A. Introduction

The treatment of cardiovascular disease has changed dramatically over the past 2 decades, allowing patients to live longer with better-quality lives. The introduction of new therapies (both drugs and devices) has contributed much to this success. Pre-clinical studies using animal models play a very important role in the initial evaluation of efficacy and safety of new medical devices (1-6) before their use in human clinical studies. Once these technologies enter the clinical area, further understanding of their therapeutic mechanisms can be also realized through testing in animal models, especially when such findings are compared with clinical effects and, when available, pathological specimens.

Chronic heart failure (CHF) is a clinical syndrome in which pathophysiologic underpinnings include left ventricular (LV) dysfunction, cardiac and vascular remodeling, autonomic dysfunction and marked neurohormonal activation. Despite numerous advances in pharmacologic strategies and surgical techniques, CHF remains a major cause of morbidity and mortality around the world. The prevalence of CHF is predicated to increase 25% by the year 2030 (7). Accordingly, the need to develop new pharmacologic treatments and therapeutic interventions for HF is of paramount importance. Large animal models that recapitulate the clinical phenotype of CHF are an important component of the foundation of the pathway from mechanistic understanding, to drug/device development, to clinical testing and, finally, to regulatory approvals of many new therapies. Models of myocardial infarction/ischemia, ischemic cardiomyopathy, ventricular pressure and volume overload, and pacing induced dilated cardiomyopathy have been created in large animal models to study basic aspects of CHF and to investigate potential therapies. Each model is associated with advantages and disadvantages (8), therefore care must be taken to choose the right model when initiating study of a new therapy and the findings derived from the study must be carefully evaluated.

Prior reports have reviewed different large animal models for cardiac device development (9-13). The purpose of this review (over and above prior reviews) is to consider various animals of CHF and their respective contributions to the development of relevant novel surgical and interventional strategies.

B. Comparative Analysis of Large Animal Models

An ideal, clinically relevant animal model of CHF should mimic both the etiology of a common form of human CHF and parallel its natural history. Each available model has advantages and limitations and none is ideal (Table 1) (8). Sheep and swine have advantages over dogs in that the coronary anatomy closely mimics that of humans, whereas dogs have an extensive coronary collateral supply and much faster heart rates. The relatively large size of sheep and swine, their consistent coronary anatomy and vasomotor responsiveness makes them suitable for application of multiple diagnostic and therapeutic strategies. The following sections will first review the commonly employed large animal species used in CHF research, followed by reviews of the methods of evaluating cardiovascular function and the ways employed to create CHF.

Canine models

Until recently, the most extensively used and widely reported large animal models of CHF were in dogs. Accordingly, there is a substantial literature related to canine physiology and pathophysiology. Multiple models of heart failure have been reported and dogs are very well suited to chronic studies with indwelling instrumentation in the conscious state (14-17). However, the canine myocardium is rich with collateral perfusion, and collateralization is further stimulated by ischemia (18); thus, dog models of ischemia-induced injury can be challenging to establish as they often do not demonstrate substantial dysfunction (19). In contrast to humans, dogs are always left-dominant (20). In addition, dogs possess a more narrow and deep chest cavity as compared to the more barrel-shaped chest of humans; this may influence surgical approach or cardiac assist device placement. However, they have very compliant skin and, when needed, electrical wires and tubes used for chronic instrumentation can easily be buried in subcutaneous pocket for repeated use over long durations of time (21).

Porcine models

Pigs are a popular species for animal model development in cardiovascular research because of their suitable cardiac and vascular anatomy. Pigs have the similar aortic valve as humans and coronary artery distribution and perfusion is right dominant, whereas 72–85% of human hearts are also right dominant (22). Pigs have only two pulmonary vein Ostia in the left atrium and two vessels that branch from the aortic arch, the brachiocephalic trunk and the left subclavian artery (23). The brachiocephalic artery branches into the right subclavian and the left and right common carotid arteries. Pigs are difficult to use in chronic studies requiring chronic instrumentation because they are very active, their skin is tough (which makes subcutaneous burying of cables and tubes difficult) and they grow fast. The peripheral vessels of pigs are small and deep, presenting some challenges in vascular access particular in studies where repeated procedures or sampling is involved.

Ovine Models

The sheep is considered to be a suitable model for cardiovascular research because of its ease of handing, the size of the heart and the chest cavity, and vascular anatomy and hemodynamic responses which bears close resemblance to the human. Sheep do not have a prevalent network of coronary collaterals; thus, infarction-based studies result in distinct injuries with sharp border zone regions of ischemic versus perfused territories (24). Sheep are left-dominant, and the left and right coronaries perfuse the LV and RV, respectively, with only minor overlap. Compared to human hearts, the normal sheep heart is spade-shaped which sometimes poses a problem for instrumentation of the ventricular cavity. Unlike dogs and pigs, however, sheep can be stanchioned, making them very suitable for chronic instrumentation with leads or tubes exiting from their dorsum. Sheep often have substantial amounts of wool and need to be shorn regularly, particularly in long-term studies or where body weight or conditioning plays a role in the data (1).

Bovine models

The structure of a bovine (calve) heart is really the same as human heart except for a bigger size. Calves (4–6 months) are more suitable than adult cattle for approximating the human body size (70–100 kg) and are physically easier to handle by study personnel. Normal calves are the industry standard to test safety, performance, reliability, and efficacy of ventricle assist devices (VAD) because phylogenetically lower species are too small to accommodate implantable cardiac assist devices (25-28). However, a stable and reproducible model of heart failure has been limited in calves due to high morbidity and mortality rates (29), inconsistent methodologies, and extensive resource requirements such the cost of acquisition and maintenance, as well as vivarium and surgical space large enough to accommodate animal model size. In addition, the significant growth of the weight and corresponding increased cardiac output of the calves are potentially limited the duration of the VAD studies.

C. Methods of Evaluating in HF Models

Animal models of CHF should be carefully characterized to ensure that they have critical features common to human CHF. There has been nearly half a century of technological development for assessing cardiovascular function in large animals. Importantly, essentially method that can be used in humans can be applied to large animal models and should be considered. The types of measurements made in any given study should depend on the goals of the analysis, which parameters are most relevant to study. Furthermore, it is noteworthy that the hemodynamic characterization is more physiological in large animal models when the data are obtained in the conscious rather anesthetized state.

