Local myocardial overexpression of growth hormone attenuates postinfarction remodeling and preserves cardiac function

CARDIOVASCULAR

Local Myocardial Overexpression of Growth Hormone Attenuates Postinfarction Remodeling and Preserves Cardiac Function Vasant Jayasankar, MD, Lawrence T. Bish, BA, Timothy J. Pirolli, Mark F. Berry, MD, Jeffrey Burdick, BS, and Y. Joseph Woo, MD Departments of Surgery and Physiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania

Background. Ventricular remodeling with chamber dilation and wall thinning is seen in postinfarction heart failure. Growth hormone induces myocardial hypertrophy when oversecreted. We hypothesized that localized myocardial hypertrophy induced by gene transfer of growth hormone could inhibit remodeling and preserve cardiac function after myocardial infarction. Methods. Rats underwent direct intramyocardial injection of adenovirus encoding either human growth hormone (n ⴝ 9) or empty null vector as control (n ⴝ 9) 3 weeks after ligation of the left anterior descending coronary artery. Analysis of the following was performed 3 weeks after delivery: hemodynamics, ventricular geometry, cardiomyocyte fiber size, and serum growth hormone levels. Results. The growth hormone group had significantly better systolic cardiac function as measured by maximum left ventricular pressure (73.6 ⴞ 6.9 mm Hg versus control 63.7 ⴞ 7.8 mm Hg, p < 0.05) and maximum dP/dt (2845 ⴞ 453 mm Hg/s versus 1949 ⴞ 605 mm Hg/s, p < 0.005), and

diastolic function as measured by minimum dP/dt (ⴚ2,520 ⴞ 402 mm Hg/s versus ⴚ1,500 ⴞ 774 mm Hg/s, p < 0.01). Ventricular geometry was preserved in the growth hormone group (ventricular diameter 12.2 ⴞ 0.7 mm versus control 13.1 ⴞ 0.4 mm, p < 0.05; borderzone wall thickness 2.0 ⴞ 0.2 mm versus 1.5 ⴞ 0.1 mm, p < 0.001), and was associated with cardiomyocyte hypertrophy (6.09 ⴞ 0.63 ␮m versus 4.66 ⴞ 0.55 ␮m, p < 0.005). Local myocardial expression of growth hormone was confirmed, whereas serum levels were undetectable after 3 weeks. Conclusions. Local myocardial overexpression of growth hormone after myocardial infarction resulted in cardiomyocyte hypertrophy, attenuated ventricular remodeling, and improved systolic and diastolic cardiac function. The induction of localized myocardial hypertrophy presents a novel therapeutic approach for the treatment of ischemic heart failure. (Ann Thorac Surg 2004;77:2122–9) © 2004 by The Society of Thoracic Surgeons

H

geometry such as aneurysmectomy, infarct restraint, and cardiomyoplasty have resulted in reduced LV size, reduced wall stress, and in many cases improved contractile function in both animal models and clinical trials [7–13]. Taken together, these results suggest that improved ventricular geometry can result in better cardiac mechanics and function, with reduced morbidity and mortality. Human growth hormone (GH) is a 191-amino acid polypeptide whose receptor is expressed on cardiomyocytes [14]. Sustained elevated serum levels of GH, as in acromegaly, result in significant cardiac hypertrophy. Systemic treatment with GH after MI can induce myocardial hypertrophy and may thereby attenuate or reverse ventricular remodeling by reducing wall stress [15]. This effect has been demonstrated in animal models, in which systemic GH administration results in modest preservation of LV systolic function and geometry [15– 18]. To achieve this benefit, however, supraphysiologic serum levels of GH have been required for extended periods of time, and global cardiac hypertrophy has been observed. Because high serum levels of GH are associated with a diabetogenic state, hypertension, and in-

eart failure often occurs after myocardial infarction (MI) despite optimal medical management with ␤-blockade and angiotensin-converting enzyme (ACE) inhibition, with 22% of men and 46% of women disabled by congestive heart failure (CHF) within 6 years of their initial MI [1]. A central component of ischemic heart failure is ventricular remodeling with wall thinning and chamber dilation, a process that places the heart at increasing mechanical disadvantage and wall stress [2, 3]. The degree of post-MI remodeling predicts morbidity and mortality, further underscoring the importance of normal ventricular geometry to myocardial function and long-term outcome [2– 4]. In clinical trials of post-MI patients, treatment with ACE inhibitors or ␤-blockers resulted in inhibition or reversal of left ventricular (LV) remodeling, improved myocardial function, and decreased morbidity and mortality [5, 6]. Surgical approaches to improve ventricular Accepted for publication Dec 10, 2003. Address reprint requests to Dr Woo, Division of Cardiothoracic Surgery, University of Pennsylvania School of Medicine, 6 Silverstein Pavilion, 3400 Spruce St, Philadelphia, PA 19104; e-mail: wooy@uphs. upenn.edu.

© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc

0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.12.043

Abbreviations and Acronyms ACE ⫽ angiotensin-converting enzyme BSA ⫽ bovine serum albumin CHF ⫽ congestive heart failure GH ⫽ growth hormone LAD ⫽ left anterior descending coronary artery LV ⫽ left ventricular MI ⫽ myocardial infarction OCT ⫽ optimum cutting temperature embedding compound RVU ⫽ relative volume unit

creased risks of breast, prostate, and colon cancer, this strategy would not be desirable in clinical application [19, 20]. In addition, global LV hypertrophy, as seen in hypertension and aortic stenosis, can result in significant diastolic dysfunction with myocardial stiffening. We therefore postulated that local myocardial overexpression of GH by adenoviral gene transfer could be targeted to the borderzone area surrounding the infarct, the region most susceptible to abnormally high wall strain and subsequent thinning and remodeling [21]. This technique could result in high levels of GH in the myocardial borderzone area with minimal systemic exposure and adverse effects on other organs. We hypothesized that this strategy for local GH overexpression would induce localized myocardial hypertrophy and thereby prevent the progression to dilated failure after a large MI in a rat model. We chose to initiate GH overexpression 3 weeks after an ischemic myocardial insult in rats that, both in our experience and as reported by others, is a time point that coincides with the transition to decompensated heart failure in this model [22]. It was our intention to model a clinically relevant scenario for patients with ischemic cardiomyopathy.

Material and Methods All animals received humane care in compliance with the “Guide for the Care And Use of Laboratory Animals,” prepared by the Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council (National Academy Press, Washington, DC, 1996) and this research protocol was approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania.

Adenoviral Vector Construction Replication-deficient (E1, E3 deleted) adenoviral vectors containing the human GH transgene driven by the cytomegalovirus promoter (Ad.GH) were made and obtained from the Institute of Human Gene Therapy, Vector Core, Dr. James M. Wilson, University of Pennsylvania. Human GH shares 75% homology with rat GH at the mRNA level, and ⬃80% homology at the amino acid level [23]. Empty replication-deficient adenovirus containing no transgene (Ad.Null) was obtained as a control from the

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

2123

University of Iowa Gene Transfer Vector core (supported in part by the National Institutes of Health and the Roy J. Carver Foundation).

In Vivo Transgene Expression by Immunoblotting and Immunohistochemistry Successful virus delivery using our method of direct intramyocardial injection was confirmed using a replication-deficient E1, E3-deleted adenovirus containing the ␤-galactosidase gene under control of the cytomegalovirus promoter. X-gal staining was performed on histologic sections by fixation in 2% paraformaldehyde plus 0.1% glutaraldehyde followed by overnight incubation in X-gal reaction solution at 37°C (1 mg/mL X-gal, 4 mmol/L K-ferrocyanide, 2 mmol/L MgCl2 in phosphate-buffered saline, pH ⫽ 7.4). Sections were then counterstained with eosin. In addition, in vivo expression of GH transgene was confirmed by performing direct intramyocardial injections of either Ad.GH or Ad.Null control virus into 2 rats using the methodology described below. The hearts were procured after 1 week, protein samples were prepared from ventricular biopsy specimens and 50 ␮g of each were electrophoresed on a 12% sodium dodecyl sulfatepolyacrylamide gel after 10 minutes of denaturation at 100°C. Proteins were transferred to Immobilon-P membranes (Millipore, Bedford, MA) and immunoblotting was performed using a mouse anti-human GH monoclonal antibody (NeoMarkers, Fremont, CA). In addition, immunohistochemistry analysis was performed on the hearts of these animals. Histologic sectioning was performed as described below, and sections were incubated with mouse antihuman GH antibody (NeoMarkers, 1:500 in 5% bovine serum albumin [BSA]) followed by rhodamine-conjugated anti-mouse secondary antibody (1:200 in 5% BSA).

