Cardio-omentopexy Reduces Cardiac Fibrosis and Heart Failure After Experimental Pressure Overload

Abstract

Background.

The pedicled greater omentum has been shown to offer benefit in ischemic heart disease for both animal models and human patients. The impact of cardio-omentopexy in a pressure overload model of left ventricular hypertrophy (LVH) is unknown.

Methods.

LVH was created in rats by banding the ascending aorta after right thoracotomy (n = 23). Sham surgery was performed in 12 additional rats. Six weeks after banding, surviving LVH rats were assigned to cardio-omentopexy by left thoracotomy (LVH + Om, n =8) or sham left thoracotomy (LVH, n = 8). Sham rats also underwent left thoracotomy for cardio-omentopexy (Sham + Om, n = 6); the remaining rats underwent sham left thoracotomy (Sham, n = 6).

Results.

Echocardiography 10 weeks after cardio-omentopexy revealed LV end-systolic diameter, cardiomyocyte diamter, and myocardial fibrosis in the LVH group were significantly increased compared with the LVH + Om, Sham + Om, and Sham groups (p < 0.01). LV ejection fraction of the LVH group was lower than theLVH + Om group (p < 0.01). Gene expression analysis revealed significantly lower levels of sarcoendoplasmic reticulum calcium adenosine triphosphatase 2b in LVH rats than in the LVH + Om, Sham + Om, and Sham groups (p < 0.01). In contrast, collagen type 1 α 1 chain, lysyl oxidase-like protein 1, nuclear protein-1, and transforming growth factor- β1 in the LVH group were significantly higher than in the LVH + Om cohort (p < 0.01), consistent with a reduced fibrotic phenotype after omentopexy. Lectin staining showed myocardial capillary density of the LVH group was significantly lower than all other groups (p < 0.01).

Conclusions.

Cardio-omentopexy reduced cardiac dilation, contractile dysfunction, cardiomyocyte hypertrophy, and myocardial fibrosis, while maintaining other molecular indicators of contractile function in this LVH model.

Cardio-omentopexy is a term first coined in the 1930s by Laurence O’Shaughnessy to describe a procedure he developed to treat angina successfully in patients with atherosclerotic coronary artery disease [1]. In supportive animal research in greyhounds, O’Shaughnessy showed that, after surgically approximating the pedicled greater omentum to the ventricular epicardium in animals with ligated left coronary arteries, vascular anastomoses developed between the omental and coronary arterial systems [1]. Interestingly, these animals sustained their racing conditioning [1]. O’Shaughnessy’s cardio-omentopexy concept was lost until it was taken up by Arthur M. Vineberg in the 1960s [2]. Vineberg also recognized and demonstrated the efficacy of cardio-omentopexy for the revascularization of ischemic myocardium. Despite this efficacy, cardio-omentopexy and the similar Vineberg procedure [3] were both rapidly superceded by the development of coronary artery bybass grafting for surgical revascularization of the ischemic heart.

These historical efforts to use cardio-omentopexy emanated from the longstanding recognition of the ability of the omentum to promote healing and regeneration in traumatic and surgical wounds. The greater omentum is a sizeable apron-like fold in the visceral peritoneum that hangs down from the stomach with a rich vascular supply [4]. It has long been noted that the human omentum can promote local angiogenic activity. The omentum also contains lymphoid aggregates, called milky spots, that are rich in macrophages and support both the innate and adaptive immune response to peritoneal antigens [5]. More recent investigation has also found the omentum to be a source of pluripotent stem and numerous growth factors [6].

As discussed above, cardio-omentopexy has been used for myocardial revascularization in ischemic heart disease [7]. Thought largely to be due to its angiogenic capability, the pedicled omental flap has been shown to be effective in coronary ligation animal models and in human patients with chronic angina [8–10]. Given the aforementioned emerging understanding of the potential cellular and molecular healing properties of the omentum, we hypothesized that cardio-omentopexy could mitigate the fibrotic and hypertrophic maladaptive response to chronic left ventricular pressure overload in a rat aortic banding model.

Material and Methods

Animal Surgical Procedures

LVH MODEL WITH AORTIC BANDING.

