Transgenic Mice Overexpressing Growth Hormone (GH) Have

0013-7227/08/$15.00/0 Printed in U.S.A.

Endocrinology 149(11):5758 –5769 Copyright © 2008 by The Endocrine Society doi: 10.1210/en.2008-0346

Transgenic Mice Overexpressing Growth Hormone (GH) Have Reduced or Increased Cardiac Apoptosis through Activation of Multiple GH-Dependent or -Independent Cell Death Pathways Fausto Bogazzi, Dania Russo, Francesco Raggi, Federica Ultimieri, Claudio Urbani, Maurizio Gasperi, Luigi Bartalena, and Enio Martino Department of Endocrinology and Metabolism (F.B., D.R., F.R., F.U., E.M.), University of Pisa, 56124 Pisa, Italy; SantAnna School of Advanced Studies (C.U.), 56127 Pisa, Italy; Department of Endocrinology (M.G.), University of Molise, 86100 Campobasso, Italy; and Department of Clinical Medicine (L.B.), University of Insubria, 21100 Varese, Italy GH has antiapoptotic effects in cardiac or noncardiac cell lines; however, increased apoptosis has been found in myocardial samples of patients with acromegaly. The aim of this study was to investigate cardiac apoptosis and underlying molecular mechanisms in transgenic mice overexpressing bovine GH [acromegalic mice (Acro)] aged 3 or 9 months. Cardiomyocyte apoptosis was evaluated by terminal deoxynucleotidyl transferase assay and annexin V; expression of pro- or antiapoptotic proteins was assessed by Western blot. Specificity of GH action was confirmed using a selective GH receptor antagonist. Apoptosis was lower in 3-month-old Acro than in controls; reduction was abolished by a GH receptor antagonist. The effects of GH were consistent with an antiapoptotic phenotype (increased Bcl2 and Bcl-XL and reduced Bad and cytochrome c levels, leading to lower activation of caspase-9 and caspase-3). In contrast, apo-

C

ARDIAC HYPERTROPHY IS a common finding of systemic manifestations of GH/IGF-I excess in patients with acromegaly (1–3). Acromegaly is often complicated by hypertension, diabetes, or dyslipidemia, all well established risk factors for cardiovascular disease; however, many observations support the concept that GH/IGF-I excess be independently responsible for a peculiar acromegalic cardiomyopathy characterized by biventricular hypertrophy and diastolic dysfunction (4 – 6). Diastolic dysfunction may progress to systolic impairment and eventually to heart failure (2). Transgenic mice overexpressing bovine GH (bGH) have extensively been used for studying cardiac features of GH excess because they develop a concentric hypertrophic cardiac phenotype with impaired cardiac function eventually

First Published Online July 10, 2008 Abbreviations: Acro, Acromegalic mice; Acro-Peg, acromegalic mice treated with pegvisomant; Apaf-1, apoptotic protease-activating factor-1; Bad, Bcl-2 antagonist of cell death; Bax, Bcl-2-associated X protein; Bcl2, B-cell CLL/lymphoma 2; Bcl-XL, Bcl-2-related protein long isoform; bGH, bovine GH; COXIV, cytochrome oxidase subunit IV; LV, left ventricle; Mek1, MAPK kinase 1; PI3K, phosphatidylinositol-3-kinase; P90RSK, ribosomial S6 kinase 90 kDa; RT, room temperature; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; Wt, wild type. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

ptosis was higher in 9-month-old Acro than in littermate controls; in addition, a GH receptor antagonist was without effect; the proapoptotic phenotype consisted in increased Bad, cytochrome c, caspase-9, and caspase-3. GH reduced apoptosis through p38 and p44/42 kinase pathways at young ages, whereas phosphatidylinositol-3-kinase was silent; on the contrary, the effects of GH on p38 and p44/42 kinase pathways were overcome by GH-independent stimuli in 9-month-old Acro. In addition, the antiapoptotic effect of GH was still present at this age as shown by phosphatidylinositol-3-kinase/ Akt pathway activation. In conclusion, chronic GH excess reduced apoptosis at a young age, whereas its antiapoptotic action was overwhelmed in older animals by GH-independent mechanisms, leading to increased cell death. (Endocrinology 149: 5758 –5769, 2008)

evolving to cardiac failure (7), thus reproducing many features of acromegalic patients. In many experimental models of hearts progressing to failure, apoptosis has been demonstrated, usually at low levels (8 –11), but it is unclear whether it has a pathogenic role or simply is an epiphenomenon of cardiac failure (12). In addition, it is not completely understood whether hypertrophic signaling is pro- or antiapoptotic. Physiological hypertrophy may lead to activation of signals, which protect against apoptosis, whereas pathological hypertrophy activates proapoptotic pathways (13). Cardiac remodeling is an important component of cardiac failure, and a link between hypertrophy and apoptosis has been suggested. In fact, overexpression of components of apoptotic pathways led to dilated cardiomyopathy (14, 15), indicating a role for apoptosis in cardiac failure. On the other hand, the GH/IGF-I axis and phosphatidylinositol-3-kinase (PI3K) pathway have been associated with cardiac hypertrophy and reduced apoptosis (12). In addition, several studies have shown a cardioprotective effect of IGF-I with decreased apoptosis mediated through PI3K and AKT during ischemia-reperfusion (16, 17). On the contrary, a dramatic increase in cardiac apoptosis has been reported in myocardial biopsies of acromegalic patients (18). In addition, a positive relationship between the degree of apoptosis and serum IGF-I concentrations or duration of acromegalic disease was reported (18). The latter findings differed from

