Pressure overload-dependent membrane type 1-matrix metalloproteinase induction: relationship to LV remodeling and fibrosis

Abstract

Increased myocardial extracellular matrix collagen represents an important structural milestone during the development of left ventricular (LV) pressure overload (PO); however, the proteolytic pathways that contribute to this process are not fully understood. This study tested the hypothesis that membrane type 1-matrix metalloproteinase (MT1-MMP) is directly induced at the transcriptional level in vivo during PO and is related to changes in LV collagen content. PO was induced in vivo by transverse aortic constriction in transgenic mice containing the full length human MT1-MMP promoter region ligated to luciferase (MT1-MMP Prom mice). MT1-MMP promoter activation (luciferase expression), expression, and activity; collagen volume fraction (CVF); and left atrial dimension were measured at 1 (n = 8), 2 (n = 12), and 4 (n = 17) wk following PO. Non-PO mice (n = 10) served as controls. Luciferase expression increased by fivefold at 1 wk, fell at 2 wk, and increased again by ninefold at 4 wk of PO (P < 0.05). MT1-MMP expression and activity increased at 1 wk, fell at 2 wk, and increased again at 4 wk after PO. CVF increased at 1 wk, remained unchanged at 2 wk, and increased by threefold at 4 wk of PO (P < 0.05). There was a strong positive correlation between CVF and MT1-MMP activity (r = 0.80, P < 0.05). Left atrial dimension remained unchanged at 1 and 2 wk but increased by 25% at 4 wk of PO. When a mechanical load was applied in vitro to LV papillary muscles isolated from MT1-MMP Prom mice, increased load caused MT1-MMP promoter activation to increase by twofold and MT1-MMP expression to increase by fivefold (P < 0.05). These findings challenge the canonical belief that PO suppresses overall matrix proteolytic activity, but rather supports the concept that certain proteases, such as MT1-MMP, play a pivotal role in PO-induced matrix remodeling and fibrosis.

the development of adverse myocardial remodeling and extracellular matrix (ECM) accumulation represent a critical turning point in the development of chronic pressure-overload (PO) (1, 7, 41). However, the mechanisms that lead to myocardial fibrosis have not been completely defined. Matrix metalloproteinases (MMPs) are conventionally considered enzymes that primarily degrade structural ECM proteins (10, 30, 34). However, the results of recent studies have challenged the canonical concept that MMPs act only to degrade collagen and decrease collagen accumulation (9, 14, 31, 32, 36, 37, 43). In fact, in vitro and in vivo studies have identified that a specific class of membrane bound MMPs, in particular membrane type 1 (MT1)-MMP, may actually facilitate the development of fibrosis (9, 14, 31, 32, 36, 37, 43). MT1-MMP has been shown to act on a diverse portfolio of profibrotic signaling molecules. For example, MT1-MMP acts on latency-associated TGF binding protein (LTBP) to release transforming growth factor (TGF)-β and results in profibrotic signaling (9, 13, 36, 37).

Unlike other MMPs, MT1-MMP is activated intracellularly and is inserted into the membrane in a fully functional form; thus its transcriptional regulation represents a critical control point (14, 31, 32, 43). In vitro studies have shown that MT1-MMP transcriptional activation is altered by mechanical stimuli. For example, the application of a mechanical force to isolated cells and isolated vascular tissue has been shown to increase MT1-MMP expression and demonstrated that there is a mechanical-molecular set point for the induction of MT1-MMP promoter activation (11, 14, 31, 32, 43). We hypothesized that the stimuli present in vivo in chronic PO is sufficient to increase MT1-MMP promoter activation and is associated with the development of ECM remodeling and fibrosis. Accordingly, the purposes of this study were as follows: first, to define the time course of changes in MT1-MMP promoter activation, expression, abundance, and activity in vivo in a murine model of PO; second, to determine whether there is a temporal relationship between MT1-MMP promoter activation and PO-induced ECM remodeling and fibrosis; and third, determine in vitro whether a change in mechanical load and neurohormonal activation act independently or in a combinatorial fashion to alter MT1-MMP promoter activation and expression.

METHODS

Overview and Rationale

The response to PO, created in a murine model by transverse aortic constriction, was examined 1, 2, and 4 wk after PO and compared with non-PO control mice. Changes in LV structure and function were assessed using echocardiography; changes in collagen volume fraction (CVF) were determined histologically.

To determine whether and to what extent in vivo PO causes induction of MT1-MMP gene promoter activity, a transgenic MT1-MMP promoter reporter mouse was developed (MT1-MMP Prom). Changes in mRNA for endogenous MT1-MMP, luciferase, and profibrotic gene expression were assessed using TaqMan QPCR and PCR SuperArray in LV myocardial samples. Changes in MT1-MMP activity were assessed using a MT1-MMP-specific quenched fluorogenic substrate.