Male Sprague-Dawley rats (250 to 300 grams) were used for the left vern-tricular hypertrophy (LVH) model. The Institutional Animal Care and Use Committee of the University of Texas Southwestern Medical Center approved this study.

The rats were intubated with a 15-gauge flexible catheter and ventilated with 2% to 3% isoflurane. The right second intercostal space was opened and a 4–0 silk was placed around the ascending aorta. A nonocclusive surgical knot was tied around both the ascending aorta and a 16-gauge blunt needle placed adjacent and parallel to the vessel (Fig 1A). The lungs were fully reinflated and the chest cavity was closed. Sham surgery was done by opening and closing the chest cavity without banding the ascending aorta.

Fig 1.

Surgical procedures of (A) ascending aortic banding and (B–D) cardio-omentopexy.

CARDIO-OMENTOPEXY.

Anesthesia was administered in the same way as for the aortic banding procedure. A left thoracotomy was made through the fourth intercostal space. The left lung was gently pushed away from the pericardium with a small sponge. The pericardium was opened. A 3-mm opening was then made through the left diaphragm parallel to the direction of its fibers. Through this incision, the omentum was pulled up into the chest, and the LV surface was covered with the omentum. To avoid later herniation of abdominal organs into the pleural cavity, mattress sutures were made from side-to-side in the diaphragmatic incision through the substance of the omentum. The omentum was sutured to the myocardium with 7–0 Prolene (Ethicon, Somerville, NJ) sutures, avoiding the branches of the coronary vessels (Figs 1B–1D).

Experimental Design

Six weeks after aortic banding and sham surgeries, rats from the banding cohort were randomly assigned to sham left thoracotomy (LVH group, n = 8) or cardio-omentopexy (LVH+Om group, n = 8). Rats from the sham cohort were divided randomly to sham left thoracotomy (Sham group, n = 6) and cardio-omentopexy (Sham+Om group, n = 6).

Ten weeks after cardio-omentopexy, the rats were euthanized after echocardiographic evaluations were completed. The rat hearts were removed and immediately processed for histology and real-time quantitative polymerase chain reaction (qPCR) analysis.

Assessment of Cardiac Function

Cardiac function was evaluated by echocardiography, beginning 6 weeks after aortic banding (baseline measurement) and then at 10 weeks after cardio-omentopexy (end point measurement). Transthoracic echocardiography (Acuson; Siemens, Malyern, PA) was performed with standard views by using an S5 probe under general anesthesia induced and maintained by mask inhalation of isoflurane (2%, 0.5 mL/min). M-mode images were obtained from the left parasternal short-axis views of the left ventricle at the level of papillary muscles to define wall thickness and internal diameters during systole and diastole. The following M-mode measurements were done: LV end-diastolic dimension, LV end-systolic dimension, interventricular septum thickness at diastole, and LV posterior wall thickness at diastole. From these measurements, LV ejection fraction (LVEF) and fractional shortening were derived.

Histology

All rat hearts were transected at the papillary muscle level. Half of the ventricular heart was fixed in 10% formalin solution. The other ventricular half was used for real-time PCR. Standard hematoxylin and eosin, trichrome, and lectin staining were performed to analyze cell size, myocardial fibrosis, and capillary density by using NIS Elements Viewer, version 4.20 (Nikon Instruments Inc, Melville, NY).

qPCR

LV tissue was dissected and snap-frozen in liquid nitrogen, then stored at –80°C. Total mRNA was isolated with TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Real-time PCR conditions were 40 cycles of denaturing at 94° C for 20 seconds and primer annealing/extension at 62°C for 60 seconds. The PCR amplicons of atrial natriuretic peptide (ANP), sar-coendoplasmic reticulum calcium adenosine triphosphatase 2a (SERCA2a) and SERCA2b, collagen type 1 α 1 chain (COL1a1), lysyl oxidase-like protein 1 (LOXL1), nuclear protein-1 (NUPR1), and transforming growth factor-β1 (TGF-β1) were analyzed. The TaqMan rodent glyceraldehyde 3-phosphate dehydrogenase (GAPDH) control reagent was used to detect rat GAPDH as the internal standard. Expression levels of the target gene were normalized to the GAPDH level in each sample. The following are the real-time PCR primers used: rCol1A1-L: 5′-cat gtt cag ctt tgt gga cct-3′, rCol1A1-R: 5′-gca gct gac ttc agg gat gt-3′, rLoxL1-L: 5′-ctg tat tcc ttg cgt tgt gc-3′, rLoxL1-R: 5′-gtg gcct cag gag cat agg-3′, rNupr1-L: 5′-tgg gat cct gga tga gta tga-3′, rNupr1-R: 5′-ggt acg tcc ttt ccg acc tc-3′, rTGFb1-L: 5′-cct gga aag ggc tca aca c-3′, and rTGFb1-R: 5′-cag ttc tct gtg gag ctg a-3′.

