Demonstration of Direct Effects of Growth Hormone on Neonatal Cardiomyocytes*

,

Several observations implicate a role for growth hormone (GH) 1 in modulation of cardiac structure and function (1). Patients with excess endogenous GH (i.e. acromegaly) suffer from cardiac complications including biventricular hypertrophy, impaired diastolic filling, and decreased cardiac performance on effort due to diastolic and systolic dysfunction (2). Patients with chronic GH deficiency also show cardiac abnormalities; in general, the data support the presence of a hypokinetic cardiac syndrome in patients with GH deficiency that can be reversed with GH replacement therapy (3)(4)(5). Fazio et al. (6) reported that GH therapy in patients with idiopathic dilated cardiomyopathy was associated with significant improvement in left ventricular ejection fraction, isovolumic relaxation time, and efficiency of myocardial energy utilization. Subsequent to these landmark findings, some studies have supported a beneficial effect of exogenous GH on cardiac function (7), whereas other investigators were unable to demonstrate salutary effects of GH on cardiac function in patients with heart failure (8).
A particularly well studied animal model is that of the transplanted GH-secreting pituitary tumor cell line, GH 3. In this model of GH excess, there is increased myocardial contractility and calcium sensitivity of myocardial contractile proteins (1). Similarly, normal rats given recombinant GH show an increase in left ventricular mass, as well as an increase in several aspects of cardiac performance (9). In the rodent model of myocardial infarction, administration of GH results in improvements in myocardial contractility, left ventricular endsystolic and end-diastolic pressures, and cardiac index with no increase in the size of the infarct (10 -13). In general, animal models of GH deficiency also support a role for GH in the maintenance of cardiac structure and function. Thus, genetically GH-deficient dwarf mice show cardiac abnormalities that can be reversed following GH therapy (14). However, hypophysectomized rats given GH show little improvement in ventricular function, indicating that GH may cooperate with other factors (e.g. thyroid hormone) in its effects on the heart (15).
In the intact animal, GH increases the circulating levels of IGF-1 by stimulating the production of IGF-1 (16). IGF-1 itself induces hypertrophy and alters gene expression in isolated cardiomyocytes and increases myocardial contractility (17,18). IGF-1 may also serve to inhibit cardiomyocytes from undergoing apoptosis following infarction (19). However, to date, studies have not been able to demonstrate direct effects of GH on isolated cardiomyocytes. Thus, Ito et al. (20) reported that, whereas IGF-1 was able to induce hypertrophy with concomitant increase in expression of muscle-specific genes in isolated rat neonatal cardiomyocytes, GH did not have an observable effect in this model. Similarly, Donath et al. (17) reported that in cardiomyocytes from adult rats IGF-1, but not GH, enhances myofibril development and down-regulates smooth muscle-␣ actin. In support of a direct (i.e. not secondary to stimulation of circulating IGF-1) effect of GH on cardiac function is the demonstration that ex vivo perfusion of the isolated rat heart with GH results in enhanced protein synthesis (21).
At present, it is not known if GH, acting through its cognate receptor (growth hormone receptor (GHR)), has direct effects on the heart or if some or all of the observed cardiovascular effects of GH administration are secondary to an increase in circulating levels of IGF-1 and/or hemodynamic changes in vascular tone and blood pressure (22). Recently, Lupu et al. (23) reported that in a knockout model of GHR deficiency characterized by severe growth retardation, the weight of the heart in the adult animal, although decreased in absolute terms, was similar to controls when expressed relative to total body weight. Moreover, relative levels of IGF-1 transcripts in the heart also were unaffected. Thus, in this model of GHR deficiency, myocardial growth and IGF-1 expression are implicitly independent of the GH/GHR axis. However, the studies of Lupu et al. (23) do not preclude a role for GH/GHR in normal growth and development of the neonatal heart, including adaptive myocardial metabolism, characterized by a transition from virtually exclusive utilization of glucose and lactate in the fetus to fatty acids in the mature animal (24).
A major impediment to investigating the role of GH/GHR in the heart is the lack of a suitable in vitro tissue culture model. We have developed such a model by exploiting the strategy of recombinant adenovirus-mediated overexpression of GHR in neonatal rat cardiomyocytes. Using this model, we demonstrate that GH stimulates hypertrophy of the isolated cardiomyocyte, and these actions of GH seem to be largely independent of changes in IGF-1 gene expression, findings that support an important independent role for the GH/GHR axis in normal cardiac growth. In addition, GH alters the metabolic profile of rat neonatal cardiomyocytes by inhibiting glucose uptake, stimulating fatty acid uptake, and increasing protein synthesis, changes that facilitate both postnatal cardiac growth and maturation of myocardial metabolism.