To determine whether the effects of increased mechanical load and neurohormonal activation were independent or interdependent determinants of MT1-MMP induction, LV papillary muscles were isolated from MT1-MMP Prom mice and studied in vitro under conditions of normal versus increased load and with or without ANG II. Changes in mRNA for endogenous MT1-MMP, luciferase, and transcription factor gene expression were assessed in the papillary muscle studies using TaqMan QPCR.

MT1-MMP Prom

To develop the MT1-MMP Prom mouse, the human genomic MT1-MMP fragment extending from −3,364 bp (relative to the transcriptional start site) to the first intron was ligated to the firefly luciferase gene (luc, 550 aa ORF) (31, 37). Transgenic founders were developed within the FVB background, and a stable, viable colony was established. Mice born into this colony did not display any phenotypic, developmental, and/or reproductive abnormalities.

Transverse Aortic Constriction (PO)

The methods used to create transverse aortic constriction have been described previously (5). All MT1-MMP Prom mice were between 12 and 14 wk of age at the time of PO. Mice were studied 1 wk (n = 8), 2 wk (n = 12), or 4 wk (n = 17) following PO. Ten mice did not undergo transverse aortic constriction and served as non-PO controls. All procedures performed were approved by the Institution Animal Care and Use Committee of the Medical University of South Carolina in accordance with National Institutes of Health guidelines.

Echocardiography

Echocardiographic measurements were made using a 40 MHz mechanical scanning transducer (707B) and a Vevo 770 echocardiograph (VisualSonics, Toronto, Canada). LV dimension, volume, wall thickness, fractional shortening, mass, and left atrial dimension were measured using the American Society of Echocardiography criteria (20). LV mass was normalized to body weight. Left atrial diameter was used to reflect chronic changes in LV diastolic pressure; diameter increases as a function of sustained increased pressure (i.e., an integration of pressure over time) (4, 19).

The ambient in vivo hemodynamic load on the LV myocardium was assessed by calculating the LV systolic wall stress using the following formula: LV systolic wall stress = 1.35 × LVSP × ESD/[4 × PWThs × (1 + PWThs/ESD)], where LVSP is LV systolic pressure, ESD is end-systolic dimension, and PWThs is LV posterior wall end-systolic thickness. LV systolic pressure was calculated as the sum of the pressure gradient across the transverse aortic constriction and the mean aortic pressure (MAP) distal to the transverse aortic constriction. MAP was measured using tail cuff pressure measurements made with a CODA system (Kent Scientific, Torrington, CT) in restrained, but nonanesthetized mice. MAP = (2 × DBP + SBP)/3, where DBP is diastolic blood pressure and SBP is systolic blood pressure. The aortic pressure gradient was determined using the modified Bernoulli equation: Pressure gradient = 4(V_peak)², where V_peak is the peak Doppler velocity measured across the transverse aortic constriction site.

Collagen Content by Light Microscopy

Fibrillar collagen content was examined qualitatively using birefringence light microscopy and picrosirius red staining and quantitatively by calculating the CVF as the area stained by picrosirius red divided by the total area of interest (3, 5, 6). Fields with large blood vessels were excluded from the analysis. Areas examined were distributed throughout the myocardium from subendocardial to subepicardial and excluded the epicardial surface.

MT1-MMP Activity Assay

Myocardial MT1-MMP activity was directly measured in LV myocardium from each mouse using an MT1-MMP-specific quenched fluorogenic substrate using methods described previously (8, 9). Upon cleavage of the peptide by active MT1-MMP, the quenching group becomes sufficiently separated from the fluorescent group, allowing a fluorescent signal to be emitted. Fluorescence was measured (ex330 nm/em405 nm) and recorded on a fluorescent microplate reader (Fluorostar Galaxy BMG Labtechnologies, Cary, NC) and compared with a standard curve using active recombinant MT1-MMP. Several in vitro validation studies were performed previously to demonstrate the specificity of the MT1-MMP substrate (8, 9).