The methods that are recommended to assess large-animal cardiovascular function include:

Invasive hemodynamic assessment in HF

Routine hemodynamic parameters, including cardiac output, ventricular systolic and diastolic pressures, chamber end-systolic and end-diastolic volume, ejection fraction (EF), LV dP/dtmax (left ventricular maximal rate of pressure rise) and LV dP/dtmin (left ventricular maximal rate of pressure decline) can be measured via either by implanted sensors (14-16,29) or by acute catheterization under anesthesia (30). In animal studies, a surgical intervention can be performed before the induction of CHF to insert sonomicrometric crystals, fluid-filled catheters, ultrasound probes and micromanometers. Recording can be made by connecting a single electric plug to a skin-button where all internal sensor wires merge, or by radio telemetry. These approaches allow cardiac structural and functional changes to be measured in conscious animals during the progression of induced disease and after a therapeutic intervention (14-16,29). Cardiac indices can be measured using catheters in a cardiac catheterization laboratory. The most common methods involve placement of balloon-tipped catheters into the pulmonary artery, and micromanometer catheters placed in arterial or ventricular cavities to measure hemodynamics. It has been reported that large animals have similar cardiac catheterization measurements as in humans (31).

The most universally accepted index of contractility used in practice is EF. However, LV EF depends on ventricular contractility, afterload, and preload. Preload is related to end-diastolic volume (EDV) and pressure; the relationship between these 2 parameters (the end-diastolic pressure-volume relationship, EDPVR) indexes the degree of ventricular remodeling. Although EF is known to correlate with mortality in the subset of heart failure patients with reduced EF, it is previously unknown which of its determining factors contribute most importantly to prognosis. Cardiac pressure-volume (PV) analysis has been the most powerful approach for addressing such a problem (32-33). The relationship between end-systolic pressure and volume from a variety of variably loaded cardiac contractions yields the end-systolic pressure–volume relationship (ESPVR), its slope being the end-systolic elastance (Ees). Ees is a load-independent index and conveys information about both contractile function and myocardial constitutive properties. Basic and clinical physiologists considered PV analysis to be the gold standard for assessing ventricular properties. The EDPVR was determined by applying non-linear regression analysis to the end-diastolic pressure and volume points (Ped and Ved, respectively). These data were fit to the following equation: Ped = ßeαVed, where α is the chamber stiffness constant, and ß is a scaling constant. In primary diastolic heart failure, diastolic pressure-volume relation (dashed line) shifts upward and to the left, indicating a disproportionate and a greater increase in diastolic pressure for any increase in diastolic volumes (Fig. 1) (34).

Figure 1. Schematic diagram of pressure-volume relations in normal, systolic and diastolic heart failure. In systolic heart failure, a downward and rightward shift of the end-systolic pressure-volume line indicates decreased contractile function. This is the principal cause of reduced ejection fraction and forward stroke volume. In primary diastolic heart failure, diastolic pressure-volume relation (dashed line) shifts upward and to the left, indicating a disproportionate and a greater increase in diastolic pressure for any increase in diastolic volumes. If there is also a decrease in end-diastolic volume, then a decrease in stroke volume also occurs.

Noninvasive imaging methodologies

Echocardiography

Echocardiography is a rapid, safe, and widely available technique that allows the assessment of chamber dimensions, wall thickness and geometry, indices of regional and global, systolic and diastolic ventricle function in conscious or sedated state. Echocardiography also provides rapid and semi-quantitative assessment of valvular function, especially mitral, tricuspid and aortic stenosis and regurgitation. LV EF is a standard and important measurement in distinguishing systolic dysfunction of HF and monitoring the effects of drugs and devices on the heart. The ASE (American Society of Echocardiography) recommends (35) that for most research purposes, EF should be measured from LV volumes with the method of discs (Simpson’s Rule) whenever good apical views can be obtained. If poor resolution apical images do not allow identification of endocardium or are foreshortened, the method of multiple diameters may be used to measure the LV volumes from the parasternal long-axis views. In animal studies, due to the different anatomy of the chest cavity, standard parasternal long- and short-axis planes and apical views are acquired in the right lateral position while conscious or under anesthesia. In addition, three-dimensional (3D) echocardiography has shown better accuracy than 2D echocardiography and the real-time 3D acquisition and measurement may become the future method of choice for measuring LV EF in preclinical and clinical studies. Measures of diastolic function are important in patients of heart failure with preserved systolic function (LVEF ≥45-50%) and hypertension, as well as in preclinical animal studies. At the present time, a multitude of Doppler-derived parameters have been used to evaluate diastolic function. Information from mitral inflow, pulmonary vein flow, LV inflow propagation velocity, and mitral annular diastolic velocity can be used to determine 4 patterns of diastolic function that represent stages of relaxation and compliance abnormality (I: abnormal relaxation; II: pseudo-normalization; III: Reversible restriction; and grade IV: Irreversible restriction) (36).

Speckle tracking echocardiography (STE) is a novel technique that enables the assessment of myocardial strain through the analysis of speckle motion inherently present in a standard, 2-D echocardiographic image. In contrast to tissue Doppler-based assessment of strains, the STE process is angle independent and is fewer operators dependent. Experimental and clinical studies have demonstrated that the STE method can assess myocardial function accurately in healthy subjects in the settings of acute and chronic ischemia, dyssynchrony, and cardiomyopathy (37-40). This new approach should be considered and used where appropriate.

Cardiac magnetic resonance imaging (CMR)

CMR is a versatile, highly accurate, and reproducible imaging technique for the assessment of left and right ventricular volumes, global function, regional wall motion, myocardial thickness, thickening, myocardial mass, and cardiac valves (41-42). CMR consists of several techniques that can be performed separately or in various combinations during an examination. Cine-CMR provides assessment of cardiac morphology and function, first-pass contrast-enhanced perfusion-CMR with and without vasodilators can provide assessment of myocardial perfusion reserve, and delayed contrast enhanced CMR (DE-CMR) can be used for non-invasive tissue characterization. DE-CMR is effective in detecting the presence, location, and extent of myocardial infarction, and predicting improvement in contractile function following revascularization in both the acute and chronic settings. Moreover, DE-CMR is capable of visualizing micro scars that cannot be detected by other imaging techniques. CMR is fewer operators dependent when compared with echo and images can be obtained when echo proves sub-optimal because of a poor acoustic window in large animals due to the more vertical orientation of the heart relative to the thoracic cavity. CMR offers a comprehensive assessment of heart failure patients and is now the gold standard imaging technique to assess myocardial anatomy, regional and global function, and viability in both pre-clinical experimental (43-44) and clinical studies.

Cardiac CT

CMR is the gold standard for cardiac structure and function evaluation. However, due to the duration and multiple prolonged breath-holds required to obtain adequate images, inability to perform the study in patients with implantable cardioverter defibrillators and pacemakers, MRI is not always feasible. The current generation of multidetector CT (MDCT) scanners with improved temporal and spatial resolution has allowed noninvasive imaging of coronary arteries and provides information on left ventricular structure and function, cardiac venous anatomy, the pulmonary venous system, and right ventricular function (45-46). Schuleri KH et al (47) assessed therapeutic effects of myocardial regenerative cell therapies in a pig MI model using MDCT. Infarct size, end-diastolic LV volume, and LVEF assessed by MDCT compared favorably with those assessed by cardiac magnetic resonance acquisitions.

