Spatiotemporal induction of matrix metalloproteinase-9 transcription after discrete myocardial injury

Abstract

Radiofrequency (RF) ablation of the myocardium causes discrete sites of injury. RF scars can expand, altering the extracellular matrix (ECM) structure and the continuity of the electrical syncytium of the adjacent myocardium. Matrix metalloproteinases (MMPs), such as MMP-9, contribute to ECM remodeling. However, whether and to what degree transcriptional induction of MMP-9 occurs after myocardial RF injury and the association with electrical conduction patterns after RF injury remains unexplored. This study examined MMP-9 gene promoter (M9PROM) activation after myocardial RF injury using mice in which the M9PROM was fused to a β-galactosidase (β-gal) reporter. RF lesions (0.5-mm probe, 80°C, 30 s) were created on the left ventricular (LV) epicardium of M9PROM mice (n=62) and terminally studied at 1 h, 1 d, 3 d, 7 d, 14 d, and 28 d after RF injury. M9PROM activation was localized through β-gal staining. The RF scar area and the area of β-gal staining were measured and normalized to LV area (planimetry). RF scar size increased from 1 h post-RF-injury values by 7 d and remained higher at 28 d. M9PROM activation became evident at 3 d and peaked at 7 d. Electrical conduction was measured (potentiometric dye mapping) at 7 d after RF injury. Heterogeneities in action potentials and electrical impulse propagation coincident with M9PROM activation were observed after RF injury. For example, conduction proximal to the RF site was slower than that in the remote myocardium (0.15±0.02 vs. 0.83±0.08 mm/ms, P<0.05). Thus, a unique spatiotemporal pattern of MMP-9 transcriptional activation occurred after discrete myocardial injury, which was associated with the development of electrical heterogeneity. Therefore, these findings suggest that changes in a key determinant of extracellular matrix remodeling, in addition to changes in myocardial structure, can contribute to arrhythmogenesis around the region of myocardial injury.—Mukherjee, R., Colbath, G. P., Justus, C. D., Bruce, J. A., Allen, C. M., Hewett, K. W., Saul, J. P., Gourdie, R. G., Spinale, F. G. Spatiotemporal induction of matrix metalloproteinase-9 transcription following discrete myocardial injury.

A common sequel to myocardial injury either due to myocardial infarction (MI), ischemia-reperfusion, or discrete thermal application of heat or cold is changes in myocardial ultrastructure, as well as in left ventricular (LV) chamber geometry (1–7). Despite the differing etiology of the injury to the myocardium, initially an adaptive remodeling of the myocardium can occur, leading to the rearrangement and/or derangement of the underlying extracellular matrix (ECM) as well as the electrical syncytium at and around the site of injury (5, 8, 9). The matrix metalloproteinases (MMPs) are proteolytic enzymes implicated in ECM remodeling (1, 8, 10), and past studies using pharmacological MMP inhibition or genetic manipulation of certain MMP types have established a cause-effect relationship between postinjury myocardial remodeling and MMP activation (6, 7, 11–13). Increased MMP activation/expression within the myocardium can occur early (within hours of the injury) and persist for longer durations in the postinjury period (13–16). Induction of one MMP type, MMP-9, has been uniformly demonstrated to occur after myocardial injury (1, 11, 16, 17). There are a number of substrates identified for MMP-9-mediated proteolysis, including denatured collagens (gelatins) and basement membrane components (10). One consequence of ECM proteolysis within the myocardium is the degradation of supporting scaffolds between myocytes, resulting in myocyte slippage and, subsequently, disruption of the intermyocyte electrical continuity (18, 19). Therefore, MMP-9 induction after myocardial injury may lead to structural as well electrical abnormalities at and around the site of injury. Because most tissue types do not constitutively elaborate MMP-9, increases in MMP-9 levels are generally considered to be due to increased MMP-9 gene transcription, putatively resulting from activation of the MMP-9 gene promoter (20, 21). Accordingly, the goals of this study were to quantitate the temporal pattern of MMP-9 gene promoter activation after focal myocardial injury and relate the spatial expression pattern of MMP-9 promoter activation to localized abnormalities in electrical conduction.

