Animal model of left atrial thrombus in congestive heart failure in rats

Abstract

The aim of the present study was to develop and study a new model of left atrial thrombus (LAT) in rat with congestive heart failure (CHF). CHF was induced by aortic banding for 2 mo, followed by ischemia-reperfusion (I/R) and subsequent aortic debanding for 1 mo. Cardiac function and the presence of LAT were assessed by echocardiography. Masson’s staining was performed for histological analysis. All CHF rats presented with significantly decreased cardiac function, fibrosis in remote myocardium, and pulmonary edema. The incidence rate of LAT was 18.8% in the rats. LAT was associated with severity of aortic constriction, aortic pressure gradient, aortic blood flow velocity, and pulmonary edema but not myocardial infarction or a degree of left ventricular depression. The progressive process of thrombogenesis was characterized by myocyte hypertrophy, fibrosis, and inflammation in the left atrial wall. Fibrin adhesion and clot formation were observed, whereas most LAT presented as a relatively hard “mass,” likely attributable to significant fibrosis in the middle and outer layers. Some LAT mass showed focal necrosis as well as fibrin bulging. Most LAT occurred at the upper anterior wall of the left atrial appendage. Aortic debanding had no significant impact on large LATs (>5 mm²) that had formed, whereas small LATs (<5 mm²) regressed 1 mo after aortic release. LAT is found in a rat model of aortic banding plus I/R followed by aortic debanding. The model provides a platform to study molecular mechanisms and potential new pathways for LAT treatment.

NEW & NOTEWORTHY It is critically important to have a rodent model to study the molecular mechanism of thrombogenesis in the left atrium. Left atrial thrombus (LAT) is not a simple fibrin clot like those seen in peripheral veins or arteries. Rather, LAT is a cellular mass that likely develops in conjunction with blood clotting. Studying this phenomenon will help us understand congestive heart failure and promote new therapies for LAT.

INTRODUCTION

Thrombosis is the formation of a blood clot inside the circulatory system and can occur in a vein, artery, or left atrium. Venous thrombosis leads to congestion of the affected part of the body, as occurs in the affected lung in pulmonary thrombosis. Deep venous thrombosis is widely acknowledged with three pathophysiological elements, hypercoagulability, hemodynamic stasis, and endothelial injury (13, 21, 40). Alternatively, arterial thrombosis usually occurs as a result of rupture of chronic atherosclerotic plaque and is characterized by inflammation and activation of cell adhesion molecules (25, 30). Left atrial thrombus (LAT) is a complication secondary to severe mitral stenosis (26–33% of cases) (22, 42), atrial fibrillation (AF, 10–30%) (14), and congestive heart failure (CHF) (~30%) (19) with most cardiac valvular disease and AF causally related with CHF (20).

The mechanism of LAT is much more complex than venous and arterial thrombosis. The left atrium is far from a simple passive transport chamber; it is highly dynamic and responds to stretch with the secretion of atrial natriuretic peptides, which can induce natriuresis and decrease cardiac preload and afterload by inhibiting sympathetic neurohormones, inflammation, and the renin-angiotensin-aldosterone pathway (3, 4). It is well known that CHF is associated with a hypercoagulable state, even during sinus rhythm (21, 31), although coagulation is not associated with the development of CHF (41). Anticoagulant medication (warfarin) is widely used to prevent and treat thrombi in both the vessels and left atrium along with the ventricles; however, >50% of LAT resulting from CHF or AF are resistant to warfarin (14, 19). Endothelial damage and dysfunction in CHF appears to be related to activation of the neuroendocrine system with reduced nitric oxide and increased angiotensin II and endothelin levels, which may promote monocyte and platelet adhesion to the endothelium and contribute to increased vasoconstriction (12). Decreased nitric oxide bioavailability and increased generation of reactive oxygen species are among the major molecular changes associated with endothelial dysfunction, even in the absence of alterations of endothelial nitric oxide synthase expression/activity (36, 39). CHF has long been recognized to predispose to inflammation, fibrosis, and microcirculatory dysfunction that accelerate ischemia, hypoxia, proliferation, and angiogenesis in the left atrium (11, 23). However, the exact pathogenesis of LAT in CHF has not been fully elucidated (13), and fundamental questions still remain. For example, why are more than half of thromboses resistant to anticoagulant agents like warfarin or rivaroxaban? Is there any difference between dissoluble LAT and indissoluble LAT?

