Mitigation of Myocardial Fibrosis by Molecular Cardiac Surgery Mediated Gene Overexpression

Abstract

Objective

Heart failure is accompanied by upregulation of transforming growth factor beta signaling, accumulation of collagen and dysregulation of sarcoplasmic reticulum calcium ATPase cardiac isoform 2a (SERCA2a). We examined the fibrotic response in small and large myocardial infarct and the effect of overexpressing the SERCA2a gene.

Methods

Ischemic cardiomyopathy was induced via creation of large infarct or small infarct in 26 sheep. All animals were divided into four groups: small infarct; large infarct with heart failure; gene treated (large infarct with heart failure followed by AAV1.SERCA2a gene construct transfer by molecular cardiac surgery with recirculating delivery); and control group.

Results

Heart failure was significantly less pronounced in the gene treated and small infarct groups than in the large infarct group. Expression of transforming growth factor beta signaling components was significantly higher in large infarct compared to small infarct or gene treated. Further, both the angiotensin II type 1 receptor and angiotensin II were significantly elevated in small and large infarcts, while gene treatment diminished this effect. Active fibrosis with de novo collagen synthesis was evident in large infarct, while small infarct and gene treatment groups showed less fibrosis with a lower ratio of de novo to mature collagen.

Conclusions

The data presented supports that the progression of fibrosis is mediated through increased transforming growth factor beta and angiotensin II signaling, which is mitigated by increased SERCA2a gene expression.

INTRODUCTION

Myocardial infarction (MI) results in extensive left ventricular (LV) remodeling in both infarcted and non-infarcted zones, with subsequent development of fibrosis and heart failure (HF). At the molecular level the remodeling is accompanied by a reparative deposition of extracellular matrix in an attempt to maintain cardiac structural integrity. Altering this process presents a significant therapeutic opportunity in the management of HF [1]. In contrast to traditional treatments, gene therapy appears promising due to the ability to alter the genetic structure of myocardial cells and the extracellular matrix. One important contributor to the development of fibrosis is the transforming growth factor beta (TGFβ1)-SMAD signaling cascade, which stimulates collagen expression and other downstream pro-fibrotic targets and is markedly up-regulated after MI [2]. It is also a potent regulator of the multiple stages of the cell cycle in the heart and is integral to infarct healing, myocardial hypertrophy, and post-infarction remodeling [3]. TGFβ1 signaling can be inhibited with antisense oligonucleotides [4] and neutralizing antibodies [5] resulting in attenuated LV remodeling and reduced interstitial fibrosis [6]. Yet, the contribution of TGFβ1-SMAD signaling on the development of cardiac fibrosis as a function of MI extent and zonal proximity to the infarct is unknown [2]. In addition the ability to manipulate these pathways with gene therapy is largely unclear and controversial [6]. Moreover, the impact of gene therapy acting on cellular structures to modify the structural integrity of myocytes is a new area of research.

The cardiac extracellular matrix is composed mostly of fibrillar collagen type I (tensile strength) and type III (elasticity and structural integrity) [7]. Both types are synthesized by cardiac myofibroblasts wherein a procollagen (a prerequisite of fibrillar collagen) forms in the sarcoplasmic reticulum (SR) and is dependent on the function of sarcoplasmic reticulum calcium ATPase 2a (SERCA2a). In HF, the ratio of type I to type III collagen as well as the ratio of mature to de novo collagen is altered [8]. However, the changes in these ratios after ischemic injury and the kinetics of de novo collagen deposition in border and remote zones of infarcted hearts are not yet understood [9].

A hallmark of HF is abnormal intracellular calcium ion (Ca²⁺) handling and down-regulation of SERCA2a. In failing hearts, there is dysfunction in excitation-contraction coupling and deficient SR Ca²⁺ uptake. We and others have demonstrated that the normalization of SERCA2a expression improves cardiac function in the infarcted heart [10, 11]. Since SERCA2a closely controls intracellular Ca²⁺, we hypothesized a potential link between the effects of overexpression of SERCA2a and MI-stimulated fibrogenesis.

