INTRODUCTION The effect of GH and IGF-I on the heart has been demonstrated in numerous experimental studies. GH and IGF-I receptors are expressed in cardiac myocytes, and IGF-I causes hypertrophy of cultured rat cardiomyocytes and delays cardiomyocyte apoptosis. In addition, GH and IGF-I have a direct effect on myocardial contractility, increasing the intracellular calcium content and enhancing the calcium sensitivity of myofilaments in cardiomyocytes. Clinical studies in patients with disorders of the GH/IGF-1 axis confirm the significant relationship between GH/IGF-I and the cardiovascular system. Interestingly, both GH excess and deficiency states are associated with abnormal cardiac function and with an attendant increased risk for cardiovascular morbidity and mortality in both. Our contention is that the apparent paradox of the relationship may be due to changes in the energy state of the myocardium in GH excess (acromegaly) and deficiency (GHD in hypopituitarism) and the inability of cellular metabolism to change appropriately between the competing demands of oxidative stress and anabolic processes. Data from the giant GH-overexpressing transgenic mouse demonstrate reduced creatinine phosphate-to-ATP ratio supporting an effect of GH on cardiac energy status. .

AMP-activated protein kinase (AMPK) is the energy sensor of the cell and has been shown to be an important regulator of cell metabolism, including cardiac cells (Dyck & Lopaschuk, 2002). AMPK is activated by rising AMP/ATP ratio, and programs intracellular metabolism to conserve energy for oxidate metabolism and to suspend anabolic processes. A substantial body of evidence testifies to the importance of AMPK and the associated regulation of myocardial energy status to cardiac function.

• Reduced activity of AMPK is a feature of inherited cardiomyopathies.
• Low energy status measured non-invasively is a predictor of death in dilated cardiomyopathy.
• Activation of AMPK reduces the injury in experimental models of ischaemia
• Reduced cardiac energy reserve is a feature of type 2 diabetes mellitus (T2DM), and cardiovascular death accounts for over 80% of mortality in T2DM.
• Drugs which are known to activate AMPK improve mortality in type 2 diabetes, such as metformin and glitazones (Kahn et al., 2005).
• Cannabinoids and ghrelin, first identified by our group to increase cardiac AMPK levels, have been shown to improve ischaemia-reperfusion injury (Frascarelli et al., 2003; Underdown et al., 2005; Shibata et al., 2005).

Taking these experimental observations together, it can be inferred that in diabetic cardiac muscle, there is impaired activation of AMPK despite low energy levels, and potentially a failure of the normal mechanism to favour oxidative metabolism and cytoprotective functions during ischaemia.

In patients, cardiac AMPK cannot be directly studied, although in a related animal study we will assess the effect of GH excess and deficiency on the activation of AMPK cardiac intracellular energy status and myocardium function. However, using 31P Magnetic Resonance Spectroscopy (MRS), in vivo measurement of high energy phosphate molecules can be assessed non-invasively and these have been shown in other diseases to have important clinical consequences. Specifically, both phosphocreatine (PCr), and ATP can be determined and ADP levels derived from the phosphocreatine:ATP ratio (Scheurmann-Freestone 2003). This approach has been recently used to demonstrate low ambient level of myocardial ADP in patients with T2DM. In the current study we will investigate the energy status in patients with acromegaly and GHD before and after normalisation of their GH status. We hypothesise that in untreated acromegaly, the unrestrained drive of anabolism will be associated with a low energy status, and particularly in inability to respond to exercise. We propose that in GHD, reduced anabolism will be associated with relatively high levels of energy but an impairment of muscle mass. We further anticipate that the normalisation of hormone levels resulting from either medical or surgical therapy will result in improvement in energy-storing phosphate molecule ratios in the myocardium.