Biomarkers in HF

Both BNP and NT-pro BNP are currently used clinically to aid the diagnosis of CHF, assessing the severity of CHF and risk stratification in patients with coronary artery diseases (48). The value of natriuretic peptides in assessment of patients with possible new CHF was evaluated by Cowie et al (49) Geometric mean concentrations of natriuretic peptides were much higher in patients with heart failure than in those with other diagnoses (29.2 vs 12.4 pmol/L for ANP; 63.9 vs 13.9 pmol/L for BNP; 1187 vs 410.6 pmol/L for NT-ANP; all p<0.001). Although all peptides correlated with the diagnosis, BNP had the highest sensitivity (97%) and specificity (84%) for the diagnosis of CHF. Plasma BNP or NT-Pro-BNP provides prognostic information in CHF (50). Fig. 2A shows the BNP concentrations and ANP concentrations in relation to the severity of heart disease according to the NYHA BNP concentrations in patients with heart disease or hypertension and classification of cardiac function (51). The plasma BNP values, but not ANP values, were higher even in patients with NYHA class I than in control subjects. In a pig tachycardia-induced HF model (Fig. 2B), BNP level was determined in sera from 3 individuals (control pig, pigs with moderate and severe HF) at 0 to 28 week of RV pacing. An increase of serum BNP was noted in moderate and severe HF pig, whereas serum BNP remained stable up to 28 week in a control pig (52), Kass P et al (53) reported that dogs with moderate to severe mitral regurgitation due to myxomatous mitral valve disease (MVD) have increased plasma BNP concentration compared to normal dogs. Plasma BNP was significantly high in the dogs with MVD and no radiographic evidence of CHF (group I) when compared to the control dogs. Additionally, plasma BNP was greater in dogs with MVD and CHF (groups II–IV) than in control dogs as well as in dogs with MVD only (group I) (Fig. 2C).

Figure 2. A: The plasma brain natriuretic peptide (BNP) concentrations and atrial natriuretic peptide (ANP) concentrations in relation to the severity of heart disease according to the NYHA classification of cardiac function. Reproduced from Hirata Y et al Cardiovasc Res 2001;51:585-591 with permission from Oxford University Press. B: Serum BNP concentration in control, moderate and severe heart failure (HF) pigs. Reproduced from Paslawska U. Int J Cardiol. 2011;153(1):36-41 with permission from Elsevier. C: Distribution of plasma BNP concentration in control dogs and dogs with heart disease. Reproduced from MacDonald KA. J Vet Intern Med. 2003;17(2):172-7 with permission from John Wiley and Sons.

D. Large Experimental Models of CHF

Ischemia/infarction induced CHF

Coronary artery micro-embolization

Coronary microembolization has been used extensively for studies of acute and chronic heart failure. Serial injections of 90 micron diameter polystyrene microspheres into the left main coronary artery over an ~8 week period lead to a reduced ejection fraction, increased LVEDP and elevated plasma norepinephrine levels. The most extensive experience with this model is in dogs, followed by sheep. In the canine model of CHF, dogs are subjected to multiple coronary artery embolic procedures performed serially over an 8 week period (54). Embolizations proceeded until LVEDP reached 15 mmHg and LV dP/dt is less than 2400 mmHg/s. After the creation of reduced LV systolic function and dilation, progression of LV dilation and dysfunction (i.e., remodeling) occurs during subsequent long term follow up. LV remodeling is apparent after embolization, as evidenced by the rightward shift of the pressure-volume loops (PV loops, Fig. 3) (54). LV PV relationship loops were measured during a transient preload reduction induced by inferior vena cava occlusion. Using the end-systolic pressure and volume data points, the end-systolic PV relationship (ESPVR) was assessed using a linear regression to the equation: Ees = Pes/ (Ves -V0), where the slope (end-systolic elastance, Ees) defines contractility and V0 its volume-axis intercept. As shown in Fig. 1, Ees reduced at CHF (after 8 weeks coronary embolization). LV remodeling was sustained in control animals without the LV passive support device (LVSD) (Fig. 3 Right panel, top) but was partially reversed by LVSD treatment, as demonstrated by the leftward shifting of the PV loops at 8 weeks (Fig. 3 Right panel, bottom). EDPVR was determined by applying non-linear regression analysis and chamber stiffness (α) increased significantly with coronary embolization and a trend to reduction was seen after 8 weeks of LVSD therapy versus control.

Figure 3. Dog coronary artery micro-embolization model treated with a Left ventricular passive support device (LVSD). Compared with control animals, LVSD proved effective in left ventricle (LV) volume reduction. Left top: Paracor device placed over left and right ventricles. Left bottom: Left ventricular end-systolic volume (ESV) and end-diastolic volume (EDV) after left ventricular support device therapy. Right: Representative pressure–volume (PV) loops at inferior vena cava occlusion; A: control; B: LVSD.

Embolization results in permanent obstruction of the coronary microvessels, with the magnitude of injury influenced by the size and volume of microspheres delivered. What start out initially as small microinfarcts evident only on histological examination ultimately turn into large areas of dense infarction evident on visual examination as the number of delivered microemboli builds over time. This finding indicates that with repeated microembolizations, the collateral circulation of dogs is eventually exhausted and results in irreversible myocardial damage, resulting in left ventricular dysfunction and compromised coronary flow reserve (14-16, 54). Because microsphere injections are into the left main coronary artery, this model reproduces a global infarct model, not a localized infarction with development of an aneurysm. In these aspects, the embolization model resembles human ischemic cardiomyopathy.

The microembolization model is technically complex to produce, requiring serial surgical interventions. An alternative is for implantation of a chronic, indwelling coronary catheter for delivery of microspheres, but this too is technically challenging to maintain over the ~8 week period of CHF induction. Recently, Koenig et al (29) reported that in calves, percutaneously injected 90 μm microspheres into the left coronary artery induced stable and reproducible CHF. 11 of the 17 animals (survival rate: 65%) survived coronary microembolization and exhibited significant echocardiographic and hemodynamic signs of severe systolic dysfunction at 60 days after coronary micro-embolization. EF decreased from a baseline conscious value of 81 ± 1% to a final conscious value of 43 ± 6%; and a significant increase in LVEDP accompanied a significant decrease in LVP, -dP/dT and AoP. The LV PV loops showed increased end-systolic and end-diastolic volume, reduced stroke volume, and reduced left ventricular work after coronary microembolization (Fig. 4) (29). The microembolization model has been used to investigate a multitude of pharmacologic, electrical and surgical therapies for the treatment of CHF.