EXPERIMENTAL PROCEDURES
Isolation of Cardiomyocytes-Neonatal ventricular myocytes were prepared from hearts of 1-3-day-old Harlan Sprague-Dawley rats using a previously described protocol (25,26). Briefly, hearts were removed and transferred to a 100-mm dish containing heparin solution (1ϫ Hanks' balanced saline solution, 0.8 M MgSO 4 , 20 mM HEPES, and 10 units of heparin). After dissecting hearts into two pieces with a razor blade, the tissue was washed once with the heparin solution and incubated with gentle rocking in a solution containing 1ϫ Hanks' balanced saline solution, 0.8 M MgSO 4 , 20 mM HEPES, 2 g/ml DNase, and 2 mg/ml trypsin. Tissue was pipetted three times every 20-min with a 10-ml-wide mouth pipette, and the supernatant was discarded at each time. Cardiomyocytes were recovered afterward by alternating rocking and pipetting of tissue every 5 min and pooled into a 50-ml tube containing 7 ml of fetal bovine serum. Cells were centrifuged at 1000 rpm for 5 min and washed once with cold DMEM containing 5% fetal bovine serum. Fibroblasts were removed by plating (100-mm dishes) cells at 37°C for 30 min in DMEM containing 5% fetal bovine serum. Cardiomyocytes were recovered by collecting the nonadherent cells and cultured in DMEM/F-12 medium containing 5% charcoal-stripped calf serum (Sigma), 10 mM HEPES, insulin-transferrin sodium selenite (5 g/ml insulin, 5 g/ml transferrin, and 5 ng/ml selinium selenite, respectively), 10 mM glutamine, 30 g/ml bromodeoxyuridine, and 10 g/ml gentamicin. Cells were plated on Pronectin (BIOSOURCE)coated tissue culture plates at a density of 1 ϫ 10 5 cells/cm 2 and grown at 37°C in 5% CO 2 . As assessed by staining with antisarcomeric antibody (MF20; Developmental Studies Hybridoma Bank, University of Iowa), these cell preparations typically contain greater than 95% cardiomyocytes, form a nearly confluent monolayer by the second day after plating, and beat spontaneously and synchronously. After 24 or 48 h of incubation, cells were washed with serum-free DMEM/F-12 culture medium and infected with recombinant adenovirus expressing either the murine GHR (GHRAdlox) or the ␤-galactosidase (␤-galAdlox) gene.
Engineering and Production of Recombinant Adenoviral Vectors-The second generation shuttle vector pADLOX was modified to contain either the ␤-galactosidase (␤-galAdlox) or the murine GHR cDNA (GHRAdlox). The in vitro coupled transcription-translation technique (Promega) was used to confirm that the plasmids coded for proteins of appropriate molecular weight (data not shown). The production of the replication-deficient adenoviral vectors expressing either the ␤-galactosidase or GHR cDNA was achieved by a protocol based on the Cre-lox principle for efficient and relatively quick identification of recombinant viral particles (27). The adenovirus was plaque-purified, and a cesium chloride preparation was used for the experiments.
Scatchard Analysis-The number of cell surface receptors was determined by competition binding using 125 I-labeled human GH. 1 ϫ 10 6 aliquots of neonatal cardiomyocytes (uninfected and infected at MOI of 12.5 plaque-forming units/cell for 48 h) were washed with binding buffer (25 mM HEPES (pH 7.4), 125 mM NaCl, 4 mM KCl, 1 mM CaCl 2 , 1.5 mM MgCl 2 , 2 mM KH 2 PO 4 , and 1% human serum albumin), incubated with 50,000 cpm 125 I-human GH (PerkinElmer Life Sciences) and various concentrations of unlabeled GH for 90 min in binding buffer before centrifugation through 100 l of dibutyl phthalate. Specific binding was determined by subtracting the amount of 125 I-human GH bound in the presence of excess unlabeled rGH (National Hormone and Pituitary Program, National Institutes of Health), and the binding characteristics were analyzed using a computer program (GraphPad).
Immunofluorescence and Planimetry-After 24 h of adenoviral infection, cells were rinsed with phosphate-buffered saline, fixed in 4% paraformaldehyde, and permeabilized with 0.1% Triton X-100. Following three phosphate-buffered saline washes, the chamber slides were incubated in 0.1% bovine serum albumin for 30 min to block nonspecific sites. Subsequently, the cells were covered with either anti-GHR antibody (GHR-2; specific for the cytoplasmic region of the GHR molecule and does not cross-react with GH-binding protein (28)) or antisarcomeric myosin antibody (MF-20). Following incubation for 1 h in the dark with the primary antibody, the cells were washed four times with phosphate-buffered saline prior to exposure to Texas Red-conjugated secondary antibodies. A Zeiss Axiovert 135 microscope equipped with a video camera system was used to capture images and to determine cell size, estimated by measuring the area to which individual MF-20positive cells attached. 20 -45 randomly selected cells from each group (␤-galAdloxor GHRAdlox-infected, with and without exposure to rGH) were chosen for such analysis, which was conducted by an operator blinded to the identity of the cells being examined. The data gathered were analyzed using ImagePro software.
Isolation of RNA-GHRAdlox-or ␤-galAdlox-infected cells were incubated with or without rGH for certain time periods. Cells were washed once with ice-cold phosphate-buffered saline, and total RNA was prepared by the Tri-Reagent method and quantitated by absorbance at 260 nm. The integrity of the RNA and the accuracy of the spectrophotometric determinations were confirmed by visual inspection of the ethidium bromide-stained 28 and 18 S ribosomal RNA bands after agarose-formaldehyde gel electrophoresis.
