Myocardial damage induced by a single high dose of isoproterenol in C57Bl/6J mice triggers a persistent adaptive immune response against the heart

Heart failure is the common final pathway of a range of conditions causing myocardial damage and a major cause of morbidity and mortality worldwide. Strategies to improve tissue repair and prevent heart failure thus remain an urgent clinical need. Recent studies have shown activation of the adaptive immune system in response to myocardial damage and implicated anti-heart autoimmunity in the development of heart failure. In an attempt to harness this new therapeutic avenue, the number of experimental studies using in vivo models of myocardial damage to study the ensuing immune response has surged. The beta-adrenergic agonist isoproterenol-hydrochloride has been used for its cardiac effects in a variety of different dosing and administration regimes. Most prominently, low doses (<10mg/kg sc) over an extended time period induce cardiac hypertrophy and fibrosis. In addition, single injections of high doses (>100mg/kg) induce cardiomyocyte necrosis and have been used to mimic acute myocardial necrotic lesions as seen in myocardial infarction (MI). However, despite significant resource and animal welfare advantages, concerns about off-target effects and clinical relevance have so far limited uptake in the cardiovascular research community. To assess suitability of the isoproterenol model for the analysis of chronic post-MI immunological readouts, we treated C57Bl/6J mice with a single intra-peritoneal bolus injection of 160mg/kg isoproterenol. Our results confirm the presence of necrotic lesions in the myocardium with significant resemblance of the histopathology of type 2 myocardial infarction. Kidneys develop mild fibrosis secondary to early cardiac damage, while other organs remain unaffected. Most importantly, we showed that isoproterenol treatment causes myocardial inflammation and fibrosis, activation of T cells in the heart-draining mediastinal lymph nodes, deposition of mature antibodies in the myocardium and the presence of auto-antibodies against the heart in the serum 12 weeks after the initial injury. In summary, this simple and cost-effective experimental model with significant animal welfare benefits induces myocardial damage reminiscent of human type 2 MI, which is followed by a persistent adaptive immune response against the heart. This makes it a suitable and high-throughput model to study pathological mechanisms of anti-heart autoimmunity as well as potential immunomodulatory therapeutic approaches.

attempt to harness this new therapeutic avenue, the number of experimental studies using in vivo models of myocardial damage to study the ensuing immune response has surged.
The beta-adrenergic agonist isoproterenol-hydrochloride has been used for its cardiac effects in a variety of different dosing and administration regimes. Most prominently, low doses (<10mg/kg sc) over an extended time period induce cardiac hypertrophy and fibrosis. In addition, single injections of high doses (>100mg/kg) induce cardiomyocyte necrosis and have been used to mimic acute myocardial necrotic lesions as seen in myocardial infarction (MI).
However, despite significant resource and animal welfare advantages, concerns about offtarget effects and clinical relevance have so far limited uptake in the cardiovascular research community.
To assess suitability of the isoproterenol model for the analysis of chronic post-MI immunological readouts, we treated C57Bl/6J mice with a single intra-peritoneal bolus injection of 160mg/kg isoproterenol. Our results confirm the presence of necrotic lesions in the myocardium with significant resemblance of the histopathology of type 2 myocardial infarction. Kidneys develop mild fibrosis secondary to early cardiac damage, while other organs remain unaffected. Most importantly, we showed that isoproterenol treatment causes myocardial inflammation and fibrosis, activation of T cells in the heart-draining mediastinal

Introduction
Clinical guidelines distinguish between myocardial injury and myocardial infarction (MI).
While myocardial injury encompasses any type of acute myocardial damage such as necrosis induced by myocarditis, sepsis or endogenous catecholamines, MI is defined as ischemic damage resulting from an insufficient myocardial oxygen supply 1 2 . A prominent cause of MI is atherosclerotic vascular disease. Disruption of an atherosclerotic plaque can prevent coronary blood flow through a given artery resulting in infarction of the dependent myocardium. This is referred to as Type 1 MI (T1MI). Conversely, Type 2 MI (T2MI) is a heterogeneous syndrome that occurs in the absence of coronary artery disease (CAD) and an acute atherothrombotic coronary event. Instead, other conditions such as hypoxemia, hypo/hypertension, tachycardia or tachyarrhythmias lead to an imbalance between oxygen supply and demand 3 4 . T2MI is a common but under-recognised clinical entity, with prevalence estimates of up to 58% of MI patients 5 .