Animal Surgery Eighteen male Lewis inbred rats (250 to 300 g, Charles River Laboratories) were used for the main portion of this study. The Lewis strain of rat was chosen for the consistent infarct size, low mortality, and predictable development of post-MI heart failure after ligation of the left anterior descending coronary artery (LAD) [22]. Rats were anesthetized with intraperitoneal doses of ketamine (50 mg/kg) and xylazine (5 mg/kg), intubated, and mechanically ventilated with 0.5% isoflurane at the initial surgery. A left thoracotomy was performed through the fourth interspace and the LAD was ligated with a 7-0 polypropylene suture at the level of the left atrial appendage, a standardized location that is easily identified and reproducible. Myocardial ischemia was confirmed by blanching of the anterolateral left ventricle at the time of ligation. The animals were subsequently closed and allowed to recover. Three weeks after the infarction procedure, as heart failure ensued, repeat thoracotomies were performed. The animals were then randomly assigned to one of two experimental groups in a blinded fashion: delivery of Ad.GH (n ⫽ 9) (GH) or Ad.Null virus (n ⫽ 9) (control). All animals received

CARDIOVASCULAR

Ann Thorac Surg 2004;77:2122–9

2124

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

CARDIOVASCULAR

direct intramyocardial injections of 5 ⫻ 109 plaqueforming units of virus into the infarction borderzone region through a 30-gauge needle. A total volume of 250 ␮L was injected into six separate areas. The animals were then closed and allowed to recover for an additional 3 weeks.

Hemodynamic Measurements Three weeks after the virus injection procedure the animals were once again anesthetized, intubated, mechanically ventilated, and a repeat thoracotomy was performed. A 2F pressure-volume conductance microcatheter was inserted into the left ventricle through the apex of the heart (Millar Instruments, Houston, TX). Hemodynamic measurements were analyzed using the ARIA 1 Pressure Volume Analysis software (Millar Instruments) in a blinded fashion. The relative volume unit (RVU) is a calculated estimation of ventricular volume based on the conductance of blood in the LV cavity as measured by the conductance microcatheter, a method that has been validated as an accurate assessment of volume [24]. The heart was then arrested in diastole by injection of 0.1 mL KCl (1 mEq/mL), then the LV cavity was filled with Optimum Cutting Temperature (OCT) embedding compound retrogradely through the transected aortic root. Hearts were filled with OCT in a blinded fashion using a standardized delivery system utilizing a 5-mL syringe and an 18-gauge angiocatheter. Hearts were then frozen in liquid nitrogen.

Ventricular Geometry and Infarct Size Ten-micrometer sections were prepared from all hearts with a cryostat from the midway point between the site of LAD ligation and the apex of the heart, an easily identified, standardized, and reproducible location. Staining with hematoxylin and eosin was then performed. Geometry measurements were performed on digitized photomicrographs using Openlab image processing software (Improvision, Lexington, MA) with standards of known length and were obtained on two representative sections for each animal. For chamber size, LV diameter was recorded in both vertical and horizontal axes and averaged. For borderzone wall thickness, measurements were obtained on two separate areas for each section and averaged. The wall thickness of remote areas of the LV free wall was also measured in two separate areas in a similar manner. Infarct size was determined as the percent of infarct scar relative to the outer circumference of the LV free wall in all animals. A single investigator blinded to the treatment groups performed all measurements.

Laminin Immunohistochemistry and Fiber Size Measurements To measure cardiomyocyte fiber width, immunohistochemical staining for laminin, an extracellular matrix and basal lamina marker, was performed on histologic sections with a primary mouse anti-laminin antibody (NeoMarkers, 1:200 in 5% BSA) and a secondary rhodamineconjugated anti-mouse antibody (1:200 in 5% BSA).

Ann Thorac Surg 2004;77:2122–9

Myocyte fiber width was measured in four representative microscopic fields in the borderzone area for each animal. Fiber width was measured at the level of the myocyte nucleus, and approximately 100 fibers in the borderzone area were counted in each heart and the average width was calculated. All measurements were performed by a single investigator blinded to the treatment group.

Serum Assays To determine the early time course of systemic exposure to the human GH transgene, a separate group of 8 rats underwent direct intramyocardial injections with either Ad.GH (n ⫽ 4) or Ad.Null (n ⫽ 4), and blood was collected after 1 week. Blood was also collected from the experimental animals in the primary study groups before procurement of the heart at the time of sacrifice. Serum levels of human GH were measured using a humanspecific ELISA kit that does not detect rat GH (ICN Pharmaceuticals, Orangeburg, NY).

Statistical Analysis All values are expressed as mean ⫾ standard deviation (SD). The unpaired Student’s t test was used to calculate the statistical significance between the means of groups. A p value of less than 0.05 was considered to be significant.

