Reduced Cardiac Remodeling and Function in Cardiac-Specific EP4 Receptor Knock-Out Mice with Myocardial Infarction

Abstract

We have previously shown that: 1) COX-2 inhibition reduces cardiac hypertrophy and fibrosis post myocardial infarction (MI) in a mouse model; and 2) prostaglandin E₂ stimulates cardiomyocyte hypertrophy in vitro through its EP4 receptor. Since the role of cardiac myocyte EP4 in cardiac function and hypertrophy in vivo is unknown, we generated mice lacking EP4 only in cardiomyocytes (CM-EP4 KO). 12-14-wk-old mice were evaluated using echocardiography and histology. There were no differences in ejection fraction (EF), myocyte cross-sectional area (MCSA) and interstitial collagen fraction (ICF) between KO mice and littermate controls. To test the hypothesis that EP4 is involved in cardiac remodeling following MI, we induced MI by ligating the left anterior descending coronary artery. Two weeks later, the mice were subjected to echocardiography and hearts removed for histology and Western blot. There was no difference in infarct size between KO mice and controls; however, KO mice showed less MCSA and ICF than controls. Also, CM-EP4 KO mice had reduced EF. Since the transcription factor Stat-3 is involved in hypertrophy and protection from ischemic injury, we tested whether it was activated in control and KO mouse hearts after MI. Western blot indicated that Stat-3 was activated in control hearts following MI but not in KO hearts. Thus, CM-EP4 deletion decreased hypertrophy, fibrosis and activation of Stat-3. However, cardiac function was unexpectedly worsened in these mice. We conclude that cardiac myocyte EP4 plays a role in hypertrophy via activation of Stat-3, a process which seems to be cardioprotective.

Introduction

Cardiac hypertrophy is a common pathological lesion induced by various extracellular stimuli, including mechanical stress (e.g. systemic hypertension), hormones (e.g. angiotensin II and β-adrenergic agents), and cytokines (e.g. TGFβ). During the early stages of various cardiac diseases, hypertrophy is compensatory, but sustained stimulation may lead to excessive cardiac remodeling and ultimately heart failure. Understanding the cellular basis of cardiac hypertrophy is important when designing an appropriate therapy to regulate the hypertrophic response.

PGE₂ is a well-known pro-inflammatory prostanoid. It is synthesized from arachidonic acid by cyclooxygenase (COX) and PGE₂ synthase (PGES). There are two COX isoforms in mammalian cells. COX-1 is expressed constitutively in almost all tissues, and its PG products mediate physiological responses such as vascular homeostasis and gastroprotection. COX-2, although often undetectable in resting cells, is readily induced as an immediate early gene in response to cytokines, growth factors, phorbol esters, and bacterial lipopolysaccharides ^1,2. In addition to inflammatory effects, PGE₂ has also been shown to promote growth. PGE₂ enhanced cell proliferation in murine aortic smooth muscle cells ³ and NIH 3T3 cells ⁴. It also increased protein synthesis, cell size and BNP expression in cardiac myocytes, all markers of hypertrophic growth ^5,6. We have also shown that myocardial infarction (MI) increased generation of PGE₂ accompanied by upregulation of COX-2 in the mouse heart, and treatment with a specific COX-2 inhibitor for 2 weeks improved cardiac function and reduced hypertrophy and fibrosis ⁷. Given that COX-2 products have both deleterious and protective effects in the heart and vasculature ^8-10, it is important to understand how PGE₂ affects the pro-hypertrophic response.

