OUP user menu

Cyclic strain-mediated regulation of endothelial matrix metalloproteinase-2 expression and activity

Nicholas von Offenberg Sweeney, Philip M Cummins, Yvonne A Birney, John P Cullen, Eileen M Redmond, Paul A Cahill
DOI: http://dx.doi.org/10.1016/j.cardiores.2004.05.008 625-634 First published online: 1 September 2004


Objective: To investigate the role of cyclic strain in controlling matrix metalloproteinase-2 (MMP-2) expression and activity in endothelial cells (ECs) in vitro. Methods: A Flexercell® Tension Plus™ FX-4000T™ system was used to apply a physiological level of equibiaxial cyclic strain (0–10% strain, 60 cycles/min, 0–24 h, cardiac waveform) to bovine aortic endothelial cells (BAECs). Cells and conditioned media were harvested for analysis of MMP-2/9 expression and activity (pro and active) using reverse-transcriptase polymerase chain reaction (RT-PCR), Western blotting and zymography techniques. Results: Cyclic strain significantly increased MMP-2 expression and activity force- and time-dependently. Pretreatment with Giα-protein inhibitors, pertussis toxin (PTX) and NF023, transient expression of inhibitory mutants of Giα-subunits, or pretreatment with RGD peptides to block RGD-dependent integrin signaling failed to attenuate strain-induced increases in MMP-2 expression in BAECs. In contrast, inhibition of Gβγ-signaling with βArk-ct or tyrosine kinase blockade with genistein reduced strain-induced MMP-2 expression while concomitantly inhibiting strain-induced p38 and ERK activity in these cells. Pretreatment with PD169316 and PD98059 to selectively inhibit p38 and ERK activity, respectively, also resulted in a significant inhibition of the strain-induced MMP-2 response. Finally, inhibition of the adaptor protein, Shc, (via Shc-SH2 transfection) resulted in a significant decrease in strain-induced MMP-2 activity concomitant with a reduction in ERK activity in BAECs. Conclusion: Cyclic strain stimulates MMP-2 expression, in part, by stimulating both p38- and ERK-dependent pathways through activation of Gβγ and tyrosine kinase in BAECs.

  • MMP-2
  • Endothelial matrix metalloproteinase
  • Cyclic strain-mediated regulation

This article is referred to in the Editorial by P. Lacolley (pages 577–579) in this issue.

1. Introduction

Vascular remodeling underlies the pathogenesis of major cardiovascular diseases, such as atherosclerosis and restenosis [1,2]. The role of specialized extracellular matrix-degrading enzymes called matrix metalloproteinases (MMPs) has become the subject of intense recent interest in relation to physiological and pathological vascular remodeling [3,4]. MMPs are zinc-dependent proteases that cleave ECM (e.g., collagen, laminin, fibronectin) as well as non-matrix substrates (e.g., growth factors, receptors) [5]. Experimental evidence suggests that the major drivers of vascular remodeling, namely hemodynamics, injury, inflammation, and oxidative stress, regulate MMP expression and activity [3,4,6–10]. In addition, non-specific MMP inhibition seems to oppose remodeling [11,12]. Indeed, an emerging concept is that MMP-related genetic variations may contribute to heterogeneity in the presentation and natural history of atherosclerosis [13–15].

Various in vivo animal models highlight the importance of hemodynamic forces in modulating MMP expression and function [16,17]. Elevation of transmural pressure in porcine arteries ex vivo induced MMP-2/9 matrix-degrading activities [18], suggesting that MMPs may also be involved in early vascular remodeling associated with hypertension. Changes in the hemodynamic environment are thought to be of major importance in the failure of saphenous vein grafts [19]. Investigation of potential MMP involvement has shown upregulation of MMP-2/9 production after transposition of porcine saphenous veins in the carotid artery position [19]. Ex vivo comparison of human saphenous vein grafts in simulated arterial vs. venous conditions indicated that arterial conditions stimulate MMP expression and activation [19]. Finally, several studies using various cell lines revealed that hemodynamic forces could both stimulate [20–22] and inhibit [23–25] MMP expression and activity in vitro.