Immunoblot Studies to Examine MT1-MMP Protein Abundance

MT1-MMP protein abundance was determined using methods published previously (37). LV myocardial samples from control or 4-wk PO mice were thawed and transferred to cold extraction/homogenization buffer [buffer volume: 1:6 (wt/vol)] containing 10 mM cacodylic acid (pH 5.0), 0.15 M NaCl, 10 mM ZnCl₂, 1.5 mM NaN₃, and 0.01% Triton X-100 (vol/vol) and homogenized. Ten micrograms of the supernatant were fractionated on a 4–12% bis-Tris gradient gel. Proteins were transferred to nitrocellulose membranes (0.45 μm; Bio-Rad) and incubated in antiserum corresponding to MT1-MMP (0.4 μg/ml, catalog no. AB815; Millipore Biosciences, Temecula, CA). Antisera was diluted in 5% nonfat dry milk-PBS. A secondary peroxidase-conjugated antibody was then applied (1:5,000, 5% nonfat dry milk-PBS), and signals were detected with a chemiluminescent substrate (Western Lighting Chemiluminescence Reagent Plus, Perkin-Elmer). Band intensity was quantified using Gel-Pro Analyzer software (version 3.1.14; Media Cybernetics, Silver Spring, MD) and reported as the percent change from the non-PO control homogenates.

Gene Expression Analysis

Changes in selected gene expression were examined in two sets of myocardial samples: LV and papillary muscles. In LV samples from control mice and mice 1, 2, and 4 wk after PO, endogenous MT1-MMP and luciferase expression were assessed using a TaqMan QPCR method. Expression of selected, representative profibrotic genes listed in Table 2 were assessed in control mice and mice 1 and 4 wk after PO using a custom designed PCR SuperArray method. Studies performed in LV papillary muscles isolated from MT1-MMP Prom mice to examine the effects of load and ANG II on MT1-MMP, luciferase, and transcription factor expression were assessed by TaqMan QPCR.

Table 2.

Effects of pressure overload on profibrotic and transcription factor gene expression

		Fold Change
Name	Abbreviation	1 wk PO	4 wk PO
Membrane type 1 matrix metalloproteinase	MT1-MMP	2.33*	2.54*
Procollagen type 1, α1	Col1α1	2.68*	1.66*
Procollagen type 1, α2	Col1α2	4.00*	1.63*
Latent transforming growth factor β-binding protein 2	LTBP2	5.04*	1.41*
Connective tissue growth factor	CTGF	2.89*	1.02
Plasminogen activator inhibitor type 1	PAI-1	3.22*	1.14
Plasminogen activator inhibitor type 2	PAI-2	1.80*	1.00
Transforming growth factor β-receptor 1	TGF-βR1	1.94*	1.28

^*Either a ≥2-fold change or a significant increase in ΔCt value at the P < 0.05 level versus non-PO controls.

RNA isolation.

In both the TaqMan QPCR and PCR SuperArray methods, mRNA was isolated using similar techniques. Myocardial homogenates were subjected to RNA extraction (RNeasy Fibrous Tissue Mini Kit; Qiagen, Valencia, CA), and the quantity and quality of the RNA were determined (Experion Automated Electrophoresis System; Bio-Rad) using methods published previously (31, 36, 37).

TaqMan QPCR.

RNA (1 μg) was reverse transcribed to generate cDNA (iScript cDNA Synthesis Kit; Bio-Rad). The cDNA was amplified with gene-specific primer/probe sets (TaqMan Universal PCR Master Mix: catalog no. 4364321; Applied Biosystems, Foster City, CA) using single-color real-time PCR (rtPCR, MyiQ, Bio-Rad). The specific TaqMan primer/probe sets (Applied Biosystems) were luciferase and MT1-MMP (catalog nos. Hs0000237119_m1 and Mm01318965_m1); transcription factors c-FOS, c-JUN, v-rel reticuloendotheliosis viral oncogene homolog A (RELA), and NF-κB1 (catalog nos.: Mm00487425_m1, Mm00495062_s1, Mm00476361_m1, Mm00501346_m1); and 18 S rRNA (catalog no. 4333760F). These transcription factors were chosen because the MT1-MMP gene contains two AP-1 binding sequences (responsive to c-FOS) and one NF-κB1 binding sequence. Negative controls were run to verify the absence of genomic DNA contamination (reverse transcription control) and the absence of overall DNA contamination in the PCR system and working environment (template control).

PCR SuperArray.

RNA (1 μg) was reverse-transcribed to generate cDNA using an RT² First Strand Kit (catalog no. C-03; SABiosciences, Frederick, MD), and the cDNA was immediately assayed for gene expression by QPCR using a custom RT² Profiler PCR Array (Custom Services; SABiosciences) designed to test multiple candidate genes, including four housekeeping control genes, in a multi-well plate format using methods published previously and manufacturer's protocol (31, 36, 37). The candidate genes examined are listed in Table 2.

Results from both TaqMan QPCR and PCR SuperArray methods are presented as either mRNA Expression = 2^−(ΔCt), where ΔCt = change in cycle threshold, or as fold change from control (see statistical analysis below).