5758

Bogazzi et al. • Apoptosis in Acromegalic Heart

Endocrinology, November 2008, 149(11):5758 –5769

5759

many cell culture studies showing an antiapoptotic effect of GH (19 –23). Thus, the question of whether sustained heart stimulation by GH has pro- or antiapoptotic effects still is unsettled. To address this issue, in this study, cardiac apoptosis and the underlying mechanisms were evaluated in transgenic mice overexpressing bGH aged 3 and 9 months, reflecting exposure to excess GH/IGF-I in young and elder ages, respectively. Materials and Methods Animals Transgenic mice overexpressing a coding sequence of bGH gene under the control of metallothionein promoter in the C57BL/6J ⫻ CBA genetic background were a generous gift of Dr. M. Bohlooly-y (University of Goteborg, Goteborg, Sweden) and have been described (24). C57BL/6J ⫻ CBA mice were purchased from Harlan Italy (Udine, Italy). The identity of transgenic mice was confirmed by PCR analysis of DNA from tail biopsy specimens using PCR primers located in the metallothionein promoter and in the bGH gene, as reported (25). Male animals, 3 and 9 months old, were used for the experiments, as representative of young and elder ages (7). The study groups included wild-type mice (Wt), acromegalic mice (Acro), acromegalic mice treated with pegvisomant (Acro-Peg), and acromegalic mice treated with adryamicin (doxorubicin chlorhydrate). Each group included five animals of each age; thus, 40 animals were included in the study. The environment of the animal rooms was controlled with a 12-h light, 12-h dark cycle, a relative humidity of 45–55%, and temperature of 20 C. Animals had free access to tap water and standard pellet chow. The study protocol was approved by the local Board for Animal Experimentation at the University of Pisa and conducted in accordance with the National Institutes of Health guidelines for the use of experimental animals (26).

Treatment Some acromegalic animals (see animals) were treated with pegvisomant (Pfizer, Rome, Italy), a specific antagonist of GH receptor. The administered dose (0.1 mg/daily, sc for 15 d) was chosen on the basis of the data previously reported (27) and of preventive dose- and timeresponse experiments (the lowest dose of pegvisomant associated with normalization of serum IGF-I concentrations was chosen). Effectiveness of pegvisomant was evaluated by measuring serum IGF-I concentrations before and at the end of treatment. Other animals were treated with adryamicin (Sigma-Aldrich, Milan, Italy), a proapoptotic drug (28). The drug was administrated at various doses (0.05, 0.5, and 2 mg/kg) for 24 h as reported (29). TABLE 1. Body and heart weight of the animals included in the study Age (months) 3

Body weight (g) Wt Acro Acro-Peg Heart weight (g) Wt Acro Acro-Peg Heart weight/body weight Wt Acro Acro-Peg

9

26 ⫾ 4 47 ⫾ 3 46 ⫾ 4

37 ⫾ 4 58 ⫾ 3 60 ⫾ 2

0.12 ⫾ 0.01 0.31 ⫾ 0.02 0.33 ⫾ 0.01

0.13 ⫾ 0.03 0.38 ⫾ 0.04 0.36 ⫾ 0.02

0.0046 0.0065 0.0071

0.0035 0.0065 0.0060

Data are shown as mean ⫾ SD; differences between Acro and Wt at any age were statistically significant (P ⬍ 0.03) as assessed by ANOVA.

FIG. 1. Histology of the heart of acromegalic animals. LVs were fixed in formalin, embedded in paraffin, and then examined using a light microscope. Serial 4-␮m tissue sections were deparaffinized and stained with hematoxylin and eosin or reticulin. Fiber diameter was determined by calculating the mean of the shortest and longest diameters as reported by Lund and Tomanek (30).

Assays Serum IGF-I concentrations were measured using a commercial kit (Diagnostic System Laboratories, Webster, TX). Sensitivity was 21 ng/ml; intra- and interassay variations were 12 and 9%, respectively.

Tissue samples Body and heart weights of Acro, Acro-Peg, and Wt were determined after animals were killed (bleeding and cervical dislocation done under ether anesthesia); ventricles were separated and then immediately frozen in liquid nitrogen until further examination.

Histology Left ventricles (LV) were fixed in 10% formalin, embedded in paraffin, and then subjected to light-microscopic examinations. Serial 4-␮m tissue sections were deparaffinized and stained with hematoxylin and eosin or reticulin. Fiber diameter was determined by calculating the mean of the shortest and longest diameters as reported (30). LV hypertrophy was defined as increased cardiac fiber diameter and size as reported (7).

Apoptosis Apoptosis was evaluated by terminal deoxynucleotidyl transferasemediated dUTP nick end labeling (TUNEL) assay (Roche Diagnostic, Penzberg, Germany) and by annexin V (Santa Cruz Biotechnology, Santa Cruz, CA) in mice heart, according to manufacturers’ protocols. Fourmicrometer LV sections were prepared, mounted on slides, and used for TUNEL or annexin V analysis. Sections were counterstained with the nucleic acid-binding dye Sytox Orange (Molecular Probe, Invitrogen, Italy) to visualize the entire population of cell nuclei in the myocardial

5760



FIG. 2. Cardiomyocyte apoptosis. Prevalence of cardiomyocyte apoptosis was determined by TUNEL (A) and annexin V (B). LV slices, obtained from five transgenic mice overexpressing bGH (Acro), five Acro treated with GH receptor antagonist (Acro-Peg), and five littermate control (Wt), aged 3 and 9 months, were counterstained with Sytox orange to visualize all cell nuclei in the myocardial section. Results represent the mean ⫾ SD of data obtained in five animals for each group and represent the number of cells simultaneously stained with ␣-sarcomeric actin (␣-actin) and TUNEL (or annexin V, as appropriate) divided by total cells stained with Sytox. Five Acro and five Wt aged 3 months were treated with doxorubicin for 24 h (C). Apoptosis was evaluated by TUNEL and expressed as mean ⫾ SD of cells stained with ␣-sarcomeric actin and TUNEL simultaneously divided by total cells stained with Sytox. ␣-Sacromeric actin was used to visualize cardiomyocytes apoptosis in all experiments (see below). section (31, 32). To distinguish cardiomyocytes from non-cardiomyocyte cell types, myocardial sections were further labeled with ␣-sarcomeric actin (18). Tissues were treated with a fixation solution (4% paraformaldehyde in PBS, pH 7.4) for 20 min at room temperature (RT), washed with PBS for 30 min, and incubated with permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate) for 2 min on ice. After PBS washing, tissues were preincubated with 1:500 monoclonal anti-␣-sarcomeric actin an-

tibody (Sigma-Aldrich) for 1 h at RT and then incubated with TUNEL reaction mixture in a humidified atmosphere for 60 min at 37 C in the dark or with annexin V for 1 h at RT, respectively; the revelation system was a goat antirabbit-Alexa 546 (for annexin V) or antimouse-Alexa Fluor 546 (for ␣-sarcomeric actin) (both from Molecular Probes, Milan, Italy). Samples were directly analyzed by Nikon fluorescence microscope using a ⫻40 objective (Nikon Eclipse 80i; Nikon, Florence, Italy). Four thousand cells were counted for each sample. Images were digitally