Results Viral Delivery and Transgene Expression An adenoviral vector containing the ␤-galactosidase transgene was delivered by direct injection to a rat heart, the rat was sacrificed after 1 week, and X-gal staining was performed (Fig 1A). Transmyocardial expression, stained blue, is seen throughout the LV free wall area of injection. Western blotting for human GH performed on myocardial specimens obtained 1 week after virus injection revealed high levels of expression in the Ad.GH animal and no detectable human GH protein in the Ad.Null control sample (Fig 1B). Immunohistochemical staining for human GH was performed with an antibody that is specific for the human protein and does not detect rat hormone. High levels of tissue expression were seen throughout the area of injection in the Ad.GH animal, with only background staining visible in the control animal (Fig 1C). No staining was seen in remote, uninjected areas of myocardium in Ad.GH-treated animals (septum, right ventricle).

Hemodynamics and Cardiac Function The GH group had significant preservation of LV systolic function as measured by maximum LV pressure (GH 73.6 ⫾ 6.9 mm Hg versus control 63.7 ⫾ 7.8 mm Hg, p ⬍ 0.05) and maximum dP/dt (2845 ⫾ 453 mm Hg/s versus 1949 ⫾ 605 mm Hg/s, p ⬍ 0.005) (Fig 2). A sensitive measure of the contractile state of the heart is the slope of the maximum dP/dt versus end-diastolic volume curve,

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

2125

Fig 1. (A) Representative histologic section of a rat heart with X-gal staining 1 week after direct injection of an adenoviral vector containing the ␤-galactosidase gene. (B) Growth hormone (22 kD) immunoblotting results for intramyocardial injection of viral vector, with actin bands shown as a loading controls. (C) Immunohistochemical staining for human growth hormone. Significant staining is seen in only the GH animal. (Original magnification ⫻200.) (GH ⫽ growth hormone.)

which was significantly better in the GH group compared with control animals (1018 ⫾ 276 mm Hg · s⫺1RVU⫺1 versus 675 ⫾ 192 mm Hg · s⫺1RVU⫺1, p ⬍ 0.01) (Fig 2). No significant difference was noted in LV end-diastolic pressure between the groups (GH 5.7 ⫾ 1.6 mm Hg versus control 7.1 ⫾ 2.6 mm Hg, p ⬎ 0.05). Diastolic function was also improved in the GH group as measured by the minimum dP/dt (GH ⫺2520 ⫾ 402 mm Hg/sec versus control ⫺1500 ⫾ 774 mm Hg/sec, p ⬍ 0.01) and the time constant of isovolumic relaxation tau (14.3 ⫾ 2.1 ms versus 21.7 ⫾ 5.8 ms, p ⬍ 0.01) (Fig 3). As tau is an index of the time needed for isovolumic relaxation, a higher value represents impaired relaxation and diastolic filling.

Ventricular Geometry and Infarct Size The GH group had a significantly lower mean LV diameter compared with the control group (GH 12.2 ⫾ 0.7 mm versus control 13.1 ⫾ 0.4 mm, p ⬍ 0.05) (Fig 4). In addition, the LV borderzone wall thickness was significantly greater in the Ad.GH treated animals (GH 2.0 ⫾ 0.2 mm versus control 1.5 ⫾ 0.1 mm, p ⬍ 0.001), confirming localized hypertrophy and attenuated wall thinning with remodeling. Remote areas of the LV had similar thickness in both of the groups (GH 1.5 ⫾ 0.1 mm versus control 1.4 ⫾ 0.1 mm, p ⬎ 0.05), and infarct size was not significantly different (GH 49% ⫾ 9% versus control 44% ⫾ 5%, p ⬎ 0.05). Fig 2. Assessment of systolic cardiac function 6 weeks after MI. (A) Maximum LV pressure. (B) Maximum dP/ dt. (C) Representative cardiac contractility curves for a single animal in each experimental group. Contractility is measured as the slope of the maximum dP/dt versus enddiastolic volume curve. *p ⬍ 0.05, †p ⬍ 0.005. (GH ⫽ growth hormone; LV ⫽ left ventricular; MI ⫽ myocardial infarction; RVU ⫽ relative volume unit.)

CARDIOVASCULAR

Ann Thorac Surg 2004;77:2122–9

2126

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

Ann Thorac Surg 2004;77:2122–9

CARDIOVASCULAR

Fig 3. Assessment of diastolic cardiac function 6 weeks after MI. (A) Minimum LV dP/dt. (B) Tau, the time constant of isovolumic relaxation. *p ⬍ 0.01. (GH ⫽ growth hormone; LV ⫽ left ventricular; MI ⫽ myocardial infarction.)

