University of Veterinary Medicine Hannover

Multimodality Molecular Imaging of Inflammation and Fibrosis in Pressure Overload-Induced Heart Failure and After Mechanical Ventricular Unloading

INAUGURAL - DISSERTATION in fulfillment of the requirements of the degree of Doctor of Veterinary Medicine -Doctor medicinae veterinariae- (Dr. med. vet.)

submitted by Aylina Glasenapp Kaltenkirchen

Hannover 2020

Scientific supervision: Prof. Dr. med. vet. Marion Bankstahl Central Animal Facility and Institute for Laboratory Animal Science, Hannover Medical School

Prof. Dr. med. Frank Michael Bengel Department of Nuclear Medicine, Hannover Medical School

PD Dr. med. Katja Derlin Department of Radiology, Hannover Medical School

James Thomas Thackeray, PhD Department of Nuclear Medicine, Hannover Medical School

1st supervisors: Prof. Dr. med. vet. Marion Bankstahl Prof. Dr. med. Frank Michael Bengel PD Dr. med. Katja Derlin

2nd supervisor: Prof. Dr. med. vet. Michael Fehr Clinic for Small Mammals, Reptiles and Birds, University of Veterinary Medicine Hannover

Day of the oral examination: 11.05.2020

Table of Contents

1.1 Previously Published Parts of This Thesis ...... V

1.2 List of Abbreviations ...... VII

1.3 List of Figures ...... IX

1.4 List of Tables ...... X

2 Introduction ...... - 1 -

3 State of the Art Review ...... - 5 -

3.1 Heart Failure ...... - 5 -

3.1.1 Epidemiology & Public Health Statistics ...... - 5 -

3.1.2 Clinical Characterization & Classification ...... - 5 -

3.1.3 Pathology & Mechanisms ...... - 8 -

3.1.4 Clinical Management ...... - 9 -

3.2 Animal Models of Heart Failure and Unloading ...... - 10 -

3.2.1 Pharmacological Induction of Heart Failure ...... - 10 -

3.2.2 Surgical Induction of Heart Failure - Transverse Aortic Constriction ...... - 11 -

3.2.3 Surgical Induction of Heart Failure – Myocardial Infarction ...... - 13 -

3.2.4 Model of Reversible Heart Failure - Reverse TAC Model ...... - 13 -

3.2.5 Summary ...... - 14 -

3.3 Left Ventricular Remodeling ...... - 16 -

3.3.1 Definition ...... - 16 -

3.3.2 Reverse Remodeling ...... - 17 -

3.3.3 Difference of TAC and MI-Induced Remodeling ...... - 18 -

3.4 Inflammation in Heart Failure ...... - 19 -

3.4.1 Inflammation & Pressure Overload ...... - 19 -

I 3.4.2 Inflammation for Prediction of Functional Outcome ...... - 22 -

3.5 Fibrosis in Heart Failure ...... - 23 -

3.5.1 Fibrosis & Pressure Overload ...... - 24 -

3.5.2 Fibrosis as a Therapeutic Target ...... - 26 -

3.6 Imaging ...... - 27 -

3.6.1 Magnetic Resonance Imaging ...... - 27 -

3.6.1.1 Principle ...... - 27 -

3.6.1.2 Clinical Relevance & Applications in Heart Failure ...... - 28 -

3.6.1.3 T1 Mapping ...... - 28 -

3.6.2 Positron Emission Tomography ...... - 30 -

3.6.3 Principle ...... - 30 -

3.6.3.1 Clinical Relevance & Applications in Heart Failure ...... - 32 -

3.6.3.2 18F-FDG & Myocardial Glucose Metabolism ...... - 32 -

3.6.3.3 68Ga-Pentixafor & Chemokine Receptor CXCR4 ...... - 34 -

3.6.4 Imaging for Targeted Therapy ...... - 35 -

4 Hypotheses & Objectives ...... - 36 -

4.1 Hypotheses...... - 36 -

4.2 Objectives ...... - 36 -

5 Experimental Design ...... - 37 -

6 Material & Methods ...... - 40 -

6.1 Animals ...... - 40 -

6.2 Surgical Models ...... - 42 -

6.2.1 Transverse Aortic Constriction ...... - 42 -

6.2.2 Reverse Transverse Aortic Constriction ...... - 45 -

II 6.2.3 Myocardial Infarction ...... - 45 -

6.3 Cardiac Magnetic Resonance Imaging ...... - 47 -

6.3.1 Scan Protocol & Image Acquisition ...... - 47 -

6.3.2 Scan Protocol ...... - 48 -

6.3.3 Functional Cardiac Magnetic Resonance Imaging ...... - 50 -

6.3.4 T1 Mapping Acquisition ...... - 50 -

6.3.5 Image Analysis ...... - 51 -

6.4 Positron Emission Tomography ...... - 53 -

6.4.1 Radiosynthesis ...... - 53 -

6.4.2 Animal Preparation ...... - 53 -

6.4.3 18F-FDG Scan Acquisition, Reconstruction & Analysis ...... - 54 -

6.4.4 68Ga-Pentixafor Scan Acquisition, Reconstruction & Analysis ...... - 55 -

6.5 Ex Vivo Tissue Workup ...... - 58 -

6.5.1 Ex Vivo Biodistribution ...... - 58 -

6.5.2 Ex Vivo Autoradiography ...... - 58 -

6.5.3 Western Immunoblotting ...... - 59 -

6.5.4 Histology ...... - 59 -

6.5.5 Immunohistochemistry ...... - 60 -

6.6 Analysis & Statistics ...... - 61 -

7 Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure ...... - 62 -

8 Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling ...... - 94 -

9 Discussion ...... - 133 -

9.1 Key Findings...... - 133 -

III 9.2 Heart Failure Animal Models ...... - 134 -

9.3 Pathology of Inflammation ...... - 137 -

9.4 Imaging of Inflammation ...... - 141 -

9.5 Pathology of Fibrosis ...... - 144 -

9.6 Imaging of Fibrosis ...... - 146 -

9.7 Imaging for Targeted Therapy and Treatment Guidance ...... - 149 -

9.8 Limitations ...... - 151 -

9.9 Clinical Perspective ...... - 151 -

9.10 Future Perspectives ...... - 153 -

9.11 Conclusions ...... - 156 -

10 Summary ...... - 157 -

11 Zusammenfassung ...... - 159 -

12 References ...... - 163 -

13 Acknowledgments...... - 196 -

IV 1.1 Previously Published Parts of This Thesis

Journal publication: Glasenapp A, Derlin K, Marcel Gutberlet, Wang Y, Bankstahl M, Wollert KC, Bengel FM, Thackeray JT “Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure.” Journal of Nuclear Medicine, 2019 Oct 25. pii: jnumed.119.232488. doi: 10.2967/jnumed.119.232488.

Oral presentations: “Multimodality molecular imaging of inflammation and cardiac dysfunction in acute stages following pressure overload-induced heart failure” Aylina Glasenapp, Katja Hueper, Marcel Gutberlet, Yong Wang, Kai C. Wollert, Frank M. Bengel, James T. Thackeray, 2nd GyMIC Molecular Imaging Symposium MoBi 2018, 15.-16.11.2018, Göttingen, Germany

“Multimodality molecular imaging of inflammation, fibrosis and cardiac dysfunction in acute and chronic stages following pressure overload-induced heart failure” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Yong Wang, Kai C. Wollert, Frank M. Bengel, James T. Thackeray, 57. Jahrestagung der Deutschen Gesellschaft für Nuklearmedizin, 03.-06.04.2019, Bremen, Germany, Abstracts der 57. DGN-Jahrestagung 2019, Abstract ID: WVP108

“Pressure overload evokes cardiac chemokine receptor CXCR4 upregulation, which predicts subsequent progression of heart failure” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Saskia Kropf, Hans-Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, ICNC 2019, Nuclear Cardiology & Cardiac CT, 12.-14.05.2019, Lisbon, Portugal, European Heart Journal - Cardiovascular Imaging, Volume 20, Issue Supplement_3, June 2019, jez150.005, https://doi.org/10.1093/ehjci/jez150.005

“Multimodal imaging identifies cardiac chemokine receptor CXCR4 upregulation in response to pressure overload as a predictor of subsequent heart failure progression” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Laura B. N. Langer, Hans-Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, SNMMI 2019 Annual Meeting, 22.- 25.06.2019 Anaheim, USA, J Nucl Med May 1, 2019 vol. 60 no. supplement 1 34

V “Multimodal imaging identifies cardiac chemokine receptor CXCR4 upregulation in response to pressure overload as a predictor of subsequent heart failure progression” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Laura B. N. Langer, Hans-Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, 3rd GyMIC Molecular Imaging Symposium MoBi 2019, 26.-27.09.2019, Münster, Germany

“Multimodal imaging identifies inflammation and fibrosis in response to pressure overload- induced heart failure and detects alleviation of cardiac remodeling following ventricular unloading” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Laura B. N. Langer, Hans- Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, TOPIM 2020, Hot Topics in Molecular Imaging, 12.-17.01.2020, Les Houches, France

Poster presentations: “Multimodality molecular imaging of inflammation, fibrosis and cardiac dysfunction in acute and chronic stages following pressure overload-induced heart failure” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Hans-Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, 2nd Small Animal MRI Symposium S.A.M.S. 2019, 13.-14.06.2019, Hannover, Germany

“Multimodal imaging identifies inflammation and fibrosis in response to pressure overload- induced heart failure and detects alleviation of cardiac remodeling following ventricular unloading” Aylina Glasenapp, Katja Derlin, Marcel Gutberlet, Laura B. N. Langer, Hans- Jürgen Wester, Tobias L. Ross, Frank M. Bengel, James T. Thackeray, TOPIM 2020, Hot Topics in Molecular Imaging, 12.-17.01.2020, Les Houches, France

Funding This study was supported by the German Research Foundation (DFG, Clinical Research Group KFO311, Excellence Cluster REBIRTH-2, and research grant TH2161/1-1).

VI 1.2 List of Abbreviations

%ID/g Percentage injected dose per gram µm Micrometer 18F-FDG 2-18F-fluoro-deoxy-D-glucose ACE Angiotensin-converting enzyme ANG II Angiotensin II CVD Cardiovascular diseases CMR Cardiac magnetic resonance CT Computed tomography ECM Extracellular matrix EDV End diastolic volume EF Ejection fraction ESV End systolic volume FACS Fluorescence-activated cell counting FAP Fibroblast activation protein FLASH Fast low-angle shot G Gauge GLUT1 Glucose transporter-1 GLUT4 Glucose transporter-4 HFpEF Heart failure with preserved ejection fraction HFrEF Heart failure with reduced ejection fraction i.v. Intravenously i.p. Intraperitoneally IL Interleukin IP-10 Interferon-γ-inducible protein-10 K/X Ketamine/xylazine LAD Left anterior descending coronary artery LGE Late gadolinium enhancement LV Left ventricle LVAD Left ventricular assist device

VII MBq Megabecquerel MI Myocardial infarction MCP-1 Monocyte chemoattractant protein-1 MIP-2 Macrophage inflammatory protein-2 MRI Magnetic resonance imaging NMR Nuclear magnetic resonance NYHA New York Heart Association OSEM3D/MAP Ordered-subset expectation-maximization 3-dimensional/maximum-a-posteriori PBS Phosphate-buffered saline PET Positron emission tomography RAAS Renin-angiotensin-aldosterone system ROI Region of interest rTAC Reverse transverse aortic constriction s.c. Subcutaneously SA Short axis Sx Surgery TAC Transverse aortic constriction TGF-β Transforming growth factor-β TNF-α Tumor necrosis factor-α

VIII 1.3 List of Figures

Figure 1. Anatomical and functional classification of heart failure...... - 6 -

Figure 2. Cardiac phenotype and myocardial structure pre and post-TAC...... - 12 -

Figure 3. Left ventricular patterns of remodeling ...... - 16 -

Figure 4. Schematic illustration of fibrogenesis ...... - 25 -

Figure 5. Principle of magnetic resonance imaging ...... - 28 -

Figure 6. Principle of positron emission tomography ...... - 31 -

Figure 7. Glucose and 18F-FDG ...... - 33 -

Figure 8. Chemical structure of 68Ga-pentixafor ...... - 34 -

Figure 9. Study 1. Imaging and histological validations in the acute phase post-TAC...... - 37 -

Figure 10. Study 2. Serial imaging and ex vivo measurements of inflammation and fibrosis

post-TAC...... - 38 -

Figure 11. Study 3. Imaging of the effect of reverse TAC...... - 39 -

Figure 12. Schematic illustration of TAC surgery ...... - 44 -

Figure 13. Schematic illustration of rTAC surgery ...... - 45 -

Figure 14. Schematic illustration of LAD ligation ...... - 46 -

Figure 15. Arrangement of short axis image slice stack ...... - 48 -

Figure 16. CMR imaging protocol ...... - 49 -

Figure 17. Representative cardiac functional analysis ...... - 52 -

Figure 18. Scan protocol 18FDG ...... - 55 -

Figure 19. Scan protocol 68Ga-pentixafor / 18F-FDG ...... - 57 -

Figure 20. 68Ga-Pentixafor image analysis...... - 57 -

IX 1.4 List of Tables

Table 1. NYHA classification of heart failure ...... - 7 -

Table 2. Overview of cardiac geometry, structure and function in HF mouse models ...... - 15 -

Table 3. Cardiac positron emission tomography radionuclides ...... - 31 -

Table 4. Animal groups study 1 ...... - 41 -

Table 5. Animal groups study 2 & 3 ...... - 42 -

Table 6. CMR imaging protocol ...... - 48 -

X Introduction

2 Introduction

Heart failure (HF) is the common endpoint of numerous cardiovascular diseases (CVD). It is a leading cause of death worldwide with an increasing prevalence in the population (MOSTERD u. HOES 2007). HF is a clinical syndrome and occurs with a variety of symptoms and severity stages. Typically, HF results from adverse cardiac remodeling, which is characterized by geometric and structural changes, often ventricular dilatation and cardiomyocyte hypertrophy. Mostly, disease progression is paralleled by impaired physical condition and a reduced quality of life. Patients suffer from dyspnea, pulmonary edema, fatigue and peripheral edema due to water retention (PONIKOWSKI et al. 2016). Therapy is frequently implemented only after clinical symptoms are already established. Drug treatment covers symptoms of HF to improve life quality but often does not treat the underlying pathological mechanisms. Mechanical circulatory support, such as left ventricular assist devices (LVADs), improves cardiac function and may even reverse remodeling (BIRKS et al. 2006). However, more research is needed to detect the early pathophysiological mechanisms of HF development, which are responsible for remodeling, HF progression, and may be targeted to support repair. Inflammation and fibrosis are thought to contribute critically to the development and progression of HF (SEGURA et al. 2014; DICK u. EPELMAN 2016), and the time course and evolution of these early processes could be targeted for preventative or corrective treatment to improve functional outcome. Inflammation in HF can be categorized as acute or chronic. Acute myocardial inflammation is initiated by cardiomyocyte cell death during ischemic injury of the myocardium, mostly following myocardial infarction (MI) (FRANGOGIANNIS et al. 2002). Chronic non-ischemic myocardial inflammation is provoked by multiple stimuli, such as pressure or volume overload, cardiomyopathy, toxicity, infections, and metabolic stress (BRIASOULIS et al. 2016). Generally, excessive inflammatory leukocyte and macrophage infiltration is suspected to aggravate HF (SHIRAZI et al. 2017). Studies suggest that early treatment targeting macrophages in pressure overload- induced HF could prevent adverse cardiac remodeling (KAIN et al. 2016; PATEL et al. 2018). On one hand, excessive inflammatory response can be deleterious, but on the other, physiological response is also needed for optimal tissue repair.

- 1 - Introduction

Additionally, fibrosis is thought to influence inflammation and cardiac disease development, but there are limited therapies to directly target adverse fibrosis (FAN u. GUAN 2016). Myocardial fibrosis can be classified as reparative or reactive fibrosis (TALMAN u. RUSKOAHO 2016). Reparative fibrosis sequentially follows acute inflammation after ischemic injury resulting in scar formation, whereas reactive fibrosis coexists with chronic myocardial inflammation and is characterized by diffuse interstitial and perivascular collagen deposits (FRANGOGIANNIS 2008). Furthermore, cardiac remodeling including inflammatory and fibrotic processes might be modified or even reversed by pressure unloading in response to therapeutic (LANDMESSER et al. 2009) or interventional treatment (WOHLSCHLAEGER et al. 2005). The crucial interplay between inflammation and fibrosis is thought to influence disease progression but is poorly characterized (TALMAN u. RUSKOAHO 2016). Consequently, the need arises to assess the temporal progression of inflammatory and fibrotic pathophysiological mechanisms in HF development. There are several limitations in longitudinal characterization of these processes, as most methods, including cell quantification by flow cytometry or histology, are terminal measurements or require invasive biopsies for tissue samples. Thus, post-mortem tissue analysis provides unique information for one time point but cannot reflect disease progression. By contrast, non-invasive multimodal imaging provides the opportunity to monitor inflammatory and fibrotic mechanisms in vivo. Moreover, imaging can be used to evaluate therapeutic response with novel interventions against these pathogenic processes (HESS et al. 2019). Using positron emission tomography (PET), molecular physiology can be visualized serially during disease progression. Clinically, cardiac PET is mostly conducted to assess myocardial perfusion (MARI u. STRAUSS 2002; SARIKAYA 2015), but also allows imaging of inflammation in CVD using the glucose analogue PET tracer 2-18F-fluoro-deoxy- D-glucose (18F-FDG), i.e. after MI (W. W. LEE et al. 2012). However, 18F-FDG has its limitations for detecting myocardial inflammation, as it is non-specific and accumulates in numerous metabolically active cells. This highlights the need for more specific inflammatory PET imaging tracers. Here, 68Ga-pentixafor might bridge the gap by targeting the chemokine receptor CXCR4, which is expressed on a broad range of leukocytes (BORCHERT et al.

- 2 - Introduction

2019), and allows for detection of active inflammation in the myocardium after ischemic injury (THACKERAY et al. 2015b; REITER et al. 2018). Additionally, cardiac magnetic resonance (CMR) imaging is used to evaluate cardiac function and tissue composition (STROHM et al. 2001; PATERSON et al. 2013). CMR imaging is non-invasive, radiation-free and provides reproducible images with high spatial resolution. Ventricular volumes during cardiac remodeling are used to calculate the most common cardiac functional parameter, left ventricular (LV) ejection fraction (EF), which is a measure of blood output into the systemic circulation. Moreover, CMR enables assessment of myocardial tissue composition. Late gadolinium enhancement (LGE) using contrast agent allows for scar detection in ischemic myocardium (KIM et al. 1999), has shown promise to characterize fibrosis in extracellular fraction (BOHL et al. 2008), and to predict late functional outcome in non-ischemic CVD (KURUVILLA et al. 2014). However, in vivo quantification of diffuse interstitial fibrosis in non-ischemic CVD remains challenging (L. M. ILES et al. 2015). CMR T1 mapping detects myocardial tissue alterations including fibrosis in mice after pressure overload HF (STUCKEY et al. 2014) and in humans (EVERETT et al. 2016; GONZALEZ et al. 2018), and may allow quantitative analysis. In this project, we aimed to characterize the time course of inflammatory and fibrotic responses to pressure overload HF induced by transverse aortic constriction (TAC) and following mechanical unloading via reverse transverse aortic constriction (rTAC) in a longitudinal multimodal imaging study. The combination of PET imaging and CMR T1 mapping visualizes molecular-level alterations in tissue composition throughout the progression of HF and in response to therapy. Specifically, we aimed to characterize the inflammatory response using 18F-FDG and 68Ga-pentixafor PET imaging and to investigate the contribution of early inflammation to late functional outcome. In addition, we aimed to assess cardiac remodeling and function using CMR and T1 mapping to identify fibrotic tissue alterations. We validated these image-based measurements by histopathology, immunoblotting, and ex vivo autoradiography. These studies provide a foundation for future work to guide a specific targeted anti- inflammatory or anti-fibrotic therapy against the early pathogenesis of remodeling prior to overt HF, at a critical juncture when late function can still be improved. The imaging assays

- 3 - Introduction

may ultimately stratify patient risk and direct individual therapeutic decisions towards improved clinical outcome.

- 4 - State of the Art Review

3 State of the Art Review

3.1 Heart Failure

3.1.1 Epidemiology & Public Health Statistics

According to the World Health Organization, CVD are the number one cause of death worldwide, with the majority of affected people live in low- and middle-income countries. Approximately 17.9 million people died from CVD in 2016, which represents 31% of all global deaths. Most CVD could be prevented by avoiding behavioral risk factors such as smoking, alcohol and drug abuse, and unhealthy diet, which is connected to obesity and physical inactivity (YANCY et al. 2013). HF is the common endpoint of many CVD. Although knowledge about HF and treatment strategies continuously improves, the prevalence and incidence of HF in the population still progressively increases with age, especially as average life expectancy rises (MOSTERD u. HOES 2007). Moreover, acute intervention during cardiovascular events improves survival, which means more patients advance to HF. Total health care costs for HF are expected to rise substantially within the next decade (ZIAEIAN u. FONAROW 2016; MAZUREK u. JESSUP 2017).

3.1.2 Clinical Characterization & Classification

HF is a multifactorial clinical syndrome with a variety of symptoms including dyspnea, fatigue and fluid retention, covering a broad range of severity. Medically, HF is a chronic, progressive condition in which the heart muscle is unable to pump sufficient blood to meet the body’s need for oxygen and nutrients (PONIKOWSKI et al. 2016). Clinically, HF is divided into left-sided, right-sided, and congestive or biventricular HF. Left- sided HF is further categorized into HF with reduced ejection fraction (HFrEF), formerly termed systolic failure, and HF with preserved ejection fraction (HFpEF), formerly known as diastolic failure (PONIKOWSKI et al. 2016) (Figure 1).

- 5 - State of the Art Review

Figure 1. Anatomical and functional classification of heart failure. Left-sided heart failure is further categorized into heart failure with reduced ejection fraction (HFrEF) and heart failure with preserved ejection fraction (HFpEF).

The functional classification system of the New York Heart Association (NYHA) categorizes patients according to their symptom severity (Table 1). However, symptom severity is not necessarily related to cardiac function, as patients with modest symptoms may have a high risk of hospitalization and mortality (DUNLAY et al. 2009). Established classes of recommendation (I-IV) in combination with levels of evidence (A-D) provide guidance for appropriate therapy selection (PONIKOWSKI et al. 2016). Treatment can attenuate or diminish symptoms in HF patients.

- 6 - State of the Art Review

Table 1. NYHA classification of heart failure

Class Patient Symptoms

No limitation of physical activity. Ordinary physical activity does not cause undue I fatigue, palpitation, dyspnea (shortness of breath).

Slight limitation of physical activity. Comfortable at rest. Ordinary physical activity II results in fatigue, palpitation, dyspnea (shortness of breath).

Marked limitation of physical activity. Comfortable at rest. Less than ordinary III activity causes fatigue, palpitation, or dyspnea.

Unable to carry on any physical activity without discomfort. Symptoms of heart IV failure at rest. If any physical activity is undertaken, discomfort increases.

Level Objective Assessment

No objective evidence of cardiovascular disease. No symptoms and no limitation in A ordinary physical activity.

Objective evidence of minimal cardiovascular disease. Mild symptoms and slight B limitation during ordinary activity. Comfortable at rest.

Objective evidence of moderately severe cardiovascular disease. Marked limitation C in activity due to symptoms, even during less-than-ordinary activity. Comfortable only at rest.

Objective evidence of severe cardiovascular disease. Severe limitations. Experiences D symptoms even while at rest.

Adapted from DOLGIN et al. 1994.

- 7 - State of the Art Review

3.1.3 Pathology & Mechanisms

Cardiac output, defined as the blood volume pumped out of the ventricle per minute, may be reduced in HF (KEMP u. CONTE 2012). Early stages of HF are characterized by cardiac remodeling due to one or more pathologies such as myocardial ischemic injury, cardiomyopathy, hypertension, renal dysfunction or diabetes (KEMP u. CONTE 2012). Cardiac remodeling involves adverse changes of ventricular geometry including LV dilatation, wall thinning or thickening, and myocardial stiffness, as well as compositional changes of myocardial tissue, such as cardiomyocyte hypertrophy and stretching (VERMA et al. 2008). Depending on the triggering injury, it can affect the left, right or both ventricles. The pathologic mechanisms leading to remodeling and HF are diverse. Ischemic injury and subsequent myocardial necrosis initiate local tissue inflammation, replacement fibrosis and extracellular matrix (ECM) reorganization. Compensatory remodeling, coupled with wall thinning and scar formation, leads to contractile dysfunction (SCHIRONE et al. 2017). In non-ischemic HF, chronic cardiac stress, pressure- or volume-overload, or hereditary disease provide the initial trigger (THAM et al. 2015). These abnormalities instigate hemodynamic changes, which can exacerbate ventricular remodeling. Activation of the neurohumoral and sympathetic nervous system is a response which aims at increasing contractility and heart rate to compensate for geometric changes, but this further worsens HF, as chronic overstimulation of the adrenergic system and renin-angiotensin-aldosterone system (RAAS) propagates ventricular remodeling. Moreover, rising aldosterone due to RAAS activation elevates blood volume and pressure, which can stimulate fibrosis. ECM, as the essential scaffolding of healthy myocardium, undergoes reorganization in response to HF due to activation of matrix metalloproteinases. Moreover, dysregulated calcium cycling is crucially involved in lowering cardiac contractility, thereby aggravating HF. Kidney dysfunction often manifests in parallel with HF, and renal sodium and water retention contribute to dyspnea and edema. In summary, all types of HF are characterized by cardiomyocyte cell death and declining cardiac output, which triggers compensatory measures such as neurohumoral upregulation, adrenergic activation, inflammation, and oxidative stress, to limit damage and spare function (BRAUNWALD 2013).

- 8 - State of the Art Review

3.1.4 Clinical Management

Clinical management of HF focuses on minimizing symptoms. First, lifestyle modifications are suggested to prevent or accompany further treatment. This includes physical factors and behaviors like body weight reduction, altered diet, smoking cessation, alcohol abstinence and exercise (KEMP u. CONTE 2012). Generally, developing or established HFrEF is pharmacologically treated with the main goal to improve quality of life and prognosis. Basic treatment strategies include diuretic drugs to reduce excess fluid and lower blood volume and pressure, resulting in reduced cardiac strain. Additionally, neuroendocrine blockade is commonly employed including angiotensin- converting enzyme (ACE) inhibitors and β-adrenergic receptor blockers (ELGENDY et al. 2019). ACE inhibitors and angiotensin receptor blockers block the action of angiotensin II (ANG II), which reduces blood pressure and afterload. Beta-Adrenergic receptor blockers depress sympathetic hyperactivity to reduce heart rate and improve contractility. These blockbuster drug therapies can delay and lessen adverse remodeling in progressive systolic dysfunction, but cannot completely prevent or arrest disease progression (FRIGERIO u. ROUBINA 2005). Mineralocorticoid receptor agonists are sometimes added to further stabilize and improve blood pressure (ROSSIGNOL et al. 2019). In terminal HFrEF with severely impaired function, including drug-refractory patients, invasive intervention with mechanical circulatory support such as LVADs may be used for bridging to cardiac transplantation or as final destination therapy. The inflow cannula of the LVAD is inserted into the apex of the LV and the outflow cannula is anastomosed to the ascending aorta and provides a continuous blood flow. This treatment leads to a significant improvement of functional outcome and overall survival (MILLER et al. 2019). Heart transplantation is the last therapeutic option for chronic end-stage HF patients (PONIKOWSKI et al. 2016) and offers the best survival for the recipients in short- and long- term (MILLER et al. 2019).

- 9 - State of the Art Review

3.2 Animal Models of Heart Failure and Unloading

Several animal models are known and used in experimental induction of HF and must be chosen carefully, based on hypothesis, applicability and clinical relevance. Small rodent models, particularly mice, are widely-used in cardiovascular research as they breed well, and the biological mechanisms of the models are translational to clinical conditions (RECCHIA u. LIONETTI 2007). The pathophysiology depends on the mouse strain, injury trigger, and measurement technique. The available animal models of HF, their phenotypic and cellular presentation and functional outcome are summarized in Table 2.

3.2.1 Pharmacological Induction of Heart Failure

HF in mice can be induced by a variety of pharmacological interventions. In 1975, chronic ethanol administration was described to induce HF symptoms in mice (BERK et al. 1975). The injection of the cardiotoxic agent doxorubicin also induces HF (DELGADO et al. 2004), modeling anthracycline cardiotoxicity. Today, chronic HF is widely induced by continuous administration of sympathomimetic drugs, such as isoproterenol (CHANG et al. 2018) or phenylephrine (AGHAJANIAN et al. 2019), via osmotic minipump. These models mimic stress-induced cardiomyopathy. The osmotic minipump is implanted subcutaneously (s.c.) and releases drug to stimulate the sympathetic nervous system. Isoproterenol is a synthetic catecholamine which stimulates β- adrenergic receptors, with high affinity for cardiac β1-adrenergic receptors, eliciting positive chronotropic, inotropic and dromotropic effects on the heart, and vasoconstriction of the vasculature. Phenylephrine targets α1-adrenoceptors, resulting in vasoconstriction and elevated blood pressure. This chronic sympathetic stress evokes progressive HF. Chronic administration of ANG II also mimics hypertension due to neurohumoral activation reminiscent of clinical HF (BACMEISTER et al. 2019). ANG II regulates blood pressure and water volume in the RAAS. In the heart, ANG II acts mostly via the ANG II type 1 receptor. During ANG II administration, mice show LV remodeling with and without hypertension, suggesting that remodeling pathways may be independent of pressure overload (BACMEISTER et al. 2019). Neurohumoral activation by continuous ANG II infusion causes sterile myocardial inflammation, reactive interstitial fibrosis (BACMEISTER et al. 2019) and cardiomyocyte

- 10 - State of the Art Review

hypertrophy (BOOZ u. BAKER 1996). Cardiac remodeling and functional outcome are strongly dependent on the administered dose. Mice with low dose treatment (0.14- 0.2mg/kg/day) develop concentric hypertrophy accompanied by diastolic dysfunction reminiscent of human HFpEF (REGAN et al. 2015; SHINOHARA et al. 2015) whereas high dose treatment (1.8mg/kg/day) results in dilative, congestive HF accompanied by reduced ejection fraction reminiscent of human HFrEF (WESTERMANN et al. 2012). A combination of ANG II and isoproterenol or phenylephrine is also often used to better simulate the multifactorial clinical condition. Pharmacologic HF models have the advantages of non-invasive or minimally invasive procedures and a simple and reproducible technique. But they are limited by a high mortality and non-cardiac effects to other organs due to systemic administration. Further, these models are less translatable and comparable to development and progression of HF in humans (BRECKENRIDGE 2010), since single signaling pathways are typically targeted, which is not the case in the clinical condition.

3.2.2 Surgical Induction of Heart Failure - Transverse Aortic Constriction

TAC is a well-established model to induce non-ischemic HF by cardiac pressure overload and was first described by ROCKMAN et al. in 1991. TAC is thought to mimic increased left ventricular afterload in patients, due to aortic stenosis or systemic hypertension. In the literature, various surgical procedures are described to perform TAC. The surgery is performed under anesthesia, either using manual ventilation (TARNAVSKI et al. 2004) or with spontaneous breathing (TAVAKOLI et al. 2017). The aortic arch can be approached by various means. Often in the conventional description, a complete sternotomy is carried out with opening of the thoracic cavity (YOO et al. 2018). Less invasive methods feature the opening of the second intercostal space and retraction of the thymus (TARNAVSKI et al. 2004). Alternatively, non-invasive access via a midline incision in the anterior neck has been described (ZAW et al. 2017). Regardless of access, the transverse aorta is constricted between the brachiocephalic trunk and left carotid artery, using a ligature (TARNAVSKI et al. 2004), clips (X. ZHANG et al. 2013) or other tools (MELLEBY et al. 2018). The desired diameter of constriction can be manually adjusted, depending on the used method. Generally, a ligature with a 6-0 suture and blunted needle with a fixed diameter

- 11 - State of the Art Review

which varies typically between 23 and 27-gauge (G) (0.573 to 0.361mm) is used. Functional outcome varies strongly with the degree of constriction (RICHARDS et al. 2019). TAC induces massive systolic wall stress on the myocardium (YOU et al. 2018) and leads to wall thickening, LV dilatation and impaired contractile function (RICHARDS et al. 2019) (Figure 2). TAC causes cardiomyocyte hypertrophy and stretching, sterile inflammation and reactive interstitial fibrosis due to cardiac pressure overload (BACMEISTER et al. 2019). This hypertrophy is characterized by elevated heart weight relative to body weight or tibia length, which is typically measured ex vivo. The TAC model can be adjusted by altering the location of constriction. Constriction of the ascending aorta more strongly elevates the LV pressure, whereas the stenosis of the descending aorta generates milder HF. The TAC model using Rockman`s technique elicits post-operative mortality rates between 35% (MERINO et al. 2018) and 45% (MOHAMMED et al. 2012). But mortality is strongly dependent on numerous factors, such as the diameter of constriction, anesthetic and surgical techniques, experience of the surgeon, and mouse strain, among others. Despite the high mortality, TAC has many advantages and is regarded as the closest animal model to human pressure overload HF. As such, pathobiology is highly translatable.

Figure 2. Cardiac phenotype and myocardial structure pre and post-TAC. Transverse aortic constriction (TAC) induces myocardial hypertrophy (top). Masson trichrome staining of left ventricular myocardial wall shows cardiomyocytes (red) and diffuse interstitial fibrosis (blue) (bottom).

- 12 - State of the Art Review

3.2.3 Surgical Induction of Heart Failure – Myocardial Infarction

MI is a widely-used animal model to induce HF by ischemic injury, first described in small rodents in 1979 (PFEFFER et al. 1979). Infarction is induced by ligation of the left anterior descending coronary artery (LAD), causing ischemic injury to the LV anterolateral myocardium. The LAD ligation can be performed either permanently or for a transient period. Transient LAD occlusion mimics patients with clinical MI with successful revascularization (LINDSEY et al. 2018). The duration of ischemia directly relates to severity of HF, as mice after transient occlusion exhibit less cardiac dysfunction and smaller infarct sizes (MICHAEL et al. 1995). By contrast, permanent LAD occlusion generates large infarcts and more severe remodeling. For both methods, mice are anesthetized by either inhalation or injection anesthesia, then intubated and mechanically ventilated. The fourth intercostal space is opened and pericardium is gently separated. The LAD is ligated with a 5-0 suture. The rib cage and incision are then closed. The severity of ischemia can be adjusted by time of occlusion (MICHAEL et al. 1995), from mild injury with occlusion <30min to severe injury with permanent ligation. Ischemic injury leads to LV dilatation, LV wall thinning and cardiac dysfunction (BACMEISTER et al. 2019). At the cellular level, ischemia causes cell death, followed by acute inflammation and subsequent reparative fibrosis and scar formation. Non-ischemic remote myocardium can also be affected by interstitial fibrosis. These models facilitate investigation of pathophysiological mechanisms of myocardial ischemia and are highly translatable, as they mimic clinical MI. The model has a low post- operative mortality, but mice are at risk of LV rupture within the acute phase after surgery (GAO et al. 2005b).

3.2.4 Model of Reversible Heart Failure - Reverse TAC Model

Reverse TAC is a model to achieve a rapid mechanical unloading of the left ventricle. To this end, the aortic banding is surgically removed after established HF using a procedure identical to TAC surgery. The clinical corollary is the implantation of a mechanical support system e.g. LVAD. The removal of the aortic constriction allows study of the impact of hemodynamic unloading on cardiac function and reverse remodeling (WEINHEIMER et al. 2018). Previous studies indicated regression of cardiomyocyte hypertrophy and interstitial fibrosis after

- 13 - State of the Art Review

removal of the aortic constriction, supporting functional recovery (MERINO et al. 2018; WEINHEIMER et al. 2018). However, application of this model is limited, as the surgery is technically demanding and requires a second thoracotomy to remove the aortic banding. Additionally, debanding causes abrupt hemodynamic unloading, modeling LVAD implantation, but translation to biological mechanisms during gradual decrease of blood pressure in response to pharmacological treatment is complicated.

3.2.5 Summary

For our studies, we chose the murine model of TAC, as it best reflects the biological mechanism of HF development and provokes cardiac phenotype development comparable to HF patients. Therefore, the pathologic mechanisms are highly translatable. TAC induces inflammation and fibrosis, resulting in adverse remodeling and HF (Table 2). The standardized procedure is highly reproducible, and the model induces modest mortality. Although rTAC is challenging due to a second surgery, it optimally mimics implantation of mechanical circulatory support in patients. Ventricular unloading provides novel insights in the mechanisms of reverse remodeling.

- 14 - State of the Art Review

Table 2. Overview of cardiac geometry, structure and function in HF mouse models

Cardiac Functional Model Mechanism Cellular Alterations Phenotype Outcome

- Cardiomyocyte HFrEF, Cardiac hypertrophy / HFpEF LV pressure TAC hypertrophy, stretching (depending overload on LV dilatation - Inflammation constriction - Reactive fibrosis diameter)

- Cardiomyocyte Cardiac HFrEF, Isoproterenol Sympathomimetic hypertrophy hypertrophy, HFpEF infusion stimulation - Inflammation (depending LV dilatation - Reactive fibrosis on dose)

- Cardiomyocyte Cardiac HFrEF, ANG II Neurohumoral hypertrophy hypertrophy, HFpEF infusion activation - Inflammation (depending LV dilatation - Reactive fibrosis on dose)

Permanent LV dilatation, - Cardiomyocyte HFrEF LAD ligation wall thinning cell death Myocardial - Inflammation ischemia HFrEF Transient - Replacement LV dilatation (depending LAD ligation fibrosis on time of occlusion)

- 15 - State of the Art Review

3.3 Left Ventricular Remodeling

3.3.1 Definition

Cardiac remodeling can either be physiological, as in normal growth and in response to exercise training, or pathological as in HF (MANN et al. 2012). Macroscopically, it is characterized by enlargement and geometric alterations of the ventricle. Microscopically, myocardial structural changes include cardiomyocyte hypertrophy and lengthening, increased ECM, and collagen deposition. Remodeling contributes to functional deterioration and may involve left, right, or both ventricles. There are three different patterns of cardiac remodeling in patients after MI based on echocardiographic measurements (Figure 3): (I) Concentric remodeling, characterized by normal LV mass index and increased relative wall thickness; (II) eccentric hypertrophy featuring increased LV mass index and normal relative wall thickness; and (III) concentric hypertrophy, described by increased LV mass index and increased relative wall thickness (VERMA et al. 2008). In non-ischemic HF, wall stress due to pressure overload is a main contributor to concentric hypertrophy (OPIE et al. 2006).

Figure 3. Left ventricular patterns of remodeling Concentric remodeling of the left ventricle exhibits increased wall thickness without hypertrophy. Eccentric hypertrophy of the left ventricle occurs without changes in wall thickness. Concentric hypertrophy of the left ventricle is accompanied by increased left ventricular wall thickness. Adapted from VERMA et al. 2008.

- 16 - State of the Art Review

In general, cardiac remodeling is initiated by pathophysiological stimuli, such as myocardial injury, pressure or volume overload, neurohumoral activation or inflammation. By consequence, necrosis or apoptosis leads to loss of cardiomyocytes. Oxidative stress and altered energy metabolism result in cardiomyocyte hypertrophy. Dying cardiomyocytes recruit circulating monocytes, which propagate a local inflammatory response. Resident fibroblasts undergo activation and transdifferentiation, culminating in myocardial fibrosis (SCHIRONE et al. 2017). In mouse models of pressure overload, LV remodeling proceeds in two successive stages. Initially, mice exhibit moderate systolic and diastolic dysfunction and concentric hypertrophy, which is followed by severe systolic and diastolic dysfunction accompanied by eccentric hypertrophy, as assessed by echocardiography (BACMEISTER et al. 2019). Fluorescence-activated cell counting (FACS) revealed that the murine heart consists of 56% myocytes, 27% fibroblasts, 7% endothelial cells and 10% vascular smooth muscle cells (BANERJEE et al. 2007). In HF, myocardial cell composition shifts initially towards higher inflammatory cell content, and fibroblast proliferation and differentiation (SHINDE u. FRANGOGIANNIS 2014; PRABHU u. FRANGOGIANNIS 2016). The main morphologic cell alterations occur in cardiomyocytes and ECM (SEGURA et al. 2014). Dying cardiomyocytes undergo apoptosis, while surviving cardiomyocytes enlarge, and diffuse interstitial collagen accumulates (KONSTAM et al. 2011). Myocyte lengthening occurs before HF symptoms appear and myocyte cross-sectional area is increased, typically reflecting concentric hypertrophy (GERDES 2002). These morphological alterations counteract the loss of cardiomyocytes to apoptosis and compensate for failing contractile function.

3.3.2 Reverse Remodeling

Reverse remodeling is a decrease of LV volume following e.g. hemodynamic unloading. Reverse remodeling has been described in a variety of cardiac diseases, such as viral myocarditis, peripartum cardiomyopathy or cardiotoxicity (MANN et al. 2012). As such, patients with chronic HF who undergo medical, device, or surgical therapies bear potential for reverse remodeling. In some cases, pharmacological therapy with ACE inhibitors, ANG II type 1 receptor blocker and β-blocker therapy has been associated with beneficial effects on

- 17 - State of the Art Review

LV remodeling in mice (ROCKMAN et al. 1994; P. MULLER et al. 2009) and humans (LANDMESSER et al. 2009). Implantation of LVADs may also be effective in inducing reverse ventricular remodeling and (SEGURA et al. 2014) attenuating cardiac fibrosis (KASS et al. 1995). Interestingly, the effect of normalization of LV geometry in HF patients persisted even when therapy was ceased, suggesting that biological mechanisms respond and adapt to strain (MANN et al. 2012). Nevertheless, the underlying molecular mechanisms that are responsible for a decrease in LV size remain poorly understood. In addition, novel treatment approaches might influence remodeling. For instance, statin therapy improves survival in CVD (HORWICH et al. 2004) and reduces progression of HF (KJEKSHUS et al. 1997), suggesting that statin treatment in early HF may attenuate further remodeling. In mice, statin treatment inhibits LV hypertrophy as assessed by echocardiography after pressure overload (KAMEDA et al. 2012). More experimentally, gene therapy may e.g. be used to directly reprogram murine and human fibroblasts to functional cardiomyocytes in vitro (Y. CHEN et al. 2017), and was demonstrated to generate new cardiac muscle and improve heart function after MI in mice in vivo (MURAOKA u. IEDA 2014).