MATERIALS AND METHODS

This study was designed to examine the spatial and temporal induction pattern of the MMP-9 gene promoter after discrete myocardial injury. In this study, transgenic mice in which a fusion gene product of the MMP-9 promoter and the β-galactosidase gene (lacz) was inserted into the genome were used (20, 21). All animals were treated and cared for in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (22), and the protocol was approved by the Institutional Animal Care and Use Committee.

MMP-9 promoter reporter mice

Mice with the MMP-9 reporter-lacz construct in the CD-1 background strain were developed and described by Mohan et al. (20) As internal controls, transgene expression was ascertained by PCR of tail clip DNA and β-galactosidase elaboration (X-gal reacted) at the cut edges of the excised tail tips.

Model of discrete myocardial injury

This study used a model of myocardial injury, in which lesions were caused through the application of radiofrequency (RF) energy, as described previously (2, 3). Equal numbers of male and female mice (n=73; age 15±1 wk; mass 32±1 g) that were positive expressers of the MMP-9 promoter-lacz transgene were used. The mice were initially anesthetized by exposure to isoflurane vapors, and the neck, chest, and back (2–3 cm², for RF current continuity) were shaved. The mice were laid supine over a stainless steel grounding plate, which was previously coated with an electroconductive gel, intubated orotracheally, and connected to a ventilator (225–250 breaths/min, tidal volume 250 μl; Hugo-Sachs, March-Hugstetten, Germany). Anesthesia was maintained by delivering 2% isoflurane in oxygen. Three needle electrodes inserted into the subcutaneous space of the upper and lower extremities were used to monitor the surface electrocardiogram in Einthoven's lead I configuration (PONEMAH; LDS Gould, Valley View, OH, USA). A 1.0-cm left lateral thoracic incision was made over the third intercostal space, and a pericardectomy was performed to provide access to the LV free wall. RF current (RFG-3C; Radionics, Boulder, CO, USA) was delivered to the myocardium through a custom designed probe (0.5 mm diameter; Omega Engineering, Stamford, CT, USA) using negative feedback temperature control to achieve a temperature of 80°C for 30 s (2, 3). RF lesion formation was confirmed visually as a discrete blanched region on the LV epicardium and electrically as an elevation of the ST segment of the electrocardiogram. The ribs were apposed with 5-0 silk sutures, air and fluid were evacuated from the thorax, and the overlying muscle layers and the skin incision were closed. The animals were rotated to the prone position, administered an analgesic (buprenorphine, 0.05 mg/kg s.c.), weaned from the ventilator, extubated, and recovered. A group of sham-operated non-RF-injured mice (n=5) served as referent controls for analysis of MMP-9 promoter activation.

Time course

Terminal studies on the mice were performed at 1 h (acute), 1 d, 3 d, 7 d, 14 d, and 28 d after injury. These time points were selected to examine the spatiotemporal expression patterns for the MMP-9 promoter during the early (acute and 1 d), mid (3 and 7 d), and late (14 and 28 d) post-RF periods. At the assigned postinjury time point, the animals were anesthetized by exposure to isoflurane inhalation and weighed, and a recording of the electrocardiogram (in lead I configuration) was obtained. The thoracic cavity was opened, the heart was arrested by injecting CdCl₂ (0.1 M in saline, 0.1 ml) through the LV apex, quickly extirpated, weighed, and fixed in 10% formalin. One tibia was isolated and immersed in a saturated KOH solution, and tibial length was measured on the following day. To normalize for differences in body size, heart mass was normalized to tibial length.

β-Galactosidase assay and quantification

MMP-9 gene promoter activation was assayed as a function of β-galactosidase elaboration by incubating the hearts overnight in a substrate buffer (X-gal; Sigma-Aldrich, St. Louis, MO, USA) as described previously (20). In brief, X-gal (5-bromo-4-chloro-3-indol; Sigma-Aldrich) was prepared as a 40 mg/ml stock solution in dimethyl sulfoxide (Sigma-Aldrich) and freshly diluted to a 2% working solution in a buffer containing 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 2 mM magnesium chloride. The hearts were fixed in 10% formalin for 30 min, washed 3 times for 10 min each in PBS, and stained by overnight incubation at 37°C in the 2% X-gal solution. The hearts were washed in three changes of PBS and then fixed in 10% formalin for 3 d followed by storage in 70% ethanol. To quantitate the area and intensity of positive β-galactosidase staining, the epicardial and endocardial surfaces of the left ventricles were photographed (Sony F707; Sony, Tokyo, Japan) using isoplanar illumination against a white background. The photographs were input into an image analysis program (SigmaScan Pro 4.0; SPSS, Inc., Chicago, IL, USA) to determine the area and intensity of β-galactosidase staining and the epicardial area of the left ventricle by digital planimetry. The area of β-galactosidase staining was normalized to total LV epicardial and endocardial area and is expressed as a percentage.