Myocyte Fiber Size Myocyte fiber width in the borderzone was significantly higher in the GH group (GH 6.09 ⫾ 0.24 ␮m versus control 4.66 ⫾ 0.25 ␮m, p ⬍ 0.005) (Fig 5).

Serum Assays One week after direct intramyocardial injection of virus, human GH was undetectable (less than 0.05 ng/mL) in the Ad.Null control animals, and was only 4.0 ⫾ 0.8 ng/mL in the Ad.GH group. Serum human GH levels were undetectable (less than 0.05 ng/mL) in both the control and GH experimental groups at the study end point 6 weeks after LAD ligation.

Comment These experiments demonstrate that after MI in rats, local targeted overexpression of human GH by adenoviral gene transfer results in significant preservation of LV Fig 4. (A) Average left ventricular chamber diameter at the study end point. (B) Average borderzone wall thickness in each group at the conclusion of the study. (C) Representative histologic cross-sections of the midleft ventricle 6 weeks after initial surgery (hematoxylin & eosin staining). Arrows depict the diameter (D) and wall thickness (WT) measurement areas. *p ⬍ 0.05, †p ⬍ 0.001. (GH ⫽ growth hormone; LV ⫽ left ventricular.)

geometry with increased borderzone wall thickness and cardiomyocyte hypertrophy confirmed by fiber size determinations. The hypertrophy appeared limited to the borderzone area of injection, as remote myocardial areas did not differ significantly in thickness between the groups. The diminished post-MI remodeling was associated with significantly improved systolic cardiac function as measured by both load-dependent (maximum LV pressure and dP/dt) and load-independent (slope of maximum dP/dt versus end-diastolic volume) indices, while no significant difference was noted in end-diastolic volume. Despite borderzone hypertrophy and increased wall thickness, diastolic function was also markedly improved in Ad.GH-treated animals. We also found that systemic exposure to GH transgene is minimal using this technique, and that overexpression is limited to the area of injection. Importantly, no significant difference was noted in infarct size between the experimental groups. Ventricular remodeling is a central component of isch-

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

2127

Fig 5. (A) Average borderzone myocyte fiber width 6 weeks after initial surgery, demonstrating cardiomyocyte hypertrophy in the growth hormone (GH) animals. (B) Representative photomicrographs of immunofluorescent laminin staining from a single animal in each experimental group. *p ⬍ 0.005. (Original magnification ⫻200.)

emic heart failure, and LV size is a predictor of clinical outcome after MI [2– 4]. Successful pharmacologic treatment of CHF is associated with attenuation or reversal of the remodeling process, in addition to improved longterm outcomes [5, 6]. Surgical maneuvers to return the LV to a more normal geometric configuration have included passive epicardial restraint [11, 12], cardiomyoplasty [8, 9], and aneurysm plication or resection [7, 10]. These treatment strategies have resulted in improvement of end-diastolic volume, end-systolic volume, ejection fraction, and LV diameter to varying degrees. However, many patients continue to experience a progressive decline in cardiac function despite optimal medical management, and surgical modalities involve additional risks for operative morbidity and mortality. Therefore, alternative methods to improve ventricular geometry and function in ischemic heart failure are needed. Growth hormone is a potent stimulus for cardiomyocyte hypertrophy when secreted in excess [25]. In addition, receptor levels for GH may be higher in the heart than in other tissues [26]. We hypothesized that local overexpression of GH targeted to the borderzone area could result in preserved myocardial geometry with reduced dilation and improved wall thickness, and thereby significantly reduced LV wall stress and improved cardiac performance. Because diastolic function can be significantly impaired by myocardial hypertrophy and the resultant decrease in LV compliance, an important finding in this study is that Ad.GH treatment actually improved diastolic function as measured by both minimum dP/dt and tau. In contrast to the pathologic hypertrophy seen in hypertension and aortic stenosis, it appears that local hypertrophy induced by GH overexpression only in the infarct borderzone targeted the area most susceptible to mechanical stress and progressive thinning after an acute MI. The borderzone areas adjacent to infarcts are subject to supraphysiologic stress because they act as a bridge between normally contract-