3.3.3 Difference of TAC and MI-Induced Remodeling

TAC induces LV concentric hypertrophy and cardiomyocyte hypertrophy (TOISCHER et al. 2010), while MI generates scar accompanied by LV wall thinning, progressive LV dilatation and cardiomyocyte hypertrophy in remote myocardium (HOUSER et al. 2012). Functionally, both models induce progressive systolic dysfunction, with a mild diastolic functional abnormality (BACMEISTER et al. 2019). Inflammation and fibrosis during remodeling in response to non-ischemic HF after TAC and ischemic HF after MI diverge considerably (BACMEISTER et al. 2019). First, the inflammatory response to MI is thought to progress faster than in non-ischemic HF after TAC. Second, inflammation after MI is more robust compared to TAC. Third, TAC is characterized by sustained fibrosis (B. CHEN u. FRANGOGIANNIS 2018). These differences are relevant, as divergent pathological mechanisms, including the inflammatory and fibrotic severity, contribute to cardiac remodeling, resulting in distinct

- 18 - State of the Art Review

disease manifestations. Reverse remodeling has been described in non-ischemic HF rather than after ischemic injury. This means that treatments of HF targeting individual pathologic stimuli, might attenuate or even reverse cardiac remodeling.

3.4 Inflammation in Heart Failure

3.4.1 Inflammation & Pressure Overload

Inflammatory leukocytes and macrophages are crucial contributors to interstitial fibrosis, contractile dysfunction (PATEL et al. 2018) and non-ischemic HF progression (HOFMANN u. FRANTZ 2013). HF patients exhibit elevated circulating white blood cells (BARRON et al. 2001) as well as higher levels of inflammatory cytokines and biomarkers (AMMIRATI et al. 2012), regardless of pathogenesis. As such, inflammation has emerged as a potential therapeutic target not only in ischemic HF, but also non-ischemic cardiomyopathies (BOZKURT et al. 2001), though the efficacy of such approaches has not been fully validated.

Following MI, the abrupt death of cardiomyocytes initializes a focal inflammatory response in the damaged region. The inflammatory response after MI begins with an early peak of neutrophils and granulocytes which are gradually replaced by infiltration of pro-inflammatory monocytes and differentiated macrophages peaking on day three to five after MI. A subsequent switch towards reparative macrophages occurs around day five to seven, which supports angiogenesis and repair. A consistent presence of infiltrating B and T lymphocytes modulates the local inflammatory cascade (FRANGOGIANNIS et al. 2002). But based on different causes of inflammation in ischemic compared to non-ischemic HF, the mechanisms during disease development and progression are not necessarily comparable. After cardiac pressure overload, cardiomyocyte necrosis is modest and several factors may play a role in inducing inflammation, including neurohumoral activation (LEUSCHNER et al. 2010), increased oxidative stress (SRIRAMULA u. FRANCIS 2015) and mechanical strain (LINDNER et al. 2014). Recent evidence suggests that inflammatory cytokines and leukocytes, particularly macrophages and T lymphocytes, play a major role in HF progression and adverse cardiac remodeling (HOFMANN u. FRANTZ 2013; AYOUB et al. 2017), and may promote interstitial fibrosis (HAIDER et al. 2019) and contractile dysfunction (PATEL et al. 2018). Pressure overload initiates an inflammatory response with upregulation of

- 19 - State of the Art Review

chemokines and cytokines which recruit leukocytes to the damaged myocardium. This inflammation is associated with a pathophysiological shift to fibroblast activation, differentiation and collagen deposition (HOFMANN u. FRANTZ 2013; HAIDER et al. 2019). High levels of circulating pro-inflammatory cytokines and chemokines are related to cardiac functional decline in HF patients (TORRE-AMIONE et al. 1996). Initial studies in mice with pressure overload-induced HF reported early pro-inflammatory cytokine and chemokine upregulation, including tumor necrosis factor-α (TNF-α) and interleukin (IL) -1β peaking on day three and macrophage inflammatory protein-2 (MIP-2), interferon-γ inducible protein-10 (IP-10) and monocyte chemoattractant protein-1 (MCP-1) peaking on day seven as measured by mRNA expression levels (XIA et al. 2009). A peak of leukocytes and preferential recruitment of pro-inflammatory macrophage subsets was demonstrated one week after TAC without statistical significance of neutrophil accumulation as quantified by flow cytometry (XIA et al. 2009; WEISHEIT et al. 2014; PATEL et al. 2017). Moreover, it is known that macrophages play a crucial role in adverse cardiac remodeling and fibrogenesis (KONG et al. 2014; CHENG et al. 2017), wherein they infiltrate coronary arteries and promote perivascular fibrosis (WEBER et al. 2013). Accordingly, circulating monocyte- derived cardiac chemokine receptor CCR2+ macrophages, which infiltrate the heart one week after pressure overload in mice, are required as main contributing cells for adverse LV remodeling in pressure overload-induced HF (PATEL et al. 2018), supporting the notion that CCR2 could be a viable immunomodulatory therapeutic target to ameliorate cardiac remodeling and improve functional outcome. As genetic deletion of the inflammatory cytokine IL-6 attenuates LV hypertrophy and dysfunction induced by TAC, it also plays a critical role in the pathogenesis of LV hypertrophy in response to pressure overload (ZHAO et al. 2016). The consensus of these studies reveals a major contribution of increased macrophage density in the pathogenesis of hypertrophy and cardiac dysfunction in pressure overload-induced HF. Nevertheless, the recruitment of peripheral and activation of cardiac resident macrophages following LV hemodynamic overload and their contribution in the development and progression of HF are not fully understood (B. CHEN u. FRANGOGIANNIS 2018). The acquired immune system including T lymphocytes also plays a role in progressive HF (SMITH u. ALLEN 1991; KAYA et al. 2008). Genetically modified lymphocyte-deficient

- 20 - State of the Art Review

mice do not develop LV dilatation in response to pressure overload, displaying preserved systolic function and reduced myocardial fibrosis compared to wild-type mice. T lymphocyte transfer to lymphocyte-deficient mice enhanced contractile dysfunction and promoted adverse ventricular remodeling (LAROUMANIE et al. 2014). CD4+ T cells in particular promote the transition from cardiac hypertrophy to HF and aggravate myocardial remodeling in pressure overload (LAROUMANIE et al. 2014). CD8+ T cells do not accelerate the progression of HF (GROSCHEL et al. 2018). Treatments to augment regulatory T cells lowered the inflammatory response in the left ventricle and inhibited HF progression in mice (H. WANG et al. 2016). As T cells infiltrate coronary arteries, they are also described to promote perivascular fibrosis (WEBER et al. 2013). Currently, the early pathogenic mechanisms leading to non-ischemic HF remain undertreated. Conventional therapeutic approaches combine multiple drugs targeting late symptoms of HF to support cardiac function and improve survival (PONIKOWSKI et al. 2016). Initial clinical studies with broadly targeted anti-inflammatory therapies including non-steroidal anti- inflammatory drugs generated inconsistent outcome regarding cardiac function, HF symptoms, and mortality (AYOUB et al. 2017). Treatment with beta-blockers, ACE inhibitors, angiotensin receptor antagonists and mineralocorticoid receptor blockers reduce clinical symptoms, but do not limit mortality and morbidity substantially (HEYMANS et al. 2009). Specific treatment of pro-inflammatory cytokines in mice yielded promising results but was not successful in patients (CHUNG et al. 2003; MANN et al. 2004) A balanced inflammatory reaction is necessary for tissue recovery, but excessive inflammation has deleterious effects. Consequently, the optimal treatment time point and target are crucial for successful therapy. In mice, modulation of T cell-mediated immunity attenuated cardiac hypertrophy after pressure overload (KALLIKOURDIS et al. 2017). Depletion of monocytes also ameliorated late adverse cardiac remodeling (PATEL et al. 2018). Consequently, inflammatory and fibrotic mechanisms are promising novel treatment targets. But it remains difficult to assess inflammation at the local tissue level, which typically requires biopsy or indirect blood biomarker measurements. Chemokines are small proteins which play a key role in the inflammatory response, as they are produced by inflammatory tissue and recruit leukocytes to sites of inflammation

- 21 - State of the Art Review

(HUGHES u. NIBBS 2018). The cell surface G protein-coupled chemokine receptor CXCR4 is mainly expressed on leukocyte subsets including T cells, B cells, monocytes and macrophages (MURDOCH 2000). CXCR4 plays a major role in lymphocyte trafficking and recruitment at sites of inflammation, where the ligand CXCL12 (SDF-1α) is a strong chemoattractant for leukocyte and progenitor recruitment (MURDOCH 2000; BORCHERT et al. 2019). In inflammatory cardiovascular diseases such as atherosclerosis and myocardial ischemia, CXCR4 is highly upregulated at the injury site due to influx of leukocytes (DORING et al. 2014) Blockade of CXCR4 alleviates the early inflammation and improves functional outcome (Y. WANG et al. 2019). Here, molecular imaging holds great promise, because it may detect these adverse alterations specifically and noninvasively.

3.4.2 Inflammation for Prediction of Functional Outcome

Severity of early inflammation predicts the degree of HF progression following ischemic injury (RISCHPLER et al. 2016; HESS et al. 2019). Since inflammation is thought to play a similar role in early cardiac remodeling in non-ischemic HF, the severity might have similar prognostic value. However, inflammatory cell infiltration and adverse ventricular alterations in chronic non-ischemic HF are poorly characterized. Although some predictors of death and hospitalization are known, accurate risk stratification in HF patients remains challenging. Blood biomarkers such as natriuretic peptide reflect myocardial wall stress and troponin T reflects cardiac stress. Recently, the inflammation and fibrosis related biomarker soluble suppression of tumorigenesis-2 demonstrated prognostic value for cardiovascular mortality (EMDIN et al. 2018). While systemic inflammation may be quantified through blood biomarkers such as C-reactive protein and leukocyte counting, local inflammatory activity in the myocardium is more difficult to assess. Non-invasive molecular imaging might identify early pathophysiological mechanisms of HF on the myocardial level, and thereby predict late functional outcome.

- 22 - State of the Art Review

3.5 Fibrosis in Heart Failure

Fibrosis is defined as an excessive accumulation of ECM and is commonly found in CVD (DE BOER et al. 2019). Resident cardiac fibroblasts within the ECM serve as support cells in the healthy heart (CAMELLITI et al. 2005) where they are involved in matrix degradation, insulation of the conduction system, cardiomyocyte electrical coupling, vascular maintenance, stress sensing, and fibrillary collagen production (IVEY u. TALLQUIST 2016). In the healthy human heart, fibrillary collagen is a major structural component of ECM (WEBER 1989), and the healthy myocardium consists of approximately 2-4% collagen (SEGURA et al. 2014). Pathological myocardial fibrosis can be caused by various factors and is divided into two distinct morphologic forms: reparative and reactive fibrosis (HARA et al. 2017). Reparative or replacement fibrosis develops primarily as a consequence of myocardial necrosis following ischemic myocardial injury (GANDHI et al. 2011). After MI, the abrupt death of cardiomyocytes instigates a focal inflammatory response in the infarct region, which is followed by collagen accumulation and scar formation at the site of damage. This process preserves myocardial structure but has adverse functional consequences due to myocardial wall thinning, LV dilatation and loss of elasticity. Conversely, reactive fibrosis is a common response to non-ischemic cardiac stress including pressure or volume overload, cardiomyopathy, cardiotoxicity, infection and metabolic stress, and is initiated by multiple pathways (KONG et al. 2014). Generally, cardiac resident fibroblasts are stimulated by several mechanisms, including cardiomyocyte cell death, mechanical stimuli or neurohumoral activation. Fibroblasts become activated, and further proliferate and transdifferentiate into myofibroblasts, the main effector cells in production of fibrillary collagen (WYNN u. RAMALINGAM 2012). However, labeling and identification of fibroblasts in vivo is complicated. The characterization of myocardial fibrosis in patients requires indirect evaluation. Non- invasively, echocardiographic measurements of ventricular filling (MOREO et al. 2009), or magnetic resonance imaging (MRI) with gadolinium-enhancement T1 mapping are used to estimate ventricular fibrosis, but do not precisely indicate tissue composition (DE BOER et al. 2019). Clinically, myocardial fibrosis can be determined by invasive myocardial biopsy (GONZALEZ et al. 2018), which is subject to sampling errors and may not be representative for the entire organ.

- 23 - State of the Art Review

Several animal studies showed that myocardial collagen content is elevated after four to seven days and peaks after three to six weeks after MI (DE BOER et al. 2019). Activated fibroblasts express fibroblast activation protein (FAP), which could be used to identify and quantify active fibrosis (AGHAJANIAN et al. 2019). It was demonstrated by morphometric analysis of rat heart sections that density of FAP-positive fibroblasts is significantly higher in the infarct border zone compared to infarct center or remote myocardium (VARASTEH et al. 2019). Further biomarkers, e.g. the matricellular protein periostin are under investigation for in vivo identification of activated fibroblasts. Periostin is suggested to be involved in transforming growth factor-β (TGF-β) signaling and is associated with pathological fibrosis; thus, it may allow characterization of cardiac fibroblasts and their regulating signals to enhance understanding of pathophysiology (SNIDER et al. 2009).

3.5.1 Fibrosis & Pressure Overload

Like inflammation, reactive fibrosis in HF is both a cause and consequence of remodeling and contractile dysfunction (WEBER et al. 2013). Following pressure overload, mice exhibit diffuse interstitial and perivascular collagen deposits within the LV myocardium (BACMEISTER et al. 2019). Multiple factors contribute to the dynamic process of cardiac fibrogenesis in response to pressure overload. Initial collagen accumulation occurs at perivascular sites (HARA et al. 2017) with later deposition in the interstitium, which directly impacts ventricle stiffness (WEBER u. BRILLA 1991; SEGURA et al. 2014). Mechanical strain by pressure overload stimulates resident cardiac fibroblasts, cardiomyocytes, resident and infiltrating inflammatory cells and endothelial cells in the myocardium, which secrete profibrotic factors. Resident cardiac fibroblasts are stimulated to proliferate and expand. The upregulation of cytokines, such as TGF-β, and RAAS activation stimulate fibroblast proliferation and differentiation to non-proliferative myofibroblasts (HUMERES u. FRANGOGIANNIS 2019). These myofibroblasts secrete procollagen types I and III and other nonstructural extracellular matrix macromolecules (GONZALEZ et al. 2018), which propagate inflammation and enhance local fibrosis (SHINDE u. FRANGOGIANNIS 2014; TRAVERS et al. 2016). The elevated activation of fibroblasts and myofibroblasts increases diffuse interstitial fibrosis around cardiomyocytes (SEGURA et al. 2014) (Figure 4), which impairs diastolic and ultimately systolic function. Recently, it was demonstrated that

- 24 - State of the Art Review

macrophages can contribute to fibroblast activation, leading to myofibroblast transdifferentiation and expansion (HAIDER et al. 2019). Myofibroblasts are described as main effector cells in cardiac fibrosis as they express contractile proteins, such as α-smooth muscle actin (KONG et al. 2014), and known to recruit inflammatory cells by cytokine production (TRAVERS et al. 2016). A recent study showed that human macrophages in culture adopt a fibroblast-like phenotype in vitro. A transgenic mouse model which allowed tracking of macrophages by specific cell markers, demonstrated that (myo)fibroblast-like cells are derived from macrophages in a HF model (HAIDER et al. 2019). This suggests the critical intersection between inflammation and fibrosis as a viable therapeutic target (HAIDER et al. 2019).

Figure 4. Schematic illustration of fibrogenesis Resident cardiac fibroblasts are activated by transforming growth factor-β (TGF-β). Activated fibroblasts express fibroblast activation protein (FAP), secrete cytokines, proliferate and differentiate into myofibroblasts. Activated fibroblasts and myofibroblasts secrete collagen, resulting in diffuse interstitial fibrosis.

- 25 - State of the Art Review

3.5.2 Fibrosis as a Therapeutic Target

The prospect of targeting fibrosis for therapy to enhance recovery has gained interest in recent years. Generally, myocardial fibrosis persists in patients treated according to the general guidelines (HEYMANS et al. 2015). Therapeutic options to directly mitigate excessive cardiac fibrosis are lacking (FAN u. GUAN 2016; FANG et al. 2017), but targeted treatment of mechanisms underlying fibrosis development could prevent adverse remodeling (GONZALEZ et al. 2018). Reactive fibrosis may be a viable therapeutic target, since it is thought to have the potential for partial reversal (KONG et al. 2014). Conventional drug therapy inhibits fibrotic mechanisms in hypertension and HF patients (HARA et al. 2017), but the underlying mechanism of this protection is unclear. Identification of the best treatment target and time point remains challenging. Direct reprogramming of cardiac fibroblasts using cardiac-specific transcription factors to convert fibroblasts to cardiomyocytes is a new strategy in regenerative medicine. Gene transfer of reprogramming factors into mouse myocardium after MI improves heart function and reduces LV remodeling (SONG et al. 2012; MURAOKA u. IEDA 2014). Recently, immunotherapy using engineered T cells demonstrated promise to directly inhibit fibroblast activity, attenuate fibrosis and achieve partial functional rescue in mice with pharmacological induced HF (AGHAJANIAN et al. 2019). These novel treatment approaches illustrate the need for further research on fibrotic mechanisms as potential therapeutic targets in HF and the potential of molecular imaging to non-invasively evaluate such therapy. However, the success of therapy requires a method to quantify fibrosis non-invasively to evaluate the efficacy of the treatment.

- 26 - State of the Art Review

3.6 Imaging

Medical imaging allows for the visualization of structure, anatomy, and physiological processes to gain insights into disease pathology. Magnetic resonance imaging provides accurate anatomic imaging with high resolution and functional and structural tissue assessment. Positron emission tomography allows for imaging of physiological and pathological molecular mechanisms by tracking the distribution of a radioactive ligand or substrate in the body. These molecular imaging techniques may address some limitations of conventional diagnostic methods.

3.6.1 Magnetic Resonance Imaging

3.6.1.1 Principle MRI is based on the principles of nuclear magnetic resonance (NMR), first described in 1938 by Isidor Rabi (RABI et al. 1992) in molecular beams and discovered independently in 1945 by Felix Bloch and Edward Mills Purcell in liquids and solids (BLOCH 1953; ANDREW 1992). Living organisms consist substantially of hydrogen, within organic matter, but differ in distribution among tissues. The nuclei of hydrogen atoms have a positive charge and spin along their own axes, which is called precession. The frequency of precession, or the Larmor frequency, depends on the surrounding magnetic field. Moreover, the spin creates an additional magnetic field, effectively turning each atom into an individual bar magnet. By placing the atoms in an external magnetic field, the atoms align parallel or antiparallel to the field. For imaging, a radiofrequency pulse with the same frequency as the Larmor frequency is introduced to induce the NMR effect, wherein nuclei of some atoms move to a higher energy state due to energy absorption. After the radiofrequency pulse, the nuclei return to their former positions in the magnetic field. This movement is termed relaxation, and the nuclei emit energy to their surroundings at the Larmor frequency (Figure 5). This energy can be detected by a receiver coil which surrounds the imaged subject. Relaxation time is based on the fading signal strength of the emitted energy and depends of the composition of the surrounding tissue of the hydrogen protons. Two types of relaxation times can be distinguished. T1 relaxation time is the time constant for the recovery of the longitudinal magnetization parallel to the external magnetic field. T2 relaxation time is also called

- 27 - State of the Art Review

transversal or spin-spin relaxation. As both relaxation times are dependent on the composition of surrounding tissue, they provide information on changes in tissue make-up (JAMES et al. 1982).

Figure 5. Principle of magnetic resonance imaging Hydrogen protons spin along their own axis (precession) creating a magnetic field. Protons align to external magnetic field. Radiofrequency pulse excites protons by energization and protons flip into transversal plane. After radiofrequency pulse, protons return to alignment with magnetic field, T1 relaxation time is the regrowth of longitudinal magnetization.

3.6.1.2 Clinical Relevance & Applications in Heart Failure CMR imaging provides high resolution images of LV geometry and structure throughout the cardiac cycle, which are used to determine cardiac function. CMR is considered the gold standard for measurement of contractile function. Accordingly, CMR is a reliable and reproducible technique for the diagnosis of heart diseases and it is often used in clinical trials to quantify functional outcome (STROHM et al. 2001).

3.6.1.3 T1 Mapping T1 relaxation time is the time constant for recovery of longitudinal magnetization, also called longitudinal relaxation or spin lattice relaxation. The T1 relaxation time is defined as the time needed to recover 63% of the original longitudinal magnetization. T1 is therefore quantitative, which allows for tissue analysis. Because the relaxation time is not consistent, T1 mapping can identify differences in biologic material and tissue composition based on water content. For example healthy myocardium, mainly comprised of cardiomyocytes with small

- 28 - State of the Art Review

extracellular space has lower T1 relaxation time compared to failing myocardium which has extensive irregular extracellular space, including collagen and infiltrating cells (KOCKOVA et al. 2016). Consequently, altered T1 relaxation time may provide information about the extent of myocardial fibrosis (BURT et al. 2014). T1 relaxation time may be measured with or without gadolinium-based contrast agents. In native (contrast free) acquisition, the T1 relaxation time is prolonged in fibrotic myocardium. Contrast agent slowly accumulates in the ECM of the myocardium, leading to enhancement of the magnetic resonance image after distribution, termed ̔delayed enhancement̕. With higher interstitial fibrotic content, more contrast agent is retained long in the tissue, resulting in shorter T1 relaxation time. In ischemic heart disease, native T1 relaxation time differentiates between normal and infarcted myocardium (EVERETT et al. 2016), where replacement fibrosis indicative of myocardial scar formation accumulates gadolinium on delayed imaging (VOGEL-CLAUSSEN et al. 2006; WINTER et al. 2016). Information on the accuracy of MRI-based detection of diffuse interstitial myocardial fibrosis in hypertrophic cardiomyopathy is limited (BURT et al. 2014), but bears potential to identify subtle disease in the heart (KELLMAN u. HANSEN 2014). T1 mapping detects myocardial tissue alterations such as fibrosis in HF patients (EVERETT et al. 2016; GONZALEZ et al. 2018) and mice after cardiac pressure overload (STUCKEY et al. 2014), providing novel diagnostic approaches. Among patients with HF, prolonged native T1 relaxation times correlated to extent of diffuse cardiac fibrosis as measured by histological examination of biopsies (KOCKOVA et al. 2016). Moreover, shorter T1 times are related to beneficial cardiac remodeling after long term treatment with standard HF drugs (BRADHAM et al. 2017). Conversely, prolonged T1 relaxation times indicate worse prognosis including cardiac mortality and cardiovascular events (PUNTMANN et al. 2016). Native T1 mapping displays prolonged T1 relaxation time in TAC mice compared to sham animals (STUCKEY et al. 2014). As such, T1 mapping provides novel insights into diffuse myocardial interstitial fibrosis in non-ischemic HF and may improve clinical assessment and diagnosis. But the timeframe for T1 relaxation time alterations, its relation to inflammation, and response to therapy remain poorly characterized in progressive HF.

- 29 - State of the Art Review

3.6.2 Positron Emission Tomography

3.6.3 Principle

The development of the first PET imaging camera for clinical use was described in 1975 (TER-POGOSSIAN et al. 1975). Currently, this non-invasive imaging technique is widely used for oncological purposes, as well as for specific applications in cardiology for perfusion and viability assessment. In research, preclinical small animal PET scanners provide high sensitivity and high spatial resolution, up to 1.5mm (VISSER et al. 2009). PET is based on the decay characteristics of positron-emitting radionuclides, mapping the spatial and temporal distribution of radionuclide-labeled compounds, or radiotracers. The radiotracer can be a ligand or substrate for a specific biological pathway. The radionuclides undergo positron decay, whereby a proton within the unstable radionuclide nucleus converts to a neutron and releases a positron and an electron neutrino. The emitted positron is the antiparticle to the electron, which has the same mass but a positive charge. Depending on the initial energy of the positron, it travels a short distance (~0.1-3mm) in the surrounding tissue before colliding with an electron and undergoing annihilation. This annihilation emits two photons with an energy of 511keV each, at 180° to each other. Within the PET scanner, a ring of position- sensitive detectors surrounding the examined object detects photons in coincidence (CHERRY u. GAMBHIR 2001) (Figure 6). Reconstruction algorithms, e.g. iterative ordered subset expectation maximization (OSEM), use these coincident detected events to identify the location of the annihilation and generate a virtual map of the tracer distribution in the body. In cardiac PET imaging, the half-life of the most commonly used radionuclides is relatively short and varies between 122s (15O) and 110min (18F) (SARIKAYA 2015). Some radionuclides like 18F are produced using a cyclotron, and others such as 68Ga are produced by a generator (68 Ge/ 68 Ga generator) (Table 3).

- 30 - State of the Art Review

Figure 6. Principle of positron emission tomography Annihilation of a positron and electron releases two gamma ray photons with the energy of 511kEV, radiating in an angle of 180° (left). Detector ring in the PET scanner around the examined mouse detects coincidence events of photons with an energy of 511kEV (right).

Table 3. Cardiac positron emission tomography radionuclides

Radionuclide Half-Life ~ Production Applications (Tracer) 82 82 Rb 75s Generator Myocardial perfusion ( RbCl) 15 15 O 122s Cyclotron Myocardial perfusion ( O-H2O) 13N 10min Cyclotron Myocardial perfusion (13N-ammonia) Myocardial metabolism (11C-acetate) 11C 20.4min Cyclotron sympathetic neuronal function (11C-hydroxyephedrine)

68Ga 68min Generator Myocardial inflammation (68Ga-pentixafor) Myocardial perfusion (18F-Flurpiridaz), viability (18F-FDG), inflammation (18F-FDG), 18F 110min Cyclotron angiogenesis (18F-Galacto-RGD), calcification (18F-NaF)

- 31 - State of the Art Review

3.6.3.1 Clinical Relevance & Applications in Heart Failure Cardiac PET can assess a variety of pathophysiological mechanisms in HF, depending on the applied tracer. Most commonly, cardiac PET measures myocardial perfusion, metabolism, and innervation. More recently, PET techniques have been applied to evaluate molecular processes such as inflammation, angiogenesis, cell death, and ventricular remodeling (THACKERAY u. BENGEL 2016). Using perfusion tracers, myocardial blood flow and global coronary flow reserve can be determined. Reduced coronary flow reserve is predictive for cardiac death, and patients suffering coronary artery disease could benefit from risk stratification by PET (WU 2016). Myocardial perfusion or viability assessment can also include measurements of LV function (YANCY et al. 2013).

3.6.3.2 18F-FDG & Myocardial Glucose Metabolism The PET radiotracer 18F-FDG is a glucose analogue and widely used in nuclear medicine. The radionuclide 18F has a half-life of 109.8min and is produced by a cyclotron. Metabolically active tissue transports 18F-FDG / glucose into the cells, where it is phosphorylated by the enzyme hexokinase and trapped within the cell, as the resulting phosphorylated deoxyglucose does not proceed through glycolysis. Because hexokinase is the rate-limiting step of glycolysis, uptake is proportional to glucose metabolism (YOSHINAGA u. TAMAKI 2007) (Figure 7). The healthy heart is an omnivore. Under basal conditions, it prefers free fatty acids as its abundant major energy source (TAMAKI et al. 2000) since more than 70% of cardiac adenosine triphosphate is derived from fatty acid oxidation, with the remainder from carbohydrates, lactate, ketone bodies and amino acids (TAEGTMEYER et al. 2016). By contrast, higher glucose utilization reflects activation of the fetal gene program. Altered cardiac metabolism in the failing heart, or metabolic remodeling (NEUBAUER 2007; WENDE et al. 2017), shifts the preferred energy substrate from fatty acids to enhanced glucose utilization. In mice after pressure overload, glucose uptake increases while fatty acid utilization declines. Similarly, glucose is preferentially utilized for biomass synthesis in cardiac hypertrophy, suggesting that supplementation of energy may restore cardiac function (UMBARAWAN et al. 2018).

- 32 - State of the Art Review

18F-FDG has also been applied to assess inflammation, as the tracer robustly accumulates in activated inflammatory leukocytes including macrophages (SATOMI et al. 2013; THACKERAY u. BENGEL 2018). 18F-FDG imaging of myocardial inflammation has demonstrated prognostic value for outcome after MI in patients and animal models (RISCHPLER et al. 2016; THACKERAY u. BENGEL 2018), but is not optimal in chronic non-ischemic HF due to robust cardiomyocyte uptake and heightened glucose metabolism (GLASENAPP et al. 2019).

Figure 7. Glucose and 18F-FDG Chemical structure (top) and cardiomyocyte metabolism (bottom) of glucose and 18F-FDG.

- 33 - State of the Art Review

3.6.3.3 68Ga-Pentixafor & Chemokine Receptor CXCR4 To overcome the limitations of 18F-FDG for inflammation imaging, alternative tracers have been developed. One example is 68Ga-pentixafor (Figure 8), which targets chemokine receptor CXCR4. The radionuclide 68Ga has a half-life of 68.1min and is produced by a 68 Ge/ 68 Ga generator. CXCR4 is expressed by a variety of leukocyte subsets and plays a major role in lymphocyte trafficking and recruitment at sites of inflammation (MURDOCH 2000). In inflammatory cardiovascular diseases such as atherosclerosis and myocardial ischemia, CXCR4 signal is elevated proportionally to leukocyte content in the affected region (DORING et al. 2014). Clinically, 68Ga-pentixafor has an emerging diagnostic potential and is used in human cancer (WESTER et al. 2015; LAPA et al. 2017). In oncologic patients, 68Ga-pentixafor uptake was observed to accumulate in inflammatory atherosclerotic lesions (LI et al. 2018a). Recently, cardiac applicability was demonstrated in a human cell uptake study (BORCHERT et al. 2019) and in mice and patients after acute MI, suggesting CXCR4 could predict subsequent cardiac dysfunction (THACKERAY et al. 2015b; REITER et al. 2018). These results demonstrate that 68Ga-pentixafor is a suitable inflammation marker in ischemic HF, supporting its potential application in non-ischemic pressure overload-induced HF.

Figure 8. Chemical structure of 68Ga-pentixafor Radiotracer 68Ga-pentixafor is a ligand for chemokine receptor CXCR4.

- 34 - State of the Art Review

3.6.4 Imaging for Targeted Therapy

Molecular imaging might provide applications beyond diagnosis and prognosis. The idea of targeted therapy and treatment guidance is to connect therapeutic drugs to molecular imaging agents targeting the identical (patho)physiological mechanism. Here, imaging could identify patients who would respond to a specific therapy and define optimal treatment time points or strategies. This paradigm is well established in oncology, where tracking response to therapy and treatment guidance has gained clinical significance (PANTEL u. MANKOFF 2017). In CVD, molecular targeted therapy and treatment guidance approaches are emerging, and bear promise to identify early pathological mechanisms that may be treated to prevent late functional decline (HESS et al. 2019).

- 35 - Hypotheses & Objectives

4 Hypotheses & Objectives

4.1 Hypotheses

The primary hypothesis of this project is that inflammation and concurrent fibrosis contribute to worse outcome in pressure overload-induced experimental HF.

Our secondary hypotheses are that:

 Myocardial inflammation occurs early in non-ischemic HF and predicts late cardiac dysfunction.  Myocardial fibrosis parallels inflammation and is related to functional impairment.  Left ventricular unloading attenuates inflammation and fibrosis and rescues cardiac function.  Non-invasive imaging using CXCR4-targeted PET for inflammation and / or T1 mapping CMR for fibrosis detects progression and regression of disease.

4.2 Objectives

Our specific aims for the project are:

 To compare the progression of inflammation and fibrosis by noninvasive multimodal imaging and assess the influence of acute pathology and chronic contractile function.  To interrogate the temporal inflammatory response in the acute and chronic phase after TAC and rTAC using 18F-FDG or 68Ga-pentixafor PET, and to validate imaging signal by ex vivo autoradiography, biodistribution, histology and immunohistochemistry.  To evaluate myocardial fibrosis in the acute and chronic phase after TAC and rTAC using T1 mapping CMR and confirm the fibrotic biomarker by histology.  To assess cardiac functional impairment and geometric ventricular alterations in the acute and chronic phase after TAC and rTAC.

- 36 - Experimental Design

5 Experimental Design

This thesis comprises three studies. In the first study, we evaluated the relationship between inflammation and contractile function in the acute phase after TAC using 18F-FDG PET to assess inflammation, and CMR imaging to evaluate cardiac function. TAC or sham surgery was performed. The acute phase within 1 week was investigated using CMR at 2 and 8 days and 18F-FDG PET at 3 and 7 days after surgery. This was complemented by tissue sampling at 3 days for histology, and 8 days for histology and Western immunoblotting (Figure 9).

Figure 9. Study 1. Imaging and histological validations in the acute phase post-TAC. TAC or sham surgery was performed, followed by 18F-FDG PET and CMR within the acute phase post-surgery. Tissue for histology was sampled at 3 and 8 days post-surgery. Western immunoblot was performed at 8 days post-surgery.

In the second study, the acute and chronic phase over a time course of 6 weeks after TAC or sham surgery were assessed by longitudinal imaging. Here, we used 68Ga-pentixafor PET to image myocardial chemokine receptor CXCR4 expression to detect inflammation. In addition, we performed CMR imaging to evaluate cardiac function including T1 mapping to assess fibrosis. CMR imaging was performed at baseline (~1 week prior to surgery) and at 2, 8, 23 and 44 days after surgery. In parallel, 68Ga-pentixafor PET imaging was conducted at 3, 7, 21

- 37 - Experimental Design

and 42 days after surgery. Tissue for histology and immunohistochemistry was taken at 3, 8, 23 and 44 days after surgery. Additionally, autoradiography and biodistribution was performed at 7 and 21 days after surgery. All animals were killed at the end of the experiment and tissues were weighed (body weight, heart, lung, spleen) and collected for correlative analysis (Figure 10).

Figure 10. Study 2. Serial imaging and ex vivo measurements of inflammation and fibrosis post-TAC. TAC or sham surgery was performed, followed by 68Ga-pentixafor PET and CMR within the acute and chronic phase post-surgery. Tissue for histology was sampled at intermediate timepoints and end of experiment. Ex vivo autoradiography and biodistribution was performed 1 and 3 weeks post-surgery.

The third study was an extension of the second, in which mechanical unloading by rTAC surgery was performed after 3 weeks of TAC. 68Ga-pentixafor PET was conducted at 7 and 21 days and CMR including T1 mapping at 8 and 23 days following TAC or sham surgery. Tissue for histology and immunohistochemistry was taken at 8 and 23 days after TAC or sham. The aortic banding was surgically removed 25 days after TAC or sham. Then, 68Ga- pentixafor PET at 7 and 21 days and CMR including T1 mapping was performed at 8 and 23 days after rTAC, sham rTAC or sham surgery. Tissue for histology and immunohistochemistry was taken 8 and 23 days after the second surgery. All animals were killed at the end of the experiment and tissues were weighed (body weight, heart, lung, spleen) and collected for correlative analysis (Figure 11).

- 38 - Experimental Design

Figure 11. Study 3. Imaging of the effect of reverse TAC. TAC or sham surgery was performed, followed by 68Ga-pentixafor PET and CMR within 3 weeks post-surgery. Then, rTAC, sham rTAC or sham surgery was performed and followed by PET and CMR within 3 weeks post-surgery. Tissue for histology was sampled at intermediate timepoints and at the end of experiment.

- 39 - Material & Methods

6 Material & Methods

6.1 Animals

All animal experiments were approved by the state authority (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit, Aktenzeichen 33.12-42502-04-17/2718) and local animal ethics review board. Male C57Bl/6N mice at the age of 9 weeks (n=180; 23.6 ± 1.6g) were purchased from Charles River (Sulzfeld, Germany) and housed in groups in a temperature-controlled facility under a 14h/10h light/dark cycle with food and water freely available. C57BL/6N is an inbred strain which is widely used in research in the field of cardiovascular diseases. Male mice were used because hormonal variations have less influence on experimental outcome than in female mice. The animals are genetically identical within the strain. Without genetic differences, biological variance, which could impact results, are reduced. Inbred mice exhibit homogeneity in their phenotypes, including appearance, behaviour and response to experimental interventions (SIMON et al. 2013). The required animal numbers for experimental groups were calculated according to preliminary data and existing literature using power analysis to demonstrate statistical significance (α=0.05, 1-β=0.80). In the first study, 10 TAC, 4 sham and 5 MI animals were included in the imaging study. Additional animals (n=29) were killed on the intermediate and end time point of experiment for tissue analysis (Table 4). These animals from the first study were used for experiments to obtain data for the manuscript “Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure”.

- 40 - Material & Methods

Table 4. Animal groups study 1

Application Number of animals Surgery

18F-FDG PET + K/X d3, d7; 10 TAC CMR d2, d8 4 Sham

18F-FDG PET + K/X d3 5 MI

4 TAC Histology d3 1 MI

8 TAC Histology d8 8 Sham

4 TAC Western immunoblotting 4 Sham Ketamine/xylazine (K/X) anesthesia

In the second study, 56 TAC and 37 sham mice were used for longitudinal multimodal imaging. Due to logistics and organization, not all animals in the longitudinal imaging group were imaged at all time points. Additional mice (n=11 TAC and n=5 sham) were killed 7 days after surgery for ex vivo autoradiography and biodistribution measurements. In the groups that underwent two surgeries for the third study, 14 TAC + rTAC, 6 TAC + Sham rTAC and 3 Sham + Sham operated animals were used. Tissue from these animals was taken at respective intermediate time points and at the end of the study for correlative measurements (Table 5). These animals from the second and third study were used for experiments to obtain data for the manuscript “Multimodality Imaging of Inflammation and Fibrosis during HF Progression and Reverse Remodeling”.

- 41 - Material & Methods

Table 5. Animal groups study 2 & 3

Application Number of animals Surgery

68 Ga-Pentixafor PET 56 TAC d3, d7, d21, d42; CMR; d2, d8, d22, d44; Histology 37 Sham

11 TAC Ex vivo autoradiography & biodistribution d7 5 Sham

14 TAC + rTAC 68Ga-Pentixafor PET Sx I / Sx II + d7, d21; 6 TAC + Sham rTAC CMR Sx I / Sx II + d8, d22; Histology 3 Sham + Sham

Surgery (Sx)

6.2 Surgical Models

6.2.1 Transverse Aortic Constriction

HF was induced by TAC to generate cardiac pressure overload. At the age of 10 weeks, mice underwent TAC or sham surgery under isoflurane anesthesia, as described previously (TARNAVSKI et al. 2004). Mice were pretreated with analgesic carprofen (Rimadyl, Zoetis, 5mg/kg s.c.) and anesthetized by isoflurane (Isofluran-Piramal, induction 4% at 3L/min oxygen). After oral intubation using a 22G venous catheter, anesthesia was maintained at 2% under mechanical ventilation (Harvard MiniVent). Ventilation was kept at 170 breaths/minute with a tidal volume of 0.15ml for a mouse of 25g body weight. Atropine sulfate (Braun) was injected s.c. in order to prevent bradycardia and tracheal mucus accumulation during surgery. Eye ointment (Bepanthen) was applied to protect cornea from desiccation. Fur was removed using animal hair clipper (Aesculap Isis) and skin was disinfected (Octeniderm, Schülke). Skin was opened parallel to ribs using scissors (FST). Intercostal muscles were anesthetized by application of one drop of local anesthesia lidocaine hydrochloride (Xylocain,

- 42 - Material & Methods

AstraZeneca). A lateral thoracotomy in the second intercostal space was performed and the thymus gently retracted using blunt forceps (S&T JFA-5b TC). The aorta was exposed by blunt dissection and gently separated from surrounding tissue between the brachiocephalic trunk and the left carotid artery. A customized blunt 23G needle served as a hook to pass a braided 6-0 silk suture under the aorta. To constrict the transverse aorta, the suture was secured around the vessel and a blunt needle to define a consistent ligature diameter. To obtain heterogeneity within the group, the aorta was narrowed to either 0.5mm using a 25G needle or 0.4mm using a 27G needle which were promptly removed after securing the suture (Figure 12). In total 4 knots fixed the aortic banding. After constriction, the rib cage was closed with one simple interrupted stitch and skin was closed with two simple interrupted everted stitches (6/0 Polyester-S grün, Catgut). Mechanical ventilation was disconnected and mice were extubated when the animals started breathing independently and frequently. When reflexes were restored, animals were allowed to recover separately under infrared light (Phillips) before returning to the home cage. For sham operation, the identical procedure was followed without securing the suture. Following surgery, analgesia was maintained orally over 5 days by tramadol hydrochloride (Tramal Tropfen, Grünenthal) in drinking water. The surgery had a duration of approximately 15min.

- 43 - Material & Methods

Figure 12. Schematic illustration of TAC surgery The transverse aorta was constricted between the brachiocephalic trunk and the left carotid artery using a blunted 25 or 27-gauge (G) needle and a 6-0 silk suture. White arrow direction represents the direction of blood flow and arrow thickness is proportional to blood pressure. Distal to the constriction, pressure is reduced while pressure to the left ventricle and proximal brachiocephalic trunk is markedly elevated. Adapted from LUO et al. 2015.

- 44 - Material & Methods

6.2.2 Reverse Transverse Aortic Constriction

To unload the ventricle, mice underwent rTAC surgery by removal of the aortic banding at 3 weeks after initial TAC surgery. The surgical procedure and the treatment pre and post- surgery were identical to the TAC model and the suture was carefully removed with fine scissors and forceps (Figure 13). The surgery has a duration of approximately 30min. As a negative control, additional TAC animals underwent sham rTAC with thoracotomy and blunt dissection without release of the aortic band. A third group underwent sham TAC and sham rTAC procedures.

Figure 13. Schematic illustration of rTAC surgery Aortic banding was carefully removed by blunt dissection of soft tissue with forceps and cutting the knot with fine scissors after 3weeks of TAC.