PGE₂ exerts its biological actions through four different G protein-coupled receptors, EP₁, EP₂, EP₃ and EP₄, which differ in structure, ligand-binding properties, activation of signal transduction pathways, and tissue distribution ¹¹. EP mRNAs have been detected in the heart of several species, including humans, and EP₄ is the most abundantly expressed EP subtype ¹². Activation of EP₄ leads to accumulation of intracellular cAMP through interaction with a cholera-sensitive G_sα protein. EP₄ has also been linked to activation of p42/44 MAPK ¹³, and in cardiac myocytes it is involved in regulation of protein synthesis and EGFR transactivation ¹⁴. All 4 EPs have been knocked out in mice with no cardiovascular phenotypes in unstressed conditions, except that EP₄ KO mice died shortly after birth from patent ductus arteriosus ¹⁵. Using the global EP4 KO mouse (EP4 knocked out in all cells), the role of EP₄ has been examined acutely (24 hours or less) in ischemia/reperfusion models. An EP₄ agonist reduced infarct size when administered prior to ischemia/reperfusion injury, while ischemia/reperfusion injury enhanced infarct size in EP₄ KO mice, suggesting that EP₄ is cardioprotective in this model ¹². It is hard to extrapolate these results to predictions of how EP₄ contributes to cardiac function and structure during chronic disease processes (for example, cardiac remodeling post-MI). Since the aforementioned study used a global EP₄ KO model, the role of EP4 in cardiac myocytes is not clear. Thus, we developed a novel mouse model where EP₄ is deleted only in cardiac myocytes, and used these mice to study the effect of EP4 on cardiac remodeling following MI.

Methods

Generation and genotyping of CM-EP₄ KO mice

The CM-EP₄ KO mouse was created by the Cre-loxP mechanism ¹⁶ through interbreeding two transgenic mice lines. The EP₄^flox/flox mouse was generated by gene targeting and possessed loxP sites flanking exon 2 of the EP₄ receptor gene (“EP₄ floxed”) (obtained from Dr. Matthew Breyer, Vanderbilt University Medical Center). This strain can generate a conditional EP₄ “knockout” in that exon 2 will be deleted in cells expressing Cre recombinase. Importantly, the loxP sites do not alter the wild-type EP₄ phenotype in the absence of Cre ¹⁷. The second line has a Cre recombinase introduced as a transgene driven by an α-myosin heavy chain promoter (obtained from Dr. Michael Schneider, Baylor College of Medicine) ¹⁸. The α-myosin heavy chain is expressed primarily in adult cardiac myocytes, obviating the complications of EP₄ deletion early during fetal development. The colony was maintained by breeding mice homozygous for the floxed EP4 allele and hemizygous for the Cre transgene to mice homozygous for the floxed EP4 allele and lacking the Cre transgene. This scheme generated roughly equal numbers of CM-EP4 KO and “wild-type floxed” littermates, and use of littermates as controls assured that the backgrounds matched. The majority of each strain of mice was alive after birth and had similar survival rates.

Total DNA was extracted from heart and other tissues using a DNeasy tissue kit (QIAGEN) following the manufacturer's instructions. 100 ng of each DNA sample was amplified by PCR (40 cycles of 30 sec at 94°C, 30 sec at 60°C, and 30 sec at 72°C). The sense primer for EP4 was 5′-TCT GTG AAG CGA GTC CTT AGG CT-3′, and the antisense primer was 5′-CGC ACT CTC TCT CTC CCA AGG AA-3′. The primers distinguished the wild-type and floxed EP4 gene by amplifying a 243-bp fragment in the wild-type and a 344-bp fragment in EP₄^flox/flox. The sense primer for Cre was 5′-CCAGCTAAACATGCTTCATCGTCGTC-3′ and the antisense primer was 5′-ATTCTCCCACCGTCAGTACGTGAG-3′. The Cre primers amplified a 300 bp fragment. We also performed Western blot for Cre (antibody 69050-3 from Novagen) to confirm its expression in CM-EP4 KO mouse hearts.

Animal Protocols

All animal experiments were approved by the Henry Ford Health System Institutional Animal Care and Use Committee. A total of 141 mice were used for this study (KO = 62; littermate controls = 79).

Protocol 1

Effect of CM-EP4 KO on cardiac phenotypes in unstressed mice. 12-14-wk-old male CM-EP4 KO and “wild-type” littermates were used in these studies. Normal, unstressed mice were subjected to two-dimensional (2D) M-mode echocardiography and then hearts were harvested. There were 2 sets of samples for both KO and WT mice, with the first set used for extraction of DNA, RNA and protein and the second for histological analysis.