In vitro Application of Myocardial Mechanical Load, Neurohormonal Activation, or Both

Papillary muscles were isolated from MT1-MMP promoter reporter mice (n = 12) and placed in an isolated muscle chamber for study using methods described previously (5, 6). Two LV papillary muscles were isolated from each mouse and attached to a force/length servo-system (model 300B; Aurora Scientific, ASI, Aurora, Canada) and controlled by custom-made software (Dynamic Muscle Control; ASI). With the use of parallel platinum electrodes, the muscle was stimulated at a rate of 0.5 Hz with a square wave pulse of 5-ms duration and voltage 10% above the threshold. Each muscle was allowed to contract at a light isotonic preload of 0.1 g for 30 min until it reached an equilibrium (baseline) state. During this preconditioning period, at 10-min intervals, each muscle was subjected to sets of three isometric contractions and isotonic contractions. After the preconditioning period, L_max was defined and one papillary muscle from each mouse was randomly assigned to a control (1 muscle from each of 12 mice) or loaded (1 muscle from each of 12 mice) groups. L_max is the muscle length at which active tension is maximal. Muscles in each of these two groups were further divided into those treated with ANG II (ANG II was added to Krebs-Henseleit buffer at a concentration of 100 nM; n = 6 muscles) and those not treated with ANG II (Krebs-Henseleit buffer; n = 6 muscles). Control muscles were held at a muscle length near slack length (very lightly preloaded muscle at 0.1 g) for 3 h. Loaded muscles were held at a muscle length of 15% greater than that at slack length (equivalent to 100% of L_max preload) for 3 h. After the 3-h time period, isometric contractions and isotonic contractions were performed to demonstrate papillary muscle viability, the muscles were removed from the study chamber, and expression of endogenous MT1-MMP, luciferase, and selected transcription factors was determined using TaqMan QPCR methods described above.

Statistical Analysis

Temporal changes in LV structure and function, CVF, luciferase expression, MT1-MMP expression, and MT1-MMP activity were compared between the non-PO control and PO groups using a one-way ANOVA; pairwise comparisons were made using the Bonferroni test to adjust for multiple comparisons. In LV papillary muscle studies, changes in MT1-MMP promoter activation and transcription factor expression were made using a two-way ANOVA examining the effects of load and ANG II; pairwise separations were made using the Bonferroni test to adjust for multiple comparisons. Correlations between luciferase expression and LV wall stress, between CVF and MT1-MMP activity, and between MT1-MMP and cFOS expression were examined using a Pearson's correlation coefficient.

Gene expression was determined by measuring the cycle thresholds (C_t) for each sample, normalizing to the 18 S signal, the final expression levels determined as a function of total RNA concentrations and calculated using the equation: Expression = 2^−(ΔCt). The expression values for each animal were then averaged, and the means ± SE were plotted. In LV myocardial samples, expression of endogenous MT1-MMP, luciferase, and selected profibrotic and transcription factors was expressed as fold change from non-PO control mice. Fold expression values greater than 2.0 or a significant difference in the ΔCt value were considered to have a significant change in gene expression.

Values of P < 0.05 were considered statistically significant. All statistical procedures were performed using the STATA statistical software package (Statacorp, College Station, TX). Results are presented as means ± SE. The authors had full access to the data and take full responsibility for its integrity.

RESULTS

Temporal Changes in LV Structure and Function

All PO mice had a minimum gradient of 100 mmHg by 2 wk after across the transverse aortic constriction. None of the PO mice had significant changes in body weight, tibial length, or heart rate after transverse aortic constriction (Table 1). LV wall thickness, mass, and mass-to-end-diastolic volume ratio increased after 1 wk of PO and then increased further after 2 wk of PO (Table 1 and Fig. 1). However, no further changes occurred between 2 and 4 wk after PO. For example, when compared with non-PO controls, LV mass increased by 22% after 1 wk and 36% after 2 wk (P < 0.05) but did not increase further at 4 wk of PO.

Table 1.

LV structure and function following transverse aortic constriction

		PO, wk
	Non-PO Controls	1	2	4
Sample size, n	10	8	12	17
Body weight, g	26 ± 1	24 ± 1	26 ± 1	26 ± 1
Tibial length, mm	19.8 ± 0.1	19.5 ± 0.1	20.0 ± 0.1	19.7 ± 0.1
Heart rate, beats/min	510 ± 11	501 ± 18	512 ± 22	537 ± 10
Mean aortic pressure, mmHg	83 ± 2	99 ± 4	98 ± 4	83 ± 2
Aortic pressure gradient, mmHg	0	95 ± 4*	105 ± 2*	104 ± 2*
End-diastolic volume, μl	50 ± 1	39 ± 2*	41 ± 3*	45 ± 1*+
Fractional shortening, %	35 ± 1	35 ± 1	34 ± 1	31 ± 1*+#
LV wall thickness, mm	0.8 ± 0.1	0.9 ± 0.1*	1.1 ± 0.1*+	1.1 ± 0.1*+
LV mass, mg	79 ± 4	88 ± 6*	114 ± 6*+	120 ± 5*+
LV mass-to-body weight ratio, mg/g	3.1 ± 0.1	3.8 ± 0.1*	4.3 ± 0.2*+	4.5 ± 0.2*+

Values are means ± SE. LV, left ventricular; PO, pressure-overload induced by transverse aortic constriction.