6.2.3 Myocardial Infarction

As a positive control group that exhibits robust macrophage infiltration, MI was induced by coronary artery ligation as described previously (THACKERAY et al. 2015b). C57Bl/6N mice at the age of 10 weeks were pretreated with butorphanol analgesic (2 mg/kg, s.c.) and anaesthetized with isoflurane (Isofluran-Piramal, induction at 4% at 3L/min oxygen). After oral intubation using a 22G venous catheter (Vasofix Braunüle, Braun), anesthesia was maintained at 2% under mechanical ventilation (Harvard MiniVent). Ventilation was kept at

- 45 - Material & Methods

170 breaths/minute with a tidal volume of 0.15ml for a mouse of 25g body weight. Eye ointment (Bepanthen) was applied to protect cornea from desiccation. Fur was removed using animal hair clipper (Aesculap Isis) and skin was disinfected (Octeniderm, Schülke). Skin was opened parallel to ribs using scissor (FST). Intercostal muscles were anesthetized by application of one drop local anesthesia (Xylocain 1%, AstraZeneca). Left thoracotomy in the fourth intercostal space was performed. Three chest retractors were inserted between the ribs to maximize surgical field of view. Then, the epicardium was gently pulled apart, and the left anterior descending coronary artery was permanently ligated (5-0 Prolene, Ethicon) (Figure 14). After ligation, the rib cage was closed with one cross stitch and skin was closed with one simple interrupted everted stitch (6/0 Polyester-S grün, Catgut). Mechanical ventilation was disconnected and mice were extubated when the animals started breathing independent and frequent. When reflexes were restored, animals were allowed to recover in a wake-up box under infrared light (Phillips) before returning to the home cage.

Figure 14. Schematic illustration of LAD ligation Left anterior descending coronary artery (LAD) was permanently ligated (blue cross) and causes ischemic injury (yellow) of the anterolateral left ventricular myocardium.

- 46 - Material & Methods

6.3 Cardiac Magnetic Resonance Imaging

6.3.1 Scan Protocol & Image Acquisition

CMR scans were performed on a 7 Tesla small animal MRI system (Pharmascan 70/16, Bruker BioSpin MRI GmbH, Ettlingen, Germany; Software ParaVision 6.0.1). A 72-mm- diameter volume transmit coil (T11070 89/72 Quad to AD, Bruker BioSpin MRI GmbH) was used in combination with an anatomically shaped four-element mouse cardiac phased-array surface receive coil (T20027V3, Bruker BioSpin MRI GmbH). The animals were positioned prone with the chest on the coil, and body temperature was maintained by a water-heated system during examination. Eye ointment (Bepanthen) was applied to protect cornea from desiccation. All scans were performed under isoflurane inhalation anesthesia (Isofluran- Piramal, 3% induction, 1-2% maintenance). Respiration rate was monitored throughout the acquisition and kept between 50-60 breaths per minute. First, a baseline scan was performed to determine functional parameters ~1 week prior surgery. Baseline scan was followed by a longitudinal imaging protocol and scans were conducted at 2, 8, 23 and 44 days after TAC or sham surgery. Mice which underwent TAC and reverse TAC surgery were imaged at 8 days and 3 weeks after TAC and again at 8 days and 3 weeks after reverse TAC (~ four and seven weeks after initial TAC surgery). Each scan started with a localizer sequence to determine the correct animal position on the coil and in the scanner. Then, localizer sequences in axial, coronal and sagittal orientation were performed to center the field of view on the heart. A cine sequence in 4 chamber view was acquired to visualize size, geometry and contractility of the LV and to identify the aortic banding around the aortic arch. These images were then used to arrange the short axis (SA) image stack (9-11 slices) covering the whole LV (Figure 15). Afterwards, the scan was followed by a fast low-angle shot (FLASH) sequence for calibration of slice position. Radial cycle sequence corrected gradient delay and self-gated radial FLASH sequence measured T1 relaxation time (Figure 16). The complete scan acquisition for one animal takes approximately 45min (Table 6).

- 47 - Material & Methods

Figure 15. Arrangement of short axis image slice stack Navigator (green) was positioned in close proximity to short axis slice stack (9-11 slices, yellow) covering the whole left ventricle.

6.3.2 Scan Protocol

Table 6. CMR imaging protocol

Scan Sequence Duration ~

1 Localizer 1min

2 Localizer axial 2min

3 Localizer sagittal 1min

4 Localizer coronal 1min

5 4 Chamber View FLASH 3min

6 Short Axis Cine Brightblood FLASH 17min

7 FLASH 1min

8 Radial Cycle 3min

9 Radial FLASH 6min

- 48 - Material & Methods

T1 mapping. T1

Angle Shot long axis cardiac of (FLASH) Angle sequences

-

Locker Inversion Recovery sequence for Recovery sequence Inversion Locker

-

CMR imaging CMR protocol

. .

16

Figure Figure by orientations followed Fast three in were Low Localizer and function Look and axis short geometry and for

- 49 - Material & Methods

6.3.3 Functional Cardiac Magnetic Resonance Imaging

Cardiac function and geometry were assessed using a navigator-based self-gated 2D spoiled gradient echo sequence (IntraGate FLASH, Bruker Biospin MRI GmbH, Ettlingen, Germany) as described previously (FRIES et al. 2012; ZUO et al. 2017). Cine images were acquired in four chamber view, in the plane of the aortic arch to visualize the aortic banding. A SA cine sequence covered the whole left ventricle to calculate functional parameters. SA sequence parameters were: 25×25mm2 field of view, 196×196 pixel matrix size, 0.9mm slice thickness, 0.9mm slice distance, 84.64ms repetition time, 1.84ms echo time, 45° flip angle, 24 reconstructed phases, ~17 minutes acquisition time.

6.3.4 T1 Mapping Acquisition

For T1 mapping, a Look-Locker Inversion-Recovery sequence with 2D radial k-space sampling was implemented (WINTER et al. 2016; WECH et al. 2018). Sequence parameters were: 35 x 35mm2 field of view, 128x128 pixel matrix size, 1mm slice thickness, 3.3ms repetition time, 1.6ms echo time, 1 slice, 2048 spokes, 20 repetitions, ~4 minutes acquisition time. Radial k-space data were corrected for eddy currents (BLOCK et al. 2007; MOUSSAVI et al. 2014). After sorting the data according to its cardiac phase (five cardiac phases for the total cardiac cycle) and its inversion time after application of the inversion pulse (64 inversion times, ~ 7ms – 6600ms) using its direct current signal (WINTER et al. 2016), images were reconstructed using a combination of parallel imaging and a total variation constraint for the inversion times and the cardiac phases (MOUSSAVI et al. 2014).

- 50 - Material & Methods

6.3.5 Image Analysis

Image analysis was performed with cvi42 (Version 5.3.6 Circle Cardiovascular Imaging, Calgary, Canada) to determine functional parameters and assess heart morphology. First, end systolic and end diastolic cardiac phase were selected from 24 reconstructed gates. Then, epicardial and endocardial contours were manually drawn into short axis slices covering base to apex. Papillary muscles were included into the volume of the left ventricle. The parameter ventricle volume at end systole (ESV, µL) and end diastole (EDV, µL) calculated stroke volume (SV, µL) ejection fraction (EF, %) and left ventricular mass (LV mass, mg) (SUINESIAPUTRA et al. 2018) (Figure 17).

Equation 1 SV = EDV − ESV

(EDV−ESV) Equation 2 EF = × 100 EDV

mg Equation 3 LV Mass = (Epicardial − Endocardial Volume) × 1.05 µL

T1 maps were calculated using a 3-parameter fit of a mono-exponential function (NEKOLLA et al. 1992). After manual segmentation of the myocardium using ImageJ (Vers.1.50e, National Institutes of Health, USA), mean T1 relaxation time was assessed for each mouse.

- 51 - Material & Methods

Figure 17. Representative cardiac functional analysis Epicardial (green) and endocardial (red) contours were manually drawn in end systolic and end diastolic cardiac phase covering the whole left ventricle from base to apex (9-11 slices) to calculate volumes and cardiac function.

- 52 - Material & Methods

6.4 Positron Emission Tomography

6.4.1 Radiosynthesis

68Ga-Pentixafor was synthesized as previously described (THACKERAY et al. 2015b), using an automated module and CPCR4.2 precursor from Scintomics GmbH (Fürstenfeldbruck, Germany). 18F-FDG was synthesized using standard kits and production methods.

6.4.2 Animal Preparation

PET scans were performed using a dedicated small animal PET-CT system (Inveon DPET, Siemens, Knoxville, USA), as previously described (THACKERAY et al. 2015b). Mice underwent a longitudinal PET imaging protocol at 3, 7, 21, and 42 days after surgery. Mice which underwent both TAC and reverse TAC surgery were imaged at 7 and 21 days after TAC and reverse TAC, respectively. Eye ointment (Bepanthen) was applied to protect cornea from desiccation. Body temperature was maintained during examination. For intravenous (i.v.) application of PET tracer, custom built catheters were inserted into a lateral tail vein prior to scan. Catheters were constructed from a 27G sharp cannula (Terumo), a 27G blunt cannula (Braun) and flexible polyethylene tubing (Tygon) with an inner diameter of 0.01 inches and a length of 20cm (dead volume = 60µL). For catheterization, animals were positioned prone on a heating pad with the nose in the anesthesia mask. Tails were warmed by water for optimal vasodilation and visualization of the veins. With padded forceps, blood flow was gently restricted, the 27G catheter was inserted and secured with transpore tape (Leukopor). For dynamic 18F-FDG scans, mice were transferred to the imaging bed (Minerve) and into the PET scanner. 18F-FDG bolus was injected just after start of the PET scan. For static 68Ga-pentixafor scans, mice were injected, catheter removed, and animals allowed to wake up for the uptake phase. They were then re-anesthetized ~10min prior to the scan, which was conducted from minute 50 to 60 post-injection. Anesthetized mice were positioned prone on the imaging bed with the heart centered in the scanner field of view.

- 53 - Material & Methods

6.4.3 18F-FDG Scan Acquisition, Reconstruction & Analysis

To image inflammatory cells in the myocardium, 18F-FDG PET scans were performed at 3 and 7 days after TAC, sham or MI surgery (Figure 18). The scans were performed under intraperitoneally (i.p.) administered ketamine (84mg/kg) / xylazine (11.2mg/kg) (K/X) anesthesia in order to suppress glucose uptake by cardiomyocytes (THACKERAY et al. 2015a). Respiration rate during the scans was monitored during acquisition with a rate between 160 and 200 breaths per minute. Blood glucose levels were measured prior to and after 18F-FDG PET scan from vena saphena and vena caudalis (tail vein), respectively. Catheters were inserted into tail veins and mice were positioned prone on the imaging bed (Minerve) with the heart centered in the scanner field of view. 12.7±0.9 MBq of 18F-FDG was injected i.v. as a 0.15mL bolus. A 60min dynamic PET scan was acquired. Dynamic images were histogrammed into 32 frames and reconstructed using an iterative ordered-subset expectation-maximization 3- dimensional/maximum-a-posteriori (OSEM3D/MAP) algorithm as described previously (THACKERAY et al. 2015a). All PET images were reconstructed to a 128 × 128 × 159 image matrix (0.78 × 0.78 × 0.80 mm) using an OSEM3D/MAP algorithm (β = 0.01, 2 OSEM iterations in 16 subsets, 18 MAP iterations) with scanner-applied corrections for decay and scatter. A 57Co transmission scan (10min) was used for attenuation correction (THACKERAY et al. 2015b). Analysis was performed with Inveon Research Workplace (Version 4.2; Siemens Medical Solutions USA) to evaluate LV volume and signal. A region of interest was defined on the LV myocardium by interactive thresholding as previously described (THACKERAY et al. 2015a). The size and position of the region of interest were verified by anatomic landmarks in the fused CT images. The signal was evaluated semiquantitatively as percentage injected dose per gram of tissue (%ID/g).

- 54 - Material & Methods

Figure 18. Scan protocol 18FDG Mice underwent 60min dynamic 18F-FDG PET scan under ketamine/xylazine (K/X) anesthesia. Blood glucose was measured prior to and after PET scan. Finally, a CT scan was conducted.

6.4.4 68Ga-Pentixafor Scan Acquisition, Reconstruction & Analysis

With 68Ga-pentixafor PET imaging, we aimed to characterize the acute and long-term inflammatory response to pressure overload. To reduce animal numbers and minimize the burden, we chose the most critical time points in the development of inflammation and fibrosis after pressure overload according to previous studies (WEISHEIT et al. 2014; PATEL et al. 2017). The scans were performed under isoflurane inhalation anesthesia (3% induction, 1-2% maintenance). Respiration rate during the scans was monitored during acquisition and kept between 50-60 breaths per minute. 12.5±1.5 MBq 68Ga-pentixafor was injected i.v. as a 0.10- to 0.15-ml bolus via the tail vein. A 10min static scan was acquired at 50min after injection under isoflurane anesthesia. This was followed by a CT scan for coregistration. For colocalization and to define left ventricle contours, 19.83±1.10 MBq 18F-FDG (0.2ml) was administered i.p. under isoflurane anesthesia after the 68Ga-pentixafor and CT scan, and a 10min static scan was performed 20min after injection (Figure 19). All PET images were reconstructed to a 128 × 128 × 159 image matrix (0.78 × 0.78 × 0.80 mm) using an OSEM3D/MAP algorithm (beta = 0.01, 2 OSEM iterations, 18 MAP iterations)

- 55 - Material & Methods

with scanner-applied scatter correction. A 57Co transmission scan (10min) was used for attenuation correction (THACKERAY et al. 2015b). Image analysis was performed with Inveon Research Workplace 4.2 (Siemens Medical Solutions USA) to evaluate left ventricular volume and signal. First, coregistration 18F-FDG images were used to manually create regions of interest (ROIs) of LV myocardium by interactive thresholding as previously described (THACKERAY et al. 2015b). These ROIs were imported into 68Ga-pentixafor images to determine left ventricular position and calculate 68Ga-pentixafor uptake. The signal was evaluated semi-quantitatively as %ID/g (Figure 20).

- 56 - Material & Methods

Figure 19. Scan protocol 68Ga-pentixafor / 18F-FDG Scans were performed under continuous isoflurane anesthesia. A static 10min 68Ga-Pentixafor PET scan was performed 50min after intravenous injection and followed by a CT scan. A second static 10min 18F-FDG scan was conducted 20min after intraperitoneal injection.

Figure 20. 68Ga-Pentixafor image analysis. 18F-FDG images were used for colocalization of left ventricular myocardium to determine myocardial CXCR4 upregulation.

- 57 - Material & Methods

6.5 Ex Vivo Tissue Workup

All animals were killed at the terminal endpoint or intermediate time points by cervical dislocation under isoflurane anesthesia. Mouse hearts were immediately removed, cleaned in phosphate-buffered saline (PBS) and weighed. The atria were then removed, the heart sliced in two along the long axis, and frozen under liquid nitrogen in optimal cutting temperature compound (O.C.T. Tissue-Tek). Tissue was stored at -70°C until use.

6.5.1 Ex Vivo Biodistribution

60min after i.v. 0.15µL bolus injection of 11.7±0.8 MBq 68Ga-pentixafor, ex vivo biodistribution measurements of various organs were performed at 7 days after surgery. Blood samples were taken by cardiac puncture at 60min post-injection and hearts were immediately removed and rinsed in PBS. Tissue was prepared as described above, with additional dissection of the right ventricle wall. Right ventricle, left ventricle, spleen, lung, liver, kidney, skeletal muscle and blood were excised and placed in pre-weighted tubes and counted for radioactivity in a gamma counter (PerkinElmer) with a known dilution of the injected dose. The left ventricle was cut in half along the ventricle long axis. One half was counted for ex vivo biodistribution, while the other was snap-frozen for sectioning and autoradiography exposure. Radioactivity counts were normalized by the injected radioactivity and the weight of tissue. Results are shown as %ID/g.

6.5.2 Ex Vivo Autoradiography

The left ventricle long axis was sliced into 10µm sections using a cryostat (LEICA CM3050 S) and air-dried. Slides were exposed for 60min to a phosphor imaging screen (PerkinElmer) in a light-impermeable cassette. The screen was then digitalized using a radiometric imager (Cyclone, OptiQuant, PerkinElmer). Quantitative analysis of the autoradiograms was performed with PMOD software (Version 3.7). Circular ROIs were drawn manually around each heart section and optical density values were calculated. Background values from the screen were subtracted and results shown as optical density.

- 58 - Material & Methods

6.5.3 Western Immunoblotting

Hearts were harvested as described and dissected to separate the left and right ventricle 8 days after surgery. Western immunoblotting was performed as described previously (THACKERAY et al. 2011). Briefly, right and left ventricular tissue samples were homogenized in lysis buffer and proteins were separated by gel electrophoresis. Primary antibodies used were mouse monoclonal anti-glucose transporter-1 GLUT1 (SPM498, 1/5000), mouse monoclonal anti-glucose transporter-4 GLUT4 (ab65267, 1µg/ml) (Abcam). Secondary antibody was rabbit polyclonal anti-mouse IgG conjugated to horseradish peroxidase (Abcam). Immunoblots were analyzed by chemiluminescence and integrated density values normalized to β-actin loading control were compared between groups.

6.5.4 Histology

Masson trichrome and picrosirius red staining protocols identified interstitial and perivascular collagen in the myocardium at 8, 23, and 44 days after TAC and sham. Adjacent cryosections of the left ventricle in long axis (6µm) sections were thaw-mounted onto charged slides (Superfrost Plus). For Masson trichrome staining, we followed the manufacturer`s protocol (Sigma-Aldrich Accustain Trichrome Stains (Masson)). Briefly, slides were first incubated in Bouin`s fixative solution for at least 4 hours. Then, slides were rinsed for 10min in running cold tap water. Slides were incubated in Working Weigert`s iron hematoxylin for 2min and washed for 5min in running tap water. After rinsing in PBS, slides were incubated in Biebrich Scarlet-Acid- Fuchsin for 5min, followed by additional rinsing in deionized water. A mixture of deionized water, phosphotungstic acid solution, phosphomolybdic acid (2/1/1) was prepared and slides were incubated for 7min in this solution. Afterwards, slides were incubated in Aniline Blue for 8min, followed by fixing in 1% acetic acid. Finally, slides were dehydrated through several changes of alcohol and cleared in xylene. After drying overnight, slides were coverslipped with Eukitt mounting medium. Masson trichrome staining stains nuclei in black, cytosol in red and collagen in blue. Picrosirius red staining followed standard protocol (SCHIPKE et al. 2017). Slides were fixed in acetone in room temperature for 15min, then washed in PBS and allowed to dry completely. 500mg direct red (Sigma-Aldrich) were diluted in 500mL picric acid (Sigma-

- 59 - Material & Methods

Aldrich), and slides were incubated in this solution for 60min. Next, acidified water (5mL glacial acetic acid / 1L tap water) washed the slides twice before dehydrating through ethanol and xylene clearance. Slides were allowed to dry overnight, and coverslipped with Eukitt mounting medium (Sigma-Aldrich). Picrosirius red staining stains myocytes in yellow and collagen in red. Overview images (2.5x) and representative high magnification images were captured at a microscope (Zeiss). For quantification, 6 images (200x magnification) were taken from selected sections from individual mice. Images were analyzed for fibrotic area (%) of the tissue using ImageJ (Version 1.50e). For picrosirius red staining, as collagen is stained red, the images were converted to 3 slice (red, green, blue) stacks. Here, the green channel is used to automatically threshold and quantify the contrast of stained collagen as a percentage of the total stained field. Mean fibrotic area (%) was calculated for each animal.

6.5.5 Immunohistochemistry

Myocardial inflammatory cell infiltration was evaluated using an immunohistochemical staining protocol. Immunostaining for CD68 identified macrophages and for Ly6G detected granulocytes. For immunostaining, slides were warmed to room temperature and fixed in acetone for 15min. After rinsing slides in PBS, they were allowed to dry completely for 30min. Each heart section was then circled with a Dako Pen to create a water repellent circle around the probe. Then, slides were incubated with 3-4 drops avidin blocking solution per slide for 10min and rinsed in PBS afterwards, and 3-4 drops biotin blocking solution for 10min (Avidin-Biotin Kit, Dako) to inhibit endogenous reactivity. Slides were rinsed in PBS and blocked in 10% horse serum for 30min. After serum removal, slides were incubated in biotin- conjugated primary antibody (rat anti mouse CD68 or rat anti mouse Ly6G, BIO-RAD; each 1:100 diluted in PBS) for 60min. Slides were then rinsed 3x for 3-5min each in PBS, then incubated with 0.3% H2O2 for 10-15min and rinsed again 3x in PBS. Afterwards, slides were incubated with streptavidin-peroxidase (1:100 diluted in PBS) for 30min. After a final rinse

(3x in PBS), 3,3′-Diaminobenzidine solution (1 tablet in 15ml PBS and 12µL H2O2 (30%) was added to slides, and allowed to react for 10-15min. Slides were finally rinsed in deionized water and counterstained with Mayer`s hematoxylin for 2min. Afterwards, slides were

- 60 - Material & Methods

washed for 10min in running tap water. Finally, slides were dehydrated through ethanol and cleared in xylene. Slides were allowed to dry overnight and coverslipped with Eukitt mounting medium (Sigma-Aldrich). For quantification, 6 images (200x magnification) per section of each heart were taken at a microscope (Zeiss). Positive cells were manually counted using ImageJ cell counter (Version 1.50e). Mean number of cells / field was calculated for each animal for group comparison.

6.6 Analysis & Statistics

Statistical analysis was carried out with Prism 6 (GraphPad, San Diego, California). All data are shown as mean ± SD. Serial imaging data obtained in individual animals were analyzed using Student’s paired t-test. Protein expression, cell content, fibrosis, imaging markers, activity and optical density were compared between TAC and sham groups using the Welch`s t-test due to unequal variance. Pearson product-moment correlation coefficients were calculated to describe the relationship between pairs of continuous variables. P < 0.05 was considered statistically significant.

- 61 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

7 Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Aylina Glasenapp,1,2 Katja Derlin,2 Yong Wang,3 Marion Bankstahl,4 Martin Meier,4 Kai C Wollert,3 Frank M Bengel,1 James T Thackeray1

Departments of 1Nuclear Medicine, 2Radiology, and 3Cardiology and Angiology; 4Central Laboratory Animal Facility and Institute for Laboratory Animal Science Hannover Medical School, Hannover, Germany

Short title: FDG imaging of inflammation after TAC Word Count: 4487

This research was originally published in JNM. Glasenapp A, Derlin K, Wang Y, Bankstahl M, Meier M, Wollert KC, Bengel FM, Thackeray JT. Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure. J Nucl Med. published on October 25, 2019 as doi:10.2967/jnumed.119.232488. © SNMMI.

AG conducted experiments, analyzed data, drafted manuscript KD designed study, analyzed data, revised manuscript YW performed surgeries MB revised manuscript MM acquired magnetic resonance imaging data KW supported biochemical analysis, interpreted data, revised manuscript FMB designed study, interpreted data, revised manuscript JTT designed study, conducted experiments, analyzed data, wrote manuscript

First Author Aylina Glasenapp, Department of Radiology, Hannover Medical School, Carl Neuberg Str 1, 30625 Hannover, Germany

Corresponding Author James T Thackeray, PhD, Department of Nuclear Medicine Hannover Medical School, Carl Neuberg Str 1, 30625 Hannover, Germany [email protected]

- 62 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Inflammation contributes to ventricular remodeling after myocardial ischemia, but its role in non-ischemic heart failure is poorly understood. Local tissue inflammation is difficult to assess serially during pathogenesis. While 18F-fluorodeoxyglucose (FDG) accumulates in inflammatory leukocytes which may identify inflammation in the myocardial microenvironment, it remains unclear whether this imaging technique can isolate diffuse leukocytes in pressure overload heart failure. We aimed to evaluate the capability to serially image inflammation with 18F-FDG in the early stages of pressure overload-induced heart failure, and to compare the timecourse to functional impairment assessed by cardiac magnetic resonance imaging (MRI). Methods. C57Bl6/N mice underwent transverse aortic constriction

(TAC, n=22) or sham surgery (n=12), or coronary ligation as an inflammation-positive control (n=5). MRI assessed ventricular geometry and contractile function at d2 and d8 after

TAC. Immunostaining identified the extent of inflammatory leukocyte infiltration early in pressure overload. 18F-FDG PET scans were acquired at d3 and d7 after TAC, under ketamine-xylazine anesthesia to suppress cardiomyocyte glucose uptake. Results. Pressure overload evokes rapid left ventricular dilation compared to sham (end systolic volume, d2:

40.6±10.2µL vs. 23.8 ±1.7µL, p<0.001). Contractile function was similarly impaired (ejection fraction, d2: 40.9 ± 9.7% vs. 59.2 ± 4.4%, p<0.001). Severity of contractile impairment was proportional to histology-defined myocardial macrophage density at d8 (r=-0.669; p=0.010).

Positron emission tomography (PET) imaging identified significantly higher 18F-FDG accumulation in the left ventricle of TAC compared to sham mice at d3 (10.5±4.1%ID/g vs.

3.8±0.9%ID/g, p<0.001) and at d7 (7.8±3.7%ID/g vs.3.0 ±0.8%ID/g p=0.006), though the efficiency of cardiomyocyte suppression was variable between TAC mice. The 18F-FDG signal correlated to ejection fraction (r=-0.75, p=0.01) and ventricle volumes (r=0.75, p<0.01). Western immunoblotting demonstrated 60% elevation of myocardial glucose

- 63 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

transporter 4 expression in the left ventricle at d8 after TAC, indicating altered glucose metabolism. Conclusion. TAC induces rapid changes in left ventricle geometry and contractile function, with a parallel modest infiltration of inflammatory macrophages.

Metabolic remodeling overshadows inflammatory leukocyte signal using 18F-FDG PET imaging. More selective inflammatory tracers are requisite to selectively identify the diffuse local inflammation in pressure overload.

Key words: Heart failure, PET, 18F-FDG, inflammation, MRI, macrophage

- 64 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

INTRODUCTION

Heart failure (HF) is a multifactorial syndrome and the common endpoint of many cardiovascular diseases. Despite improvements in early treatment strategies, HF remains a prevalent cause of death worldwide and a significant burden on the healthcare system (1).

Conventional therapeutic approaches combine multiple drugs targeting nonspecific symptoms and common mechanisms of HF, to support cardiac function and improve survival (2). But various specific pathogenetic mechanisms contributing to development and progression of HF remain undertreated, particularly in non-ischemic heart failure (3). Recent evidence suggests that inflammatory leukocytes and macrophages may play a role in non-ischemic HF progression (4), and contribute to interstitial fibrosis and contractile dysfunction (4). HF patients exhibit elevated circulating white blood cells, inflammatory cytokines and biomarkers, regardless of pathogenesis (5,6). As such, inflammation has emerged as a potential therapeutic target for not only ischemic heart failure, but also non-ischemic cardiomyopathies (7).

Acute infiltration of inflammatory leukocytes predicts functional decline and progression of HF following ischemic injury (8). Conversely, inflammatory cell infiltration and adverse ventricular remodeling in non-ischemic pressure overload is less well characterized, but the temporal dynamics and involved cell population are thought to be distinct from ischemic injury (9). At 7d after aortic banding, total macrophages and leukocyte content of the left ventricle is maximal (10-12), and may contribute to ventricular remodeling

(13,14). While systemic inflammation may be quantified through biomarkers or leukocyte count, local myocardial inflammatory activity is more difficult to assess. Positron emission tomography (PET) using 18F-deoxyglucose (18F-FDG) robustly accumulates in activated inflammatory leukocytes including macrophages (15). In the myocardium, however, this is

- 65 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

complicated by physiologic uptake of FDG into viable myocytes, which is further enhanced by ischemia or other injury (16). Acute pressure overload instigates rapid changes in myocyte glucose metabolism, which is believed to contribute to ventricular hypertrophy (17,18).

Nevertheless, inflammation-targeted 18F-FDG imaging, using preparation protocols for suppression of myocyte uptake, has demonstrated prognostic value for outcomes after myocardial infarction in patients and animal models (8,19). But the feasibility of imaging myocardial inflammation in non-ischemic heart failure is undefined.

We hypothesized that 18F-FDG PET imaging would identify infiltrating macrophages and inflammatory leukocytes in the myocardium early in the pathogenesis of non-ischemic pressure overload-induced HF. Accordingly, we compared the uptake of 18F-FDG in the left ventricle after transverse aortic constriction in mice (using a myocyte uptake suppression protocol) to the progressive changes in left ventricle geometry and contractile function assessed by cardiac magnetic resonance imaging (MRI). This was complemented by histological and biochemical tissue workup to determine factors underlying the imaging signal.

METHODS

Animals

All animal experiments were approved by the state authority (Niedersächsisches

Landesamt für Verbraucherschutz und Lebensmittelsicherheit) and local animal ethics review.

Male C57Bl/6N mice (n=39, 23.7±1.5g) were purchased from Charles River (Sulzfeld,

Germany) and housed in groups in a temperature-controlled facility under a 14h/10h light/dark cycle with food and water freely available.

- 66 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Transverse aortic constriction (TAC)

Mice underwent TAC (n=22) or sham surgery (n=12) under isoflurane anesthesia, as described previously (20). Briefly, mice were pretreated with analgesic carprofen (5mg/kg sc) and anesthetized by isoflurane (induction 3-4%, 3L/min oxygen; maintained 1.5-2% after oral intubation under mechanical ventilation). Atropine sulfate (0.05mg/kg, sc) prevented bradycardia and tracheal mucus accumulation. A lateral thoracotomy in the second intercostal space was performed and the thymus gently retracted. The aorta was exposed by blunt dissection and gently separated from surrounding tissue between the brachiocephalic trunk and the left carotid artery. To constrict the transverse aorta, a 6-0 silk suture was secured around the vessel using a blunt 25G needle to define a consistent ligature diameter. After securing the suture, the needle was removed, the rib cage and skin closed, and the animal allowed to recover before returning to the home cage. For sham operation, the identical procedure was followed without securing the suture. Following surgery, oral analgesia was maintained with Tramadol (2.5mg/100ml in drinking water).

Myocardial infarction (MI)

As a positive control that exhibits robust macrophage infiltration, MI was induced by coronary artery ligation as described (21). Briefly, mice (n=5) pretreated with butorphanol analgesic (2mg/kg sc) were anaesthetized with isoflurane (induction at 3-4%, 3L/min oxygen; maintained at 1.5-2% after oral intubation under mechanical ventilation). Left thoracotomy was performed and the left anterior descending coronary artery was permanently ligated.

Study Design

Ventricular geometry and contractile function early in pressure overload heart failure was assessed by cardiac MRI at d2 and d8 after surgery. Inflammation was serially assessed by 18F-FDG positron emission tomography (PET) under ketamine-xylazine anesthesia to

- 67 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

suppress cardiomyocyte glucose uptake at d3 and d7 after surgery. These imaging timepoints were selected based on the expected maximal inflammatory cell infiltration (d7) based on flow cytometry studies (10,14), with an intermediate timepoint to allow evaluation of changes over the first week after TAC. All mice were killed by cervical dislocation at d8-9 for ex vivo tissue workup. An additional group was killed at d3 for histologic workup. 18F-FDG images were acquired at d3 after surgical MI, which evokes robust inflammatory cell infiltration of the infarct territory.

Histopathology

Mice were killed at d3 (n=4) after TAC or after MI (n=1) and at d8 (n=8) after TAC or after sham surgery (n=8), for histologic assessment of inflammation. Hearts were removed, cleaned in phosphate buffered saline, and frozen under optimal cutting temperature compound

(Tissue Tek). Adjacent long axis 10µm sections were sliced using a cryostat (Shandon) and thaw-mounted onto charged slides (Superfrost Plus). Immunostaining for CD68 and Ly6G identified macrophages and granulocytes, respectively.

Small animal magnetic resonance imaging

Mice (TAC n=10, sham n=4) underwent MRI at 2d and 8d after surgery using a 7

Tesla small animal MRI system (Pharmascan 70/16, Bruker BioSpin GmbH, Ettlingen,

Germany; Software ParaVision 6.0.1). Cardiac function and geometry were evaluated by using a navigator-based self-gated cine MRI sequence (IntraGate FLASH, Bruker Biospin,

Ettlingen, Germany) as described previously (22).

MR image analysis

Image analysis was performed with CVI (Vers. 5.3.6) to determine ventricle volumes and assess heart function. Endocardial and epicardial contours were defined on short axis

- 68 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

sections and used to calculate ventricular mass at end systole and end diastole; ejection fraction was calculated from endocardial volumes.

Small animal positron emission tomography

Mice (TAC n=10, sham n=4; MI n=5) underwent PET imaging using a dedicated small animal PET camera (Inveon DPET, Siemens), as previously described (21). Briefly, mice were anesthetized with ketamine (84mg/kg ip) / xylazine (11.2mg/kg ip) to suppress glucose uptake by cardiomyocytes (21), and positioned prone on the imaging bed with the heart centered in the scanner field of view. 18F-FDG (12.7±0.9 MBq) was administered as a

0.15mL bolus through a lateral tail vein catheter. A 60 min dynamic listmode PET scan. 18F-

FDG was synthesized using standard kits and production methods. One mouse died between the first and second PET acquisitions.

PET Image Analysis

Images were dynamically histogrammed and reconstructed using iterative

OSEM3D/MAP as described previously (21). Analysis was performed with Inveon Research

Workplace 4.2 (Siemens Medical Solutions USA) to evaluate left ventricular volume and signal. A region of interest (ROI) was defined on the left ventricle myocardium by interactive thresholding as previously described (21). Size and position of the ROI was verified by anatomic landmarks in the fused CT images. The signal was evaluated semi-quantitatively as percent injected dose per gram of tissue (%ID/g). Additional details are provided in the online supplement.

Western Immunoblotting

Hearts from a subset of animals (TAC n=4, sham n=4) were processed for immunoblotting using lysates of left and right ventricle. Additional details are provided in the online supplement.

- 69 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Statistics

Statistical analysis was conducted using Prism 6 (GraphPad). All data are shown as mean ± SD. Serial imaging data obtained in individual animals were compared using

Student’s paired t-test. TAC and sham groups were compared using the Welch t-test for unequal variance. Pearson product-moment correlation coefficients described the relationship between pairs of continuous variables. P<0.05 was considered statistically significant.

RESULTS

Pressure Overload Evokes Early Left Ventricular Dilation and Cardiac Dysfunction

Within 2d of TAC, cardiac MRI demonstrated 30% increase in left ventricle mass and a corresponding elevation in ventricular volumes compared to sham (Fig. 1A). Ventricular mass was consistently elevated in the first week after TAC surgery (d2: 87.5±14.6 mg vs.

66.3±8.4mg, p=0.007; d8: 89.3±7.8mg vs. 67.3±9.5mg, p=0.010) (Fig. 1B). Similarly, a progressive increase in end systolic volumes and end diastolic volumes were identified from 2 and 8d post-surgery (ESV d2: 40.6±10.2µL vs. 23.8 ±1.7µL, p<0.001; d8: 47.2±11.1µL vs.

17.8±3.1 p<0.001) (EDV d2: 68.5±7.2µL vs. 58.5 ±2.4µL, p=0.002; d8: 75.4±9.9 vs.

56.3±5.9, p=0.002) compared to sham (Fig. 1B). TAC animals exhibited a marked reduction in ejection fraction at 2d (40.9±9.7% vs. 59.2±4.4%, p<0.001) and 8d post-surgery (37.5±11.8 vs. 68.3±3.4, p<0.001) compared to sham (Fig. 1B), consistent with early and persistently impaired contractile function after pressure overload.

Left Ventricle Macrophage Infiltration Correlates with Contractile Dysfunction

To investigate myocardial inflammation, CD68 and Ly6G immunostaining identified inflammatory leukocyte infiltration. CD68-positive macrophages were noted throughout the

- 70 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

left ventricle after TAC, at a diffuse density compared to myocardial infarction (Fig. 2A).

Semi-quantification of cell infiltration revealed a threefold elevation of CD68+ cells in the left ventricle of TAC mice compared to sham at d3 and twice as many cells at d8 (d3:

111.9±61.8 cells/field, p=0.094; d8: 74.4±24.3 cells/field, p=0.003; vs. sham: 37.6±11.2 cells/field) (Fig. 2B). Macrophage density after TAC was much lower compared to extensive staining after MI, where quantitative calculation is complicated by the large abundance of

CD68+ staining (Fig. 2A). Nevertheless, in TAC and sham mice, the CD68 cell count inversely correlated with ejection fraction (r=-0.669; p=0.010) (Fig. 2B), suggesting a relationship between inflammatory cell infiltration and functional decline. Ly6G-positive granulocytes were present but at lower density than macrophages after TAC (Supplemental

Fig. 1).

Myocardial 18F-FDG Signal under Ketamine-Xylazine (K/X) Suppression is Elevated after TAC

K/X anesthesia significantly lowered cardiomyocyte 18F-FDG uptake in sham operated animals to background levels (<5%ID/g) (Supplemental Fig. 2). In inflammation positive control MI mice, cardiomyocyte uptake of 18F-FDG was effectively suppressed, with global myocardial PET signal <10%ID/g, and a hotspot localized to the infarct territory, consistent with histology findings (Fig. 3A). By contrast, TAC-operated animals exhibited a range of modest to robust 18F-FDG accumulation throughout the left ventricle, despite significantly lower inflammatory cell content. The left ventricular 18F-FDG uptake in TAC operated mice exceeded that of MI at 3d (8/10 mice) and at 7d (4/9 mice), suggesting ineffective cardiomyocyte suppression despite identical K/X protocol (Supplemental Fig. 3).

Semi-quantitative analysis confirmed higher 18F-FDG signal in the left ventricle of TAC mice

- 71 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

compared to sham at d3 (+176%, 10.5±4.1%ID/g vs. 3.8±0.9%ID/g, p<0.001) and at d7

(+160%, 7.8±3.7%ID/g vs.3.0 ±0.8%ID/g p=0.006) (Fig. 3B). The inconsistency of the PET signal localization and intensity necessitated further interrogation of the source of this 18F-

FDG uptake.

TAC Induces Altered Myocardial Glucose Transporter Expression

Since 18F-FDG uptake was more robust than expected from histologic evidence of leukocyte infiltration, we further investigated the efficacy of K/X to suppress glucose uptake after TAC. Blood glucose measurements at the time of acquisition confirmed hyperglycemia in both TAC and sham-operated animals (Fig. 4A), and was comparable between mice above and below the 10%ID/g threshold (p=0.35). Western immunoblotting demonstrated no difference in insulin-independent GLUT1 expression between TAC and sham operated animals in right or left ventricle. However, GLUT4 expression was elevated by 60% in TAC compared to sham animals, specifically in the left ventricle (Fig. 4B). Accordingly, metabolic changes of substrate utilization in pressure overloaded cardiomyocytes are a major contributor to the increased 18F-FDG PET signal after TAC.

Elevated Myocardial 18F-FDG-Uptake is Related to Early Functional Impairment

As both inflammation and metabolic changes may influence ventricular remodeling and cardiac function, we assessed the relationship between left ventricle 18F-FDG signal and cardiac function. At 3d post-TAC a strong correlation to ejection fraction (r=-0.75, p=0.01),

ESV (r=0.75, p<0.01) and EDV (r=-0.67, p<0.05) (Fig. 5A-C) was observed, which may suggest that animals with worse cardiac function exhibit ineffective cardiomyocyte suppression reflecting altered metabolic substrate utilization. This relationship was less

- 72 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

prominent for LV mass and stroke volume at 3d (Fig. 5D), and all parameters at 7d

(Supplemental Fig. 4).

- 73 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

DISCUSSION

Myocardial inflammation in cardiovascular disease is a powerful prognosticator of functional decline and an emerging therapeutic target. While strongly implicated in acute myocardial infarction and ischemic HF, inflammatory leukocyte contribution to non-ischemic pressure overload HF is less well characterized, owing in part to the challenge in non-invasive interrogation of local inflammation. Here, we evaluated the temporal relationship between ventricular geometric change, contractile function, and inflammation, and the capability of

18F-FDG to non-invasively image inflammatory cell content. Histological analysis revealed mild diffuse inflammatory cell infiltration of the left ventricle early after pressure overload, but while 18F-FDG uptake was increased, the intensity of the signal was disproportionate to the severity of inflammation. Tissue immunoblotting demonstrated a selective upregulation of

GLUT4 in the overloaded left ventricle. The strain incurred by the myocardium during pressure overload renders K/X suppression of cardiomyocyte glucose transport ineffective, such that the 18F-FDG signal early after TAC, while correlating with cardiac dysfunction, may partly reflect inflammation, but is probably dominated by metabolically compromised cardiomyocytes.

The role and time course of inflammatory cell invasion in myocardial infarction is well characterized, describing dynamic mobilization of leukocytes (23). By comparison, the role of inflammation in non-ischemic heart failure remains equivocal, though flow cytometry has identified inflammatory cell content in late-stage pressure overload heart failure (10).

Previous studies demonstrate transient pro-inflammatory cytokines upregulation, and macrophage, but not granulocyte, recruitment, reaching maximum at 7d after aortic constriction (10,11). Monocyte-derived cardiac macrophages are considered main contributors to adverse remodeling after TAC (14). In the present study, we observe modest

- 74 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

diffuse CD68+ macrophage infiltration of the left ventricle beginning from 3d after TAC and persisting to 7d. Notably, MRI revealed elevated ventricle mass, dilatation, and contractile dysfunction in parallel to or preceding inflammatory cell infiltration. It is difficult to distinguish the role of inflammation in ventricular remodeling considering need for multiple timepoint tissue sampling.