Acromegaly - Cardiac function Acromegalic cardiomyopathy is the specific myocardial disease of acromegaly (Clayton, 2003; Sacca et al., 2003; Colao et al., 2004b). Active acromegaly is associated with a 2- to 3-fold increase in mortality, mainly from cardiovascular disease, although with effective treatment the excess mortality can be reversed (Sheppard, 2005). In its early stages, acromegalic cardiomyopathy is characterised by increased left ventricular mass, with eccentric remodeling and normal diastolic function with a high cardiac output state with reduction of systemic vascular resistance (Fazio et al., 2000). In the next stage, increased heart mass and diastolic dysfunction, attributed to direct injury of the myocardium with GH hypersecretion, occur in the absence of associated diseases like hypertension, diabetes mellitus, and thyroid dysfunction. These abnormalities can progress to dilated cardiomyopathy and heart failure if acromegaly remains untreated for many years. There are specific structural changes in the myocardium with increased myocyte size and interstitial fibrosis of both ventricles. Left ventricular hypertrophy is common (in 64% of newly diagnosed cases) even in young patients with short duration of disease (Colao et al., 2003). Functionally, the main consequence of these changes is impaired left ventricular diastolic function, particularly when exercising, such that exercise tolerance is reduced (Colao et al., 2002). Myocardial perfusion is impaired in patients with active acromegaly as assessed by single photon emission computed tomography (SPECT), thus representing an early stage of cardiac involvement in acromegaly that may be directly mediated by GH excess (Herrmann et al., 2003). Some of these structural and functional changes can be reversed by effective treatment to lower serum GH levels to less than 1-2 ng/ml (glucose suppressed or random, respectively) and normalize IGF-I and long-term outcome and survival is improved (Colao et al., 2002; Jaffrain-Rea et al., 2003). Diastolic function improves with treatment, but the effect on exercise tolerance is more variable, and more longitudinal data are required to assess the benefits.

Acromegaly - Coronary atherosclerosis Although coronary disease has a prevalence of 3-40% in acromegalic patients, there is considerable controversy whether this is directly related to GH excess. According to a well-controlled study (Otsuki et al., 2001) in which intima thickness was measured, the extent of atherosclerosis in the acromegalic patients was not higher than that in non-acromegalic subjects, and considering their increased number of atherosclerotic risk factors, such as lipid abnormalities, diabetes, increased homocystein, lipoprotein a, fibrinogen and platelet activator inhibitor 1 levels which are all strong predictors of coronary vascular disease, GH might therefore be even protective (Matta & Caron, 2003). In a recently described study with 79 acromegalic patients with 22 age-matched controls again no difference was observed in intima thickness (Paisley et al., 2006). Increased concentration of IGF-I – via its vasodilatotary effect mediated by nitric oxide - might be involved in the lack of susceptibility to atherosclerosis in some acromegalic patients. There are no previously published studies using electron beam coronary CT in this group of patients.

Acromegaly - Skeletal muscle It is recognised that skeletal muscle in acromegaly fatigues rapidly but this has not been adequately explained, particularly in relation to the increase in muscle mass. It has been suggested that this may result from associated metabolic derangements (diabetes or thyroid abnormalities), or a direct effect of GH excess on muscle. Increase in muscle specific creatinine kinase levels have been detected in patients with active disease which improves with a reduction of serum GH (McNab & Khandwala, 2005) . Microscopic examination of skeletal muscle from acromegalic patients shows type 1 fibre hypertrophy (Nagulesparen et al., 1976).

GHD - Cardiac function Patients with GH deficiency have an increased mortality due to cardiovascular disease. Most studies show decreases in heart function such as reduced left ventricular mass and cardiac output, and reduced left ventricular ejection fraction particularly during exercise; in addition there is an increased incidence of coronary artery disease (Colao et al., 2004a; Colao et al., 2004b; Svensson et al., 2005). The mechanism of cardiac abnormalities in GHD is debated, but most data support the hypothesis that GHD leads to reduced cardiac mass, reduced contractility, reduced pre-load and increased after-load, all of which could lead to reduced stroke-volume and reduced cardiac output (Colao et al., 2004b). Treatment with GH results in improvement in cardiac function and reduction of peripheral resistance due to direct anabolic actions of the myocardium but increased dimensions could also be secondary to an increased cardiac output and stroke volume as a result of increased contractility and increased pre-load. In conclusion, GHD is associated with cardiac dysfunction related to the degree and duration of GHD and GH replacement seems to enhance ventricular mass and cardiac function and to reverse diastolic abnormalities (Juul, 1996).

GHD - Coronary sclerosis The evidence linking coronary sclerosis and GHD is robust although it is unclear whether this is a direct effect of GH deficiency on cardiac endothelium, or an indirect effect of the increase in cardiovascular risk factors documented in patients with hypopituitarism and GH deficiency. These include increased abdominal adiposity, abnormal LDL and triglyceride levels, high fibrinogen and plasminogen activator inhibitor activity, increased markers of inflammation (such as interleukin-6 and C-reactive protein) and consequently increased intima-thickness (Colao et al., 2004a; Colao et al., 2004b). In our studies, we will use electron-beam CT (EBCT) as a non-invasive means of determining coronary artery disease, which is a sensitive indicator of IHD. This will allow us to control for changes in the onset and/or recovery from ischaemia that might be due to macrovascular ischemia rather than reduced intracellular capacitance. There are no previously published studies using electron beam coronary CT in these groups of patients.