5761

FIG. 3. Cytochrome c, caspase-9, and caspase-3 expression in cardiomyocytes from 3-month-old animals. Representative Western blot of mitochondrial and cytosolic extracts of LV obtained from 3-month-old transgenic mice (Acro), controls (Wt), or Acro treated with a GH antagonist (Acro-Peg) (A). The expression of cytosolic proteins was normalized for ␣-sarcomeric actin (␣-actin), whereas that of mitochondrial cytochrome c was normalized for COXIV. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group. Reduced cytosolic cytochrome c levels (A) were associated with lower caspase-9, caspase-3, and Apaf-1 degree of expression (B) in Acro than in Wt.

photographed and analyzed by ACT-2U Nikon software (Nikon, Italy) for numbering cell nuclei. Only cells labeled with TUNEL and identified as cardiomyocytes by ␣-sarcomeric actin were included in the cardiomyocyte apoptosis count. In each microscopic field, the number of cardiomyocyte nuclei stained by TUNEL was divided by the total number of Sytox orange-labeled nuclei and expressed as a percentage. The effect of GH on apoptosis was also evaluated in a rat cardiac cell line (H9c2, see supplemental data, published as supplemental data on The Endocrine Society’s Journals Online web site at http:// endo. endojournals.org).

Tissue extracts Tissue extracts were obtained by homogenizing LV with lysis buffer [150 mm NaCl, 10 mm Tris-HCl (pH 7.4), 1 mm EGTA, 1 mm EDTA, 1% Triton X-100, and protease inhibitors cocktail tablets (benzamidine, phenanthroline, aprotinin, leupeptin, pepstatin, and PMSF)]. After incubation on ice for 30 min and subsequent centrifugation, supernatants were stored at ⫺80 C. Protein concentration was measured by Bradford assay using the Bio-Rad reagent (Bio-Rad Laboratories, Hercules, CA). Cytosolic and mitochondrial proteins were isolated using a commercial

5762



FIG. 4. Myocardial level of proteins belonging to the p38, p 44/42, and PI3K pathways. Proteins belonging to the p38 (p38 and p53) and to p44/42 (Mek1, ERK1/2, and p90RSK) kinase pathway were lower in transgenic mice overexpressing bGH (Acro) than in controls (Wt); on the contrary, the expression of p110␣ and p110␥ (belonging to the PI3K pathway) in Acro was indistinguishable from that of Wt. Acro treated with a GH antagonist of GH receptor (Acro-Peg) had expression of most proteins belonging to the p38 and p44/42 pathway indistinguishable from that of Wt. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group. mitochondrial/cytosol fractioning buffer system (IMAGENEX; Analitica De Mori, Milan, Italy) and used for evaluating the expression of cytochrome c.

Western blotting Mitochondrial (20 ␮g), cytosolic (50 ␮g), or total (50 ␮g) myocardial protein extracts were resolved on a 12% SDS-PAGE, transferred onto nitrocellulose membrane, and stained with Ponceau red to verify the amount of proteins per lane. Transferred proteins were incubated overnight at 4 C in 50% Tris-buffered saline [200 mm Tris-HCl (pH 7.6) and 1.4 m NaCl] and 50% TTBS (Tris-buffered saline with 0.05% Tween 20) containing 5% nonfat dry milk and subsequently incubated with the appropriate primary antibody for 1 h at RT. After TTBS washing, a chicken antigoat IgG horseradish peroxidase-conjugated secondary antibody was added for 1 h at RT; positive proteins were detected using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech, Piscataway, NJ). Membranes were incubated at 70 C for 10 min in stripping buffer [5 mm Tris HCl (pH 6.8), 2% SDS, 67.5 ml ultrapure water, 0.8% ␤-mercaptoethanol] and reprobed for ␣-sarcomeric actin (for total and cytoplasmic proteins) or cytochrome oxidase subunit IV (COXIV) (for mitochondrial proteins) for loading normalization.

Films were scanned on densitometry (Bio-Rad Life Science, Milan, Italy), and band intensity was evaluated using GmbH software (Interfocus GmbH, Sonnenblumenring, Mering, Germany). Each sample value was normalized for loading errors (dividing by intensity of ␣-sarcomeric actin or COXIV, as appropriate); data were expressed as arbitrary units (A.U.) that represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein). The resulting values within the same mouse group but from different blots were then combined.

Antibodies The following antibodies were used: p38 (C-20) rabbit IgG antibody, p53 (FL-393) rabbit polyclonal IgG antibody, Bcl-2-associated X protein (Bax; N-20) rabbit polyclonal IgG antibody, Bcl-2 antagonist of cell death(Bad; C-7) mouse monoclonal IgG antibody, Bcl-2-related protein long isoform (Bcl-XL; H-5) mouse monoclonal IgG antibody, Bcl-2 (C-2) mouse monoclonal IgG antibody, cytocrome c (7H8) mouse monoclonal IgG antibody, apoptotic protease-activating factor 1 (Apaf-1; H-324) rabbit polyclonal IgG antibody, caspase-9 (H170) rabbit polyclonal IgG antibody, inhibitor of apoptosis protein 1/2 (c-IAP1/2; A-13) goat polyclonal IgG antibody, caspase-3 (E-8) rabbit polyclonal IgG antibody, AKT1 (B-1) mouse monoclonal IgG anti-



5763

FIG. 5. Myocardial level of pro- and antiapoptotic proteins in 3-month-old mice. A representative Western blot is shown of changes in the expression of cardiomyocytes levels of Bad, Bax, Bcl-2, and Bcl-XL in Acro. The level of expression of proapoptotic proteins (Bad and Bax) was lower and that of antiapoptotic proteins (Bcl-2 and Bcl-xL) higher in Acro than in controls (Wt). Treatment of Acro with a GH receptor antagonist (Acro-Peg) restored a tissue level of Bax, Bad, and Bcl-XL indistinguishable from that of Wt. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group.

body, MAPK kinase 1 (Mek1; H-8) mouse monoclonal IgG antibody, ERK3 (MAPK3, Erk1; K-23) rabbit polyclonal IgG antibody, and ribosomal S6 kinase 90 kDa (p90-RSK; H-60) rabbit polyclonal IgG antibody (all from Santa Cruz Biotechnology); monoclonal ␣-sarcomeric actin (clone 5C5) (Sigma-Aldrich); and anti-COXIV (Molecular Probes, Invitrogen, Milan, Italy) rabbit antimouse Alexa Fluor 546 (Molecular Probes).