^*P < 0.05 vs. non-PO controls;

⁺P < 0.05 vs. 1 wk PO;

^#P < 0.05 vs. 2 wk PO.

Fig. 1.

Serial changes in left ventricular (LV) structure and function. PO, pressure-overload created by transverse aortic constriction; EDV, LV end-diastolic volume. *P < 0.05 vs. non-PO control; #P < 0.05 vs. 1 wk after PO; +P < 0.05 vs. 2 wk after PO. LV mass/EDV increased over 2 wk of PO but did not significantly change after 2 wk of PO (A). Significant increases in left atrial diameter (an index of LV diastolic function) occurred after 2 wk of PO (B).

Fractional shortening was unchanged after 1 and 2 wk of PO but fell slightly after 4 wk of PO (35 ± 1% non-PO controls vs. 31 ± 1%, P < 0.05); however, these fractional shortening values after 4 wk of PO remained in the normal range (Table 1 and Fig. 1). By contrast, left atrial dimension remained unchanged at 1 and 2 wk of PO but increased significantly by 25% at 4 wk of PO (P < 0.05).

Temporal Changes in CVF and MT1-MMP Activity

CVF was increased by 80% at 1 wk after PO compared with control, remained relatively unchanged at 2 wk, but was increased by 200% at 4 wk (P < 0.05; Fig. 2). MT1-MMP activity increased by 75% 1 wk after PO compared with control, then fell by 15% 2 wk after PO compared with 1 wk after PO, and then rose again by 15% 4 wk after PO compared with 2 wk after PO (Fig. 2). MT1-MMP abundance measured by immunoblot was increased after 4 wk of PO by 35% compared with control (Fig. 2). There was a strong positive correlation between PO-induced changes in MT1-MMP activity and CVF (r = 0.80, P < 0.05).

Fig. 2.

Serial changes in membrane type 1-matrix metalloproteinases (MT1-MMP) activity (A), MT1-MMP abundance (B), and collagen content (C). MT1-MMP activity and collagen volume fraction increased 1 wk after PO, decreased, or did not change between 1 and 2 wk after PO, and subsequently increased between 2 and 4 wk after PO. There was a strong positive correlation between PO-induced changes in MT1-MMP activity and collagen volume fraction (r = 0.80, P < 0.05). *P < 0.05 vs. non-PO control; #P < 0.05 vs. 1 wk after PO; +P < 0.05 vs. 2 wk after PO.

Temporal Changes in MT1-MMP Gene Promoter Activation

One week after PO, both luciferase and MT1-MMP mRNA expression increased by five- and threefold, respectively, compared with non-PO control (Fig. 3). However, 2 wk after PO, both luciferase and MT1-MMP mRNA were decreased compared with values 1 wk after PO; luciferase fell to values comparable with non-PO control, and MT1-MMP fell but remained elevated compared with non-PO control. Four weeks after PO, both luciferase and MT1-MMP mRNA increased by nine- and 1.5-fold, respectively, compared with 2-wk values. Luciferase at 4 wk was higher than at 1 wk, whereas MT1-MMP was equivalent to 1-wk values.

Fig. 3.

Serial changes in MT1-MMP gene promoter activation (measured by Luciferase mRNA expression) (A), endogenous MT1-MMP mRNA expression (B), and LV systolic wall stress (an index of in vivo LV load) (C). MT1-MMP expression, luciferase expression, and LV systolic wall stress increased 1 wk after PO, decreased between 1 and 2 wk after PO, and subsequently increased between 2 and 4 wk after PO. There was a strong positive correlation between in vivo changes in LV systolic wall stress and luciferase expression (r = 0.81, P < 0.05). *P < 0.05 vs. non-PO control; #P < 0.05 vs. 1 wk after PO; +P < 0.05 vs. 2 wk after PO.

The relationship between MT1-MMP gene promoter activation and changes in in vivo LV load (as measured by systolic wall stress) was also examined. When compared with non-PO control, there was a 110% increase in LV systolic wall stress 1 wk after PO, then a 23% decrease at 2 wk and then a further 25% increase at 4 wk (Fig. 3). There was a strong positive correlation between in vivo changes in load (LV systolic wall stress) and MT1-MMP promoter activation (luciferase expression) (r = 0.81, P < 0.05).