Ribonuclease Protection Assay for IGF-1-IGF-I gene expression was studied by RNase protection assay as previously reported (29,30). Briefly, 10 -50-g aliquots of total RNA were hybridized to 32 P-labeled riboprobes generated from templates containing exon 4 of the rat IGF-I gene (a 376-base pair Sau3AI-EcoRI fragment) or 18 S rRNA (Ambion). Protected probes were denatured, electrophoresed on an 8% polyacrylamide gel, and exposed to X-Omat AR film for 1-2 days. The protected bands corresponding to IGF-I mRNA and 18 S rRNA were scanned using an Agfa Arcus II scanner, and densitometric analysis was performed using the MacBAS version 2.31 computer program (Fuji Photo Film Co., Tokyo, Japan). The level of IGF-I mRNA was normalized for the amount of 18 S rRNA and analyzed as relative abundance to untreated control samples.
Real Time Semiquantitative RT-PCR Assay-Real-time quantitative RT-PCR using the ABI Prism 7700 Sequence Detection System (PE Biosystems, Foster City, CA) was carried out using established protocols (31)(32)(33)(34)(35). The primers (synthesized by Life Technologies, Inc.) and TaqMan probes (synthesized by Applied Biosystems) for the quantitation of the GHR, IGF-1, fatty acid transport protein (FATP), c-fos, and ANF transcripts were designed using the primer design software Primer Express (Applied Biosystems) ( Table I). The primers and Taq-Man probe for 18 S rRNA were purchased from a commercial vendor (PerkinElmer Life Sciences). The 18 S probe was labeled with reporter fluorescent dyes VIC and the GHR, IGF-1, FATP, c-fos, and ANF probes with 6-carboxyfluorescein. The relative efficiencies of the GHR and IGF-1 primers/probe sets and the 18 S primer/probe pair were tested by subjecting serial dilutions of a single RNA sample from each of the tissues analyzed to real time RT-PCR analysis. The plot of log input versus ⌬C T was Ͻ0.1, which satisfies the previously established criterion for equivalence of efficiency of amplification (34). C T , or threshold cycle, represents the PCR cycle at which an increase in reporter fluorescence above a base-line signal can first be detected, and ⌬C T refers to the difference between the threshold cycles for the target and the reference. After confirming that the efficiencies of amplification of the gene of interest (e.g. GHR) and 18 S transcripts were approximately equal, the amount of the transcripts for the specific gene relative to the 18 S transcript was determined by using the comparative C T (separate tube) method (34). The comparative C T method is similar to the conventional standard curve method, except it uses arithmetic formulas to achieve the same result for relative quantitation. The amount of target, normalized to an endogenous reference and relative to a calibrator is given by the following formula: -fold induction ϭ 2 ⌬⌬CT , where ⌬⌬C T ϭ (C T GI (unknown sample) Ϫ C T 18 S (unknown sample)) Ϫ (C T GI (adult heart) Ϫ C T 18 S (adult heart)); GI represents the gene of interest (e.g. GHR). Briefly, 2-ng aliquots of total RNA were analyzed using the One-Tube RT-PCR protocol (Applied Biosystems). Following reverse transcription at 48°C for 30 min, the samples were subjected to PCR analysis using the following cycling parameters: 95°C for 10 min; 95°C for 15 s 3 60°C for 1 min for 40 cycles. Each sample was analyzed in triplicate in individual assays performed on two or more occasions.
Antibodies-Anti-GHR used for immunostaining was a gift of Dr. F. Talamantes (28). Antisarcomeric myosin antibody (MF-20) was obtained from the Developmental Studies Hybridoma Bank (University of Iowa). Antibodies against Glut1 and Glut4 have been described previously (39), and anti-actin antibody was purchased from Sigma. Antiphosphotyrosine monoclonal antibody (4G10) and anti-MAPK affinitypurified rabbit antibody (directed at residues 333-367 of rat ERK1; recognizes both ERK1 and ERK2) were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Anti-activated (anti-phospho-) MAPK affinity-purified rabbit antibody (recognizing the dually phosphorylated Thr 183 and Tyr 185 residues that correspond to the active forms of ERK1 and ERK2) was purchased from Promega (Madison, WI). Anti-phospho-Akt (specifically recognizing Akt phosphorylated at Ser 473 ) affinity-purified rabbit antibodies were purchased from New England Biolabs (Beverly, MA). Rabbit anti-phosphotyrosine-STAT5 polyclonal antibody (raised against a phosphopeptide surrounding phosphorylated Tyr 694 of murine STAT5A, which is conserved in both STAT5A and STAT5B) was obtained from Zymed Laboratories Inc. (San Francisco, CA). The rabbit polyclonal serum, anti-JAK2 AL33 (referred to herein as anti-JAK2), has been described previously (40).
Anti-GHR cyt-AL47 (referred to as anti-GHR in immunoblotting experiments) is a new antibody developed in the Frank laboratory. This rabbit serum was raised against a bacterially expressed N-terminally His-tagged fusion protein incorporating the entire cytoplasmic domain of the human GHR (residues 271-620) (41). The cDNA encoding this fusion was created by PCR in the pET vector system (Novagen) (PCR primer sequences available upon request). Molecular biology techniques, cDNA sequencing for verification of fidelity, bacterial fusion protein expression, and preparation of Ni 2ϩ -agarose-purified fusion protein were performed according to our previous methods (36,37,40,42) and the manufacturer's suggestions.