Statistics Results were expressed as mean ⫾ sd. ANOVA was used to evaluate differences in the prevalence of cardiomyocyte apoptosis, degree of expression of pro- and antiapoptotic proteins, caspase-3, and caspase-9 among groups of animals. A P value ⬍ 0.05 was considered statistically significant.

Results Heart weights and histology and serum IGF-I concentrations

As expected, mean body weight and heart weight of Acro mice was greater than that of Wt (Table 1). Cardiac hyper-

trophy was determined by measuring the diameter of 30 fibers, as reported (25); cardiac hypertrophy was found at histology (Fig. 1) in 9-month-old Acro but not in 3-monthold Acro, confirming previous data (7, 25). Mean serum IGF-I concentrations were 293 ⫾ 44 ng/ml in Wt, 691 ⫾ 73 ng/ml in Acro, and 289 ⫾ 94 ng/ml in Acro-Peg (P ⬍ 0.04) and 278 ⫾ 56 ng/ml in Wt, 757 ⫾ 44 ng/ml in Acro, and 301 ⫾ 45 ng/ml in Acro-Peg mice in 3- and 9-month-old animals, respectively. Cardiac apoptosis in 3- and 9-month-old transgenic mice

Apoptosis was 0.29 ⫾ 0.014 and 0.25 ⫾ 0.008% in 3- and 9-month-old Wt, respectively, and 0.13 ⫾ 0.03 and 0.52 ⫾ 0.009% in young and elder Acro, respectively (P ⬍ 0.0001) when measured by TUNEL (Fig. 2A); the trend was indistinguishable (0.32 ⫾ 0.02 and 0.25 ⫾ 0.016% in 3- and 9 month-old Wt, respectively, and 0.11 ⫾ 0.015 and 0.44 ⫾ 0.012% in young and old Acro, respectively (P ⬍ 0.0001)

5764



FIG. 6. Cytochrome c, caspase-9, and caspase-3 expression in 9-month-old animals. A representative Western blot is shown of cytosolic and mitochondrial extracts of LVs obtained from 9-month-old transgenic mice (Acro), controls (Wt), or Acro treated with a GH receptor antagonist (Acro-Peg) (A). The expression of cytosolic proteins was normalized for ␣-sarcomeric myosin, whereas that of mitochondrial cytochrome c was normalized for COXIV. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group. Nine-month-old Acro had a proapoptotic phenotype consisting in higher cytosolic cytochrome c, (A) caspase-9, Apaf-1, and caspase-3 levels (B) than those of Wt. It is worth noting that a GH receptor antagonist had no effect on the expression of the above mentioned proteins.

when apoptosis was evaluated by annexin V (Fig. 2B). Overall, 3-month-old Acro had 55% lower apoptosis than Wt (P ⬍ 0.001), whereas 9-month-old Acro had about 80% higher apoptosis degree than littermate controls (P ⬍ 0.0001). Importantly, young Acro had a degree of cardiac apoptosis that was indistinguishable, after treatment with a GH receptor an-

tagonist, from that of littermate controls, suggesting a GHdependent mechanism in the low degree of apoptosis observed in young Acro. On the contrary, treatment with a GH receptor antagonist had no effect in 9-month-old Acro, suggesting GHindependent mechanisms of increased apoptosis in elder ages (Fig. 2, A and B).



5765

FIG. 7. Myocardial expression of proteins belonging to p38, p44/42 and PI3K pathways in 9-month-old animals. A representative Western blot shows the expression of proteins of the p38, p44/42, and PI3K pathways. Most proteins belonging to the p38 (p38 and p53), p44/42 (Mek1, ERK1/2, and p90RSK), and PI3K (p110␣, p110␥, and Akt) pathway were higher in transgenic mice (Acro) than in controls (Wt). Treatment with a GH receptor antagonist was without effect (Acro-Peg). Only the expression of p110␥ was restored to that of Wt after treatment with pegvisomant. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group.

Protective role of GH excess on doxorubicin-induced cardiotoxicity

To evaluate whether GH-dependent reduction in cardiac apoptosis might have a functional significance, Acro or Wt aged 3 months were exposed to increasing doses of doxorubicin chlorhydrate (0.05– 0.5–2 mg/kg body weight). Acro treated with doxorubicin at 0.05 or 0.5 mg/kg body weight had about 50 and 35% lower apoptosis, respectively, than Wt (P ⬍ 0.005). Higher doses of doxorubicin were associated with higher degree of apoptosis without a difference between Acro and Wt (8.5 ⫾ 0.71 and 9.1 ⫾ 0.72%, respectively, P value not significant; Fig. 2C). Mechanisms for attenuation of cardiac apoptosis in 3-month-old transgenic mice

Myocardial cytochrome c release. To investigate mechanisms involved in attenuation of cardiomyocyte apoptosis, we first evaluated the release of cytochrome c from mitochondria in 3-month-old animals. As shown in Fig. 3, mitochondrial levels of cytochrome c were higher in Acro than in Wt (P ⬍ 0.0001), whereas those of cytosolic cytochrome c followed an inverse pattern (P ⬍ 0.002). Importantly, Acro-Peg had mitochondrial and cytosolic cytochrome c levels indistinguishable from those of Wt (Fig. 3). Hence, young Acro had lower GH-dependent mitochondrial cytochrome c release. Caspase-9 and caspase-3 expression. Next, we investigated whether changes in cytochrome c expression affected caspase-9 and caspase-3 expression: Acro had lower caspase-9, Apaf-1 (the activator of caspase-9), and caspase-3 expression than that of Wt (P ⬍ 0.0001, Fig. 3). Blockade of GH action with a GH receptor antagonist prevented reduction of caspase-9, Apaf-1, and caspase-3 (Fig. 3). Myocardial levels of pro- and antiapoptotic proteins. To further explore the mechanisms underlying the reduced release of cytochrome c and caspases in young Acro, the level of expression of proteins belonging to the p38, p44/42, and PI3K pathway was then measured (Fig. 4). p38/p53/ERK1–2/p90-Rsk expression was lower in Acro than in littermate controls (P ⬍ 0.0001); these changes were