Figure 4. Calf coronary artery micro-embolization model. Left: During a typical microembolization session, 90-μm microspheres injected into the left main coronary artery produced aortic pressure (A), electrocardiographic (B) and pressure-volume relationship changes (C). Right: Echocardiography shows left ventricular ejection fraction and volume changes after coronary microembolization.

Coronary artery ligation / coil-embolization

Coronary ligation is a popular means of inducing MI and can be performed on a selected coronary artery or arteries to infarct targeted regions of myocardium or specifically induce ischemic-based mitral valve dysfunction. In pigs and sheep, coronary ligation yields infarcts with sharp, distinct border zones, which are useful for studies examining differences between infarcted myocardium, border zone regions, and remote tissue areas within the same heart (55-57). Although coronary ligation results in an immediate ischemic insult to the heart, the placement of coronary ligation materials and ameroid constrictors requires invasive surgical procedures. The thoracotomy or sternotomy performed results in a significant chest incision, which requires attentive care of the animal to minimize pain, stress, and the potential for infection and disrupts the chest cavity. In contrast to surgical procedures (55) (Fig. 5, top), percutaneous transcatheter complete occlusion of coronary arteries via coil-embolization (6) (Fig. 4, bottom) has been developed to establish large animal models of ischemic HF and apical aneurysms for preclinical left ventricle support and reconstruction therapy. Although widely used, this model has historically been plagued by a 30% mortality rate due to fatal arrhythmias, most commonly ventricle fibrillation. We have reported that by using a standardized protocol, the mortality rate can be greatly reduced (58). Under fluoroscopic guidance, the 2.0~3.5 mm coronary coils were delivered to the middle LAD and the complete and persistent occlusion was confirmed by coronary angiography (Fig. 5, bottom: A) 8 weeks after MI creation, large transmural infarction areas were shown in the LV anterior and anteroseptal wall (Fig. 5, bottom: B). This MI model was used to evaluate an epicardial catheter-based ventricular reconstruction procedure (6,40). Speckle tracking echocardiographic analysis showed the longitudinal and circumferential strains were significantly reduced after anteroapical infarct and significantly improved after device implantation (Fig. 6 Top) (40). Histological examination shows the scar was complete exclude with the device implantation (Fig. 6. Bottom).

Figure 5. Sheep myocardial infarct (MI) model. Top: Ligate the left anterior descending (LAD) and its diagonal branches 40% of the distance from the apex to the base of the heart. Reprocuded from Gorman RC et al. Ann Thorac Surg 2009; 87:794-802 with permission from Elsevier. Bottom: MI was induced by percutaneous coil embolization of LAD and its diagonal branches. Reproduced from Cheng et al. J Thorac Cardiovasc Surg. 2014 July;148(1):225-31 with permission from Elsevier.

Figure 6. Epicardial catheter-based ventricular reconstruction procedure in a chronic coil-embolization induced anteroapical aneurysm ovine model. Top: The longitudinal (A) and circumferential (B) strains changes before and after device implantation by speckle tracing echocardiography. Bottom: Microscopic tissue examination of the implant sites (H&E Stain, 10x). Reproduced from Cheng Y et al. Interact Cardiovasc Thorac Surg. 2013 Dec;17(6):915-22 with permission from Oxford University Press.

Ligation of the second and third oblique branches (OM2 and OM3) from the circumflex coronary artery (LCX) (59-60) induced a ~20% infarction in the posterior wall and the papillary muscle. Post-infarction LV remodeling causes mitral leaflet tethering and annular dilatation. Three-dimensional echocardiography examination of mitral valve geometry showed that tethering length (distance from mitral valve to infarcted PM), tethering volume and leaflet closing area significantly increased with LV dilation (Fig. 7 Top right & Bottom). Moderate mitral valve regurgitation commonly develops within 8 weeks (Fig. 7 Top left) (60). This model reproduces the acute and chronic changes of the ischemic-based functional mitral valve dysfunction without damage to the mitral valve apparatus.

Figure 7. Sheep ischemic mitral regurgitation (MR) model induced by ligation of the second and third circumflex obtuse marginal branches. Top Left: moderate ischemic MR developed after ligation of left circumflex branches and the MR decreased after injection of polyvinyl-alcohol (PVA) hydrogel into the infarcted myocardium. Top Right: Real-time 3D echocardiography showing PVA visualized within the myocardium (red circle). Bottom left: Ejection fraction and LV volume changes. Right: Changes in mitral valve geometry at baseline, post–chronic MR, and PVA stages. Reproduced from Solis J et al, Circ Cardiovasc Interv. 2010 Oct;3(5):499-505 with permission from Wolters Kluwer Health.

Ischemia / reperfusion

Coronary flow can be occluded internally via balloon catheters passed minimally invasively into the coronary artery of choice under fluoroscopic guidance. Depending on the occlusion time and how proximal the occlusion is or the absence of coronary collaterals, the resultant infarct can be different. Ghugre et al. (61) reported that in a pig model, animals were subjected to coronary balloon occlusion for either 90 or 45 min, followed by reperfusion. Week-6 post-procedure, the 90-min model produced large transmural infarcts with hemorrhage and microvascular obstruction, while the 45 min produced small nontransmural and nonhemorrhagic infarction. In addition, signs of adverse remodeling were evident in the 90-min occlusion group only (Fig. 8).

Figure 8. Pig balloon-occlusion and reperfusion myocardial infarct (MI) model. Top: Ejection fraction (EF) and end diastolic volume (EDV) evolution. Cumulative time course of EF (a) and EDV (b) post-acute myocardial infarct (AMI) compared in the 90- and 45-min groups; error bars show standard deviation. *P = 0.05 compared with control (day 0); P <0.05 for comparison between 90- and 45-min models at rest. Bottom: Histology and MRI. Representative short-axis LGE images containing the infarct zone are shown for the 90- and 45-min models. Reproduced from Ghugre NR et al. Magnetic Resonance in Medicine 2013 70:1095–1105 with permission from John Wiley and Sons.

Tachycardia-induced CHF

Dilated cardiomyopathy (DCM) is characterized by structural hallmarks of LV dilation and increased LV chamber radius-to-wall thickness ratio in the absence of coronary artery disease of infarction. The most well characterized large animal non-infarct-related model of CHF is the pacing induced tachycardia model achieved through implantation and programming of specially modified pacemakers (63-68).

Figure 9. Myocardial infarct (MI) was induced by percutaneous balloon occlusion of the LAD and followed by reperfusion in sheep. The infarct and border zones were identified at 4 weeks after MI creation by 2D and 3D echocardiography (left). Stem cells were injected into the target region via a delivery catheter under 3D echo guidance (left: A-D). Right: The wall thickness change before and after cells injection were seen in the region of injections (arrows). Reproduced from Cheng Y et al. Cell Transplant 2013;22(12):2299-309 with permission from Cognizant Communication Corporation.