Non-invasive interrogation of inflammatory leukocytes is therefore attractive to monitor disease progression. In several preclinical and clinical studies, 18F-FDG uptake identifies inflammation in cardiovascular diseases e.g. acute MI (19), preferentially targeting metabolically active pro-inflammatory macrophages in cardiac tissue (15,24). However, the robust uptake of 18F-FDG by cardiomyocytes limits the ability to specifically target inflammation in the heart. Previous studies established significantly higher 18F-FDG uptake by the left ventricle within 24h of aortic constriction, which was effectively inhibited by beta blocker therapy (18). Elevated glucose utilization is associated with increased glucose-6 phosphate expression (17), which would be consistent with inability to suppress the cardiomyocyte 18F-FDG uptake in the present study. Moreover, the increased rate of glucose transport gradually increases over weeks of pressure overload (17), which may reflect the higher success of K/X cardiomyocyte suppression at the earlier timepoint (d3) of the current study. Accordingly, despite comparably elevated blood glucose levels following K/X anesthesia, upregulation of GLUT4 in the overloaded left ventricle prevents the normal suppression pattern. Notably, the chronic stages of pressure overload lead to regional hypometabolism, wherein the extent of myocardial metabolic defects correlates with change in ventricular volume (25).

As such, while diffuse inflammatory cells invade the myocardium in pressure overload, enhanced cardiomyocyte glucose utilization obscures the delineation of

- 75 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

inflammation using 18F-FDG. Of note, contractile function declined with rising 18F-FDG uptake, consistent with altered metabolic substrate utilization in progressive remodeling. The interplay between inflammation and fibrosis believed to contribute to the progression of pressure overload heart failure (26,27), could be more effectively evaluated by specific inflammation agents (28).

Some limitations to the present study should be considered. Firstly, as we aimed to characterize the acute development of inflammation in pressure overload, we did not assess the long-term functional outcome. We were primarily interested in the acute stage of disease, when inflammation is thought to predicate remodeling and fibrosis prior to terminal heart failure. The rapid change in ventricle geometry and decline in contractile function in parallel with inflammation and metabolic perturbation suggests we have captured active stage of disease. Secondly, as we used separate camera systems, we could not assess cardiac function and 18F-FDG distribution on the same day, such that correlations may be imprecise. Despite gradual increase in the ventricle volumes and mass, the ejection fraction remains relatively consistent from 2d to 8d, supporting the correlation between PET and MRI measurements.

Finally, we evaluated only K/X anesthesia as a means of suppressing cardiomyocyte 18F-FDG uptake. Earlier studies indicate that continuous administration of propranolol prevents glucose hypermetabolism (18). Because the sympathetic nervous system is also known to modulate ventricular remodeling and inflammatory leukocyte mobilization (29), and the temporal pattern of leukocyte infiltration is poorly characterized, we focused specifically on K/X suppression.

In conclusion, TAC induces rapid changes in left ventricle geometry and contractile function with a parallel modest increase in acute inflammatory cell infiltration. 18F-FDG imaging cannot effectively resolve inflammatory cells in the presence of extensive

- 76 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

cardiomyocyte metabolic remodeling, as in pressure-overload heart failure. A more selective molecular tracer targeting inflammation may provide a clearer illustration of the local inflammatory response and its contribution to pathogenesis.

- 77 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

ACKNOWLEDGMENTS

This study was partly supported by the German Research Foundation (DFG, Clinical

Research Group KFO311, Excellence Cluster REBIRTH-2, and research grant TH2161/1-1).

The authors thank the Preclinical Molecular Imaging and the Small Animal MRI Centre for their technical assistance with these experiments.

Disclosure

No potential conflicts of interest relevant to this article exist.

KEY POINTS

QUESTION: Can 18F-FDG imaging identify inflammatory leukocyte infiltration of the myocardium early in pressure overload-induced heart failure?

PERTINENT FINDINGS: Cardiac 18F-FDG uptake is elevated after transverse aortic constriction, but is disproportionate to the diffuse infiltration of inflammatory macrophages due to increased glucose metabolism by the overloaded heart.

IMPLICATIONS FOR PATIENT CARE: Assessment of inflammation with 18F-FDG may be imprecise in non-ischemic heart failure, necessitating more specific imaging agents.

- 78 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

FIGURE LEGENDS

Figure 1. (A) Representative cardiac 4-chamber view (4CV) and short axis (SA) images show elevated heart volume and dilatation of the left ventricle after TAC. (B) Quantitative functional analysis of left ventricular mass (upper left), end systolic volume (upper right), end diastolic volume (lower left) and ejection fraction (lower right).

Figure 2. (A) CD68 immunostaining of representative long axis sections at 3d and 8d after TAC and 3d after MI. (B) Quantitative analysis of stained sections shows elevation of CD68+ cells/field 3d and 8d after TAC (top) which inversely correlates to impaired ejection fraction at 8d after TAC (lower).

Figure 3. (A) Representative 18F-FDG short axis (SA) and horizontal long axis (HLA) images under effective ketamine-xylazine suppression of cardiomyocyte glucose uptake displays global myocardial inflammation at 3d and 7d after TAC. Elevated 18F-FDG uptake is localized to the infarct territory after MI. (B) Semi-quantitative assessment of 18F-FDG mean percent injected dose per gram (%ID/g).

Figure 4. (A) Blood glucose levels are comparably elevated under ketamine-xylazine anesthesia in TAC and sham mice. (B) Quantification of Western immunoblots display elevated GLUT4 expression in left ventricle (LV) but not right ventricle (RV) after cardiac pressure overload, and no change in insulin-independent GLUT1 at 8d after TAC.

Figure 5. Correlation of cardiac geometry and function with 18F-FDG accumulation in left ventricle at 3d after TAC, displaying inverse correlation to (A) ejection fraction, and direct correlation to (B) left ventricular end systolic volume, (C) end diastolic volume and (D) left ventricular mass.

- 79 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

REFERENCES

1. Savarese G, Lund LH. Global Public Health Burden of Heart Failure. Card Fail Rev. 2017;3:7-11.

2. Ponikowski P, Voors AA, Anker SD, et al. 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC)Developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J. 2016;37:2129-2200.

3. Bengel FM, George RT, Schuleri KH, Lardo AC, Wollert KC. Image-guided therapies for myocardial repair: concepts and practical implementation. Eur Heart J Cardiovasc Imaging. 2013;14:741-751.

4. Hofmann U, Frantz S. How can we cure a heart "in flame"? A translational view on inflammation in heart failure. Basic Res Cardiol. 2013;108:356.

5. Ammirati E, Cannistraci CV, Cristell NA, et al. Identification and predictive value of interleukin-6+ interleukin-10+ and interleukin-6- interleukin-10+ cytokine patterns in ST- elevation acute myocardial infarction. Circ Res. 2012;111:1336-1348.

6. Barron HV, Harr SD, Radford MJ, Wang Y, Krumholz HM. The association between white blood cell count and acute myocardial infarction mortality in patients > or =65 years of age: findings from the cooperative cardiovascular project. J Am Coll Cardiol. 2001;38:1654- 1661.

7. Bozkurt B, Torre-Amione G, Warren MS, et al. Results of targeted anti-tumor necrosis factor therapy with etanercept (ENBREL) in patients with advanced heart failure. Circulation. 2001;103:1044-1047.

8. Rischpler C, Dirschinger RJ, Nekolla SG, et al. Prospective Evaluation of 18F- Fluorodeoxyglucose Uptake in Postischemic Myocardium by Simultaneous Positron Emission Tomography/Magnetic Resonance Imaging as a Prognostic Marker of Functional Outcome. Circ Cardiovasc Imaging. 2016;9:e004316.

9. Brenes-Castro D, Castillo EC, Vazquez-Garza E, Torre-Amione G, Garcia-Rivas G. Temporal Frame of Immune Cell Infiltration during Heart Failure Establishment: Lessons from Animal Models. Int J Mol Sci. 2018;19.

10. Patel B, Ismahil MA, Hamid T, Bansal SS, Prabhu SD. Mononuclear Phagocytes Are Dispensable for Cardiac Remodeling in Established Pressure-Overload Heart Failure. PLoS One. 2017;12:e0170781.

- 80 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

11. Weisheit C, Zhang Y, Faron A, et al. Ly6C(low) and not Ly6C(high) macrophages accumulate first in the heart in a model of murine pressure-overload. PLoS One. 2014;9:e112710.

12. Xia Y, Lee K, Li N, Corbett D, Mendoza L, Frangogiannis NG. Characterization of the inflammatory and fibrotic response in a mouse model of cardiac pressure overload. Histochem Cell Biol. 2009;131:471-481.

13. Chen B, Frangogiannis NG. The Role of Macrophages in Nonischemic Heart Failure. JACC Basic Transl Sci. 2018;3:245-248.

14. Patel B, Bansal SS, Ismahil MA, et al. CCR2(+) Monocyte-Derived Infiltrating Macrophages Are Required for Adverse Cardiac Remodeling During Pressure Overload. JACC Basic Transl Sci. 2018;3:230-244.

15. Satomi T, Ogawa M, Mori I, et al. Comparison of contrast agents for atherosclerosis imaging using cultured macrophages: FDG versus ultrasmall superparamagnetic iron oxide. J Nucl Med. 2013;54:999-1004.

16. Young LH, Russell RR, 3rd, Yin R, et al. Regulation of myocardial glucose uptake and transport during ischemia and energetic stress. Am J Cardiol. 1999;83:25H-30H.

17. Sen S, Kundu BK, Wu HC, et al. Glucose regulation of load-induced mTOR signaling and ER stress in mammalian heart. J Am Heart Assoc. 2013;2:e004796.

18. Zhong M, Alonso CE, Taegtmeyer H, Kundu BK. Quantitative PET imaging detects early metabolic remodeling in a mouse model of pressure-overload left ventricular hypertrophy in vivo. J Nucl Med. 2013;54:609-615.

19. Thackeray JT, Bengel FM. Molecular Imaging of Myocardial Inflammation With Positron Emission Tomography Post-Ischemia: A Determinant of Subsequent Remodeling or Recovery. JACC Cardiovasc Imaging. 2018;11:1340-1355.

20. Tarnavski O, McMullen JR, Schinke M, Nie Q, Kong S, Izumo S. Mouse cardiac surgery: comprehensive techniques for the generation of mouse models of human diseases and their application for genomic studies. Physiol Genomics. 2004;16:349-360.

21. Thackeray JT, Bankstahl JP, Wang Y, Wollert KC, Bengel FM. Clinically relevant strategies for lowering cardiomyocyte glucose uptake for 18F-FDG imaging of myocardial inflammation in mice. Eur J Nucl Med Mol Imaging. 2015;42:771-780.

22. Zuo Z, Subgang A, Abaei A, et al. Assessment of Longitudinal Reproducibility of Mice LV Function Parameters at 11.7 T Derived from Self-Gated CINE MRI. Biomed Res Int. 2017;2017:8392952.

23. Frangogiannis NG, Smith CW, Entman ML. The inflammatory response in myocardial infarction. Cardiovasc Res. 2002;53:31-47.

- 81 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

24. Lee WW, Marinelli B, van der Laan AM, et al. PET/MRI of inflammation in myocardial infarction. J Am Coll Cardiol. 2012;59:153-163.

25. Todica A, Beetz NL, Gunther L, et al. Monitoring of Cardiac Remodeling in a Mouse Model of Pressure-Overload Left Ventricular Hypertrophy with [(18)F]FDG MicroPET. Mol Imaging Biol. 2018;20:268-274.

26. Kallikourdis M, Martini E, Carullo P, et al. T cell costimulation blockade blunts pressure overload-induced heart failure. Nat Commun. 2017;8:14680.

27. Patel VB, Bodiga S, Fan D, et al. Cardioprotective effects mediated by angiotensin II type 1 receptor blockade and enhancing angiotensin 1-7 in experimental heart failure in angiotensin-converting enzyme 2-null mice. Hypertension. 2012;59:1195-1203.

28. Thackeray JT, Derlin T, Haghikia A, et al. Molecular Imaging of the Chemokine Receptor CXCR4 After Acute Myocardial Infarction. JACC Cardiovasc Imaging. 2015;8:1417-1426.

29. Dutta P, Courties G, Wei Y, et al. Myocardial infarction accelerates atherosclerosis. Nature. 2012;487:325-329.

- 82 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Figure 1

- 83 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Figure 2

- 84 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Figure 3

- 85 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Figure 4

Figure 5

- 86 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Methods

Cardiac MRI Acquisition. Mice (TAC n=10, sham n=4) underwent MRI at 2d and 8d after surgery using a 7 Tesla small animal MRI system (Pharmascan 70/16, Bruker BioSpin

GmbH, Ettlingen, Germany; Software ParaVision 6.0.1). Cardiac function and geometry were evaluated by using a navigator-based self-gated cine MRI sequence (IntraGate FLASH,

Bruker Biospin, Ettlingen, Germany) as described previously (15). A 72-mm-diameter volume transmit coil (T11070 89/72 Quad to AD, Bruker BioSpin MRI GmbH) was used in combination with an anatomically shaped four-element mouse cardiac phased-array surface receive coil (T20027V3, Bruker BioSpin MRI GmbH). The animals were positioned prone and body temperature was maintained by a water-heated system during examination. All scans were performed in isoflurane inhalation anesthesia (3% induction, 2% maintenance).

Respiration rate was monitored and kept between 50-60 breaths per minute.

Cine images were acquired in 4 chamber view to visualize the aortic banding and short-axis

(SA) orientation that covered the whole left ventricle (LV). Sequence parameters (SA) were: field of view (FOV) 25×25mm, matrix 196×196, slice thickness 0.9mm, slice distance

0.9mm, repetition time (TR) 84.64ms, echo time (TE) 1.84ms, flip angle 45°, number of reconstructed phases 24.

MRI Image Analysis. Image analysis was performed with CVI (Vers. 5.3.6 Circle

Cardiovascular Imaging, Calgary, Alberta) to determine ventricle volumes and assess heart function. Endocardial and epicardial contours were defined on short axis sections and used to calculate ventricular mass at end systole and end diastole; ejection fraction was calculated from endocardial volumes.

- 87 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

PET image reconstruction and analysis. All PET images were reconstructed to a

128 × 128 × 159 image matrix (0.78 × 0.78 × 0.80 mm) using an OSEM3D/MAP algorithm

(β = 0.01, 2 OSEM iterations in 16 subsets, 18 MAP iterations) with scanner-applied decay- and scatter-correction. A 57Co transmission scan was used for attenuation correction (14).

Western Immunoblotting. Hearts were excised and dissected to separate the left and right ventricle, then flash frozen under liquid nitrogen. Tissue was stored at -80°C until use.

Western immunoblotting was performed as described previously (17). Briefly, tissue samples were homogenized in lysis buffer separated by gel electrophoresis. Primary antibodies were mouse monoclonal anti-glucose transporter GLUT1 (SPM498, 1/5000), mouse monoclonal anti-glucose transporter GLUT4 (ab65267, 1µg/ml) (Abcam). Secondary antibody was rabbit polyclonal anti-mouse IgG conjugated to horseradish peroxidase (Abcam). Immunoblots were analysed by chemiluminescence and integrated density values normalized to β-actin loading control were compared between groups.

- 88 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Figure Captions

Supplemental Figure 1. Immunohistologic Ly6G+ staining of representative long axis sections demonstrates no change in Ly6G-positive cells in left ventricular myocardium 8d after TAC compared to sham.

Supplemental Figure 2. Ketamine-xylazine (K/X) anesthesia suppresses cardiomyocyte glucose uptake effectively in healthy control and sham mice

(A) Robust 18F-FDG uptake in healthy myocardium under isoflurane anesthesia without K/X suppression. K/X suppression of cardiomyocyte 18F-FDG uptake lowered tracer uptake in healthy control mice to background levels, comparable to sham-operated animals. (B)

Quantitative 18F-FDG signal shows elevation in healthy mice under isoflurane anesthesia

(31.4±7.7 %ID/g, n=6). Signal under K/X anesthesia is comparable in healthy control mice

(3.8±1.5 %ID/g, n=5) and sham (3.4±1.0 %ID/g, n=8).

Supplemental Figure 3. Cardiac functional parameter are not related to elevated 18F-FDG signal at 8d after surgery (A) Correlation of stroke volume (µL) at d2 and d8 to 18F-FDG signal. (B) Elevated left ventricular end systolic volumes (µL), end diastolic volumes (µL), left ventricular mass and ejection fraction (%) at 8d do not correlate to 18F-FDG signal at 7d.

Supplemental Figure 4. Ineffective K/X suppression of cardiomyocyte glucose uptake visualizes altered cardiac glucose metabolism in response to cardiac pressure overload

Diffuse 18F-FDG signal in SA and HLA images under effective K/X suppression of cardiomyocyte glucose uptake reflects global myocardial inflammation, whereas robust 18F-

FDG signal in the LV shows insufficient suppression of cardiomyocyte glucose uptake.

- 89 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Figure 1

- 90 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Figure 2

- 91 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Figure 3

- 92 - Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure

Supplemental Figure 4

- 93 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

8 Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Aylina Glasenapp,1,2 Katja Derlin,2 Marcel Gutberlet,2 Annika Hess,1 Laura B N Langer,1 Hans-Jürgen Wester,3 Frank M Bengel,1 James T Thackeray1

Departments of 1Nuclear Medicine and 2Experimental Radiology, Hannover Medical School, Hannover, Germany 3Radiopharmaceutical Chemistry, Technical University of Munich, Munich, Germany

Manuscript prepared for submission

Short title: Imaging biomarkers detect inflammation and fibrosis

AG conducted experiments, performed surgeries, analyzed and interpreted data, wrote manuscript KD designed study, revised manuscript MG programmed T1 mapping sequence, analyzed T1 maps, revised manuscript AH conducted experiments LBNL produced radiopharmaceutical H-JW invented radiopharmaceutical FMB designed study, interpreted data, revised manuscript JTT designed study, conducted experiments, interpreted data, revised manuscript

First Author Aylina Glasenapp, Department of Radiology, Hannover Medical School, Carl Neuberg Str 1, 30625 Hannover, Germany

Corresponding Author James T Thackeray, PhD, Department of Nuclear Medicine Hannover Medical School, Carl Neuberg Str 1, 30625 Hannover, Germany [email protected]

- 94 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Abstract

Inflammation and fibrosis contribute to ventricular remodeling, but the precise contribution and timecourse of these processes are not fully elucidated in non-ischemic heart failure. Positron emission tomography imaging of chemokine CXC-motif receptor-4 (CXCR4) identified diffuse inflammatory cell infiltration of the left ventricle beginning from 7d after transverse aortic constriction (TAC) in mice, receding over 6 weeks to sham levels. Cardiac magnetic resonance imaging revealed prolonged T1 relaxation time at 7d and 3 weeks after TAC, correlating with rising interstitial fibrosis. The imaging markers reflected concurrent and progressive ventricle dilatation and predicted chronic contractile function. Alleviation of pressure overload by release of the aortic banding restored cardiac function and geometry, with a parallel reduction of T1 relaxation and fibrosis and delayed attenuation of inflammatory PET signal. These findings indicate the interrelationship between inflammation and fibrosis in pressure overload heart failure and provide a foundation for non-invasive monitoring of novel heart failure therapies.

- 95 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Introduction. Heart failure (HF) is a leading cause of death worldwide with an increasing prevalence in the aging population (1). In ischemic heart disease, excessive inflammation after myocardial infarction is a critical mediator of adverse left ventricular (LV) remodeling and functional impairment in the progression of HF (2), but the role of inflammation to cardiac remodeling and dysfunction in non-ischemic HF is less well characterized. Increasing evidence indicates a contribution of the inflammatory response to the activation and transdifferentiation of cardiac myofibroblasts in the progression of HF (3-6). Monocytes and macrophages are thought to play a key role in adverse cardiac remodeling (7,8). Recent studies indicate that cardiac macrophages are required for fibroblast activation, and may even adopt a fibroblast-like phenotype under certain conditions, such as ischemia (9). But the precise role of macrophages for the interstitial fibrosis in non-ischemic HF warrants further investigation. Current HF therapy relies on blockbuster drugs, focused on management of symptoms (10). Broadly targeted therapies such as non-steroidal anti-inflammatory drugs have borne limited success in management of acute myocardial infarction and HF (11), owing in part to the dynamic process involving multiple cell types. Notably, therapies against fibrosis have not been marketed (12), but novel and repurposed anti-inflammatory therapies canakinumab against interleukin-1β and colchicine have shown promise to improve outcomes early after insult. For advanced HF, ventricular unloading and device implantation may attenuate or reverse remodeling (13-15), but the molecular mechanisms underlying this improvement are not fully elucidated (16). Consequently, evaluating the temporal dynamics of inflammation and fibrosis may provide insights into disease progression and response to therapy. Molecular imaging facilitates longitudinal assessment of inflammation using positron emission tomography (PET) to visualize chemokine receptor CXC motif receptor 4 (CXCR4). Therefore, we evaluated temporal and spatial inflammation and fibrosis in acute and chronic stages following non-ischemic pressure overload-induced HF in a mouse model of transverse aortic constriction (TAC) and after mechanical ventricle unloading (reverse TAC) using positron emission tomography (PET) und cardiac magnetic resonance (CMR) imaging. Generally, 18F-FDG is a useful PET tracer to detect inflammation and macrophage metabolism (17). But in non-ischemic HF 18F-FDG does not provide sufficient information

- 96 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

about myocardial inflammation due to high glucose metabolism of cardiomyocytes and the unspecific nature of 18F-FDG (18). To detect a more specific inflammatory signal we were using 68Ga-pentixafor PET targeting chemokine receptor CXCR4 (19). Expressed on a variety of leukocytes and activated by the ligand CXCL12 (20), CXCR4 is upregulated in inflammatory diseases (21) and allows monitoring of the acute inflammatory cell infiltration after myocardial infarction in mice and patients (22). Therefore, 68Ga-pentixafor could be suitable do detect myocardial cell infiltration after pressure overload-induced HF as well. Parallel, CMR imaging was conducted to determine heart function, including T1-mapping as potential fibrosis biomarker to assess myocardial fibrosis and remodeling (23, 24). First, we hypothesized that multimodal imaging biomarker would identify the time course of inflammation and fibrosis after pressure overload-induced HF and predict late cardiac dysfunction. Second, we hypothesized that mechanical ventricular unloading results in functional improvement and attenuates molecular mechanisms of cardiac remodeling. The time course of molecular changes assessed by imaging was complemented by histological tissue workup and ex vivo measurements at corresponding time points.

- 97 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Methods

Animals. All animal experiments were approved by the state authority (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit) and local animal ethics review board. Male C57Bl/6N mice (n=132, 23.5 ± 1.7 g) were purchased from Charles River (Sulzfeld, Germany) and housed in groups in a temperature-controlled facility under a 14h/10h light/dark cycle with food and water freely available.

Study Design. Animals underwent serial multimodality imaging using CMR at baseline, 2d, 8d, 3wk and 6wk after surgery to determine heart function and geometry. To evaluate tissue characteristics, native T1-mapping was conducted at baseline, 7d, 3wk and 6wk after surgery. Parallel, PET imaging to assess myocardial inflammation was conducted using 68Ga- pentixafor targeting chemokine receptor CXCR4 at 3d, 7d, 3wk and 6wk after surgery followed by 18F-FDG PET for LV colocalization (Suppl. Fig.1). Further, the imaging protocol was also performed in a subgroup of animals 7d and 3wk post TAC or sham, as well as post reverse TAC, sham reverse TAC or second sham surgery, respectively (Suppl. Fig.2). Subgroups of animals were killed at the intermediate imaging timepoints 3d, 7d and 3wk for ex vivo tissue workup and correlative analysis. All other mice were killed by cervical dislocation at 6wk after TAC or sham surgery or rather 3wk after rTAC or sham rTAC for ex vivo tissue workup and correlative analysis.

Surgical Models.

Transverse aortic constriction (TAC). Mice underwent TAC (n=56) or sham (n=37) surgery under isoflurane anesthesia as described previously (25). Briefly, mice were pretreated with analgesic carprofen (5mg/kg s.c.) and anesthetized by isoflurane (induction 4% at 3L/min oxygen, maintained at 2% after oral intubation under mechanical ventilation) and 0.05mg/kg atropine sulfate was injected subcutaneously. A lateral thoracotomy in the second intercostal space was performed and the thymus gently retracted. The transverse aorta was constricted with a 6-0 silk suture between the brachiocephalic trunk and the left carotid artery To obtain heterogeneity in the severities of HF within the groups, the aorta was narrowed to either 0.5mm using a 25G needle or 0.4mm using a 27G needle and promptly removed after securing the suture. After constriction, the rib cage and skin were closed. For sham operation,

- 98 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

the identical procedure was followed without securing the suture. Following surgery analgesia was maintained orally by Tramadol 2.5mg/100ml drinking water.

Reverse TAC (rTAC). A subgroup of animals underwent rTAC (n=14), sham rTAC (n=6) or second sham (n=3) surgery. For this, the aortic banding was surgically removed 3weeks after TAC surgery. The surgical procedure and the treatment pre and post-surgery were identic as described in the TAC model and the silk suture was carefully removed.

Small animal cardiac magnetic resonance (CMR) imaging. Functional cardiac MR images were acquired using a 7 Tesla small animal MRI system (Pharmascan 70/16, Bruker BioSpin GmbH, Ettlingen, Germany; Software ParaVision 6.0.1). A 72-mm-diameter volume transmit coil (T11070 89/72 Quad to AD, Bruker BioSpin MRI GmbH) was used in combination with an anatomically shaped four-element mouse cardiac phased-array surface receive coil (T20027V3, Bruker BioSpin MRI GmbH). The animals were positioned prone and body temperature was maintained by a water-heated system during examination. All scans were performed under isoflurane inhalation anesthesia (3% induction, 1-2% maintenance). Respiration rate was monitored and maintained between 50-60 breaths per minute.

Functional CMR. Cardiac function and geometry were assessed using a navigator-based self- gated 2D spoiled gradient echo sequence (IntraGate FLASH, Bruker Biospin, Ettlingen, Germany) as described previously (26, 27). Cine images were acquired in 4 chamber view, in the plane of the aortic arch to visualize the aortic banding. A short-axis (SA) cine sequence covered the whole left ventricle (LV) to calculate functional parameter. Sequence parameters (SA) were: field of view (FOV) 25×25mm2, matrix size 196×196, slice thickness 0.9mm, slice distance 0.9mm, repetition time (TR) 84.64ms, echo time (TE) 1.84ms, flip angle 45°, number of reconstructed phases 24, acquisition time ~17 minutes.

T1 mapping. For T1 mapping, a Look-Locker Inversion-Recovery sequence with 2D radial k-space sampling was implemented (28,29). Sequence parameters were: FOV= 35 x 35mm2, matrix size= 128 x 128, slice thickness= 1mm, TR=3.3ms, TE=1.6ms, 1 slice, 2048 spokes, 20 repetitions, acquisition time ~4 minutes). Radial k-space data were corrected for eddy currents (30). After sorting the data according to its cardiac phase (five cardiac phases for the total cardiac cycle) and its inversion time after application of the inversion pulse (64 inversion times, ~ 7ms – 6600ms) using its DC signal (28), images were reconstructed using a

- 99 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

combination of parallel imaging and a total variation constraint for the inversion times and the cardiac phases (30).

CMR image analysis. Image analysis was performed with cvi42 (Version 5.3.6 Circle Cardiovascular Imaging, Calgary, Canada) to determine functional parameter, and assess heart morphology. Systolic and diastolic heart phase were selected out of 24 reconstructed phases and endocardial and epicardial contours were manually drawn into short axis slices covering base to apex including papillary muscles. Ventricle volume at end systole (ESV, µL) and end diastole (EDV, µL) calculated stroke volume (SV, µL) ejection fraction (EF, %) and left ventricular mass (LV mass, mg). T1 maps were calculated using a 3-parameter fit of a mono-exponential function (31). After manual segmentation of the myocardium using ImageJ (Vers.1.50e, National Institutes of Health, USA), mean T1 relaxation time was assessed for each mouse.

Small animal 68Ga-Pentixafor positron emission tomography (PET). Mice underwent longitudinal PET imaging protocol at 3d, 7d, 3wk, and 6wk using a dedicated small animal PET camera (Inveon DPET, Siemens, Knoxville, USA), as previously described (22). Mice which underwent TAC and reverse TAC surgery were imaged at 7d and 3wk after surgery, respectively. Briefly, anesthetized mice were positioned prone on the imaging bed (Minerve) with the heart centered in the scanner field of view. As a specific marker of inflammatory leukocytes, 68Ga-pentixafor (12.6±1.3 MBq) was injected intravenously as a 0.10- to 0.15-ml bolus via the tail vein. A 10-min static scan was acquired at 50 min after injection under isoflurane anesthesia. Then, 0.2ml 18F-FDG (19.9±1.8 MBq) was administered intraperitoneally under anesthesia for localization of the heart. After 20min under continued isoflurane, a second 10-min static scan was performed.

Radiosynthesis. 68Ga-Pentixafor was synthesized as previously described (22), using an automated module and CPCR4.2 precursor provided by Scintomics GmbH (Fürstenfeldbruck, Germany). 18F-FDG was synthesized using standard kits and production methods.

PET image reconstruction and analysis. All PET images were reconstructed to a 128 × 128 × 159 image matrix (0.78 × 0.78 × 0.80 mm) using an OSEM3D/MAP algorithm (β = 1, two OSEM iterations, 18 MAP iterations) with scanner-applied scatter correction. A 57Co transmission scan (10 min) was used for attenuation correction (22). Image analysis

- 100 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

was performed with Inveon Research Workplace 4.2 (Siemens Medical Solutions USA) to evaluate LV volume and signal. First 18F-FDG images were used to manually create regions of interest (ROIs) of LV myocardium by interactive thresholding as previously described (22). These ROIs were imported onto co-registered 68Ga-pentixafor images to calculate 68Ga- pentixafor uptake. The signal was semi-quantitatively analyzed as percent injected dose per gram of tissue (%ID/g).

Ex vivo Autoradiography and biodistribution. At 7d after surgery a subgroup of mice (TAC n=11, sham n= 5) was killed 60min after i.v. 0.15µL bolus injection of 11.7±0.8 MBq 68Ga-pentixafor for ex vivo validation. After cervical dislocation, blood samples were acquired by cardiac puncture at 60min post injection and hearts were immediately removed, rinsed in PBS and weighed. The atria and the right ventricle were removed and the left ventricle was cut in the long axis orientation. One half of the left ventricle was frozen under OCT compound (Tissue Tek) in the long axis for autoradiography. Adjacent sections were sliced into 10µm sections using a cryostat (LEICA CM3050 S) and air-dried. The screen (PerkinElmer) was exposed for 60min and digital autoradiography was conducted using a radiometric imager (Cyclone, PerkinElmer). Quantitative analysis of the autoradiograms was performed with PMOD (Version 3.7). For biodistribution, right ventricle, left ventricle, spleen, lung, liver, kidney, skeleton muscle and blood were excised and placed in pre- weighted tubes and counted with a known dilution of the injected dose for radioactivity in a gamma counter (PerkinElmer). Radioactivity counts were normalized by the injected radioactivity, animal weight and the weight of tissue. Results are shown as the percent injected dose per gram of tissue (%ID/g).

Histopathology for inflammation. Correlative measurements were conducted in tissue obtained from mice at the intermediate 3d (TAC n=4), 8d (TAC n=8) and 3wk (n=12) timepoint after surgery and at the endpoint of the experiments at 6wk after TAC (n=9), 3wk after rTAC (TAC n=14) and sham (n=14). Myocardial inflammatory cell infiltration was histologically evaluated. Mouse hearts were removed, cleaned in PBS, atria removed and frozen under OCT compound (Tissue Tek) in the long axis. Adjacent sections were sliced into 6µm sections using a cryostat (company) and thaw-mounted onto charged slides (SuperFrost Plus). Immunostaining for CD68+ cells identified macrophages. For quantification, 6 images

- 101 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

out of 1 section of each heart stained with CD68+ immunostaining were analyzed for CD68+ cells/field using ImageJ (Version 1.50e).

Histopathology for fibrosis. Myocardial fibrosis was histologically evaluated parallel to imaging time points on 8d (TAC n=4), 3wk (n=6) and 6wk (n=18) after TAC, 3wk after rTAC (n=14) and sham (n=11) surgery. Mouse hearts were removed, cleaned in PBS, atria removed and frozen under OCT compound (Tissue Tek) in the long axis. Adjacent sections were sliced into 6µm sections using a cryostat (company) and thaw-mounted onto charged slides (SuperFrost Plus). To confirm cardiac remodeling after TAC, Masson Trichrome & Picrosirius Red staining were conducted to identify interstitial collagen deposits in the myocardium. For quantification, 6 images out of 1 section of each heart stained with PSR were analyzed for fibrotic area (%) using ImageJ (Version 1.50e).

Statistics. Statistical analysis was carried out with Prism 6 (GraphPad, San Diego, California). All data are shown as mean ± SD. Serial imaging data obtained in individual animals were analyzed using Student’s paired t-test. TAC and sham groups were compared using the Welch t-test due to unequal variance. Pearson product-moment correlation coefficient was calculated to describe the relationship between pairs of continuous variables. P <0.05 was considered statistically significant.

Results TAC induced rapid and progressive ventricle enlargement and functional decline. CMR imaging demonstrated significant LV hypertrophy early after TAC surgery compared to sham animals, with progressive remodeling over 6wk (Fig. 1A). Ventricle hypertrophy corresponded to functional impairment, where EF was reduced by 29% compared to age- matched sham at 3d after TAC (Fig. 1B). EF declined slightly over the subsequent 6 weeks. Additionally, ventricular geometry displayed dilatation, with a 94% increase in systolic and 19% increase in diastolic volume at 1 week after TAC compared to sham (p<0.001). LV mass was elevated by 37% (p<0.001) after 1wk compared to sham, reflecting early remodeling and dilatation of the left ventricle with substantial rise over 6wk. TAC-induced hypertrophy was confirmed at necropsy (Suppl. Fig. 3). SV was significantly impaired in the acute phase by 19%, but demonstrated compensation at 6wk. By contrast, ventricular geometry remained uniform across the timecourse in sham operated animals (Suppl. Fig. 4).

- 102 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

TAC evoked acute and diffuse inflammatory cell infiltration of the left ventricle. 68Ga- Pentixafor PET images describe diffuse CXCR4 signal indicating myocardial inflammation after pressure-overload (Fig. 2A). Semi-quantitative analysis identified escalating 68Ga- pentixafor signal in TAC hearts compared to sham, elevated by 17% at 3d (p=0.02), 34% at 7d (p<0.001), and 18% at 3wk (p=0.02). The signal declined to sham levels at 6wk after surgery (Fig. 2B). Immunostaining for CD68+ identified significantly elevated macrophage content in the LV myocardium over a similar timecourse, with higher cell density at 3d, 8d (p=0.001), and 3wk (p<0.001) after injury compared to sham. Cell content was comparable at 6wk after surgery (Fig. 2C & 2D). CXCR4 68Ga-pentixafor signal correlated to CD68+ cell density in the myocardium throughout the timecourse (r=0.459, p=0.027) (Suppl. Fig. 5). The macrophage content at 8d after surgery correlated with contractile function at 8d (r=-0.710, p=0.003) (Suppl. Fig. 6). Ex vivo autoradiography confirmed globally higher myocardial 68Ga-pentixafor signal compared to sham at 7d (p=0.006) (Fig. 2E & 2F). To ensure that the in vivo signal was not influenced by ventricle hypertrophy and increased LV wall thickness, tissue was sampled for ex vivo biodistribution. While blood activity was significantly elevated at 7d after TAC compared to sham consistent with a systemic inflammatory response, the tracer retention in left ventricle was elevated twofold further, supporting the selective CXCR4 upregulation in the pressure overload myocardium (Suppl. Table 1).

Pressure overload induced progressive interstitial fibrosis detected in vivo by T1 mapping. Serial T1-maps display gradual increase in T1 relaxation time over 6 weeks after TAC (Fig. 3A, B). By contrast, sham mice displayed no change in T1 relaxation time over the same timeframe (Fig. 3B). Masson trichrome staining identified diffuse interstitial collagen in the left ventricle beginning from 8d after TAC (Fig. 3C & 3D). Fibrosis was quantified by picorosirius red staining and significantly elevated after TAC compared to sham, peaking on 8d and demonstrating slightly regression over 6 weeks (Fig. 3D & 3E). The mean myocardial T1 relaxation time correlated directly with the fibrotic area on histology (Suppl. Fig. 7), suggesting that T1 mapping may estimate cardiac fibrosis.

Imaging biomarkers predict subsequent decline in cardiac function. To determine the significance of the early imaging markers on function, we evaluated the capability of early imaging measurements to predict subsequent function. Early 68Ga-pentixafor signal at 7d correlated to systolic volume at 8d (r=0.609 p<0.001), 3wk (r=0.576 p<0.001) and 6wk

- 103 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

(r=0.817 p<0.001) (Fig. 4A-4C) Hence, the degree of inflammatory leukocyte infiltration appeared to predict severity of subsequent LV remodeling. Importantly, the 7d 68Ga- pentixafor signal correlated inversely with EF on 8d (r=-0.573 p<0.001), 3wk (r=-0.567 p<0.01) and 6wk (r=-0.860 p<0.001) (Fig.4A-4C). T1 relaxation time correlated directly with ESV (r=0.798 p<0.001) and inversely with EF (r=-0.795 p<0.001) over the timecourse after TAC (Fig. 4D). There was a direct correlation between 68Ga-pentixafor and T1 relaxation time (r=0.756 p<0.001) (Fig. 4E) reflecting the interaction between inflammation and fibrosis.

Mechanical ventricular unloading restored contractile function and was associated with attenuated inflammation and fibrosis. To determine the ability of inflammation and fibrosis molecular imaging to monitor therapeutic response, mechanical unloading was performed by removal of the aortic banding after 3 weeks. Debanding resulted in marked reduction of ESV and recovery of EF to sham values (Fig. 5). Serial imaging with 68Ga-pentixafor demonstrated regression of CXCR4 PET signal at 7d (p=0.005) and 3 weeks (p=0.016) after rTAC compared to pre-debanding (Fig. 6A & 6B). By contrast, sham rTAC did not reduce the CXCR4 signal. Histopathology demonstrated modestly lower CD68+ macrophage content in the LV at 3wk post rTAC compared to 3wk post TAC (Fig. 6C). Ventricle unloading also alleviated interstitial fibrosis. Debanding lowered T1 relaxation time compared to non-debanded TAC animals at 6 weeks after initial surgery (Fig. 7A & 7B). Histological examination confirmed 53% less myocardial collagen content at 3 weeks after debanding compared to age-matched TAC animals (Fig. 7C & 7D).

