pathway inhibitors were also based on our previous studies (44) according to the pharmacological properties of the compounds.

BrdU uptake analysis.

At the end of the culture treatment period for BrdU analysis cardiomyocytes were fixed in ice-cold acidified ethanol for staining procedures. Cells were permeabilized by incubating with 5 μg/ml DNase as previously described (45) for 30 min at 37°C. Fixed cultured cardiomyocytes were double-stained with mouse anti-myosin heavy chain α/β(1:5,000, ab15, Abcam, Cambridge, MA) and rat anti-BrdU (1:500, ab6326, Abcam) antibodies overnight at 4°C. This method identifies BrdU positive cardiomyocytes. Following washes with 1X phosphate-buffered saline (PBS), cells were incubated in anti-mouse rhodamine red (1:200, Jackson ImmunoResearch, West Grove, PA), and anti-rat FITC (1:200, Jackson ImmunoResearch, West Grove, PA) secondary antibodies for 2 h at room temperature (RT). Coverslips were mounted onto slides using Vectashield Hardset mounting medium with DAPI (Vector Laboratories, Burlingame, CA) and stored in the dark overnight at 4°C to allow the mounting medium to set (7). The portion of BrdU positive myocytes was determined in a random sample of 300 (minimum) myocytes using fluorescence microscopy (400 × magnification; Zeiss Axiophot, Bartels and Stout, Bellevue, WA) with FITC (excitation 450–490 nm, emission 515–565 nm) and TRITC filters (excitation 485 nm; emission 515–530 nm). BrdU positive cardiomyocytes were identified as those that stained positive for both myosin (rhodamine) and punctuate nuclei (FITC) indicating BrdU incorporation.

Thyroid hormone signaling.

We determined the rapid activation (10 min) profile of ERK, Akt and p70^S6K by phosphorylation (pERK, pAkt, p-p70^S6K) by Western blot in fetal cardiomyocytes (100 dGA and 135 dGA). To isolate the activity of signaling cascades apart from active molecules in serum, studies were performed in serum-free media (5, 19, 27). Higher pharmacological concentrations of T₃ (7, 8) and IGF-1 (44) did not significantly increase pERK or pAkt activity; thus the stimulatory doses we used were on the activity plateau while remaining within the relevant physiological range for both ages (32, 33, 35). The cells for Western blot analysis were incubated with T₃ alone (1.5 nM) and in combination with the pathway inhibitors (U0, LY, R; pretreatment for 20 min) in SF media for 10 min at 39°C. To complement protein analysis, passage 1 cells from the LV, as previously described (31), were plated onto coverslips and treated in a similar manner prior to fixation.

Western blot analysis.

After the appropriate incubation period, media was aspirated, cardiomyocytes were rinsed in ice-cold 1X PBS, lysed [5 mM Tris-HCl, 5 mM EGTA, 5 mM EDTA, 0.06% SDS, protease inhibitor Mini-complete tablet (Roche, IN), and phosphatase inhibitor cocktail I and II (Sigma-Aldrich)], and collected into prechilled tubes. Protein concentration was quantified by BCA assay (Pierce, Rockford, IL). Equal amounts of total protein/sample (10 μg) were separated by SDS-PAGE on a 10% Tris-glycine gel and transferred to a nitrocellulose membrane (Optitran BA-S 83, Whatman, NJ). Membranes were blocked with 5% milk in 1X Tris-buffered saline + 0.01% Tween 20 (TBS-T) buffer for 1 h at RT. Membranes were incubated with primary antibodies (1:1,000) overnight at 4°C in 4% bovine serum albumin (Sigma-Aldrich) in 1X TBS-T buffer. Membranes were washed in large volumes of TBS-T before exposure to the secondary antibody (1:5,000) in 5% milk-TBS-T for 1 h at RT. Antibody binding was detected using chemiluminescence (SuperSignal, Pierce, IL); protein expression was quantified from a digitized image of the blot using NIH ImageJ (version 1.4; NIH). Signal density was normalized to total protein of interest and expressed as phospho-protein/total protein (ERK, Akt, p70^S6K). Alpha-tubulin staining was used to normalize for loading.

Immunohistochemistry for pERK and pAkt.