Different pacing protocols have been implemented to induce CHF. Wang and associates (69) induced heart failure in dogs with rapid pacing at 210 bpm for 3 weeks and then 240 bpm during week 4. These dogs develop CHF characterized by typical hemodynamic abnormalities, blunted endothelium-mediated vasodilator function in coronary and femoral circulations, and decreased gene expression of endothelial constitutive nitric oxide synthase (ECNOS) (Fig. 10). In addition to the structural and functional changes in the myocardium, chronic tachycardia also induces abnormalities in gene expression, disruption of the extracellular matrix, and neurohormonal changes very similar to those observed in human heart failure. Additionally, apoptosis has been reported as one of the causes of myocyte loss and ventricular dysfuction.

Figure 10. Dog rapid pacing Myocardial infarct (MI) model. Left: Responses of epicardial coronary diameter (CD) and coronary blood flow (CBF) to treadmill exercise (EX) challenge were blunted in dogs with cardiac pacing alone. Right: Northern blot shows that in dog with cardiac pacing plus daily exercise training (Pacing+Exercise), aortic ECNOS gene expression is preserved compared with untrained dog (Pacing). Reproduced from Wang J. Circulation. 1997; 96(8):2683-92 with permission from Wolters Kluwer Health.

However, it is very important to note that ventricular systolic and diastolic function recover in this model within 2 to 3 weeks of cessation of rapid pacing, though hypertrophy and a certain degree of ventricular dilation typically persists. On the other hand, continued rapid pacing for more than 10-12 weeks results in progressively deteriorating heart failure and high mortality.

Advantages of the rapid pacing model include neurohormonal alterations consistent with those observed in humans, generation of predictable degrees of LV dilation and pump dysfunction, and the ability to test pharmacologic strategies aimed at attenuating progression of LV dysfunction to HF.

Pressure-overload induced hypertrophy

Pressure overload induces cardiac dysfunction that progresses into compensatory hypertrophy and eventual HF via increased afterload. Techniques used to impose a pressure overload in large animal models have included renal artery construction, pulmonary artery or aortic banding (70-74). Production of an aortic banding model (73) involves dissecting the ascending aorta free from the pulmonary artery and placing an encircling Dacron (C.R. Bard, Covington, GA) patch 1.5 to 2.0 cm wide and 6 to 7 cm long to reduce the diameter of the aorta by ~50%. Aortic and left ventricular pressures are then rechecked to confirm that a 50- to 60-mm Hg systolic pressure gradient has been created. An alternative is to banding is implantation of a hydraulic occlude that can be adjusted as desired over time. An advantage of these pressure-overload models is that the constriction can be reversed with either removal the band or alleviation of the hydraulic constrictor. Although this model produces marked left ventricular hypertrophy, it does not reproduce neurohormonal activation or left ventricular systolic dysfunction. Nevertheless, the increased LV stiffness and reduced LV relation and filling associated with these models render them appropriate for studies of myocardial hypertrophy. The relations among hypertrophy, fibrosis and diastolic performance were examined in an early hypertension dog model (75). The perinephritic hypertension was induced by silk wrap of one kidney with contralateral nephrectomy. After in 12 weeks, blood pressure increased from 148 to 235 mm Hg. In association with hypertension, there were increases posterior wall thickness and left ventricular mass (90 ± 26 to 113 ± 20 g). There were no changes in left ventricular cavity size or systolic performance. Diastolic function was impaired in dogs with left ventricular hypertrophy, with decreased Doppler early to atrial inflow velocity ratio (E/A) (1.35 versus 1.72), increased atrial Dling fraction (35% versus 29%), decreased sonomicrometric peak rates of wall thinning (-2.01 versus -3.37Iiters/s) and filling (4.33 versus 6.64 Iiters/s) and prolonged time constant of isovolumetric relaxation (tau; 34.3 versus 28.1 ms) (75).

Volume-overload induced CHF

Volume overload of the heart is used to simulate the progressive response of the myocardium, vasculature, and neurohormonal system to augmented preload as a result of pump failure. Several experimental techniques used to induce volume overload include the surgical creation of an anteriovenous fistula (carotid artery to jugular vein or via femoral shunt), and induction of mitral valve regurgitation (59-60, 76-78). Carabello (79) reported a dog model through disruption of the mitral valve chordate and creation of mitral regurgitation (Fig. 11). In our lab, a chronic LV volume overload model was created in a sheep model of aortic regurgitation through balloon-dilation of the aortic leaflets using a catheter-based technique under fluoroscopic guidance. In 10 weeks, LVEDV increased from 41 ml to 192ml and EF decreased from 59% to 32% (Fig. 12; Data not published). Chronic volume overload has been recognized as an important determinant of the progression of heart failure; however, these models do not have alterations in the myocardial structure observed in CHF due to ischemia, or hypertrophy.

Figure 11. A canine model of volume overload produced by mitral regurgation. Left: A flexible grasping forceps was positioned in left ventricle (LV) via a 7.5 Fr sheath in carotid artery to disrupt mitral chordae or leaflet. Right: LV end diastolic volume (EDV) (A) and end systolic volume (ESV) (B) over time. Reproduced from Kleaveland JP. Am J Physiol. 1988 Jun;254(6 Pt 2):H1034-41 with permission from The American Physiological Society.

Figure 12. A sheep model of volume overload produced by aortic regurgitation (AI). Left: A 12 mm balloon dilated one of the aortic leaflet and created severe aortic reguigation. Right: LV end diastolic volume (EDV) and end systolic volume (ESV) increased at 10 weeks after AI creation.

E. Summary

Experimental studies are an essential part of the successful development of HF device technologies. Non-diseased models are adequate to the validation and development of trans-catheter techniques (5). However, CHF is a clinical syndrome in which pathophysiologic underpinnings included LV dysfunction, remodeling, and increased neurohormonal activations. CHF animal models have been used in gene therapy, stem cell therapy, left ventricular reshaping devices and mechanical support devices. According, large animal constructs must be developed that mimic this disease process in order to demonstrate efficacy, investigate physiologic response, elucidate genetic, molecular, and cellular mechanisms, and develop new treatment strategies. In vivo testing of a new device requires two noteworthy components: First, relevant translational animal models should be carefully selected getting as close as possible to the human condition. It must be recognized that there is no one perfect animal model of HF and carefully individualized analysis is required to determine the appropriate model for the specific issue being investigated. Second, adverse outcomes should not be disregarded (i.e., thrombosis), usually experimental outcomes predict safety events. Large animal models recapitulating the clinical HF phenotype and translating basic science to clinical applications have successfully traveled the journey from bench to bedside. Appropriate design and execution of the pre-clinical studies is essential, rarely, a well conducted pre-clinical validation program leads to adverse clinical and financial outcomes.