abolished in Acro-Peg, whereas expression of Mek1 was not affected by GH (Fig. 4). At variance with changes in the expression of proteins of the p38 and p44/42 kinase pathways, cellular levels of p110␣ and p110␥ (belonging to the PI3K pathway) did not change in young Acro (P value not significant), suggesting that the PI3K pathway was probably not involved in the regulation of apoptosis mediated by GH in Acro aged 3 months (Fig. 4). To further demonstrate the mechanisms involved in the reduced apoptosis in Acro, we measured the expression of several apoptotic proteins. Figure 5 shows a representative Western blot for proapoptotic Bad and Bax proteins, the level of which was lower in Acro than in Wt (P ⬍ 0.001). On the contrary, cellular levels of antiapoptotic Bcl2 and Bcl-XL proteins was significantly higher in Acro than in littermate controls (P ⬍ 0.0001; Fig. 5); it is worth noting that changes of either pro- or antiapoptotic proteins were abolished in Acro-Peg. Thus, it looks that convergence of p38 and p44/42 kinase pathways on coordinate variation of Bax/Bad/Bcl2/Bcl-XL expression, driven by GH, may lead to reduced cytochrome c release and lower caspases levels. Mechanisms for increased cardiac apoptosis in 9-month-old transgenic mice

Myocardial cytochrome c release. Acro aged 9 months had higher (P ⬍ 0.0001) cytoplasmic cytochrome c levels and lower mitochondrial cytochrome c (P ⬍ 0.0001) than Wt (Fig. 6), which persisted in Acro-Peg. Caspase-9 and caspase-3 expression. To investigate whether changes in cytochrome c levels were associated with corresponding variations in caspases expression, we measured the cellular levels of caspase-9 and caspase-3. Expression of caspase-9, caspase-3, and Apaf-1 was higher in elder Acro than in Wt (P ⬍ 0.0001) and not affected by a GH receptor antagonist (Fig. 6). Myocardial levels of pro- and antiapoptotic proteins. The expression of p38 was not different in elder Acro and Wt, whereas that of p53/ERK1–2/Mek1–2/p90-Rsk increased in Acro (P ⬍ 0.0001; Fig. 7). Changes of the latter proteins were not

5766



FIG. 8. Myocardial level of pro- and antiapoptotic proteins in 9-month-old mice. A representative Western blot shows changes in the expression of cardiomyocytes levels of Bax, Bcl-2, Bad, and Bcl-XL in transgenic mice (Acro), contrrols (Wt), or Acro treated with a GH receptor antagonist (Acro-Peg). The level of expression of antiapoptotic proteins (Bcl-2 and Bcl-xL) was higher in Acro than in Wt; proapoptotic proteins were either reduced (Bax) or increased (Bad). The expression of most proteins was not modified by a GH receptor antagonist; only the degree of the proapoptotic Bax increased in Acro-Peg, suggesting that GH still had antiapoptotic action also in elder Acro. Data are expressed as A.U., which represent the ratio between the intensity of the band of interest and the intensity of the band corresponding to the control protein and represent the mean ⫾ SD of measurements obtained in five animals of each group.

affected by a GH receptor antagonist (Fig. 7), suggesting that in Acro aged 9 months, other factors besides GH intervene in regulating expression of proteins belonging to p44/42 and p38 kinase pathway; alternatively, these pathways might develop GH resistance, becoming no longer GH dependent in elder transgenic mice. Expression of p110␣ protein (belonging to the PI3K pathway) was lower in Acro than in Wt (P ⬍ 0.0001) and not affected by pegvisomant. Conversely, expression of p110␥ protein was higher in Acro than in Wt, and this increase was abolished by a GH receptor antagonist (P ⬍ 0.0001). These findings suggest that the PI3K pathway might be involved in cardiac apoptosis in elder Acro (differently from young Acro) and that it is, at least in part, regulated by GH (Fig. 7). The proapoptotic phenotype of 9-month-old Acro consisted in lower Bax and higher Bad/Bcl2/Bcl-XL/Akt levels than in Wt (P ⬍ 0.0001; Fig. 8). These changes were also found in Acro-Peg, indicating a GH insensitivity. The only exception was represented by Bax, the reduction of which was abolished by treating Acro with a GH receptor antagonist. To summarize data, p38 and p44/42 pathways mediated the increased apoptosis in elder Acro through a GH-independent mechanism; GH still had antiapoptotic actions as

suggested by the lower proapoptotic Bax and higher antiapoptotic Akt levels. Discussion

The relationship between the GH/IGF-I axis and the heart is supported by several clinical and experimental observations (6). Receptors for GH and IGF-I have been demonstrated on cardiomyocytes (33), and IGF-I induces hypertrophy in rat cell cultures (13), protects from myocyte death after infarction (1), and reduces apoptosis of cardiomyocytes in dilated cardiomyopathy (34). Patients with acromegaly frequently have concentric biventricular hypertrophy; however, early stages of acromegalic cardiomyopathy are characterized by improved cardiac performance at rest and decreased performance only during exercise (4, 6). In the intermediate stage of the disease, biventricular hypertrophy and impaired diastolic filling are commonly found and may eventually progress to systolic failure (6). Hypertrophy, interstitial fibrosis, myofibrillar abnormalities, and monocyte necrosis are found in end-stage acromegalic cadiomyopathy and considered to be, at least partially, due to an increase in myocyte apoptosis associated with GH excess (18). However, the proapoptotic effects of longstand-