Protocol 2

Effect of CM-EP4 KO on cardiac remodeling post-MI. To induce MI, 12-14 week-old (22-25g) male mice were anesthetized with sodium pentobarbital (50 mg/kg IP), and a left thoractomy performed. The left anterior descending coronary artery (LAD) was ligated with an 8-0 silk suture placed near its origin at the edge of the left atrium as described previously ⁷. Ligation was deemed successful when the anterior wall of the left ventricle became pale. The ligature was positioned so as to produce a moderate infarct (20-50%). Mice were kept on a heating pad and monitored until awake. Two weeks post-MI, mice were subjected to two-dimensional echocardiography and then hearts were harvested. In all groups of KO and WT mice, half of the hearts were used for extraction of DNA, RNA and protein and half for histological analysis. The mortalities after MI (from surgery to sacrifice) were similar between KO and WT mice (42%).

Two-dimensional transthoracic echocardiography was performed on conscious mice using an Acuson 256 system (Mountain View, CA) with a 15-MHz linear transducer as reported previously ⁷. M-mode images from both short- and long-axis were used to determine left ventricular end-diastolic/systolic dimension (LVDd/LVDs) and shortening fraction (SF). Ejection fraction (EF) was obtained from the short-axis view and calculated by the formula: diastolic area – systolic area/ diastolic area. Diastolic measurements were made at the maximum left ventricle cavity dimension, whereas systolic parameters were measured during maximum anterior motion of the posterior wall.

Histological assessment of interstitial collagen fraction, myocyte cross-sectional area, infarct size and wall thickness of the infarct area

Mouse hearts were harvested and sectioned transversely into four slices from apex to base. Three sections (Sections A, B, and D) were frozen in isopentane and stored at −70°C for determination of interstitial collagen fraction (ICF), myocyte cross-sectional area (MCSA) and infarct size (IS), as previously described⁷. The wall thickness of the infarct area (WTIA) was measured from the section with the longest infarct zone. The infarct length of the epicardium was evenly divided into 6 portions and five evenly-spaced points were selected for measurement. The distance between the epicardium and endocardium of each point was measured and averaged to obtain WTIA. Measurements were made using MicroSuite Biological Imaging Software (Melville, NY).

RT-PCR analysis of EP_1-4 and BNP mRNAs

The infarcted region of the mouse heart was discarded and the remaining portion of the left ventricle was homogenized. Total RNA was extracted from tissue homogenates using an RNeasy Fibrous Tissue Mini Kit (QIAGEN) following the manufacturer's instructions. RT was performed using the Omniscript RT kit (QIAGEN) with 2 μg total RNA, 1 μg random primer (Invitrogen), and 10 U/μl RNAsin (Promega) in a 20-μl volume for 1 hr at 37°C. Primers were as follows: EP₁ sense: TCC CCA ATA CAT CTG TGG TGC; EP₁ antisense: TCC CCA GCG CCG GCG ATC ATC (718 bp); EP₂ sense: GTG GCC CTG GCT CCC GAA AGT C; EP₂ antisense: GGC AAG GAG CAT ATG GCG AAG GTG (536 bp); EP₃ sense: CAT GAT GGT CAC TGG CTT CGT;EP₃ antisense: GTC ACC ACC AGA GCC AGC AAG (438 bp); EP₄ sense: TGC TTC TGT GAA CCC CAT C; EP₄ antisense: GTG GTG TCT GCT TGG GTC A (326 bp); BNP sense: TGA AGG TGC TGT CCC AGA TGA; BNP antisense: GTG CTG CCT TGA GAC CGA A (241 bp); GAPDH sense: ATT CAA CGG CAC AGT CAA GG; GAPDH antisense: TGG ATG CAG GGA TGA TGT TC (482 bp). PCR conditions were: 94°C for 30 sec, 58–62°C for 30 sec, 72°C for 90 sec and a final extension at 72°C for 5 min. Cycle number was 40 for EP₃, 35 for EP₁, EP₂ and EP₄, 30 for BNP and 25 for GAPDH. The GAPDH signal was used for evaluating input RNA values. The PCR bands were quantified by densitometry and normalized by GAPDH, and ratios between different groups were compared.