References

1. Monreal G, Sherwood LC, Sobieski MA, Giridharan GA, Slaughter MS, Koenig SC. Large animal models for left ventricular assist device research and development. ASAIO J. 2014 Jan-Feb;60(1):2-8.
2. Zheng Y, Sampaio LC, Li K, Silva GV, Cabreira-Hansen M, Vela D, Segura AM, Bovè C, Perin EC. Safety and feasibility of mapping and stem cell delivery in the presence of an implanted left ventricular assist device: a preclinical investigation in sheep. Tex Heart Inst J. 2013;40(3):229-34
3. Fukamachi K, Shiose A, Massiello AL, Horvath DJ, Golding LA, Lee S, Starling RC. Implantable continuous-flow right ventricular assist device: lessons learned in the development of a cleveland clinic device. Ann Thorac Surg. 2012 May;93(5):1746-52.
4. Fukamachi K, Inoue M, Doi K, Schenk S, Nemeh H, Dessoffy R, Kopcak MW Jr, McCarthy PM. Device-based left ventricular geometry change for heart failure treatment: developmental work and current status. J Card Surg. 2003 Jul-Aug;18 Suppl 2:S43-7.
5. Yi GH, Cheng Y, Aboodi MS, Farnan R, Kar S, Burkhoff D, Farnan G, Granada JF. Safety and feasibility of percutaneous delivery of a novel circulatory assist device (CircuLite® SYNERGY®) in the swine model. EuroIntervention. 2013 Jun 22;9(2):259-68
6. Cheng Y, Aboodi MS, Wechsler AS, Kaluza GL, Granada JF, Van Bladel K, Annest LS, Yi GH. Epicardial catheter-based ventricular reconstruction: a novel therapy for ischaemic heart failure with anteroapical aneurysm. Interact Cardiovasc Thorac Surg. 2013 Dec;17(6):915-22.
7. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, et al. Executive summary: heart disease and stroke statistics–2014 update: a report from the American Heart Association. Circulation. 2014 Jan 21;129(3):399-410.
8. Katz MG, Fargnoli AS, Tomasulo CE, Pritchette LA, Bridges CR. Model-specific selection of molecular targets for heart failure gene therapy. J Gene Med. 2011 Oct;13(10):573-86.
9. Dixon JA, Spinale FG. Large animal models of heart failure: a critical link in the translation of basic science to clinical practice. Circ Heart Fail. 2009 May;2(3):262-71.
10. Yarbrough WM, Spinale FG. Large animal models of congestive heart failure: a critical step in translating basic observations into clinical applications. J Nucl Cardiol. 2003 Jan-Feb;10(1):77-86
11. Recchia FA, Lionetti V. Animal models of dilated cardiomyopathy for translational research. Vet Res Commun. 2007 Aug;31 Suppl 1:35-41
12. Houser SR, Margulies KB, Murphy AM, Spinale FG, Francis GS, Prabhu SD, et al. Animal models of heart failure: a scientific statement from the American Heart Association. Circ Res. 2012 Jun 22;111(1):131-50.
13. Savino JA 3rd, Kosmas CE, Wagman G, Vittorio TJ. Evolution of the chronic congestive heart failure paradigm. Cardiol Rev. 2013 May-Jun;21(3):121-6.
14. Wolff MR, de Tombe PP, Harasawa Y, Burkhoff D, Bier S, Hunter WC, Gerstenblith G, Kass DA. Alterations in left ventricular mechanics, energetics, and contractile reserve in experimental heart failure. Circ Res. 1992 Mar;70(3):516-29
15. Todaka K, Leibowitz D, Homma S, Fisher PE, Derosa C, Stennett R, et al. Characterizing ventricular mechanics and energetics following repeated coronary microembolization. Am J Physiol. 1997 Jan;272(1 Pt 2):H186-94.
16. Todaka K, Wang J, Yi GH, Knecht M, Stennett R, Packer M, Burkhoff D. Impact of exercise training on ventricular properties in a canine model of congestive heart failure. Am J Physiol. 1997 Mar;272(3 Pt 2):H1382-90.
17. Shimizu M, Tanaka R, Fukuyama T, Aoki R, Orito K, Yamane Y. Cardiac remodeling and angiotensin II-forming enzyme activity of the left ventricle in hamsters with chronic pressure overload induced by ascending aortic stenosis. Vet Med Sci. 2006 Mar;68(3):271-6
18. Scheuer J. Effects of physical training on myocardial vascularity and perfusion. Circulation. 1982 Sep;66(3):491-5.
19. Kim WG, Shin YC, Hwang SW, Lee C, Na CY. Comparison of myocardial infarction with sequential ligation of the left anterior descending artery and its diagonal branch in dogs and sheep. Int J Artif Organs. 2003 Apr;26(4):351-7.
20. Oliveira CLS, David GS, Carvalho MO, Dornelas D, Araújo S, et al. Anatomical Indicators of Dominance between the Coronary Arteries of Dogs. Int. J. Morphol. 2011, 29:845-849.
21. Yi GH, Burkhoff D, Zhang H, Zhu SM, Zwas D, Wang J. Hemodynamic effects of a calcium channel promoter, BAY y 5959, are preserved after chronic administration in ischemic heart failure in conscious dogs. J Pharmacol Exp Ther. 1998 Aug;286(2):760-6
22. Weaver ME, Pantely GA, Bristow JD, Ladley HD. A quantitative study of the anatomy and distribution of coronary arteries in swine in comparison with other animals and man. Cardiovasc Res. 1986 Dec;20(12):907-17.
23. Crick SJ, Sheppard MN, Ho SY, Gebstein L, Anderson RH. Anatomy of the pig heart: comparisons with normal human cardiac structure. J Anat. 1998 Jul;193 ( Pt 1):105-19.
24. Gorman JH 3rd, Gorman RC, Plappert T, Jackson BM, Hiramatsu Y, St John-Sutton MG, et al. Infarct size and location determine development of mitral regurgitation in the sheep model. J Thorac Cardiovasc Surg. 1998 Mar;115(3):615-22.
25. Koenig SC, Giridharan GA, Ewart DL, Schroeder MJ, Ionan C, Slaughter MS, et al. Human, bovine and porcine systematic vascular input impedances are not equivalent: implications for device testing and xenotransplantation in heart failure. J Heart Lung Transplant. 2008 Dec;27(12):1340-7.
26. Koenig SC, Litwak KN, Giridharan GA, Pantalos GM, Dowling RD, Prabhu SD, et al. Acute hemodynamic efficacy of a 32-ml subcutaneous counterpulsation device in a calf model of diminished cardiac function. ASAIO J. 2008 Nov-Dec;54(6):578-84.