METHODS

Animals

All animals received humane care in compliance with the National Institutes of Health and the local Institutional Animal Care and Use Committee. Dorsett male sheep (n=26) weighing 46.1±3.6 kg were used. All animals were divided into four groups. Group 1 (n=6): animals with small MI without clinical signs of HF after proximal ligation of the first branch of the circumflex artery (OM1); Group 2 (n=10): animals with large MI with clinical HF after proximal ligation of the first two branches of the circumflex artery (OM1+OM2); Group 3 (n=7): animals with large MI followed by gene construct (AAV1.CMV.SERCA2a) transfer by molecular cardiac surgery with recirculating delivery (MCARD) after 4 weeks (large MI/SERCA); and Group 4 (n=3): control animals, with tissue collected for molecular studies before any procedure. Animals in groups 1–3 were euthanized at 12 weeks.

All data presented in this article is new; however, we have published some data from subsets of these animals previously [12,13]

AAV1/SERCA2a Vector Production

A recombinant single stranded adeno-associated viral vector (AAV), serotype 1, encoding SERCA2a under the control of the human immediate early cytomegalovirus (CMV) gene promoter with a splice donor/acceptor sequence and polyadenylation signal from the human globin gene was constructed in a validated pre-clinical grade system resulting in high quality ssAAV1.CMV.SERCA2a titers by the Penn Vector Core (http://www.med.upenn.edu/gtp/vectorcore/).

Surgical Procedures and Gene Delivery

Infarct Model Creation

Surgical details of the creation of small and large myocardial infarction have been described previously [10, 12]. Molecular Cardiac Surgery with Recirculating Delivery (MCARD) – As with the infarct creation procedures, the basic surgical parts of the MCARD procedure have been published [10, 144]. We briefly describe the main features. The heart was isolated and closed-loop cardiac circuit flow was initiated. The virus solution consisting of 10¹³ genome copies of ssAAV1.CMV.SERCA2a was injected into the coronary sinus catheter and recirculated. The circuit was then flushed to wash out residual vector and the chest was closed. All animals received postoperative veterinary intensive care unit management.

Hemodynamic Evaluation and Cardiac Function Assessment

Magnetic resonance imaging (MRI) (Signa LX, GE Healthcare, UK) acquisition and analysis of cardiac hemodynamics were generated for each animal at different timepoints [15] More details are described in supplemental materials. [].

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) and Western Blot

For mRNA analysis, total RNA was isolated using Trizol Reagent and cDNA synthesis performed as previously described [16]. Western blots were done as previously described [12] and normalized to glyceraldyhde-3-phosphate dehydrogenase (GAPDH, Santa Cruz Biotechnology, sc-48166). More details are described in supplemental materials.

Assessment of Collagen Content by Quantitative multispectral imaging of Herovici’s Polychrome

Quantification of collagen content and distinguishing between De Novo and Mature collagen were performed using Herovici’s polychrome [17]. More details are described in supplemental materials.

Myocardial Fibrosis Assessment

Magnetic Resonance Imaging

After localization of the heart and acquisition of electrocardiographically gated cine images, gadolinium-diethylenetriaminepentacetate was administered to the animal. To quantify the precise amount of fibrotic tissue, hyperenhanced areas were manually traced on the short-axis images. Scar tissue was expressed as percentage of the LV [15].

Morphological assessment of collagen volume fraction

LV sections were stained with Sirius Red. Quantifying of fibrosis was performed with computer-assisted morphometry measurement using a digital camera for microscopy. Each field was scanned with a computer-generated microscale and analyzed blindly using 9–11 randomly selected tissue sections from each of the hearts with image analysis software [18].

Electron Microscopy Analysis and Histology

A minimum of 10 random fields for border zone (BZ) and remote zone (RZ) were utilized for electron microscopy analysis of myofibrillar disruption in each group [19]. More details are described in supplemental materials.