Statistical Analysis

All results were expressed as mean ± SE of the mean. Comparisons of groups were made by one-way analysis of variance with a Tukey-Kramer test. Statistical analysis was performed with GraphPad Prism 7 (GraphPad Software, La Jolla, CA). A value of p less than 0.05 was considered statistically significant.

Results

Pressure overload and cardiac hypertrophy were induced by ascending aortic constriction. Seven rats (30%) died within 48 hours of aortic banding. There was no subsequent death during the experimental protocol.

Baseline Echocardiographic Data

Six weeks after aortic banding, as shown in Table 1, M-mode echocardiography revealed an increase of interventricular septum thickness at end diastole/interventricular septum thickness at diastole (p < 0.01) compared with sham rats. The LV posterior wall thickness at end diastole/LV posterior wall thickness at diastole was also elevated (p < 0.05) relative to sham rats. These data were consistent with successful banding-induced hypertrophy. No other statistically significant differences were found between the experimental groups at baseline.

Table 1.

Echocardiography Before Omentopexy

Variable	Sham	Sham+Om	LVH	LVH+Om	p Value
LVEDD, mm	7.62 ± 0.27	7.57 ± 0.32	7.21 ± 0.27	7.16 ± 0.17	0.49
LVESD, mm	4.28 ± 0.35	4.15 ± 0.33	4.25 ± 0.32	4.18 ± 0.19	0.74
IVSd, mm	1.95 ± 0.07	1.82 ± 0.08	2.5 ± 0.08	2.43 ± 0.08	<0.01
LVPWd, mm	2.43 ± 0.19	2.18 ± 0.13	2.88 ± 0.16	3.05 ± 0.21	<0.05
LVEF, %	74.17 ± 3.79	77.78 ± 3.41	76.36 ± 2.91	79.49 ± 1.58	0.62
FS, %	39.13 ± 2.28	40.6 ± 2.91	40.84 ± 2.49	43.55 ± 1.51	0.58

Values are presented as mean ± SE of the mean.

FS = fractional shortening; IVSd = interventricular septum diastolic dimension; LVH = left ventricular hypertrophy; LVPWd = left ventricular posterior wall diastolic dimension; LVEDD = left ventricle end-diastole diameter; LVEF = left ventricular ejection fraction; LVESD = left ventricle end-systole diameter; Om = cardio-omentopexy by left thoracotomy.

Heart Weight, Body Weight, and Echocardiographic Data After Cardio-omentopexy

As shown in Table 2, the LVH+Om group demonstrated a significantly lower heart weight/body weight ratio compared with the LVH group (p < 0.05). Serial changes in LV function after omentopexy were assessed by echocardiography. At 10 weeks after omentopexy, the LV end-systolic dimension of the LVH group was significantly larger than the LVH+Om, Sham+Om, and Sham groups (p < 0.01); LVEF and fractional shortening of the LVH group were also lower than the other three groups (p < 0.01). Taken together, these data indicated that cardio-omentopexy of LVH hearts preserved ventricular systolic function.

Table 2.

