Heart Failure and Mouse Models

Heart failure and mouse models

Ross Breckenridge1,*

Heart failure is a common, complex condition with a poor prognosis and increasing couraged in heart failure for many years on the basis of studies showing increased mor- incidence. The syndrome of heart failure comprises changes in electrophysiology, tality when given to acutely unwell patients contraction and energy metabolism. This complexity, and the interaction of the (Braunwald and Chidsey, 1965). Subse- clinical syndrome with very frequently concurrent medical conditions such as quently,  blockers have been found to be diabetes, means that animal modelling of heart failure is difficult. The current animal beneficial when used in relatively low doses models of heart failure in common use do not address several important clinical in stable heart failure patients (CIBIS, 1999). problems. There have been major recent advances in the understanding of cardiac An increasingly commonly recognised biology in the healthy and failing myocardium, but these are, as yet, unmatched variant of heart failure is ‘diastolic’ or ‘sys- tolic function preserved’ heart failure, char- by advances in therapeutics. Arguably, the development of new animal models acterised by resistance to ventricular filling of heart failure, or at least adaptation of existing models, will be necessary to fully rather than defective contraction (Dodek et translate scientific advances in this area into new drugs. This review outlines the al., 1972; Zile et al., 2004). As the com-

DMM mouse models of heart failure in common usage today, and discusses how monest causes of diastolic heart failure are adaptations in these models may allow easier translation of animal experimentation ischaemia, obesity, hypertension, diabetes into the clinical arena. and ageing (Owan and Redfield, 2005), the incidence of this condition is expected to increase with time. Fundamentally, the Heart failure: the medical problem the case study. Our understanding of the mechanisms underlying this variant of heart Heart failure is an increasingly common heart failure disease phenotype is not yet failure are unknown. diagnosis, with a dismal prognosis that is sufficient to appreciate whether aspects of There have been major recent advances worse even than many types of cancer (Ho the final heart failure syndrome differ with in the clinical assessment of heart failure et al., 1993). There are few therapeutic the causative aetiologies. Preliminary evi- patients – the aim being earlier diagnosis options. The estimated cost of heart failure dence, however, demonstrates that different and risk stratification (i.e. the identification is currently between 1-2% of the total gene expression patterns (Huang et al., of high-risk patients). Clinical scoring based healthcare spend in developed economies 2005) and prognoses (Felker et al., 2000) are on a patient’s assessment of their own exer- and is expected to rise (McMurray and associated with heart failure resulting from cise capacity and basic clinical observations Stewart, 2000). differing causes. has proved of use in identifying high-risk Clinical presentation may be insidious or Heart failure comprises changes in car- patients and guiding therapy (Goldman et Disease Models & Mechanisms acute, with decreased exercise tolerance diac contractility, electrical conduction and al., 1981). More recently, imaging technol- and shortness of breath. Cardiac arrhyth- energy metabolism, leading to an inability ogy has been used to improve diagnosis and mias may accompany heart failure, leading of the heart to meet circulatory demands prognostication in heart failure. Echocar- to high rates of sudden death. Current treat- (Jessup and Brozena, 2003; Stanley et al., diography has been the historical ‘gold stan- ment includes simultaneous administration 2005). This activates neurohormonal com- dard’ for non-invasive evaluation of the fail- of angiotensin-converting enzyme (ACE) pensatory mechanisms, such as vasocon- ing heart, and is safe and relatively cheap. inhibitors (acting principally by vasodilata- striction, which are thought to be helpful in The recent development of tissue Doppler tion),  blockers (slowing the heart rate) and maintaining general organ perfusion in the measurement to evaluate myocardial strain spironolactone (vasodilatation and diure- short term, but maladaptive with respect to has improved sensitivity with respect to sis). This combination reduces death result- long-term cardiac function. In this way, the early detection of abnormalities. Further ing from heart failure, but these drugs do physiological and gene expression changes adaptations, such as ultrasonographic track- not ‘cure’ the condition. There are no truly observed in heart failure are not constantly ing of acoustic markers (‘speckle tracking’) ‘disease modifying’ drugs for heart failure. ‘adaptive’ or ‘maladaptive’ (and are, thus, not and intravascular ultrasound contrast (Flu Many conditions eventually lead to heart easily amenable to pharmacological modu- et al., 2009), have shown promise at their failure (Table 1), several of which are asso- lation). This has introduced therapeutic relatively early stage of development. Car- ciated with each other, such as hyperten- confusion over the years. For example, use diac magnetic resonance imaging (cMRI) is sion, obesity and diabetes, as exemplified in of -adrenoceptor antagonists was dis- increasingly used in heart failure and, as well as being the most accurate method to deter- 1MRC National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA, UK mine left ventricular mass (an important *Author for correspondence ([email protected]) clinical measurement in certain types of car-