After the treatment period, media was aspirated and cells were fixed using fresh 4% paraformaldehyde (Sigma-Aldrich) for 15 min at room temperature. Fixative was aspirated and slides rinsed 2 times in 1X PBS. Slides were postfixed in ice-cold 100% methanol. Fixed cardiomyocytes were incubated with rabbit anti-pERK (1:200) or rabbit anti-pAkt (1:200) antibodies overnight at 4°C. Following PBS washes, cells were incubated in anti-rabbit secondary antibody (1:200) for 1 h at room temperature (RT). Samples were then incubated with avidin/biotin complex (1:200 in PBS, Vectastain ABC Kit, VectorLabs, CA) for 1 h at room temperature. Positive signal were stained with 3,3′-diaminobenzidine chromagen (DAB) in substrate buffer (DakoCytomation, CA) for 2 min and cells were counterstained with 0.1% methylene blue. The proportion of pERK or pAkt positive myocytes was determined in a random sample of 300 (minimum) myocytes light microscopy (400 × magnification; Zeiss Axiophot, Bartels and Stout, Bellevue, WA). The number of binucleated cells was noted for each slide and positive mononucleated and binucleated cells were scored separately. Data are presented as percentage of positive mononucleated and binucleated cells in the 300 cells/slide. Three hundred was determined to be the optimal number required to obtain a constant SD. Binucleated cardiomyocytes were not detected in cardiomyocytes sampled from the 100 dGA group.

Statistical analysis.

Only data from left ventricular cells are reported because responses in the right ventricle were similar and replication offered no new information. On average, BrdU incorporation was slightly lower in RV, but the trends between ventricles were otherwise similar. A minimum of 300 cells per treatment group were assessed for BrdU incorporation and reported as a percent of total number of cardiomyocytes counted. A minimum of 300 cells per treatment group were also assessed for pERK and pAkt positive cardiomyocytes and reported as a percent of total mononucleated or binucleated cell. One-way analysis of variance (ANOVA) was used for all analyses. If justified by ANOVA, differences were further analyzed by Tukey's multiple comparison post hoc test for differences between treatment groups. Statistical significance was set at P < 0.05. Data are presented as means ± SE. The 100 dGA cells were analyzed separately from 135 dGA cells. We performed two-way analyses across gestational age for BrdU studies. BrdU uptake was found to be greater in the younger cardiomyocytes as we have previously published. The interaction using age and treatment has a P value = 0.049 with a greater percentage of change attributed to the effect of age (50% for age vs. 24% for treatment).

RESULTS

BrdU uptake.

Figure 2 shows that BrdU incorporation, an index of cell proliferation, was found in a higher portion of the population in the younger 100 dGA cardiomyocytes (Fig. 2A) than in older 135 dGA myocytes (Fig. 2B) under serum-free conditions. In the presence of IGF-1, proliferation rates nearly doubled from ∼18% to 33% in 100 dGA cardiomyocytes (Fig. 2A) and nearly quadrupled in the 135 dGA cardiomyocytes, from ∼2% to ∼8% (Fig. 2B). In cells exposed to combined IGF-1 and T₃, BrdU incorporation rates decreased by 43% in younger myocytes (Fig. 2A) and 75% in older myocytes (Fig. 2B). As we have reported in previous studies, T₃ was shown to have no influence on BrdU uptake under the nonphysiological serum-free condition at either age (7, 8).

Fig. 2.

BrdU uptake in proliferating fetal sheep cardiomyocytes in vitro. After 48 h incubation with IGF-1 (1 μg/ml) and 5-bromo-2′-deoxyuridine (BrdU; 10 μM), IGF-1 increases BrdU uptake in both 100 days gestation (dGA) (A) and 135 dGA cardiomyocytes (B). T₃ at 1.5 nM does not alter BrdU uptake in basal conditions. At both ages, addition of T₃ (0.37 nM, 1.5 nM) inhibits IGF-1-induced BrdU uptake; this was a 30% reduction in 100 dGA cells and a 75% reduction in 135 dGA. SF, serum free. Data are means ± SE; each bar n = 7. ***P < 0.001 vs. SF; +P < 0.05, ++P < 0.01, +++P < 0.001 vs. IGF-1.

Role of MAPK and Akt pathways in proliferation.