Cyclic circumferential strain, the repetitive pulsatile distention of the arterial wall, is closely linked to vascular disease, although the signal transduction pathways by which cyclic strain elicits functional and structural responses are only now being elucidated [26–28]. Considering the close association between MMP-2/9 function and vascular remodelling, we hypothesize that chronic cyclic strain of vascular endothelial cells (ECs) modulates expression/function of these MMPs through specific signaling pathways. In this study, we subsequently report on the regulation by cyclic strain of MMP-2 in bovine aortic endothelial cells (BAECs). Our data indicates that cyclic strain promotes a significant increase in MMP-2 expression and activity. Moreover, we demonstrate that strain stimulates MMP-2 expression, in part by stimulating both p38- and ERK-dependent pathways through activation of Gβγ and tyrosine kinase in these cells.

2. Methods

All materials, unless otherwise indicated, were supplied by Sigma-Aldrich (Poole, UK) and were of the highest purity commercially available.

2.1. Cell culture and cyclic strain

BAECs (Corriell Cell Repository, Camden, NJ) were cultured in RPMI 1640 media supplemented with 10% fetal bovine serum and antibiotics (50 U/ml penicillin, 50 μg/ml streptomycin). Cells (P5–P15) were grown in a humidified atmosphere of 5% CO2/95% air at 37 °C. For cyclic strain studies, BAECs were seeded into 6-well Bioflex® plates (Dunn Labortechnik, Germany) at 6 × 105 cells/well. Bioflex® plates contain a pronectin-coated silicon membrane bottom which enables precise deformation of cultured cells by microprocessor-controlled vacuum [29]. When cells reached confluency, media was removed and replaced with serum-free media and the cells exposed to cyclic strain. A Flexercell® Tension Plus™ FX-4000T™ system (Flexcell International Corp. Hillsborough, NC) was used to apply physiological equibiaxial cyclic strain (0–10% strain, 60 cycles/min, 0–24 h, cardiac waveform). For inhibitor studies, BAECs were placed in serum-free media containing pharmacological agents (Calbiochem, San Diego, CA) to attenuate specific signalling pathways prior to strain; protein tyrosine kinase (50 μM genistein), Gi-protein (100 ng/ml pertussis toxin (PTX) or 10 μM NF023) and mitogen-activated protein kinase (MAPK) (10 μM PD98059 or PD169316). Cells were subsequently exposed to 0% or 5% cyclic strain for 24 h. Media and cells were harvested post-strain and cells placed in ice-cold lysis buffer (as previously described [33]) and subjected to ultrasonication. All samples were stored at −80 °C. Protein concentration was measured by the method of Bradford [30] with BSA as a standard (Bio-Rad Laboratories, Hercules, CA).

2.2. Lactate dehydrogenase assay

Enzymatic activity of lactate dehydrogenase (LDH), a well-known marker for cellular damage, was monitored in both cells and media as described [31].

2.3. Western blotting

Prior to Western blotting, conditioned media samples were concentrated 10-fold by centrifugal filtration (10 kDa cutoff). Both lysate and conditioned media samples were subsequently resolved by 10% SDS-PAGE under reducing conditions according to the method of Laemmli [32]. Gels were electroblotted and immunostained as previously described [33]. Primary antisera: 1:200 anti-MMP-2 polyclonal IgG (Chemicon, Temecula, CA). Secondary antisera: 1:1000–1:10,000 HRP-conjugated IgG (Amersham Biosciences). For quantitative comparisons between bands, scanning densitometry was performed using a Kodak 1D image analysis software system (Scientific Imaging Systems, Rochester, NY). Bands were identified based on molecular weight and co-migration with known MMP standards (Chemicon). Total protein/well was used for normalization. Membranes were routinely stained with Ponceau S to normalize for protein loading/transfer.

2.4. MAPK/p38 activity

MAPK/p38 activity was determined by measuring phosphorylated ERK-1/2 and p38 expression in cell lysates following SDS-PAGE. Antiphospho-ERK-1/2 and antiphospho-p38 primary polyclonal antibodies (1:500) (Calbiochem) were used as described above.

2.5. In-gel zymography

Following experimentation, cells and conditioned media were collected and media samples centrifuged at 5000 × g for 5 min to remove cells. Conditioned media (equal volume) and cellular lysates (equal protein) were subsequently mixed with loading buffer. Pro-MMP-2 and MMP-2 activity levels were determined by gelatin zymography, combined with scanning densitometry, as previously described [34]. Bands were identified using gelatinase standards (Chemicon). Total protein/well was used for normalization.