GHD - Skeletal muscle Adult growth hormone deficiency (AGHD) is associated with fatigue, tiredness and isometric muscle strength is reduced to 76%. After 6 months of GH therapy, muscle mass increases 5% (Bengtsson et al., 1993; Juul, 1996). Neuromuscular dysfunction is also observed (abnormal electromyogram and interference pattern analysis) which improves after initiating GH therapy (Webb et al., 2003). The reduced exercise capacity observed in GHD is multifactorial but reduced skeletal muscle mass has an important contribution to it (Juul, 1996).

In summary, existing evidence demonstrates that patients with both GH deficiency and GH excess have abnormal cardiac and skeletal muscle function, although the nature of the cardiac abnormality differs. In acromegaly there is hypertrophy and impaired contractility with dysfunction amplified during exercise. In GHD, risk factors for IHD cause premature occlusive coronary artery disease, coupled with a poor anabolic stimulus and reduced muscle mass. Skeletal muscle echoes these processes. . In transgenic mice overexpressing GH (an acromegalic mouse) the phosphocreatine -to-ATP ratio measured by MRS is significantly lower in GH overexpressing mice than controls, suggesting impaired energy level in the myocardium (Bollano et al., 2000). We hypothesise that chronic growth hormone excess may lead to reduced energy level within the myocardium, and that restoration of normal GH level will ameliorate this abnormality. Our model would also predict that in GHD resting energy levels would be normal or even high, and that the primary defect is a reduction in muscle mass due to reduced anabolic stimulus. In patients with GH excess and GHD we will assess cardiac energy levels with MRS, a non-invasive technique of assessing myocardial energy metabolism before and after treatment, a technique that has been successfully used to establish abnormal heart energy levels in patients with diabetes and heart failure (Scheuermann-Freestone et al., 2003).

Objectives

1. Determine cardiac (at rest) and skeletal muscle (at rest, peak exercise and recovery) energy metabolism in patients with GH excess and detect changes after normalisation of GH and IGF-1 levels
2. Determine cardiac and skeletal muscle energy metabolism in patients with GH deficiency and detect changes after normalisation of IGF-1 levels
3. To correlate muscle energy metabolism parameters to GH and IGF-1 status in the control subjects and in both patient groups
4. Determine the prevalence of coronary artery calcifications in patients with GH excess and GH deficiency and correlate with their risk factors and other surrogate markers of CHD (serum levels of high sensitivity CRP, lipoprotein (a), homocysteine, and insulin resistance (HOMA analysis).
5. Determine the anatomical distribution of the atheromatous plaques within the coronary arteries and the extent of isolated plaques compared to diffuse disease.