ing myocardium and noncontractile infarct. The targeted hypertrophy induced by GH overexpression was sufficient to attenuate the remodeling process and preserve cardiac function after MI. Systemic GH administration after acute MI is associated with preserved LV function in rodent studies [15– 18]. However, supraphysiologic serum levels exceeding 1,700 ng/mL (normally less than 7 ng/mL) have been required to demonstrate positive effects. Systemic GH administration in human trials for cardiomyopathy have resulted in only marginal benefits with regard to improving LV geometry or function [27–30]. These clinical studies likely suffered from substantial underdosing, however, because serum GH was not significantly increased with the dosing regimens used (2 to 4 IU GH subcutaneous daily). Clinical dosing of GH is limited by its potential to induce a diabetogenic state or hypertension with prolonged exposure [20] and associated increased rates of breast, prostate, and in particular, colon cancer [19]. Therefore, our strategy of local overexpression through a gene transfer technique produces a high local level of GH within the targeted myocardial territory adjacent to the infarcts, while minimizing systemic exposure and potential morbidity. Serum levels of human GH were very low 1 week after virus administration (4.0 ⫾ 0.8 ng/mL) and undetectable at the time of sacrifice 3 weeks after injection. Transgene expression was limited to the targeted borderzone areas by immunohistochemistry, and no evidence of human GH staining was seen in remote myocardial territories. In our study virus delivery was performed 3 weeks after creation of a large MI. This time point was specifically chosen to maximize the clinical relevance while minimizing animal mortality. Previous work by our laboratory (unpublished data) and others using the Lewis rat model has shown that cardiac function remains stable between 2 and 4 weeks from infarction, but declines rapidly between 4 and 8 weeks [22]. Therefore, we in-

CARDIOVASCULAR

Ann Thorac Surg 2004;77:2122–9

2128

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

CARDIOVASCULAR

jected the virus during a period of relative stability of cardiac function, with maximal virus transgene expression, which occurs between 1 and 2 weeks after delivery [31], during the period of greatest decline in cardiac function. Our intent was to model a clinical scenario in which a patient presents with symptomatic heart failure and declining cardiac function and undergoes a specific therapy. To provide a reliable and reproducible means of achieving high local levels of cardiac transgene expression, we chose an adenoviral gene transfer strategy. Alternative methods of local overexpression include transcoronary delivery of adenoviral vectors, injection of naked plasmid DNA, myocardial implantation of drugeluting beads, or catheter-based myocardial injections. It has been shown that adenoviral vectors provide that highest level of gene transfer and expression compared with naked DNA and liposomal strategies [32]. We have found that local adenoviral injection provides a means to achieve consistently high local levels of growth factor in our model, and thus serves as an important investigative tool. As a control injection we chose to utilize Ad.Null virus to introduce the same viral construct as in the GH group but without the actual transgene, and thereby maximize the similarities in treatment as much as possible between the two experimental groups. One potential concern with this strategy is that the control virus may actually enhance the progression into heart failure and thereby confound analysis of any beneficial effects of Ad.GH. However, we have performed experiments with saline injections as an alternative control after LAD ligation, and found no difference in cardiac function or geometry between the Ad.Null and saline groups (unpublished data). Therefore, we believe that the differences between the experimental and control groups in our study are due solely to the overexpression of GH. Several important limitations to this study are worth noting. Although using blinded and standardized techniques we have shown an improvement in ex vivo cardiac geometry, an in vivo assessment using echocardiography or MRI would be helpful to demonstrate differences under physiologic loading conditions. In addition, these noninvasive modalities offer the opportunity to track postinfarction remodeling over the course of weeks and may allow further insight into the effects of local GH overexpression. Measurement of ejection fraction by echocardiography would have provided closedchest data on cardiac performance and may therefore be a better method to assess myocardial function. In conclusion, we have demonstrated in this study that targeted borderzone overexpression of GH in a rat model of postinfarction heart failure resulted in regional cardiomyocyte hypertrophy, diminished LV remodeling, preserved systolic and diastolic function, and minimal systemic exposure to GH protein. These results underscore the importance of normal ventricular geometry to cardiac function, and demonstrate that attenuation of remodeling can prevent functional decline. The induction of targeted regional myocardial hypertrophy is a novel approach to treating heart failure.

Ann Thorac Surg 2004;77:2122–9

This work was supported in part by grant HL69597 from the National Heart, Lung, and Blood Institute, National Institutes of Health (Bethesda, MD).