2.6. Reverse-transcriptase polymerase chain reaction

Extraction of total BAEC mRNA and subsequent monitoring of mRNA expression was performed by reverse-transcriptase polymerase chain reaction (RT-PCR) as previously described [33]. Genes monitored were MMP-2 (N=35 cycles) and glyceraldehyde phosphate dehydrogenase (GAPDH, included for normalization, N=22 cycles). Primer pairs (MWG Biotech, UK) used were (i) MMP-2(534bp): 5′-gtggcaaccccgacgtgg-3′ (For), 5′-gcagggctgtccgtcgg-3′ (Rev), and (ii) GAPDH(341bp): 5′-tcctgcaccaccaactgctt-3′ (For), 5′-tgcttcaccaccttcttgat-3′ (Rev). PCR reaction products were analysed by agarose gel electrophoresis with scanning densitometry.

2.7. Transfection studies

For transient transfection, dominant negative plasmids were used to selectively ablate Giα (i.e., Giα1-G202T, Giα2-G203T and Giα3-G202T, Guthrie cDNA Resource Centre, Sayre, PA) and Shc (i.e., Shc-SH2, Professor John Shyy, UC Riverside) signalling pathways [35,36]. Transfection of βArk-ct (a selective Gβγ scavenger) was also used to attenuate Gβγ signalling [35]. LipofectAMINE™ Reagent (Invitrogen) was employed throughout (protocol as defined by manufacturer). Following recovery, transfected BAECs were seeded into Bioflex® plates (8 × 105 cells/cm2) for straining experiments. Transfection efficiency was routinely monitored by co-transfection with either LacZ or green fluorescent protein (GFP). Western blotting also confirmed over-expression of transfected genes. Mock transfections were included in all experiments.

2.8. Statistical analysis

Results are expressed as mean±S.E.M. Experimental points were performed in triplicate with a minimum of three independent experiments (n=3). Statistical comparisons between groups of normalized densitometric data were performed using both unpaired Student's t-test and Wilcoxon signed rank test. A value of P≤0.05 was significant.

3. Results

3.1. Cyclic strain increases pro-MMP-2 and MMP-2 expression and activity in BAECs

The effects of cyclic strain on MMP expression in BAECs were determined by measuring MMP-2/9 levels (pro and active) in cell lysates and conditioned media. Cyclic strain (5%) stimulated BAECs to preferentially secrete pro-MMP-2 and MMP-2 in conditioned media (Fig. 1a–b) without any significant release of pro-MMP-9 or MMP-9 (data not shown). Western blotting subsequently confirmed the strain-dependent increase in MMP-2 protein in conditioned media (Fig. 1d). Moreover, cyclic strain increased pro-MMP-2 activity in cell lysates (Fig. 1c) in addition to MMP-2 mRNA levels (Fig. 1e). The changes in pro-MMP-2 activity and MMP-2 protein expression (both monitored in conditioned media) were force- and time-dependent with significant increases observed after 8 h and at 2.5% strain, respectively (Fig. 2a–b). Similar results were also observed for pro-MMP-2 activity in lysates (data not shown). No increase in LDH release was observed following cyclic strain (up to 10%) indicating that strain was not inducing cellular injury. For all subsequent studies, strain-dependent effects on pro-MMP-2 and MMP-2 were monitored in BAEC-conditioned media, unless otherwise indicated.

Fig. 2

Time- and force-dependent increase in BAEC pro-MMP-2 and MMP-2 in response to cyclic strain. BAECs exposed to (a) 0% or 5% strain for 3–24 h and (b) 0–10% strain for 24 h. Conditioned media was analyzed for pro-MMP-2 activity (zymography) and MMP-2 protein expression (Western blot). Histograms represent normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained controls.

Fig. 1

Cyclic strain-induced increases in pro-MMP-2 and MMP-2 expression and activity. Following cyclic strain (5%, 24 h), conditioned media, cell lysates and total RNA were harvested. (a) Pro-MMP-2 and MMP-2 activities identified by gelatin zymography (gel inverted for clarity). MMP activities were monitored in (b) conditioned media (upper panel, MMP-2; lower panel, pro-MMP-2) and (c) cellular lysates (pro-MMP-2). Strain-dependent increases in (d) MMP-2 protein (in conditioned media) and (e) MMP-2 mRNA expression are also shown (mRNA gels inverted for clarity). (f) Histogram represents normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained controls. All gels are representative.