Methods Cardiac and skeletal investigations

• MRI at 1.5T (Siemens Sonata) will be used to measure right and left ventricular morphology and global function (Sandstede et al., 2005) (i.e. left and right ventricular systolic and end-diastolic volumes and ejection fractions). Tissue phase mapping will be used to measure radial and rotational tissue velocities for regional wall motion abnormalities and abnormal contractile patterns.
• 31P MRS will be used as previously described (Crilley et al., 2000; Roest et al., 2001) to monitor intracellular PCr and ATP in heart muscle. Acquisition will be triggered using electrocardiographic gating. Cardiac diastolic function will be measured using cine MRI volume-time curves and tissue phase mapping parameters.
• Heart failure severity (NYHA status) will be determined from the degree of dyspnoea (6 min walk test).
• Peak O2 uptake (MVO2) will be estimated from a metabolic gas exchange analysis performed during maximal treadmill cardiopulmonary exercise testing.
• Skeletal muscle MR imaging and spectroscopy: The protocol for the 31P MRS determination of calf muscle (gastrocnemius and soleus) metabolism at rest and during dynamic exercise is well established (Scheuermann-Freestone et al., 2003). A T1-weighted MR image will be used to determine the muscle cross-sectional area and 31P MRS of the gastrocnemius muscle will be performed at rest, during and after exercise. Each subject will lie in the 3T superconducting magnet with their calf overlying a 6 cm diameter surface coil. The muscle will be exercised by plantar flexion of the right ankle, lifting a weight of 10% lean body mass a distance of 7 cm at a rate of 30/min. After acquisition of four spectra (each 1.25 min), the weight will be incrementally increased by 2% of lean body mass every other minute. The subject will exercise until stopping when fatigued and the muscle will then be studied during recovery. On retesting after therapy, exercise will continue again until the subject is stopped by fatigue (treatment may lengthen this time). Relative concentrations of Pi, PCr and ATP will be measured from their signal intensities in the spectra as described previously (Adamopoulos et al., 1999; Butterworth et al., 2000; Scheuermann-Freestone et al., 2003).
• Electron beam coronary CT (EBCT) to assess coronary disease. As calcium is deposited in the earliest stages of atheroma formation, EBCT can detect sub-clinical coronary disease many years before it results in ischaemia. The extent of coronary artery calcification reflects the overall impact of risk factors, both known and unknown, on the end organ, the arterial wall, and as such gives a superior predictive value over the Framingham risk assessment (Thompson & Partridge, 2004; Pletcher et al., 2004). EBCT scanning will be performed using a GE-IMATRON 300 scanner using a conventional protocol. Calcium deposition within coronary arteries will be assessed by total volume and by the Agatson correction. The number of plaques and calcium score will be measured in the left main artery, left anterior descending, left circumflex, and right coronary artery, in addition to an overall score.

Other investigations will include: Demographic and clinical data (incl. medication) will be obtained for all patients and control subjects; ECG, Epworth Sleepiness Scale questionnaire and 5 point test for sleep apnoea, blood samples will be taken after a 12 h fast, after 30 min rest, for glucose, lactate, free fatty acids, ketone bodies, insulin, triglyceride, cholesterol, HDL, LDL, ANP, BNP, noradrenalin, electrolytes, creatinine, creatinine kinase, uric acid, TNF , leptin, Hb and glycosylated Hb determinations. Apoprotein A/B, High sensitivity C-reactive protein, Lipoprotein (a), Homocysteine, Insulin (HOMA analysis), biochemistry and liver function, pregnancy test. Endocrine investigations will include Glucose tolerance tests (for acromegaly patient; for GHD if necessary), GH day-curve (a mean of 5 estimations throughout the day), serum IGF-I, pituitary and peripheral hormones (LH, FSH, cortisol, TFT, testo/E2, PRL), estimation of duration of the disease (the presumed duration of acromegaly will be estimated by comparison of patients’ photographs taken over a 1–3-decade span and by interviews to date the onset of acral enlargement and other clinical symptoms (Damjanovic et al., 2002; Colao et al., 2003)) Patient selection: Patients will be recruited at St. Bartholomew’s Hospital (Dr P. Jenkins and Prof. A. Grossman), King’s Hospital (Dr S. Aylwin) and St Thomas’s Hospital (Dr P. Carroll) in London, Royal Free Hospital (Prof P. Boloux), the John Radcliffe Hospital Oxford (Prof J. Wass), Sheffield (Dr J. Newell-Price), and Stroke-on-Trent (Prof R. Clayton) from the Endocrine Wards and outpatient clinics. This constitutes a large recruitment base. We estimate that 45 new acromegaly patients and 60-80 new GHD patients per year will be screened. Patients will be selected on the basis of the clinical diagnosis of acromegaly [Acromegaly: typical signs and symptoms and biochemical evidence (GH > 1μg/l during OGTT, IGF-I above age-related reference range); GHD: a structural hypothalamo-pituitary defect, at least one other pituitary hormone deficiency, AGHDA score >11 (NICE criterion) and evidence of severe biochemical GHD (usually a peak GH response to an insulin or glucagon stress test of <3 μg/l)]. Patients will be managed according to the clinical protocols of the referring centre.

Inclusion criteria for acromegaly

• Clinical and biochemical diagnosis of acromegaly. Thyroid and glucocorticoid replacement if necessary stable for at least 4 weeks before the study. Gonadotrophin status will be recorded and whenever possible patients will be studied in the same status
• Males and females aged 18-70 years willing to give informed consent
• At least 6 months after the onset of symptoms of acromegaly and on stable medication for heart failure treatment (if any) for at least 4 weeks prior to inclusion into the study
• Systolic blood pressure < 180 mmHg, diastolic blood pressure < 110 mmHg.