Optical mapping of electrical activation

Optical mapping studies were performed in a subset of the mice (n=11) after 7 d of RF injury (23). A separate cohort of non-RF-treated mice (n=7) was used to provide referent control values for measurements of electrical activation of the left ventricle. The mice were anesthetized with 5% isoflurane, and the heart was removed to allow for cannulation of the aorta, after which the heart was perfused with oxygenated Tyrode's solution at 37°C at a constant pressure of 100 mm Hg. The heart was placed in a thermostatically controlled tissue chamber (37±1°C) on the stage of a dissecting microscope (Leica MZ FLIII). The potentiometric dye, di-4-ANEPPS (12 μM; Molecular Probes, Eugene, OR, USA), was perfused for 5 min. Cytochalasin D (20 μM) was added to the perfusate to uncouple mechanical contraction (24). Hearts were paced using bipolar electrodes from the apex at a rate of 300 bpm. The stimulation pulses were 2 ms in duration, and the voltage was adjusted to 2× threshold. Changes in membrane voltage (action potentials) were recorded as chromatographic changes in di-4-ANNEPPS excitation using a high-speed 80 × 80 pixel charge-coupled device camera (1 kHz; Red Shirt Imaging, Fairfield, CT, USA). Action potential upstrokes were marked using Cardioplex software (Red Shirt Imaging) (25). The peak of the first derivative of the action potential upstroke was used to mark activation time. Conduction velocities were derived from the gradients of activation times (23, 25). Parameters determined from the action potentials included upstroke velocity and action potential amplitude.

Histochemical staining

The left ventricles previously stained for β-galactosidase were embedded in paraffin, and thin sections (4 μm) were obtained. The sections were placed on slides, deparaffinized, and stained with either hematoxylin and eosin (H&E) or used for immunohistochemical analysis. The H&E-stained sections were used to determine in situ LV wall thicknesses at the injured and remote regions by digital planimetry of photomicrographs [1× objective; Zeiss Axioskop (Carl Zeiss, Oberkochen, Germany)and SigmaScan Pro 4.0]. Immunohistochemical staining for MMP-9 (AB19047, 1:100; Chemicon International, Temecula, CA, USA) was performed to determine protein and promoter reporter colocalization. In brief, myocardial sections that contained the scar were blocked and then incubated for 60 min at room temperature with primary antibodies. After incubation with secondary antisera, detection was performed by visualization of a 3′,3′-diaminobenzidine-hydrogen peroxide substrate (Vector Laboratories, Burlingame, CA, USA). For immunohistochemical staining for connexin43 (Cx43), the sections were incubated with antigen retrieval solution, an anti-Cx43 antibody (1:2000; Sigma-Aldrich), and an Alexa Fluor 488-conjugated secondary antibody. Qualitative assessment was performed on 5 random fields from the injured and remote regions of each section. Negative controls included substitution with nonimmune primary antisera.

In vitro cleavage of Cx43 by MMP-9

To determine whether MMP-9 could proteolyze Cx43, increasing amounts of active recombinant MMP-9 (0–20 μg/ml, PF024; Oncogene Research Products, San Diego, CA, USA) was added to 5 μg of Cx43 recombinant protein fragment (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and assayed in triplicate. After 3 h of incubation at 37°C, the samples were loaded on a 10–20% polyacrylamide gel (Bio-Rad Laboratories, Hercules, CA, USA), electrophoretically separated, and immunoblotted for Cx43 (Sigma-Aldrich).