Heart Weight/Body Weight Rate and Echocardiography

Variable	Sham	Sham+Om	LVH	LVH+Om	p Value
HW/BW, mg/g	2.64 ± 0.09	2.82 ± 0.05	3.61 ± 0.21a	3.11 ± 0.05	<0.05
LVEDD, mm	7.72 ± 0.08	7.93 ± 0.29	7.68 ± 0.18	7.61 ± 0.19	0.72
LVESD, mm	4.33 ± 0.21	4.38 ± 0.16	5.35 ± 0.21a	4.04 ± 0.16	<0.05
IVSd, mm	1.98 ± 0.07	2.03 ± 0.08	2.3 ± 0.14	2.23 ± 0.08	0.15
LVPWd, mm	2.33 ± 0.08	2.4 ± 0.11	2.65 ± 0.17	2.77 ± 0.22	0.38
LVEF, %	78.85 ± 1.88	79.33 ± 1.72	62.93 ± 2.86b	83.79 ± 1.65	<0.01
FS, %	42.82 ± 1.81	43.27 ± 1.64	30.36 ± 1.91b	45.6 ± 2.14	<0.01

^ap < 0.05;

^bp < 0.01.

Values are presented as mean ± SE of the mean.

BW = body weight; FS = fractional shortening; HW = heart weight; IVSd = interventricular septum diastolic dimension; LVEDD = left ventricle end-diastole diameter; LVEF = left ventricular ejection fraction; LVESD = left ventricle end-systole diameter; LVPWd = left ventricular posterior wall diastolic dimension; LVH = left ventricular hypertrophy; Om = cardio-omentopexy by left thoracotomy.

Histology

Cardiomyocyte diameter of the LVH group was significantly higher than the LVH+Om, Sham, and Sham+Om groups (p < 0.05) (Fig 2). Cardiac fibrosis in the LVH group was greater than the LVH+Om, Sham, and Sham+Om groups (p < 0.05). Lectin staining revealed that the myocardial capillary density of the LVH group was significantly lower than the other three groups (p < 0.05) (Fig 3), consistent with decreased blood supply in the LVH hearts.

Fig 2.

(A–P) Representative hematoxylin and eosin– and trichrome-stained left ventricular cross sections at papillary muscle level show marked myocyte hypertrophy and myocardial fibrosis in the LVH heart.(Scale bar 4 mm [A–D and I–L], 100 μm [E–H and M–P].) (LVH = left ventricular hypertrophy; Om = cardio-omentopexy by left thoracotomy.)

Fig 3.

(A–D) Lectin staining of the left ventricular cross sections at papillary muscle level showed significant lower capillary density in the LVH group compared with LVH+Om, Sham, and Sham+Om groups (p < 0.5). (Scale bar 100 um.) (LVH = left ventricular hypertrophy; Om = cardioomentopexy by left thoracotomy.)

Gene Expression

To examine changes at the molecular level, we next performed qPCR of omentopexy-treated heart tissue. The LVH and LVH+Om groups showed higher ANP mRNA expression than the Sham and Sham+Om groups (p < 0.01), but no difference was found between the LVH and LVH+Om groups. SERCA2a expression, an indication of proper cardiomyocyte calcium handling, was not substantially altered in the four groups. The LVH group, however, had substantially lower SERCA2b expression relative to the other three groups (Fig 4). COL1a1, LOXL1, NUPR1, and TGF-β1showed substantially higher expression in the LVH group than in the other three groups (Fig 5), consistent with the reduced fibrotic appearance of the LVH+Om cohort.

Fig 4.

mRNA expression of (A) atrial natriuretic peptide (ANP), (B) sarcoendoplasmic reticulum calcium adenosine triphosphatase 2a (SERCA2a), and (C) sarcoendoplasmic reticulum calcium adenosine triphosphatase 2b (SERCA2b). (LVH = left ventricular hypertrophy; Om = cardioomentopexy by left thoracotomy.)

Fig 5.

mRNA expression of (A) collagen type 1 α 1 chain (COL1a1), (B) lysyl oxidase-like protein 1 (LOXL-1), (C) nuclear protein-1 (NUPR1), and (D) transforming growth factor-β1 (TGF-β1). (LVH left ventricular hypertrophy; Om = cardio-omentopexy by left thoracotomy.)

Comment

In this study we demonstrated the treatment effects of cardio-omentopexy in a rat model of LV pressure overload. Banding of the ascending aorta is the most common method for creating a rat model of LVH, which causes aortic constriction and an increase in LV afterload. The normal average diameter of a rat’s ascending aorta has been measured as 2.16 mm [11]. We constricted it to 1.65 mm (with a 16-gauge needle) and observed the development of LVH with increased cardiomyocyte size, decreased LVEF, and increased myocardial fibrosis in the LVH group.