Western blotting

Mouse hearts were homogenized in sample buffer (Tris 47.5 mmol/L (mM), 2% SDS, 10% glycerol, pH 6.8) or Tris-HCl 25 mmol/L, EDTA 0.5 mmol/L and EGTA 0.5 mmol/L-pH 7.5 containing a proteinase and phosphatase inhibitor cocktail (Roche). 40 μg protein per sample was electrophoresed and transferred to a PVDF membrane. Phospho(p)-Stat-3 and total(t)-Stat proteins were detected with the appropriate antibodies (Cell Signaling). The antibody-antigen reaction was detected using a secondary antibody linked to horseradish peroxidase. The protein bands were scanned, quantified and p-Stat normalized to total Stat-3 and ratios between different groups compared.

Statistics

Data were expressed as mean ± SE and analyzed by unpaired one-tailed Student's t-test (to compare data from WT littermates with MI vs CM-EP4 KO mice with MI). P < 0.05 was considered significant.

Results

Characterization of CM-EP4 KO mice

Using specific PCR primers and DNA extracted from the heart and kidneys, we tested whether EP₄ was knocked out only in the heart. In the top panel of Fig. 1, EP4^flox/flox gene (344 bp amplification product) was detected in the hearts of “wild-type floxed” littermates but at a lower level in CM-EP₄ KO mice. However, the EP4^flox/flox gene was detected in kidneys of both “wild-type floxed littermates” and CM-EP4 KO mice at similar levels (middle panel). Densitometry of the PCR data (lower panel) indicated that the EP4^flox/flox gene level was reduced by 92% in the CM-EP4 KO mouse hearts compared with “wild-type” littermates (P<0.001) but there were no differences in the kidneys. Next, we extracted DNA from multiple mouse tissues to further investigate cardiac-specific deletion of the floxed EP4 gene. The EP4^flox/flox gene was detected in the aorta, brain, kidney and liver of both CM-EP4 KO and WT mice, but not in the heart of CM-EP4 KO mouse (Fig. 2A and B). Finally, using RT-PCR we evaluated EP4^flox/flox mRNA extracted from mouse heart and kidney and found that the mRNA was decreased by 83% (p<0.01) in the hearts of CM-EP4 KO mice vs “wild-type floxed” littermates, but there were no differences in EP4^flox/flox mRNA in kidney from the two strains of mice (Fig. 3). Hence, the EP4 gene and its gene product were deleted only in the hearts of CM-EP4 KO mice.

Figure 1. PCR of EP4 gene in mouse heart and kidney.

The upper panel is a representative agarose gel of PCR products from the hearts of “wild-type floxed” and CM-EP4 KO mice. “EP4^flox/flox” is a 344 bp PCR product generated from the EP4^flox/flox gene. “EP4 WT” represents a 243 bp PCR product of the wild-type EP4 gene generated from tail DNA of a C57Bl6/J mouse. GAPDH is the control gene. The DNA ladder lane shows the position of the 500 bp DNA fragment. The graph is the densitometric quantification of PCR-amplified DNA from the hearts of “wild-type floxed” littermates and CM-EP4 KO mice (n=5-6).

Figure 2. PCR of EP4 gene in mouse tissues.

The figure is a representative agarose gel of PCR detection of the floxed EP4 gene in multiple tissue samples of CM-EP4 KO mice (panel A) and littermate controls (panel B). The DNA ladder shows the position of the 500 bp DNA fragment.

Figure 3. RT-PCR of EP4 mRNA in CM-EP4 KO mouse heart and kidney.