5767

FIG. 9. Proposed mechanisms regulating cardiomyocyte apoptosis driven by GH in 3- and in 9-month-old animals. GH modulates activity of p38 and p 44/42 kinase pathways but has no effect on PI3K pathway in Acro aged 3 months (A). These pathways (p38 and p44/42) likely converge at the level of Bax/Bcl2/Bad/Bcl-XL protein expression leading to an antiapoptotic phenotype (lower degree of the proapoptotic Bax and Bad proteins and higher degree of antiapoptotic Bcl2 and Bcl-XL proteins), leading to reduced cytochrome c release and low apoptosis. These changes, including the expression of caspase-9 and caspase-3, could reverse by blocking GH activity with a specific GH receptor antagonist. The net effect of GH excess in young ages consisted in reduced apoptosis. Nine-month-old Acro have higher cardiac apoptosis than the corresponding controls (B). The underlying molecular findings were the following: 1) expression of most proteins belonging to the p38 and p44/42 kinase pathways was higher in Acro than in Wt (at variance with Acro aged 3 months) leading to increased expression of the proapoptotic Bad protein; 2) cytosolic cytochrome c, caspase-9, and caspase-3 degree of expression was higher leading to increased apoptosis; 3) changes in pro- or antiapoptotic proteins were mostly unaffected by a GH receptor antagonist, suggesting that other stimuli superimpose to GH in the regulation of these pathways; 4) GH maintained its antiapoptotic action as suggested by the GH-dependent reduction of the proapoptotic Bax protein and by the increased Akt expression (through the PI3K pathway).

ing exposure to GH excess are at variance with the commonly reported protective role of IGF-I overexpression on cardiomyocyte death (19 –21). Our study showed that the degree of cardiomyocyte apoptosis changes during the lifespan of mice overexpressing bGH, being lower in young Acro and higher in elder Acro, respectively, than in littermate controls. Reduced apoptosis

in young Acro was due to GH because it was abolished by treatment with a specific GH receptor antagonist. On the other hand, the increased cardiomyocyte apoptosis was not inhibited by blocking GH receptor in elder Acro, indicating that at this stage, GH excess does not per se have a proapoptotic effect on heart. It is worth noting that, at variance with young Acro, elder Acro had cardiac hypertrophy at histol-

5768


ogy, which might be triggered by activation of PI3K pathways (35). Overall, adaptive hypertrophy results from exercise conditioning, whereas maladaptive hypertrophy develops in response to excess hemodynamic load (36). Underlying mechanisms of the former include activation of p110␣ subunit of the PI3K pathway by the GH/IGF-I system (37). On the contrary, activation of the p110␥ is mainly involved in maladaptive hypertrophy (37). The fact that elder Acro had a GH-dependent increase in p110␥ levels leads to the speculation that longstanding GH excess might be involved in maladaptive hypertrophy; alternatively, chronic GH excess stimuli might activate signaling pathways not involved in young ages. However, some pro- and antiapoptotic proteins maintained GH sensitivity also in elder Acro, suggesting that GH continues its antiapoptotic action and other proapoptotic factors (for example, triggered by hypertrophic hearts) superimposed on those regulated by GH. The reduced apoptosis, driven by GH in young Acro, was mediated mainly by p38 and p44/42 kinase pathways. These two pathways merged to reduce mitochondrial release of cytochrome c through cellular variations in Bax/Bad/Bcl2/ Bcl-XL cellular proteins, favoring an antiapoptotic phenotype. This pattern was reported in an experimental model of skeletal muscle atrophy (38); GH prevented muscle cell apoptosis in rats with cardiac heart failure through increase of Bcl2 and reduced Bax and caspase levels (38). The association between GH, cardiomyocyte apoptosis, and variation of proand antiapoptotic proteins in young animals was further demonstrated by treating Acro with a specific GH receptor inhibitor that abolished the aforementioned changes. However, it cannot be excluded that production of IGF-I might partially contribute to the antiapoptotic effect of GH; in fact, Acro-Peg had serum IGF-I levels indistinguishable from those of Wt due to the competitive action of pegvisomant on GH receptor. Thus, blocking GH receptor was associated with reduced GH action, including IGF-I production: the two aspects could not be separated in the present study. It is intriguing that PI3K proteins were not modified by GH in young Acro; this pathway has been involved in IGF-I-mediated survival of cardiomyocytes and in heart hypertrophy (35), which is not evident at histology in mice of that age. It is tempting to speculate that long-lasting GH signaling becomes somehow less efficient in cardiomyocytes or is overcome by other (proapoptotic) signals deriving, for example, from hypertrophic myocardial cells. In fact, sustained GH signaling was associated with loss of GH sensitivity of antiapoptotic proteins (Bcl-2 and Bcl-XL), which in turn contributed to an increased mitochondrial release of cytochrome c. The ratio between pro- and antiapoptotic proteins of the Bcl-2 family conveys a final decision for the mitochondrial release of cytochrome c (39); it is interesting to note that Acro/Wt ratio of antiapoptotic proteins was very similar in young and elder animals; in contrast, the level of expression of the proapoptotic protein Bad was very low in young Acro and highly increased in elder Acro, suggesting that regulation of this protein might be a crucial step in the regulation of GH-dependent cardiac apoptosis. It has been proposed that the cardioprotective effect of IGF-I is mediated through PI3K-Akt signaling (35), and Akt was reported to reduce apoptosis by inactivating Bad or