The in-depth cellular and molecular mechanisms of LAT need more experimental supports because most of the available information is from clinical human cases (15, 20, 26). Likewise, whereas venous and arterial models of thrombus in rodents have facilitated hemorheologic studies of thrombosis, no animal model of endothelial damage in the left atrium has been reported (16, 18). LAT was reportedly induced by high-cholesterol diet in Osborne-Mendel rats (2), by spontaneously running (6,000 revolutions/day × ~80–100 wk) in SPORTS rats (35), and with toxic chemical compounds in mice (43); however, an appropriate LAT model that reflects the complexity of the underlying pathophysiology of CHF is crucial to understand the mechanism of thrombogenesis in the left atrium and the causal relationship between LAT, CHF, and pulmonary failure (27, 32). Most clinical cases have reported that the LAT was a “mass” (24, 38), rather than a blood clot faithful to venous or arterial thrombosis, highlighting the importance of these distinctions.

In the present study, we describe in detail the surgical processes of inducing a rat model of aortic banding (Ab) plus ischemia-reperfusion (I/R) followed by aortic debanding (DeAb), which results in LAT. We provide clear data demonstrating the pathophysiological characters of LAT. This model and study of LAT will provide a new platform to explore molecular mechanisms and drug discovery.

METHODS

All procedures were followed by the recommendations of the Guide for the Care and Use of Laboratory Animals (Department of Health and Human Services publication number NIH 78-23, 1996) and were approved by the Mount Sinai School of Medicine Animal Care and Use Committee.

Animal protocol.

CHF was induced by sequentially performing Ab for 2 mo, followed by acute I/R and DeAb 1 mo later as previously described (Fig. 1A) (7). Surgical points are as follows: 1) Ab was performed in male Sprague-Dawley rats (150–180 g) by constricting the ascending aorta with a 4-0 suture against a 0.98-mm diameter PE-50 tube through right thoracotomy at the second intercostal space. The best diameter for Ab to induce pressure overload is 0.96 ± 0.02 mm, which may be attained by gently pulling tension on the PE-50 tube. The slip knot against aorta should not be loose as the PE tube is removed. Aortic flow velocity (≥4.0 m/s) is the optimal method to determine whether constriction is sufficient, whereas diameter as measured by two-dimensional echocardiography model is often inaccurate (Fig. 1B). 2) For I/R 2 mo following Ab, rats underwent left coronary artery ligation for 30 min followed by reperfusion to induce I/R injury. Ligation of left coronary artery after Ab was more difficult than in naïve animals because the artery was mostly invisible in the hypertrophic heart. Duration of ischemia was 30 min, as longer than 45 min was found to increase mortality. Following Ab, the left ventricle had stronger resistance to ischemia than naïve rats, leading to less severe submyocardial infarction (Fig. 1C). 3) For DeAb 1-mo post-I/R, the rats underwent a third thoracotomy at the upper-right side of the sternum. The roots of two to four ribs were cut along the right side of the sternum to expose the aorta. The knot end of the banding suture (>0.5 cm) covered by the thymus fat should be cut gradually with microdissecting spring scissors (BRI, 11-1395) and separated with microdissecting spring forceps (Roboz, RS-5135), as the aortic wall is very thin and easily broken, leading to severe bleeding (Fig. 1, E and F). Blood clotting spray (First Aid Only M529) was useful in preventing and reducing bleeding. Five ~8-ml aliquots of 0.9% normal saline were injected via tail vein after skin closure. For all surgical procedures, anesthesia was induced by intraperitoneal administration of ketamine (65 mg/kg) + xylazine (13 mg/kg) + acepromazine (1 mg/kg). Animals underwent intratracheal intubation and mechanical ventilation.

Fig. 1.