The upper 2 panels are representative agarose gels of EP4 cDNA amplified by RT-PCR of RNA extracted from hearts and kidneys of “wild-type floxed” littermates and CM-EP4 KO mice. The lower panel is a graph of densitometric quantification of EP4 mRNA from the hearts and kidneys of “wild-type floxed” littermates and CM-EP4 KO mice (n=5-6).

Deletion of a receptor may result in a compensatory change in other receptors ^19,20. To investigate the effect of EP4 deletion on other EPs, we performed RT-PCR to amplify EP1, EP2 and EP3 mRNAs. The mRNA levels of EP1, EP2 and EP3 were similar in the hearts of both CM-EP4 KO and “wild-type” littermates, indicating that other EPs were not changed to compensate for the loss of EP₄ (Fig. 4).

Figure 4. RT-PCR of EP mRNAs.

RNA was extracted from mouse hearts. RT-PCR was performed. (A) Representative agarose gels of the RT-PCR products EP1, EP2, EP3 and GAPDH; (B) RT-PCR products were quantified by densitometry and EPs normalized to GAPDH as a control. The control values were normalized to 1 and the KO values compared to them (n=5-6).

We performed 2D-M-mode echocardiography of mice, and found no difference in those parameters that measure the pumping function of the heart (e.g., EF) and its shape (LVDs, LVDd) between CM-EP4 KO and “wild-type floxed” littermates (Table. 1). Thus, EP4 gene deletion does not change cardiac function in the absence of stress.

Table 1. Baseline Characteristics of CM-EP4 KO Mice.

Parameters Studied	Littermate Controls	CM-EP4 KO
MCSA (μm2)	165.00 ± 9.66	152.13 ± 4.18
ICF (%)	4.91 ± 0.12	4.80 ± 0.68
	n = 5	n = 5
EF (%)	80.1 ± 1.35	80.3 ± 1.99
LVDs (mm)	1.00 ± 0.03	1.10 ± 0.09
LVDd (mm)	2.90 ± 0.08	3.00 ± 0.11
	n = 6	n = 5

Echocardiography parameters were measured and compared between groups: ejection fraction (EF); left ventricular end-systolic dimension (LVDs); left ventricular end-diastolic dimension (LVDd). Hearts were harvested for quantification of myocyte cross-sectional area (MCSA) and interstitial collagen fraction (ICF) as described in Methods.

Finally, we used histological methods to evaluate the size of myocytes (MCSA) and degree of fibrosis (ICF) in hearts of CM-EP4 KO mice and “wild-type” littermates. As shown in Table 1, neither MCSA nor ICF was different between KO and littermate mice. Thus EP₄ deletion does not affect myocyte size and fibrosis in the unstressed mouse heart.

Effect of EP4 deletion on cardiac remodeling

We investigated the effect of EP4 deletion on infarct size, cardiac hypertrophy and fibrosis following MI. There was no difference in the infarct size and the thickness of the anterior wall scar (WTIA) following MI in CM-EP4 KO mice and littermate controls (Table 2). In contrast, MI increased MCSA significantly in controls (from 160 to 200 μm²) but hypertrophy was reduced 40% in CM-EP4 KO mice (from 200 to 184 μm²). In addition, ICF was attenuated 38% in CM-EP4 KO mice versus controls (Table 2).

Table 2. Histology and Echocardiography of CM-EP4 KO Mice after MI.

Parameters Studied	Littermate Controls		CM-EP4 KO
	Sham	MI	Sham	MI
IS (%)		31.4 ± 2.9		37.3 ± 4.3
WTIA (μm)		447.9 ± 26.9		446.4 ± 38.2
		n = 11		n = 7
MCSA (μm2)	160.6 ± 0.6	200.3 ± 4.1	160.0 ± 5.1	184.3 ± 2.3*
ICF (%)	5.4 ± 0.4	9.3 ± 0.3	5.0 ± 0.5	7.8 ± 0.2†
	n = 8	n = 9	n = 5	n = 6
EF (%)	81.42 ± 2.66	43.08 ± 3.55	80.06 ± 1.09	34.37 ± 3.66*
SF (%)	67.67 ± 1.26	48.48 ± 4.30	65.05 ± 0.94	31.58 ± 3.41†
LVDs (mm)	1.04 ± 0.06	2.05 ± 0.36	1.01 ± 0.04	2.98 ± 0.36*
LVDd (mm)	3.19 ± 0.07	3.77 ± 0.27	3.17 ± 0.11	4.22 ± 0.32
	n = 14	n = 12	n = 8	n = 10