Full-length Cx43 protein was expressed in HeLa cells, a cell line that is constitutively connexin-deficient, using adenoviral transfection as described previously (26). The cells were grown in plastic flasks to confluence (2.3×10⁶ cells/ml), trypsinized, and transferred to reaction chambers to which either 0.41 μg/ml recombinant active MMP-9 (in DMEM, n=4, PF024; Oncogene Research Products) or DMEM (vehicle, n=5) was added. After 3 h of incubation at 37°C, the cells were centrifuged, resuspended in 1× PBS, and homogenized (0.5 ml of Sigma Reagent 3; Sigma-Aldrich). The protein concentration was determined using a Bradford assay. Cx43 levels were determined by immunoblotting and quantitated by normalizing the integrated optical density to the protein concentration of the cell lysate.

Data analysis

Temporal changes in heart mass indices and β-galactosidase staining were compared using ANOVA. Conduction velocity and action potential parameters were compared between the post-RF injury and control groups using 2-way ANOVA. Pairwise separation was performed by a Bonferroni adjusted t test. Results are presented as mean ± se. P < 0.05 was considered to be statistically significant.

RESULTS

LV morphometrics

All mice in which discrete myocardial injury was induced through the application of RF current survived to the assigned post-RF current time point. The final sample size for each post-RF current time point is provided in Table 1. Heart mass indexed to tibial length was increased from acute values by d 3 after RF injury and was higher than acute and 3-d post-RF-injury values at d 7, 14, and 28 post-RF injury (Table 1). At d 28 after RF injury, heart mass and indexed heart mass were higher than d 7 and 14 post-RF-injury values. LV wall thickness at and adjacent to the site of RF injury was lower than acute values at d 14 and 28 after RF injury. Conversely, LV wall thickness diametrically opposite the area of RF injury was increased from acute values by 3 d after RF injury and further increased at d 14 and 28 after RF injury. The area of the epicardial and endocardial surfaces of the LV was larger than acute values by d 1 after RF injury and remained increased at all subsequent post-RF current time points. The area of the RF scar was increased from acute values by 3 d post-RF injury and further increased from 1- and 3-d post-RF-injury values at longer post-RF-injury durations.

Table 1.

Time-dependent changes in heart mass indices and intensity of positive β-galactosidase staining in MMP-2 and MMP-9 reporter mice after RF injury

Parameter	Day post-RF injury
	0; acute	1	3	7	14	28
Heart mass (mg)	147 ± 3	155 ± 7	160 ± 5#	169 ± 3#	168 ± 7#	190 ± 12#,+,a,b
Heart mass/body mass (mg/g)	4.3 ± 0.1	5.2 ± 0.2#	5.6 ± 0.2#	5.8 ± 0.2#	5.8 ± 0.3#	5.8 ± 0.2#,+
Heart mass/tibial length (mg/mm)	8.2 ± 0.2	8.7 ± 0.3	9.2 ± 0.2#	9.4 ± 0.1#,+	9.5 ± 0.4#,+	10.5 ± 0.6#,+,a,b
Wall thickness adjacent to injury (mm)	0.90 ± 0.05	0.88 ± 0.02	0.86 ± 0.03	0.85 ± 0.02	0.75 ± 0.03#,+,a,b	0.77 ± 0.03#,+,a,b
Wall thickness remote to injury (mm)	0.90 ± 0.01	0.89 ± 0.01	0.99 ± 0.02#,+	0.96 ± 0.02#,+	1.11 ± 0.02#,+,a	1.18 ± 0.03#,+,a,b
LV area (mm2)	69.2 ± 2.5	80.6 ± 4.2#	79.1 ± 3.9#	79.9 ± 3.3#	79.9 ± 3.7#	79.8 ± 4.0#
Area of injury/scar area (mm2)	2.29 ± 0.25	2.79 ± 0.46	3.57 ± 0.52#	5.31 ± 0.67#,+,a	5.73 ± 0.78#,+,a	6.72 ± 0.60#,+,a
Area of β-galactosidase staining (mm2)	0.09 ± 0.04	0.22 ± 0.09	3.21 ± 1.58#,+	9.29 ± 2.41#,+,a	4.65 ± 1.58#,+,b	0.19 ± 0.11a,b,c
Sample size (n)	11	9	10	10	11	11

Data are means ± se.

^#P < 0.05 vs. d 0 (acute);

⁺P < 0.05 vs. d 1;

^aP < 0.05 vs. d 3;

^bP < 0.05 vs. d 7;

^cP < 0.05 vs. d 14.