Figure 3 shows BrdU uptake in the 100 dGA cardiomyocytes in the presence of the MAPK inhibitor U0126 (10 μM; Fig. 3A), Akt inhibitor LY294002 (10 μM; Fig. 3B), and mTOR inhibitor, rapamycin (10 nM; Fig. 3C) during treatment with IGF-1 alone and in combination with T₃. IGF-1 alone increased the portion of cells that incorporated BrdU (P < 0.05 vs. SF) as in Fig. 3 but that increase was completely suppressed in the presence of any of the three inhibitors (P < 0.05 vs. IGF-1). T₃ did not suppress BrdU uptake beyond the levels seen by the blocking agents. Figure 4 shows BrdU uptake rates in the older cardiomyocytes (135 dGA) using the same experimental design as in Fig. 3. The outcomes were similar at the two myocyte ages. Again the increase in proliferation rates caused by IGF-1 in the 135 dGA cardiomyocytes was completely suppressed in the presence of any of the three tested inhibitors of the MAPK and PI3K pathways (P < 0.05 vs. IGF-1, Fig. 4) whether T₃ was present or not (P < 0.001 vs. IGF-1, Fig. 4). The relative suppression was more severe in the older myocytes even though the absolute rates of BrdU incorporation were relatively low.

Fig. 3.

Proliferation studies (48 h) in 100 dGA cardiomyocytes in the presence of 10 μM U0126 (A), 10 μM LY294002 (LY) (B), and 10 nM rapamycin (R) (C). Each blocker on its own does not alter BrdU uptake compared with SF levels. IGF-1 (1 μg/ml)-stimulated BrdU (10 μM) uptake is inhibited by each blocker and is not further inhibited by the addition of T₃. Data are means ± SE; each bar n = 7. **P < 0.01 vs. SF; +P < 0.05, ++P < 0.01 vs. IGF-1.

Fig. 4.

Proliferation studies (48 h) in 135 dGA cardiomyocytes in the presence of 10 μM U0126 (A), 10 μM LY294002 (B), and 10 nM rapamycin (C). Each blocker on its own does not alter BrdU uptake compared with SF levels. IGF-1 (1 μg/ml)-stimulated BrdU (10 μM) uptake is inhibited by each blocker and is not further inhibited by the addition of T₃. Data are means ± SE; each bar n = 7. ***P < 0.001 vs. SF; +++P < 0.001 vs. IGF-1.

Developmental comparison of ERK and Akt signaling.

Phosphorylated ERK (pERK), Akt (pAkt), and p70^S6K (p-p70^S6K) were determined by Western blot analysis of lysates of cardiomyocytes that had been exposed for 10 min to IGF-1 (1 μg/ml), T₃ (1.5 nM), the combination of T₃ and IGF and inhibitors of ERK (U0, 10 μM) and Akt (LY, 10 μM) phosphorylation. Figure 5 shows that in the younger cells, phosphoprotein levels for ERK, Akt, and p70^S6K increased with IGF-1 alone and also with T₃ alone relative to their unphosphorylated protein levels (Fig. 5, A–C; P < 0.05 vs. SF). For ERK and Akt activation, IGF-1 stimulation was greater than for T₃ (⁺P < 0.05 vs. IGF-1). The combination of IGF and T₃ showed a less robust stimulation of ERK (Fig. 5A) compared with IGF-1 alone (P < 0.05) but no less than T₃ alone. Addition of LY did not further suppress pERK levels, while the addition of the U0 suppressed ERK phosphorylation to yield levels found under serum-free conditions and nearly undetectable (P < 0.001 vs. IGF-1). Figure 5B shows that the combination of IGF-1 and T₃ inhibits Akt phosphorylation to levels equivalent to serum-free conditions compared with IGF-1 and T₃ separately (P < 0.05). Cardiomyocytes treated with LY showed reduced Akt phosphorylation compared with cells treated with U0 (P < 0.05). Figure 5C shows a similar phosphorylation pattern for p-p70^S6K in cardiomyocytes subjected to either IGF-1 or T₃ alone as seen for the other signaling molecules studied. In the presence of IGF-1, T₃, and LY, p-p70^S6K levels were higher than those found under the same conditions with the MEK/ERK inhibitor, U0 (P < 0.05). These data are compatible with the hypothesis that p-p70^S6K is more tightly linked to the ERK cascade than to the PI3K cascade at this stage of development.

Fig. 5.