References 1. American Heart Association. 2003 Heart disease and stroke statistical update. Dallas, TX: American Heart Association, 2003. 2. White HD, Norris RM, Brown MA, et al. Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation 1987; 76:44 –51. 3. Pfeffer MA, Braunwald E. Ventricular remodeling after myocardial infarction: experimental observations and clinical implications. Circulation 1990;81:161–72. 4. Kostuk WJ, Kazamias TM, Gander MP, Simon AL, Ross J Jr. Left ventricular size after acute myocardial infarction: serial changes and their prognostic significance. Circulation 1973; 47:1174 –9. 5. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction: results of the survival and ventricular enlargement trial. The SAVE Investigators. N Engl J Med 1992;327:669 –77. 6. Teerlink J, Massie B. Beta-adrenergic blocker mortality trials in congestive heart failure. Am J Cardiol 1999;84:94R–102R. 7. Dymond DS, Stephens JD, Stone DL, Elliott AT, Rees GM, Spurrell RA. Combined exercise, radionuclide and hemodynamic evaluation of left ventricular aneurysmectomy. Am Heart J 1982;104:977–87. 8. Kass DA, Baughman KL, Pak PH, et al. Reverse remodeling from cardiomyoplasty in human heart failure. External constraint versus active assist. Circulation 1995;91:2314 –8. 9. Hagege AA, Desnos M, Fernandez F, et al. Clinical study of the effects of latissimus dorsi muscle flap stimulation after cardiomyoplasty. Circulation 1995;92:210 –5. 10. Batista RJV, Santos JLV, Takeshita N, et al. Partial left ventriculectomy to improve left ventricular function in endstage heart disease. J Cardiac Surg 1996;11:96 –7. 11. Oh JH, Badhwar V, Mott BD, Li CM, Chiu RC. The effects of prosthetic cardiac binding and adynamic cardiomyoplasty in a model of dilated cardiomyopathy. J Thorac Cardiovasc Surg 1998;116:148 –53. 12. Moainie SL, Guy TS, Gorman JH, et al. Infarct restraint attenuates remodeling and reduces chronic ischemic mitral regurgitation after postero-lateral infarction. Ann Thorac Surg 2002;74:444 –9. 13. Konertz W, Dushe S, Hotz H, et al. Safety and feasibility of a cardiac support device. J Card Surg 2001;16:113–7. 14. Adamafio NA, Towns RJ, Kostyo JL. Growth hormone receptors and action in BC3H-1 myocytes. Growth Regul 1991;1: 17–22. 15. Tivesten A, Caidahl K, Kujacic V, et al. Similar cardiovascular effects of growth hormone and insulin-like growth factor-I in rats after experimental myocardial infarction. Growth Horm IGF Res 2001;11:187–95. 16. Tajima M, Weinberg EO, Bartunek J, et al. Treatment with growth hormone enhances contractile reserve and intracellular calcium transients in myocytes from rats with postinfarction heart failure. Circulation 1999;99:127–34. 17. Yang R, Bunting S, Gillett N, Clark T, Jin H. Growth hormone improves cardiac performance in experimental heart failure. Circulation 1995;92:262–7. 18. Jin H, Yang R, Lu H, et al. Effects of early treatment with growth hormone on infarct size, survival, and cardiac gene expression after acute myocardial infarction. Growth Horm IGF Res 2002;12:208 –15. 19. Melmed S, Jackson I, Kleinberg D, Klibanski A. Current treatment guidelines for acromegaly. J Clin Endocrinol Metab 1998;83:2646 –52.

20. Fukuda I, Hizuka N, Murakami Y, et al. Clinical features and therapeutic outcomes of 65 patients with acromegaly at Tokyo Women’s Medical University. Intern Med 2001;40: 987–92. 21. Jackson BM, Gorman JH, Salgo IS, et al. Border zone geometry increases wall stress after myocardial infarction: contrast echocardiographic assessment. Am J Physiol Heart Circ Physiol 2003;284:H475–9. 22. Liu Y-H, Yang X-P, Nass O, Sabbah HN, Peterson E, Carretero OA. Chronic heart failure induced by coronary artery ligation in Lewis inbred rat. Am J Physiol 1997;272:H722–7. 23. Roskam WG, Rougeon F. Molecular cloning and nucleotide sequence of the human growth hormone structural gene. Nucleic Acids Res 1979;7:305–20. 24. Feldman MD, Erikson JM, Mao Y, et al. Validation of a mouse conductance system to determine LV volume: comparison to echocardiography and crystals. Am J Physiol Heart Circ Physiol 2000;279:H1698 –707. 25. Sacca L, Cittadini A, Fazio S. Growth hormone and the heart. Endocr Rev 1994;15:555–73. 26. Mathews LS, Enberg B, Norstedt G. Regulation of rat growth

JAYASANKAR ET AL GROWTH HORMONE ATTENUATES REMODELING

27. 28. 29.