27. Jahanmir S, Hunsberger AZ, Heshmat H, Tomaszewski MJ, Walton JF, Weiss WJ, et al. Performance characterization of a rotary centrifugal left ventricular assist device with magnetic suspension. Artif Organs. 2008 May;32(5):366-75.
28. Kindo M, Radovancevic B, Gregoric ID, Conger JL, Kadipasaoglu K, Tamez DA, et al. Biventricular support with the Jarvik 2000 ventricular assist device in a calf model of pulmonary hypertension. ASAIO J. 2004 Sep-Oct;50(5):444-50.
29. Bartoli CR, Sherwood LC, Giridharan GA, Slaughter MS, Wead WB, Prabhu SD, et al. Bovine model of chronic ischemic cardiomyopathy: implications for ventricular assist device research. Artif Organs. 2013 Dec;37(12):E202-14.
30. Liu Y1, Yue WS, Liao SY, Zhang Y, Au KW, Shuto C, et al. Thoracic spinal cord stimulation improves cardiac contractile function and myocardial oxygen consumption in a porcine model of ischemic heart failure. J Cardiovasc Electrophysiol. 2012 May;23(5):534-40.
31. Milani-Nejad N, Janssen PM. Small and large animal models in cardiac contraction research: advantages and disadvantages. Pharmacol Ther. 2014 Mar;141(3):235-49.
32. Borlaug BA, Kass DA. Invasive hemodynamic assessment in heart failure. Cardiol Clin. 2011 May;29(2):269-80.
33. Burkhoff D, Mirsky I, Suga H. Assessment of systolic and diastolic ventricular properties via pressure-volume analysis: a guide for clinical, translational, and basic researchers. Am J Physiol Heart Circ Physiol. 2005 Aug;289(2):H501-12.
34. Chatterjee K, Massie B. Systolic and diastolic heart failure: differences and similarities. J Card Fail. 2007 Sep;13(7):569-76.
35. Gottdiener JS1, Bednarz J, Devereux R, Gardin J, Klein A, Manning WJ, et al. American Society of Echocardiography recommendations for use of echocardiography in clinical trials. J Am Soc Echocardiogr. 2004 Oct;17(10):1086-119.
36. Nishimura RA, Tajik AJ. Evaluation of diastolic filling of left ventricle in health and disease: Doppler echocardiography is the clinician’s Rosetta Stone. J Am Coll Cardiol 1997;30:8-18.
37. Wang A, Cabreriza SE, Quinn TA, Richmond ME, Cheng B, Spotnitz HM. Regional and global strain changes during biventricular pacing in a porcine model of acute left ventricular volume overload. J Ultrasound Med. 2013 Apr;32(4):675-82.
38. Park HE, Chang SA, Kim JH, Oh IY, Choi EK, Oh S. Left ventricular dyssynchrony in pre-excitation syndrome: effect of accessory pathway location and reversibility after ablation therapy. Heart Vessels. 2013 Mar;28(2):199-207.
39. Arita T, Sorescu GP, Schuler BT, Schmarkey LS, Merlino JD, Vinten-Johansen J, et al. Speckle-tracking strain echocardiography for detecting cardiac dyssynchrony in a canine model of dyssynchrony and heart failure. Am J Physiol Heart Circ Physiol. 2007 Jul;293(1):H735-42.
40. Cheng Y, Aboodi MS, Annest LS, Wechsler AS, Kaluza GL, Granada JF, Yi GH. Off-pump epicardial ventricular reconstruction restores left ventricular twist and reverses remodeling in an ovine anteroapical aneurysm model. J Thorac Cardiovasc Surg. 2014;148(1):225-231.
41. Bellenger NG1, Davies LC, Francis JM, Coats AJ, Pennell DJ. Reduction in sample size for studies of remodeling in heart failure by the use of cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2000;2(4):271-8.
42. Alfakih K, Reid S, Jones T, Sivananthan M. Assessment of ventricular function and mass by cardiac magnetic resonance imaging. Eur Radiol. 2004 Oct;14(10):1813-22.
43. Witschey WR, Zsido GA, Koomalsingh K, Kondo N, Minakawa M, Shuto T, et al. In vivo chronic myocardial infarction characterization by spin locked cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2012 Jun 15;14:37.
44. Abegunewardene N, Schmidt KH, Vosseler M, Kreitner KF, Schreiber LM, Lehr HA, et al. Gene therapy with iNOS enhances regional contractility and reduces delayed contrast enhancement in a model of postischemic congestive heart failure. Clin Hemorheol Microcirc. 2011;49(1-4):271-8.
45. Mahnken AH. Computed tomography imaging in myocardial infarction. Expert Rev Cardiovasc Ther. 2011 Feb;9(2):211-21.
46. Mahnken AH, Bruners P, Bornikoel CM, Krämer N, Guenther RW. Assessment of myocardial edema by computed tomography in myocardial infarction. JACC Cardiovasc Imaging. 2009 Oct;2(10):1167-74.
47. Schuleri KH, Centola M, Choi SH, Evers KS, Dawoud F, George RT, et al. CT for evaluation of myocardial cell therapy in heart failure: a comparison with CMR imaging. JACC Cardiovasc Imaging. 2011 Dec;4(12):1284-93.
48. Palazzuoli A, Caputo M, Calabrò A, Nuti R. Clinical impact of BNP and other emerging biomarkers in heart failure evaluation and management. Minerva Cardioangiol. 2012 Apr;60(2):183-94.
49. Cowie MR, Struthers AD, Wood DA, Coats AJ, Thompson SG, Poole-Wilson PA, Sutton GC. Value of natriuretic peptides in assessment of patients with possible new heart failure in primary care. Lancet 1997;350:1349–1353.
50. Chen HH, Burnett Jr JC. Clinical application of the natriuretic peptides in heart failure. EHJ Supplements. 2006, 8: E18–E25.
51. Hirata Y, Matsumoto A, Aoyagi T, Yamaoki K, Komuro I, Suzuki T, et al. Measurement of plasma brain natriuretic peptide level as a guide for cardiac overload. Cardiovasc Res. 2001 Aug 15;51(3):585-91.
52. Paslawska U, Gajek J, Kiczak L, Noszczyk-Nowak A, Skrzypczak P, Bania J, et al. Development of porcine model of chronic tachycardia-induced cardiomyopathy. Int J Cardiol. 2011 Nov 17;153(1):36-41.
53. MacDonald KA, Kittleson MD, Munro C, Kass P. Brain natriuretic peptide concentration in dogs with heart disease and congestive heart failure. J Vet Intern Med. 2003 Mar-Apr;17(2):172-7.
54. George I, Cheng Y, Yi GH, He KL, Li X, Oz MC, Holmes J, Wang J. Effect of passive cardiac containment on ventricular synchrony and cardiac function in awake dogs. Eur J Cardiothorac Surg. 2007 Jan;31(1):55-64.