Quantification of infarct size (IS) and thickness (WTIA), myocyte cross-sectional area (MCSA) and interstitial collagen fraction (ICF) were performed as described in the Methods. Echocardiography parameters were measured: ejection fraction (EF); shortening fraction (SF); left ventricular end-systolic dimension (LVDs); left ventricular end-diastolic dimension (LVDd). Statistical significance:

^*p ≤ 0.05 and

^†p < 0.005 for littermate control MI versus CM-EP4 KO MI.

Since hypertrophy and fibrosis were reduced in CM-EP4 KO mouse hearts following MI, we expected that cardiac function would be improved. In littermate controls, MI reduced cardiac function as indicated by decreases in EF and SF and increases in LVDs and LVDd. However, all functional parameters were worse and chamber dilatation tended to be more severe in CM-EP4 KO mice, indicating that EP4 deletion exacerbated MI-induced cardiac dysfunction and LV dilatation (Table 2).

Effect of EP4 deletion on BNP gene expression

Since we have previously shown that PGE₂, acting though EP4, stimulates BNP gene expression ²¹, we used RT-PCR to assess the effect of cardiac-specific EP4 deletion on this marker gene of hypertrophy. As shown in Fig. 5, MI induced BNP gene expression in littermate controls, but this effect was abrogated in KO mice.

Figure 5. RT-PCR of BNP.

RNA was extracted from mouse hearts. RT-PCR for BNP and GAPDH was performed (lower panel is a gel of PCR products), and the BNP/GAPDH ratio was determined after quantification by densitometry as shown in the graph (n=4-10).

Effect of EP4 deletion on activation of Stat-3

Previous studies in cardiac myocytes have shown that PGE₂ induces IL-6 expression ²² and activation of Stat-3 ⁶. In addition, IL-6 and Stat-3 are involved in hypertrophy and protection from ischemic injury ²³. To investigate a potential mechanism underlying the effect of cardiac myocyte-specific EP4 deletion on cardiac hypertrophy, we performed Western blot for phosphorylated Stat-3 (p-Stat-3). As shown in Fig. 6, MI increased Stat-3 phosphorylation 1.7-fold in littermate controls but not in CM-EP4 KO mouse hearts.

Figure 6. Western blot for Stat-3.

Two weeks after MI, hearts were removed, protein extracted and Western blot performed. The lower panel is a representative Western blot of p-Stat-3 and t-Stat-3, and the upper graph is the densitometric quantification of the p-Stat-3/t-Stat-3 ratio (n=12).

Discussion

PGE₂ is a pro-inflammatory prostanoid with proliferative and hypertrophic properties ²⁴. Recently we and others found that PGE₂ acting through EP4 is involved in cardiomyocyte hypertrophy in vitro ^5,6,14. The availability of an EP4 KO mouse would represent the best model to test whether EP4 plays a role in cardiac remodeling in vivo. Global EP4 gene KO mice have been reported, but 95% of homozygotes for the null allele have patent ductus arteriosus after birth and soon die ^25,26. Experiments have been done on these mice by separately breeding F2 generation global EP4 KO survivors and F2 wild-type mice ¹². However, this results in colonies of different genetic backgrounds and artificially selects for those EP4 KO mice that have adapted to survive this genetic deletion. Thus understanding the role of EP4 in the heart requires a conditional knock-out approach, as well as a breeding scheme that produces KO and littermate controls of the same overall background.