caspase-9 (40). However, these mechanisms have been questioned, and other pathways, such as Akt-mediated changes in nitric oxide production, have been suggested (41). We observed that Akt expression was higher in 9-month-old Acro than in Wt, suggesting that the PI3K/Akt pathway was active although not sufficient to lower apoptosis. In addition, the PI3K/Akt signaling pathway is activated in response to both adaptive and maladaptive heart growth (37), likely depending on the intensity of hormonal and hypertrophic stimuli as well as on the relative proportion of the downstream [mammalian target of rapamycin (mTOR) and glycogen synthase kinase (GSK)-3␤] effectors that may also be activated independently by GH (37). The down-regulation of MAPKs (ERK and Mek in 3-month-old Acro (with absent LV hypertrophy) strongly support the antiapoptotic effect of GH; on the contrary, the GH-independent increase of these MAPKs in hypertrophic hearts of elder Acro is in keeping with the reported recruitment of stress-activated MAPKs in maladaptive hypertrophy (37, 42). Finally, our results might explain, at least in part, the reported increased apoptosis in the hearts of patients with longstanding acromegaly (18); all patients of that series had end-stage acromegalic cardiomyopathy in which, based on the present data, a GH-independent apoptosis may occur. It is interesting to note that treatment with a GH receptor antagonist for 2 wk was associated with changes in apoptosis degree and in cardiomyocyte levels of several apoptogens but was not sufficient to modify heart weight. This was likely due to the short course of pegvisomant, which in patients with acromegaly, reduced LV hypertrophy after at least a 6-month-course treatment and in most patients after 18 months (43). In conclusion, the results of this study suggest that the degree of apoptosis changes during the lifespan of acromegalic mice and switches from lower to higher levels. The cardioprotective effect of GH is evident in young ages, whereas in elder ages, it is likely overwhelmed by GH-independent proapoptotic signals (Fig. 9) possibly deriving from hypertrophied myocardial cells. However, the latter mechanisms presently remain speculative and need further research. Acknowledgments We thank Professor A. Pinchera (University of Pisa) for his continuous encouragement and advice. Received March 13, 2008. Accepted July 1, 2008. Address all correspondence and requests for reprints to: Fausto Bogazzi, M.D., Ph.D., Department of Endocrinology and Metabolism, University of Pisa, Ospedale Cisanello, Via Paradisa 2, 56124 Pisa, Italy. E-mail: [email protected] or [email protected]. This work was supported by grants from the University of Pisa (Fondi d’Ateneo) to F.B., from the Ministry of Education, University and Research (MUR, Rome) to E.M., and from the University of Insubria at Varese (Fondi d’Ateneo) and from MUR to L.B. Disclosure Statement: There is no conflict of interest that would prejudice impartiality of reported data. F.B., D.R., F.R., F.U., C.U., M.G., L.B., and E.M. have nothing to declare.

References 1. Lie JT, Grossman SJ 1980 Pathology of the heart in acromegaly: anatomic findings in 27 autopsied patients. Am Heart J 100:41–52


2. Sacca` L, Cittadini A, Fazio S 1994 Growth hormone and the heart. Endocr Rev 15:555–573 3. Melmed S 2006 Medical progress: acromegaly. N Engl J Med 355:2558 –2573 4. Fazio S, Cittadini A, Cuocolo A, Merola B, Sabatini D, Colao A, Biondi B, Lombardi G, Sacca` L 1994 Impaired cardiac performance is a distinct feature of uncomplicated acromegaly. J Clin Endocrinol Metab 79:441– 446 5. Ciulla M, Arosio M, Barelli MV, Paliotti R, Porretti S, Valentini P, Tortora G, Buonamici V, Moraschi A, Capiello V, Magrini F 1999 Blood pressureindependent cardiac hypertrophy in acromegalic patients. J Hypertens 17: 1965–1969 6. Colao A, Ferone D, Marzullo P, Lombardi G 2004 Systemic complications of acromegaly: epidemiology, pathogenesis, and management. Endocr Rev 25: 102–152 7. Bollano E, Omerovic E, Bohlooly-y M, Kujacic V, Madhu B, Tornell J, Isaksson O, Soussi B, Schulze W, Fu ML, Matejka G, Waagstein F, Isgaard J 2000 Impairment of cardiac function and bioenergetics in adult transgenic mice overexpressing the bovine growth hormone gene. Endocrinology 141: 2229 –2235 8. Bryant D, Becker L, Richardson J, Shelton J, Franco F, Peshock R, Thompson M, Giroir B 1998 Cardiac failure in transgenic mice with myocardial expression of tumor necrosis factor-␣ (TNF). Circulation 97:1375–1381 9. Kubota T, McTiernan CF, Frye CS, Slawson SE, Lemster BH, Koretsky AP, Demetris AJ, Feldman AM 1997 Dilated cardiomyopathy in transgenic mice with cardiac specific overexpression of tumor necrosis factor-␣. Circ Res 81: 627– 635 10. Condorelli G, Morisco C, Stassi G, Notte A, Farina F, Sgaramella G, de Rienzo A, Roncarati R, Trimarco B, Lembo G 1999 Increased cardiomyocyte apoptosis and changes in proapoptotic and antiapoptotic genes bax and bcl-2 during left ventricular adaptations to chronic pressure overload in the rat. Circulation 99:3071–3078 11. Li Z, Bing OHL, Long X, Robinson KG, Lakatta EG 1997 Increased cardiomyocyte apoptosis during the transition to heart failure in the spontaneously hypertensive rat. Am J Physiol 272:H2313–H2319 12. Kang PM, Izumo S 2003 Apoptosis in heart: basic mechanisms and implications in cardiovascular diseases. Trends Mol Med 9:177–182 13. Eto Y, Yonekura K, Sonoda M, Arai N, Sata M, Sugiura S, Takenaka K, Gualberto A, Hixon ML, Wagner MW, Aoyagi T 2000 Calcineurin is activated in rat hearts with physiological left ventricular hypertrophy induced by voluntary exercise training. Circulation 101:2134 –2137 14. Yussman MG, Toyokawa T, Odley A, Lynch RA, Wu G, Colbert MC, Aronow BJ, Lorenz JN, Dorn 2nd GW 2002 Mitochondrial death protein Nix is induced in cardiac hypertrophy and triggers apoptotic cardiomyopathy. Nat Med 8:725–730 15. Wencker D, Chandra M, Nguyen K, Miao W, Garantziotis S, Factor SM, Shirani J, Armstrong RC, Kitsis RN 2003 A mechanistic role for cardiac myocyte apoptosis in heart failure. J Clin Invest 111:1497–1504 16. Aikawa R, Nawano M, Gu Y, Katagiri H, Asano T, Zhu W, Nagai R, Komuro I 2000 Insulin prevents cardiomyocytes from oxidative stress-induced apoptosis through activation of PI3 kinase/Akt. Circulation 102:2873–2879 17. Matsui T, Tao J, del Monte F, Lee KH, Li L, Picard M, Force TL, Franke TF, Hajjar RJ, Rosenzweig A 2001 Akt activation preserves cardiac function and prevents injury after transient cardiac ischemia in vivo. Circulation 104:330 –335 18. Frustaci A, Chimenti C, Setoguchi M, Guerra S, Corsello S, Crea F, Leri A, Kajstura J, Anversa P, Maseri A 1999 Cell death in acromegalic cardiomyopathy. Circulation 99:1426 –1434 19. Costoya JA, Finidori J, Moutoussamy S, Sea˜ris R, Devesa J, Arce VM 1999 Activation of growth hormone receptor delivers an antiapoptotic signal: evidence for a role of Akt in this pathway. Endocrinology 140:5937–5943 20. Gu Y, Zou Y, Aikawa R, Hayashi D, Kudoh S, Yamauchi T, Uozumi H, Zhu W, Kadowaki T, Yazaki Y, Komuro I 2001 Growth hormone signalling and apoptosis in neonatal rat cardiomyocytes. Mol Cell Biochem 223:35– 46 21. Segard HB, Moulin S, Boumard S, Augier de Cre´miers C, Kelly PA, Finidori J 2003 Autocrine growth hormone production prevents apoptosis and inhibits differentiation in C2C12 myoblasts. Cell Sign 15:615– 623 22. Baixeras E, Jeay S, Kelly PA, Postel-Vinay MC 2001 The proliferative and antiapoptotic actions of growth hormone and insulin-like growth factor-1 are