Discussion Inflammation and fibrosis are theorized to work in concert to promote adverse cardiac remodeling in cardiomyopathy, culminating in HF, but the timeframe of these processes has not been fully elucidated. Taking advantage of longitudinal multimodality molecular imaging, we characterized the inflammatory and fibrotic response in the myocardium following pressure overload-induced HF, revealing three major findings. First, macrophage infiltration was identified by CXCR4 PET imaging in the first week after aortic banding, with highest leukocyte content at 7d after pressure overload, which predicted worse subsequent ventricular dilatation and hypertrophy. Second, interstitial fibrosis was increased early, peaking on 8d after TAC in parallel with inflammation, and maintaining that level with slightly decrease

- 104 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

over time, as measured by elongated T1 relaxation time. Third, ventricle unloading not only improved contractile function, but reversed interstitial fibrosis and tended to lower chronic inflammation, which could be effectively monitored by PET and MR imaging biomarkers. Taken together, inflammation and fibrosis appear to contribute to ventricular remodeling in heart failure, and molecular imaging may allow precise monitoring of novel targeted therapeutic strategies to counter HF progression. While the prevalence of HF as the terminal manifestation of cardiovascular disease continues to rise, clinical management relies on well-established evidence-based therapies that treat chronic symptoms to support cardiac function. (10). The early pathogenic mechanisms underlying HF development and progression are not well understood and are generally undertreated. Basically, developing or established HF is pharmacologically treated with the main goal to attenuate symptoms and improve prognosis and survival. Recent evidence suggests that both inflammatory leukocytes, especially macrophages , play a major role in HF progression and adverse cardiac remodeling (32), as they contribute to interstitial fibrosis (9) and contractile dysfunction (8). In mouse models, anti-inflammatory therapy has demonstrated greater success. Blockade of T cell-mediated immunity in mice using rheumatoid arthritis drug abatacept attenuated cardiac hypertrophy secondary to pressure overload (4). Likewise, early antibody-mediated depletion of CCR2+ monocytes ameliorated late adverse cardiac remodeling 4 weeks after TAC (8). However, initial experience with anti- inflammatory therapies in the setting of non-ischemic HF generated mixed results for cardiac outcomes (33), which may relate to heterogeneous inflammatory response in the patient population. The variable inflammatory response can be difficult to detect in patients, due to reliance on plasma biomarkers that provide limited information on the local tissue microenvironment, especially early in the disease process (Hess et al 2020 JACC Img). The capacity to non-invasively monitor the inflammatory status of the heart in the progression of disease is therefore desirable. In the present study, CXCR4 PET signal is modestly elevated throughout the left ventricle within the first week of pressure overload, suggesting a role in the early stages of ventricular remodeling. (20,34). Notably, the CXCR4 PET signal directly corresponded with the CD68+ macrophage content in the left ventricle, supporting the validity of the imaging marker for local tissue inflammation. Likewise, in other inflammatory cardiovascular diseases such as atherosclerosis and myocardial ischemia CXCR4 PET signal

- 105 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

also correlated with leukocyte content in plaque and infarct territory, respectively (35,36). Importantly the severity of inflammatory cell infiltration is significantly lower than in myocardial infarction (18), but this approach still localized the signal to the left ventricle, as validated by ex vivo analysis. Moreover, the intensity of the signal in the first 7d after TAC surgery predicted subsequent systolic hypertrophy, whereby greater inflammation early in pressure overload corresponded to more severe remodeling. Indeed, the CXCR4 PET signal tended to be higher in animals with greater constriction of the transverse aorta. By extension, imaging-based delineation of high inflammatory activity may therefore predict adverse outcome, though this will require verification in clinical studies with greater heterogeneity. Reactive fibrosis in HF is likewise implicated in cardiac remodeling and dysfunction (37), but clinical assessment is often restricted to echocardiography measurements. Pressure overload HF in mice generates diffuse interstitial and perivascular collagen (38). New therapies to modulate fibrosis are being developed (12), such that non-invasive characterization of fibroblast activity and adverse interstitial fibrosis will be advantageous. T1 mapping on CMR identifies changes in tissue density and extracellular volume, and has been suggested to reflect increased fibrosis in heart (23,24,39). In the present study, we found prolonged T1 relaxation times as early as 7d after TAC, which was strikingly and rapidly reversed by debanding. The T1 relaxation time correlated with the 68Ga-pentixafor signal representing inflammation at early timepoints, but tended to persist later in HF with increased histology-defined collagen deposition. Thus, while the early signal may reflect a composite of edema and fibrosis, T1 mapping appears to identify the presence of interstitial collagen in the TAC model. The direct relationship with contractile function supports rapid compensatory remodeling in response to pressure overload. The rapid response to unloading is surprising, but suggests that the reactive fibrosis may be reversed by more directed therapy. A recent study demonstrated that immunotherapy using engineered T cells against fibroblast activation could prevent fibrosis in an angiotensin II infusion mouse model of cardiac fibrosis (40). Severe chronic HF may be treated by invasive therapeutic interventions including mechanical assist devices. The reverse TAC model rapidly relieved fibrosis in the left ventricle in mice. Similarly, HF patients with a supporting device exhibit attenuated adverse LV remodeling or even reversal (41). Although several mechanisms have been identified, the precise cellular processes required for ventricular recovery remain elusive (16). The removal of the aortic

- 106 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

constriction is a method to study the impact of hemodynamic unloading of the ventricle to cardiac remodeling (14). In our unloading study, we observed attenuated fibrosis, confirming previous studies which indicated regression of cardiomyocyte hypertrophy and interstitial fibrosis after removal of the aortic constriction such as functional recovery (13,14). Importantly, both inflammation and fibrosis followed a similar timecourse in pressure overload heart failure. Recent studies have identified the close association between inflammatory cells and fibroblast expansion. The failing heart initiates an inflammatory response including secretions of chemokines and cytokines to recruit leukocytes to the myocardium which directly influence fibrosis (32), including the transdifferentiation of myofibroblasts. Indeed, various components of the innate and adaptive immune system promote fibroblast activation and lead to myofibroblast activation through a myriad of cytokines. In fact, prolonged culture of human peripheral blood-derived macrophages resulted in the adoption of fibroblast-like morphology and expression of fibroblast markers including α-smooth muscle actin, P4H and collagen I (9). Moreover, mouse hearts after myocardial infarction exhibit fibroblasts that appear to derive from macrophage origin, based on lineage tracking studies (9). Interestingly, cardiac fibroblasts have also been described to augment inflammation, by release of inflammatory cytokines and chemokines (42). The parallel timecourse of 68Ga-pentixafor signal and prolonged T1 relaxation time observed in our study supports the notion of a complex interaction between inflammation and fibrosis. As such, early treatment of inflammation may impede interstitial fibrosis. Accordingly, inflammation and fibrosis are critically linked, suggesting that macrophages are able to transition to a fibroblast-like phenotype, therapeutic intervention during the shift from macrophages to fibroblasts suggests the possibility to modulate cardiac remodeling in HF (9). Some limitations in the present study should be considered. Firstly, we restricted our analysis to two acute and two chronic timepoints for logistical reasons, such that other intermediate changes in morphology, inflammation, and fibrosis are not intimately tracked. The timepoints were selected based on inflammation timecourse in flow cytometry assessment of TAC mice (44,45), but dedicated assessment of the very early timepoints may provide further insights into the cause-effect relationship with acute function. Secondly, we did not directly assess the specific binding of 68Ga-pentixafor in the left ventricle after pressure overload. Previous studies have identified lower affinity to the mouse CXCR4 isoform compared to human (19,

- 107 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

22), and blocking studies in mice have previously been demonstrated (46). We did not perform partial volume corrections, and the thickening of the ventricle wall may contribute to higher 68Ga-pentixafor signal observed at later timepoints. However, ex vivo analysis confirmed the signal increase in tissue, and the tracer accumulation declined between 3 and 6 weeks, despite continued increase in LV mass, As such, we do not believe these technical limitations influenced the interpretation of the PET data. Finally, the exact cell composition responsible for 68Ga-pentixafor signal has not been determined, but previous flow cytometry analysis identified elevated leukocyte subsets after TAC with mainly macrophages (45) and 68Ga-pentixafor uptake has been shown in multiple leukocyte subtypes (34). In this study, the tracer accumulation directly correlated with CD68 macrophage content in the left ventricle. In conclusion, inflammation and fibrosis develop early in the development of pressure overload-induced HF along a similar continuum and can be dynamically imaged by PET and CMR. The intensity of inflammation identified by CXCR4-targeted PET predicts subsequent ventricular dilatation and severity of HF and CMR T1 mapping identifies progressive fibrosis. Mechanical ventricular unloading rapidly rescues cardiac function and alleviates myocardial fibrosis, which is quantitatively detected by PET and CMR. These sensitive imaging approaches may provide useful insights into the progression of non-ischemic heart failure and non-invasively monitor the efficacy of novel targeted therapies to alleviate parallel inflammation and fibroblast activity.

- 108 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure Legends. Figure 1. CMR reveals acute and progressive cardiac dysfunction and geometric change after TAC (A) Representative morphologic CMR images in 4 chamber view in systole of baseline and TAC and position of aortic banding (arrow). (B) Quantitative assessment of cardiac functional parameter at baseline and longitudinal imaging timepoints after TAC: EF (%), end systolic volume (µL), end diastolic volume (µL), LV mass (mg), stroke volume (µL) at 2d, 8d, 3wk and 6wk. ***p < 0.001 **p < 0.01 * p < 0.05 NS = not significant

Figure 2. 68Ga-Pentixafor defines diffuse myocardial inflammation after pressure-overload in vivo and ex vivo (A) Representative images of myocardial 68Ga-pentixafor signal, 18F-FDG signal and both signals as fusion in sham and 3d, 7d, 3wk and 6wk after TAC. (B) Quantitative 68Ga-pentixafor signal (%ID/g) after TAC compared to sham at corresponding timepoints. (C) Immunhistological CD68+ cell staining detects elevation of myocardial macrophages. (D) Quantification of CD68+ cells/field confirms increased cell density up to 6 wk after TAC. (E) Ex vivo autoradiography shows myocardial 68Ga-pentixafor binding. (F) Quantification of autoradiography as optical density confirms myocardial 68Ga-pentixafor binding. ***p < 0.001 **p < 0.01 * p < 0.05 NS = not significant

Figure 3. Myocardial T1 relaxation time is elevated at all imaging timepoints after TAC (A) Representative T1 maps at baseline, 8d, 3wk and 6wk after TAC. (B) Mean T1 relaxation time is elevated on 8d, 3wk and 6wk after TAC compared to sham and baseline. (C) Overview of Masson trichrome stained heart sections of sham and 8d and 3wk after TAC demonstrating cardiomyocytes in red and collagen in blue (arrows). (D) Representative images of Masson trichrome and picrosirius red staining at 6wk after TAC and sham. (E) Quantification of % fibrotic area / field confirms collagen presence at 8d, 3wk, and 6wk after TAC compared to sham. ***p < 0.001 **p < 0.01 * p < 0.05 NS = not significant

Figure 4. Imaging biomarkers correlate to cardiac dysfunction and predict late functional outcome. (A) Correlation of 68Ga-pentixafor (%ID/g) at 7d to ESV (µL) and EF (%) at 8d (sham n=18, TAC n=48). (B) Correlation of 68Ga-pentixafor (%ID/g) at 7d to ESV (µL) and EF (%) at 3wk (sham n=11, TAC n=37). (C) Correlation of 68Ga-pentixafor (%ID/g) at 7d to ESV (µL) and EF (%) at 6wk (sham n=6, TAC n=8). (D) Correlation of mean T1 relaxation time (ms) to ESV (µL) and EF (%) at multiple timepoints (Baseline n=3, sham n=8, TAC

- 109 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

n=19). (E) Correlation of 68Ga-pentixafor (%ID/g) to mean T1 relaxation time (ms) at multiple timepoints (sham n=4, TAC n=19).

Figure 5. Mechanical ventricular unloading leads to functional recovery (A) Quantitative assessment of cardiac functional parameter at 3wk after TAC (n=20) and both 8d and 3wk after rTAC (n=14) show alleviation of cardiac volumes (ESV, EDV,µL) LV mass (LV mass, mg) and improvement of EF (%) compared to sham rTAC (n=6) and sham (n=3). (B) Stroke volume (µL) after rTAC is on sham level and elevated compared to TAC and sham rTAC. *** p<0.001 ** p<0.01 * p>0.05; NS = not significant; vs matched sham ††† p<0.001 †† p<0.01 † p<0.05; NS = not significant; vs matched sham reverse TAC

Figure 6. (A) Representative images of myocardial 68Ga-pentixafor signal, 18F-FDG signal and both signals as fusion in sham, 3wk after TAC and 8d after rTAC. (B) Quantitative 68Ga- pentixafor signal analysis (%ID/g) 7d and 3wk after TAC and sham; 7d and 3wk after rTAC and Sham rTAC at corresponding timepoints show attenuated signal following rTAC but not after sham rTAC (paired t test) (rTAC n=10, sham rTAC n=6).

Figure 7. T1 mapping detects attenuated T1 relaxation time after rTAC. (A) Representative T1 maps after rTAC. (B) Quantification of mean T1 relaxation time at 8d and 3wk after rTAC show alleviation compared to sham rTAC. (C) % Fibrotic area / field is alleviated after rTAC. (D) Representative picrosirius red staining images of diffuse interstitial collagen. *** p<0.001 ** p<0.01 * p>0.05; NS = not significant; vs matched sham

- 110 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

References

1. Ziaeian B, Fonarow GC. Epidemiology and aetiology of heart failure. Nat Rev Cardiol. 2016;13(6):368-78.

2. Rischpler C, Dirschinger RJ, Nekolla SG, Kossmann H, Nicolosi S, Hanus F, et al. Prospective Evaluation of 18F-Fluorodeoxyglucose Uptake in Postischemic Myocardium by Simultaneous Positron Emission Tomography/Magnetic Resonance Imaging as a Prognostic Marker of Functional Outcome. Circ Cardiovasc Imaging. 2016;9(4):e004316.

3. Bosch L, de Haan J, Seijkens T, van Tiel C, Brans M, Pasterkamp G, et al. Small molecule-mediated inhibition of CD40-TRAF6 reduces adverse cardiac remodelling in pressure overload induced heart failure. Int J Cardiol. 2019;279:141-4.

4. Kallikourdis M, Martini E, Carullo P, Sardi C, Roselli G, Greco CM, et al. T cell costimulation blockade blunts pressure overload-induced heart failure. Nat Commun. 2017;8:14680.

5. Nevers T, Salvador AM, Grodecki-Pena A, Knapp A, Velazquez F, Aronovitz M, et al. Left Ventricular T-Cell Recruitment Contributes to the Pathogenesis of Heart Failure. Circ Heart Fail. 2015;8(4):776-87.

6. Nevers T, Salvador AM, Velazquez F, Ngwenyama N, Carrillo-Salinas FJ, Aronovitz M, et al. Th1 effector T cells selectively orchestrate cardiac fibrosis in nonischemic heart failure. J Exp Med. 2017;214(11):3311-29.

7. Richards DA, Aronovitz MJ, Calamaras TD, Tam K, Martin GL, Liu P, et al. Distinct Phenotypes Induced by Three Degrees of Transverse Aortic Constriction in Mice. Sci Rep. 2019;9(1):5844.

8. Patel B, Bansal SS, Ismahil MA, Hamid T, Rokosh G, Mack M, et al. CCR2(+) Monocyte-Derived Infiltrating Macrophages Are Required for Adverse Cardiac Remodeling During Pressure Overload. JACC Basic Transl Sci. 2018;3(2):230-44.

9. Haider N, Bosca L, Zandbergen HR, Kovacic JC, Narula N, Gonzalez-Ramos S, et al. Transition of Macrophages to Fibroblast-Like Cells in Healing Myocardial Infarction. J Am Coll Cardiol. 2019;74(25):3124-35.

- 111 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

10. Ponikowski P, Voors AA, Anker SD, Bueno H, Cleland JG, Coats AJ, et al. 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail. 2016;18(8):891-975.

11. Frangogiannis NG. The immune system and cardiac repair. Pharmacol Res. 2008;58(2):88-111.

12. Travers JG, Kamal FA, Robbins J, Yutzey KE, Blaxall BC. Cardiac Fibrosis: The Fibroblast Awakens. Circ Res. 2016;118(6):1021-40.

13. Merino D, Gil A, Gomez J, Ruiz L, Llano M, Garcia R, et al. Experimental modelling of cardiac pressure overload hypertrophy: Modified technique for precise, reproducible, safe and easy aortic arch banding-debanding in mice. Sci Rep. 2018;8(1):3167.

14. Weinheimer CJ, Kovacs A, Evans S, Matkovich SJ, Barger PM, Mann DL. Load- Dependent Changes in Left Ventricular Structure and Function in a Pathophysiologically Relevant Murine Model of Reversible Heart Failure. Circ Heart Fail. 2018;11(5):e004351.

15. Birks EJ, Tansley PD, Hardy J, George RS, Bowles CT, Burke M, et al. Left ventricular assist device and drug therapy for the reversal of heart failure. N Engl J Med. 2006;355(18):1873-84.

16. Schirone L, Forte M, Palmerio S, Yee D, Nocella C, Angelini F, et al. A Review of the Molecular Mechanisms Underlying the Development and Progression of Cardiac Remodeling. Oxid Med Cell Longev. 2017;2017:3920195.

17. Wollenweber T, Roentgen P, Schafer A, Schatka I, Zwadlo C, Brunkhorst T, et al. Characterizing the inflammatory tissue response to acute myocardial infarction by clinical multimodality noninvasive imaging. Circ Cardiovasc Imaging. 2014;7(5):811-8.

18. Glasenapp A, Derlin K, Wang Y, Bankstahl M, Meier M, Wollert KC, et al. Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure. J Nucl Med. 2019.

19. Buck AK, Stolzenburg A, Hanscheid H, Schirbel A, Luckerath K, Schottelius M, et al. Chemokine receptor - Directed imaging and therapy. Methods. 2017;130:63-71.

- 112 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

20. Murdoch C. CXCR4: chemokine receptor extraordinaire. Immunol Rev. 2000;177:175-84.

21. Salvatore P, Pagliarulo C, Colicchio R, Napoli C. CXCR4-CXCL12-dependent inflammatory network and endothelial progenitors. Curr Med Chem. 2010;17(27):3019-29.

22. Thackeray JT, Derlin T, Haghikia A, Napp LC, Wang Y, Ross TL, et al. Molecular Imaging of the Chemokine Receptor CXCR4 After Acute Myocardial Infarction. JACC Cardiovasc Imaging. 2015;8(12):1417-26.

23. Stuckey DJ, McSweeney SJ, Thin MZ, Habib J, Price AN, Fiedler LR, et al. T(1) mapping detects pharmacological retardation of diffuse cardiac fibrosis in mouse pressure- overload hypertrophy. Circ Cardiovasc Imaging. 2014;7(2):240-9.

24. Gonzalez A, Schelbert EB, Diez J, Butler J. Myocardial Interstitial Fibrosis in Heart Failure: Biological and Translational Perspectives. J Am Coll Cardiol. 2018;71(15):1696-706.

25. Tarnavski O, McMullen JR, Schinke M, Nie Q, Kong S, Izumo S. Mouse cardiac surgery: comprehensive techniques for the generation of mouse models of human diseases and their application for genomic studies. Physiol Genomics. 2004;16(3):349-60.

26. Fries P, Massmann A, Seidel R, Muller A, Stroeder J, Custodis F, et al. Comparison of retrospectively self-gated and prospectively triggered FLASH sequences for cine imaging of the aorta in mice at 9.4 Tesla. Invest Radiol. 2012;47(4):259-66.

27. Zuo Z, Subgang A, Abaei A, Rottbauer W, Stiller D, Ma G, et al. Assessment of Longitudinal Reproducibility of Mice LV Function Parameters at 11.7 T Derived from Self- Gated CINE MRI. Biomed Res Int. 2017;2017:8392952.

28. Winter P, Kampf T, Helluy X, Gutjahr FT, Meyer CB, Bauer WR, et al. Self- navigation under non-steady-state conditions: Cardiac and respiratory self-gating of inversion recovery snapshot FLASH acquisitions in mice. Magn Reson Med. 2016;76(6):1887-94.

29. Wech T, Tran-Gia J, Rutzel F, Klink T, Bley TA, Kostler H. Single-shot late Gd enhancement imaging of myocardial infarction with retrospectively adjustable contrast and heart-phase. Magn Reson Imaging. 2018;47:48-53.

30. Moussavi A, Untenberger M, Uecker M, Frahm J. Correction of gradient-induced phase errors in radial MRI. Magn Reson Med. 2014;71(1):308-12.

- 113 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

31. Nekolla S, Gneiting T, Syha J, Deichmann R, Haase A. T1 maps by K-space reduced snapshot-FLASH MRI. J Comput Assist Tomogr. 1992;16(2):327-32.

32. Hofmann U, Frantz S. How can we cure a heart "in flame"? A translational view on inflammation in heart failure. Basic Res Cardiol. 2013;108(4):356.

33. Ayoub KF, Pothineni NVK, Rutland J, Ding Z, Mehta JL. Immunity, Inflammation, and Oxidative Stress in Heart Failure: Emerging Molecular Targets. Cardiovasc Drugs Ther. 2017;31(5-6):593-608.

34. Borchert T, Beitar L, Langer LBN, Polyak A, Wester HJ, Ross TL, et al. Dissecting the target leukocyte subpopulations of clinically relevant inflammation radiopharmaceuticals. J Nucl Cardiol. 2019.

35. Doring Y, Pawig L, Weber C, Noels H. The CXCL12/CXCR4 chemokine ligand/receptor axis in cardiovascular disease. Front Physiol. 2014;5:212.

36. Hyafil F, Pelisek J, Laitinen I, Schottelius M, Mohring M, Doring Y, et al. Imaging the Cytokine Receptor CXCR4 in Atherosclerotic Plaques with the Radiotracer (68)Ga- Pentixafor for PET. J Nucl Med. 2017;58(3):499-506.

37. Weber KT, Sun Y, Bhattacharya SK, Ahokas RA, Gerling IC. Myofibroblast- mediated mechanisms of pathological remodelling of the heart. Nat Rev Cardiol. 2013;10(1):15-26.

38. Bacmeister L, Schwarzl M, Warnke S, Stoffers B, Blankenberg S, Westermann D, et al. Inflammation and fibrosis in murine models of heart failure. Basic Res Cardiol. 2019;114(3):19.

39. Everett RJ, Stirrat CG, Semple SI, Newby DE, Dweck MR, Mirsadraee S. Assessment of myocardial fibrosis with T1 mapping MRI. Clin Radiol. 2016;71(8):768-78.

40. Aghajanian H, Kimura T, Rurik JG, Hancock AS, Leibowitz MS, Li L, et al. Targeting cardiac fibrosis with engineered T cells. Nature. 2019;573(7774):430-3.

41. Kass DA, Baughman KL, Pak PH, Cho PW, Levin HR, Gardner TJ, et al. Reverse remodeling from cardiomyoplasty in human heart failure. External constraint versus active assist. Circulation. 1995;91(9):2314-8.

- 114 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

42. Shinde AV, Frangogiannis NG. Fibroblasts in myocardial infarction: a role in inflammation and repair. J Mol Cell Cardiol. 2014;70:74-82.

43. Lee WW, Marinelli B, van der Laan AM, Sena BF, Gorbatov R, Leuschner F, et al. PET/MRI of inflammation in myocardial infarction. J Am Coll Cardiol. 2012;59(2):153-63.

44. Patel B, Ismahil MA, Hamid T, Bansal SS, Prabhu SD. Mononuclear Phagocytes Are Dispensable for Cardiac Remodeling in Established Pressure-Overload Heart Failure. PLoS One. 2017;12(1):e0170781.

45. Weisheit C, Zhang Y, Faron A, Kopke O, Weisheit G, Steinstrasser A, et al. Ly6C(low) and not Ly6C(high) macrophages accumulate first in the heart in a model of murine pressure-overload. PLoS One. 2014;9(11):e112710.

46. Gourni E, Demmer O, Schottelius M, D'Alessandria C, Schulz S, Dijkgraaf I, et al. PET of CXCR4 expression by a (68)Ga-labeled highly specific targeted contrast agent. J Nucl Med. 2011;52(11):1803-10.

- 115 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 1

- 116 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 2

- 117 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 3

- 118 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 4

- 119 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 5

- 120 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 6

- 121 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Figure 7

- 122 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 1. Study design longitudinal imaging protocol after TAC/sham surgery. CMR was performed at baseline, 2d, 8d, 3wk and 6wk after surgery. PET was conducted at 3d, 7d, 3wk and 6wk after surgery. Tissue samples for histology were taken at 3d, 8d, 3wk and 6wk after surgery. Autoradiography and biodistribution measurements were performed at 7d after surgery.

Supplemental Figure 2. Study design longitudinal imaging protocol prior to and after rTAC / sham rTAC / sham surgery. CMR was performed at 8d and 3wk after each surgery. PET was conducted at 7d and 3wk after each surgery. rTAC / sham rTAC / sham surgery was performed 3wk after TAC /sham. Tissue samples were taken at 8d and 3wk after surgery, respectively.

Supplemental Figure 3. Heart weight (mg) after sham and TAC, grouped (sham n=37, TAC n=56); Heart weight (mg) after sham (n=37) and TAC with 25G (n=46) and 27G (n=10); Heart weight (mg) after sham (n=37) and TAC at 3d (n=4), 8d (n=15), 3wk (n=21) and 6wk (n=16); Heart weight (mg) 3 weeks after rTAC (n=14) and sham rTAC (n=6); Heart weight / body weight ratio after sham and TAC, grouped Lung weight (mg) after sham and TAC, grouped(sham n=37, TAC n=56).

Supplemental Figure 4. Quantitative assessment of cardiac functional parameter at baseline compared to 8d, 3wk and 6wk after sham show no statistical significant differences.

Supplemental Figure 5. Correlation of 68Ga-pentixafor signal (%ID/g) to CD68+ cells/field at multiple timepoints (sham n=8 TAC n=15).

Supplemental Figure 6. Correlation of CD68+ cells/field to EF on 8d (sham n=8 TAC n=7) and grouped for all imaging matched timepoints (sham n=8 TAC n=29).

Supplemental Figure 7. Correlation of T1 relaxation time (ms) to fibrotic area (%) (sham n=7, TAC n=12) and 68Ga-pentixafor signal (%ID/g) to fibrotic area (%) (sham n=5, TAC n=15) at matched multiple timepoints.

Supplemental Figure 8. Correlation of T1 relaxation time (ms) to fibrotic area (%) and 68Ga- pentixafor signal (%ID/g) to fibrotic area (%). Correlation of fibrotic area (%) to ESV (µL), EF (%) and LV mass (mg) at matched multiple timepoints.

- 123 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 9. Correlation of 68Ga-pentixafor signal (%ID/g) to T1 relaxation time after rTAC (n=17), Correlation of 68Ga-pentixafor signal (%ID/g) and T1 relaxation time (ms) to fibrotic area (%) after rTAC (n=10), sham rTAC (n=6) and sham (n=6) at multiple timepoints.

Supplemental Table 1 Ex vivo biodistribution measurements (%ID/g) 7d after surgery (sham n=5, TAC n=11).

- 124 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 1

Supplemental Figure 2

- 125 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 3

- 126 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 4

- 127 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 5

Supplemental Figure 6

- 128 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 7

- 129 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 8

- 130 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Figure 9

- 131 - Multimodality Imaging of Inflammation and Fibrosis during Heart Failure Progression and Reverse Remodeling

Supplemental Table 1

Tissue Sham+7d (%ID/g) (n=5) TAC+7d (%ID/g) (n=11) %

Right Ventricle 0.448±0.055 0.566±0.105 +26 Left Ventricle 0.382±0.037 0.589±0.170 +54 Spleen 6.335±2.771 6.181±2.970 -2 Lung 1.831±0.368 2.503±0.679 +37 Liver 2.674±0.453 3.351±0.839 +25 Kidney 3.983±0.843 5.006±1.698 +26 Skeletal Muscle 0.252±0.051 0.316±0.072 +25 Blood 1.262±0.175 1.603±0.362 +27

- 132 - Discussion

9 Discussion

9.1 Key Findings

Inflammation and fibrosis play key roles in numerous diseases and are known to contribute critically in the development and progression of HF. In this thesis, we identified molecular mechanisms of myocardial inflammation and fibrosis in the acute and chronic phase after pressure overload HF and mechanical ventricular unloading using PET and CMR imaging. First, we found that 18F-FDG uptake is elevated in the acute stage after pressure overload. The early signal correlated to cardiac dysfunction. The robust 18F-FDG signal one week after TAC was disproportionate to severity of myocardial inflammation on histology, and relative to the extensive inflammation mice after MI. This elevated 18F-FDG signal reflected a composite of both, early pro-inflammatory macrophage infiltration and altered glucose metabolism of cardiomyocytes, as shown by histology and immunoblotting. Accordingly, 18F-FDG poorly delineates the inflammatory cell signal, as suppression of cardiomyocyte glucose uptake by K/X anesthesia in stressed and metabolically altered myocardium is ineffective. By contrast, chemokine receptor CXCR4 upregulation identified by 68Ga-pentixafor PET, better reflected diffuse inflammatory leukocyte infiltration in the myocardium after pressure overload HF, and we established that the signal intensity reflecting myocardial macrophage infiltration at seven days post-TAC predicted severity of late functional remodeling. Accordingly, 68Ga-pentixafor as ligand for chemokine receptor CXCR4 may provide insights into novel therapeutic approaches. Furthermore, we detected prolonged myocardial T1 relaxation time after pressure overload HF by CMR T1 mapping, reflecting diffuse interstitial and perivascular myocardial fibrosis. We found that extent of myocardial collagen related to functional impairment. Finally, we demonstrated that surgical mechanical unloading of the left ventricle by removal of the aortic banding resulted in partial recovery of cardiac function and reverse remodeling. Moreover, inflammation and fibrosis are attenuated after unloading, which could be detected and monitored by multimodal imaging. This supports the feasibility of targeted imaging of molecular inflammatory and fibrotic mechanisms to monitor therapeutic responses. We demonstrated that inflammation and fibrosis occur early in the pathogenesis of pressure overload HF and follow a parallel time course, contributing to ventricular remodeling. This

- 133 - Discussion

underlines their complex interactions and involvement in HF progression. Based on these results, we may define time points for therapeutic interventions targeting one or both early pathomechanisms, which may have complementary benefit.

9.2 Heart Failure Animal Models

Transverse aortic constriction is a well-established model to induce non-ischemic HF by cardiac pressure overload (ROCKMAN et al. 1991). Nevertheless, the description of functional outcome and pathophysiological mechanisms in response to pressure overload vary greatly in the literature, especially in the early stages of disease. In our experiments, we observed a rapid functional decline, as the EF was already significantly impaired after two days and ventricular volumes and mass rise. In contrast, some studies describe a sustained heart function within the first weeks after TAC with functional decline only later in progression (RICHARDS et al. 2019). The most commonly used method to measure cardiac function in mice after TAC is echocardiography. Functional measurements are strongly dependent on the method used to obtain them, and ultrasound calculations exhibit higher variance than MRI. Moreover, echocardiography is highly operator dependent, and calculations can vary greatly between research groups and investigators. Differences in functional outcome may also relate to mortality after the surgical procedure. If animals with severely impaired heart function die short-term after surgery, the examination group is dominated by animals with moderate HF, which adapt successfully to the increased pressure. To overcome this effect and to create heterogeneous groups representative of the patient population, we employed two severities of aortic constriction. Animals with a 27G stenosis, narrowed to ~40% of the original aortic diameter, exhibited significantly worse late functional outcome (min.18%, max. 58%; 34±13% EF) and more advanced cardiac hypertrophy at three weeks and later. Acute post-surgical mortality was also higher. However, we observed that even assessing solely 'mild' TAC with 25G stenosis, narrowed to ~50% of the original aortic diameter, resulted in a high range of functional outcome (min. 14%, max. 66%; 46±10% EF), as some animals showed only mild and some severe functional impairment. This could mean that not only the needle diameter contributes to severity, but also the applied tension on the suture while tightening. All mice were operated at the age of ten weeks, but we cannot

- 134 - Discussion

exclude biological variability within the strain from the supplier, or slight fluctuation in the age at surgery. These variations could result in differences of functional response to TAC, as younger mice with ongoing heart growth might react more strongly to pressure overload than mice with mature myocardium. A recent study showed that intra- and inter-surgeon reproducibility as measured by coefficient of variation of heart weights and echocardiography can be optimized by using an o-ring with a fixed inner diameter and internal suture for aortic banding (MELLEBY et al. 2018). Notably, this study used only modest constrictions, the narrowest was 0.61mm, resulting in early and rapid functional decline as measured by CMR EF, whereas o-rings with 0.66 and 0.71mm only exhibited significant impaired EF after eight weeks or later. The 25G constriction, when fully tightened, narrows the transverse aortic diameter to 0.5mm, or 20% greater than the narrowest o-ring. However, the o-ring method induces smaller variations in hypertrophic response compared to conventional TAC, which may be of benefit for treatment studies. Other factors contributing to heterogeneity may include fat tissue and fascia surrounding the aorta, which could influence the diameter and functional outcome. In our experiments, we found that EF was rapidly impaired and remained at a consistently reduced level over six weeks, whereas ventricle volumes and mass continued to rise. Removal of aortic banding to allow mechanical unloading also directly affected ventricle geometry and heart function. Ventricle volumes and mass decreased, and EF partially recovered within one week of debanding. The degree of recovery was dependent on the initial cardiac impairment. Animals with moderately impaired heart function with EF>40% fully recovered to EF comparable with healthy animals. By contrast, animals with EF<40% only recovered partially. Accordingly, not only the severity but also the time point of the removal of the aortic banding is essential for functional recovery. Removal of the aortic banding after eight weeks of TAC was less successful to regain function than earlier intervention (i.e. at four weeks) (GAO et al. 2005a). This observation suggests a threshold of damage and remodeling from which recovery is more limited. Human HF patients are susceptible to repeat cardiac events, such as non-ischemic HF compounded with infarction, or vice versa (YANCY et al. 2013; MARINO et al. 2019). To model this phenomenon, a mouse model combining pressure overload and ischemic injury was recently developed (WEINHEIMER et al. 2018). Mice underwent both interventions in a

- 135 - Discussion

dual surgical procedure, first TAC using a 26G cannula, followed by a small apical infarct by distal left anterior coronary ligation. These animals were studied with and without aortic debanding two weeks after TAC + MI, to mimic HF comparable to patients with a reduced EF treated with drug and/or device therapies. Although heart function was impaired by the ischemic injury, hemodynamic unloading resulted in LV structure and geometry normalization with partial functional recovery, suggesting that pathological mechanisms in chronic HF might be reversible despite ischemic injury. Inflammatory and fibrotic gene pathways were measured by Ingenuity Pathway Analysis to define recovered and persistently dysregulated HF genes. Most normalized genes were allocated to contractile function and metabolism, but a notable contingent was related to fibrosis and inflammation. Among persistently dysregulated genes, many of the inflammation and fibrosis gene pathways were detected, though not as many were identified as compared to normalized HF genes, suggesting that these pathways are involved in progression and regression of pressure overload-induced HF. Moreover, they observed that reverse remodeling after debanding, measured by echocardiography and cardiomyocyte cell size, did not relate to fibrillary collagen content measured by picrosirius red staining. This suggests that changes in fibrillary collagen content are not required for reverse LV remodeling. Importantly, collagen density was quite low compared to other experiments and our own, raising questions as to the accuracy of the staining or quantification method (WEINHEIMER et al. 2015). By contrast, we observed significantly higher collagen content in TAC mice compared to the combined HF model, as well as greater differences compared to sham group. We also noted a significant decrease of collagen after debanding. Deviation of outcome might also relate to gender differences, as female C57Bl/6 mice were used in the TAC-MI dual injury model studies, whereas exclusively male mice were used in our experiments. Due to differences in severity of constriction, divergent methods for measuring hypertrophy, and deviation in mouse strains, age, sex and origin, the cardiac phenotype after TAC is highly variable (MOHAMMED et al. 2012; RICHARDS et al. 2019). For in vivo assessment of LV mass and geometry, mostly echocardiography and CMR imaging are used. Ex vivo, the heart or LV can be weighed and related to body weight or tibial length (YIN et al. 1982). In our experiments, all examined mice had the same age and only minor deviations in body weight. According to numerous studies, we observed that LV mass and volumes rise over time after

- 136 - Discussion

TAC compared to sham. We and others (HOUSER et al. 2012) described early impairment of EF, whereas other studies observed early decompensation with late functional impairment as measured by EF (BACMEISTER et al. 2019), which strongly relates to severity of constriction (RICHARDS et al. 2019). We confirmed cardiac hypertrophy by measurements of ex vivo heart weights and calculation of heart weight / body weight ratio as well as LV mass measured by CMR imaging which correlated well (CMR mass vs. necropsy mass r=0.81, p<0.001). As described, not all animals which underwent TAC developed HF with severe impaired heart function, described in the literature as compensated HF (MOHAMMED et al. 2012). We and others observed rapid reverse remodeling with decreased LV mass within one week after aortic debanding (WEINHEIMER et al. 2018). We measured lung and spleen weights but did not observe significant differences in organ masses. In contrast, other groups observed elevated lung weights in the chronic phase after TAC as a further indication of fluid retention due to HF (MOHAMMED et al. 2012). This difference may relate to the range of severity in the present study. Histologically, cardiomyocyte hypertrophy can be measured using wheat germ agglutinin staining and interstitial fibrosis by Masson trichrome or picrosirius red staining (BENSLEY et al. 2016). Here, the quantification depends on the used analysis method and might be not comparable. In summary, we detected rapid cardiac dysfunction and ventricular alterations which were aggravated over time after TAC compared to sham, assessed by functional CMR measurements and on necropsy. Removal of the aortic banding resulted in normalization or partial recovery of EF, dependent on the initial cardiac impairment. Baseline and sham animals did not exhibit differences in cardiac function over the time course.

9.3 Pathology of Inflammation

Inflammation is a critical contributor to the development and progression of HF and a potential predictor of cardiac dysfunction. Early excessive inflammation is thought to aggravate late cardiac dysfunction (SCHIRONE et al. 2017), meaning that treatment of early inflammatory cell infiltration or activity may improve functional outcome.

- 137 - Discussion

There are numerous stimuli of inflammatory activation in pressure overload HF. First, activation of the neurohumoral system is known to instigate bone marrow and spleen monocyte mobilization, initiating a systemic inflammatory response (B. CHEN u. FRANGOGIANNIS 2018). Second, oxidative stress has been demonstrated to promote recruitment of inflammatory cells via cytokine secretion, especially TNF-α, as shown in mice after ANG II-induced HF (SRIRAMULA u. FRANCIS 2015). Third, mechanosensitive signaling could promote fibrotic reprogramming of cardiomyocytes and fibroblasts to secrete pro-inflammatory cytokines, which in turn recruit monocytes from spleen and bone marrow. The strain by pressure overload increases myocardial stretch driving secretion of pro- inflammatory factors. In vitro, mechanical stretch induces chemokine expression in murine and human cardiac fibroblasts, with augmented release of TGF-β (LINDNER et al. 2014). These mechanisms indicate that cell death is not required for activation of inflammation in non-ischemic HF. That said, elevated ventricle pressure leads to modest cardiomyocyte cell death via apoptosis and local release of danger-signals (FRANGOGIANNIS et al. 2002). These danger-signals stimulate toll-like receptor signaling which induces expression of pro- inflammatory cytokines, including TNF-α, thereby recruiting leukocytes to sites of myocardial damage (RAMADAN et al. 2017; YOU et al. 2018). Two weeks after TAC in mice, cardiomyocyte apoptosis was reported, related to TNF-α elevation relative to sham mice, as quantified by terminal deoxynucleotidyl transferase- mediated dUTP-biotin nick end labeling (TUNEL) staining method. By contrast, volume overload as generated in a model of aortic regurgitation, did not exhibit appreciable cardiomyocyte apoptosis (YOU et al. 2018). In the present study, we were surprised to find a similar time course of CXCR4 upregulation and T1 relaxation time elongation, suggesting parallel inflammatory cell infiltration and fibrosis. Using 18F-FDG, which is metabolized by pro-inflammatory macrophages, we detected a composite signal of inflammation and metabolically altered cardiomyocytes. As 18F-FDG did not identify inflammatory cells efficiently, subsequent analysis using the specific tracer 68Ga-pentixafor characterized the inflammatory response over six weeks after TAC. We observed infiltration of leukocytes on day three and seven and mild inflammation after three weeks which was resolved by six weeks. These findings were confirmed by quantitative immunohistology for CD68+ cells, predominantly macrophages, and ex vivo

- 138 - Discussion

autoradiography. Staining for Ly6G+ cells did not reveal significant granulocyte enrichment after TAC. These observations align with previous descriptions of inflammatory cell infiltration after TAC. XIA et al. (2009) characterized the inflammatory and fibrotic response to pressure overload-induced HF by TAC in C57BL6J mice (Jackson). At three days after TAC, maximal inflammatory cytokines such as TNF-α and IL-1β were detected, whereas IL-6 remained low, distinct from infarct models. Inflammatory cytokines peak after three days and decrease to sham level after seven days, based on mRNA measurements. Upregulation of the CXC chemokines MIP-2, IP-10 (CXCL10), and the CC chemokine MCP-1 (CCL2) is maximal at seven days after TAC. Immunohistochemistry revealed modest neutrophil infiltration on day three, and significantly elevated macrophage density on day seven, which resolved after four weeks (XIA et al. 2009). Our results are consistent with these findings, as we detected no neutrophils, but significant macrophage infiltration, both by PET imaging and immunohistology. In contrast to these findings, we also observed a trend to higher macrophage infiltration at three days after TAC, albeit based on a relatively small number of investigated animals. The discrepancy in these findings could originate from distinct mouse suppliers, responses of individuals, different methods and/or cell quantification techniques. In our experiments, some animals reacted with more severe inflammation than others, likely due to the severity of the aortic banding and therefore the strain. We did not directly assess myocardial strain, though this could be interrogated by MR imaging, but is complex to achieve. The relation of quantitatively assessed inflammatory cell infiltration to severity of TAC has not been described. However, quantification methods in published studies rely predominantly on flow cytometry or immunohistological staining, which is restricted to a single terminal time point. Outcome data are therefore difficult to assess serially. Flow cytometry is more quantitatively accurate, reliable and reproducible. Staining methods and analysis vary between studies, which complicates comparisons. Further, multiple studies which investigate inflammation after TAC do not provide information about the severity of constriction. On histology, we observed consistent patterns of macrophage cell infiltration, tending to accumulate in nests in the base and mid-ventricular myocardium, such that regional sampling can affect quantification. Inconsistencies might derive from different methods. XIA et al. (2009) quantified macrophages by counting Mac-2+ cells/mm2. Conversely, we counted CD68+ cells/field in six randomized regions of left ventricle long axis sections. Both

- 139 - Discussion

antibodies are used to detect monocytes and macrophages, whereas Mac-2 is expressed on murine macrophages (SATO u. HUGHES 1994) and CD68 is known to be expressed on murine and human macrophages (RAMPRASAD et al. 1996). Interestingly, both antibodies against monocytes and macrophages are described to stain fibroblasts as well, but with lower intensity (INOUE et al. 2005). WEISHEIT et al. (2014) described macrophage infiltration adjacent to cardiac blood vessels in the mid-myocardial region of the LV, similar to our observations. Macrophage quantification by flow cytometry demonstrated Ly6Clow/M2 macrophages peaking on day six after TAC, and a gradual decline over three weeks, which remained higher than in sham. Notably, flow cytometry analysis showed a trend of Ly6Clow/M2 macrophages elevation on day three, as observed in our study. Conversely, Ly6Chigh/M1 macrophages did not show significant alterations, though numbers progressively increased over three weeks (WEISHEIT et al. 2014). These results are reflected by the same pattern that we observed by CXCR4 upregulation by 68Ga-pentixafor PET imaging and CD68+ immunohistology. 68Ga-pentixafor identifies a broad range of leukocytes which express CXCR4, including both M1- and M2- like macrophages (BORCHERT et al. 2019), meaning this tracer detects inflammation but does not resolve the exact cell composition. By contrast, 18F-FDG is preferentially taken up by M1 pro-inflammatory macrophages (BORCHERT et al. 2019), suggesting that 18F-FDG can distinguish inflammatory subtypes, in the absence of cardiomyocyte uptake and metabolic perturbation. Ly6Clow/M2 macrophages are responsible for tissue regeneration processes such as myofibroblast accumulation, angiogenesis and collagen deposition (NAHRENDORF et al. 2010). These findings are in contrast to MI, where the inflammatory response is dominated early by Ly6Chigh/M1 macrophages (NAHRENDORF et al. 2007). Both macrophage subtypes express cell surface marker CD68 (ROSZER 2015). Additionally, WEISHEIT et al. (2014) described a significant increase of neutrophils peaking on day three after TAC, which we and others did not observe (XIA et al. 2009; PATEL et al. 2017). Moreover, PATEL et al. (2017) describe a significant elevation of total leukocytes and a significant elevation of M1 and a trend to increased M2 macrophages in the myocardium one week after TAC, identified by FACS. Both studies showed that Ly6Chigh and Ly6Clow monocyte cell numbers are elevated in the blood after TAC (WEISHEIT et al. 2014; PATEL et al. 2017). In patients with HF, leukocyte number in the blood correlates with mortality

- 140 - Discussion

(BARRON et al. 2001). In a subsequent study, PATEL et al. (2018) identified one week after TAC monocyte-derived macrophages expressing chemokine C-C motif receptor 2 (CCR2+), which derive from Ly6Chigh monocytes, to be responsible for chronic inflammation and adverse cardiac remodeling. Blockade of CCR2 inhibited macrophage infiltration and reduced cardiac hypertrophy, suggesting these cells as possible therapeutic target for immunomodulation to prevent adverse cardiac remodeling (PATEL et al. 2018). Accordingly, depletion of macrophages resulted in attenuated remodeling and 80% less fibrosis in mice after TAC (KAIN et al. 2016). The presence of T lymphocytes in the myocardium was recently demonstrated early (one week) and late (four weeks) after TAC. T cells are involved in the maintenance of long-term immune response. Treatment with the arthritis drug abatacept beginning two days post-TAC, inhibits T cell activation, cardiac macrophage maturation, and reduces myocardial T cell and macrophage infiltration, resulting in attenuated cardiac hypertrophy after pressure overload. Interestingly, in the absence of IL-10 which is produced by B lymphocytes, the beneficial treatment effect was lacking (KALLIKOURDIS et al. 2017). Taken together, macrophages peak in the myocardium around one week after pressure overload-induced HF. These data suggest that early intervention against specific inflammatory cell types could prevent adverse remodeling, fibrosis and HF. Accordingly, we conclude that the 68Ga-pentixafor signal mainly represents myocardial macrophage infiltration after TAC, as the dominant cell population in this model.