Animal model of left atrial thrombus (LAT) induced by aortic banding (Ab) plus ischemia-reperfusion (I/R) following by aortic debanding (DeAb) in rats. A and B: illustration of access of 3 surgeries, 1 = Ab, 2 = I/R, 3 = DeAb. Sa, sacrifice animals. C: aortic constriction. D: I/R. a = 2-0 suture, b = 6-0 suture. A surgeon’s knot is more secure than noose knot. E: process of debanding. a = aorta, s = suture. F. debanded aorta. G: LAT.

Time points.

Animal age at Ab is between 1 and 1.5 mo, with body weight ~150–180 g. It is suggested that body weights exceeding 250 g should not be used. I/R should be conducted within 2 mo after Ab, as an interval greater than this will increase the mortality of I/R surgery. Most thrombosis occurred ~2–4 wk after I/R (Fig. 1G). Periods longer than 1 mo after I/R injury could increase LAT incidence slightly in animals with aortic flow velocity ≥3.7 m/s. The time points are an empirical observation. We do not have sufficient data to fully quantify this claim.

Echocardiography.

The animals were sedated by intraperitoneal injection of ketamine (40 mg/kg). Echocardiograms (GE healthcare VIVID 7) equipped with 14-MHz probes were performed as previously described (6, 8). Left atrial area was measured at parasternal short axis view. Left ventricular fractional shortening (FS) data were obtained by using M-mode following the vendor’s instructions. Aortic valve velocity was obtained with a GE P2d probe (TE100024, 2 MHz) at continuous wave Doppler.

Visualization of blood flow on left atrium in vivo with Evans blue injection.

Evans blue in PBS was injected into the left ventricle to observe the pattern of blood flow in the left atrium in 6 rats (2 ml of 0.2%) and to observe blood flow in the left ventricle in 10 mice (100 µl of 0.2%) recorded by in vivo video.

Histology.

Heart, lung, and atrium were harvested after perfusion with 60 ml of cold PBS and 0.2 ml of 1% heparin. The samples were photographed, frozen, and cryostated at 8 μm. Masson’s trichrome staining of collagen fibers and myocytes was performed as previously described (10).

Statistics.

Variables are expressed as means ± SE. Student’s t-test was performed to compare experimental groups using GraphPad Prism software. P values <0.05 were considered statistically significant.

RESULTS

Cardiac and pulmonary pathophysiological changes.

I/R injury is essential to formation of LAT in rats after pressure overloading. However, our data demonstrate that LAT did not depend on severity in left ventricular injury (myocardial infarct size) and dysfunction but rather associated with left atrial expansion (blood stasis) and pulmonary edema (Table 1). Ventricular and pulmonary fibrosis and inflammation were severe in all LAT rats (Fig. 2). The pulmonary edema was scored according to lung weight/body weight ratio, lung size, color, and severity of alveolar exudation on slides stained with Masson’s trichrome. Right ventricles also showed significant hypertrophy and fibrosis (Fig. 2).

Table 1.

LAT with left ventricular myocardial infarction and pulmonary edema

		LAT Negative		LAT Positive
	(1) Control	(2) No MI	(3) Plus MI	(4) Small LAT	(5) Large LAT
Animals, n	14	26	11	8	9
LA area, mm2	25 ± 1.4	35 ± 3	38 ± 4a	53 ± 7b,d	71 ± 3b,d,e
LAT area, mm2	0	0	0	4.1 ± 1.18b,c	28 ± 3b,d,f
MII, %	0	0	100b	63b	22
MI size, %	0	0	15 ± 3b	9 ± 2b	2 ± 1
L/BW, g/kg	3.14 ± 0.26	3.34 ± 0.46	3.59 ± 0.52	3.87 ± 0.65b,d,f	4.15 ± 0.73b,d,e,g
PE score, 0–4	0	0	1 ± 0	3.0 ± 0.33d,g	3.7 ± 0.24d,g
FS, %	69 ± 3	47 ± 2b	45 ± 3b	45 ± 4b	46 ± 3b

Values are means ± SE. LA, left atrium; LAT, left atrial thrombus; small LAT, LAT <5 mm²; large LAT, LAT >5 mm²; MI, myocardial infarction; MII, MI incidence; L/BW, lung/body weight; PE, pulmonary edema; PE score, 0 = no PE, 4 = very serious PE, based on lung color and Masson’s trichrome staining on slides; FS, fractional shortening.