MMP-9 promoter activation

Representative photographs of whole hearts showing the epicardial and endocardial surface at each post-RF current time point are depicted in Fig. 1. MMP-9 promoter activity, evidenced as β-galactosidase (blue-green) staining, was localized to the region of RF injury and at the borders of the injured area in 18% of left ventricles at 1 h after RF injury and in 84, 68, 100, 82, and 64% of left ventricles at d 1, 3, 7, 14, and 28 post-RF injury, respectively. The proportion of hearts in which β-galactosidase staining was observed was higher at d 1, 7, 14, and 28 post-RF injury than at the acute post-RF current time point (P<0.05, Fisher's exact test). For analysis of β-galactosidase staining, the left ventricles in which no β-galactosidase staining was grossly evident were assigned a value of 0 for the area. In all the other left ventricles in which β-galactosidase staining was observed, epicardial and endocardial staining areas were planimetered and expressed as a total. The results for β-galactosidase staining area are summarized in Table 1. The RF scar area and area of β-galactosidase staining were normalized to the LV surface area (Fig. 2). MMP-9 promoter activity was increased from acute and 1-d post-RF-injury values by 3 d and peaked at 7 d after RF injury (Fig. 2). Representative low- and high-power photomicrographs that show colocalization of MMP-9 protein and β-galactosidase elaboration are provided in the left panels of Fig. 3. H&E staining (right panels of Fig. 3) clearly showed disruption of the ECM at the border as well as within the injured region.

Figure 1.

Photographs from the epicardial surface (top) and endocardial surface (bottom) of hearts from MMP-9 gene promoter reporter mice showing regions of positive β-galactosidase staining (dark regions on lighter myocardium) at indicated time points after RF current-induced injury. Positive β-galactosidase staining was observed at 3 d post-RF injury and peaked at 7 d after MI. Scale grid at left represents a square with 2-mm sides.

Figure 2.

Time course for changes in RF scar/injury area and β-galactosidase (β-Gal) staining after RF injury. RF scar/injury area was increased over acute and 1 d post-RF-injury values by 7 d after RF injury and remained higher at subsequent time points. The area of β-galactosidase staining peaked at 7 d after RF injury and declined to near nondetectable levels by 28 d after RF injury. Sample sizes at each post-RF current time point are presented in Table 1. ^#P < 0.05 vs. acute (1 h after RF injury); ⁺P < 0.05 vs. 1 d after RF injury; ^aP < 0.05 vs. 3 d after RF injury; ^bP < 0.05 vs. 7 d after RF injury; ^cP < 0.05 vs. 14 d after RF injury.

Figure 3.

Left panels: representative photomicrograph of the RF-injured and border regions from hearts of MMP-9 promoter mice at 7 d after RF current stained using antibodies against MMP-9. Inset: higher-power micrograph from the highlighted area of the scar border. Positive staining for MMP-9 protein (brown) was localized at and around cells that elaborated β-galactosidase (blue). Right panels: Representative H&E-stained serial sections from the same mouse heart showing increased cellularity at the border region. Inset: higher-power photomicrograph of the region shows disruption of the ECM within the region of myocardial injury. Scale bars = 100 μm (lower power); 20 μm (higher power).