9.4 Imaging of Inflammation

With molecular imaging, specific markers can detect inflammatory sites in the body for diagnosis. Here, we aimed to detect early pathomechanisms of inflammation related to late cardiac dysfunction. Specific targeted treatment of these adverse mechanisms might rescue long term cardiac function and could be precisely guided using molecular imaging. The widely used clinical PET tracer 18F-FDG has been applied in a variety of inflammatory CVD, including atherosclerosis (OSBORN u. JAFFER 2013), sarcoidosis (SKALI et al. 2013), and after MI (W. W. LEE et al. 2012). Atherosclerosis is characterized by lipid-rich atherosclerotic plaques narrowing the vessel wall, which may be infiltrated by inflammatory

- 141 - Discussion

leukocytes, and reflect plaque stability. Here, several imaging tracers targeting different aspects of atherosclerotic plaque have been described, including calcification, neoangiogenesis and inflammation (WILDGRUBER et al. 2013). The most common method to image inflammatory plaques in atherosclerosis is 18F-FDG PET which accumulates in macrophages in the vessel lesions (TARKIN et al. 2014). In sarcoidosis, a systemic inflammatory disease, inflammatory sites in the myocardium can be detected by elevated 18F- FDG uptake (SARASTE u. KNUUTI 2017), based on specific patterns of distribution. Moreover, 18F-FDG identifies inflammatory cell infiltration into the myocardium after MI in mice (W. W. LEE et al. 2012) and in patients (RISCHPLER et al. 2015). MI leads to ischemia and apoptosis of cardiomyocytes, which initiates an immediate inflammatory response with neutrophil granulocytes peaking on day one, followed by pro-inflammatory macrophages peaking on day three to five (FRANGOGIANNIS et al. 2002). These infiltrating leukocytes accumulate 18F-FDG. However, as a glucose analogue, 18F-FDG is not specific to macrophages and is taken up by a variety of metabolically active cells, including cardiomyocytes. The 18F-FDG uptake of cardiomyocytes has to be suppressed for delineation of inflammatory signal. This may be accomplished using several methods. In patients, a high fat diet, fasting and heparin injection prior to the scan suppress cardiomyocyte glucose uptake (TANG et al. 2016). In mice, injection anesthesia with K/X is often used. Notably, standard inhalation anesthesia with isoflurane amplifies cardiomyocyte 18F-FDG uptake (K. H. LEE et al. 2005; THACKERAY et al. 2015a). The partial α2-adrenergic receptor agonist xylazine suppresses insulin secretion from pancreas (WINDELOV et al. 2016) leading to hypoinsulinemia and consequentially rising blood glucose levels (ABDEL EL MOTAL u. SHARP 1985). Thus, 18F-FDG transport by GLUT4 into the cardiomyocytes is reduced, as GLUT4 translocation is stimulated by insulin (LETO u. SALTIEL 2012). This suppression technique works well in healthy mice (THACKERAY et al. 2015a). We aimed to image inflammatory myocardial macrophages after TAC using 18F-FDG. At three days after surgery, the 18F-FDG signal correlated well with the impaired heart function, but the signal was significantly more robust than expected relative to the number of inflammatory cells. After quantification of CD68+ leukocytes in the myocardium, we found that the number of leukocytes was lower compared to the inflammation positive MI group, but the 18F-FDG signal was higher comparing to the signal in the early phase after MI. We found GLUT4

- 142 - Discussion

overexpression in the stressed myocardium, reflecting increased glycolysis in the failing heart. Assuming that with time after TAC induced HF, cardiomyocyte metabolism changes and adapts, GLUT4 expression rises with further disease progression (SEN et al. 2013). These results suggest that the 18F-FDG signal is not derived from leukocytes alone, but rather a composite of leukocyte infiltration and altered cardiomyocyte metabolism (GLASENAPP et al. 2019). Consequently, as glucose utilization in the stressed myocardium increases (KARWI et al. 2018) the glucose uptake in metabolically altered and functionally compromised cardiomyocytes after TAC cannot be suppressed effectively anymore. Accordingly, a more specific inflammatory PET imaging tracer to detect myocardial inflammation independent of cardiomyocyte metabolism is desirable. 68Ga-Pentixafor targets chemokine receptor CXCR4, which is expressed on a variety of leukocytes. Imaging of CXCR4 allows tracking of physiological processes including cell migration, homing of bone marrow-derived stem cells to sites of injury, angiogenesis and cell proliferation. Moreover, it enables imaging of inflammation in oncological and cardiovascular diseases (BUCK et al. 2017). 68Ga-Pentixafor displayed selective accumulation in human leukocyte subsets including monocytes, macrophages, and lymphocytes in vitro (BORCHERT et al. 2019). In vivo, 68Ga-pentixafor identified inflammation in the infarct territory in mice and humans after MI (THACKERAY et al. 2015b). 68Ga-Labeled radio-tracers have the advantage that they are more readily available than cyclotron products, as they are using a generator and chemistry is relatively straightforward. However, 68Ga has a high positron range (3-4mm in water and 6- 8mm in air), which can complicate imaging of small organs and structures, especially in mice. The small size of the left ventricle wall (~1-1.5mm) in combination with the camera resolution (~1.5-2.0mm) and positron range, renders the diffuse inflammatory signal difficult to resolve. Indeed, 68Ga-pentixafor images post-TAC did not describe clear ventricle contours. To overcome this limitation, we performed sequential 18F-FDG imaging under isoflurane anesthesia to localize myocardial contours in the same anatomic space for analysis. Interestingly, we observed also in the sham operated group a slightly elevated early inflammatory signal, which is likely due to the inflammatory response caused by surgical intervention in general. In the present project, 68Ga-pentixafor successfully detected inflammatory upregulation over six weeks after TAC, with peak signal at day seven after surgery, recovering to sham levels

- 143 - Discussion

after six weeks. Notably, the degree of diffuse signal was more consistent with the number of leukocytes in the LV than was the case for 18F-FDG under K/X. These results were confirmed by immunohistology, ex vivo autoradiography and biodistribution. We found that early inflammation after seven days predicts late cardiac dysfunction. Moreover, the PET signal sensitively identified recurrent inflammation after debanding surgery, further supporting the notion of a systemic inflammatory response to general surgical intervention. Three weeks after removal of the aortic banding, we observed alleviated CXCR4 signal compared to age- matched animals which underwent sham rTAC. Our results suggest that CXCR4 targeted imaging in non-ischemic HF may be a valuable tool for identification of high-risk patients for HF progression.

9.5 Pathology of Fibrosis

In non-ischemic pressure overload-induced HF, inflammation and fibrosis develop in parallel and coexist (BACMEISTER et al. 2019). Established myocardial fibrosis does not respond sufficiently to conventional treatment and is associated with inflammation and adverse remodeling (HEYMANS et al. 2009). Non-invasive imaging of early fibrotic mechanisms in response to pressure overload HF could depict active fibrosis and determine targets and time points for novel molecular therapies to arrest or reverse fibrosis progression. Collagen deposition in the myocardium begins by three days after TAC and reaches maximum at day seven, as quantified by picrosirius red histology (XIA et al. 2009). Collagen deposition follows an early upregulation of matricellular proteins, such as osteopontin, three days after TAC and tenascin-C seven days after TAC. By one week, fibrosis tends to be localized to perivascular sites, with beginnings of diffuse interstitial infiltration of spindle- shaped myofibroblasts. Quantification of collagen content demonstrated a homogenous distribution over the whole left ventricle as well as in the right ventricle. Interestingly, collagen content was significantly decreased after four weeks without treatment, suggesting a compensatory adjustment by the murine heart. As inflammation initiates fibrosis, reactive fibrosis in response to pressure overload might partly resolve as the inflammation decreases. These observations were reflected by our own results. We observed significantly elevated fibrosis at seven days, identified by picrosirius red staining, which manifested as perivascular

- 144 - Discussion

and diffuse interstitial fibrosis. After three and six weeks of TAC, we observed a modest but not statistically significant attenuation of collagen content relative to day seven. The quantification of fibrosis in our longitudinal study aligns with the T1 relaxation time. Furthermore, the interplay between inflammation and fibrosis and the shift of leukocytes to fibroblasts with subsequent fibrosis and vice versa are crucial contributors in HF. Pro-fibrotic factors such as TGF-β activate resident cardiac fibroblasts. Activated fibroblasts proliferate and differentiate into myofibroblasts which produce collagen, causing interstitial fibrosis. In non-ischemic HF, myofibroblasts and fibrous tissue continue to secrete ANG II and fibrogenic growth factors including TGF-β, stimulating further fibrogenesis. As macrophages are able to transition to a fibroblast-like phenotype (HAIDER et al. 2019), and fibroblasts secrete pro-inflammatory cytokines under certain conditions (SHINDE u. FRANGOGIANNIS 2014), both mechanisms are interconnected and accelerate each other. Myofibroblasts are known as main contributing cells in cardiac fibrosis, but also leukocytes including macrophages and lymphocytes, as well as vascular endothelial cells and cardiomyocytes are thought to contribute to fibrogenesis (KONG et al. 2014). Moreover, T lymphocytes and macrophages infiltrate the adventitia of intramural coronary arteries, activating fibroblast differentiation to myofibroblasts and promoting reactive perivascular fibrosis (WEBER et al. 2013). It has been suggested that human myofibroblasts in the setting of HF have the capacity to revert to fibroblasts. Specifically, inhibition of isolated LV myofibroblasts from transplant recipients by TGF-β receptor kinase initiated dedifferentiation of myofibroblasts to a fibroblast phenotype, characterized by loss of α-smooth muscle actin and reduced expression of profibrotic factors and cytokines (NAGARAJU et al. 2019). Furthermore, direct reprogramming of myofibroblasts and fibroblasts into induced cardiomyocyte-like cells was demonstrated by forced overexpression of cardiac transcription factors in both replacement and reactive fibrosis in vitro and in vivo, suggesting the potential of active fibrosis reversibility (TALMAN u. RUSKOAHO 2016). Accordingly, pharmacological anti-inflammatory treatments have been suggested to attenuate myocardial fibrosis. In one study, mice were treated four weeks after TAC with non-steroidal anti-inflammatory drug aspirin. Here, aspirin reduced interstitial fibrosis in vivo as assessed by Masson trichrome staining compared to untreated TAC and control mice. In addition, aspirin significantly suppressed the expression of α-smooth muscle actin and attenuated

- 145 - Discussion

collagen production after TAC in vivo, assessed by Western blotting, and in vitro in cardiac fibroblasts (LI et al. 2018b). Another study demonstrated that treatment with pirfenidone, an anti-fibrotic drug with anti-inflammatory properties, reduced myocardial fibrosis and inflammation and increased survival rate after TAC (Y. WANG et al. 2013). It has been reported that ventricular remodeling is attenuated in patients after medical treatment or implantation of LVAD, as determined by echocardiography. Some patients sustain acceptable cardiac function even after removal of LVAD (WOHLSCHLAEGER et al. 2005), suggesting an additive therapeutic effect. However, it remains controversial whether collagen content or fibrosis can be attenuated by medical or interventional treatments (MANN et al. 2012). Some studies described lower collagen and reduced cardiomyocyte size (J. MULLER et al. 1997; KATO et al. 2011), whereas others report higher collagen deposits and cardiomyocyte elongation after LVAD implantation (MCCARTHY et al. 1995; LOK et al. 2015). In mice, hemodynamic unloading results in recovery of LV structure, but does not result in complete normalization of LV function after TAC. Moreover, the recovery depends on the duration of pressure overload on the myocardium before removal of the banding. Interstitial collagen content was demonstrated to decrease significantly after aortic debanding compared to TAC (GAO et al. 2005a). Here, the aortic debanding resulted in a rapid partial recovery of function, decrease of dilatation and reduced mass, as measured by CMR and necropsy. We observed a significant attenuation of fibrosis and T1 relaxation time after rTAC. Our observations suggest that unloading interrupts the continuous cycle of inflammation and fibrosis. Taken together, TAC seems to induce chronic cardiac stress which is maintained by continuous inflammatory and fibrotic remodeling processes and can be visualized using multimodal imaging. These processes can be attenuated and partially reversed by mechanical unloading.

9.6 Imaging of Fibrosis

As we detected parallel inflammation and fibrosis by histology, and both mechanisms are known to trigger each other (SHINDE u. FRANGOGIANNIS 2014; HAIDER et al. 2019), an investigation of fibrotic alterations after TAC with a specific fibrosis marker would provide

- 146 - Discussion

additional insights into fibrogenesis and its relationship to cardiac dysfunction following pressure overload HF. Generally, small heart size and rapid heart rate complicate CMR imaging in mice. Using LGE, focal myocardial fibrosis after MI can be detected in patients (KIM et al. 1999) and mice (BOHL et al. 2009), but the accuracy to detect diffuse interstitial fibrosis is highly reduced. In a rat model of cardiotoxicity (LIGHTFOOT et al. 2010) and mouse model of muscular dystrophy (STUCKEY et al. 2012) diffuse myocardial fibrosis has been identified by LGE. However, the sensitivity of LGE is limited and not advisable for diffuse myocardial fibrosis quantification (L. ILES et al. 2008). Native T1 mapping allows for non-invasive detection of focal (DALL'ARMELLINA et al. 2012) and diffuse interstitial fibrosis in the myocardium (GONZALEZ et al. 2018). T1 measurements pre- and post-administration of contrast agent enable calculation of extracellular volume fraction, which serves as quantitative indicator for myocardial fibrosis (BURT et al. 2014). Nevertheless, T1 is not specific to fibrosis as it measures differences in tissue composition, such that inflammation or edema could also evoke prolonged T1 relaxation time (GERMAIN et al. 2014). In patients after MI, pre-contrast T1 relaxation time was highest in the acute stage, and higher T1 relaxation time was observed in infarct compared to remote myocardium (MESSROGHLI et al. 2007). These results align with serial myocardial T1 mapping in dogs peaking five days after coronary occlusion (WISENBERG et al. 1988), suggesting that the mechanism behind this phenomenon may be early myocardial edema, which is known to evoke prolonged T1 relaxation time (HIGGINS et al. 1983). However, acute response to MI and TAC are not equivalent due to different biological mechanisms. In a prospective longitudinal study including more than 600 patients diagnosed with non- ischemic dilatated cardiomyopathy, the prognostic relevance of T1 mapping parameters was compared to conventional outcome markers, such as presence of LGE and LVEF<35%. On multivariate stepwise analysis, native T1 relaxation time was the only independent predictor of HF events and mortality (PUNTMANN et al. 2016). Furthermore, the relationship of native T1 relaxation time and myocardial fibrosis in patients with aortic stenosis and heart valve diseases was investigated. Fibrosis was measured by targeted myocardial biopsy of matched myocardial segments and stained with picrosirius red for quantification of collagen. A

- 147 - Discussion

significant correlation was observed between T1 relaxation time and LV mass, indicating that native T1 relaxation time is a precise marker of diffuse myocardial fibrosis in non-ischemic valvular heart disease (KOCKOVA et al. 2016). Both studies relied on septal T1 relaxation times. By contrast, we analyzed T1 relaxation time in acquired T1 maps for both whole myocardium and septum-only to exclude deviations due to lung signal, but did not detect differences between these measurements. In another study, interstitial myocardial fibrosis was measured before and after treatment with mineralocorticoid receptor antagonists using T1 mapping in a group of twelve patients with non-ischemic dilated cardiomyopathy on conventional therapy. Myocardial T1 relaxation time in a single midventricular slice was demonstrated to predict beneficial remodeling, as determined by LVEF and ESV, after treatment with mineralocorticoid receptor antagonist in patients with non-ischemic dilated cardiomyopathy (BRADHAM et al. 2017). This emphasizes T1 relaxation time as a non- invasive measure to monitor response to therapy. In rats with HF after two weeks of continuous ANG II administration, T1 maps pre- and post- contrast agent were used to calculate extracellular volume fraction to detect diffuse myocardial fibrosis, suggesting extracellular volume fraction as quantitative measure for diffuse myocardial fibrosis (MESSROGHLI et al. 2011). Serial T1 mapping in TAC mice with and without ANG II type 1 receptor antagonist losartan treatment revealed elevated native T1 relaxation time compared to sham after seven days which declined by four weeks. Post-contrast T1 and extracellular volume fraction were increased after four weeks but not after seven days. Losartan treatment lowered collagen content after four weeks but did not affect T1 relaxation times or extracellular volume fraction calculation. The authors speculated that early T1 prolongation may relate to edema, and absence of early increased ECV relates to increased intracellular water content (STUCKEY et al. 2014). However, they emphasized that native T1 could quantify tissue damage. Indeed, histological quantification of fibrosis was only performed at the end point of experiment after four weeks and not at the intermediate time point after seven days. These results are partly in contrast to our findings, as we detected prolonged native T1 relaxation time at all imaging time points after TAC up to six weeks but with a trend to declination. We and others (XIA et al. 2009) demonstrated the presence of fibrosis after TAC by histology already after one week. It should be noted that T1 sequences

- 148 - Discussion

vary greatly between investigations and lack standardization. The specific sequences may influence sensitivity of the measurement. In summary, we cannot fully determine which tissue changes are reflected by altered T1 relaxation times. We found native prolonged T1 relaxation after TAC in the early and chronic phase which correlated well to cardiac dysfunction and collagen content at matched time points. The pattern of elevated T1 relaxation time in the time course after TAC was reflected by quantification of collagen but might also reflect the inflammatory upregulation occurring in parallel post-TAC. The concurrent correlation of T1 relaxation time to 68Ga-pentixafor signal could reflect the parallel development of inflammation and fibrosis or it may indicate that T1 relaxation time is influenced by inflammatory cell content. Interestingly, we observed a significant and distinct decline of T1 relaxation time as consequence to removal of the aortic banding, suggesting that relief of myocardial strain leads to rapid myocardial tissue recovery. We detected a greater change after debanding with ~50% less fibrosis content, but only slight changes in inflammation signal and cell numbers. This observation supports the notion that T1 relaxation time reflects diffuse myocardial fibrosis rather than inflammation.

9.7 Imaging for Targeted Therapy and Treatment Guidance

The parallel development of inflammation and fibrosis suggests synergistic pathology. Strategies to mitigate one or the other process may impact the complementary mechanism. The rationale of targeted therapy and treatment guidance is to link drugs and molecular imaging agents to the same (patho)physiological target. Using this approach, suitable patients who would benefit from a specific targeted treatment can be identified and the optimal time point for treatment or intervention may be determined, i.e. the maximal imaging signal (HESS et al. 2019). Monitoring of response to therapy and guidance of treatment by imaging has gained traction especially in cancer (PANTEL u. MANKOFF 2017). Beyond oncology, this paradigm has the potential to treat early pathological mechanisms, e.g. in inflammatory CVD (HESS et al. 2019). In this sense, chemokine receptor CXCR4 is a valuable target for imaging and therapy. Pharmacologic CXCR4 blockade administered by repeated bolus injection of selective antagonist POL5551 within one week after reperfused MI in mice lowered inflammation,

- 149 - Discussion

promoted angiogenesis and tissue repair, and attenuated LV remodeling and cardiac dysfunction (Y. WANG et al. 2019). Moreover, in a porcine model of reperfused MI the analog clinical-stage compound POL6326 reduced infarct volume and promoted LVEF recovery (VAN DER SPOEL et al. 2011). These studies emphasize the high translational potential of CXCR4-targeted treatment, as the therapeutic is clinically applicable. As described previously, CXCR4 is upregulated in patients after MI and can be identified by 68Ga-pentixafor PET imaging (THACKERAY et al. 2015b), reflecting a broad range of leukocytes (BORCHERT et al. 2019). Likewise, in the present study, increased CXCR4 signal after TAC suggests that specific receptor blockade may alter the inflammatory response in non-ischemic HF. Indeed, in mice with dilated cardiomyopathy, administration of AMD3100 significantly reduced perivascular and interstitial fibrosis (CHU et al. 2019). These results indicate that the anti-inflammatory action of CXCR4 blockade could have the ancillary benefit of attenuated cardiac remodeling and fibrosis, improving cardiac functional outcome. T1 mapping could provide a complementary non-invasive estimate of diffuse fibrosis to test subsequent therapeutic effects of novel and emerging anti-fibrotic treatments. While no targeted anti-fibrosis therapeutics are clinically available, several common HF drugs affect fibrosis such as ACE inhibitors or ANG II type 1 receptor blockers (HARA et al. 2017). Native T1 mapping was demonstrated to correlate to biopsy-quantified LV fibrosis in patients with aortic stenosis who underwent aortic valve replacement (BULL et al. 2013). Similar correlations were demonstrated in HF due to chronic heart valve diseases (KOCKOVA et al. 2016), suggesting non-invasive T1 mapping without administration of contrast agent as suitable to monitor therapies. But, experience in broader HF is limited and would require dedicated study. In summary, both targeted imaging and treatment of the molecular mechanisms of inflammation and fibrosis could assist in identifying therapeutic efficacy, selecting candidate patients for targeted therapies, and monitoring the treatment effects longitudinally. Studies to support this notion should be a next step, for which our current work provides a foundation.

- 150 - Discussion

9.8 Limitations

Some limitations of the conducted studies have to be considered. As we aimed to characterize the acute inflammatory response in pressure overload by 18F-FDG PET imaging, we did not assess the long-term cardiac functional outcome in these animals. We were primarily interested in the acute stage of disease, when inflammation is thought to peak after pressure overload, and focused the 18F-FDG examination on the timeframe. The lack of effective suppression precluded further study with this method. A second limitation is that 68Ga-pentixafor affinity for the murine CXCR4 receptor isotype is lower compared to the human form (THACKERAY et al. 2015b; BUCK et al. 2017), which limits maximal receptor binding. We did not perform a blocking study in mice after TAC. However, the tracer specificity has been confirmed in blocking studies using coinjection of CXCR4 antagonist AMD3100 in mice after MI (THACKERAY et al. 2015b). Finally, for CXCR4 evidence and colocalization in tissue, immunostaining antibodies are not reliable in mice. The exact cell composition representing the 68Ga-pentixafor signal has not yet been firmly determined, but flow cytometry analysis revealed elevation of all leukocyte subsets after TAC with mainly macrophages as the dominant subtype (WEISHEIT et al. 2014). 68Ga- Pentixafor uptake was verified in all leukocyte subsets (BORCHERT et al. 2019). The promiscuous binding of 68Ga-pentixafor suggests a readout of total leukocyte infiltrate. It should also be noted that T1 relaxation time is not entirely specific for fibrosis, as it reflects tissue alterations in general and is dependent of molecular tissue composition. Thus, prolonged T1 relaxation times might also reflect inflammatory cell infiltration, edema or other cellular myocardial alterations. However, T1 relaxation time has the advantage of quantitative assessment, and we found that the T1 relaxation times show a significant correlation with fibrosis. The time course of T1 mapping more closely reflects the pattern of myocardial tissue fibrosis than inflammation after aortic debanding.

9.9 Clinical Perspective

Non-invasive imaging techniques are widely used for diagnosis and surveillance of treatment success. The spectrum of applications is growing, and emerging techniques provide optimized and personalized imaging approaches, depending on patients and diseases.

- 151 - Discussion

In the clinics, PET imaging is widely used to detect cardiac disease with a focus on measurement of myocardial perfusion and viability (SARIKAYA 2015). Moreover, imaging of inflammatory mechanisms in CVD is gradually gaining clinical acceptance. 18F-FDG is often used clinically as it is easy to synthesize, widely available and has proven its utility in imaging of local cardiovascular inflammation, such as atherosclerosis, sarcoidosis and after ischemic injury. Overall, the accuracy of imaging inflammation and fibrosis in non-ischemic HF is complicated by the global and diffuse distribution of leukocytes and fibroblasts over the whole myocardium, which may pose a formidable challenge to isolate the precise signal source, due to a lack of internal normal reference region. We showed that 18F-FDG is not suitable to detect diffuse myocardial inflammation in pressure overload HF. 68Ga-Pentixafor effectively targets CXCR4 upregulation in cancer patients and has also demonstrated its utility in inflammatory CVD such as atherosclerosis and ischemic myocardial injury. In mice and human patients, 68Ga-pentixafor identifies local inflammatory myocardial cell infiltration after ischemia, but not all patients after MI show elevated 68Ga-pentixafor uptake in the infarct region in contrast to mice (THACKERAY et al. 2015b). Whereas imaging of local inflammation in CVD has gained a variety of valuable clinical applications, more research for global diffuse myocardial tissue alterations in CVD is required. As we detected diffuse myocardial cell infiltration in non-ischemic HF after TAC in mice by 68Ga-pentixafor PET and predicted late functional outcome by early inflammation, targeted treatment of the early excessive inflammation could complement conventional therapy strategies by treatment guidance. Therapeutic blockade of CXCR4 by clinically relevant antagonists was successfully demonstrated after MI in mice and pigs (Y. WANG et al. 2019), suggesting that early CXCR4 blockade may attenuate LV remodeling also in non-ischemic cardiomyopathies. Treatment prior to peak inflammation in response to TAC might prevent migration of deleterious CXCR4+ leukocyte subpopulations to the myocardium and prevent adverse remodeling. Such treatment strategies have yet to be evaluated in clinics. Additionally, T1 mapping might provide additional insights on molecular changes (BURT et al. 2014) and indicate cardiac diffuse interstitial fibrosis (EVERETT et al. 2016). This technique has been applied to detect diffuse myocardial fibrosis in patients with HF due to aortic stenosis and heart valvular disease (BULL et al. 2013; KOCKOVA et al. 2016), but

- 152 - Discussion

broader application in failing hearts is limited. As we detected diffuse interstitial fibrosis in the myocardium after pressure overload-induced HF by T1 mapping, it may be beneficial to assess cellular abnormalities which might predict cardiac dysfunction, risk of hospitalization or higher risk of mortality. Importantly, the mechanical unloading procedure confirmed attenuation of inflammation and reversal of myocardial fibrosis, which has also been described in some patients after LVAD implantation. We demonstrated that the pathophysiological mechanisms of these processes can be sensitively monitored by imaging. As we defined the pathogenesis of the inflammatory and fibrotic response to pressure overload-induced HF and after mechanical unloading by imaging, these mechanisms and methods could be used to manage HF therapy. In brief, with both imaging methods, patients at high risk could be identified and assigned to inflammation- or fibrosis-targeted therapy to maximally improve functional outcome. This individualized therapy could then be monitored by imaging.

9.10 Future Perspectives

With an established time course of inflammation and fibrosis after TAC and in response to ventricular unloading, non-invasive imaging provides the opportunity to monitor therapies targeted to pathogenic processes and investigate the interaction between these processes. Testing of different drugs and time points in a longitudinal imaging study after TAC and in combination with rTAC would facilitate this approach. One option would be a long-term medical treatment in mice after TAC with clinically relevant drugs, e.g. ACE inhibitors. Treatment with the ACE inhibitor captopril has been established to attenuate TAC induced myocardial hypertrophy in mice (P. MULLER et al. 2009; Y. ZHANG et al. 2019). This positive remodeling effect occurs only when the treatment is administered during sleep time of mice, illustrating the importance of diurnal timing for therapies (MARTINO et al. 2011). In patients with HF, ACE inhibitors including enalapril are widely used to manage HF symptoms, as part of a first stage therapeutic regimen (PONIKOWSKI et al. 2016). In addition to anti-hypertensive and remodeling efficacies, enalapril is known to inhibit monocyte release from the spleen, lowering leukocyte infiltration to the myocardium (LEUSCHNER et al. 2010). Oral enalapril treatment reduced the number

- 153 - Discussion

of CD45+ leukocytes in the myocardium within one week of coronary artery occlusion in mice, measured by flow cytometry and reduced 68Ga-pentixafor signal (THACKERAY et al. 2015b). Imaging of the molecular mechanisms of inflammation and fibrosis in mice following continuous administration of ACE inhibitor enalapril could provide additional insights on the impact of medical treatment on the inflammatory response to pressure overload-induced HF and ancillary fibrotic response. However, enalapril treatment has a wide range of systemic effects which should be considered. Furthermore, recent studies focused on the role of the chemokine receptor CXCR4 and its ligand CXCL12 in inflammatory cardiovascular diseases such as atherosclerosis and myocardial ischemia as a possible therapeutic target. Chemokine CXCL12 is highly expressed after MI by cardiomyocytes, fibroblasts and endothelial cells (HU et al. 2007) and induces leukocyte recruitment to the affected region (DORING et al. 2014). Chemokine receptor CXCR4 is mainly expressed on leukocyte subsets and plays a major role in lymphocyte trafficking and recruitment at sites of inflammation with a high concentration of CXCL12 (MURDOCH 2000; BORCHERT et al. 2019). Cardioprotective properties of the CXCR4/CXCL12 axis in the human adult heart are described as supporting progenitor cell invasion at sites of myocardial ischemia (DORING et al. 2014). As we determined peaks of cardiac CXCR4 upregulation following pressure overload-induced HF, blockade using CXCR4 antagonists like AMD3100 may suppresses excessive infiltration of inflammatory cells and rescue cardiac function (CHU et al. 2015; Y. WANG et al. 2019). In MI, a single dose treatment was described as cardioprotective (JUJO et al. 2013), whereas continuous administration had adverse effects on cardiac function (JUJO et al. 2010). Cardiomyocyte-specific CXCR4 knockout mice develop progressive HF (LAROCCA et al. 2019), and deletion of CXCR4 in cardiomyocytes exacerbates cardiac dysfunction in a model of isoproterenol administration (E. R. WANG et al. 2014). In addition, cardiac fibrosis was successfully reduced by treatment with a CXCR4 antagonist in a diabetic rat model (CHU et al. 2015) and in mice with dilative cardiomyopathy (CHU et al. 2019). Consequently, in patients with progressive cardiomyopathy and HF, the CXCR4 axis could serve as a therapeutic target (LAROCCA et al. 2019). Here, molecular imaging could monitor the response to therapy and evaluate therapeutic efficacy.

- 154 - Discussion

As macrophages are major contributors to adverse LV remodeling and fibrosis, targeted treatments to suppress macrophage activity may be beneficial. Specific depletion of macrophages by i.v. injection of clodronate-loaded liposomes results in less cardiac remodeling, reduced fibrosis, and downregulation of pro-fibrotic genes after TAC (KAIN et al. 2016). Using this method, cells of the mononuclear phagocytic system, including macrophages, are depleted. Clodronate-loaded liposomes are ingested by macrophages via endocytosis, and induce cell death by apoptosis (VAN ROOIJEN u. SANDERS 1994; VAN ROOIJEN u. HENDRIKX 2010). Suppression of the early myocardial macrophage infiltration after pressure overload-induced HF may rescue late cardiac function after TAC. The effect could be monitored by imaging. Furthermore, the interplay between fibrosis and inflammation could be evaluated more thoroughly by specific characterization of the fibrotic response after TAC by multimodal imaging. A novel radiotracer has been developed to image FAP, which is upregulated on activated fibroblasts and strongly expressed by myofibroblasts. PET imaging of FAP during the acute and chronic phase after TAC would provide complementary insights into fibroblast dynamics and fibrosis development. FAP is expressed in human hearts after myocardial stress or injury, but is generally absent from the healthy heart (AGHAJANIAN et al. 2019). Thus, FAP also has a potential as a therapeutic target (HAIDER et al. 2019). 68Ga-Labeled fibroblast activation protein inhibitors were recently demonstrated to be a suitable tracer for imaging of cancer-associated fibroblasts in human patients (GIESEL et al. 2019). After coronary ligation, an acute inflammatory reaction is initiated followed by replacement fibrosis with scar formation in which fibroblasts differentiate into myofibroblasts. Elevated FAP expression has been detected in rats after coronary artery ligation and in human patients after acute MI and is suggested to be involved in myocardial wound healing and remodeling (TILLMANNS et al. 2015). Thus, FAP is a potential therapeutic target. 68Ga-Labeled FAP inhibitor accumulated in the infarct territory after permanent coronary artery ligation in rats peaking on day six, and specificity of the signal was confirmed by blocking and immunofluorescence staining (VARASTEH et al. 2019). As FAP is also highly expressed following TAC (AGHAJANIAN et al. 2019), longitudinal imaging of FAP upregulation following pressure overload-induced HF would provide novel insights into fibrogenesis in response to hemodynamic stress.

- 155 - Discussion

Finally, investigation of the inflammatory and fibrotic pathomechanisms in other non- ischemic HF models such as ANG II or isoproterenol infusion may be useful to understand the pathogenesis of fibrosis in the absence of overt inflammation due to injury.

9.11 Conclusions

In conclusion, myocardial inflammation occurs early in non-ischemic pressure overload- induced heart failure in mice. Early myocardial macrophage infiltration predicts late cardiac dysfunction. Moreover, myocardial fibrosis parallels inflammation and is related to functional impairment, suggesting an interaction between inflammatory cells and fibroblasts. Ventricular unloading attenuates myocardial inflammation and significantly reduces interstitial fibrosis, underscoring the potential role in pathology. Unloading evokes reverse remodeling and partially rescues cardiac function. CXCR4 PET imaging sensitively detects myocardial inflammation in response to pressure overload-induced HF and response to unloading. T1 mapping provides a non-invasive read- out of interstitial fibrosis in progression and regression of HF. These studies provide the foundation for image-guided therapy, in which the response to targeted treatment of early inflammatory and/or fibrotic mechanisms, which may prevent late cardiac dysfunction, could be monitored by imaging.

- 156 - Summary

10 Summary

Aylina Glasenapp

Multimodality Molecular Imaging of Inflammation and Fibrosis in Pressure Overload- Induced Heart Failure and After Mechanical Ventricular Unloading

HF is a multifactorial syndrome and a leading cause of death worldwide with increasing prevalence. Inflammation and fibrosis are thought to contribute to adverse cardiac remodeling early in HF and are associated with prognosis. Early modulation of these pathophysiological mechanisms during adverse myocardial remodeling might rescue heart function. Regional characterization of inflammation and/or fibrosis is challenging in HF due to its diffuse and dynamic pattern, which is difficult to assess serially. Multimodality molecular imaging may facilitate visualization of these processes to predict outcome and monitor therapy. Here, we aimed to characterize the temporal progression of inflammation and fibrosis and their relation to cardiac dysfunction in non-ischemic HF. Using molecular imaging to serially assess chemokine receptor CXCR4 (inflammation) and T1 relaxation time (fibrosis), we evaluated the early and chronic tissue alterations after TAC in comparison with CMR-derived function over three to six weeks. We further evaluated these imaging parameters in response to ventricular unloading. In our first study, we identified early pro-inflammatory macrophage infiltration and altered glucose metabolism in the myocardium in the first days after TAC using 18F-FDG PET imaging. The tracer 18F-FDG is a glucose analogue and accumulates in pro-inflammatory macrophages and cardiomyocytes. Despite suppression of cardiomyocyte glucose transport by K/X anesthesia, increased 18F-FDG was observed throughout the LV beyond the signal expected for inflammatory cells alone. Indeed, glucose transporter expression was upregulated in LV myocardium, reflecting metabolic alteration. Accordingly, 18F-FDG PET is not effective to delineate the inflammatory cell signal. Consequently, we applied a more selective molecular tracer targeting inflammation to illustrate the local inflammatory response to pressure overload and its contribution to pathogenesis. In our second study we investigated the acute and chronic phase after pressure overload- induced HF targeting chemokine receptor CXCR4 by 68Ga-pentixafor PET. CXCR4 was upregulated after TAC compared to sham, representing inflammatory leukocyte infiltration in

- 157 - Summary

the myocardium. Quantitative immunohistological staining for CD68+ cells reflects time course of CXCR4 upregulation. Ex vivo autoradiography and biodistribution measurements after TAC confirmed LV 68Ga-pentixafor uptake. We found myocardial macrophage infiltration peaking on day seven, predicting late functional outcome on correlation analysis. Therefore, 68Ga-pentixafor as ligand to CXCR4 may provide insights into novel therapeutic approaches. Parallel to inflammation, we also detected myocardial fibrosis by CMR T1 mapping. T1 relaxation time was prolonged from one to six weeks after TAC compared to sham. Quantitative histological analysis of fibrosis reflects time course of prolonged T1 relaxation time. We found T1 relaxation time related to functional impairment and fibrotic area. Additionally, we found that extent of myocardial fibrosis related to functional impairment. In our third study, we evaluated the response to surgical mechanical unloading of the left ventricle. Contractile function was rapidly improved by removal of the aortic constriction, which virtually normalized EF. One week after debanding, T1 relaxation time was restored towards sham values, and inflammation CXCR4 PET measurements gradually declined by three weeks after debanding. In conclusion, we found that inflammation and fibrosis developed in parallel early in the pathogenesis of pressure overload-induced HF, as detected by multimodal molecular imaging. While 18F-FDG could not distinguish between metabolically perturbed cardiomyocytes and inflammatory cells, chemokine receptor CXCR4 upregulation, visualized by 68Ga-pentixafor PET, depicted the inflammatory response peaking on day seven after TAC, and receding over three to six weeks. The CXCR4 signal at seven days predicted chronic ejection fraction at three to six weeks. Prolonged T1 relaxation time reflected myocardial fibrosis and was related to cardiac dysfunction. Ventricular unloading alleviated both inflammation and fibrosis and partially restored cardiac function which was sensitively monitored by imaging. However, the cause-effect relationship between mechanistic and pathological processes cannot yet be fully determined. Our studies provide a foundation for image-guided therapy, in which the response to targeted treatment of early inflammatory and/or fibrotic mechanisms may prevent late cardiac dysfunction and could be monitored by noninvasive molecular imaging.