In this study, we have reported the generation and characterization of mice with reduced EP4 in the heart. An explanation for the less than 100% reduction of EP4 in the heart is that EP4 should still be present in fibroblasts, endothelial cells and smooth muscle cells. Also, previous studies using this technology have described some variability in the percentage of myocytes expressing Cre and thus not all myocytes are assumed to have excised the floxed EP4 gene ^18,27.

As there were no cardiac phenotypes, we used the CM-EP4 KO mice to study the effect of EP4 deletion on several parameters of MI-induced cardiac remodeling, including infarct size and scar thickness, hypertrophy, fibrosis and function. Cardiac remodeling subsequent to MI involves a number of different processes, including infarct expansion, pressure overload and volume overload ²⁸. Early remodeling occurs in the first 3 or so days after MI and involves the infarct and peri-infarct zones. Remodeling also occurs in the noninfarcted regions of the left ventricle and involves myocyte hypertrophy, accumulation of extracellular matrix and changes in the size and geometry of the entire left ventricle. The initial adaptive hypertrophic remodeling of the noninfarcted myocardium is hemodynamically beneficial to reduce wall stress. Late remodeling involves changes in the extracellular matrix and in the size and shape of the heart. When the changes in the left ventricle are no longer compensatory, left ventricular remodeling becomes maladaptive, leading to ventricular dilatation and overt heart failure ^29,30.

We found that EP4 signaling in cardiac myocytes had no effect on infarct size in our model. In contrast, Xiao et al ¹² reported that systemic EP4 KO mice had a larger infarct size than wild-type controls in an acute model of ischemia/reperfusion (24 hrs of injury), suggesting that EP4 is cardioprotective. It is possible that EP4 has a different function in the very early events resulting from ischemia/reperfusion injury than in the late remodeling that we are studying in our permanent LAD ligation model. Another possibility is that in the global EP4 KO mouse subjected to ischemia/reperfusion injury, the absence of EP4 in other cells, such as inflammatory cells, fibroblasts, vascular smooth muscle cells and endothelial cells in the heart and other tissues, may have an influence on the extent of ischemia/reperfusion injury. Finally, there is also the possibility that differences in the genetic backgrounds of the global EP4 KO mice and the controls could influence infarct size differences.

Our data suggest that EP4 signaling is involved in part in cardiac hypertrophy in vivo. The results are supported by publications showing that PGE₂ acting through EP4 increased cardiomyocyte hypertrophy and the hypertrophic marker gene BNP in myocytes in vitro ^6,14,21. Our data are also consistent with our previous publication showing that COX-2 inhibition reduced MCSA after MI ⁷. However the studies with the COX-2 inhibitor could not identify which COX-2 products and prostanoid receptors were involved in the effect. Thus our present studies with the CM-EP4 KO mouse point to EP4 signaling in cardiac myocytes as part of the hypertrophic growth program.

Regarding fibrosis post-MI, the increase in ICF was less in the KO mouse hearts than in the littermate controls, implying that cardiac myocyte EP4 is somehow involved in fibrosis. Given that PGE₂ levels increase in the heart post-MI ⁷, the PGE₂ may also act on myocytes in an EP4-dependent manner to stimulate the synthesis and secretion of growth factors. For example, PGE₂ has been shown to stimulate IL-6 expression in neonatal cardiac myocytes ²², and IL-6 stimulates fibroblast proliferation ³¹. Thus it will be of interest to use our CM-EP4 KO mice to study the role of myocyte EP4 in fibroblast proliferation.