^aP < 0.05 compared with control, column 1;

^bP < 0.01 compared with control, column 1;

^cP < 0.05 compared with column 2;

^dP < 0.01 compared with column 2;

^eP < 0.05 compared with column 4;

^fP < 0.01 compared with column 4;

^gP < 0.01 compared with column 3.

Fig. 2.

Fibrosis and inflammation in lung and left and right ventricle-associated left atrial thrombogenesis in congestive heart failure. A: control lung, lung/body weight (g/kg) = 3.15. B: pulmonary edema. Lung/body weight = 4.23. C: Masson’s trichrome staining showed that small vessels, bronchial and alveolar, were clear in controls. D: lung in left atrial thrombus (LAT)-positive rat bronchial and alveolar edema. E: left ventricle in control. F: perivascular fibrosis in LAT-positive left ventricle. G: inflammation in the left ventricle. H: perivascular fibrosis and neointima formation in the right ventricle.

Hemodynamic changes in left atrium.

Figure 3 shows hemodynamic characteristics of CHF with LAT. The severity of aortic constriction had a dominant impact on left ventricular afterload. The difference in constricted aortic diameter between LAT-negative (Ab + I/R) and LAT-positive (Ab + I/R + LAT) groups was very small (0.14 mm). Aortic blood flow velocity and pressure gradient were significantly increased after Ab plus I/R injury (Ab + IR) as blood flow resistance was elevated, whereas mitral valve blood flow velocity decreased as cardiac contractility weakened. The LAT-positive hearts had more pressure loading, slower left ventricular filling, and significant left atrial enlargement. Although debanding could reduce aortic flow velocity, it did not result in the recovery of mitral valve flow velocity (left ventricle contractility). The left atrium expanded as left ventricular function decreased. However, LAT was apparently associated with severity of aortic constriction, rather than correlating directly with cardiac dysfunction. After DeAb, left atrial area decreased, whereas the LAT remained (Fig. 4). AF was not observed in the LAT rats. However, there were cases of ectopy and arrhythmias, including premature ventricular contractions and paroxysmal supraventricular tachycardia. The correlation between arrhythmia and LAT remained unclear, however, because of the small number of animals. The incidence rate of LAT was 18.8% in 90 animals over the 4-mo period (2-mo Ab, 1-mo I/R, 1-mo DeAb).

Fig. 3.

A–D: echocardiographic characteristics of congestive heart failure with left atrial thrombus (LAT). Means ± SE, *P < 0.05, **P < 0.01 compared with control. ##P < 0.01 compared with aortic banding (Ab) + ischemia-reperfusion (IR). Ab + IR, aortic banding 2 mo plus ischemia 30 min and reperfusion 1 mo; AIRD, Ab + IR followed with debanding 1 mo; +LAT: with left atrial thrombus.

Fig. 4.

Left atrial expansion in congestive heart failure (CHF) with left atrial thrombus (LAT). A: control left atrium (LA). B: dilated LA in CHF. C: LAT in CHF. D: left atrial area in different groups. Means ± SE, **P < 0.01 compared with control. Ab + IR, aortic banding for 2 mo plus ischemia for 30 min and reperfusion for 1 mo; AIRD, Ab + IR followed with debanding 1 mo; +LAT, with LAT; LV, left ventricle.

General morphology and histology of LAT.

Echocardiography with a 14-MHz probe (GE VIVID 7) had a good sensitivity of LAT ≥1 mm². LATs are heterogeneous, in that their size, number, and fabric vary even within the same time point. Most LATs identified were solid “mass,” with only 2 of them being blood clots (Fig. 5D). Small LAT could be irregular on the surface (Fig. 5E). Large LAT are mostly spherical and have a membrane (Fig. 5F). Some animals showed multiple spherical large LATs (Fig. 5G). LAT showed fatty content after breaking the membrane. The size and number of LAT had no direct correlation with the severity of left ventricular dysfunction (Fig. 5).

Fig. 5.