- 158 - Zusammenfassung

11 Zusammenfassung

Aylina Glasenapp

Multimodale molekulare Bildgebung von Inflammation und Fibrose bei kardialer Drucküberlastung und nach mechanischer ventrikulärer Entlastung

Die Herzinsuffizienz ist ein multifaktorielles klinisches Syndrom und das terminale Stadium von vielen kardiovaskulären Erkrankungen. Sowohl global als auch in Deutschland sind kardiovaskuläre Erkrankungen die häufigste Todesursache mit einer altersabhängig steigenden Prävalenz und Inzidenz. Zudem erhöht die Anzahl der erfolgreichen akuten Interventionen bei kardiovaskulären Ereignissen die Anzahl der Patienten, welche im späteren Verlauf eine Herzinsuffizienz entwickeln können. Die Herzinsuffizienz ist durch eine verminderte Pumpleistung des Herzens gekennzeichnet, der Körper kann nicht mehr mit dem erforderlichen Herzzeitvolumen bei normalem enddiastolischen Ventrikeldruck versorgt werden. Klinisch wird zwischen einer Links-, Rechts-, und Globalinsuffizienz unterschieden, wobei die Linksinsuffizienz wiederum in die Herzinsuffizienz mit reduzierter Ejektionsfraktion (HFrEF) und die Herzinsuffizienz mit erhaltener Ejektionsfraktion (HFpEF) unterteilt ist. Die reduzierte Ejektionsfraktion mit einer linksventrikulären Pumpfunktion <50% bei erhöhtem linksventrikulärem enddiastolischem Druck und Volumen wird auch als systolische Herzinsuffizienz bezeichnet. Patienten mit einer Herzinsuffizienz können an einer Vielzahl von Symptomen leiden, unter anderem dem Leitsymptom Dyspnoe, sowie Lungenödemen und Flüssigkeitsretention im Körper, insbesondere in den Extremitäten. Zudem treten häufig weitere Begleiterkrankungen auf. Generell wird die Herzinsuffizienz mit einer Kombination von verschiedenen Medikamenten behandelt, meistens ACE-Inhibitoren, Angiotensin 1 Rezeptorantagonisten, Betablockern, Mineralokortikoiden und Diuretika, um die Symptome abzuschwächen. Bei erfolgsloser medikamentöser Therapie und schwerer Linksherzinsuffizienz kann die Implantation eines linksventrikulären Unterstützungssystems (LVAD) die Pumpfunktion des Herzens unterstützen oder die Zeit bis zu einer Herztransplantation unterstützen. Das kardiale oder auch ventrikuläre Remodeling sind geometrische und strukturelle Veränderungen, insbesondere gekennzeichnet durch eine Dilatation der Ventrikel und Veränderungen der Myokardwanddicke, die zu einer Herzinsuffizienz führen können. Die

- 159 - Zusammenfassung

mechanische Druckbelastung des Myokards, z.B. durch Herzmuskel- oder Herzklappenerkrankungen oder Hypertonie, ist ein Faktor in der Entwicklung des kardialen Remodelings und der daraus resultierenden Herzinsuffizienz. Die Implantation eines LVADs kann ein reversibles Remodeling initiieren, indem die kardiale Geometrie, Struktur und Funktion in Folge der Druckentlastung partiell oder vollständig normalisiert werden. Auch eine medikamentöse Therapie wird teilweise mit reparativen myokardialen Umbauprozessen assoziiert. Die pathophysiologischen Mechanismen von Inflammation und Fibrose tragen zu einem kardialem Remodeling bei und sind mit schlechterer Herzfunktion assoziiert. Eine frühe Behandlung von überschießenden inflammatorischen oder fibrotischen Prozessen könnte einem schlechteren Verlauf vorbeugen oder das Fortschreiten der Erkrankung einschränken. Allerdings haben systemische anti-inflammatorische Therapien keine eindeutigen Ergebnisse erzielt und anti-fibrotische Therapien sind klinisch nicht etabliert. Die Identifikation von lokaler Inflammation und Fibrose im Myokard ist zudem erschwert, da Biomarker im Blut keine eindeutige Aussage über die inflammatorischen und fibrotischen Zellzahlen oder den Kollagenanteil im Herzgewebe geben können. Eine Myokardbiopsie ist invasiv und kann einem Stichprobenfehler unterliegen, da die Probe möglicherweise nicht repräsentativ für das gesamte Organ ist. Mit nicht-invasiver molekularer Bildgebung könnten die frühen Pathomechanismen von Inflammation und Fibrose detektiert werden. Diese könnten eine Prognose für den späteren Krankheitsverlauf geben. Darüber hinaus könnten diese Pathomechanismen in Zukunft zielgerichtet behandelt werden, um die Krankheitsentstehung und den -verlauf präventiv zu behandeln oder zu mildern. Das Ziel unserer Studien war es, die molekularen Mechanismen von Inflammation und Fibrose während der Entstehung und Progression von kardialer Drucküberlastung und nach mechanischer Entlastung zu evaluieren. Zudem wollten wir den prognostischen Wert von früh auftretenden inflammatorischen und fibrotischen Prozessen für die später resultierende Herzfunktionseinschränkung bestimmen. Dafür haben wir verschiedene Bildgebungsmarker im Mausmodell der transversen Aortenkonstriktion (TAC) und nach der mechanischen ventrikuläreren Entlastung (rTAC) mittels multimodaler longitudinaler Bildgebung charakterisiert. Die Bildgebungszeitpunkte legten wir anhand publizierter Daten von

- 160 - Zusammenfassung

inflammatorischen Zellzahlen im Myokard, welche mittels Durchflusszytometrie bestimmt wurden, fest. Um die myokardiale Inflammation im akuten Stadium nach kardialer Drucküberlastung zu identifizieren, führten wir serielle Positronen-Emissions-Tomographie (PET) mit dem Radiopharmakon 18F-FDG durch. 18F-FDG wird identisch wie Glukose in eine Vielzahl von Zellen transportiert, z.B. Kardiomyozyten und auch pro-inflammatorische Makrophagen, dort aber nicht weiter verstoffwechselt. Parallel dazu führten wir die Herzfunktionsmessungen mittels Magnetresonanztomographie (MRT) durch. Dabei fanden wir heraus, dass das 18F- FDG PET Signal mit der kardialen Dysfunktion am frühen Zeitpunkt nach drei Tagen korrelierte, aber nicht nach einer Woche. Im weiteren Verlauf erhöhte sich das 18F-FDG Signal disproportional zu der histologisch bestimmten Zelldichte und der inflammatorisch positiven Kontrollgruppe nach Myokardinfarkt. Um die Aufnahme von 18F-FDG in die Kardiomyozyten zu supprimieren und das Signal von inflammatorischen Zellen, insbesondere pro-inflammatorischen Makrophagen, im Myokard zu lokalisieren, wurden die Tiere in Ketamin/Xylazin Anästhesie untersucht. Xylazin supprimiert die Aufnahme von 18F-FDG in die Kardiomyozyten. Ex vivo Untersuchungen identifizierten eine signifikant erhöhte Expression von Glukosetransporter-4 im linken Ventrikel acht Tage nach TAC. Die Ergebnisse dieses Experiments zeigten, dass eine ausreichende Unterdrückung der 18F-FDG Aufnahme mittels Ketamin/Xylazin Anästhesie im drucküberlasteten, gestressten Myokard durch einen veränderten Glukosemetabolismus erschwert ist. Wir konnten daraus schließen, dass das 18F-FDG Signal früh nach TAC Operation Makrophagen im Myokard detektieren kann. Allerdings setzte sich das Signal aus der Stoffwechselaktivität der inflammatorischen Zellen und der gestressten Kardiomyozyten mit verändertem Glukosemetabolismus zusammen. Eine Woche nach TAC ließ sich die 18F-FDG Aufnahme in die Kardiomyozyten schlechter unterdrücken als drei Tage nach TAC, was vermutlich aus deren erhöhter Glukose- Utilisation und verändertem Metabolismus resultierte. Demzufolge führten wir unsere Experimente mit einem spezifischeren PET Tracer zur Detektion von Leukozyten fort. Aufbauend auf unseren Ergebnissen verwendeten wir den Tracer 68Ga-Pentixafor, welcher an den Chemokinrezeptor CXCR4 bindet, um die myokardiale Inflammation in der akuten und chronischen Phase nach Drucküberlastung und Entlastung in einer longitudinalen Studie zu detektieren. CXCR4 wird auf einer Vielzahl von Leukozyten exprimiert und ist somit ein

- 161 - Zusammenfassung

geeigneter Inflammationsmarker. In unserer Studie war das CXCR4 Signal bis drei Wochen nach TAC signifikant erhöht. Die histologische Quantifizierung zeigte zudem signifikant erhöhte CD68+ Zellzahlen, welche den Verlauf der myokardialen CXCR4 Expression widerspiegelten. Das CXCR4 Signal zeigte eine maximale Ausprägung eine Woche nach TAC Operation und korrelierte zu diesem Zeitpunkt mit der histologisch quantifizierten Makrophagendichte. Darüber hinaus war das CXCR4 Signal prädiktiv für eine spätere kardiale Dysfunktion. Zusätzlich führten wir MRT Untersuchungen durch. Mittels T1- Mapping detektierten wir die diffuse myokardiale Fibrose und mittels Cine Messungen der kurzen Achse bestimmten wir die Herzfunktion. Die longitudinale T1 Relaxationszeit war bis sechs Wochen nach TAC verlängert. Dieser Verlauf wurde vom histologisch quantifizierten myokardialen Fibrosegehalt reflektiert. Die T1 Relaxationszeit korrelierte mit der Fibrose und der eingeschränkten Herzfunktion zum jeweiligen Zeitpunkt. Darüber hinaus stellten wir fest, dass sich nach einer mechanischen ventrikulären Entlastung die kardiale Hypertrophie, ventrikuläre Dilatation und Funktion teilweise oder komplett normalisierten, abhängig vom Schweregrad der vorherigen Funktionseinschränkung. Zusätzlich resultierte die Entlastung in einer Abschwächung des CXCR4 Inflammationssignals, einer kürzeren T1 Relaxationszeit und einem deutlich reduziertem Kollagengehalt. Zusammenfassend fanden wir heraus, dass Inflammation und Fibrose früh nach kardialer Drucküberlastung auftraten und eine Woche nach TAC Operation zeitgleich ihren Höhepunkt erreichten. Die Bildgebungsmarker CXCR4 und T1 Relaxationszeit waren prognostisch für eine spätere kardiale Dysfunktion und korrelierten mit den jeweiligen histologischen Parametern. Die mechanische ventrikuläre Entlastung führte zu einer schnellen Normalisierung der Herzgeometrie sowie zu einer Abschwächung der durch die Bildgebung dargestellten inflammatorischen und fibrotischen Mechanismen. Daraus konnten wir schließen, dass eine gezielte frühe Behandlung der untersuchten Mechanismen der weiteren Ausprägung von kardialem Remodeling und Dysfunktion vorbeugen könnte. Mit unseren Ergebnissen konnten wir Zeitpunkte und Zielmechanismen für eine mögliche Behandlung identifizieren. Zukünftig könnten wir mit molekularer Bildgebung den therapeutischen Effekt einer zielgerichteten Behandlung überwachen.

- 162 - References

12 References

ABDEL EL MOTAL, S. M. u. G. W. SHARP (1985): Inhibition of glucose-induced insulin release by xylazine. Endocrinology 116, 2337-2340

AGHAJANIAN, H., T. KIMURA, J. G. RURIK, A. S. HANCOCK, M. S. LEIBOWITZ, L. LI, J. SCHOLLER, J. MONSLOW, A. LO, W. HAN, T. WANG, K. BEDI, M. P. MORLEY, R. A. LINARES SALDANA, N. A. BOLAR, K. MCDAID, C. A. ASSENMACHER, C. L. SMITH, D. WIRTH, C. H. JUNE, K. B. MARGULIES, R. JAIN, E. PURE, S. M. ALBELDA u. J. A. EPSTEIN (2019): Targeting cardiac fibrosis with engineered T cells. Nature 573, 430-433

AMMIRATI, E., C. V. CANNISTRACI, N. A. CRISTELL, V. VECCHIO, A. G. PALINI, P. TORNVALL, A. M. PAGANONI, E. A. MIENDLARZEWSKA, L. M. SANGALLI, A. MONELLO, J. PERNOW, M. BJORNSTEDT BENNERMO, G. MARENZI, D. HU, N. G. UREN, D. CIANFLONE, T. RAVASI, A. A. MANFREDI u. A. MASERI (2012): Identification and predictive value of interleukin-6+ interleukin-10+ and interleukin-6- interleukin-10+ cytokine patterns in ST-elevation acute myocardial infarction. Circ Res 111, 1336-1348

ANDREW, E. R. (1992): Nuclear magnetic resonance and the brain. Brain Topogr 5, 129-133

AYOUB, K. F., N. V. K. POTHINENI, J. RUTLAND, Z. DING u. J. L. MEHTA (2017): Immunity, Inflammation, and Oxidative Stress in Heart Failure: Emerging Molecular Targets. Cardiovasc Drugs Ther 31, 593-608

BACMEISTER, L., M. SCHWARZL, S. WARNKE, B. STOFFERS, S. BLANKENBERG, D. WESTERMANN u. D. LINDNER (2019): Inflammation and fibrosis in murine models of heart failure. Basic Res Cardiol 114, 19

BANERJEE, I., J. W. FUSELER, R. L. PRICE, T. K. BORG u. T. A. BAUDINO (2007): Determination of cell types and numbers during cardiac development in the neonatal and adult rat and mouse. Am J Physiol Heart Circ Physiol 293, H1883-1891

- 163 - References

BARRON, H. V., S. D. HARR, M. J. RADFORD, Y. WANG u. H. M. KRUMHOLZ (2001): The association between white blood cell count and acute myocardial infarction mortality in patients > or =65 years of age: findings from the cooperative cardiovascular project. J Am Coll Cardiol 38, 1654-1661

BENSLEY, J. G., R. DE MATTEO, R. HARDING u. M. J. BLACK (2016): Three-dimensional direct measurement of cardiomyocyte volume, nuclearity, and ploidy in thick histological sections. Sci Rep 6, 23756

BERK, S. L., P. J. BLOCK, P. A. TOSELLI u. W. C. ULLRICK (1975): The effects of chronic alcohol ingestion in mice on contractile properties of cardiac and skeletal muscle: a comparison with normal and dehydrated-malnourished controls. Experientia 31, 1302-1303

BIRKS, E. J., P. D. TANSLEY, J. HARDY, R. S. GEORGE, C. T. BOWLES, M. BURKE, N. R. BANNER, A. KHAGHANI u. M. H. YACOUB (2006): Left ventricular assist device and drug therapy for the reversal of heart failure. N Engl J Med 355, 1873-1884

BLOCH, F. (1953): The Principle of Nuclear Induction. Science 118, 425-430

BLOCK, K. T., M. UECKER u. J. FRAHM (2007): Undersampled radial MRI with multiple coils. Iterative image reconstruction using a total variation constraint. Magn Reson Med 57, 1086-1098

BOHL, S., C. A. LYGATE, H. BARNES, D. MEDWAY, L. A. STORK, J. SCHULZ- MENGER, S. NEUBAUER u. J. E. SCHNEIDER (2009): Advanced methods for quantification of infarct size in mice using three-dimensional high- field late gadolinium enhancement MRI. Am J Physiol Heart Circ Physiol 296, H1200-1208

BOHL, S., R. WASSMUTH, H. ABDEL-ATY, A. RUDOLPH, D. MESSROGHLI, R. DIETZ u. J. SCHULZ-MENGER (2008): Delayed enhancement cardiac magnetic resonance imaging reveals typical patterns of myocardial injury in patients with various forms of non-ischemic heart disease. Int J Cardiovasc Imaging 24, 597-607

- 164 - References

BOOZ, G. W. u. K. M. BAKER (1996): Role of type 1 and type 2 angiotensin receptors in angiotensin II-induced cardiomyocyte hypertrophy. Hypertension 28, 635-640

BORCHERT, T., L. BEITAR, L. B. N. LANGER, A. POLYAK, H. J. WESTER, T. L. ROSS, D. HILFIKER-KLEINER, F. M. BENGEL u. J. T. THACKERAY (2019): Dissecting the target leukocyte subpopulations of clinically relevant inflammation radiopharmaceuticals. J Nucl Cardiol

BOZKURT, B., G. TORRE-AMIONE, M. S. WARREN, J. WHITMORE, O. Z. SORAN, A. M. FELDMAN u. D. L. MANN (2001): Results of targeted anti-tumor necrosis factor therapy with etanercept (ENBREL) in patients with advanced heart failure. Circulation 103, 1044-1047

BRADHAM, W. S., S. P. BELL, D. W. ADKISSON, H. M. SMITH, F. E. HARRELL, JR., M. A. LAWSON, H. OOI, D. B. SAWYER u. M. W. KRONENBERG (2017): Myocardial T1 Measurement Predicts Beneficial LV Remodeling After Long-Term Heart Failure Therapy. J Card Fail 23, 262-265

BRAUNWALD, E. (2013): Heart failure. JACC Heart Fail 1, 1-20

BRECKENRIDGE, R. (2010): Heart failure and mouse models. Dis Model Mech 3, 138-143

BRIASOULIS, A., E. ANDROULAKIS, T. CHRISTOPHIDES u. D. TOUSOULIS (2016): The role of inflammation and cell death in the pathogenesis, progression and treatment of heart failure. Heart Fail Rev 21, 169-176

BUCK, A. K., A. STOLZENBURG, H. HANSCHEID, A. SCHIRBEL, K. LUCKERATH, M. SCHOTTELIUS, H. J. WESTER u. C. LAPA (2017): Chemokine receptor - Directed imaging and therapy. Methods 130, 63-71

- 165 - References

BULL, S., S. K. WHITE, S. K. PIECHNIK, A. S. FLETT, V. M. FERREIRA, M. LOUDON, J. M. FRANCIS, T. D. KARAMITSOS, B. D. PRENDERGAST, M. D. ROBSON, S. NEUBAUER, J. C. MOON u. S. G. MYERSON (2013): Human non-contrast T1 values and correlation with histology in diffuse fibrosis. Heart 99, 932-937

BURT, J. R., S. L. ZIMMERMAN, I. R. KAMEL, M. HALUSHKA u. D. A. BLUEMKE (2014): Myocardial T1 mapping: techniques and potential applications. Radiographics 34, 377-395

CAMELLITI, P., T. K. BORG u. P. KOHL (2005): Structural and functional characterisation of cardiac fibroblasts. Cardiovasc Res 65, 40-51

CHANG, S. C., S. REN, C. D. RAU u. J. J. WANG (2018): Isoproterenol-Induced Heart Failure Mouse Model Using Osmotic Pump Implantation. Methods Mol Biol 1816, 207-220

CHEN, B. u. N. G. FRANGOGIANNIS (2018): The Role of Macrophages in Nonischemic Heart Failure. JACC Basic Transl Sci 3, 245-248

CHEN, Y., Z. YANG, Z. A. ZHAO u. Z. SHEN (2017): Direct reprogramming of fibroblasts into cardiomyocytes. Stem Cell Res Ther 8, 118

CHENG, B., H. C. CHEN, I. W. CHOU, T. W. TANG u. P. C. HSIEH (2017): Harnessing the early post-injury inflammatory responses for cardiac regeneration. J Biomed Sci 24, 7

CHERRY, S. R. u. S. S. GAMBHIR (2001): Use of positron emission tomography in animal research. ILAR J 42, 219-232

CHU, P. Y., M. S. JOSHI, D. HORLOCK, H. KIRIAZIS u. D. M. KAYE (2019): CXCR4 Antagonism Reduces Cardiac Fibrosis and Improves Cardiac Performance in Dilated Cardiomyopathy. Front Pharmacol 10, 117

- 166 - References

CHU, P. Y., K. WALDER, D. HORLOCK, D. WILLIAMS, E. NELSON, M. BYRNE, K. JANDELEIT-DAHM, P. ZIMMET u. D. M. KAYE (2015): CXCR4 Antagonism Attenuates the Development of Diabetic Cardiac Fibrosis. PLoS One 10, e0133616

CHUNG, E. S., M. PACKER, K. H. LO, A. A. FASANMADE, J. T. WILLERSON u. T. N. F. T. A. C. H. F. I. ANTI (2003): Randomized, double-blind, placebo-controlled, pilot trial of infliximab, a chimeric monoclonal antibody to tumor necrosis factor-alpha, in patients with moderate-to-severe heart failure: results of the anti-TNF Therapy Against Congestive Heart Failure (ATTACH) trial. Circulation 107, 3133-3140

DALL'ARMELLINA, E., S. K. PIECHNIK, V. M. FERREIRA, Q. L. SI, M. D. ROBSON, J. M. FRANCIS, F. CUCULI, R. K. KHARBANDA, A. P. BANNING, R. P. CHOUDHURY, T. D. KARAMITSOS u. S. NEUBAUER (2012): Cardiovascular magnetic resonance by non contrast T1-mapping allows assessment of severity of injury in acute myocardial infarction. J Cardiovasc Magn Reson 14, 15

DE BOER, R. A., G. DE KEULENAER, J. BAUERSACHS, D. BRUTSAERT, J. G. CLELAND, J. DIEZ, X. J. DU, P. FORD, F. R. HEINZEL, K. E. LIPSON, T. MCDONAGH, N. LOPEZ-ANDRES, I. G. LUNDE, A. R. LYON, P. POLLESELLO, S. K. PRASAD, C. G. TOCCHETTI, M. MAYR, J. P. G. SLUIJTER, T. THUM, C. TSCHOPE, F. ZANNAD, W. H. ZIMMERMANN, F. RUSCHITZKA, G. FILIPPATOS, M. L. LINDSEY, C. MAACK u. S. HEYMANS (2019): Towards better definition, quantification and treatment of fibrosis in heart failure. A scientific roadmap by the Committee of Translational Research of the Heart Failure Association (HFA) of the European Society of Cardiology. Eur J Heart Fail 21, 272-285

DELGADO, R. M., 3RD, M. A. NAWAR, A. M. ZEWAIL, B. KAR, W. K. VAUGHN, K. K. WU, N. ALEKSIC, N. SIVASUBRAMANIAN, K. MCKAY, D. L. MANN u. J. T. WILLERSON (2004): Cyclooxygenase-2 inhibitor treatment improves left ventricular function and mortality in a murine model of doxorubicin-induced heart failure. Circulation 109, 1428-1433

DICK, S. A. u. S. EPELMAN (2016): Chronic Heart Failure and Inflammation What Do We Really Know? Circulation Research 119, 159-176

- 167 - References

DOLGIN M ASSOCIATION NYHA, F. A., GORLIN R, LEVIN RI, NEW YORK HEART ASSOCIATION. CRITERIA COMMITTEE (1994): Nomenclature and criteria for diagnosis of diseases of the heart and great vessels. 9th ed. Boston, MA: Lippincott Williams and Wilkins March 1, 1994

DORING, Y., L. PAWIG, C. WEBER u. H. NOELS (2014): The CXCL12/CXCR4 chemokine ligand/receptor axis in cardiovascular disease. Front Physiol 5, 212

DUNLAY, S. M., M. M. REDFIELD, S. A. WESTON, T. M. THERNEAU, K. HALL LONG, N. D. SHAH u. V. L. ROGER (2009): Hospitalizations after heart failure diagnosis a community perspective. J Am Coll Cardiol 54, 1695-1702

ELGENDY, I. Y., D. MAHTTA u. C. J. PEPINE (2019): Medical Therapy for Heart Failure Caused by Ischemic Heart Disease. Circ Res 124, 1520-1535

EMDIN, M., A. AIMO, G. VERGARO, A. BAYES-GENIS, J. LUPON, R. LATINI, J. MEESSEN, I. S. ANAND, J. N. COHN, J. GRAVNING, L. GULLESTAD, K. BROCH, T. UELAND, S. H. NYMO, H. P. BRUNNER-LA ROCCA, R. A. DE BOER, H. K. GAGGIN, A. RIPOLI, C. PASSINO u. J. L. JANUZZI, JR. (2018): sST2 Predicts Outcome in Chronic Heart Failure Beyond NT-proBNP and High-Sensitivity Troponin T. J Am Coll Cardiol 72, 2309-2320

EVERETT, R. J., C. G. STIRRAT, S. I. SEMPLE, D. E. NEWBY, M. R. DWECK u. S. MIRSADRAEE (2016): Assessment of myocardial fibrosis with T1 mapping MRI. Clin Radiol 71, 768-778

FAN, Z. u. J. GUAN (2016): Antifibrotic therapies to control cardiac fibrosis. Biomater Res 20, 13

FANG, L., A. J. MURPHY u. A. M. DART (2017): A Clinical Perspective of Anti-Fibrotic Therapies for Cardiovascular Disease. Front Pharmacol 8, 186

- 168 - References

FRANGOGIANNIS, N. G. (2008): The immune system and cardiac repair. Pharmacol Res 58, 88-111

FRANGOGIANNIS, N. G., C. W. SMITH u. M. L. ENTMAN (2002): The inflammatory response in myocardial infarction. Cardiovasc Res 53, 31-47

FRIES, P., A. MASSMANN, R. SEIDEL, A. MULLER, J. STROEDER, F. CUSTODIS, J. REIL, G. SCHNEIDER u. A. BUECKER (2012): Comparison of retrospectively self-gated and prospectively triggered FLASH sequences for cine imaging of the aorta in mice at 9.4 Tesla. Invest Radiol 47, 259-266

FRIGERIO, M. u. E. ROUBINA (2005): Drugs for left ventricular remodeling in heart failure. Am J Cardiol 96, 10L-18L

GANDHI, M. S., G. KAMALOV, A. U. SHAHBAZ, S. K. BHATTACHARYA, R. A. AHOKAS, Y. SUN, I. C. GERLING u. K. T. WEBER (2011): Cellular and molecular pathways to myocardial necrosis and replacement fibrosis. Heart Fail Rev 16, 23-34

GAO, X. M., H. KIRIAZIS, X. L. MOORE, X. H. FENG, K. SHEPPARD, A. DART u. X. J. DU (2005a): Regression of pressure overload-induced left ventricular hypertrophy in mice. Am J Physiol Heart Circ Physiol 288, H2702-2707

GAO, X. M., Q. XU, H. KIRIAZIS, A. M. DART u. X. J. DU (2005b): Mouse model of post-infarct ventricular rupture: time course, strain- and gender-dependency, tensile strength, and histopathology. Cardiovasc Res 65, 469-477

GERDES, A. M. (2002): Cardiac myocyte remodeling in hypertrophy and progression to failure. J Card Fail 8, S264-268

- 169 - References

GERMAIN, P., S. EL GHANNUDI, M. Y. JEUNG, P. OHLMANN, E. EPAILLY, C. ROY u. A. GANGI (2014): Native T1 mapping of the heart - a pictorial review. Clin Med Insights Cardiol 8, 1-11

GIESEL, F. L., C. KRATOCHWIL, T. LINDNER, M. M. MARSCHALEK, A. LOKTEV, W. LEHNERT, J. DEBUS, D. JAGER, P. FLECHSIG, A. ALTMANN, W. MIER u. U. HABERKORN (2019): (68)Ga-FAPI PET/CT: Biodistribution and Preliminary Dosimetry Estimate of 2 DOTA- Containing FAP-Targeting Agents in Patients with Various Cancers. J Nucl Med 60, 386-392

GLASENAPP, A., K. DERLIN, Y. WANG, M. BANKSTAHL, M. MEIER, K. C. WOLLERT, F. M. BENGEL u. J. T. THACKERAY (2019): Multimodality Imaging of Inflammation and Ventricular Remodeling in Pressure Overload Heart Failure. J Nucl Med

GONZALEZ, A., E. B. SCHELBERT, J. DIEZ u. J. BUTLER (2018): Myocardial Interstitial Fibrosis in Heart Failure: Biological and Translational Perspectives. J Am Coll Cardiol 71, 1696-1706

GROSCHEL, C., A. SASSE, S. MONECKE, C. ROHRBORN, L. ELSNER, M. DIDIE, V. REUPKE, G. BUNT, A. H. LICHTMAN, K. TOISCHER, W. H. ZIMMERMANN, G. HASENFUSS u. R. DRESSEL (2018): CD8(+)-T Cells With Specificity for a Model Antigen in Cardiomyocytes Can Become Activated After Transverse Aortic Constriction but Do Not Accelerate Progression to Heart Failure. Front Immunol 9, 2665

HAIDER, N., L. BOSCA, H. R. ZANDBERGEN, J. C. KOVACIC, N. NARULA, S. GONZALEZ-RAMOS, M. FERNANDEZ-VELASCO, S. AGRAWAL, M. PAZ-GARCIA, S. GUPTA, K. DELEON-PENNELL, V. FUSTER, B. IBANEZ u. J. NARULA (2019): Transition of Macrophages to Fibroblast-Like Cells in Healing Myocardial Infarction. J Am Coll Cardiol 74, 3124-3135

HARA, H., N. TAKEDA u. I. KOMURO (2017): Pathophysiology and therapeutic potential of cardiac fibrosis. Inflamm Regen 37, 13

- 170 - References

HESS, A., J. T. THACKERAY, K. C. WOLLERT u. F. M. BENGEL (2019): Radionuclide Image-Guided Repair of the Heart. JACC Cardiovasc Imaging

HEYMANS, S., A. GONZALEZ, A. PIZARD, A. P. PAPAGEORGIOU, N. LOPEZ- ANDRES, F. JAISSER, T. THUM, F. ZANNAD u. J. DIEZ (2015): Searching for new mechanisms of myocardial fibrosis with diagnostic and/or therapeutic potential. Eur J Heart Fail 17, 764-771

HEYMANS, S., E. HIRSCH, S. D. ANKER, P. AUKRUST, J. L. BALLIGAND, J. W. COHEN-TERVAERT, H. DREXLER, G. FILIPPATOS, S. B. FELIX, L. GULLESTAD, D. HILFIKER-KLEINER, S. JANSSENS, R. LATINI, G. NEUBAUER, W. J. PAULUS, B. PIESKE, P. PONIKOWSKI, B. SCHROEN, H. P. SCHULTHEISS, C. TSCHOPE, M. VAN BILSEN, F. ZANNAD, J. MCMURRAY u. A. M. SHAH (2009): Inflammation as a therapeutic target in heart failure? A scientific statement from the Translational Research Committee of the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail 11, 119-129

HIGGINS, C. B., R. HERFKENS, M. J. LIPTON, R. SIEVERS, P. SHELDON, L. KAUFMAN u. L. E. CROOKS (1983): Nuclear magnetic resonance imaging of acute myocardial infarction in dogs: alterations in magnetic relaxation times. Am J Cardiol 52, 184-188

HOFMANN, U. u. S. FRANTZ (2013): How can we cure a heart "in flame"? A translational view on inflammation in heart failure. Basic Res Cardiol 108, 356

HORWICH, T. B., W. R. MACLELLAN u. G. C. FONAROW (2004): Statin therapy is associated with improved survival in ischemic and non-ischemic heart failure. J Am Coll Cardiol 43, 642-648

- 171 - References

HOUSER, S. R., K. B. MARGULIES, A. M. MURPHY, F. G. SPINALE, G. S. FRANCIS, S. D. PRABHU, H. A. ROCKMAN, D. A. KASS, J. D. MOLKENTIN, M. A. SUSSMAN, W. J. KOCH, C. O. C. C. AMERICAN HEART ASSOCIATION COUNCIL ON BASIC CARDIOVASCULAR SCIENCES, G. COUNCIL ON FUNCTIONAL u. B. TRANSLATIONAL (2012): Animal models of heart failure: a scientific statement from the American Heart Association. Circ Res 111, 131-150

HU, X., S. DAI, W. J. WU, W. TAN, X. ZHU, J. MU, Y. GUO, R. BOLLI u. G. ROKOSH (2007): Stromal cell derived factor-1 alpha confers protection against myocardial ischemia/reperfusion injury: role of the cardiac stromal cell derived factor-1 alpha CXCR4 axis. Circulation 116, 654-663

HUGHES, C. E. u. R. J. B. NIBBS (2018): A guide to chemokines and their receptors. FEBS J 285, 2944-2971

HUMERES, C. u. N. G. FRANGOGIANNIS (2019): Fibroblasts in the Infarcted, Remodeling, and Failing Heart. JACC Basic Transl Sci 4, 449-467

ILES, L., H. PFLUGER, A. PHROMMINTIKUL, J. CHERAYATH, P. AKSIT, S. N. GUPTA, D. M. KAYE u. A. J. TAYLOR (2008): Evaluation of diffuse myocardial fibrosis in heart failure with cardiac magnetic resonance contrast-enhanced T1 mapping. J Am Coll Cardiol 52, 1574-1580

ILES, L. M., A. H. ELLIMS, H. LLEWELLYN, J. L. HARE, D. M. KAYE, C. A. MCLEAN u. A. J. TAYLOR (2015): Histological validation of cardiac magnetic resonance analysis of regional and diffuse interstitial myocardial fibrosis. Eur Heart J Cardiovasc Imaging 16, 14-22

INOUE, T., D. PLIETH, C. D. VENKOV, C. XU u. E. G. NEILSON (2005): Antibodies against macrophages that overlap in specificity with fibroblasts. Kidney Int 67, 2488-2493

- 172 - References

IVEY, M. J. u. M. D. TALLQUIST (2016): Defining the Cardiac Fibroblast. Circ J 80, 2269-2276

JAMES, A. E., JR., C. L. PARTAIN, G. N. HOLLAND, J. C. GORE, F. D. ROLLO, S. E. HARMS u. R. R. PRICE (1982): Nuclear magnetic resonance imaging: the current state. AJR Am J Roentgenol 138, 201-210

JUJO, K., H. HAMADA, A. IWAKURA, T. THORNE, H. SEKIGUCHI, T. CLARKE, A. ITO, S. MISENER, T. TANAKA, E. KLYACHKO, K. KOBAYASHI, J. TONGERS, J. RONCALLI, Y. TSURUMI, N. HAGIWARA u. D. W. LOSORDO (2010): CXCR4 blockade augments bone marrow progenitor cell recruitment to the neovasculature and reduces mortality after myocardial infarction. Proc Natl Acad Sci U S A 107, 11008-11013

JUJO, K., M. II, H. SEKIGUCHI, E. KLYACHKO, S. MISENER, T. TANAKA, J. TONGERS, J. RONCALLI, M. A. RENAULT, T. THORNE, A. ITO, T. CLARKE, C. KAMIDE, Y. TSURUMI, N. HAGIWARA, G. QIN, M. ASAHI u. D. W. LOSORDO (2013): CXC-chemokine receptor 4 antagonist AMD3100 promotes cardiac functional recovery after ischemia/reperfusion injury via endothelial nitric oxide synthase-dependent mechanism. Circulation 127, 63-73

KAIN, D., U. AMIT, C. YAGIL, N. LANDA, N. NAFTALI-SHANI, N. MOLOTSKI, V. AVIV, M. S. FEINBERG, O. GOITEIN, T. KUSHNIR, E. KONEN, F. H. EPSTEIN, Y. YAGIL u. J. LEOR (2016): Macrophages dictate the progression and manifestation of hypertensive heart disease. Int J Cardiol 203, 381-395

KALLIKOURDIS, M., E. MARTINI, P. CARULLO, C. SARDI, G. ROSELLI, C. M. GRECO, D. VIGNALI, F. RIVA, A. M. ORMBOSTAD BERRE, T. O. STOLEN, A. FUMERO, G. FAGGIAN, E. DI PASQUALE, L. ELIA, C. RUMIO, D. CATALUCCI, R. PAPAIT u. G. CONDORELLI (2017): T cell costimulation blockade blunts pressure overload-induced heart failure. Nat Commun 8, 14680

KAMEDA, Y., H. HASEGAWA, A. KUBOTA, H. TADOKORO, Y. KOBAYASHI, I. KOMURO u. H. TAKANO (2012): Effects of pitavastatin on pressure overload-induced heart failure in mice. Circ J 76, 1159-1168

- 173 - References

KARWI, Q. G., G. M. UDDIN, K. L. HO u. G. D. LOPASCHUK (2018): Loss of Metabolic Flexibility in the Failing Heart. Front Cardiovasc Med 5, 68

KASS, D. A., K. L. BAUGHMAN, P. H. PAK, P. W. CHO, H. R. LEVIN, T. J. GARDNER, H. R. HALPERIN, J. E. TSITLIK u. M. A. ACKER (1995): Reverse remodeling from cardiomyoplasty in human heart failure. External constraint versus active assist. Circulation 91, 2314-2318

KATO, T. S., A. CHOKSHI, P. SINGH, T. KHAWAJA, F. CHEEMA, H. AKASHI, K. SHAHZAD, S. IWATA, S. HOMMA, H. TAKAYAMA, Y. NAKA, U. JORDE, M. FARR, D. M. MANCINI u. P. C. SCHULZE (2011): Effects of continuous-flow versus pulsatile-flow left ventricular assist devices on myocardial unloading and remodeling. Circ Heart Fail 4, 546-553

KAYA, Z., S. GOSER, S. J. BUSS, F. LEUSCHNER, R. OTTL, J. LI, M. VOLKERS, S. ZITTRICH, G. PFITZER, N. R. ROSE u. H. A. KATUS (2008): Identification of cardiac troponin I sequence motifs leading to heart failure by induction of myocardial inflammation and fibrosis. Circulation 118, 2063-2072

KELLMAN, P. u. M. S. HANSEN (2014): T1-mapping in the heart: accuracy and precision. J Cardiovasc Magn Reson 16, 2

KEMP, C. D. u. J. V. CONTE (2012): The pathophysiology of heart failure. Cardiovasc Pathol 21, 365-371

KIM, R. J., D. S. FIENO, T. B. PARRISH, K. HARRIS, E. L. CHEN, O. SIMONETTI, J. BUNDY, J. P. FINN, F. J. KLOCKE u. R. M. JUDD (1999): Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation 100, 1992-2002

- 174 - References

KJEKSHUS, J., T. R. PEDERSEN, A. G. OLSSON, O. FAERGEMAN u. K. PYORALA (1997): The effects of simvastatin on the incidence of heart failure in patients with coronary heart disease. J Card Fail 3, 249-254

KOCKOVA, R., P. KACER, J. PIRK, J. MALY, L. SUKUPOVA, V. SIKULA, M. KOTRC, L. BARCIAKOVA, E. HONSOVA, M. MALY, J. KAUTZNER, D. SEDMERA u. M. PENICKA (2016): Native T1 Relaxation Time and Extracellular Volume Fraction as Accurate Markers of Diffuse Myocardial Fibrosis in Heart Valve Disease- Comparison With Targeted Left Ventricular Myocardial Biopsy. Circ J 80, 1202-1209

KONG, P., P. CHRISTIA u. N. G. FRANGOGIANNIS (2014): The pathogenesis of cardiac fibrosis. Cell Mol Life Sci 71, 549-574

KONSTAM, M. A., D. G. KRAMER, A. R. PATEL, M. S. MARON u. J. E. UDELSON (2011): Left ventricular remodeling in heart failure: current concepts in clinical significance and assessment. JACC Cardiovasc Imaging 4, 98-108

KURUVILLA, S., N. ADENAW, A. B. KATWAL, M. J. LIPINSKI, C. M. KRAMER u. M. SALERNO (2014): Late gadolinium enhancement on cardiac magnetic resonance predicts adverse cardiovascular outcomes in nonischemic cardiomyopathy: a systematic review and meta-analysis. Circ Cardiovasc Imaging 7, 250-258

LANDMESSER, U., K. C. WOLLERT u. H. DREXLER (2009): Potential novel pharmacological therapies for myocardial remodelling. Cardiovasc Res 81, 519-527

LAPA, C., M. SCHREDER, A. SCHIRBEL, S. SAMNICK, K. M. KORTUM, K. HERRMANN, S. KROPF, H. EINSELE, A. K. BUCK, H. J. WESTER, S. KNOP u. K. LUCKERATH (2017): [(68)Ga]Pentixafor-PET/CT for imaging of chemokine receptor CXCR4 expression in multiple myeloma - Comparison to [(18)F]FDG and laboratory values. Theranostics 7, 205-212

- 175 - References

LAROCCA, T. J., P. ALTMAN, A. A. JARRAH, R. GORDON, E. WANG, L. HADRI, M. W. BURKE, G. E. HADDAD, R. J. HAJJAR u. S. T. TARZAMI (2019): CXCR4 Cardiac Specific Knockout Mice Develop a Progressive Cardiomyopathy. Int J Mol Sci 20,

LAROUMANIE, F., V. DOUIN-ECHINARD, J. POZZO, O. LAIREZ, F. TORTOSA, C. VINEL, C. DELAGE, D. CALISE, M. DUTAUR, A. PARINI u. N. PIZZINAT (2014): CD4+ T cells promote the transition from hypertrophy to heart failure during chronic pressure overload. Circulation 129, 2111-2124

LEE, K. H., B. H. KO, J. Y. PAIK, K. H. JUNG, Y. S. CHOE, Y. CHOI u. B. T. KIM (2005): Effects of anesthetic agents and fasting duration on 18F-FDG biodistribution and insulin levels in tumor-bearing mice. J Nucl Med 46, 1531-1536

LEE, W. W., B. MARINELLI, A. M. VAN DER LAAN, B. F. SENA, R. GORBATOV, F. LEUSCHNER, P. DUTTA, Y. IWAMOTO, T. UENO, M. P. BEGIENEMAN, H. W. NIESSEN, J. J. PIEK, C. VINEGONI, M. J. PITTET, F. K. SWIRSKI, A. TAWAKOL, M. DI CARLI, R. WEISSLEDER u. M. NAHRENDORF (2012): PET/MRI of inflammation in myocardial infarction. J Am Coll Cardiol 59, 153-163

LETO, D. u. A. R. SALTIEL (2012): Regulation of glucose transport by insulin: traffic control of GLUT4. Nat Rev Mol Cell Biol 13, 383-396

LEUSCHNER, F., P. PANIZZI, I. CHICO-CALERO, W. W. LEE, T. UENO, V. CORTEZ- RETAMOZO, P. WATERMAN, R. GORBATOV, B. MARINELLI, Y. IWAMOTO, A. CHUDNOVSKIY, J. L. FIGUEIREDO, D. E. SOSNOVIK, M. J. PITTET, F. K. SWIRSKI, R. WEISSLEDER u. M. NAHRENDORF (2010): Angiotensin-converting enzyme inhibition prevents the release of monocytes from their splenic reservoir in mice with myocardial infarction. Circ Res 107, 1364-1373

LI, X., D. HEBER, T. LEIKE, D. BEITZKE, X. LU, X. ZHANG, Y. WEI, M. MITTERHAUSER, W. WADSAK, S. KROPF, H. J. WESTER, C. LOEWE, M. HACKER u. A. R. HAUG (2018a): [68Ga]Pentixafor-PET/MRI for the detection of Chemokine receptor 4 expression in atherosclerotic plaques. Eur J Nucl Med Mol Imaging 45, 558-566

- 176 - References

LI, X., G. WANG, M. QILI, H. LIANG, T. LI, X. E, Y. FENG, Y. ZHANG, X. LIU, M. QIAN, B. XU, Z. SHEN, S. C. GITAU, D. ZHAO u. H. SHAN (2018b): Aspirin Reduces Cardiac Interstitial Fibrosis by Inhibiting Erk1/2-Serpine2 and P-Akt Signalling Pathways. Cell Physiol Biochem 45, 1955-1965

LIGHTFOOT, J. C., R. B. D'AGOSTINO, JR., C. A. HAMILTON, J. JORDAN, F. M. TORTI, N. D. KOCK, J. JORDAN, S. WORKMAN u. W. G. HUNDLEY (2010): Novel approach to early detection of doxorubicin cardiotoxicity by gadolinium-enhanced cardiovascular magnetic resonance imaging in an experimental model. Circ Cardiovasc Imaging 3, 550-558

LINDNER, D., C. ZIETSCH, J. TANK, S. SOSSALLA, N. FLUSCHNIK, S. HINRICHS, L. MAIER, W. POLLER, S. BLANKENBERG, H. P. SCHULTHEISS, C. TSCHOPE u. D. WESTERMANN (2014): Cardiac fibroblasts support cardiac inflammation in heart failure. Basic Res Cardiol 109, 428

LINDSEY, M. L., R. BOLLI, J. M. CANTY, JR., X. J. DU, N. G. FRANGOGIANNIS, S. FRANTZ, R. G. GOURDIE, J. W. HOLMES, S. P. JONES, R. A. KLONER, D. J. LEFER, R. LIAO, E. MURPHY, P. PING, K. PRZYKLENK, F. A. RECCHIA, L. SCHWARTZ LONGACRE, C. M. RIPPLINGER, J. E. VAN EYK u. G. HEUSCH (2018): Guidelines for experimental models of myocardial ischemia and infarction. Am J Physiol Heart Circ Physiol 314, H812-H838

LOK, S. I., F. M. NOUS, J. VAN KUIK, P. VAN DER WEIDE, B. WINKENS, H. KEMPERMAN, A. HUISMAN, J. R. LAHPOR, R. A. DE WEGER u. N. DE JONGE (2015): Myocardial fibrosis and pro-fibrotic markers in end-stage heart failure patients during continuous-flow left ventricular assist device support. Eur J Cardiothorac Surg 48, 407-415

LUO, T., B. CHEN u. X. WANG (2015): 4-PBA prevents pressure overload-induced myocardial hypertrophy and interstitial fibrosis by attenuating endoplasmic reticulum stress. Chem Biol Interact 242, 99-106

MANN, D. L., P. M. BARGER u. D. BURKHOFF (2012): Myocardial recovery and the failing heart: myth, magic, or molecular target? J Am Coll Cardiol 60, 2465-2472

- 177 - References

MANN, D. L., J. J. MCMURRAY, M. PACKER, K. SWEDBERG, J. S. BORER, W. S. COLUCCI, J. DJIAN, H. DREXLER, A. FELDMAN, L. KOBER, H. KRUM, P. LIU, M. NIEMINEN, L. TAVAZZI, D. J. VAN VELDHUISEN, A. WALDENSTROM, M. WARREN, A. WESTHEIM, F. ZANNAD u. T. FLEMING (2004): Targeted anticytokine therapy in patients with chronic heart failure: results of the Randomized Etanercept Worldwide Evaluation (RENEWAL). Circulation 109, 1594-1602