138 dmm.biologists.org Heart failure and mouse models CLINICAL PUZZLE

legal difficulties in obtaining high-quality Case study biopsies from relevant patient groups have A 72-year-old man was admitted to hospital complaining of a progressive three-day history of been addressed by a working group of the shortness of breath at rest, exacerbated by lying flat. He had suffered a myocardial infarction four American Heart Association and European years earlier, which was treated by primary angioplasty. His past medical history was significant for Society of Cardiology (Cooper et al., 2007). hypertension, hypercholesterolaemia, and type II diabetes, which was diagnosed at the time of his In the absence of widely available biopsy infarct. Following the angioplasty, he was prescribed a  blocker, an ACE inhibitor, a statin, anti- platelet agents (all of which have been shown to reduce mortality following myocardial infarction) material, clinical trials and medical practice and oral anti-diabetic medication. He had remained well in the intervening four years. On have relied on the use of biomarkers that examination, he was found short of breath at rest with crackles audible in both lung fields, consistent are essentially unvalidated. A comparison of with left ventricular failure. Echocardiography was performed, revealing a significantly depressed left the gene expression changes in animal mod- ventricular ejection fraction and normal valvular function. He was treated with intravenous diuretics els and human failing hearts is vital for val- to good symptomatic effect, and had the doses of his other cardiac medications increased. Over the idating the findings from animal models of next six months, he had two further admissions with left ventricular failure, both of which were heart failure, and eventually for assaying treated with diuretics. biomarkers used in humans.

diomyopathy and heart failure), can also niques are still experimental in that they Unanswered clinical questions in determine the cause of heart failure by have not yet been proven to predict clini- heart failure quantifying microvascular perfusion, cal events. It is largely unknown whether the individ- myocardial iron and fibrosis (Karamitsos et The gene expression changes that lead to ual changes in gene expression and physi- al., 2009). Diagnostic imaging of the heart failure are not definitively known. ology that are observed in heart failure myocardium using radio- or nanoparticle- Many changes in gene expression have been patients are adaptive or maladaptive, and labelled probes (‘molecular imaging’) is an described in biopsies taken from human how this changes with the evolution of the

DMM area of active development (Saraste et al., heart failure patients. However, biopsies disease. Further questions include whether 2009) and is potentially one of the quickest are necessarily small and usually taken there are novel biomarkers (Lainscak and ways to translate advances in basic science blindly through the endocardial route via a Anker, 2009), and which imaging modali- into the clinical arena. Many of these tech- vascular catheter. The logistical, ethical and ties are optimal that will facilitate clinical decision-making in heart failure patients. Two obvious deficiencies are hindering Table 1. Commonest documented causes of heart failure the development of new heart failure ther- Cause of heart failure apies: (1) high-resolution, longitudinal phe- No single identifiable cause found in 50% of cases notyping of heart failure patients (i.e. at sev- Ischaemia eral stages in the evolution of the condition) has not yet been carried out; and (2) the Pressure overload Hypertension development of new animal models that Aortic valve stenosis more closely mimic the medical treatment Aortic coarctation of heart failure, and the common causes of Volume overload R-L (right-to-left) shunting heart failure (such as obesity and hyper- Mitral valve regurgitation tension) that interact with treatment, would Disease Models & Mechanisms Infection Viral cardiomyopathy/myocarditis facilitate convergence between clinical and animal modelling fields. Given the huge HIV recent advances in our understanding of Infiltration Sarcoid murine biology, mouse models would Amyloid arguably be of the most use to answer these Peripartum cardiomyopathy questions. Connective tissue disease Recent insights into cardiac biology Toxins Alcohol The last two decades have seen rapid Cocaine advances in our understanding of the Anthracyclines genetic and physiological processes Other drugs (corticosteroids, lithium) involved in the development and mainte- Diabetes nance of a healthy mammalian heart, and how perturbations may result in disease; for Endocrine Thyroid disease example, the discovery of autologous car- Corticosteroid excess diac repair through pluripotent cells, which Growth hormone excess has been extensively reviewed elsewhere Phaeochromocytoma (Chien et al., 2008; Hansson et al., 2009; Inherited cardiomyopathy Nakano et al., 2008). Development of this Table adapted from Felker et al., 2000. technology towards the clinic requires reli-