MAPK and PI3K signaling analysis in 100 dGA cardiomyocytes (10 min). Western blot analysis shows that IGF-1 (1 μg/ml) and T₃ (1.5 nM) each separately leads to phosphorylation of ERK (A), Akt (B), and p70^S6K (C), but IGF-1 activates these proteins to a greater degree than T₃. The combination of T₃ and IGF-1 reduces phosphorylation of each protein to half that of IGF-1 alone. U0126 (U0; 10 μM) blocked T₃-stimulation of ERK with no effect on Akt or p70^S6K. LY294002 (10 μM) more greatly blocked T₃-stimulation of Akt compared with U0126 with no effect on ERK. T₃ in combination with LY294002 did not decrease T₃-stimulated phosphorylation of p70^S6K, which differs from the response of T₃ + U0126 (P < 0.05). α-Tubulin was used as a loading control. p-, phospho; t-, total. Data are means ± SE; each bar n = 7. *P < 0.05, **P < 0.01 vs. SF; +P < 0.05, ++P < 0.01, +++P < 0.001 vs. IGF-1; #P < 0.05 vs. T_3. + IGF-1.

Figure 6 describes the same experimental design for older cardiomyocytes (135 dGA). The older myocytes showed one surprising difference compared with the 100 dGA cardiomyocytes. While T₃ and IGF-1 alone both increased the ratios of pERK (Fig. 6A), pAkt (Fig. 6B), and p-p70^S6K (Fig. 6C), as found in the younger cells, the combination of T₃ and IGF-1 stimulated pERK and pAkt to levels that exceeded either IGF-1 or T₃ alone. U0 and LY both nearly abolished their respective signaling pathways, MAPK and Akt. Both inhibitors resulted in a ∼50% reduction in the signaling of the alternate pathway suggesting that when cells are treated with T₃ and IGF-1 together, both the pERK and pAkt pathways are important in bringing about a cellular response.

Fig. 6.

MAPK and PI3K signaling analysis in 135 dGA cardiomyocytes (10 min). Western blot analysis shows that IGF-1 (1 μg/ml) and T₃ (1.5 nM) each separately leads to phosphorylation of ERK (A), Akt (B), and p70^S6K (C). The combination of T₃ and IGF-1 stimulates further activation of ERK and Akt. U0126 (10 μM) blocked T₃ + IGF-1-stimulated ERK and Akt. LY294002 (10 μM) inhibited T₃ + IGF-1 activation of ERK and Akt. Inhibition of ERK and Akt in combination with T₃ had no effect on p70^S6K. α-Tubulin was used as a loading control. Data are means ± SE; each bar n = 7. *P < 0.05, **P < 0.01, ***P < 0.001 vs. SF; +P < 0.05 vs. IGF-1; #P < 0.05, ##P < 0.01, ###P < 0.001 vs. T₃ + IGF-1.

Determination of ERK and Akt signaling in binucleated cardiomyocytes.

The maturation of signal transduction between ages is best seen by comparing Figs. 7 and 8. They show the percentage of cardiomyocytes in a population sample that were stained positively for pERK and pAkt by immunohistochemistry, following stimulation with T₃, IGF-1, or the combination of T₃ and IGF-1. Figure 7 shows 100 dGA cardiomyocytes, a stage when all cardiomyocytes are mononucleated. The percentage of pERK and pAkt positive cells is consistent with the Western blot profile where T₃ and IGF-1 separately stimulate their activity (P < 0.01 vs. SF) and where the combination blocks an increase in activity (i.e., not different to SF). Figure 8 shows the percentage of the older 135 dGA cardiomyocytes that have phosphorylated forms of ERK (Fig. 8A) and Akt (Fig. 8B). The percentages are shown separately for mononucleated and binucleated cardiomyocyte populations because both myocyte phenotypes are found at this stage. The finding reflects the ERK and Akt activity shown in Fig. 6 describing phosphoprotein responses to a similar experiment. The percentages of the population that were pERK and pAkt positive were increased above serum-free levels in the presence of either T₃ or IGF-1. However, when both hormones were present the percentages of cells that were pERK and pAkt positive increased dramatically (P < 0.01 vs. T₃ + IGF-1), exceeding 75%. Responses of the populations of binucleated cells were similar to those of the mononucleated cells. Among older cells, a larger pool of both mono- and binucleated 135 dGA myocytes were activated vs. cells at the younger age. This “recruitment” effect explains the increase in ERK and Akt activation found by Western blot analysis.

Fig. 7.