3.2. Cyclic strain-mediated changes in pro-MMP-2 and MMP-2 are Giα-protein-independent

In order to determine the involvement of Giα-subunits in the strain-induced modulation of MMP-2, BAECs were initially pretreated for 4 h with pertussis toxin (PTX) or NF023 before exposure to cyclic strain (5%) for a further 24 h in the absence or presence of inhibitor. Pretreatment with either inhibitor had no significant effect on strain-induced increases in either pro-MMP-2 activity or MMP-2 protein expression (both monitored in conditioned media) (Fig. 3a—data shown for PTX only).

Fig. 3

The effect of Gi-protein inhibitors and dominant negative mutants on cyclic strain-induced increases in pro-MMP-2 and MMP-2. BAECs were exposed to cyclic strain (5%, 24 h). (a) Effect of PTX on strain-induced pro-MMP-2 activity (zymography) and MMP-2 protein expression (Western blot) levels in conditioned media. (b) Effect of Giα-subunit dominant negative mutants on strain-induced pro-MMP-2 activity levels (conditioned media). (c) Effect of βArk-ct on strain-induced pro-MMP-2 activity levels (upper gel, conditioned media; lower gel, lysate). Histogram represents normalized band densitometry readings averaged from three independent βArk-ct experiments±S.E.M.; *P≤0.05 vs. unstrained controls; **P≤0.05 vs. mock transfected cells at 5% strain. All gels are representative.

To allow for a more definitive evaluation of the role of Gi-proteins in mediating these effects, selective inhibition of Giα(1–3) subunits by transfection of Giα-specific dominant negative mutants prior to strain was performed [35]. Transfection efficiencies of up to 50% were attained, as demonstrated by GFP/LacZ cotransfections (data not shown). Enhanced expression of specific Giα mutants was achieved in the absence of any effect on other Giα-proteins (data not shown). Furthermore, overexpression of Giα mutants failed to alter basal levels or cyclic strain-induced increases in either pro-MMP-2 activity (Fig. 3b) or MMP-2/pro-MMP-2 protein expression (data not shown) as compared to mock transfectants.

3.3. Cyclic strain-mediated changes in pro-MMP-2 and MMP-2 are Gβγ-dependent

In order to examine the possible role of Gβγ subunits in strain-mediated regulation of MMP-2, BAECs were transfected with βArk-ct, a Gβγ-sequestering agent [35]. Transfection with βArk-ct resulted in a minor decrease in basal pro-MMP-2 activity levels in conditioned media. Following 5% strain however, levels of pro-MMP-2 activity were reduced by 83±3% as compared to mock transfectants (Fig. 3c, upper gel). A similar result was also observed for pro-MMP-2 in cell lysates (Fig. 3c, lower gel).

3.4. Cyclic strain increases ERK and p38 activity in a Giα-independent, Gβγ-dependent manner

As strain-induced pathways appear to involve Ras and mitogen-activated protein kinases (MAPKs) [37,38], the effect of cyclic strain on MAPK activity was subsequently examined in BAECs by determining phospho-ERK-1/2 and phospho-p38 activity. Exposure to cyclic strain (5%, 24 h) resulted in a significant increase in phospho-ERK-1/2 activity without affecting total ERK-1/2 protein expression (Fig. 4a, upper gel, 1.7±0.13-fold in combined levels of pp-ERK-1/2). However, over-expression of Giα(1–3) dominant negative mutants (data not shown) or βArk-ct (Fig. 4b, left histogram), failed to significantly alter strain-induced phospho-ERK-1/2 levels in these cells as compared to mock-transfectants. In parallel studies, cyclic strain significantly increased phospho-p38 activity without affecting total p38 expression (Fig. 4a, lower gel). Moreover, transfection with βArk-ct significantly decreased the strain-induced phospho-p38 activity as compared to mock-transfectants (Fig. 4b, right histogram). We also noted an increase in phospho-p38 activity under non-strained conditions following βArk-ct transfection (relative to mock), although this was statistically insignificant.

Fig. 4

The effect of βArk-ct on cyclic strain-induced increases in ERK and p38 MAPK activity. BAECs were exposed to cyclic strain (5%, 24 h). (a) Effects of cyclic strain on phospho-ERK-1/2 (pp-ERK-1 and pp-ERK-2) (upper gel) and phospho-p38 (lower gel) levels. (b) Histograms showing effects of βArk-ct on strain-induced pp-ERK-1/2 (left) and phospho-p38 (right) levels in BAECs (normalized against total ERK-1/2 and p38 protein loading, respectively). These represent normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained control. **P≤0.05 vs. 5% cyclic strain. All gels are representative. NS=not significant.