Disease Models & Mechanisms 139 CLINICAL PUZZLE Heart failure and mouse models

able animal models, which arguably do not adaptive with respect to survival, eventually exist yet. results in cardiac dysfunction when acti- Clinical terms Strikingly, the failing adult heart resem- vated chronically (Olson and Schneider, ACE inhibitors – drugs that block angiotensin- bles the foetal heart in many ways. Several 2003). In addition, pharmacological converting enzyme (ACE) to inhibit the gene expression changes that have been inhibitors of HDACs, part of the ‘stress conversion of angiotensin I to angiotensin II. reported in failing hearts are consistent pathways’, are being investigated as poten- Less angiotensin II reduces blood pressure, increases parasympathetic activity in the heart tially novel treatments for heart failure (Kee with ‘re-expression of the foetal gene to control cardiac arrhythmia, and inhibits expression program’. The expression of et al., 2006). It has been pointed out that cardiac muscle wasting (cachexia) in heart numerous transcription factors that are stress pathways are also activated in the failure patients associated with heart development (Rajabi foetal heart, adapting cardiac function to a  blocker/-adrenoceptor antagonists – et al., 2007), gene activation programmes via low-oxygen environment (Rajabi et al., propranolol-based compounds that inhibit histone deacetylases (HDACs) (Molkentin 2007). sympathetic neuron (epinephrine)-induced et al., 1998), and foetal-type microRNA Transgenic mouse experiments are now cardiac stimulation, which are used to control (miRNA) expression profiles (Rooij et al., investigating whether these changes in gene cardiac arrhythmias and elevated blood pressure 2006; Thum et al., 2007) have been reported expression drive the heart failure phenotype Cardiac contractility – the intrinsic ability of a in failing hearts. Furthermore, ‘foetal’ con- (and, thus, whether they are therapeutic tar- heart muscle fibre to contract at a certain tractile protein isoforms and ion channels gets), or whether they are protective in the length, it is a measure of muscle performance have been documented in human failing context of heart failure. Arguably, more Cardiac output – the blood volume pumped hearts. An important question is whether effective models of heart failure would be out from the heart this gene expression shift contributes to the of great use in this area. Cardiac MRI – cardiac imaging modality heart failure phenotype, or whether it is a providing direct visualisation of the heart chambers and myocardium without the risk of protective response. Recently, it has been Animal models of heart failure potentially harmful ionising radiation

DMM postulated that foetal isoforms of the con- The aim of animal modelling of heart fail- Coronary CT scan – cardiac imaging modality tractile protein myosin are beneficial in the ure is to simplify an extremely complex syn- that is used primarily to evaluate coronary failing heart as they reduce oxygen con- drome into manageable research questions. atherosclerotic disease sumption, albeit at the expense of reduced A key decision is the choice of animal sys- Echocardiography – sonogram or ultrasound contractile function (Krenz and Robbins, tem – often this is a trade off between con- of the heart, it allows real-time visualisation of 2004; Lowes et al., 1997; Rajabi et al., 2007; venience/cost and physiological applicabil- blood dynamics and wall motion in the heart Heart failure – inability of the heart to pump a Tardiff, 2006). It is perhaps a paradox that ity. Animal models of heart failure, as sufficient blood supply to meet tissue reducing cardiac contraction in a condition opposed to isolated organs or cells, do consumption needs defined by an insufficiency in cardiac out- enable analysis of the physiological effects Primary angioplasty – dilation of a narrowed put should be beneficial. However, there is of cardiac dysfunction, which are of great or stenosed coronary artery, with the goal of supporting evidence from human adult importance in the overall heart failure phe- increasing blood flow through the artery therapeutics for this hypothesis. When notype. administered appropriately to heart failure The relatively complete annotation and mouse hearts (Swynghedauw, 1986). Hav- patients,  blockers afford a clear survival simple manipulation of the mouse genome ing said this, the common features of advantage and are negatively inotropic, but have allowed significant mechanistic murine and human hearts have allowed for reduce cardiac ATP consumption (CIBIS, insights into human disease. Mouse mod- many important observations regarding Disease Models & Mechanisms 1999; Marian, 2006). It is now becoming els are indispensable tools in many areas of embryology, physiology, cell signalling, clear that understanding the alterations in medical research and the mouse is, most energetics and stem cell function. energy generation that are seen in heart fail- often, the animal system of choice when dis- ure may be pivotal in advancing our knowl- ease modelling. Surgical models of heart failure edge of heart failure. The failing heart shifts Although mice are relatively inexpensive The most commonly used group of mouse energy generation from predominantly lipid and convenient, there are significant dif- models of heart failure uses the response to oxidation, seen in the healthy adult heart, ferences between mouse and human a surgical intervention, such as banding an to glycolysis, which predominates in the hearts. Mouse hearts are obviously very aorta or clipping a coronary artery, to model foetal heart. This is another aspect in which small, and beat very quickly [400-600 beats the multisystem effects of heart failure failing hearts seem to resemble foetal per minute (bpm)] compared with humans (Table 2) (Mayosi et al., 2006). These surgi- hearts, and a key area of research has been (60-90 bpm), leading to important differ- cal models of heart failure have the advan- the possible mechanisms controlling this ences in calcium handling and ion currents tage of very closely replicating specific dis- switch. between mice and human hearts. For ease situations of myocardial infarction There are a number of biochemical ‘stress example, heterozygous deletion of the cal- (coronary artery ligation) and heart failure pathways’ that are activated in the failing cium channel SERCA2a in mice leads to a owing to hypertension or aortic valve steno- heart in response to a number of stimuli, contractile dysfunction, whereas in sis (aortic banding). However, heart failure and that may lead to changes such as meta- humans there is no apparent phenotype occurs suddenly post-surgery in the context bolic substrate switching (Pu and Izumo, (Mayosi et al., 2006). There are also impor- of a relatively young heart, whereas in 2001). It is now thought that activation of tant differences in the predominant myosin humans, the onset may be insidious over these signalling axes, although initially isoforms expressed in adult human and several years in the context of comorbidi-