Quantification of Glucose Transport-Glucose uptake was assayed at 37°C as 2-deoxy-D-glucose (2-G-3 H-labeled) internalization. Briefly, cardiomyocytes exposed to either rGH (1250 ng/ml) or the vehicle were incubated in DMEM (1 g/liter glucose) containing 1 ci/ml 2-deoxy-Dglucose (2-3 H-G-labeled) (PerkinElmer Life Sciences). At specific time points, cultures were removed from the incubator, and cytochalasin B (250 M final concentration) was added to halt further hexose uptake. The monolayer was then washed three times with ice-cold PBS, the washed cardiomyocytes were solubilized by exposure to 0.5 N NaOH for 10 min and neutralized with equal volume of 0.5 N HCl, and aliquots were taken for measurement of 3 H activity by scintillation counting and estimation of protein content by the Bradford assay. The glucose uptake was calculated by subtracting nonspecific cell-associated radioactivity  trichloroacetic acid at 4°C for 10 min. After washing with ice-cold ethanol, the cells were solubilized in 0.5 N NaOH for 10 min and then neutralized with an equal amount of 0.5 N HCl. The uptake of 3 H was determined by scintillation counting. Palmitic Acid Uptake Assay-Palmitate uptake studies were performed in infected and uninfected cardiomyocytes that were treated or untreated with rGH (0.5-1250 ng/ml) for 4 h. Albumin-bound palmitate was prepared by dissolving unlabeled sodium palmitate (hexadecanoic acid; Sigma) and [9,10-3 H]palmitic acid (American Radiolabeled Chemicals) in 10 ml of water, to give a concentration of 320 M (60 Ci/ml). Fatty acid-free bovine serum albumin was added to the above solution to obtain a palmitate/PBS ratio of 4.0. This stock solution was diluted 1:8 with PBS before adding to the cells, and the incubation was carried out for 2 min. Palmitate uptake was terminated by adding cold PBS containing 0.1% PBS and 200 M phloretin. Cells were then washed five times with cold PBS containing 0.1% bovine serum albumin, solubilized with 0.5 N NaOH, and then neutralized with 0.5 N HCl. Palmitate uptake was determined by scintillation counting.
Data Analysis-Data are presented as either mean Ϯ S.D. or mean Ϯ S.E. as indicated. Statistical differences between groups were determined by ANOVA. p values equal to or less than 0.05 were considered significant.
The study protocol was approved by the Children's Hospital of Pittsburgh Animal Care and Use Committee, and the animals received humane care in compliance with the National Research Council's criteria as outlined in the Guide for the Care and Use of Laboratory Animals prepared by the National Institutes of Health.

Dysregulation of Expression of GHR during Isolation of Neo-
natal Cardiomyocytes-Isolated cardiomyocytes from the neonatal rat heart are a standard and accepted cell culture model for studying mammalian cardiac cell physiology that has been used to examine the role of the GH/IGF-1 system in the heart (20). To ascertain if cultured cardiac cells are an appropriate model in which to study the effects of GH, qualitative RT-PCR was performed using RNA isolated directly from rat heart or from cultured cardiomyocytes (Fig. 1A). Comparison of GHR transcript levels in isolated neonatal cells (Fig. 1A, lane 9) versus that found in neonatal whole heart (Fig. 1A, lane 7) indicates that either during isolation and/or culture, GHR mRNA levels decrease significantly in isolated cardiomyocytes. To confirm these qualitative data, a quantitative analysis was next performed using a fluorescent 5Ј-nuclease (TaqMan) assay. These data confirmed that the expression of GHR mRNA was decreased in isolated rat neonatal cardiomyocytes by ϳ70 -80% compared with either the intact neonatal or the adult heart, respectively (Table II). Indirect immunofluorescence microscopy revealed that most of the endogenous GHR in the cultured cardiomyocyte was perinuclear and not in the cellular membranes, as seen in tissues, including the heart, in the intact animal (Fig. 1B) (43, 44). Thus, the dual findings of low levels of GHR mRNA and the aberrant localization of the GHR protein in these cells may explain the observation of others that GH treatment has little effect on the rat primary neonatal cardiomyocyte in culture (17,20). Furthermore, adult whole heart expresses about 35% more GHR transcripts than the neonatal heart ( Fig. 1A and Table II), thus increasing the possibility that GH has direct effects on adult cardiomyocytes in vivo.
Adenovirus-mediated Overexpression of GHR-To determine whether cultured cardiomyocytes could be made sensitive to the effects of GH, we explored means to increase the expression of GHR in rat neonatal cardiomyocytes. Pilot experiments revealed that strategies to obtain a homogenous population of cardiomyocytes overexpressing GHR, such as a selection process based on co-transfection of the plasmid pHook-2 (Invitrogen), resulted in poor yields. Hence, we chose to employ an alternate strategy using recombinant adenovirus (45,46). The second generation shuttle vector pADLOX was modified to contain either the ␤-galactosidase (␤-galAdlox) or the murine GHR cDNA (GHRAdlox) under the control of the cytomegalovirus promoter. The production of replication-deficient adenoviral vectors expressing either ␤-galactosidase or GHR cDNA was achieved by a protocol based on the Cre-lox principle (27). A multiplicity of infection of ϳ12.5 plaque-forming units/cell was routinely used in this study because pilot experiments determined this multiplicity of infection to be the upper limit of tolerance of the isolated cardiomyocytes to adenoviral infection. Analysis of cell lysates by Western blot with an antiserum against the GHR cytoplasmic domain shows the expected GHR protein product (lower panel of Fig. 2A and upper panel of Fig.  2B). The number of cell surface receptors was determined by competition binding using 125 I-labeled human GH. Whereas there was no specific binding with noninfected cardiomyocytes, 140,000 Ϯ 23,000 receptors/cell with a K d of 4.9 Ϯ 1.1 nM were detected on the infected cardiomyocytes.