Myocardial infarct size quantification

Planimetry

After harvesting, the LV was dissected and photographed. The endocardial surface and the infarcted scar were traced onto a transparency and quantified using computed planimetry. Sections were then embedded in paraffin for immunohistochemistry, and for molecular and genetic studies were stored at a temperature of −80°C until use. MRI described in section: Hemodynamic Evaluation and Cardiac Function Assessment.

Angiotensin II (Ang II) and Angiotensin II Type I receptor (AT₁R)

Expression of Ang II and AT₁R are described in supplemental materials.

Definition of Cardiac Zones

After heart harvesting and sectioning through the LV, scarred IZ was identified as discolored fibrous tissue in the postero-lateral part of the LV. Samples from a 1-cm zone extending from the outer discolored sections were taken as BZ. Right ventricle wall myocardium was identified as RZ.

Statistical Analysis

Most data were presented as average ± SEM. The Kolmogorov-Smirnov test for normality was applied to each data set prior to further statistical testing. Single ANOVA was performed across time for each parameter, then paired t-tests (P<0.05 significance) were used to assess any difference between any two time points or differences at the same time point between groups. Statistical comparisons for whisker plot are made based on one way ANOVA comparing the median/50% percentile with Tukeyat multiple comparison test.

RESULTS

Hemodynamic data in large and small myocardial infarct model and SERCA2a overexpression

Hemodynamics at baseline and 12 weeks post MI are listed in Table 1. Baseline cardiac function did not differ significantly between the groups. At 12 weeks large MI hearts revealed decreased ejection fraction (EF) (31.0±7.8%) compared to the large MI/SERCA2a delivery group (43.3±11.1%) and the small MI group (42.6±6.9%), both p<0.05. A significant improvement in cardiac output (CO), end diastolic volume (EDV) and end systolic volume (ESV), stroke volume index (SVi) and wall thickening was observed in the large MI/SERCA group compared to the large MI group. Infarct size for large MI (20.4±2.6%) and for large MI/SERCA (19.8±2.4%) did not differ significantly from each other but were significantly greater than for small MI (7.3±0.7%), both p<0.05.

Table 1.

Hemodynamics data. Baseline and 12-week hemodynamics for animals in each group. CO, cardiac output; EDV, end diastolic volume; EF, ejection fraction; SVi, stroke volume index; MI, myocardial infarct; SERCA, sarcoendoplasmic reticulum calcium ATPase cardiac isoform 2a.

	Baseline			12 weeks post-MI
	Large MI/SERCA	Small MI	Large MI	Large MI/SERCA	Small MI	Large MI
CO (mL/min)	2984 ± 123	2650 ± 258	2931 ± 178	2732 ± 220*†	2246 ± 158*‡	1861 ± 114‡
EDV	49.8 ± 1.9	46.4 ± 4.5	51.4 ± 4.7	83.2 ± 9.3†‡	56.2 ± 5.2*‡	94.8 ± 13.6†‡
ESV	19.5 ± 1.0	21.0 ± 4.3	22.1 ± 3.7	48.1 ± 7.5‡	32.5 ± 3.9*‡	66.2 ± 10.9†‡
EF (%)	60.6 ± 2.7	56.4 ± 4.9	58.0 ± 4.6	43.3 ± 4.2*‡	42.6 ± 2.6*‡	31.0 ± 3.5†‡
SVi	27.1 ± 2.1	22.6 ± 1.3	24.5 ± 2.1	30.1 ± 2.8*†	19.4 ± 2.1	20.9 ± 2.7
Wall Thickness	11.1 ± 0.6	11.8 ± 1.0	10.5 ± 0.8	13.8 ± 0.5‡	14.0 ± 0.5‡	12.6 ± 0.7
Wall Thickening (%)	26.9 ± 3.1	20.5 ± 5.1	26.5 ± 4.7	17.1 ± 2.4*‡	18.6 ± 6.6	7.4 ± 4.9‡
dP/dt Max (mmHg/s)	1151 ± 92	1246 ± 62	1165 ± 69	1084 ± 73	1237 ± 178	931 ± 41
dP/dt Min (mmHg/s)	−1769 ± 217†	−1235 ± 103*	−1856 ± 199†	−1218 ± 35	−1318 ± 115	−1259 ± 72‡
Heart Rate (bpm)	93±5	95±5	94±4	95±5	95±6	95±4
Hematocrit (%)	31±2	31±2	31±2	29±2	29±2	27±1
Mean Arterial Pressure (mmHg)	94±3	90±6	94±4	95±4	92±4	94±3