General morphology of left atrial thrombus (LAT) in rats with congestive heart failure. A: control left atrium left (LA) anterior view. B: interior of control LA. C: in situ view of LAT and aortic debanding (DeAb) suture. D: blood clot in LA. E: small LAT. F: big LAT covered with a clear membrane. G: multiple LAT. H: big LAT shows fatty content after breaking the membrane. It is difficult to distinguish the LAT from atrial myxoma in gross view.

The formation and development of LAT are not uniform in different stages and rats. Early fibrin adhesion and blood clot formation could be found in left atrium despite no obvious evidence of endothelial damage as in arterial thrombus induced by a filament injury (Fig. 6, A–F). Almost all massive LATs have fibrosis at the base with inflammation on the left atrial wall (Fig. 6, G–I), but the causal relationship of a fibrin clot and a fibrotic cellular “mass” is not clear. Moreover, it is possible that the fibrin clot is a precursor of a cellular massive LAT but is not resolved (Fig. 6, H–J). It was difficult to observe the progress of micro-LAT (<1 mm²) in the very early stage using the present techniques of echocardiography and MRI in vivo in rodents. Additionally, not all cross sections clearly showed the connection between micro-LAT (<1 mm²) and pathological change on the atrial wall (Fig. 7, A and B).

Fig. 6.

General histology of thrombogenesis from artery to left atrium (LA). A: control common carotid artery. B: thrombus in common carotid artery 3 days after endothelial injury with a filament. C: control LA. D: fibrosis in LA wall in congestive heart failure. E: fibrin adheres to the LA wall. F: resolvable emboli in the LA; there are not nuclear cells in the blood clot. G: inflammation and damaged endothelium of LA. H: fibrotic protrusion on the LA wall. I: microemboli with myocyte hypertrophy and fibrosis in LA. J: small fibrinogen thrombus with a fibrous base.

Fig. 7.

Prone site of thrombogenesis in the left atrium (LA). A: microemboli disconnected with atrial wall and not measurable with echocardiography. B: small left atrial thrombus (LAT), potentially measurable by echocardiography. C: LAT in the atrial pectinate muscle compartments. The thrombi “mass” was small. D: mass was rooted in the outside wall of the LA. E: thrombi “mass” had clear and integral membrane and a base. F and G: large thrombi “mass” did not have a complete membrane. The fibrin-like irregular edge of the “mass” faced the LA chamber, indicated its potential trend of growth. Dotted red line = missed LA wall. Arrow indicates LAT.

Predisposed location of thrombogenesis in left atrium.

Most LAT occurred on anterior superior outer wall of the left atrial appendage (Fig. 1G). Not all LAT were measurable by echocardiography or visible under surgical microscope without staining on the sectional slide (Fig. 7, A and B). Thrombus in the pectinate muscular compartment of left atrial appendage was usually small (<2 mm²) (Fig. 7C), while thrombus facing the left atrium chamber could be bigger (Fig. 7, D, F–H). Some LAT had an integral membrane with apparent focal necrosis and calcification (Fig. 7F), whereas others had incomplete or disrupted membranes with a protruding part that looked similar to a fibrin clot, indicating continuous growth of this LAT type (Fig. 7, G and H). It is unclear what affects or determines this trend of LAT-predisposed location and the uneven characteristics although we hypothesize that it is mainly attributable to the unique pattern of blood circulation through the left atrial wall. To answer this question, we conducted a primary experiment of in vivo imaging of microcirculation on the left atrium. The major artery (Ø >50 um) in the left atrial appendage was relatively less patent compared with the left coronary artery in the left ventricle. The pattern of blood perfusion is a ring type, from the periphery to the center left atrial appendage, instead of tree type as in left ventricle. Blood perfusion velocity in the left atrium (30 ± 4.5 s, n = 6) is overwhelmingly slower than in the left ventricle (4.3 ± 1.2 s, n = 10).

Effect of DeAb on LAT.