Immunohistochemical analysis of ERK and Akt in 100 dGA cardiomyocytes (10 min). Data from mononucleates are shown; at this age the number of binucleated myocytes is negligible. IGF-1 (1 μg/ml) stimulates ERK (A) and Akt (B) in 20–25% of cells above SF levels. T₃ (1.5 nM) also stimulates ERK and Akt but the combination with T₃ inhibits both signaling pathways. Cells positive for pERK (A1, 400×) show diffuse cytoplasmic staining and for pAkt (B1, 400×) show perinuclear staining. These cells are indicated by arrows. Data are means ± SE; each bar n = 4. **P < 0.01 vs. SF; +P < 0.05, ++P < 0.01 vs. IGF-1; ##P < 0.01 vs. T₃ + IGF-1.

Fig. 8.

Immunohistochemical analysis of ERK and Akt in 135 dGA cardiomyocytes (10 min). T₃ (1.5 nM) and IGF-1 (1 μg/ml) equally stimulates ERK (A, A1) and Akt (B, B1) in 40% of mononucleated and binucleated myocytes. T₃ in combination with IGF-1 stimulates 75% of the myocyte population. Cells positive for pERK (A1, 400×) show diffuse cytoplasmic staining and for pAkt (B1, 400×) show perinuclear staining. These cells are indicated by arrows. Data are means ± SE; each bar n = 4; *P < 0.05, **P < 0.01 vs. SF; ##P < 0.01 vs. T₃ + IGF-1.

DISCUSSION

Our previous studies have shown that IGF-1 and T₃ are key hormone regulators of fetal cardiomyocyte proliferation and maturation (6–8, 44). However, the details of how these two hormones regulate cell number and maturation in the fetal myocardium is unknown. We tested the conservative null hypothesis that activation of the MAPK-ERK and PI3K signaling patterns in ovine cardiomyocytes would not change as a function of gestational age. We compared signaling patterns at two stages of gestation. Cardiomyocytes of 100 days and 135 days of gestation were exposed to a combination of levels of IGF-1 and T₃, and phospho-protein quantification was used to compare the activation in key proteins in the ERK/Akt pathways.

We found that the combination of IGF-1 and T₃ led to dramatic increases in pERK and pAkt levels in older cardiomyocytes, well beyond those due to either IGF-1 or T₃ alone. In contrast, among the younger, more proliferative fetal cardiomyocytes, the combination of T₃ and IGF-1 reduced phosphorylation of ERK, Akt and p70^S6K levels by at least 50%. All of the major findings of the study are summarized in Table 1.

Table 1.

Maturation of signaling in fetal ovine cardiomyocytes

	Proliferation		Signaling Protein Activation
	IGF-1	T3 + IGF-1	T3 or IGF-1	T3 + IGF-1
100 dGA (early 3rd trimester)	Doubles proliferation	IGF-1-stimulated BrdU uptake reduced by 43%	Increased activation of ERK, Akt, and p70S6K	Reduced stimulation of ERK, Akt, and p70S6K vs. T3 or IGF-1 alone
			Approximately 20% cardiomyocytes positive for pERK or pAkt	Approximately 3% cardiomyocytes positive for pERK or pAkt
135 dGA (late 3rd trimester)	Quadruples proliferation	IGF-1-stimulated BrdU uptake reduced by 75%	Increased activation of ERK, Akt, and p70S6K	Increased activation of ERK and Akt vs. T3 or IGF-1; no change to p70S6K
			Approximately 30–50% cardiomyocytes positive for pERK or pAkt	Approximately 75% mononucleated and binucleated cardiomyocytes positive for pERK or pAkt

dGA, days gestational age.

To explain the more robust stimulatory effects of IGF-1/T₃ on ERK and Akt activation in the older cells, we sought to distinguish between a possible “superstimulatory” effect on the ERK/Akt by the IGF-1 and T₃ combination within a small population of myocytes vs. an expansion of the number of cells that were stimulated by the combination of hormones. The concentrations of exogenous T₃ and IGF-1 used in the experiments have been previously validated in previous studies (7, 8, 44). The immunohistochemical analysis (Figs. 7 and 8) clearly pointed to a maturation-dependent expansion of the number of cells stimulated in the presence of the growth factor/hormone combination. Surprisingly, even nonproliferating binucleated cells were affected to the same extent as the proliferation-capable mononucleated cells.