140 dmm.biologists.org Heart failure and mouse models CLINICAL PUZZLE

ties and age-relate changes. All surgical evidence from animal experimentation that be a risk factor for the development of models are relatively expensive and techni- the nature of the stimulus for cardiac hyper- heart failure in humans (Herrmann et al., cally demanding with high rates of intra- trophy (which often precedes heart failure 2006), have been induced in mice by het- operative mortality, which reduces repro- and is associated with activation of the erozygous deletion of cystathionine--syn- ducibility. Interventions such as long-term same ‘stress pathways’) determines the type thase (Vacek et al., 2009), again modelling pacing of the heart at high rates are also of hypertrophic response, and potentially a specific cause of heart failure. Isopro- used commonly in larger animals, such as the clinical outcome (for example exercise terenol infusion in rodents models the car- dogs and rabbits, but this is not (yet) tech- training versus intermittent transverse aor- diotoxic effects of chronic activation of the nically feasible in mice. tic banding) (Perrino et al., 2006; Tardiff, sympathetic system (Balazs and Herman, Transverse aortic banding seeks to model 2006). 1976; Oudit et al., 2003). Technically, these heart failure resulting from aortic stenosis. Surgical models of heart failure model models are simple and reproducible, but the A ligature or clip is placed across the specific cardiac conditions that represent a question of general applicability of these ascending or descending aorta, inducing portion of the heart failure cases that are experiments to all forms of heart failure abnormally increased intracardiac pressure encountered clinically. The major disease arises. (‘pressure overload’). Typically, intraopera- burden of heart failure in the future is tive mortality is high and operator-depen- expected to come from patients with the Genetic models of heart failure dent, with heart failure developing within complex phenotype cluster of hyperten- One approach has been to individually weeks of the procedure in the surviving sion/hyperlipidaemia/obesity/diabetes. It is investigate the changes in gene expression mice. This procedure is assumed to model not obvious how closely the heart failure associated with heart failure by targeted heart failure resulting from aortic stenosis resulting from, for example, hypertension mutagenesis/transgenesis or pharmacol- in humans, although the time scale of the and diabetes resembles the current animal ogy. This has proved very fruitful, but any development of heart failure differs signif- models. There are increasing numbers of such reductive approach to heart failure is

DMM icantly. How much relevance this model has spontaneously occurring or engineered liable to charges of oversimplification. to other mechanisms of heart failure is mouse strains that exhibit hypertension, Over 5000 cardiac transgenic/knockout unknown. diabetes or obesity (Russell and Proctor, studies have now been published Surgical clipping of a coronary artery, 2006). Combination of these models may (http://circres.ahajournals.org/cgi/collec- typically the left anterior descending artery, provide ‘polygenic’ models of heart failure, tion/animal_models_of_human_disease). models heart failure following myocardial perhaps in conjunction with existing surgi- Although not in themselves disease mod- infarction. In the postoperative days and cal techniques. els, they have proven indispensable for weeks, a myocardial scar and dilated car- investigating individual signalling path- diomyopathy develop. This clearly models Toxin-induced heart failure ways, or to investigate the role of individ- the effect of a completed human myocar- Administration of a single toxin or drug is ual genes. This approach has been criti- dial infarction, but has the experimental a theoretically attractive method of model- cised on several grounds – most drawbacks of relatively high perioperative ling heart failure. For example, ethanol importantly, that the study of individual mortality, relatively poor reproducibility, (Berk et al., 1975) and the cytotoxic drug gene manipulations has little relevance out and the technical challenge of such a deli- doxorubicin (Delgado et al., 2004) are of context (Cook et al., 2009). cate procedure. known cardiotoxins in humans and have In their simplest forms, genetic alter- Surgical models of heart failure/hyper- been used in mice to induce heart failure ations can be induced by overexpression Disease Models & Mechanisms trophy have afforded many insights into syndromes. Elevated levels of the non-pro- under the control of heart-specific pro- human disease. For example, there is direct tein amino acid homocysteine, suggested to moters, or deletion by targeted deletion of