30. 31.

32.

2129

hormone receptor gene expression. J Biol Chem 1989;264: 9905–10. Fazio S, Sabatini D, Capaldo B, et al. A preliminary study of growth hormone in the treatment of dilated cardiomyopathy. N Engl J Med 1996;334:809 –14. Genth-Zotz S, Zotz R, Geil S, Viogtlander T, Meger J, Darius H. Recombinant growth hormone therapy in patients with ischemic cardiomyopathy. Circulation 1999;99:18 –21. Spallarossa P, Rossettin P, Minuto F, et al. Evaluation of growth hormone administration in patients with chronic heart failure secondary to coronary artery disease. Am J Cardiol 1999;84:430 –3. Perrot A, Ranke MB, Dietz R, Osterziel KJ. Growth hormone treatment in dilated cardiomyopathy. J Card Surg 2001;16: 127–31. Gilgenkrantz H, Duboc D, Juillard V, et al. Transient expression of genes transferred in vivo into heart using firstgeneration adenoviral vectors: role of the immune response. Hum Gene Ther 1995;6:1265–74. Wong LF, Wells S, Murphy D. Gene transfer strategies for the physiologist. Exp Physiol 2000;85:735–45.

INVITED COMMENTARY Even though current treatments for heart failure have made significant progress in prolonging the survival of affected patients, complete recovery of ventricular function is still elusive. With the advent of novel intracellular targets involving cell contractility and survival and increasingly efficient gene transfer methodologies, gene based therapies are emerging as a promising therapeutic strategy. Different viral vectors have been used to transduce myocardial cells with great efficiency including recombinant adenoviruses, adeno-associated viruses, and lentivirus. However, these vectors have many limitations, including the immune response they elicit, the variability in specificity for the targeted organ, the length of expression and the difficulty in synthesizing some of these vectors at high enough titers. Heart failure is characterized at the cellular level by impaired contractility and abnormal calcium homeostasis. Restoration of a key enzyme that controls intracellular Ca2⫹ handling, the sarcoplasmic reticulum Ca2⫹ ATPase, induces functional improvement and prolonged survival in animal models of heart failure. Failing hearts are also characterized by increased cardiomyocyte programmed cell death (apoptosis), especially in models of ischemia and in post-myocardial infarction remodeling. Multiple studies suggest that apoptosis, and the caspase proteases central to this process, play an important role in the pathogenesis of heart failure. Thus, apoptosis is a relevant therapeutic target. An important anti-apoptotic pathway is the one involving insulin growth factor 1 (IGF1) and its downstream effectors PI-3 kinase and akt. Growth hormone (GH) induces an increase in IGF1 and should be effective in preventing cardiomyocyte apoptosis. Indeed, in small clinical studies, patients with dilated cardiomyopathy treated with systemic growth hormone showed improved left ventricular function and hemody-

© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc

namics. However, elevated levels of serum growth hormone are associated with diabetes and hypertension, while elevated IGF-1 levels have been associated with several forms of common cancers. Developing ways to induce local overexpression of growth factors may prevent these systemic complications. In this study, the local overexpression of GH in a rat model of myocardial infarction by adenoviral gene transfer resulted in attenuated ventricular remodeling and improvement in systolic and diastolic cardiac function. Interestingly, cardiomyocytes in rats with overexpression of GH showed hypertrophy consistent with the well described stimulatory effects of GH and IGF-1 on cardiac contractile protein synthesis. This study however leaves a number of questions unanswered. Was the efficacy of GH overexpression due to its effects on protein synthesis or on its ability to decrease cardiomyocyte apoptosis by inducing IGF-1 and akt in the infarcted rat hearts? Does the timing of GH expression post-MI play a role in the extent of recovery? The authors used adenovirus which has an expression period of 7–10 days. It would be interesting to check whether longer expression periods may be of further benefit or in fact may worsen the GH induced hypertrophic response. Finally, since the GH effects are probably mediated by IGF-1, PI-3 kinase and akt, will targeting these downstream signaling proteins bring further specificity to the desired therapeutic effect? Joren C. Madsen, MD, DPhil Roger J. Hajjar, MD Massachusetts General Hospital 55 Fruit St Boston, MA 02114-2696 e-mail: [email protected]; [email protected]

0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2004.02.054

CARDIOVASCULAR

Ann Thorac Surg 2004;77:2122–9