The LVH+Om group maintained normal echocardio-graphic cardiac function and dimensions. Histologically, cardiomyocyte size and myocardial fibrosis in the LVH+Om group were similar to the un-banded groups. Cardio-omentopexy appeared to preserve sarcoplasmic reticulum calcium transport function as evidenced by normalized expression of SERCA2b mRNA. Omentopexy also reduced the development of cardiac interstitial fibrosis as evidenced by both histologic examination and the lower expression of COL1a1, LOXL1, NUPR1, and TGF-β1 in the LVH+Om group versus LVH rats.

Interestingly, the gene expression of ANP after aortic banding was not substantially altered by the cardio-omentopexy. ANP secretion by atrial cardiac myocytes increases during hemodynamic overload in an effort to reduce volume by elevating renal sodium excretion. In this study, ANP expression was increased in both the LVH and LVH+Om groups compared with the Sham and Sham+Om groups. This is consistent with increased stretch of atrial walls after aortic banding. Because we did not intervene to remove the band around the ascending aorta for the duration of the study, the aortic stenosis was constant, and LV stress was maintained at a high level in both the LVH and LVH+Om groups. We speculate that this explains the similar expression of ANP in these two groups.

In the present study, we did not find evidence for decreased mRNA expression of SERCA2a in hypertrophied hearts. However, it appears that cardio-omentopexy has the potential to upregulate, or normalize, the expression of SERCA2b in hypertrophied hearts. SERCA plays a central role in regulating intracellular calcium homeostasis and myocardial contractility. SERCA2a is the major cardiac isoform, whereas SERCA2b is a minor cardiac isoform. A number of studies have reported a decrease in SERCA2a in heart failure, and its down-regulation is associated with impaired calcium cycling. Interestingly, a decrease in SERCA mRNA occurred in LV myocardium from failing animals after 20 weeks of banding but not in non-failing hypertrophied hearts [11]. In our study, the period of aortic banding was 16 weeks, and we found the similar result of SERCA2a expression. Silveira and colleagues [12] also observed that quantified SERCA2a was similar between sham and aortic banding. One could speculate that SERCA2a expression is maintained in compensated hypertrophy, characterized by mild abnormalities of systolic function as seen in this study.

SERCA isoforms are distinguished by subcellular localization as well as affinity for calcium (2b > 2a). In addition to the SERCA2b isoform displaying a higher affinity for Ca²⁺, it also has a lower turnover rate compared with SERCA2b, and many of these properties may be associated with the longer SERCA2b carboxyl terminal [13]. Of importance, experimental overexpression of SERCA2b leads to an increase in sarcoplasmic reticulum calcium transport function and increased cardiac contractility [14]. These findings point to the importance of SERCA2b, relative to SERCA2a, and suggest that increased expression of SERCA2b appears to be sufficient to improve myocardial contractility. This rationale is in line with our study, in which cardio-omentopexy normalized SERCA2b expression, which may be associated with the enhanced mechanical function observed in the LVH+Om rats. Future studies will directly test the association of calcium handling and contractility with elevated SERCA2b after cardio-omentopexy.

Myocardial fibrosis plays an important role in the LV dysfunction of patients with LVH, particularly with regard to diastolic dysfunction. In the present study, diffuse myocardial fibrosis was demonstrated by trichrome staining in the LVH group. Cardiac fibrosis is characterized by the induction of profibrotic growth factors and activation of cardiac fibroblasts, which have an important role in the development of the myocardial remodeling process [15]. Activated cardiac fibroblasts change their phenotype into myofibroblasts by production of extracellular matrix proteins. Collagen type 1 and extracellular matrix proteins can be excessively deposited, which occurs in fibrotic disease, resulting in myocardial dysfunction and failure. TGF-β1, a key mediator of cardiac fibroblast activation, has a major influence on collagen type 1 production. Enhanced TGF-β1 expression is often accompanied by increased collagen synthesis, deposition, and myocardial fibrosis [16]. Collagen type 1, the major component of extracellular matrix, is composed of two a1 chains and one a2 chain, which are encoded by the COL1a1 and COL1a2 genes [17]. TGF-β1 upregulates the expression of collagen type 1 in mRNA and protein levels through the DNA demethylation of COL1a1 promoter regions in rat cardiac fibroblasts [18].