23.

24. 25. 26. 27.

28. 29.

30. 31. 32.

33. 34.

35. 36. 37. 38.

39. 40. 41.

42. 43.

5769

mediated through distinct signaling pathways in the Pro-B Ba/F3 cell line. Endocrinology 142:2968 –2977 Jeay S, Sonenshein GE, Kelly PA, Postel-Vinay MC, Baixeras E 2001 Growth hormone exerts antiapoptotic and proliferative effects through two different pathways involving nuclear factor-␬B and phosphatidylinositol 3-kinase. Endocrinology 142:147–156 Bohlooly-Y M, Carlson L, Olsson B, Gustafsson H, Andersson IJ, Tornell J, Bergstrom G 2001 Vascular function and blood pressure in GH transgenic mice. Endocrinology 142:3317–3323 Fu ML, Tornell J, Schulze W, Hoebeke J, Isaksson OG, Sandstedt J, Hjalmarson A 2000 Myocardial hypertrophy in transgenic mice overexpressing the bovine growth hormone (bGH) gene. J Intern Med 247:546 –552 Institute of Laboratory Animal Resources 1996 Guide for the care and use of laboratory animals. Washington, DC: National Academies Press Liao L, Dearth RK, Zhou S, Britton OL, Lee AV, Xu J 2006 Liver-specific overexpression of the insulin-like growth factor-I enhances somatic growth and partially prevents the effects of growth hormone deficiency. Endocrinology 147:3877–3988 Menna P, Salvatorelli E, Minotti G 2007 Doxorubicin degradation in cardiomyocytes. J Pharmacol Exp Therap 322:408 – 419 Spallarossa P, Altieri P, Garibaldi S, Ghigliotti G, Barisione C, Manca V, Fabbi P, Ballestrero A, Brunelli C, Barsotti A 2006 Matrix metalloproteinase-2 and -9 are induced differently by doxorubicin in H9c2 cells: the role of MAP kinases and NAD(P)H oxidase. Cardiovasc Res 69:736 –745 Lund DD, Tomanek RJ 1978 Myocardial morphology in spontaneously hypersensitive and aortic-constricted rats. Am J Anat 152:141–151 Yan X, Habbersett RC, Yoshida TM, Nolan JP, Jett JH, Marrone BL 2005 Probing the kinetics of SYTOX Orange stain binding to double-stranded DNA with implications for DNA analysis. Anal Chem 77:3554 –3562 Yan X, Habbersett RC, Cordek JM, Nolan JP, Yoshida TM, Jett JH, Marrone BL 2000 Development of a mechanism-based, DNA staining protocol using SYTOX orange nucleic acid stain and DNA fragment sizing flow cytometry. Anal Biochem 286:138 –148 Delafontaine P 1995 Insulin-like growth factor I and its binding proteins in the cardiovascular system. Cardiovasc Res 30:825– 834 Lee WL, Chen JW, Ting CT, Ishiwata T, Lin SJ, Korc M, Wang PH 1999 Insulin-like growth factor I improves cardiovascular function and suppresses apoptosis of cardiomyocytes in dilated cardiomyopathy. Endocrinology 140: 4831– 4840 Lanning NJ, Carter-Su C 2006 Recent advances in growth hormone signaling. Rev Endocrinol Metab Dis 7:225–235 Selvetella G, Hirrsch E, Notte A, Tarone G, Lembo G 2004 Adaptive and maladaptive hypertrophic pathways: points of convergence and divergent. Cardiovasc Res 63:373–380 Dorn GW, Force T 2005 Protein kinase cascades in the regulation of cardiac hypertrophy. J Clin Invest 115:527–537 Dalla Libera L, Ravera B, Volterrana M, Gobbo V, Della Barbera M, Angelini A, Danieli Betto D, Germinario E, Vescovo G 2004 Beneficial effects of GH/IGF1 on skeletal muscle atrophy and function in experimental heart failure. Am J Physiol Cell Physiol 286:C138 –C144 Gustafsson AB, Gottlieb RA 2007 Bcl-2 family members and apoptosis, taken to heart. Am J Physiol Cell Physiol 292:C45–C51 Cardone M H, Roy N, Stennicke HR, Salvesen GS, Franke TF, Stanbridge E, Frisch S, Reed JC 1998 Regulation of cell death protease caspase-9 by phosphorylation. Science 282:1318 –1321 Gao F, Gao E, Yue T, Ohlstein EH, Lopez BL, Christopher TA, Ma X 2002 Nitric oxide mediates the antiapoptotic effect of insulin in myocardial ischemia-reperfusion: the roles of PI3-kinase, Akt, and endothelial nitric oxide synthase phosphorylation. Circulation 105:1497–1502 Petrich BG, Wang Y 2004 Stress-activated MAP kinases in cardiac remodeling and heart failure; new insights from transgenic studies. Trends Cardiovasc Med 14:50 –55 Pivonello R, Galderisi M, Auriemma RS, De Martino MC, Galdiero M, Ciccarelli A, D’Errico A, Kourides I, Burman P, Lombardi G, Colao A 2007 Treatment with growth hormone receptor antagonist in acromegaly: effect on cardiac structure and performance. J Clin Endocrinol Metab 92:476 – 482

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.