This study suggests a major difference between the functional role of the ERK and Akt pathways in the cells of different ages. In the younger cells, the two signaling pathways are both associated with proliferation and are IGF-1 sensitive. In the older cells some 50%-70% of the myocytes have exited the cell cycle as indicated by their binucleated status (23). Furthermore, at the older age, only a few percent of the population is stimulated to take up BrdU in the presence of serum or IGF-1. Yet nearly 75% of the cells in the entire population showed a robust staining for pERK and pAkt when IGF-1 and T₃ were simultaneously elevated. What characterizes the cells that were positive compared with those that were not remains a mystery.

It is well known that the MAPK and PI3K pathways can be stimulated in adult cardiomyocytes (33) and that the recruitment of the ERK/Akt pathway in adult cardiomyocytes has a protective benefit (54). Thus we speculate that the high levels of stimulation reflect the need to protect a cell population that is required for successful transition from fetal to neonatal life under difficult physiological circumstances including hypoxemia and increasing work load at birth. Thus at the younger age, when physiological levels of T₃ are normally low, cells are designed to proliferate under the influence of IGF-1 with the combination of hormones found only as a rare anomaly. On the other hand, as the older cardiomyocytes become exposed to both hormones as T₃ levels rise, we speculate that they benefit from the activation of both signaling pathways by stimulation of normal maturation and further protection against adversity.

In addition to the findings discussed above, similarities between the signaling properties of the 100 dGA and 135 dGA ovine cardiomyocytes were found in this study. These include 1) BrdU uptake is stimulated among myocytes by IGF-1 alone and suppressed at both ages when T₃ is present, and 2) IGF-1 and T₃ each independently activate ERK, Akt, and p70^S6K at both gestational ages studied. IGF-1 also stimulated both the ERK and Akt pathways in 100 dGA cardiomyocytes, and as previously reported, the same was seen in 135 dGA cardiomyocytes (44). Activation of ERK and Akt pathway constituents is usually associated with proproliferative actions in populations of dividing cells (30, 38, 43, 46). Yet in the current study, ERKs were activated even during suppression of cell cycle activity and under conditions where most cells have exited the cell cycle.

The signaling that we have studied is likely to be important in the long-term health of the heart. Twenty-five years ago, Barker and colleagues (2, 3) proposed that poor placental growth was prognostic for adult onset ischemic heart disease. Barker and colleagues also showed that small body size at birth predicts both increases in thyroglobulin and thyroid peroxidase autoantibodies in adult women (34) and spontaneous hypothyroidism among women (25). Ovine studies show that either high or low T₃ concentrations are powerful regulators of myocardial maturation (6) and alter the number of cardiomyocytes in the heart at birth (Chattergoon and Thornburg; unpublished data). Therefore, we speculate that maternal thyroid hormone concentrations affect cardiomyocyte endowment in the offspring which may increase their long-term disease risk. IGF-1 is a powerful stimulus for cardiomyocyte proliferation in the fetus but this study shows that T₃ acts as a brake on IGF-1's promotion of proliferation. Thus it appears that the healthy myocardium is determined by appropriate concentrations of IGF-1 and the appropriate timing of increases in T₃ in the last weeks of gestation.

Conclusions.

The results of the current study confirm that elevated T₃ suppresses IGF-1 mediated cardiomyocyte proliferation in the ovine fetus. It also supports our previous finding that signaling pathways (MAPK and PI3K) are important to T₃'s inhibition of IGF-1 action. In the near-term cardiomyocytes, but not in the midgestation myocytes, the combination of IGF-1 and thyroid hormone (T₃) caused activation of ERK and Akt in most myocytes regardless of their degree of nucleation and cell cycle phase. There is a clear and profound maturational effect leading to opposing signaling responses at the two ages studied. These differing responses are likely to be related to distinct changes in fetal cardiomyocyte maturation as the myocardium prepares for terminal differentiation (100 dGA) or for ex utero life (135 dGA).

GRANTS

This study was supported by National Institutes of Health Grants P01-HD-34430, R37-NS-045737, and R01-HL-102763. K. L. Thornburg was supported by the M. Lowell Edwards Endowment. N. N. Chattergoon was supported by National Heart, Lung, and Blood Institute Training Grant T32-HL-094294.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: N.N.C. and K.L.T. conception and design of research; N.N.C. performed experiments; N.N.C. analyzed data; N.N.C., S.L., P.S., and G.D.G. interpreted results of experiments; N.N.C. and P.S. prepared figures; N.N.C. drafted manuscript; N.N.C., S.L., P.S., G.D.G., and K.L.T. edited and revised manuscript; N.N.C., S.L., and K.L.T. approved final version of manuscript.