Electrical activation/conduction

Because one of the goals of this study was to examine the role of MMP-9 activation in electrical conduction abnormalities, myocardial electrical conduction patterns at and around the site of injury were determined only at 7 d after RF injury, as MMP-9 promoter activation was observed to have achieved a peak at this post-RF time point. Using the optical mapping system, an array of action potentials were recorded from control and post-RF injury left ventricles, and representative images are shown in the left panel of Fig. 4. In control left ventricles, consistent and homogeneous action potentials were recorded over the entire epicardial surface. In contrast, the morphology of action potentials recorded from all left ventricles at and adjacent to the region of RF injury was heterogeneous (Fig. 4). The amplitude of the action potential and the rate of upstroke depolarization at the region of the RF scar (Fig. 4, right panels) were both reduced, not only from control values but also from those recorded in the remote regions of the post-RF injury left ventricles. By marking activation times of the action potentials, high-resolution optical maps of myocardial depolarization were constructed to determine electrical conduction patterns and to generate isochrones of wavefront propagation at the remote region and at and around the site of injury (Fig. 5). In the control hearts, the depolarization wavefront progressed rapidly and in a homogeneous manner. In all of the injured hearts, propagation of the depolarization wavefront was heterogeneous with respect to differing conduction velocities around the site of injury and the remote, viable myocardium. Specifically, the site of the RF injury was devoid of any changes in electrical activation, and progression of the depolarization wavefront was slower at the borders of the RF injury than in the adjacent myocardium. The results of the conduction velocity measurements are summarized in the middle panel of Fig. 5. In 6 of the 11 post-RF preparations for which electrical conduction patterns were examined, a localized electrical reentry circuit was present around the RF lesion site (Fig. 5, bottom). This pattern of electrical reentry was observed in the hearts of 2 of 5 female mice and in 4 of 6 male mice, with no difference with respect to gender (P=0.38, Pearson χ²). The absence of electrical conduction at the site of RF injury and the slower conduction velocity at the border of the injury site were associated with an absence or derangement of intermyocyte electrical connectivity. Specifically, the site of RF injury was devoid of myocytes and/or the gap junction protein, Cx43 (Fig. 6). At the borders of the RF injury site, islets of myocytes were observed, but there was a clear lack of normal end-to-end Cx43 staining (Fig. 6).

Figure 4.

Action potentials on the LV epicardium were recorded in a 40 × 40 array using a potentiometric dye. Action potentials from the area of the array encompassing the scar area (contained within an 8×8 array in all hearts) were analyzed to determine amplitude and the rate of depolarization. Action potential morphology in the scar region was abnormal, with evidence of biphasic patterns, diminution of amplitude, and action potential prolongation. Summary data for action potential amplitude and rate of depolarization were reduced at the scar region but were similar to control values in myocardium remote from the scar. *P < 0.05 vs. control; ⁺P < 0.05 vs. remote region.

Figure 5.

Top left panel: representative isochronal plots for the propagation of depolarization wavefronts in the left ventricle from a sham-operated control heart and at 7 d after RF injury. Conduction of depolarization from the site of pacing (asterisk) near the LV apex to the base was complete within 7 ms for the control left ventricle. In the post-RF-treated heart, however, a distinct region of conduction block was observed with slower conduction of the wavefront distal to the block. Top right panel: summary of conduction velocity, computed as the time required for propagation of the depolarization wavefront. *P < 0.05 vs. control; ⁺P < 0.05 vs. remote region. Bottom panel: temporal series of images spaced 10 ms apart. Before LV depolarization (0 ms), most of the left ventricle was at resting membrane potential (−80 mV), but the scar region was partially depolarized (demarcated by ellipse and A). At 10 ms, when the LV free wall was completely depolarized, a portion of the scar region (B) remained partially depolarized. However, during repolarization of the left ventricle (20–40 ms), a finite region around the scar demonstrated delayed depolarization (A), and this region of depolarization propagated in counterclockwise fashion at and around the region of the scar. Similar patterns of delayed depolarization and regional propagation were observed in 6 of the 11 post-RF-treated hearts.

Figure 6.

Representative confocal photomicrograph for Cx43 (white; arrows), actin (asterisks), and nuclei (arrowheads) at the scar and border regions at 7 d after RF injury. Although there were cells present within the scar region (evidenced by nuclear staining), there were no myocytes within the scar. At the border there were islets of myocytes with sparse distribution of Cx43. Compared with the remote myocardium (inset), Cx43 distribution was abnormal in that positive Cx43 staining occurred at the lateral edges of the cells. Scale bar = 40 μm.

In vitro Cx43 proteolysis

Incubation of a recombinant Cx43 C-terminal fragment with increasing concentrations of recombinant active MMP-9 did not change the immunodetectable amount of Cx43 at any concentration (Fig. 7, top). Coincubation of HeLa cells that were transfected to express the full-length Cx43 protein with recombinant active MMP-9 yielded similar results with respect to there being no discernible effect of Cx43 proteolysis mediated by MMP-9 (Fig. 7, bottom).

Figure 7.