GH-GHR Signaling Pathways in Isolated Cardiomyocytes-GH signaling is initiated by ligand-induced GHR dimerization and activation of the receptor-associated tyrosine kinase, JAK2 (38,47,48). This results in tyrosine phosphorylation of JAK2, the GHR, and other intracellular proteins (see Refs. 49 and 50 and references therein). With the exception of calcium influx (51), the GH-dependent activation of intracellular signal cascades, such as the STAT5, phosphatidylinositol 3-kinase/Akt, and MAPK pathways requires JAK2 activation (49,50). To determine if GH signaling in our in vitro model system was similarly associated with intracellular tyrosine phosphorylation, cells stimulated with GH or vehicle were detergent-solubilized, and proteins resolved by SDS-polyacrylamide gel electrophoresis were immunoblotted with anti-phosphotyrosine antibodies ( Fig. 2A, upper panel). Stimulation of isolated cardiomyocytes transduced with the GHR-overexpressing virus with GH for 15 or 30 min resulted in the ap-  7). B, pattern of expression of GHR in cultured cardiomyocytes. Isolated primary rat neonatal cardiomyocytes were cultured for 48 h before fixation in 4% paraformaldehyde. Cells were analyzed by indirect immunofluorescence using primary antibody specific for the cytoplasmic region of the GHR (GHR-2). Note perinuclear localization of GHR fluorescence (arrow) with little membrane-associated fluorescence.
pearance of two prominent tyrosine-phosphorylated bands in the 100 -125-kDa range, which were not detected in extracts of the cells transduced with the ␤-galactosidase overexpressing virus (lanes 4 -6 versus lanes 1-3). By analogy to our previous studies of rodent cells (36,38), these GH-induced tyrosine phosphoproteins (a sharp band designated by an arrow and a diffuse, more rapidly migrating band designated by a bracket) probably include JAK2 and the GHR. To further their identification, the nitrocellulose filter was sequentially stripped and reprobed with anti-JAK2 AL33 (Fig. 2A, middle panel) and anti-GHR AL47 (Fig. 2A, lower panel) sera. As expected, endogenous JAK2 (arrow) was detectable at similar levels in each sample, while GHR (bracket that includes the fully glycosylated and faster migrating incompletely glycosylated GHR forms (52)) was present only in the GHR-overexpressing cells. The tyrosine phosphoproteins present in lanes 5 and 6 of the upper panel exactly comigrated with JAK2 and the fully glycosylated GHR and are thus designated P-JAK2 and P-GHR, as indicated. These data strongly suggest that GH acutely promotes JAK2 and GHR tyrosine phosphorylation in the GHR-reconstituted cardiomyocytes.
To further characterize signaling pathways downstream of JAK2 activated by GH in our cardiomyocyte model system, we performed a similar experiment in which ␤-galactosidaseor GHR-overexpressing cardiomyocytes were treated with vehicle or GH for 15 min. Extracted proteins were separated by SDSpolyacrylamide gel electrophoresis and sequentially immunoblotted with activation state-specific antibodies that detect the phosphorylated forms of STAT5, Akt, and MAPK (ERK1 and ERK2) (Fig. 2B). We observed robust GH-induced phosphorylation of each signaling molecule in the GHR-overexpressing cells, while the ␤-galactosidase-overexpressing cells exhibited a more modest degree of activation of STAT5 and Akt phosphorylation and minimal MAPK phosphorylation (with a higher basal degree of MAPK activation in these cells). Similarity of loading was demonstrated by assaying for total MAPK (Fig. 2B,  bottom panel). In other experiments (not shown), GH-dependent activation of the STAT5, Akt, and MAPK pathways was consistently observed in GHR-transduced cells, in contrast to inconsistent and weak responses in the ␤-galactosidase-overexpressing cells. These data suggest that the canonical pathways of GH-GHR signaling cascades are reliably intact in cardiomyocytes overexpressing the GHR.
GH Modulates Metabolic Parameters in Isolated Cardiomyocytes-GH is known to alter metabolic parameters of GH-responsive tissues. Having established that the recombinant adenovirus-infected cardiomyocytes expressed GHR, we next tested the effect of GH on metabolic parameters such as incorporation of amino acids, fatty acids, and glucose uptake of cardiomyocytes in this model. GH did not alter the uptake of any of the metabolites in nontransduced or ␤-galactosidaseoverexpressing adenovirus-transduced cardiomyocytes. However, in GHR-overexpressing neonatal cardiomyocytes, GH increased the incorporation of L- [3,4, H]leucine in a dose-dependent manner: 127 Ϯ 7, 144 Ϯ 14, and 163 Ϯ 28% at doses of 0.5, 50, and 1250 ng/ml, respectively (mean Ϯ S.E.; n ϭ 3) (Fig.  3). Likewise, GH increased the uptake of [9,10-3 H]palmitic acid, although a statistically significant effect (151 Ϯ 10%) was observable only at a GH concentration of 1250 ng/ml (Fig. 4). In contrast, the uptake of 2-deoxy-D-glucose (2-G-3 H-labeled) was inhibited by 60 Ϯ 11% after 4 h of exposure to a GH concentration of 1250 ng/ml (Fig. 5).