Temporal Changes in LV Myocardial Profibrotic Gene Expression

PCR SuperArray was used to examine activation of profibrotic pathways downstream from MT1-MMP. After 1 wk, in vivo PO caused a significant increase in the expression of multiple collagens, LTBP, TGF-β receptor, connective tissue growth factor, and serine proteinase inhibitors (Table 2). Some, but not all of these, remained increased 4 wk after PO.

Effects of Changes in In Vitro Myocardial Mechanical Load, Neurohormonal Stimulation, or Both

The effects of an in vitro application of a mechanical load on papillary muscles isolated from MT1-MMP Prom mice were examined (Fig. 4). An increase in mechanical load alone increased MT1-MMP expression by 526% compared with unloaded controls. The in vitro treatment of unloaded papillary muscles with ANG II caused an increase in MT1-MMP expression by 446% compared with unloaded muscles not treated with ANG II. However, the addition of ANG II to the loaded muscles did not augment the effects of the applied mechanical load; MT1-MMP expression was increased 309% in the loaded muscles treated with ANG II compared with unloaded muscles not treated with ANG II. To determine whether load-induced MT1-MMP expression was caused by increased promoter activation, luciferase expression was measured in a similar manner. The results demonstrated that increased load caused an induction of luciferase expression by 146% compared with the unloaded control muscles, whereas ANG II alone induced expression by 87% compared with unloaded control muscles not treated with ANG II. However, the addition of ANG II to the loaded muscles again did not augment the effects of the applied mechanical load, inducing luciferase expression by 48% compared with unloaded control muscles not treated with angiotensin II.

Fig. 4.

In vitro application of myocardial mechanical load, neurohormonal activation, or both. LV papillary muscles isolated from the transgenic MT1-MMP promoter reporter mice were subjected in vitro to an increased load (Load +, hatched bar) and compared with control muscles subjected to no change in load (Load−, white bar). In addition, the effects of the presence of angiotension II either without (gray bar) or with (black bar) an increase in mechanical load were examined. The results demonstrated that increased load was sufficient to induce MT1-MMP expression (A). To establish that the increase in MT1-MMP expression was due to promoter activation, Luciferase expression was also measured and found to be increased in a similar manner (B). Additionally, the expression of several transcription factors capable of binding within the MT1-MMP promoter region was examined in response to increased load in the presence and absence of angiotensin II. Increased mechanical load alone was sufficient to induce the expression of MT1-MMP, luciferase, c-Fos (C), NF-κB1 (D), and v-rel reticuloendotheliosis viral oncogene homolog A (RELA) (E). Likewise, angiotensin II was sufficient to induce the expression of c-Fos, NF-κB1, and RELA; however, the addition of angiotension II did not augment the effects of the applied mechanical load. There was a strong positive correlation between MT1-MMP expression and c-Fos expression in the papillary muscles exposed to increased load in the absence of angiotensin II (r = 0.62, P < 0.05). *P < 0.05 vs. control no angiotension II.

To further examine potential mechanisms mediating the effects of hemodynamic load and neurohormonal activation, changes in mRNA expression of several transcription factors with putative binding sites in the MT1-MMP promoter region were examined. The expression of RELA, NF-κB1, c-FOS, and c-JUN in isolated LV papillary muscles treated with a mechanical stress in the presence and absence of ANG II are presented in Fig. 4. There was a significant increase in RELA, NF-κB1, and c-FOS, but not c-JUN, in LV papillary muscles when an increased load was applied. When unstressed muscles were treated with ANG II, there was increased transcription factor expression, albeit to a smaller extent than the effect of load. In addition, the effects of ANG II remained constant despite the addition of the mechanical stress. There was a strong positive correlation between MT1-MMP expression and c-Fos expression in the papillary muscles exposed to increased load in the absence of ANG II (r = 0.62, P < 0.05).

DISCUSSION

Interstitial proteases play an important role in the regulation of myocardial fibrillar collagen content and composition in normal and diseased hearts (7, 10, 21, 31). During the development of dilated cardiomyopathies, increased MMP induction and activity has been shown to lead to collagen degradation, and abnormal collagen structure and function (35, 38). By contrast, during chronic PO, an increase in fibrillar collagen accumulation has been postulated to reflect a general decrease in MMP induction and activation leading to a decrease in MMP-induced collagen degradation (1, 7, 41). However, recent studies have challenged the canonical concept that MMPs act only to degrade collagen and decrease collagen accumulation (9, 14, 31, 32, 36, 37, 43). For example, membrane-bound MMPs, particularly MT1-MMP, have been shown to activate profibrotic signaling following a murine myocardial infarction (MI) and during the aging process in mice (8, 9, 36, 37). Furthermore, the application of mechanical stimulus to endothelial cells and fibroblasts increased MT1-MMP expression (11, 14, 32, 43). These studies lead us to hypothesize that the stimulus present in vivo in chronic PO is sufficient to increase MT1-MMP promoter activation and produce a temporal relationship between MT1-MMP promoter activation and ECM remodeling and fibrosis. Results from the current study support several novel findings.