Top panel: coincubation of recombinant Cx43 with increasing concentrations of recombinant, active MMP-9 did not cause a reduction in Cx43 immunoblotting. Middle panel: summary of immunoblot data. Bottom panel: Cx43 levels in HeLa cells transfected to express full length Cx43. There was no difference in Cx43 levels between HeLa cell cultures incubated with (n=4) or without (n=3) recombinant MMP-9. IOD, immunodetectable.

DISCUSSION

To determine tissue specificity or conditions that cause transcription of a gene of interest, past studies have used transgenic mice in which the promoter sequence of that particular gene has been fused to reporter cassettes (20, 21, 27). Using a similar transgenic approach, in the present study we examined the time-dependent activation of the MMP-9 gene promoter after discrete myocardial injury created through the application of RF current (2, 3). In addition, the present study examined whether the consequence of matrix disruption at and around the region of the scar altered electrical conduction properties and whether these changes were related to induction of the MMP-9 gene promoter. The unique findings of this study were 3-fold: first, MMP-9 promoter induction occurred at and around the site of injury and was primarily localized to the border of the injured region; second, there was a distinct temporal pattern of MMP-9 promoter induction, being increased over acute values at 3 d after RF injury, achieving a peak at 7 d after RF injury, and then declining to nondetectable levels at 28 d after RF injury; and third, a disruption in the localization of the gap junction protein, Cx43, was observed at the region of MMP-9 promoter activation, which was associated with the formation of conduction blocks and a slowing of the propagation of the depolarization wavefront around the site of injury. These findings demonstrated that a unique spatiotemporal pattern of MMP-9 transcriptional activation occurred after discrete myocardial injury, which was associated with the development of electrical heterogeneity. Therefore, these findings suggest that changes in a key determinant of ECM remodeling, in addition to changes in myocardial structure, can contribute to arrhythmogenesis around the region of myocardial injury.

The MMPs constitute a family of closely related proteases that are key contributors to tissue remodeling in physiological as well as pathological conditions.(1–3, 5–8, 10, 14, 16, 20) Although a number of MMPs have been identified to date, changes in MMP-9 expression and protein levels have been uniformly reported to occur concomitant to changes in LV ultrastructure and geometry after MI or discrete injury (2, 3, 5–8, 10, 14, 16). Past studies have demonstrated that there is an early increase in myocardial MMP-9 levels after MI (1). This initial increase in MMP-9 levels after MI is due to the release of preformed MMP-9 from infiltrating inflammatory cell types, such as macrophages and neutrophils (1). It is generally considered that resident cells under ambient conditions typically do not synthesize and/or release MMP-9 (20). Therefore, “on-demand” MMP-9 gene transcription after the activation of the gene promoter represents an important regulatory step for changes in MMP-9 levels (1, 20). To determine in vivo changes in MMP-9 gene promoter activation, this study used a transgenic reporter construct in which the promoter region of the MMP-9 gene was fused to the bacterial lacz cassette (1, 20, 21). Using this unique transgenic construct, a past study demonstrated that there was a unique spatial and temporal pattern to MMP-9 gene promoter activation after MI. Specifically, MMP-9 gene promoter activation was localized at the border region of the MI and achieved peak values at 7 d after MI (1). The present study builds on these past findings by demonstrating that a similar spatial and temporal pattern of MMP-9 gene promoter activation occurred at and around the site of a discrete myocardial injury (as opposed to widespread necrosis concomitant with MI). In light of the findings of past in vitro and in vivo studies reporting that MMP-9 promoter activation paralleled expression of the endogenous MMP-9 gene (19), the temporal correlation between MMP-9 promoter activation and expansion of the area of injury that was observed in the present study suggests that increased MMP-9 transcription contributed, at least in part, to remodeling of the myocardium after injury. Moreover, the role of other MMPs, such as that of MMP-8, which is also released in the early period of post-MI myocardial remodeling, needs to be elaborated.