Table 2. Advantages and limitations of some commonly used animal models of heart failure Model Advantages Limitations Surgical Aortic banding Model of human pressure overload High mortality; technically demanding; relevance to other conditions? Coronary ligation Directly applicable to human disease Cryoinjury Technically simple Relevance to human disease? Toxic Ethanol Non-invasive; technically simple; reproducible High mortality; non-cardiac effects; not generally applicable? Doxorubicin Homocysteine Genetic Transgenic Directly reproduces gene expression changes seen in Non-specific effects of overexpression disease Targeted Developmental effects of null allele Inducible null/transgenic Control of time course of induction/deletion Control of level of transgene often impossible

Disease Models & Mechanisms 141 CLINICAL PUZZLE Heart failure and mouse models

a locus or crucial part of it (Yutzey and Rob- Conclusion bins, 2007). Temporospatial control of Opportunities for making mouse Clinical and basic research genetic manipulations is now possible using models of heart failure more opportunities tet-inducible transgenic systems (Gulick clinically applicable • Investigate whether heart failure from and Robbins, 2005; Sanbe et al., 2003), and The end points used in mouse studies different causes benefits from different inducible deletion using hormone-fused should, where possible, be clinically valid treatments Cre recombinase constructs (Sohal et al., (for example, by using the same imaging • Research to understand the complex 2001). modalities when studying cardiac mor- elements that contribute to heart failure. Relating the expression of a transgene to phology in mice and heart failure patients) Central to this, is how changes in gene endogenous levels (and human expression (Allen et al., 2009). It is worth pointing out expression and energy dynamics induced by data) is absolutely crucial – this applies as that many of the clinical end points that are the disease influence its progression • Define reliable biomarkers to diagnose and much to deletion experiments as transgenic important to doctors, patients and health- phenotype heart failure overexpression ones. Very high levels of care systems alike, such as quality of life, • Establishment of relevant and informative even ‘inert’ proteins such as Cre recombi- exercise tolerance and hospital admission, indices that correlate well between animal nase or green fluorescent protein (GFP) can are unlikely to be modelled adequately in models and human patients with heart lead to dose-related phenotypes (Buerger et any animal system. The use of surrogate failure al., 2006). Therefore, misexpression or dele- markers (measured end points in clinical tion of a single gene has a high chance of trials to substitute for clinical events such of novel disease targets. Relatively simple leading to artefactual phenotypes. Consti- as death) in heart failure clinical trials is modifications in the experimental design of tutive (i.e. non-inducible) misexpression complex and controversial (Gheorghiade et animal models of heart failure have the runs the risk of generating artefacts owing al., 2003). As the field advances, it will be potential to allow increased understanding to developmental effect. Additionally, important to establish which of the indices of the complexity of the clinical heart fail-