MARI, C. u. W. H. STRAUSS (2002): Detection and characterization of hibernating myocardium. Nucl Med Commun 23, 311-322

MARINO, A., Y. ZHANG, L. RUBINELLI, M. A. RIEMMA, J. E. IP u. A. DI LORENZO (2019): Pressure overload leads to coronary plaque formation, progression, and myocardial events in ApoE-/- mice. JCI Insight 4

MARTINO, T. A., N. TATA, J. A. SIMPSON, R. VANDERLAAN, F. DAWOOD, M. G. KABIR, N. KHAPER, C. CIFELLI, P. PODOBED, P. P. LIU, M. HUSAIN, S. HEXIMER, P. H. BACKX u. M. J. SOLE (2011): The primary benefits of angiotensin-converting enzyme inhibition on cardiac remodeling occur during sleep time in murine pressure overload hypertrophy. J Am Coll Cardiol 57, 2020-2028

MAZUREK, J. A. u. M. JESSUP (2017): Understanding Heart Failure. Heart Fail Clin 13, 1-19

MCCARTHY, P. M., S. NAKATANI, R. VARGO, K. KOTTKE-MARCHANT, H. HARASAKI, K. B. JAMES, R. M. SAVAGE u. J. D. THOMAS (1995): Structural and left ventricular histologic changes after implantable LVAD insertion. Ann Thorac Surg 59, 609-613

MELLEBY, A. O., A. ROMAINE, J. M. ARONSEN, I. VERAS, L. ZHANG, I. SJAASTAD, I. G. LUNDE u. G. CHRISTENSEN (2018): A novel method for high precision aortic constriction that allows for generation of specific cardiac phenotypes in mice. Cardiovasc Res 114, 1680-1690

- 178 - References

MERINO, D., A. GIL, J. GOMEZ, L. RUIZ, M. LLANO, R. GARCIA, M. A. HURLE u. J. F. NISTAL (2018): Experimental modelling of cardiac pressure overload hypertrophy: Modified technique for precise, reproducible, safe and easy aortic arch banding-debanding in mice. Sci Rep 8, 3167

MESSROGHLI, D. R., S. NORDMEYER, T. DIETRICH, O. DIRSCH, E. KASCHINA, K. SAVVATIS, O. H.-I. D, C. KLEIN, F. BERGER u. T. KUEHNE (2011): Assessment of diffuse myocardial fibrosis in rats using small-animal Look-Locker inversion recovery T1 mapping. Circ Cardiovasc Imaging 4, 636-640

MESSROGHLI, D. R., K. WALTERS, S. PLEIN, P. SPARROW, M. G. FRIEDRICH, J. P. RIDGWAY u. M. U. SIVANANTHAN (2007): Myocardial T1 mapping: application to patients with acute and chronic myocardial infarction. Magn Reson Med 58, 34-40

MICHAEL, L. H., M. L. ENTMAN, C. J. HARTLEY, K. A. YOUKER, J. ZHU, S. R. HALL, H. K. HAWKINS, K. BERENS u. C. M. BALLANTYNE (1995): Myocardial ischemia and reperfusion: a murine model. Am J Physiol 269, H2147-2154

MILLER, L., E. BIRKS, M. GUGLIN, H. LAMBA u. O. H. FRAZIER (2019): Use of Ventricular Assist Devices and Heart Transplantation for Advanced Heart Failure. Circ Res 124, 1658-1678

MOHAMMED, S. F., J. R. STORLIE, E. A. OEHLER, L. A. BOWEN, J. KORINEK, C. S. LAM, R. D. SIMARI, J. C. BURNETT, JR. u. M. M. REDFIELD (2012): Variable phenotype in murine transverse aortic constriction. Cardiovasc Pathol 21, 188-198

MOREO, A., G. AMBROSIO, B. DE CHIARA, M. PU, T. TRAN, F. MAURI u. S. V. RAMAN (2009): Influence of myocardial fibrosis on left ventricular diastolic function: noninvasive assessment by cardiac magnetic resonance and echo. Circ Cardiovasc Imaging 2, 437-443

MOSTERD, A. u. A. W. HOES (2007): Clinical epidemiology of heart failure. Heart 93, 1137-1146

- 179 - References

MOUSSAVI, A., M. UNTENBERGER, M. UECKER u. J. FRAHM (2014): Correction of gradient-induced phase errors in radial MRI. Magn Reson Med 71, 308-312

MULLER, J., G. WALLUKAT, Y. G. WENG, M. DANDEL, S. SPIEGELSBERGER, S. SEMRAU, K. BRANDES, V. THEODORIDIS, M. LOEBE, R. MEYER u. R. HETZER (1997): Weaning from mechanical cardiac support in patients with idiopathic dilated cardiomyopathy. Circulation 96, 542-549

MULLER, P., A. KAZAKOV, P. JAGODA, A. SEMENOV, M. BOHM u. U. LAUFS (2009): ACE inhibition promotes upregulation of endothelial progenitor cells and neoangiogenesis in cardiac pressure overload. Cardiovasc Res 83, 106-114

MURAOKA, N. u. M. IEDA (2014): Direct reprogramming of fibroblasts into myocytes to reverse fibrosis. Annu Rev Physiol 76, 21-37

MURDOCH, C. (2000): CXCR4: chemokine receptor extraordinaire. Immunol Rev 177, 175-184

NAGARAJU, C. K., E. L. ROBINSON, M. ABDESSELEM, S. TRENSON, E. DRIES, G. GILBERT, S. JANSSENS, J. VAN CLEEMPUT, F. REGA, B. MEYNS, H. L. RODERICK, R. B. DRIESEN u. K. R. SIPIDO (2019): Myofibroblast Phenotype and Reversibility of Fibrosis in Patients With End-Stage Heart Failure. J Am Coll Cardiol 73, 2267-2282

NAHRENDORF, M., M. J. PITTET u. F. K. SWIRSKI (2010): Monocytes: protagonists of infarct inflammation and repair after myocardial infarction. Circulation 121, 2437-2445

- 180 - References

NAHRENDORF, M., F. K. SWIRSKI, E. AIKAWA, L. STANGENBERG, T. WURDINGER, J. L. FIGUEIREDO, P. LIBBY, R. WEISSLEDER u. M. J. PITTET (2007): The healing myocardium sequentially mobilizes two monocyte subsets with divergent and complementary functions. J Exp Med 204, 3037-3047

NEKOLLA, S., T. GNEITING, J. SYHA, R. DEICHMANN u. A. HAASE (1992): T1 maps by K-space reduced snapshot-FLASH MRI. J Comput Assist Tomogr 16, 327-332

NEUBAUER, S. (2007): The failing heart--an engine out of fuel. N Engl J Med 356, 1140-1151

OPIE, L. H., P. J. COMMERFORD, B. J. GERSH u. M. A. PFEFFER (2006): Controversies in ventricular remodelling. Lancet 367, 356-367

OSBORN, E. A. u. F. A. JAFFER (2013): The advancing clinical impact of molecular imaging in CVD. JACC Cardiovasc Imaging 6, 1327-1341

PANTEL, A. R. u. D. A. MANKOFF (2017): Molecular imaging to guide systemic cancer therapy: Illustrative examples of PET imaging cancer biomarkers. Cancer Lett 387, 25-31

PATEL, B., S. S. BANSAL, M. A. ISMAHIL, T. HAMID, G. ROKOSH, M. MACK u. S. D. PRABHU (2018): CCR2(+) Monocyte-Derived Infiltrating Macrophages Are Required for Adverse Cardiac Remodeling During Pressure Overload. JACC Basic Transl Sci 3, 230-244

PATEL, B., M. A. ISMAHIL, T. HAMID, S. S. BANSAL u. S. D. PRABHU (2017): Mononuclear Phagocytes Are Dispensable for Cardiac Remodeling in Established Pressure- Overload Heart Failure. PLoS One 12, e0170781

- 181 - References

PATERSON, I., L. M. MIELNICZUK, E. O'MEARA, A. SO u. J. A. WHITE (2013): Imaging heart failure: current and future applications. Can J Cardiol 29, 317-328

PFEFFER, M. A., J. M. PFEFFER, M. C. FISHBEIN, P. J. FLETCHER, J. SPADARO, R. A. KLONER u. E. BRAUNWALD (1979): Myocardial infarct size and ventricular function in rats. Circ Res 44, 503-512

PONIKOWSKI, P., A. A. VOORS, S. D. ANKER, H. BUENO, J. G. CLELAND, A. J. COATS, V. FALK, J. R. GONZALEZ-JUANATEY, V. P. HARJOLA, E. A. JANKOWSKA, M. JESSUP, C. LINDE, P. NIHOYANNOPOULOS, J. T. PARISSIS, B. PIESKE, J. P. RILEY, G. M. ROSANO, L. M. RUILOPE, F. RUSCHITZKA, F. H. RUTTEN, P. VAN DER MEER, M. AUTHORS/TASK FORCE u. R. DOCUMENT (2016): 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail 18, 891-975

PRABHU, S. D. u. N. G. FRANGOGIANNIS (2016): The Biological Basis for Cardiac Repair After Myocardial Infarction: From Inflammation to Fibrosis. Circ Res 119, 91-112

PUNTMANN, V. O., G. CARR-WHITE, A. JABBOUR, C. Y. YU, R. GEBKER, S. KELLE, R. HINOJAR, A. DOLTRA, N. VARMA, N. CHILD, T. ROGERS, G. SUNA, E. ARROYO UCAR, B. GOODMAN, S. KHAN, D. DABIR, E. HERRMANN, A. M. ZEIHER, E. NAGEL u. T. M. C. M. R. O. S. INTERNATIONAL (2016): T1-Mapping and Outcome in Nonischemic Cardiomyopathy: All-Cause Mortality and Heart Failure. JACC Cardiovasc Imaging 9, 40-50

RABI, II, J. R. ZACHARIAS, S. MILLMAN u. P. KUSCH (1992): Milestones in magnetic resonance: 'a new method of measuring nuclear magnetic moment' . 1938. J Magn Reson Imaging 2, 131-133

RAMADAN, A., W. G. LAND u. S. PACZESNY (2017): Editorial: Danger Signals Triggering Immune Response and Inflammation. Front Immunol 8, 979

- 182 - References

RAMPRASAD, M. P., V. TERPSTRA, N. KONDRATENKO, O. QUEHENBERGER u. D. STEINBERG (1996): Cell surface expression of mouse macrosialin and human CD68 and their role as macrophage receptors for oxidized low density lipoprotein. Proc Natl Acad Sci U S A 93, 14833-14838

RECCHIA, F. A. u. V. LIONETTI (2007): Animal models of dilated cardiomyopathy for translational research. Vet Res Commun 31 Suppl 1, 35-41

REGAN, J. A., A. G. MAURO, S. CARBONE, C. MARCHETTI, R. GILL, E. MEZZAROMA, J. VALLE RALEIGH, F. N. SALLOUM, B. W. VAN TASSELL, A. ABBATE u. S. TOLDO (2015): A mouse model of heart failure with preserved ejection fraction due to chronic infusion of a low subpressor dose of angiotensin II. Am J Physiol Heart Circ Physiol 309, H771-778

REITER, T., M. KIRCHER, A. SCHIRBEL, R. A. WERNER, S. KROPF, G. ERTL, A. K. BUCK, H. J. WESTER, W. R. BAUER u. C. LAPA (2018): Imaging of C-X-C Motif Chemokine Receptor CXCR4 Expression After Myocardial Infarction With [(68)Ga]Pentixafor-PET/CT in Correlation With Cardiac MRI. JACC Cardiovasc Imaging 11, 1541-1543

RICHARDS, D. A., M. J. ARONOVITZ, T. D. CALAMARAS, K. TAM, G. L. MARTIN, P. LIU, H. K. BOWDITCH, P. ZHANG, G. S. HUGGINS u. R. M. BLANTON (2019): Distinct Phenotypes Induced by Three Degrees of Transverse Aortic Constriction in Mice. Sci Rep 9, 5844

RISCHPLER, C., R. J. DIRSCHINGER, S. G. NEKOLLA, H. KOSSMANN, S. NICOLOSI, F. HANUS, S. VAN MARWICK, K. P. KUNZE, A. MEINICKE, K. GOTZE, A. KASTRATI, N. LANGWIESER, T. IBRAHIM, M. NAHRENDORF, M. SCHWAIGER u. K. L. LAUGWITZ (2016): Prospective Evaluation of 18F-Fluorodeoxyglucose Uptake in Postischemic Myocardium by Simultaneous Positron Emission Tomography/Magnetic Resonance Imaging as a Prognostic Marker of Functional Outcome. Circ Cardiovasc Imaging 9, e004316

- 183 - References

RISCHPLER, C., N. LANGWIESER, M. SOUVATZOGLOU, A. BATRICE, S. VAN MARWICK, J. SNAJBERK, T. IBRAHIM, K. L. LAUGWITZ, S. G. NEKOLLA u. M. SCHWAIGER (2015): PET/MRI early after myocardial infarction: evaluation of viability with late gadolinium enhancement transmurality vs. 18F-FDG uptake. Eur Heart J Cardiovasc Imaging 16, 661-669

ROCKMAN, H. A., R. S. ROSS, A. N. HARRIS, K. U. KNOWLTON, M. E. STEINHELPER, L. J. FIELD, J. ROSS, JR. u. K. R. CHIEN (1991): Segregation of atrial-specific and inducible expression of an atrial natriuretic factor transgene in an in vivo murine model of cardiac hypertrophy. Proc Natl Acad Sci U S A 88, 8277-8281

ROCKMAN, H. A., S. P. WACHHORST, L. MAO u. J. ROSS, JR. (1994): ANG II receptor blockade prevents ventricular hypertrophy and ANF gene expression with pressure overload in mice. Am J Physiol 266, H2468-2475

ROSSIGNOL, P., A. F. HERNANDEZ, S. D. SOLOMON u. F. ZANNAD (2019): Heart failure drug treatment. Lancet 393, 1034-1044

ROSZER, T. (2015): Understanding the Mysterious M2 Macrophage through Activation Markers and Effector Mechanisms. Mediators Inflamm 2015, 816460

SARASTE, A. u. J. KNUUTI (2017): PET imaging in heart failure: the role of new tracers. Heart Fail Rev 22, 501-511

SARIKAYA, I. (2015): Cardiac applications of PET. Nucl Med Commun 36, 971-985

SATO, S. u. R. C. HUGHES (1994): Control of Mac-2 surface expression on murine macrophage cell lines. Eur J Immunol 24, 216-221

- 184 - References

SATOMI, T., M. OGAWA, I. MORI, S. ISHINO, K. KUBO, Y. MAGATA u. T. NISHIMOTO (2013): Comparison of contrast agents for atherosclerosis imaging using cultured macrophages: FDG versus ultrasmall superparamagnetic iron oxide. J Nucl Med 54, 999-1004

SCHIPKE, J., C. BRANDENBERGER, A. RAJCES, M. MANNINGER, A. ALOGNA, H. POST u. C. MUHLFELD (2017): Assessment of cardiac fibrosis: a morphometric method comparison for collagen quantification. J Appl Physiol (1985) 122, 1019-1030

SCHIRONE, L., M. FORTE, S. PALMERIO, D. YEE, C. NOCELLA, F. ANGELINI, F. PAGANO, S. SCHIAVON, A. BORDIN, A. CARRIZZO, C. VECCHIONE, V. VALENTI, I. CHIMENTI, E. DE FALCO, S. SCIARRETTA u. G. FRATI (2017): A Review of the Molecular Mechanisms Underlying the Development and Progression of Cardiac Remodeling. Oxid Med Cell Longev 2017, 3920195

SEGURA, A. M., O. H. FRAZIER u. L. M. BUJA (2014): Fibrosis and heart failure. Heart Fail Rev 19, 173-185

SEN, S., B. K. KUNDU, H. C. WU, S. S. HASHMI, P. GUTHRIE, L. W. LOCKE, R. J. ROY, G. P. MATHERNE, S. S. BERR, M. TERWELP, B. SCOTT, S. CARRANZA, O. H. FRAZIER, D. K. GLOVER, W. H. DILLMANN, M. J. GAMBELLO, M. L. ENTMAN u. H. TAEGTMEYER (2013): Glucose regulation of load-induced mTOR signaling and ER stress in mammalian heart. J Am Heart Assoc 2, e004796

SHINDE, A. V. u. N. G. FRANGOGIANNIS (2014): Fibroblasts in myocardial infarction: a role in inflammation and repair. J Mol Cell Cardiol 70, 74-82

SHINOHARA, K., T. KISHI, Y. HIROOKA u. K. SUNAGAWA (2015): Circulating angiotensin II deteriorates left ventricular function with sympathoexcitation via brain angiotensin II receptor. Physiol Rep 3

- 185 - References

SHIRAZI, L. F., J. BISSETT, F. ROMEO u. J. L. MEHTA (2017): Role of Inflammation in Heart Failure. Curr Atheroscler Rep 19, 27

SIMON, M. M., S. GREENAWAY, J. K. WHITE, H. FUCHS, V. GAILUS-DURNER, S. WELLS, T. SORG, K. WONG, E. BEDU, E. J. CARTWRIGHT, R. DACQUIN, S. DJEBALI, J. ESTABEL, J. GRAW, N. J. INGHAM, I. J. JACKSON, A. LENGELING, S. MANDILLO, J. MARVEL, H. MEZIANE, F. PREITNER, O. PUK, M. ROUX, D. J. ADAMS, S. ATKINS, A. AYADI, L. BECKER, A. BLAKE, D. BROOKER, H. CATER, M. F. CHAMPY, R. COMBE, P. DANECEK, A. DI FENZA, H. GATES, A. K. GERDIN, E. GOLINI, J. M. HANCOCK, W. HANS, S. M. HOLTER, T. HOUGH, P. JURDIC, T. M. KEANE, H. MORGAN, W. MULLER, F. NEFF, G. NICHOLSON, B. PASCHE, L. A. ROBERSON, J. ROZMAN, M. SANDERSON, L. SANTOS, M. SELLOUM, C. SHANNON, A. SOUTHWELL, G. P. TOCCHINI-VALENTINI, V. E. VANCOLLIE, H. WESTERBERG, W. WURST, M. ZI, B. YALCIN, R. RAMIREZ-SOLIS, K. P. STEEL, A. M. MALLON, M. H. DE ANGELIS, Y. HERAULT u. S. D. BROWN (2013): A comparative phenotypic and genomic analysis of C57BL/6J and C57BL/6N mouse strains. Genome Biol 14, R82

SKALI, H., A. R. SCHULMAN u. S. DORBALA (2013): (18)F-FDG PET/CT for the assessment of myocardial sarcoidosis. Curr Cardiol Rep 15, 352

SMITH, S. C. u. P. M. ALLEN (1991): Myosin-induced acute myocarditis is a T cell-mediated disease. J Immunol 147, 2141-2147

SNIDER, P., K. N. STANDLEY, J. WANG, M. AZHAR, T. DOETSCHMAN u. S. J. CONWAY (2009): Origin of cardiac fibroblasts and the role of periostin. Circ Res 105, 934-947

SONG, K., Y. J. NAM, X. LUO, X. QI, W. TAN, G. N. HUANG, A. ACHARYA, C. L. SMITH, M. D. TALLQUIST, E. G. NEILSON, J. A. HILL, R. BASSEL-DUBY u. E. N. OLSON (2012): Heart repair by reprogramming non-myocytes with cardiac transcription factors. Nature 485, 599-604

- 186 - References

SRIRAMULA, S. u. J. FRANCIS (2015): Tumor Necrosis Factor - Alpha Is Essential for Angiotensin II-Induced Ventricular Remodeling: Role for Oxidative Stress. PLoS One 10, e0138372

STROHM, O., J. SCHULZ-MENGER, B. PILZ, K. J. OSTERZIEL, R. DIETZ u. M. G. FRIEDRICH (2001): Measurement of left ventricular dimensions and function in patients with dilated cardiomyopathy. J Magn Reson Imaging 13, 367-371

STUCKEY, D. J., C. A. CARR, P. CAMELLITI, D. J. TYLER, K. E. DAVIES u. K. CLARKE (2012): In vivo MRI characterization of progressive cardiac dysfunction in the mdx mouse model of muscular dystrophy. PLoS One 7, e28569

STUCKEY, D. J., S. J. MCSWEENEY, M. Z. THIN, J. HABIB, A. N. PRICE, L. R. FIEDLER, W. GSELL, S. K. PRASAD u. M. D. SCHNEIDER (2014): T(1) mapping detects pharmacological retardation of diffuse cardiac fibrosis in mouse pressure-overload hypertrophy. Circ Cardiovasc Imaging 7, 240-249

SUINESIAPUTRA, A., M. M. SANGHVI, N. AUNG, J. M. PAIVA, F. ZEMRAK, K. FUNG, E. LUKASCHUK, A. M. LEE, V. CARAPELLA, Y. J. KIM, J. FRANCIS, S. K. PIECHNIK, S. NEUBAUER, A. GREISER, M. P. JOLLY, C. HAYES, A. A. YOUNG u. S. E. PETERSEN (2018): Fully-automated left ventricular mass and volume MRI analysis in the UK Biobank population cohort: evaluation of initial results. Int J Cardiovasc Imaging 34, 281-291

TAEGTMEYER, H., M. E. YOUNG, G. D. LOPASCHUK, E. D. ABEL, H. BRUNENGRABER, V. DARLEY-USMAR, C. DES ROSIERS, R. GERSZTEN, J. F. GLATZ, J. L. GRIFFIN, R. J. GROPLER, H. G. HOLZHUETTER, J. R. KIZER, E. D. LEWANDOWSKI, C. R. MALLOY, S. NEUBAUER, L. R. PETERSON, M. A. PORTMAN, F. A. RECCHIA, J. E. VAN EYK, T. J. WANG u. S. AMERICAN HEART ASSOCIATION COUNCIL ON BASIC CARDIOVASCULAR (2016): Assessing Cardiac Metabolism: A Scientific Statement From the American Heart Association. Circ Res 118, 1659-1701

- 187 - References

TALMAN, V. u. H. RUSKOAHO (2016): Cardiac fibrosis in myocardial infarction-from repair and remodeling to regeneration. Cell Tissue Res 365, 563-581

TAMAKI, N., K. MORITA, Y. KUGE u. E. TSUKAMOTO (2000): The role of fatty acids in cardiac imaging. J Nucl Med 41, 1525-1534

TANG, R., J. T. WANG, L. WANG, K. LE, Y. HUANG, A. J. HICKEY u. L. EMMETT (2016): Impact of Patient Preparation on the Diagnostic Performance of 18F-FDG PET in Cardiac Sarcoidosis: A Systematic Review and Meta-analysis. Clin Nucl Med 41, e327-339

TARKIN, J. M., F. R. JOSHI u. J. H. RUDD (2014): PET imaging of inflammation in atherosclerosis. Nat Rev Cardiol 11, 443-457

TARNAVSKI, O., J. R. MCMULLEN, M. SCHINKE, Q. NIE, S. KONG u. S. IZUMO (2004): Mouse cardiac surgery: comprehensive techniques for the generation of mouse models of human diseases and their application for genomic studies. Physiol Genomics 16, 349-360

TAVAKOLI, R., S. NEMSKA, P. JAMSHIDI, M. GASSMANN u. N. FROSSARD (2017): Technique of Minimally Invasive Transverse Aortic Constriction in Mice for Induction of Left Ventricular Hypertrophy. J Vis Exp TER-POGOSSIAN, M. M., M. E. PHELPS, E. J. HOFFMAN u. N. A. MULLANI (1975): A positron-emission transaxial tomograph for nuclear imaging (PETT). Radiology 114, 89-98

THACKERAY, J. T., J. P. BANKSTAHL, Y. WANG, K. C. WOLLERT u. F. M. BENGEL (2015a): Clinically relevant strategies for lowering cardiomyocyte glucose uptake for 18F-FDG imaging of myocardial inflammation in mice. Eur J Nucl Med Mol Imaging 42, 771-780

- 188 - References

THACKERAY, J. T. u. F. M. BENGEL (2016): Translational Molecular Nuclear Cardiology. Cardiol Clin 34, 187-198

THACKERAY, J. T. u. F. M. BENGEL (2018): Molecular Imaging of Myocardial Inflammation With Positron Emission Tomography Post- Ischemia: A Determinant of Subsequent Remodeling or Recovery. JACC Cardiovasc Imaging 11, 1340-1355

THACKERAY, J. T., T. DERLIN, A. HAGHIKIA, L. C. NAPP, Y. WANG, T. L. ROSS, A. SCHAFER, J. TILLMANNS, H. J. WESTER, K. C. WOLLERT, J. BAUERSACHS u. F. M. BENGEL (2015b): Molecular Imaging of the Chemokine Receptor CXCR4 After Acute Myocardial Infarction. JACC Cardiovasc Imaging 8, 1417-1426

THACKERAY, J. T., M. PARSA-NEZHAD, M. KENK, S. L. THORN, M. KOLAJOVA, R. S. BEANLANDS u. J. N. DASILVA (2011): Reduced CGP12177 binding to cardiac beta-adrenoceptors in hyperglycemic high-fat-diet- fed, streptozotocin-induced diabetic rats. Nucl Med Biol 38, 1059-1066

THAM, Y. K., B. C. BERNARDO, J. Y. OOI, K. L. WEEKS u. J. R. MCMULLEN (2015): Pathophysiology of cardiac hypertrophy and heart failure: signaling pathways and novel therapeutic targets. Arch Toxicol 89, 1401-1438

TILLMANNS, J., D. HOFFMANN, Y. HABBABA, J. D. SCHMITTO, D. SEDDING, D. FRACCAROLLO, P. GALUPPO u. J. BAUERSACHS (2015): Fibroblast activation protein alpha expression identifies activated fibroblasts after myocardial infarction. J Mol Cell Cardiol 87, 194-203

TOISCHER, K., A. G. ROKITA, B. UNSOLD, W. ZHU, G. KARARIGAS, S. SOSSALLA, S. P. REUTER, A. BECKER, N. TEUCHER, T. SEIDLER, C. GREBE, L. PREUSS, S. N. GUPTA, K. SCHMIDT, S. E. LEHNART, M. KRUGER, W. A. LINKE, J. BACKS, V. REGITZ-ZAGROSEK, K. SCHAFER, L. J. FIELD, L. S. MAIER u. G. HASENFUSS (2010): Differential cardiac remodeling in preload versus afterload. Circulation 122, 993-1003

- 189 - References

TORRE-AMIONE, G., S. KAPADIA, C. BENEDICT, H. ORAL, J. B. YOUNG u. D. L. MANN (1996): Proinflammatory cytokine levels in patients with depressed left ventricular ejection fraction: a report from the Studies of Left Ventricular Dysfunction (SOLVD). J Am Coll Cardiol 27, 1201-1206

TRAVERS, J. G., F. A. KAMAL, J. ROBBINS, K. E. YUTZEY u. B. C. BLAXALL (2016): Cardiac Fibrosis: The Fibroblast Awakens. Circ Res 118, 1021-1040

UMBARAWAN, Y., M. SYAMSUNARNO, N. KOITABASHI, A. YAMAGUCHI, H. HANAOKA, T. HISHIKI, Y. NAGAHATA-NAITO, H. OBINATA, M. SANO, H. SUNAGA, H. MATSUI, Y. TSUSHIMA, M. SUEMATSU, M. KURABAYASHI u. T. ISO (2018): Glucose is preferentially utilized for biomass synthesis in pressure-overloaded hearts: evidence from fatty acid-binding protein-4 and -5 knockout mice. Cardiovasc Res 114, 1132-1144

VAN DER SPOEL, T. I., S. J. JANSEN OF LORKEERS, P. AGOSTONI, E. VAN BELLE, M. GYONGYOSI, J. P. SLUIJTER, M. J. CRAMER, P. A. DOEVENDANS u. S. A. CHAMULEAU (2011): Human relevance of pre-clinical studies in stem cell therapy: systematic review and meta- analysis of large animal models of ischaemic heart disease. Cardiovasc Res 91, 649-658

VAN ROOIJEN, N. u. E. HENDRIKX (2010): Liposomes for specific depletion of macrophages from organs and tissues. Methods Mol Biol 605, 189-203

VAN ROOIJEN, N. u. A. SANDERS (1994): Liposome mediated depletion of macrophages: mechanism of action, preparation of liposomes and applications. J Immunol Methods 174, 83-93

VARASTEH, Z., S. MOHANTA, S. ROBU, M. BRAEUER, Y. LI, N. OMIDVARI, G. TOPPING, T. SUN, S. G. NEKOLLA, A. RICHTER, C. WEBER, A. HABENICHT, U. A. HABERKORN u. W. A. WEBER (2019): Molecular Imaging of Fibroblast Activity After Myocardial Infarction Using a (68)Ga- Labeled Fibroblast Activation Protein Inhibitor, FAPI-04. J Nucl Med 60, 1743-1749

- 190 - References

VERMA, A., A. MERIS, H. SKALI, J. K. GHALI, J. M. ARNOLD, M. BOURGOUN, E. J. VELAZQUEZ, J. J. MCMURRAY, L. KOBER, M. A. PFEFFER, R. M. CALIFF u. S. D. SOLOMON (2008): Prognostic implications of left ventricular mass and geometry following myocardial infarction: the VALIANT (VALsartan In Acute myocardial iNfarcTion) Echocardiographic Study. JACC Cardiovasc Imaging 1, 582-591

VISSER, E. P., J. A. DISSELHORST, M. BROM, P. LAVERMAN, M. GOTTHARDT, W. J. OYEN u. O. C. BOERMAN (2009): Spatial resolution and sensitivity of the Inveon small-animal PET scanner. J Nucl Med 50, 139-147

VOGEL-CLAUSSEN, J., C. E. ROCHITTE, K. C. WU, I. R. KAMEL, T. K. FOO, J. A. LIMA u. D. A. BLUEMKE (2006): Delayed enhancement MR imaging: utility in myocardial assessment. Radiographics 26, 795-810

WANG, E. R., A. A. JARRAH, L. BENARD, J. CHEN, M. SCHWARZKOPF, L. HADRI u. S. T. TARZAMI (2014): Deletion of CXCR4 in cardiomyocytes exacerbates cardiac dysfunction following isoproterenol administration. Gene Ther 21, 496-506

WANG, H., L. HOU, D. KWAK, J. FASSETT, X. XU, A. CHEN, W. CHEN, B. R. BLAZAR, Y. XU, J. L. HALL, J. B. GE, R. J. BACHE u. Y. CHEN (2016): Increasing Regulatory T Cells With Interleukin-2 and Interleukin-2 Antibody Complexes Attenuates Lung Inflammation and Heart Failure Progression. Hypertension 68, 114-122

WANG, Y., K. DEMBOWSKY, E. CHEVALIER, P. STUVE, M. KORF-KLINGEBIEL, M. LOCHNER, L. C. NAPP, H. FRANK, E. BRINKMANN, A. KANWISCHER, J. BAUERSACHS, M. GYONGYOSI, T. SPARWASSER u. K. C. WOLLERT (2019): C-X-C Motif Chemokine Receptor 4 Blockade Promotes Tissue Repair After Myocardial Infarction by Enhancing Regulatory T Cell Mobilization and Immune-Regulatory Function. Circulation 139, 1798-1812

WANG, Y., Y. WU, J. CHEN, S. ZHAO u. H. LI (2013): Pirfenidone attenuates cardiac fibrosis in a mouse model of TAC-induced left ventricular remodeling by suppressing NLRP3 inflammasome formation. Cardiology 126, 1-11

- 191 - References

WEBER, K. T. (1989): Cardiac interstitium in health and disease: the fibrillar collagen network. J Am Coll Cardiol 13, 1637-1652

WEBER, K. T. u. C. G. BRILLA (1991): Pathological hypertrophy and cardiac interstitium. Fibrosis and renin-angiotensin-aldosterone system. Circulation 83, 1849-1865

WEBER, K. T., Y. SUN, S. K. BHATTACHARYA, R. A. AHOKAS u. I. C. GERLING (2013): Myofibroblast-mediated mechanisms of pathological remodelling of the heart. Nat Rev Cardiol 10, 15-26

WECH, T., J. TRAN-GIA, F. RUTZEL, T. KLINK, T. A. BLEY u. H. KOSTLER (2018): Single-shot late Gd enhancement imaging of myocardial infarction with retrospectively adjustable contrast and heart-phase. Magn Reson Imaging 47, 48-53

WEINHEIMER, C. J., A. KOVACS, S. EVANS, S. J. MATKOVICH, P. M. BARGER u. D. L. MANN (2018): Load-Dependent Changes in Left Ventricular Structure and Function in a Pathophysiologically Relevant Murine Model of Reversible Heart Failure. Circ Heart Fail 11, e004351

WEINHEIMER, C. J., L. LAI, D. P. KELLY u. A. KOVACS (2015): Novel mouse model of left ventricular pressure overload and infarction causing predictable ventricular remodelling and progression to heart failure. Clin Exp Pharmacol Physiol 42, 33-40

WEISHEIT, C., Y. ZHANG, A. FARON, O. KOPKE, G. WEISHEIT, A. STEINSTRASSER, S. FREDE, R. MEYER, O. BOEHM, A. HOEFT, C. KURTS u. G. BAUMGARTEN (2014): Ly6C(low) and not Ly6C(high) macrophages accumulate first in the heart in a model of murine pressure-overload. PLoS One 9, e112710

WENDE, A. R., M. K. BRAHMA, G. R. MCGINNIS u. M. E. YOUNG (2017): Metabolic Origins of Heart Failure. JACC Basic Transl Sci 2, 297-310

- 192 - References

WESTER, H. J., U. KELLER, M. SCHOTTELIUS, A. BEER, K. PHILIPP-ABBREDERIS, F. HOFFMANN, J. SIMECEK, C. GERNGROSS, M. LASSMANN, K. HERRMANN, N. PELLEGATA, M. RUDELIUS, H. KESSLER u. M. SCHWAIGER (2015): Disclosing the CXCR4 expression in lymphoproliferative diseases by targeted molecular imaging. Theranostics 5, 618-630

WESTERMANN, D., P. M. BECHER, D. LINDNER, K. SAVVATIS, Y. XIA, M. FROHLICH, S. HOFFMANN, H. P. SCHULTHEISS u. C. TSCHOPE (2012): Selective PDE5A inhibition with sildenafil rescues left ventricular dysfunction, inflammatory immune response and cardiac remodeling in angiotensin II-induced heart failure in vivo. Basic Res Cardiol 107, 308

WILDGRUBER, M., F. K. SWIRSKI u. A. ZERNECKE (2013): Molecular imaging of inflammation in atherosclerosis. Theranostics 3, 865-884

WINDELOV, J. A., J. PEDERSEN u. J. J. HOLST (2016): Use of anesthesia dramatically alters the oral glucose tolerance and insulin secretion in C57Bl/6 mice. Physiol Rep 4,

WINTER, P., T. KAMPF, X. HELLUY, F. T. GUTJAHR, C. B. MEYER, W. R. BAUER, P. M. JAKOB u. V. HEROLD (2016): Self-navigation under non-steady-state conditions: Cardiac and respiratory self-gating of inversion recovery snapshot FLASH acquisitions in mice. Magn Reson Med 76, 1887-1894

WISENBERG, G., F. S. PRATO, S. E. CARROLL, K. L. TURNER u. T. MARSHALL (1988): Serial nuclear magnetic resonance imaging of acute myocardial infarction with and without reperfusion. Am Heart J 115, 510-518

WOHLSCHLAEGER, J., K. J. SCHMITZ, C. SCHMID, K. W. SCHMID, P. KEUL, A. TAKEDA, S. WEIS, B. LEVKAU u. H. A. BABA (2005): Reverse remodeling following insertion of left ventricular assist devices (LVAD): a review of the morphological and molecular changes. Cardiovasc Res 68, 376-386

- 193 - References

WU, C. K. Y. (2016): Chapter 13 The Clinical Value of Cardiac PET in Heart Failure. In: Perspectives on Nuclear Medicine for Molecular Diagnosis and Integrated Therapy. Springer, Tokyo

WYNN, T. A. u. T. R. RAMALINGAM (2012): Mechanisms of fibrosis: therapeutic translation for fibrotic disease. Nat Med 18, 1028-1040

XIA, Y., K. LEE, N. LI, D. CORBETT, L. MENDOZA u. N. G. FRANGOGIANNIS (2009): Characterization of the inflammatory and fibrotic response in a mouse model of cardiac pressure overload. Histochem Cell Biol 131, 471-481

YANCY, C. W., M. JESSUP, B. BOZKURT, J. BUTLER, D. E. CASEY, JR., M. H. DRAZNER, G. C. FONAROW, S. A. GERACI, T. HORWICH, J. L. JANUZZI, M. R. JOHNSON, E. K. KASPER, W. C. LEVY, F. A. MASOUDI, P. E. MCBRIDE, J. J. MCMURRAY, J. E. MITCHELL, P. N. PETERSON, B. RIEGEL, F. SAM, L. W. STEVENSON, W. H. TANG, E. J. TSAI, B. L. WILKOFF, F. AMERICAN COLLEGE OF CARDIOLOGY u. G. AMERICAN HEART ASSOCIATION TASK FORCE ON PRACTICE (2013): 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 62, e147-239

YIN, F. C., H. A. SPURGEON, K. RAKUSAN, M. L. WEISFELDT u. E. G. LAKATTA (1982): Use of tibial length to quantify cardiac hypertrophy: application in the aging rat. Am J Physiol 243, H941-947

YOO, J., V. CHEPURKO, R. J. HAJJAR u. D. JEONG (2018): Conventional Method of Transverse Aortic Constriction in Mice. Methods Mol Biol 1816, 183-193

YOSHINAGA, K. u. N. TAMAKI (2007): Imaging myocardial metabolism. Curr Opin Biotechnol 18, 52-59

- 194 - References

YOU, J., J. WU, Q. ZHANG, Y. YE, S. WANG, J. HUANG, H. LIU, X. WANG, W. ZHANG, L. BU, J. LI, L. LIN, J. GE u. Y. ZOU (2018): Differential cardiac hypertrophy and signaling pathways in pressure versus volume overload. Am J Physiol Heart Circ Physiol 314, H552-H562

ZAW, A. M., C. M. WILLIAMS, H. K. LAW u. B. K. CHOW (2017): Minimally Invasive Transverse Aortic Constriction in Mice. J Vis Exp

ZHANG, X., H. JAVAN, L. LI, A. SZUCSIK, R. ZHANG, Y. DENG u. C. H. SELZMAN (2013): A modified murine model for the study of reverse cardiac remodelling. Exp Clin Cardiol 18, e115-117

ZHANG, Y., L. ZHANG, X. FAN, W. YANG, B. YU, J. KOU u. F. LI (2019): Captopril attenuates TAC-induced heart failure via inhibiting Wnt3a/beta-catenin and Jak2/Stat3 pathways. Biomed Pharmacother 113, 108780

ZHAO, L., G. CHENG, R. JIN, M. R. AFZAL, A. SAMANTA, Y. T. XUAN, M. GIRGIS, H. K. ELIAS, Y. ZHU, A. DAVANI, Y. YANG, X. CHEN, S. YE, O. L. WANG, L. CHEN, J. HAUPTMAN, R. J. VINCENT u. B. DAWN (2016): Deletion of Interleukin-6 Attenuates Pressure Overload-Induced Left Ventricular Hypertrophy and Dysfunction. Circ Res 118, 1918-1929

ZIAEIAN, B. u. G. C. FONAROW (2016): Epidemiology and aetiology of heart failure. Nat Rev Cardiol 13, 368-378

ZUO, Z., A. SUBGANG, A. ABAEI, W. ROTTBAUER, D. STILLER, G. MA u. V. RASCHE (2017): Assessment of Longitudinal Reproducibility of Mice LV Function Parameters at 11.7 T Derived from Self-Gated CINE MRI. Biomed Res Int 2017, 8392952

- 195 - Acknowledgments

13 Acknowledgments

Finally, I would like to express my sincere gratitude to the following people, without whom I would not have been able to complete this research.

First, I would like to thank my supervisors

Prof. Dr. Marion Bankstahl, veterinary specialist for pharmacology and toxicology in the Central Animal Facility and Institute for Laboratory Animal Science, Prof. Dr. Frank M. Bengel, head of the Department of Nuclear Medicine, PD Dr. Katja Derlin, senior physician in the Department of Radiology, and James T. Thackeray, PhD, research group leader Translational Cardiovascular Molecular Imaging in the Department of Nuclear Medicine; Hannover Medical School,

for providing me the opportunity to work on this exciting project, for excellent support, valuable feedback, continuous guidance, and encouragement through each stage of the project.

I thank the German Research Foundation (DFG, Clinical Research Group KFO311), Excellence Cluster REBIRTH-2, and research grant TH2161/1-1for funding this project.

I thank Prof. Dr. Frank Wacker, head of the Department of Radiology, for providing me the opportunity to work on this exciting project. Many thanks go to all current and former colleagues in the Experimental Radiology research group of the Department of Radiology, especially Lea Behrendt, Dr. Marcel Gutberlet, Arnd Obert, and Dr. Hinrich Winther for their help and support.

I would like to thank all current and former colleagues in the Preclinical Molecular Imaging research group of the Department of Nuclear Medicine. Many thanks particularly to Annika Hess for her help and support during my experiments. Moreover, I would like to thank Pablo Bascuñana, PhD, Dr. Tobias Borchert, Nele Hermanns, Dr. Ina Jahreis, Dr. Mariella Kessler, Dr. Rudolf Werner, and Dr. Bettina Wolf for their help and support.

I would like to thank the team of the Preclinical Molecular Imaging Department of the Department of Nuclear Medicine, especially Daniel Ahrens, Dr. Jens Bankstahl, Silvia Eilert,

- 196 - Acknowledgments

Petra Felsch, and Alexander Kanwischer for excellent assistance during PET imaging and animal handling and their continuous support.

Furthermore, I would like to thank the team of the Small Animal MRI Centre of the Central Animal Facility and Institute for Laboratory Animal Science, especially Christian Bergen and Dr. Martin Meier for excellent assistance during MRI scans and their continuous support.

Finally, I wish to thank my family and friends for their love, unconditional support and encouragement.

- 197 -

- 198 -