To ascertain the possible mechanism for the impaired glucose uptake, we next measured the levels of glucose transport proteins in these samples. Western blot analysis of cell lysates determined that GH decreased the levels of Glut1 by ϳ25-30% in GHR-overexpressing cardiomyocytes compared with car-

FIG. 2. GH stimulates GH-dependent post-receptor signaling pathways in GHR-overexpressing rat neonatal cardiomyocytes.
A and B, isolated rat neonatal cardiomyocytes overexpressing either ␤-galactosidase or GHR were stimulated with rGH (500 ng/ml) or vehicle (0 min) for the indicated durations. Detergent-solubilized proteins were resolved by SDS-polyacrylamide gel electrophoresis and sequentially immunoblotted with the indicated antibodies. In A, the position of the prestained 97-kDa molecular mass marker is indicated. P-GHR, P-JAK2, and P-STAT5, tyrosine-phosphorylated GHR, JAK2, and STAT5, respectively. P-Akt, serine-phosphorylated (activated) Akt; P-MAPK, dually (threonine and tyrosine) phosphorylated (activated) ERK-1 and -2, as described under "Experimental Procedures." The abundance of total MAPK is indicated in the bottom panel. diomyocytes transduced with the ␤-galactosidase adenovirus (Fig. 6, A and B). In contrast, there was no change in the levels of Glut4 transport protein in these samples. In addition, estimation of hexokinase activity did not reveal any significant changes in cells after exposure to GH (data not shown). These results indicate that the GH-induced decrease in glucose uptake by cardiomyocytes is in part due to alteration in levels of the Glut1 protein.
Fatty acid transport across membranes can be due to both diffusion and carrier-mediated processes (53). Many proteins are thought to participate in the transmembrane translocation of fatty acids. These include fatty acid-binding proteins, fatty acid translocase, and fatty acid transport proteins (54). Having demonstrated that GH increased the uptake of fatty acids in the GHR-overexpressing cardiomyocytes, we used the fluorescent 5Ј-nuclease (TaqMan) assay to measure the levels of mRNA for one of the known FATPs in this model system (55). These studies revealed that the GH increased the levels of mRNA for FATP by ϳ50% in GHR-overexpressing cardiomyocytes in comparison with ␤-galactosidase-overexpressing cardiomyocytes (Fig. 7).

GH Stimulates Hypertrophy of Isolated Cardiomyocytes-
IGF-1 is known to induce hypertrophy of the isolated cardiomyocyte, and previous reports have concluded that GH does not have a similar effect on these cells (20). We reexamined the issue of direct effects of GH in our model of GHR-overexpressing cardiomyocytes. For this purpose, we measured the area of cells exposed to GH or the vehicle only. These experiments revealed that cardiomyocytes transduced with the ␤-galactosidase-overexpressing adenovirus did not undergo changes in cell size on exposure to rGH (GH(Ϫ) versus GH(ϩ), 1083 Ϯ 36 Cardiac hypertrophy is associated with characteristic changes in gene expression. Thus, it is well established that the expression of early response genes (e.g. c-fos, c-myc, c-jun, egr-1) are activated within 30 -60 min after exposure to a hypertrophic stimulus. Additionally, "late" markers of cardiac hypertrophy (e.g. ANF, myosin light chain 2, ␣-skeletal actin, ␤-myosin heavy chain) are also variably activated in various forms of cardiac hypertrophy. We utilized the TaqMan assay to profile the GHdependent expression of representative genes from the "early" (i.e. c-fos) and from the "late" (i.e. ANF, myosin light chain 2, ␣-skeletal actin) response group of genes in our model system (Fig. 8). Our results indicate that in GHR-overexpressing cells, GH elicited a robust (500 -600%) increase in expression of c-fos mRNA and a modest (40 -60%) increase in expression of myosin light chain 2 and ␣-skeletal actin mRNA. In contrast, GH did not alter the expression of ANF mRNA in these cells.
GH's Actions on Isolated Cardiomyocytes Are Independent of IGF-1-At the cellular level, the actions of GH can be either attributed to direct effects of GH or due to indirect effects, since GH induces the synthesis of IGF-1 in many tissues and IGF-1 has potent actions in these tissues (56). To determine if the effects of GH observed in our model system were due to direct effects of GH or due to stimulation of IGF-1 synthesis, we estimated the levels of IGF-1 mRNA by ribonuclease protection and fluorescent 5Ј-nuclease (TaqMan) assays. The RPA analysis revealed that there was no change in IGF-1 levels in response to GH stimulation for 24 h (Fig. 9A). To exclude the possibility that changes in IGF-1 could have preceded the observed effects of GH on the metabolic parameters and cell size, we used the fluorescent 5Ј-nuclease (TaqMan) assay to investigate the temporal profile of IGF-1 in our model system. These data reveal that, in concert with the RPA data, there are no significant changes in IGF-1 mRNA levels in these cells in response to GH stimulation for 15 min, 60 min, or 4 h (Fig. 9B). DISCUSSION We demonstrate that GH has effects on the isolated neonatal cardiomyocyte expressing GHR. Our results indicate that GH causes hypertrophy of the cardiomyocyte and alters its metabolic profile by decreasing glucose uptake while increasing fatty acid uptake and protein accretion rate. Both the morpho-logical and metabolic effects are not temporally associated with changes in IGF-1 gene expression in these cells.