^*p<0.05 vs. Large MI

^†p<0.05 vs. Small MI

^‡p<0.05 vs. Baseline of same group

SERCA2a gene expression

Western blot showed a significant increase in SERCA2a protein levels, especially in large MI/SERCA animal BZ (0.28±0.018 RU) compared to small MI (0.14±0.023 RU) and large MI (0.16±0.045 RU) BZ, p<0.05. The same trend was found in the IZ but data did not reach significance (p>0.05) (Figure 1). At 12 weeks all animals had detectable levels of transgene in the BZ (4372±156.7 genome copies, GC, per 100 ng DNA) and IZ (1524±109.5 GC). This was significantly higher than in the RZ (246±99.8 GC) and liver (607±53.9 GC).

Figure 1.

Western blot quantifying SERCA2a protein expression in the BZ of Small MI, Large MI, and Large MI/SERCA groups * = p<0.01 vs. Small MI, Large MI.

Overexpression of SERCA2a reduces expression of TGFβ/SMAD signaling components in large MI hearts

TGFβ1 pro-fibrotic signaling genes [TGFβ1, TGFβRII, SMAD3, colα1] were examined from cardiac tissue isolated from different MI zones (IZ, BZ and RZ) among the three experimental groups. TGFβ1 signaling components were significantly up-regulated in large MI compared to small MI and SERCA2a overexpression mitigated these molecular alterations (Figure 2 A–C). Up-regulation of the TGFβ1 signaling cascade was particularly marked in the IZ (scar tissue) but many of these genes were also up-regulated in the BZ and RZ of large MI animals. Western blots of TGFβRII and colα1 in IZ and BZ revealed increased expression in both small and large MI compared to control and large MI/SERCA.

Figure 2.

(A–C) mRNA expression levels of TGFβ1 signaling components and downstream target [Colα1(I)] were assessed in heart tissue from (A) infarct zone, (B) border zone and (C) remote zone. (D) Protein content in infarct zone (IZ) and border zone (BZ) of Colα1(I), TGFβ1, TGFβRII and Smad3 as measured by Western blot for Control, Large MI/SERCA, Small MI, and Large MI groups. GAPDH protein expression measured for normalization. * = p<0.05 vs. Small MI, Large MI/SERCA2a.

Plasma Ang II and myocardial AT₁R

Ang II is able to directly stimulate fibrogenesis in the heart. The actions of Ang II can also be inhibited with AT₁R antagonists, which are currently used in HF therapy and thus these represent important parameters to assess after gene therapy. Ang II and AT₁R expression, assessed by immunohistochemistry, both increased significantly from control in small MI and large MI. Delivery of SERCA2a significantly reduced plasma Ang II levels and myocardial AT₁R expression compared to large MI (Supplemental Figure 1A–F).

SERCA2a decreases activation of myofibroblasts and fibronectin

Ventricular fibroblasts normally maintain low turnover of fibrillar collagens but transform to an activated myofibroblast phenotype (myoFb) in response to injury, wherein αSMA and fibronectin expression are markers of this transformation. αSMA protein expression was significantly increased in the BZ of large MI (p<0.05) and small MI (p<0.05) groups compared to large MI/SERCA2a and control (Figure 3A). Furthermore, fibronectin protein expression was significantly increased only in large MI (p<0.05) compared to large MI/SERCA, with large MI/SERCA also showing a nonsignificant decrease compared to both control and small MI (Figure 3B). This same trend was observed in IZ and RZ; however, the data did not reach significance (p>0.05).