Exclusion criteria:

• Change in medication in the preceding 4 weeks
• Patients on subcutaneous insulin therapy
• Hyperthyroidism
• Not being in sinus rhythm
• Unstable angina pectoris and decompensated heart failure (define as NYHA 3-4)
• Clinically significant valvular disease, clinically significant chronic obstructive pulmonary disease
• History of myocardial infarction or stroke within the last 6 months, major cardiac surgery within the last 6 months
• Significant history of drug- or alcohol abuse or unable to give informed consent
• Any other significant surgical or medical condition which would considerably affect results in view of the identifying clinician
• Typical contraindication for MR (e.g. metal implants in delicate positions, aneurysm clips, shrapnel injuries, pacemakers, internal defibrillators and severe claustrophobia)
• Pregnancy

Inclusion criteria for GHD

• Clinical and biochemical diagnosis of GHD. All hormones replaced (if clinically necessary) except GH. Thyroid and glucocorticoid replacement if necessary stable for at least 4 weeks before the study. Gonadotrophin status will be recorded and whenever possible patients will be studied in the same status
• Males and females aged 18-70 years willing to give informed consent
• At least 6 months after the onset of symptoms of acromegaly and on stable medication for heart failure treatment (if any) for at least 4 weeks prior to inclusion into the study
• Systolic blood pressure < 180 mmHg, diastolic blood pressure < 110 mmHg.

Exclusion criteria:

• Change in medication in the preceding 4 weeks
• Previous history of acromegaly
• Child-hood onset GHD
• Patients on subcutaneous insulin therapy metformin probably an exclusion
• Hyperthyroidism
• Not being in sinus rhythm
• Unstable angina pectoris and decompensated heart failure (define as NYHA 3-4)
• Clinically significant valvular disease, clinically significant chronic obstructive pulmonary disease
• History of myocardial infarction or stroke within the last 6 months, major cardiac surgery within the last 6 months?
• Significant history of drug- or alcohol abuse or unable to give informed consent
• Any other significant surgical or medical condition which would considerably affect results in view of the identifying clinician
• Typical contraindication for MR (e.g. metal implants in delicate positions, aneurysm clips, shrapnel injuries, pacemakers, internal defibrillators and severe claustrophobia)
• Pregnancy

POST-TREATMENT ASSESSMENT Three months is estimated to be sufficient time to reach altered (improved) cardiac metabolism as assessed by MRS based on previous studies (K. Clarke manuscript submitted)). Therefore the first post-treatment assessment will be 3 months after biochemical cure.

In acromegalic patients 3 months after documented biochemical remission (IGF-1 levels reached upper limit of age-related normal range and day-curve for GH mean < 2.5 μg/l or GH < 1 μg/l during OGTT) repeat MRI, MRS and biochemical investigations. If patient is not in biochemical remission by 24 months, then repeat investigation at 24 months after the start of therapy will be performed as the final assessment.

In GHD patients 3 months after recorded completed dose titration (IGF-I values in the upper third of the age-related normal range) repeat MRI, MRS and biochemical investigations. If patient has not completed GH dose titration by 24 months, then repeat investigation at 24 months after the start of GH therapy anyway.

24 months after start of therapy repeat EBCT. There are data to suggest that for example changing lipid status can improve EBCT after 12 months of therapy (Achenbach et al., 2002) Age and sex matched control subjects will be recruited and studied according to the pre-treatment protocol. Males and females aged 18 to 75 years willing to give informed consent will be selected without endocrine or coronary disease, with further exclusion criteria as above.

Power calculation:

The primary endpoint will be a change in cardiac energetics, as assessed by 31P magnetic resonance spectroscopy (MRS) (expressed as PCr/ATP ratio). Power calculations using data from our cardiac 31P MRS study (Scheuermann-Freestone et al., 2003) showed a highly significant reduction in cardiac PCr/ATP from 2.49  0.31 to 1.55  0.37, a difference of 0.94. In order to find a biologically significant difference between groups of one standard deviation (SD) with 90% power at P< 0.05, we would need to include 21 subjects in each group. For a paired study, in which the effects of therapy were tested on the same patient, we would need to include 11 patients in each group to find a significant difference between groups of 1 SD with 90% power at P< 0.05, while for unpaired comparison with control subjects 15 patients are necessary. To allow for patient drop-outs we will include a minimum of 30 subjects in the studies. Secondary endpoints for all studies will include cardiac function (MRI) and skeletal muscle function and energetics (MRI and MRS, respectively).