It is claimed that IGF-1 has direct actions on cardiomyocytes, whereas GH does not (17,20). In this report, we demonstrate that the process of isolation and culture of neonatal cardiomyocytes results in loss of expression of GHR mRNA and aberrant localization of the GHR protein. In contrast to this loss of expression of GHR mRNA, IGF-1 receptor mRNA is increased during the process of isolation and culture (17). Thus, these experimental artifacts could be responsible for the finding that these cells respond to IGF-1 but are unresponsive to GH (20). Our data indicate that the strategy of using recombinant adenovirus-mediated overexpression of GHR can be used to establish an in vitro model of isolated cardiomyocytes that respond to GH. It is noteworthy that in our model system we did observe some GH-dependent effects in cardiomyocytes transfected with the ␤-galactosidase-overexpressing adenovirus. The only consistent response following GH stimulation of cardiomyocytes not overexpressing GHR was an increase in phospho-STAT5. This effect possibly reflects signaling through GHR or prolactin receptors that remained intact during the isolation and culture. Activation of phospho-Akt and c-fos was also detectable in the ␤-galactosidase-overexpressing cells. However, these effects were minimal and inconsistent, possibly reflecting variations in harvesting and culture of these cells.
Whereas the actions of GH to alter the metabolic parameters in various cell types are well established, the current report is the first to demonstrate effects of GH on the metabolic parameters of the isolated cardiomyocyte. In general, the effects of GH on carbohydrate metabolism can be grouped into acute and chronic effects. The acute effects of GH, commonly referred to as insulin-like actions, include decrease in blood glucose concentration as well as stimulation of glucose uptake by muscle and isolated adipocytes in vitro. In contrast, the chronic effects of exposure to GH include increase in blood glucose concentration, inhibition of glucose uptake, and insulin resistance (16). Our results parallel observations in fat (57) and skeletal muscle (58) that demonstrate that GH decreases basal glucose transport via a decline in GH-induced Glut1 mRNA without affecting total cellular content of Glut4 protein. However, since we only analyzed total cellular Glut4, the possibility that alterations in subcellular localization of Glut4 could contribute or GHR (solid bar) overexpressing cardiomyocytes were exposed to rGH (500 ng/ml) for 4 h, and total RNA was extracted for analysis of FATP gene expression using a fluorescent 5Ј-nuclease (TaqMan) assay and employing the comparative C T method. Expression of the FATP gene was normalized for expression of the 18 S gene. Results represent mean Ϯ S.E. of three separate experiments. ૺ, p Ͻ 0.05 (ANOVA) compared with ␤-galactosidase-overexpressing cardiomyocytes.
FIG. 8. GH-induced alterations in genetic markers of hypertrophy in GHR-overexpressing rat neonatal cardiomyocytes. GHRoverexpressing cardiomyocytes were exposed to rGH (500 ng/ml) for the indicated time periods, and total RNA was extracted for analysis of c-fos, myosin light chain 2, skeletal ␣-actin, or ANF gene expression using fluorescent 5Ј-nuclease (TaqMan) assays and employing the comparative C T method. Expression of the specific mRNA was normalized for expression of the 18 S gene. Results represent the mean Ϯ S.E. of three independent experiments. ૺ, p Ͻ 0.05, not significant (ns) ϭ p Ͼ 0.05 (ANOVA) compared with cells exposed to vehicle only.
toward GH-induced perturbations in insulin-responsive glucose transport cannot be excluded (59). Our results indicate that GHR increases in the adult as compared with the neonatal heart and that GH can increase the fatty acid uptake by cardiomyocytes expressing GHR. In higher mammals, the ability of cardiomyocytes to transport fatty acids increases with age, and there is an adaptive transition from glucose to fatty acids as the preferred oxidative fuel for the mammalian heart during maturation from the fetus to the adult (24). Likewise, in the rodent there is a 3-fold increase in fatty acid oxidative rates of the heart muscle during maturation from the neonate to the adult and a contemporaneous increase in expression of fatty acid transport proteins (54,60). Since GHR expression increases postnatally, we postulate that this ontogenic profile of expression of GHR in the heart plays a role in the transition of metabolism of the mammalian heart, from dependence on glu-cose in the fetus to preferential use of fatty acids postnatally.
Cardiac hypertrophy is a hallmark of GH excess in both animal models and humans (1). In contrast to prior studies that concluded that GH does not stimulate hypertrophy of cardiomyocytes, our results indicate that GH does induce hypertrophy of isolated cardiomyocytes. Switches in isoform expression of actin and myosin are observed with many forms of cardiac hypertrophy, with each hypertrophic stimulus eliciting its characteristic profile of actin/myosin isoforms. In general, cardiac hypertrophy is associated with induction of fetal isoforms (i.e. skeletal ␣-actin, ␤-myosin heavy chain), expression of other fetal proteins (i.e. ANF), and up-regulation of constitutively expressed genes (i.e. myosin light chain 2, cardiac ␣-actin) (61). In the current model system, GH increased the expression of c-fos, myosin light chain 2, and skeletal ␣-actin but did not alter the expression of ANF. It is of interest to note that combined administration of IGF-1 and GH in the intact animal failed to alter expression of skeletal ␣-actin or ANF gene expression (62). Similarly, IGF-1 decreased ANF gene expression in isolated adult rat cardiomyocytes in culture (63). The differential effects of GH on fetal gene expression with induction of skeletal ␣-actin but not of ANF support the hypothesis that activation of the GH/IGF-1 axis induces a more physiological or nonpathological form of cardiac hypertrophy (62). However, further studies are needed to substantiate this hypothesis, and the current experimental paradigm of adenovirus-mediated GHR-overexpression will facilitate further analysis of the effects of GH/IGF-1 in both neonatal and adult cardiomyocytes.