Figure 3.

Changes in αSMA protein (A) as well as Fibronectin protein (B) measured by Western blot, normalized to GAPDH. * = p<0.05 vs. Large MI/SERCA. † = p<0.05 vs. Control

Delivery of SERCA2a to large MI animals halts synthesis of new collagen

We performed multispectral imaging of Herovici stained samples to differentiate between mature and de novo collagen synthesis. The degree of myocardial collagen deposition was quantified in heart tissue from each zone of damage (IZ, BZ, and RZ) for each experimental group (small MI, large MI, large MI/SERCA). There was a significant increase in the ratio of de novo (blue staining) to mature collagen (red staining) (Figure 4J) in all zones of the large MI animals (Figure 4B, E, H) compared to small MI (Figure 4A, D, G). Further, overexpression of SERCA2a reduced the ratio of de novo (blue fibrils) to mature (red fibrils) collagen in IZ and BZ (Figure 4C, F) compared to large MI (Figure 4B, E). There was less overall collagen fibril deposition in RZ of large MI/SERCA, thus there was limited amounts of mature collagen observed in this zone (Figure 4I). These data suggest a persistent fibrotic response (i.e., increased deposition of new collagen fibrils) in large MI while fibrosis in the small MI and large MI/SERCA has been mitigated.

Figure 4.

Herovici staining was performed in IZ, BZ and RZ in each group. (A–C) infarct zone (IZ), (D–F) border zone (BZ), (G–I) remote zone (RZ). Arrows mark collagen fibrils. 20× magnification. Blue staining indicates new collagen fibrils; red staining indicates mature collagen fibrils. (J) Quantification of de novo (new) to mature collagen deposition assessed by Herovici staining. * = p<0.05 vs. Small MI. † = p<0.05 vs. Large MI/SERCA.

Delivery of SERCA2a to large MI preserves myocardial structure

The significant difference was in the specimens from the BZ between MI/SERCA group and large MI. In MI/SERCA group at gross histology myocytes appeared nearly normal with nuclei in the center and widely spaced intercalated discs, signs of ischemic injury were less. In the BZ of the large MI group, transmission electron microscopy revealed substantial cardiac damage associated with pronounced sarcomere fragmentation including longitudinal disruption and disorganization of myofibrils as well as disrupted Z-disks. Mitochondrial damage was indicated by indistinct, fragmented, or vesiculated cristae in both the BZ and remote zone (RZ) of the large MI specimens. Notable twin T tubule dilatation was present in all BZ fields, but only in large MI of the RZ fields. In large MI/SERCA and small MI samples, morphological structure was largely preserved (Figure 5A–F). Semiquantitative analysis showed increased levels of myofibrillar disruption in large MI BZ but not small MI or large MI/SERCA BZ (Figure 5G).

Figure 5.

(A–F) Transmission electron microscopy images of border zone (BZ) and remote (RZ) tissue from each group. Examples of mitochondria (m) and myofibrils (mf) with clearly delineated sarcomeres are labeled. Varying degrees of mitochondrial fragmentation and cristae destruction are shown (black arrows) along with twin T tubules (white arrows). Images 10,500× magnified, scale bar 2µm. (G) Semiquantitative analysis of myofibrillar disruption seen in images A–F.

DISCUSSION

We demonstrate that MCARD-mediated SERCA2a gene delivery disrupts activation of the TGFβ1/SMAD signaling cascade, inhibiting de novo collagen synthesis and downregulating Ang II and AT₁R.

A growing body of evidence from studies conducted in the heart indicates that post myocardial infarction-activated TGFβ1 signaling triggers a system involving SMAD proteins, which are phosphorylated and subsequently translocated to the nucleus to regulate target gene transcription [20]. We found that TGFβ1 signaling genes are dramatically upregulated in large MI in all myocardial zones compared to small MI and large MI/SERCA groups. In response to the fibrogenic stimuli resulting from ischemic stress, cardiac fibroblasts switch to a myofibroblast phenotype and express αSMA as well as fibronectin. We found that in the large MI/SERCA group the expression of αSMA and fibronectin was markedly decreased relative to other groups.