Effects of In Vivo PO on Temporal Changes in LV Structure and Function

Chronic in vivo PO caused time-dependent, nonuniform changes in LV structure and function. LV hypertrophy, primarily a cardiomyocyte response, occurred rapidly, was essentially complete by 2 wk of PO, and did not progress between 2 and 4 wk of PO. By contrast, the structural response in the ECM, primarily a fibroblast response, occurred in a different temporal pattern. CVF increased modestly during the rapid hypertrophic response, but between 2 and 4 wk when LV mass was unchanged, CVF increased in a marked fashion. Increased left atrial diameter, an index of increased LV diastolic pressure, followed a temporal pattern similar to that of CVF and did not increase significantly until after 2 wk of PO. In this murine model of in vivo PO, the hemodynamic load applied to the LV also followed a temporal pattern; LV systolic wall stress increased immediately after the creation of PO, and stress fell during the development of LV hypertrophy, but then rose again late in the course of PO. The use of transgenic MT1-MMP promoter reporter mouse allowed us to test the hypothesis that there is a relationship between the temporal response in LV structure and function to PO and the induction of MT1-MMP promoter activation, expression, abundance, and activity.

Effects of In Vivo PO on Temporal Changes in MT1-MMP

Data from the current study showed that chronic PO applied to the transgenic MT1-MMP promoter reporter mouse caused induction of the MT1-MMP promoter in LV myocardial samples as evidenced by the increase in luciferase expression. This induction followed a temporal pattern over the course of 4 wk of PO similar to that seen in LV wall stress. In addition, MT1-MMP expression, abundance, and activity followed similar temporal patterns. In fact, there were significant relationships between increased wall stress and MT1-MMP activity and between MT1-MMP activity and CVF. Together these data suggest that the increased MT1-MMP activity and abundance are a direct result of increased gene promoter activation rather than MT1-MMP protein or mRNA stabilization. There was a specific profile of mRNAs that emerged from an array analysis that suggested that the observed changes in MT1-MMP were associated with increased downstream profibrotic signaling, including activation of TGF-β-dependent pathways. Thus these data suggest that MT1-MMP induction plays a pivotal role in chronic in vivo PO-induced matrix remodeling and fibrosis.

Effects of Changes in In Vitro Myocardial Mechanical Load, Neurohormonal Stimulation, or Both

Because a chronic in vivo process that results in a change in LV load may also alter neurohormonal activation, it may be difficult to separate direct hemodynamic effects from the indirect neurohormonal effects. Therefore, papillary muscles isolated from MT1-MMP Prom mice were exposed to an increased in vitro mechanical load. Data indicated that increased load alone was sufficient to induce MT1-MMP expression. To establish that the increase in MT1-MMP expression was due to promoter activation, luciferase expression was also measured and was found to be increased in a similar manner. Additionally, the expression of several transcription factors with putative binding sites within the MT1-MMP promoter region was examined in response to in vitro increased load; increased mechanical load alone was sufficient to induce the expression of c-Fos, NF-κB1, and RELA. In addition, there was a significant relationship between an increase in c-Fos and MT1-MMP promoter activation. The effects of an isolated increase in neurohormonal activation were examined in vitro by treating papillary muscles with ANG II. Although ANG II alone was sufficient to induce the MT1-MMP promoter activation, MT1-MMP expression, and an increase in transcription factors, the addition of ANG II to an increase in load did not augment the effects of the applied mechanical load itself. Thus, although these data suggest that there is a coregulatory interaction between mechanical load and neurohormonal activation, changes in mechanical load are likely to serve as at least one of the important determinants of the MT1-MMP induction in PO.