DMM mosaicism of expression of the transgene, measured in mice most truly model human ure syndrome. This may lead to clinical or of deletion, is invariably present and disease. In some areas, such as mouse cMRI, advances using currently available tech- rarely assessed. there are still technological limits, but nologies. Future animal models of heart fail- The targeting strategy is also crucial; echocardiography in mice has now reached ure will hopefully give mechanistic insights Yutzey and Robbins use the example of the point where human echocardiographers that lead to novel classes of therapies. myogenic regulatory factor 4 (MRF4) to can easily interpret mouse data. Currently, COMPETING INTERESTS show that three different ‘null’ mutants have the limiting factor is the size of the mouse The authors declare no competing financial interests. three different phenotypes resulting from heart – it is not clear whether this will differences in target loci and targeting vec- remain the case. Molecular imaging of REFERENCES tors (Baker et al., 2000; Stull et al., 2002; mouse hearts is developing in parallel with Allen, L. A., Hernandez, A. F., O’Connor, C. M. and Felker, G. M. (2009). End points for clinical trials in Yutzey and Robbins, 2007). human technology. acute heart failure syndromes. J. Am. Coll. Cardiol. 53, It is difficult to see how signalling path- Validation of the effect of drug treatment 2248-2258. ways could be dissected, and the effects of in mouse models against human heart fail- Baker, L. C., London, B., Choi, B. R., Koren, G. and specific disease alleles tested, without the ure is vital. For example, experimental inter- Salama, G. (2000). Enhanced dispersion of repolarisation and refractoriness in transgenic mouse use of genetically modified mice. These ventions in surgical models should be made hearts promotes re-entrant tachycardia. Circ. Res. 87, technologies have allowed the investigation on a background of standard medical ther- 275-281. Disease Models & Mechanisms of many cardiac signalling pathways, and apy ( blockers and ACE inhibitors) in the Balazs, T. and Herman, E. H. (1976). Toxic very many individual gene products, with same way as in patients in clinical trials. cardiomyopathies. Ann. Clin. Lab. Sci. 6, 467-476. Berk, S. L., Block, P. J., Toselli, P. A. and Ullrick, W. C. respect to their potential role in heart fail- The effect of comorbidities on the devel- (1975). The effects of chronic alcohol ingestion in ure and their potential as therapeutic tar- opment and treatment of heart failure is of mice on contractile properties of cardiac and skeletal gets. importance, because human heart failure muscle: a comparison with normal and dehydrated- Models of heart failure in rabbits, dogs, most often occurs as a cluster of related malnourished controls. Experientia 31, 1302-1303. Braunwald, E. and Chidsey, C. A. (1965). The sheep and larger animals have had varying medical conditions whose contribution to adrenergic nervous system in the control of the degrees of success (Halapas et al., 2008). the development of the phenotype and normal and failing heart. Proc. R Soc. Med. 58, 1063- These all have their own advantages and dis- response to treatment is unknown. In 1066. advantages, and several are arguably more future, mammalian systems other than Buerger, A., Rozhitskaya, O., Sherwood, M. C., Dorfman, A. L., Bisping, E., Abel, E. D., Pu, W. T., physiologically applicable to humans, but mouse may be needed to model complex Izumo, S. and Jay, P. Y. (2006). Dilated they are uniformly more expensive than metabolic interactions such as obesity and cardiomyopathy resulting from high-level myocardial mice and with fewer molecular biology dyslipidaemias. expression of Cre-recombinase. J. Card. Fail. 12, 392- resources. It has been estimated that, for The number of heart failure patients is 398. Chien, K. R., Domian, I. J. and Parker, K. K. (2008). example, the housing costs of experimental increasing owing to the lack of necessary Cardiogenesis and the complex biology of rabbits are 20 times those of mice, and that therapeutic tools to treat them. A dialogue regenerative cardiovascular medicine. Science 322, heart failure phenotypes may take years to between clinicians and scientists is neces- 1494-1497. develop in a rabbit, as opposed to sary to develop animal models of heart fail- CIBIS (1999). The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised trial. Lancet 353, 9-13. weeks/months in a mouse (Marian, 2005; ure that accurately replicate a complicated Cook, S. A., Clerk, A. and Sugden, P. H. (2009). Are Marian, 2006). clinical syndrome, and to allow exploration transgenic mice the ‘alkahest’ to understanding