Importantly, improved intervention and treatment after myocardial damage have significantly increased immediate survival. However, the regenerative capacity of the adult heart is minimal and healing is achieved by fibrotic repair which replaces damaged myocardium with non-contractile scar tissue. This impairs heart function and 62.7% of MI patients develop heart failure within 6 years after infarction 6 . The early immune response to myocardial necrosis is crucial for quick tissue repair, but excessive inflammation is involved in the development of heart failure 7 8 9 10 . Large amounts of cardiac self-antigens, including myosin and troponin, in a highly inflammatory environment activate the adaptive immune system. Auto-reactive T and B cells can be long-lived and have been suggested to cause ongoing low level tissue destruction 11 12 hampering regenerative efforts and exacerbating development towards heart failure 13 14 . There is therefore an urgent need for better recognition and understanding of these autoimmune processes to prevent further myocardial damage.
Investigation of the adaptive immune response to cardiac damage is a highly complex multidimensional process, which to date still needs in vivo experimentation. In rodents, surgical coronary artery ligation most prominently of the left anterior descending artery (LAD), is the most commonly used experimental model for human T1MI 15 . LAD ligation is induced during open chest surgery, an invasive low-throughput procedure associated with high mortality rates. The use of the ß-adrenergic agonist isoproterenol to induce cardiotoxicity and infarctlike lesions has been described over 50 years ago in rats 16 17 18 . It is now routinely used as a long term low-dose treatment to induce cardiac hypertrophy 19 20 21 , and few research groups have also used it to induce acute infarct-like lesions by a one-off injection of higher doses 22 23 24 . Importantly, the isoproterenol model recapitulates several of the salient feature of T2MI.
Positive inotropic and chronotropic effects lead to increased myocardial oxygen consumption, while tachycardia leads to a reduction in diastolic perfusion time and coronary blood flow, combining to result in patchy acute myocardial damage and necrosis 25 . Notably, it also mechanistically and pathologically models a pattern of injury termed inotrope-induced cardiac damage or pressor-injury or catecholamine-heart, seen in the cardiac transplant and terminal heart failure setting 26 .
Here we characterise the immunological effects of a one-off intra-peritoneal injection of isoproterenol as a simple, resource-efficient, highly reproducible and significantly less invasive alternative to surgical MI induction. We show that T cell activation in the mediastinal lymph nodes and anti-heart auto-antibody production can be investigated without the need for invasive surgical procedures, as the adaptive immune system will respond to cardiac antigens released from necrotic tissue, irrespective of the original trigger of necrosis.

Materials and Methods:
Mice: All animal procedures carried out at Imperial College London were approved by the Histology and scoring of damage parameters: Organs of treated and untreated mice were excised after perfusion as described above, fixed in 10% neutral buffered formalin overnight, and stored in 70% ethanol. For wax-embedding and histology, tissue samples were dehydrated in an increasing gradient of ethanol and embedded in paraffin. Five µm sections were cut and de-waxed and rehydrated in an ethanol gradient. Sections were stained with hematoxylin and eosin (H&E) and Picrosirius Red. All reagents were purchased from Sigma Aldrich (Sigma-Aldrich, Dorset, UK). Semi-quantitative scoring of heart sections was performed as established previously 27 . Hematoxylin & eosin stained sections were used to analyse and score mononuclear cell infiltration in hearts. Picrosirius Red staining was used to analyse and score fibrosis in hearts and kidneys. Individual parameters were scored on a scale of 0 (none), 1 (mild), 2 (moderate) to 3 (severe). Scores were obtained from 4 areas each on two midline cross sections per animal by a blinded researcher. Images were captured using a LMD7000 microscope (Leica microsystems, Milton Keynes, UK) and processed for quantification of nuclei (cell count) and area of fibrosis using the public domain software ImageJ (NIH; http://rsb.info.nih.gov,) 28 . Statistical analysis: Statistical analysis was performed using GraphPad Prism 8 and data were presented as mean±s.e.m throughout. Comparison between 2 groups was performed using student's t-test. Comparison between multiple experimental groups was performed using one-or two-way ANOVA with Dunnett's multiple comparisons post hoc test to obtain multiplicity-adjusted p-values. Non-parametric scoring data were analysed using Kruskal-Wallis test with Dunn's multiple comparisons post hoc test to obtain multiplicity-adjusted pvalues. Differences were considered significant at p<0.05.