From our hypothesis the mechanism appears as follows (Figure 6 The overexpression of SERCA2a in ischemic cardiomyopathy mitigates the otherwise MI-associated cytosolic Ca²⁺ overload, prolongation of the cytosolic Ca²⁺ transient time and the consequent increase in end-diastolic Ca²⁺ concentration [11]. Developing cytosolic Ca²⁺ overload increases the activity of Ca²⁺ dependent serine/threonine phosphatase calcineurin [21]. The activated calcineurin binds and dephosphorylates calcium-sensitive, calcineurin-nuclear factor of the activated T-cell (NFATc) [21,22], which translocates from the cytosol to the nucleus and interacts directly with the cardiac transcription factor GATA4 [23,24] This triggers a signal transduction pathway: renin splits the protein angiotensinogen, producing angiotensin Iwhich is converted into Ang II, [24]. In cardiac fibroblasts, Ang II stimulates expression of extracellular matrix proteins such as αSMA and fibronectin, mainly through AT₁R [25]. There is strong evidence that Ang II regulates TGFβ1 expression in adult ventricular cardiomyocytes [26,27]. We postulate that SERCA2a overexpression normalizes calcium signaling and decreases activation of the Ang II, providing a mechanism for the effect of SERCA2a overexpression on the TGFβ1-SMAD pathway and attenuation of fibrosis.

Figure 6.

Proposed mechanism of action from initial Ca²⁺ imbalance due to SERCA2a downregulation through Ang II and TGF-β signaling to fibrosis. NFATc: Nuclear factor of activated T-cells, cytoplasmic. GATA4: GATA family zinc finger transcription factor 4. NAD(P)H oxidase: Nicotinamide adenine dinucleotide phosphate-oxidase. PKC: Protein kinase C. P38 MAPK: p38 mitogen-activated protein kinases. AP1: transcription factor. Smad: SMAD family signal transduction proteins.

These data confirmed that SERCA2a overexpression reduces the secretion of both fibrogenic mediators, support that SERCA2a overexpression diminishes fibroblast activation.

It is still controversial whether advanced fibrotic tissue can be reverted to normal tissue architecture. Indeed, if fibrosis is accompanied with changes in the mechanical environment including tissue segment length and tension, leading to disruptive changes in myofibrils with increased end systolic and diastolic dimentions, then full reversal is not possible [28]. []. As is well known, the SR is closely associated with the myofibrils. The action potential spreads along the sarcolemma and transverse tubular system into the myofibrils, activating the ryanodine receptor that releases Ca²⁺ from the SR. Fibrogenesis changes these processes. So we sought to trace the impact of fibrosis on the structural integrity of myofibrils and we found that myocardial structural integrity is preserved after SERCA2a gene therapy.

Collagen accumulation is an integral part of fibrosis, resulting from a shift in the balance between collagen synthesis and degradation. The expression of these collagens in the ischemic cardiac muscle remains elevated long term [29]. Reversal of fibrosis is prevented by incomplete ECM degradation when the appropriate cellular mediators are no longer present [30]. Area of fibrosis assessed by MRI and quantitative morphometry demonstrated no significant difference between the large MI and large MI/SERCA groups at 12 weeks. However, the ratio of de novo to mature collagen was significantly different in BZ and RZ. Thus the SERCA gene is associated with a previously undemonstrated alteration of the composition of the cardiac extracellular matrix.