LOX family proteins are copper-dependent, extracellular matrix-embedded amine oxidases that have critical roles in the crosslinking of collagen and elastin fibrils, resulting in the deposition of insoluble collagen and elastic fibers [19]. Mammalian genomes have five isoforms encoding the prototypic LOX and LOXL1 through LOXL4 [20]. LOXL1 mRNA is upregulated in pressure overloaded hypertrophied hearts after abdominal aortic constriction [21]. LOX expression is directly correlated with LV stiffness in humans with heart failure [22]. LOX family members have also been shown to be upregulated by a variety of external factors including TGF-β1 [23], oxidative stress, hypoxia, and mechanical stress.

NUPR1, which was reduced in gene expression after omentopexy, is a stress-inducible protein involved in gene transcription, and a key player in the cellular stress response, as well as in apoptosis and autophagy [24]. NUPR1 is required for endothelin and angiotensin induction of cardiomyocyte hypertrophy; in cardiac fibroblasts, NUPR1 expression is also necessary and sufficient for tumor necrosis factor α-mediated induction of matrix metalloproteinase 9 and 13. NUPR1 levels are substantially increased in failing human heart samples, and therapeutic intervention can cause a reduction in the level of cardiac NUPR1 [25]. These findings suggest that NUPR1 might represent an excellent marker for the development of cardiac hypertrophy and be indicative of the ongoing remodeling process in the failing heart. Furthermore, the expression of NUPR1 is activated by TGF-β1 at the transcriptional level and is also regulated by the SMAD proteins, which are transcription factors specifically involved in the signaling of TGF-β1 superfamily members [26]. Thus, our findings of reduced TGF-β1 expression, as well as cardiac fibrosis, after omentopexy, are consistent with lower omentopexy-associated NUPR1 expression, as described further below.

In our LVH rats, substantially elevated LV mRNA expression of profibrotic biomarkers (TGF-β1, COL1a1, LOXL-1, and NUPR1) suggest a maladaptive response in this group. Cardio-omentopexy dramatically down-regulated the fibrosis formation in the hypertrophied heart as evidenced by both histologic examination and the gene expression of these profibrotic factors in the LVH+Om group, which was substantially lower than in the LVH group and not different from the unbanded sham rats.

Here, we demonstrate the functional and histologic benefit from cardio-omentopexy in a pressure overload, rather than ischemic, model of myocardial injury. We think that this initial observation of therapeutic efficacy warrants further investigation. For example, it will be interesting in future studies to determine whether improved regulation of angiogenic factors are associated with changes in capillary density and oxygen delivery to cardiac muscle. The identification of putative proangiogenic cellular populations by single cell mRNA sequencing could help identify the necessary cell subsets that are causally responsible for delayed disease progression. Of note, many young patients with congenital heart defects are predictably subjected to persistent hemodynamic overload that results in ventricular dysfunction. We speculate that early cardio-omentopexy may hold promise as an adjunct to conventional staged single ventricle palliation, for example, forestalling the maladaptive ventricular responses commonly observed as these patients age.

This is an initial observational study aimed at evaluating the therapeutic potential of cardio-omentopexy in a rat model of LV pressure overload. Further studies with direct pressure measurements can quantify the actual increase in LV pressure in banded rats versus sham rats. The study is based on a limited number of small animals of a single species. The pressure overload in this protocol created a statistically significant, but not overly dramatic, reduction in LVEF. Further recapitulation in genetically altered mice also could potentially allow us to examine underlying genetic and molecular factors in the omentalmediated amelioration of cardiac pathophysiology. Additional studies will need to be performed to assess the impact of cardio-omentopexy in larger animals, as well as the age-related variation of these beneficial effects. This study was not designed to elucidate the mechanism of the beneficial impact of cardio-omentopexy in pressure overload LVH, which will be a goal of future investigation.