After debanding, aortic constriction (aorta diameter at banded side/aorta diameter post banding side) was released from 30% to 46%, and pressure overload was accordingly decreased (Table 2). Left atrial area also was decreased ~10 mm², but the improvement in cardiac pumping function was still not significant. 67% (8/12) of the larger thromboses (11 ± 3.6 mm²) remained after DeAb, whereas 33% of small LAT (4.1 ± 1.2 mm²) decreased in size after DeAb, as assessed with echocardiography, and was not measurable by direct imaging with a standard digital camera (Fig. 8). Differences in baseline metrics between these two groups were not significant.

Table 2.

Effects of de-aortic banding (DeAb) on LAT in rats

	n	Model	LAT, mm2	LA Area, mm2	Ad/AD, %	AFV, m/s	APG, mmHg	MVFV, m/s
Control	10	Control	0 ± 0	26 ± 1.44	100 ± 0	1.21 ± 0.03	5.69 ± 0.30	1.01 ± 0.04
AIRD + LAT	8	(1) Before DeAb	11 ± 3.65d	54 ± 6.13d	31 ± 1.54d	4.18 ± 0.26d	72 ± 8.89d	0.75 ± 0.05d
		(2) After DeAb	10 ± 3.79a	44 ± 7.19a	46 ± 1.86c	2.13 ± 0.25c	21 ± 4.42c	0.59 ± 0.05a
		(3) Direct imaging	11 ± 4.34	46 ± 6.62
AIRD +/− LAT	4	(4) Before DeAb	4.10 ± 1.2a	44 ± 4.21a	34 ± 3.52	4.25 ± 0.09	72 ± 2.94	0.68 ± 0.02
		(5) After DeAb	1.70 ± 1.18b	36 ± 7.41b	47 ± 0.17	2.21 ± 0.32	21 ± 6.43	0.64 ± 0.07
		(6) Direct imaging	0	32 ± 8.07

Values are means ± SE. Lines 1, 2, 4, and 5 data are from Echo. Lines 3 and 6 are data from direct imaging on tissues or slides. AIRD + LAT, aortic banding + ischemia-reperfusion + aortic debanding with left atrial thrombus (LAT) (LAT was consistently positive via Echo and direct imaging before and after debanding); AIRD +/− LAT, LAT was negative by direct imaging after debanding; LA, left atrium; Ad, aorta diameter at banded side; AD, aorta diameter post banding side; Ad/AD, aortic constricted ratio; AFV, aortic flow velocity; APG, aortic pressure gradient; MVFV, mitral valve flow velocity.

^aP > 0.05 compared with line 1;

^bP > 0.05 compared with line 4;

^cP < 0.01 compared with line 1,

^dP < 0.01 compared with control.

Fig. 8.

Effect of aortic debanding on left atrial thrombus (LAT). A: left atrium (LA) 2 mo after aortic banding. B: thrombogenesis in LA 1 mo after aortic banding plus ischemia 30 min and reperfusion. C: LAT regressed 1 mo after aortic debanding. D: thrombus could not be found in opened LA. E: thrombus also could not be found in cross-sectional staining. A–E: data are from same rat. MVF, mitral valve flow; LV, left ventricle. White arrow indicates LAT.

DISCUSSION

In the present study, we explored how to establish a model of LAT in rat by Ab plus I/R followed with DeAb. We describe the operative procedure in detail and characterize pathophysiological changes in the left atrium, as well as in the left ventricle and lung. We found that thrombogenesis in the left atrial appendage resulting from CHF in rats was not a simple formation of a fibrinous blood clot, as in the vein or artery. LAT apparently correlates with blood stasis and I/R injury, but the molecular mechanism of LAT in CHF remained unclear.

The left atrium represents a midway of blood circulation between lung and left ventricle and is not a key player in the propulsion of blood flow (1). LAT mainly resulted from severe mitral stenosis, AF, and CHF (19); however, the three are not totally causally correlated with each other, at least based on our present results. Paroxysmal ventricular premature beats were observed in three CHF rats (3.3%), but only one of the left ventricular arrhythmia rats had an LAT. Aortic stenosis played a key role in the development of LAT but is not the only determining factor. LAT was found in four rats (4.4%) 2 mo after Ab. In contrast, the incidence of LAT was 18.8% (by echocardiography) in Ab plus I/R injury rats. The rate might be higher if the diagnosis for LAT was based on microscopic imaging on section slides. The role and mechanism of I/R injury in the thrombogenesis was not clear. It is likely that I/R was just a promoter, as LAT was not observed in any rat with I/R alone. However, LAT did not correlate closely with the size of myocardial infarction or severity of left ventricular dysfunction (Table 1). The subtleties in relationship between thrombogenesis, I/R injury, and cardiac dysfunction are attractive for future study.