142 dmm.biologists.org Heart failure and mouse models CLINICAL PUZZLE

myocardial hypertrophy and failure? J. Mol. Cell using gene expression profiles. BMC Bioinformatics 6, (2006). Intermittent pressure overload triggers Cardiol. 46, 118-129. 205. hypertrophy-independent cardiac dysfunction and Cooper, L. T., Baughman, K. L., Feldman, A. M., Jessup, M. and Brozena, S. (2003). Heart failure. N. Engl. vascular rarefaction. J. Clin. Invest. 116, 1547-1560. Frustaci, A., Jessup, M., Kuhl, U., Levine, G. N., J. Med. 348, 2007-2018. Pu, W. T. and Izumo, S. (2001). Transcription factors and Narula, J., Starling, R. C., Towbin, J. et al. (2007). The Karamitsos, T. D., Francis, J. M., Myerson, S., heart failure: does the stressed heart need a hand? J. role of endomyocardial biopsy in the management of Selvanayagam, J. B. and Neubauer, S. (2009). The Mol. Cell Cardiol. 33, 1765-1767. cardiovascular disease: a scientific statement from the role of cardiovascular magnetic resonance imaging in Rajabi, M., Kassiotis, C., Razeghi, P. and Taegtmeyer, American Heart Association, the American College of heart failure. J. Am. Coll. Cardiol. 54, 1407-1424. H. (2007). Return to the fetal gene program protects Cardiology, and the European Society of Cardiology. Kee, H. J., Sohn, I. S., Nam, K. I., Park, J. E., Qian, Y. R., the stressed heart: a strong hypothesis. Heart Fail. Rev. Endorsed by the Heart Failure Society of America and Yin, Z., Ahn, Y., Jeong, M. H., Bang, Y. J., Kim, N. et 12, 331-343. the Heart Failure Association of the European Society al. (2006). Inhibition of histone deacetylation blocks Rooij, E. v., Sutherland, L., Liu, N., Williams, A., of Cardiology. J. Am. Coll. Cardiol. 50, 1914-1931. cardiac hypertrophy induced by angiotensin II McAnally, J., Gerard, R., Richardson, J. and Olson, Delgado, R. M., 3rd, Nawar, M. A., Zewail, A. M., Kar, infusion and aortic banding. Circulation 113, 51-59. E. (2006). A signature pattern of stress-responsive B., Vaughn, W. K., Wu, K. K., Aleksic, N., Krenz, M. and Robbins, J. (2004). Impact of beta- microRNAs that can evoke cardiac hypertrophy and Sivasubramanian, N., McKay, K., Mann, D. L. et al. myosin heavy chain expression on cardiac function heart failure. Proc. Natl. Acad. Sci. USA 103, 18255- (2004). Cyclooxygenase-2 inhibitor treatment during stress. J. Am. Coll. Cardiol. 44, 2390-2397. 18260. improves left ventricular function and mortality in a Lainscak, M. and Anker, S. D. (2009). Prognostic factors Russell, J. C. and Proctor, S. D. (2006). Small animal murine model of doxorubicin-induced heart failure. in chronic heart failure. A review of serum biomarkers, models of cardiovascular disease: tools for the study Circulation 109, 1428-1433. metabolic changes, symptoms, and scoring systems. of the roles of metabolic syndrome, dyslipidemia, and Dodek, A., Kassebaum, D. G. and Bristow, J. D. (1972). Herz 34, 141-147. atherosclerosis. Cardiovasc. Pathol. 15, 318-330. Pulmonary edema in coronary-artery disease without Lowes, B. D., Minobe, W., Abraham, W. T., Rizeq, M. Sanbe, A., Gulick, J., Hanks, M. C., Liang, Q., Osinska, cardiomegaly. Paradox of the stiff heart. N. Engl. J. N., Bohlmeyer, T. J., Quaife, R. A., Roden, R. L., H. and Robbins, J. (2003). Reengineering inducible Med. 286, 1347-1350. Dutcher, D. L., Robertson, A. D., Voelkel, N. F. et al. cardiac-specific transgenesis with an attenuated Felker, G. M., Thompson, R. E., Hare, J. M., Hruban, R. (1997). Changes in gene expression in the intact myosin heavy chain promoter. Circ. Res. 92, 609-616. H., Clemetson, D. E., Howard, D. L., Baughman, K. L. human heart. Downregulation of alpha-myosin heavy Saraste, A., Nekolla, S. G. and Schwaiger, M. (2009). and Kasper, E. K. (2000). Underlying causes and long- chain in hypertrophied, failing ventricular Cardiovascular molecular imaging: an overview. term survival in patients with initially unexplained myocardium. J. Clin. Invest. 100, 2315-2324. Cardiovasc. Res. 83, 643-652. cardiomyopathy. N. Engl. J. Med. 342, 1077-1084. Marian, A. J. (2005). On mice, rabbits, and human heart Sohal, D. S., Nghiem, M., Crackower, M. A., Witt, S. A., Flu, W. J., van Kuijk, J. P., Bax, J. J., Gorcsan, J., 3rd failure. Circulation 111, 2276-2279. Kimball, T. R., Tymitz, K. M., Penninger, J. M. and DMM and Poldermans, D. (2009). Three-dimensional Marian, A. J. (2006). Beta-adrenergic receptors signaling Molkentin, J. D. (2001). Temporally regulated and speckle tracking echocardiography: a novel approach and heart failure in mice, rabbits and humans. J. Mol. tissue-specific gene manipulations in the adult and in the assessment of left ventricular volume and Cell Cardiol. 