55. Hamamoto H, Gorman JH 3rd, Ryan LP, Hinmon R, Martens TP, Schuster MD, et al. Allogeneic mesenchymal precursor cell therapy to limit remodeling after myocardial infarction: the effect of cell dosage. Ann Thorac Surg. 2009 Mar;87(3):794-801.
56. Muscari C, Bonafè F, Martin-Suarez S, Valgimigli S, Valente S, Fiumana E, et al. Restored perfusion and reduced inflammation in the infarcted heart after grafting stem cells with a hyaluronan-based scaffold. J Cell Mol Med. 2013 Apr;17(4):518-30.
57. Ko SF, Yip HK, Lee CC, Sheu JJ, Sun CK, Ng SH, et al. Immediate intramyocardial bone marrow-derived mononuclear cells implantation in minipig myocardium after permanent coronary artery ligation: magnetic resonance imaging with histopathologic and immunochemical correlation. Invest Radiol. 2011 Aug;46(8):495-503.
58. Dardenne A, Fernandez C, Wagner A, Milewski K, Ordanes DR, Mount PA, et al. Benefits of standardizing the treatment of arrhythmias in the sheep (Ovis aries) model of chronic heart failure after myocardial infarction. J Am Assoc Lab Anim Sci. 2013;52(3):290-4.
59. Llaneras MR, Nance ML, Streicher JT, Lima JA, Savino JS, Bogen DK, et al. Large animal model of ischemic mitral regurgitation. Ann Thorac Surg 1994;57:432–9
60. Solis J, Levine RA, Johnson B, Guerrero JL, Handschumacher MD, Sullivan S, et al. Polymer injection therapy to reverse remodel the papillary muscles: efficacy in reducing mitral regurgitation in a chronic ischemic model. Circ Cardiovasc Interv. 2010 Oct;3(5):499-505.
61. Ghugre NR, Pop M, Barry J, Connelly KA, Wright GA. Quantitative magnetic resonance imaging can distinguish remodeling mechanisms after acute myocardial infarction based on the severity of ischemic insult. Magn Reson Med. 2013 Oct;70(4):1095-105.
62. Cheng Y, Yi G, Conditt GB, Sheehy A, Kolodgie FD, Tellez A, et al. Catheter-based endomyocardial delivery of mesenchymal precursor cells using 3D echo guidance improves cardiac function in a chronic myocardial injury ovine model. Cell Transplant. 2013; 22(12):2299-309.
63. Moe GW, Stopps TP, Angus C, Forster C, De Bold AJ, Armstrong PW. Alterations in serum sodium in relation to atrial natriuretic factor and other neuroendocrine variables in experimental pacing induced heart failure. J Am Coll Cardiol 1989;13:173–179.
64. Takaseya T, Shiose A, Saraiva RM, Fumoto H, Arakawa Y, Juravic M, et al. Novel epicardial off-pump device for mitral regurgitation: acute evaluation. Eur J Cardiothorac Surg. 2010 Jun;37(6):1291-6.
65. Mahmoudabady M, Mathieu M, Dewachter L, Hadad I, Ray L, Jespers P, et al. Activin-A, transforming growth factor-beta, and myostatin signaling pathway in experimental dilated cardiomyopathy. J Card Fail. 2008 Oct;14(8):703-9.
66. Power JM, Raman J, Byrne MJ, Alferness CA. Efficacy of the Acorn Cardiac Support Device in animals with heart failure secondary to high rate pacing. Heart Fail Rev. 2005 Jun;10(2):117-23.
67. Feindt P, Litmathe J, Boeken U, Gams E. Reverse remodeling by net cardioplasty in a model of dilated cardiomyopathy: results of an animal study. Int J Artif Organs. 2004 Oct;27(10):891-7.
68. Kaye DM1, Byrne M, Alferness C, Power J. Feasibility and short-term efficacy of percutaneous mitral annular reduction for the therapy of heart failure-induced mitral regurgitation. Circulation. 2003 Oct 14;108(15):1795-7.
69. Wang J, Yi GH, Knecht M, Cai BL, Poposkis S, Packer M, Burkhoff D. Physical training alters the pathogenesis of pacing-induced heart failure through endothelium-mediated mechanisms in awake dogs. Circulation. 1997 Oct 21;96(8):2683-92.
70. Yarbrough WM1, Mukherjee R, Stroud RE, Rivers WT, Oelsen JM, Dixon JA, et al. Progressive induction of left ventricular pressure overload in a large animal model elicits myocardial remodeling and a unique matrix signature. J Thorac Cardiovasc Surg. 2012 Jan;143(1):215-23.
71. Hirsch JC1, Borton AR, Albayya FP, Russell MW, Ohye RG, Metzger JM. Comparative analysis of parvalbumin and SERCA2a cardiac myocyte gene transfer in a large animal model of diastolic dysfunction. Am J Physiol Heart Circ Physiol. 2004 Jun;286(6):H2314-21.
72. Zile MR, Brutsaert DL. New concepts in diastolic dysfunction and diastolic heart failure; part I: diagnosis, prognosis, and measurements of diastolic dysfunction. Circulation 2002;105:1387–93.
73. Chien SF, Diana JN, Brum JM, Bove AA. A simple technique for producing supravalvular aortic stenosis in animals. Cardiovasc Res 1988;22:739 –45.
74. Mirsky I, Laks MM. Time course of changes in the mechanical properties of the canine right and left ventricles during hypertrophy caused by pressure overload. Circ Res. 1980 Apr;46(4):530-42.
75. Douglas PS, Tallant B. Hypertrophy, fibrosis and diastolic dysfunction in early canine experimental hypertension. J Am Coll Cardiol. 1991 Feb;17(2):530-6.
76. Kawase Y1, Ly HQ, Prunier F, Lebeche D, Shi Y, Jin H, et al. Reversal of cardiac dysfunction after long-term expression of SERCA2a by gene transfer in a pre-clinical model of heart failure. J Am Coll Cardiol. 2008 Mar 18;51(11):1112-9.
77. Chekanov VS. A stable model of chronic bilateral ventricular insufficiency (dilated cardiomyopathy) induced by arteriovenous anastomosis and doxorubicin administration in sheep. J Thorac Cardiovasc Surg 1999;117:198 –9.
78. Tessier D, Lajos P, Braunberger E, et al. Induction of chronic cardiac insufficiency by arteriovenous fistula and doxorubicin administration. J Card Surg 2003;18:307–11.
79. Kleaveland JP1, Kussmaul WG, Vinciguerra T, Diters R, Carabello BA. Volume overload hypertrophy in a closed-chest model of mitral regurgitation. Am J Physiol. 1988; 254:H1034-41.