The role of hypercoagulation in the LAT was conjectured even though a fibrin net and blood clot could be observed both in macro view (Fig. 5D) and under microscope (Fig. 6, E and F). It is of the utmost importance to reeliminate the possibility of insufficient heparinization in those animals. Most LAT in our results were solid “mass,” as reported in most clinic patients. The mass is heterogeneous and has a fibrotic and inflammatory base on the atrial wall, a membrane, and a core with calcification, focal necrosis, and hyperplasia (Fig. 7). A combination of fibrin clot and myocardial fibrosis (Fig. 6J) could be observed in one animal, but it is not clear whether those two phenomena are causally correlated. Fibrin adhesion is probably very necessary for the growing of the “massive” thrombosis, evidenced by Fig. 7, G and H. The mechanism of transition of the thrombus from fibrin clot to solid cellular mass is not clear. There may be two phases of one event, or two different events, involving platelet, leukocyte activation, fibrosis, inflammation, angiogenesis, and microcirculation disorder in the left atrium induced by CHF (11, 28).

The left atrial appendage has a very unique structure (1, 33); it is unknown why most LATs in our study were formed on the exterior and anterior upper outer wall of the left atrial appendage, which is not consistent with clinical reports of LAT and also distinct from atrial myxoma, which could occur anywhere but mostly on atrial septum (29). Inflammation and fibrosis observed in the left ventricle and lung (Fig. 2) may play important roles in the genesis of myofibroblastic tumors or pseudotumors (17, 34, 37). Our previous studies showed that microcirculatory disorders occurred in chronic heart failure (11, 28). The unique characteristic of microcirculation on the left atrial wall probably provides a basis for the trend of predisposed location and hyperplasia of LAT in stimulation of fibrosis and inflammation. It is a crucial challenge to distinguish whether the “mass” in the left atrium is a fibrin clot or cellular pseudotumor and, more importantly, to determine whether it is a process of passive adhesion or active growth. More studies on this issue are necessary to deepen our knowledge of CHF beyond its characterization as a hemodynamic disorder.

Thrombus induced by endothelial denudation in the common carotid artery (Fig. 6B) mostly dissolved 2 wk after filament injury, and neointima also tend to be regressed or absent 4 wk later if a Western-type diet (9) was not provided. We noticed that small LAT (4.1 ± 1.2 mm², by echocardiography) has the potential to degenerate after DeAb. Meanwhile, large thrombi (11 ± 4 mm²) were not affected by DeAb (Table 2). The reason for this phenomenon is not clear although we posit that contributing factors include 1) artificial error in echocardiography or cryostat sectioning and 2) autodissolution of the thrombus. In the case of the latter, it is worthy to investigate the mechanism of LAT regression (even in early stage) to develop new medicine for treatment of the thrombi.

In the present study, only male rats were used, mainly because of our previous experience that male rodents are more likely than female ones in tolerating three open chest surgeries (5, 7). Sex and species variation can occur. The molecular pathway for LAT seems to involve fibrosis, inflammation (37), microcirculatory disorder (24), and pseudotumors (17, 34). The present study provides a model to further our understanding of the association between thrombogenesis and chronic cardiac dysfunction, which may lead to more directed therapies.

GRANTS

This work was supported in part by grants from the National Institutes of Health Grants R01-HL-128099, R01-HL-129814, and R01-HL-131404 (to R. J. Hajjar).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.C. conceived and designed research; J.C. and L.L. performed experiments; J.C. and B.S. analyzed data; J.C. interpreted results of experiments; J.C. prepared figures; J.C. and B.S. drafted manuscript; B.S. edited and revised manuscript; R.J.H. approved final version of manuscript.