41, 11-13. embryonic heart using a tamoxifen-inducible Cre function? Eur. Heart J. 30, 2304-2307. Mayosi, B. M., Kardos, A., Davies, C. H., Gumedze, F., protein. Circ. Res. 89, 20-25. Gheorghiade, M., Adams, K. F., Jr, Gattis, W. A., Hovnanian, A., Burge, S. and Watkins, H. (2006). Stanley, W. C., Recchia, F. A. and Lopaschuk, G. D. Teerlink, J. R., Orlandi, C. and O’Connor, C. M. Heterozygous disruption of SERCA2a is not associated (2005). Myocardial substrate metabolism in the (2003). Surrogate end points in heart failure trials. Am. with impairment of cardiac performance in humans: normal and failing heart. Physiol. Rev. 85, 1093-1129. Heart J. 145, S67-S70. implications for SERCA2a as a therapeutic target in Stull, L. B., Leppo, M. K., Marban, E. and Jannsen, P. Goldman, L., Hashimoto, B., Cook, E. F. and Loscalzo, heart failure. Heart 92, 105-109. M. (2002). Physiological determinants of contractile A. (1981). Comparative reproducibility and validity of McMurray, J. J. and Stewart, S. (2000). Epidemiology, force generation and calcium handling in the mouse systems for assessing cardiovascular functional class: aetiology, and prognosis of heart failure. Heart 83, myocardium. J. Mol. Cell Cardiol. 34, 1367-1376. advantages of a new specific activity scale. Circulation 596-602. Swynghedauw, B. (1986). Developmental and 64, 1227-1234. Molkentin, J. D., Lu, J. R., Antos, C. L., Markham, B., functional adaptation of contractile proteins in cardiac Gulick, J. and Robbins, J. (2005). Regulation of Richardson, J., Robbins, J., Grant, S. R. and Olson, and skeletal muscles. Physiol Rev. 66, 710-771. transgene expression using tetracycline. Curr. Protoc. E. N. (1998). A calcineurin-dependent transcriptional Tardiff, J. C. (2006). Cardiac hypertrophy: stressing out Mol. Biol. 23, 12. pathway for cardiac hypertrophy. Cell 93, 215-228. the heart. J. Clin. Invest. 116, 1467-1470. Halapas, A., Papalois, A., Stauropoulou, A., Nakano, A., Nakano, H. and Chien, K. R. (2008). Thum, T., Galuppo, P., Wolf, C., Fiedler, J., Kneitz, S., Philippou, A., Pissimissis, N., Chatzigeorgiou, A., Multipotent islet-1 cardiovascular progenitors in van Laake, L. W., Doevendans, P. A., Mummery, C. Kamper, E. and Koutsilieris, M. (2008). In vivo development and disease. Cold Spring Harb. Symp. L., Borlak, J., Haverich, A. et al. (2007). MicroRNAs models for heart failure research. In Vivo 22, 767-780. Quant. Biol. 73, 297-306. in the human heart: a clue to fetal gene Disease Models & Mechanisms Hansson, E. M., Lindsay, M. E. and Chien, K. R. (2009). Olson, E. N. and Schneider, M. D. (2003). Sizing up the reprogramming in heart failure. Circulation 116, 258- Regeneration next: toward heart stem cell heart: development redux in disease. Genes Dev. 17, 267. therapeutics. Cell Stem Cell 5, 364-377. 1937-1956. Vacek, T. P., Sen, U., Tyagi, N., Vacek, J. C., Kumar, M., Herrmann, M., Taban-Shomal, O., Hubner, U., Bohm, Oudit, G. Y., Crackower, M. A., Eriksson, U., Sarao, R., Hughes, W. M., Passmore, J. C. and Tyagi, S. C. M. and Herrmann, W. (2006). A review of Kozieradzki, I., Sasaki, T., Irie-Sasaki, J., Gidrewicz, (2009). Differential expression of Gs in a murine model homocysteine and heart failure. Eur. J. Heart Fail. 8, D., Rybin, V. O., Wada, T. et al. (2003). of homocysteinemic heart failure. Vasc. Health Risk 571-576. Phosphoinositide 3-kinase gamma-deficient mice are Manag. 5, 79-84. Ho, K. K., Anderson, K. M., Kannel, W. B., Grossman, protected from isoproterenol-induced heart failure. Yutzey, K. E. and Robbins, J. (2007). Principles of W. and Levy, D. (1993). Survival after the onset of Circulation 108, 2147-2152. genetic murine models for cardiac disease. Circulation congestive heart failure in Framingham Heart Study Owan, T. E. and Redfield, M. M. (2005). Epidemiology 115, 792-799. subjects. Circulation 88, 107-115. of diastolic heart failure. Prog. Cardiovasc. Dis. 47, 320- Zile, M. R., Baicu, C. F. and Gaasch, W. H. (2004). Huang, X., Pan, W., Grindle, S., Han, X., Chen, Y., Park, 332. Diastolic heart failure-abnormalities in active S. J., Miller, L. W. and Hall, J. (2005). A comparative Perrino, C., Naga Prasad, S. V., Mao, L., Noma, T., Yan, relaxation and passive stiffness of the left ventricle. N. study of discriminating human heart failure etiology Z., Kim, H. S., Smithies, O. and Rockman, H. A. Engl. J. Med. 350, 1953-1959.

Disease Models & Mechanisms 143