Results:
(1) A single dose of 160mg/kg isoproterenol induces significant cardiac damage in C57Bl/6J mice. Different mouse strains show variable susceptibility to induction of cardiac damage and fibrosis, and C57Bl/6J mice are considered to have a very robust cardiac phenotype and to be resistant to myocardial damage and cardiac fibrosis 29 30 . Yet, they are among the best characterised mouse strains and commonly used for genetic modification 31 .
We therefore performed a dose titration experiment to investigate if C57Bl/6J mice were susceptible to developing cardiac lesions while maintaining high animal welfare standards.
Mice were treated with a one-off bolus intra-peritoneal injection of increasing doses of isoproterenol starting with 40mg/kg up to 320mg/kg. The highest dose of 320mg/kg was administered as 2 separate 160mg/kg doses on subsequent days. Isoproterenol effects are evident from 15mins after injection, when mice start to stop moving and increased respiration rates are observed. In general, mice recover within 2 hours as judged by the return of normal feeding and grooming behaviour. 160mg/kg isoproterenol induced significant increase in the cardiac damage biomarker cardiac Troponin I in the serum ( Figure 1A) as well as myocardial infiltration and fibrosis ( Figure 1B). Spleen/body ratio as a measure of systemic inflammation and body weight as indicator of overall health was not affected with doses up to 160mg/kg ( Figure 1C, D). 160mg/kg each on 2 subsequent days however increased spleen/body ratio and resulted in a drop in body weight. A single injection of 160mg/kg was thus chosen as standard dose for subsequent experiments. Histopathologically, cardiac damage was accompanied by early patterns of necrosis followed by a maturing repair response resulting in granulation tissue and eventually fibrosis over an appropriate time-span of weeks.
Consistent with catecholaminergic damage, and a strong component of ischemia, injury was most prominent in the subendocardium of the left ventricular free wall, but was essentially confluent and transmural in the septum ( Figure 1E).
To further improve animal welfare benefits of the isoproterenol model, we included pretreatment with 0.05mg/kg buprenorphine for pain relief. Buprenorphine was injected subcutaneously 30 minutes before isoproterenol administration. Manual mouse grimace scale (MGS) scoring was deemed not suitable, as no orbital tightening, or changes in nose and cheek bulge and whisker position was notable 32 . Instead, we measured activity (moving, feeding, drinking, grooming) as a reliable sign of both the onset of and the recovery from acute isoproterenol effects (Figure 2A). Adding buprenorphine to the isoproterenol treatment shortened the average time to recovery by 20 minutes without affecting the degree of cardiac inflammatory and fibrotic damage ( Figure 2B).
(2) Kidney, liver, lung and skeletal muscle are protected from direct isoproterenol-induced damage. To define the degree of off-target damage to other organs, we performed a thorough histopathological assessment of kidneys, liver, lung and skeletal muscle at week 1 and 2 after injection of 160mg/kg isoproterenol. Myocardial mononuclear infiltration and fibrosis was confirmed ( Figure 3A). No pathological changes were observed in any of the other tested organs in any of the mice (n=5) at week 1 ( Figure 3B, C). At week 2, mild acute tubular injury and fibrosis was detectable in the kidneys ( Figure 3B, C) which is likely secondary to cardiac damage as previously observed after surgical LAD ligation 33 . (4) A single dose of 160mg/kg isoproterenol triggers production of anti-heart autoantibodies. Activated autoreactive CD4+ helper T cells are able to induce B cells to generate autoantibodies, which are a hallmark sign of an established autoimmune response. We therefore tested week 2 (acute) and week 12 (chronic) post-isoproterenol serum for the presence of anti-heart autoantibodies. We found a significant increase of auto-antibody levels at chronic stage after isoproterenol damage ( Figure 4A). Importantly, these are of the mature class-switched IgG isotype indicating an established and persistent adaptive immune response against the heart. We also found a significant amount of IgG as well as immature IgM deposited in the myocardium ( Figure 4B, C) indicating in vivo deposition. The deposition pattern follows the myocardial fibres and the inside of blood vessels indicating the potential for endothelial and cardiomyocyte damage.