MT1-MMP

There are six different members of the membrane type-MMPs; MT1-MMP is the prototypical and most studied example (30). Unlike other classes of MMPs, MT-MMPs are activated intracellularly and have full proteolytic capabilities when they are inserted into the cell membrane. MT-MMPs process a wide portfolio of substrates, which include structural components of the myocardial ECM and bioactive proteins and peptides (15–18, 21, 22, 25, 26, 29, 40, 42). For example, MT-MMPs contain a substrate recognition site for soluble pro-MMPs and form an important pathway for proteolytic activation (17, 18). In this example, MT-MMPs fulfill the canonical notion that MMPs act by degrading ECM structural proteins and decrease fibrillar collagen content (2, 17, 18, 27, 28, 33, 39). However, it has recently become clear that MT-MMPs also proteolytically process membrane bound and interstitial signaling molecules that do not lead to ECM structural protein degradation but actually result in a net increase in fibrillar collagen content (9, 14, 31, 32, 36, 37, 42, 43). These include cytokines, growth factors (such as TGF-β), and ECM glycoproteins (42). Data from the current study suggest that the increased MT1-MMP promoter activation, expression, abundance, and activity seen in PO were also associated with time-dependent activation of some of these mediators of ECM fibrosis. For example, TGF-β is synthesized as an inactive precursor bound to LTBP through disulfide bonds and a rich cysteine-rich motif found within LTBP-1 (13, 36, 37). MT1-MMP has been shown to cleave LTBP-1, releasing TGF-β to activate the profibrotic TGF-β signaling pathway. The observed increased expression in Colla1, Colla2, and PAI-1 and 2, which are classical markers of TGF-β pathway activation, supports our hypothesis that induction of MT1-MMP contributes to PO-induced myocardial fibrosis, at least in part, through TGF-β activation. There are also changes in expression of transcription factors, which are likely to act on the MT1-MMP promoter. The MT1-MMP gene contains two AP-1 binding sequences (responsive to c-FOS) and one NF-κB1 binding sequence. In the current study, both c-FOS and NF-κB1 were increased in response to an increase in in vitro load.

Effects of MT1-MMP on LV Structural Remodeling

Increased LV myocardial levels of MT1-MMP have been found in both patients and animal models of LV failure (8, 9, 31, 35–38). In particular, the ECM remodeling found in advanced aging and MI has been found to be associated with increased MT1-MMP expression and activity (9, 31, 36, 37). In mice during advanced age, there was a persistent increase in MT1-MMP expression and activity that was associated with an increase in interstitial fibrillar collagen (36). In addition, aging was associated with an increase in the processing of profibrotic molecules, specifically LTBP-1 and activation of the TGF-β pathway (36). Similar patterns were seen both in the peri-MI and remote regions following an MI (9, 37). Progressive and persistent increases in MT1-MMP expression and activity were associated with increase in interstitial fibrillar collagen, processing of profibrotic molecules, and activation of profibrotic signaling pathways (9, 37). By contrast, in the current study, rather than a persistent and progressive pattern, MT1-MMP promoter activation, expression, and activity appear to follow a time-dependent pattern following PO. In at least one other study, MT1-MMP abundance was increased following PO (23). Data presented in the study of Lin et al. (23) serve to support and are largely concordant with the results of the current study.

Limitations

There are additional regulatory control processes that can affect MT1-MMP proteolytic activity and targeting that were not examined in this study. These include posttranscriptional control by microRNAs and posttranslational effects of MT1-MMP phosphorylation that can regulate localization to the membrane and substrate specificity and may help target selected profibrotic signaling molecules (7). Addressing these additional processes in an in vivo PO models represents an important component of future studies. In addition, the temporal association between MT1-MMP activation and the structural and functional outcomes in chronic PO do not prove a cause and effect relationship. This step will require global or regional alterations in MT1-MMP activity during the development of chronic PO. However, the current study does provide a novel rationale for pursuing further studies defining the mechanistic role of MT1-MMP in PO-induced remodeling.

Conclusions

Chronic PO, studied in an in vivo murine model, caused a time-dependent induction of MT1-MMP promoter activation, expression, abundance, and activity, which temporally paralleled myocardial collagen accumulation. These findings indicate that PO does not suppress all matrix proteolytic activity, but in fact results in an increase in MT1-MMP, which appears to play a pivotal role in PO-induced matrix remodeling and fibrosis.

GRANTS

This work is supported by National Heart, Lung, and Blood Institute Grants HL-057952, HL-059165, and HL-095608 (to F. G. Spinale) and the Research Service of the Department of Veterans Affairs (to M. R. Zile and F. G. Spinale).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.R.Z., J.A.J., and F.G.S. conception and design of research; M.R.Z., C.F.B., R.E.S., A.O.V.L., and J.A. performed experiments; M.R.Z., C.F.B., R.E.S., A.O.V.L., J.A., R.M., J.A.J., and F.G.S. analyzed data; M.R.Z., C.F.B., R.E.S., R.M., J.A.J., and F.G.S. interpreted results of experiments; M.R.Z. and C.F.B. prepared figures; M.R.Z. drafted manuscript; M.R.Z., C.F.B., R.E.S., A.O.V.L., J.A., R.M., J.A.J., and F.G.S. edited and revised manuscript; M.R.Z., C.F.B., R.E.S., A.O.V.L., J.A., R.M., J.A.J., and F.G.S. approved final version of manuscript.