Disruptions in electrical conduction through the myocardium can occur when a portion of the myocardium is replaced by nonconductive fibrous tissue (18, 19). For example, fibrosis in the atrial myocardium is associated with multiple electrical propagation pathways and is considered to form a substrate for the development and progression of atrial fibrillation (28–30). Therefore, the formation of a scar during the wound healing response after myocardial injury can represent an anatomical block to conduction. An important consideration with respect to slower conduction and later depolarization of myocytes distal to the site of myocardial fibrosis is the potential for setting up electrical reentry (18, 19). Specifically, late depolarization of myocytes distal to the anatomical block can then conduct the depolarization wavefront to myocytes proximal to the area of block which could be past the refractory period and ready to receive another depolarization stimulus. In the present study, localized electrical reentry was detected around the site of injury in half of the study cohort and was localized to the region of MMP-9 promoter activation. Given that both male and female mice were used in the present study and that past findings have implicated an accentuated post-MI remodeling response in male mice (31, 32), additional analyses was performed to determine whether gender of the mice contributed to the development of electrical reentry around the site of the myocardial injury. The results from this set of analyses revealed that electrical reentry around the site of the myocardial injury occurred in an equal proportion of hearts from male and female mice. These findings suggest that MMP-mediated matrix modification, more than a gender-specific remodeling response, may play an important role in the formation of an arrhythmogenic substrate.

Propagation of electrical wavefronts through the LV myocardium occurs preferentially through intermyocyte connections, the gap junctions, that are located at the intercalated disks (25, 26, 33, 34). Indeed, disruption in the amount of gap junctions and/or Cx43 can lead to an increased susceptibility to arrhythmias (35, 36). Therefore, the myocardium, under normal conditions, acts as an electrically continuous syncytium (33, 34), and electrical impulses are usually transmitted along the long axis of myocardial fibers because of the lower resistivity to transmission along that direction. An important constituent of the gap junctions are the connexins, and the predominant connexin of the mammalian heart is Cx43 (25, 26, 34). Intercellular uncoupling either due to a loss of gap junctions and/or loss of myocyte alignment can contribute to the slowing of electrical conduction (34). An association between decreased Cx43 levels and reduction in conduction velocity has been previously demonstrated using mice with cardiac-restricted Cx43 deletion (36–41). In the present study, a loss of Cx43 and disruption of Cx43 localization was observed at the region of MMP-9 promoter activation around the area of the injury. To determine whether the loss of Cx43 occurred because of MMP-9-mediated proteolysis, recombinant active MMP-9 was coincubated with a recombinant C-terminal fragment of Cx43 or with a cell line made to overexpress full-length Cx43 (26). The results from this set of studies demonstrated that there was no direct cleavage of Cx43 by MMP-9. This finding suggests that MMP-9 may not directly cause Cx43 proteolysis; therefore, changes in MMP-9 levels in the border region of the injured region may not directly contribute to abnormalities in myocardial electrical propagation. However, other MMP subtypes probably contribute to ECM remodeling, and Cx43 may be a substrate for one of these other MMPs, such as MMP-7 (23). Nevertheless, it must be recognized that ECM proteolysis mediated by MMP-9 may disrupt normal myocyte alignment and thereby contribute to increased myocardial anisotropy at and around the injured region. However, these issues remain speculative, and future studies that examine the direct role of MMP-9 in conduction abnormalities around the region of injury through targeted MMP-9 deletion are warranted.

CONCLUSIONS

There are several limitations of the present study that must be recognized. First, the mice that were used in the present study had a reporter construct for the MMP-9 gene promoter inserted into the genome. Therefore, although these transgenic mice were useful in detecting the induction of MMP-9 transcription, functionality of MMP-9 with respect to postinjury scar remodeling and/or arrhythmogenic potential could not be determined. Second, MMP-9 promoter induction was studied in separate cohorts of mice at each of the designated post-RF current time points, precluding the possibility of performing longitudinal measurements in the same mice. Finally, in this study, electrical conduction patterns were examined at an “early” postinjury time point at which the fibrotic deposition around the scar was probably incomplete. In light of past findings that increased fibrosis is associated with greater propensity of arrhythmias (28–30), whether and to what degree a higher fibrotic content at later postinjury time points would further exacerbate electrical conduction abnormalities warrant further investigation. These limitations notwithstanding, the findings of the present study demonstrate that MMP-9 gene promoter induction occurred in a region- and time-specific manner after RF injury and heterogeneities in electrical conduction occurred coincident with the region of MMP-9 promoter induction. These findings not only emphasize the role MMP-mediated matrix modification could play in the formation of an arrhythmogenic substrate but may also lead to the development of specific strategies for modification of MMP-9 activation to modulate adverse structural and electrical remodeling after myocardial injury.