Infarct expansion includes the recruitment of adjacent BZ. This fully perfused, normally contractile myocardium undergoes advanced remodeling that leads to heart failure. Many of the molecular processes taking place in BZ are not well defined [31]. Preventing the breakdown of the extracellular matrix could stiffen the IZ, arresting infarct expansion and remodeling. Gene therapy with MCARD/SERCA2a in developing HF significantly increases gene delivery to the BZ, restores contractile function, and mitigates fibrosis in this zone. Our ovine model of large MI shows extensive LV remodeling in IZ and BZ accompanied with a permanent fibrogenic response. Contrary to previous data obtained in a murine model showing that fibrosis is independent of the severity of ischemia [32], fibrotic signaling ceases after initial scar formation in our small MI model – a major difference in fibrotic response due to infarct size. Due to the lack of significant LV remodeling in the small MI group, we did not administer gene therapy under this paradigm. These data provide novel insights into why in the presence of a small injury, remodeling is compensatory and limited whereas after a larger insult it is progressive and maladaptive.

Myocardial infarction is associated with dynamic changes in myocytes and fibroblasts and in the composition of the extracellular matrix in all cardiac zones. Many studies have demonstrated that the scar and BZ are metabolically active tissues with phenotypic modulation of fibroblasts, secretion of extracellular matrix proteins and activation of angiogenic pathways.

The highly efficient cardiac-specific MCARD delivery platform may have been vital in achieving sufficiently high SERCA2a expression to produce these results. MCARD/SERCA2a gene delivery may be a translatable approach to limit post-MI remodeling through halting the fibrogenic response. We chose to perform gene delivery at four weeks after MI due to the following reasons: First, at this time the proliferative and maturation processes of fibrogenesis overlap each other [33]. Myofibroblasts begin to appear early after infarction, remain abundant over 4 weeks, and are still present in the infarcted heart at week 8 after MI [29]. Next, the TGF-β system receptors and downstream signaling components are elevated up to eight weeks after MI [34] and, the plateau of AAV1 expression occurs at 4–6 weeks post viral infection into the cardiomyocytes. And finally, we wanted to eliminate all side effects associated with the surgical creation of MI.

Limitations of the study

We did not use a small MI/MCARD group due to preliminary data showing no upregulation of the TGF-β/SMAD system or Ang II in small MI as was expected. We also did not incorporate a control group with MCARD-null vector due to limited resources and that will be clarified by further studies. However, we believe that it is highly unlikely that the MCARD procedure - itself - is responsible for the dramatic attenuation of fibrogenesis observed with overexpression of SERCA2a. Breakdown of the extracellular matrix and collagen turnover after myocardial infarction occurs with many different matrix metalloproteinases (MMPs), zinc-dependent enzymes, interleukins etc. and it is very difficult to assay every signaling pathway involved in fibrogenesis with or without gene therapy. The mechanism responsible for fibrosis is still not well understood and may involve either a direct effect of stretch or a diffusion of signals that originate in the ischemic area through the extracellular matrix. Fibroblasts and myocytes appear to be closely connected via intracellular communications in cardiac injury thus we cannot completely eliminate role of myocyte-fibroblast interrelationship in the event of fibroblast proliferation/activation and ECM turnover. For RZ we chose right ventricle, the most distant area from infarction which is not subject to the same stress as the IZ/BZ undergoing geometric distortion and infarct expansion. This partly explains the low gene expression due to the anatomical pattern of the coronary venous drainage in the sheep heart.

Abbreviations

TGFβ

transforming growth factor beta

SERCA2a

sarcoendoplasmic reticulum calcium ATPase isoform 2a

MCARD

molecular cardiac surgery with recirculating delivery

AT₁R

angiotensin II receptor 1

Ang II

angiotensin II

SMAD

an intracellular signaling protein family

Colα1

collagen alpha 1 type 1

SR

sarcoplasmic reticulum

OM1/2

first/second obtuse marginal coronary arteries

IZ

infarct zone

BZ

border zone

RZ

remote zone

EDV

end diastolic volume

ESV

end systolic volume

SVi

stroke volume index

GC

genome copies

DAB

3,3’-diaminobenzidine

myoFb

myofibroblast

NFATc

calcineurin-nuclear factor of the activated T-cell