After MI, cardiac function was worsened in CM-EP4 KO mice compared with littermate controls. This reduction in function occurred despite the reductions in fibrosis and hypertrophy. This result is in contrast with our previous study showing that COX-2 inhibition improved cardiac function in C57BL6/J mice after MI ⁷. There may be several reasons for these disparate results. The first is that the species of mice is different in the two studies. A second potential reason is COX-2 inhibition reduces all prostanoid generation, including reductions in other deleterious prostanoids such as PGF_2α and thromboxane A₂, but these are likely still present in the CM-EP4 KO mouse heart. Thirdly, the experimental design of the studies is slightly different. In the COX-2 inhibition study, the COX-2 inhibitor was administered either 2 days or 2 weeks post MI, whereas in the present study, EP4 was deleted in the mouse hearts prior to inducing MI. This timing issue could be important as cardiac remodeling has both an early and a late component, and early inflammatory responses are necessary for wound healing and infarct repair. Finally, the COX-2 inhibitor, but not cardiac-specific EP4 deletion, would have systemic anti-inflammatory effects. In a study in rats, it was shown that treatment with a COX-2 inhibitor during MI reduced fibroblast proliferation, macrophage infiltration and inflammation in the heart ³². Thus we hypothesize that the reduction in hypertrophy and fibrosis may not have a beneficial effect if the overall level of inflammation in the heart is not appropriately regulated. This is an area for future studies.

Other studies have described reductions in fibrosis with further impairment of systolic function. For example, Cingolani et al ³³ showed that the small peptide Ac-SDKP infused into SHR rats reduced LV collagen content but not hypertrophy and further reduced systolic performance. The authors suggested that the reduction in systolic function might be caused by the lack of a proper extracellular matrix support (either collagen type or cross-linked collagen) for the hypertrophied myocytes. Additional studies are required to study the role of EP4 in systolic function.

There are many signaling mechanisms involved in cardiac hypertrophy. Activation of the transcription factor Stat-3 and the p42/44 MAPK and phosphoinositide 3-kinase (PI3Kα) signaling pathways are involved in compensatory hypertrophy and appear protective ^6,23,34-38. PI3K, which lies downstream of many receptor tyrosine kinases, activates the kinase Akt, which in turn activates Stat-3 ^37,36. Activated Stat-3 promotes cardiomyocyte survival by upregulating antiapoptotic and cytoprotective proteins. Overexpressed EP4 has been shown to activate PI3K/Akt ³⁹. Also in vitro, PGE₂ acting via EP4 induces cardiomyocyte hypertrophy through activation of Stat-3 and p42/44 MAPK ^6,14. Normally Stat-3 is activated by IL-6 and other cytokines binding to receptors (e.g., gp130) which are coupled to a kinase called JAK. The effect of PGE₂ to activate Stat-3 in myocytes may occur indirectly through stimulation of synthesis and secretion of IL-6 or result from the effects of other signaling kinases, such as p42/44 MAPK or PI3K/Akt. Cardiovascular effects of knockout of components of the Stat-3 signaling pathway have been studied. Cardiomyocyte-restricted KO of Stat-3 increased cardiac dysfunction, sensitivity to inflammation and fibrosis in aged mice ¹⁶. Similarly, cardiomyocyte-specific deletion of the gp130 receptor prevented hypertrophy in response to thoracic aortic constriction. These gp130 knockout mice showed a rapid onset of cardiac dysfunction with dilated cardiomyopathy and massive myocyte apoptosis ⁴⁰. Thus, results from our present study showing that CM-EP4 KO mice are defective in their ability to activate Stat-3, have reduced cardiac hypertrophy and a worsening of systolic function after MI are mostly consistent with the aforementioned studies on KO of gp130 and Stat-3, and suggest a role for both EP4 and Stat-3 in a growth/survival pathway to protect the heart from pathophysiological stimuli. Additional studies will be required to investigate the role of myocyte EP4 in the regulation of IL-6, inflammatory mediators, signaling kinases and apoptosis in the mouse heart after MI.

Perspectives

In summary, we are the first to have generated CM-EP4 KO mice in which the systemic complications of EP4 deletion are obviated. This is also the first report on cardiac myocyte EP4's effect on cardiac hypertrophy, fibrosis and function. Following MI, EP4 is involved in cardiac remodeling, activation of Stat-3 and cardiac function. Moreover, we suggest that EP4 is involved in compensatory hypertrophy, perhaps coupled to Stat-3 activation, and when this is disrupted, cardiac function is impaired. Therefore, an EP4 agonist may have a therapeutic benefit to enhance compensatory hypertrophy and improve function in patients with MI.