The somatomedin hypothesis states that the anabolic and growth-promoting effects of GH are principally mediated by stimulation of production of IGF-1 by the liver (endocrine action) and peripheral tissues (autocrine/paracrine action) (16). Contemporary studies using techniques such as targeted disruption and conditional tissue-specific knockout of specific proteins involved in this pathway have begun to reexamine this hypothesis (23,64). Most recently, Lupu et al. (23), using a total knockout model of GHR, demonstrate that there must be independent and additive effects of GH and IGF-1 on overall and specific tissue growth. Our results complement these findings, since observed changes in the size of cardiomyocytes and their metabolic profile in response to GH were not accompanied by changes in IGF-1 gene expression. Furthermore, using radioimmunoassay, we were unable to detect changes in concentration of IGF-1 in the conditioned medium (data not shown), suggesting that in this model system the effects of GH are, at least in part, independent of IGF-1. It is also noteworthy that cardiac hypertrophy has been observed in IGF-1 knockout mice (65). In these IGF-1 null mice, circulating levels of GH are elevated, and one of the mechanisms resulting in cardiac hypertrophy could be direct action of the elevated circulating levels of GH on the heart, consistent with our results in the isolated neonatal cardiomyocyte expressing GHR.
Lupu et al. (23) demonstrated that divergence of the growth patterns of normal and GHR knockout animals occurs only after 2 weeks of life. Our studies extend the observations of Lupu et al. (23) by demonstrating that the relative lack of GH's effect on the neonatal heart is, at least in part, due to the paucity of GHR in the neonatal heart and that expression of GHR increases significantly in the adult heart. Our model of neonatal cardiomyocytes transduced with the GHR-expressing adenovirus thus tends to recapitulate the scenario in the adult cardiomyocyte with respect to GHR expression. Hence, our studies support the hypothesis that when the heart exhibits sufficient GHR expression, as in the adult, it is a target of direct GH action. This report describes a model based on rat neonatal cardiomyocytes. Cardiomyocytes from immature ani- FIG. 9. GH-induced IGF-1 gene expression in GHR-overexpressing rat neonatal cardiomyocytes. A, ribonuclease protection assay. RNA was isolated from neonatal cardiomyocytes (uninfected (lanes 2 and 3), infected with ␤-galactosidase-overexpressing adenovirus (lanes 4 and 5), or infected with GHR-overexpressing adenovirus (lanes 6 and 7)), and the cells were either exposed for 24 h to rGH (500 ng/ml) (lanes 3, 5, and 7) or vehicle only (lanes 2, 4, and 6). Lane 1, size markers; lanes 8 and 9, control RNA from adult rat heart (lane 8) or liver (lane 9). 18 S RNA was also assayed to ensure equality of RNA loading. Three protected bands are visible: a major band of 224 bases and a minor band of 100 bases of IGF-Ia mRNA (30), both being specific protections to different regions of the probe, and an 80-base band corresponding to the 18 S rRNA. Densitometric analysis of the protected bands confirmed the absence of effect of GH on IGF-1 mRNA expression in GHR overexpressing cells (lane 6 compared with lane 7). A, fluorescent 5Ј-nuclease assay. ␤-Galactosidase-overexpressing (open bars) or GHR-overexpressing (solid bars) cardiomyocytes were exposed to rGH (500 ng/ml) for the indicated time periods, and total RNA was extracted for analysis of IGF-1 gene expression using a fluorescent 5Ј-nuclease (TaqMan) assay and employing the comparative C T method. Expression of the IGF-1 gene was normalized for expression of the 18 S gene. Results represent mean Ϯ S.E. of three separate experiments. Not significant (ns) ϭ p Ͼ 0.05 (ANOVA) compared with ␤-galactosidaseoverexpressing cardiomyocytes. mals (embryonic, fetal, and neonatal) are easier to isolate and culture and thus have been widely used in in vitro studies. Adult ventricular cardiomyocytes, however, differ from immature cells. Thus, adult cells are morphologically different in that they have a larger, more developed transverse tubular system, grow more slowly in culture, and have a restricted potential for differentiation (66). These differences should be taken into account when interpreting the results of the present study obtained with neonatal cardiomyocytes, and future studies carried out with adult cardiomyocytes will shed light on the effects of GH on the adult heart. It is noteworthy that our strategy of using adenovirus-mediated overexpression of GHR will be suitable for use with adult cardiomyocytes (67).
In summary, we have established a model suitable for examining the effects of GH on cardiomyocytes expressing GHR. Our results indicate that GH induces hypertrophy of these cells and causes alterations in their metabolic profile independently of IGF-1. These studies add to evidence that the effects of GH and IGF-1 on somatic growth may be additive yet independently mediated and provide the first direct proof of independent actions of GH on the heart.