Discussion:
The definition of myocardial infarction differentiates patients with MI due to plaque rupture (type 1) from those with myocardial necrosis due to oxygen supply-demand imbalance secondary to other acute illnesses (type 2). Myocardial necrosis without symptoms or signs of myocardial ischaemia is classified as acute or chronic myocardial injury. Both myocardial injury and T2MI are common, yet these patients have poor short-term and long-term outcomes 34 with heart failure risk comparable to T1MI 35 . Activation of the adaptive immune system and persistent autoimmunity against the heart have recently been suggested as pathological factors contributing to heart failure in response to MI 36 14 . The number of corresponding experimental studies has therefore surged and is anticipated to increase substantially in the near future as the cardiovascular community attempts to characterise the striking complexities of the post-MI immune response.
Here we show that a single dose of isoproterenol induces necrotic lesions in the heart. The initial trigger of necrotic cell death is most comparable to catecholamine-induced damage as seen during stress-related cardiomyopathies including takotsubo syndrome 37 38 as well as in severely-ill patients receiving catecholamines for cardiac pacing or inotropic support 39 . High dose catecholamines also induce an imbalance between myocardial oxygen supply and demand, and the degree and pattern of inflammatory damage and replacement fibrosis is reminiscent of the myocardial damage incurred by a T2MI. Mild kidney fibrosis is also observed in isoproterenol-treated mice starting from week 2, which may be secondary to cardiac damage 33 as not apparent at earlier timepoints. Other organs are not affected, which confirms that off-target tissue damage is not a limiting factor in this model. Importantly, isoproterenol-induced myocardial necrosis releases antigenic peptides which induce T cell activation in the heart-draining mediastinal lymph nodes. These activate B cells to produce mature IgG anti-heart autoantibodies, a sign of a bona fide autoimmune response against the heart. Anti-heart auto-antibodies indicating autoimmunity has thus far been demonstrated in a wide range of human heart conditions 40 and is likely due to high amounts of cardiac antigen in an inflammatory environment overwhelming immunological tolerance. Importantly, the adaptive immune system is activated by antigens released from necrotic cells irrespective of the initial cause of necrosis. Accordingly, we find anti-heart auto-antibodies in  (Table 1). In particular, T2MI and other cases of more subtle ischaemic cardiac tissue injury are not reflected well by surgical LAD ligation. In addition, the isoproterenol model is particularly useful for readouts including chronic and peripheral immunological measures, e.g. serum levels of cytokines and anti-heart autoantibodies, phenotype and function of immune cells in peripheral lymphoid organs, as well as effects of therapeutic approaches on these aspects.  strong remodelling response, dilation and progression to heart failure 57 58 moderate remodelling response, dilation and progression to heart failure 53 44 Adaptive immune stimulus Cardiac antigen (e.g. myosin, troponin) in circulation 59 60 Cardiac antigen (e.g. myosin, troponin) in circulation 61 62 Adaptive immune response Long-term auto-reactivity, anti-heart auto-antibodies 63 anti-heart auto-antibodies 14 Anticipated: anti-heart auto-antibodies Importantly, besides resource efficiency, replacing surgical MI procedures carries significant animal welfare benefits. Isoproterenol injections cause transient effects and are significantly less invasive than MI surgeries. In our hands, isoproterenol treatment carries an approximate mortality of 1% mostly due to acute arrhythmias in the first 24-48h after injection. Visible effects of appropriately titrated isoproterenol injections wear off within 1-2 hours after injection. Several opportunities to reduce necessary animal numbers also arise through the use of isoproterenol as alternative to LAD ligation surgery. Invasive thoracotomy causes a significant amount of tissue damage and inflammation and risks flawed interpretation of immunological readouts. Thus there is scientific need for sham operated animals, while simple PBS injected controls are suitable for isoproterenol experiments. Isoproterenol is administered by intra-peritoneal injection, a simple technique that can be learned easily, avoiding increased mortality rates due to inexperienced researchers. Isoproterenol effects are also more homogenous between treated animals than surgical infarcts, allowing significant reduction of group sizes. Susceptibility to isoproterenol-induced damage varies between mouse strains. However, after appropriate dose studies, isoproterenol can be used in both sexes of all mouse strains, avoiding surplus animals and allowing the use of transgenic mice on their respective background.
In summary, a single high-dose isoproterenol injection is a suitable method to induce adaptive immune responses against the myocardium with significant additional resource and animal welfare benefits.