3.5. Cyclic strain increases pro-MMP-2 and MMP-2 in a MAPK-dependent manner

As cyclic strain increased both phospho-ERK and -p38 activity in BAECs, the effect of pharmacological inhibition of ERK and p38 activity on strain-induced changes in MMP-2 expression was determined. BAECs were exposed to cyclic strain (5%, 24 h) in the absence or presence of PD98059 and PD169316, specific inhibitors of mitogen-activated protein kinase (MEK) and p38 kinase, respectively [39,40]. Both inhibitors significantly attenuated strain-induced increases in pro-MMP-2 activity (conditioned media) (Fig. 5a, upper gel). In parallel studies, PD98059 and PD169316 significantly decreased strain-induced increases in MMP-2 protein expression (conditioned media) by 64±17% and 57±8%, respectively (Fig. 5a, lower gel).

Fig. 5

The effect of MAPK inhibition on strain-induced increases in pro-MMP-2 and MMP-2. BAECs were pretreated with either PD98059 or PD169316 and exposed to cyclic strain (5%, 24 h). (a) Pro-MMP-2 activity (upper gel) and MMP-2 protein expression (lower gel). (b) Histogram represents normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained control; **P≤0.05 vs. 5% strain. Both gels are representative.

3.6. Cyclic strain stimulates pro-MMP-2 activity in an RGD-independent manner

Integrins play an important role in transducing mechanical stimuli into intracellular signals [41,42]. We examined the effects of integrin inhibition using synthetic linear (H–Arg–Asp–Gly–OH) and cyclic (cyclo-Arg–Gly–Asp–D–Phe–Val) RGD peptides (Bachem, UK) [43] on strain-induced changes in MMP-2 expression and activity. Pretreatment of cells with either peptide (0.5 mM) failed to significantly inhibit strain-induced increases in pro-MMP-2 activity in conditioned media (Fig. 6—data shown for linear RGD peptide only).

Fig. 6

The effect of linear RGD peptide on cyclic strain-induced increases in pro-MMP-2 activity. Effect of linear RGD peptide on pro-MMP-2 activity in conditioned media following strain (5%, 24 h). Histogram represents normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained control; NS=not significant. Gel is representative.

3.7. Cyclic strain stimulates pro-MMP-2 and MMP-2 in a tyrosine kinase-dependent manner

Since activation of ERK-1/2 and p38 is involved in strain-induced increases in pro-MMP-2 and MMP-2 in BAECs (see Fig. 5), we investigated the upstream signaling mechanism(s) that lead to strain-mediated activation of these MAPKs (see Fig. 4). Protein tyrosine kinase (PTK) phosphorylation is central to many signal transduction events [44]. Recent evidence suggests that PTK may mediate cellular responses through multiple intracellular signaling pathways including the Shc/Grb2/Sos/MAPK pathway [45,46]. We subsequently examined the role of PTK phosphorylation and the Shc/MAPK pathway on strain-induced changes in pro-MMP-2 and MMP-2. To determine if PTK plays a role in the activation of ERK and MMP-2, BAECs were preincubated with 50 μM genistein (a selective PTK inhibitor) prior to strain (5%, 24 h). Genistein caused a significant reduction in strain-induced phospho-ERK-2 (but not phospho-ERK-1) activity (Fig. 7a). Moreover, genistein significantly attenuated strain-induced pro-MMP-2 activity (in conditioned media) without any significant effect on baseline unstrained levels (Fig. 7b).

Fig. 7

The effect of PTK and Shc inhibition on cyclic strain-induced increases in pro-MMP-2 activity. Following pretreatment with genistein or transfection with Shc-SH2, BAECs were exposed to cyclic strain (5%, 24 h). Effect of PTK inhibition on (a) strain-induced phospho-ERK-1/2 expression in lysates and (b) strain-induced pro-MMP-2 activity in conditioned media. Effect of Shc on strain-induced pro-MMP-2 activity in media [(c) upper gel] and ERK-1/2 activity in lysates [(c) lower gel] is also shown. Histograms represent normalized band densitometry readings averaged from three independent experiments±S.E.M.; *P≤0.05 vs. unstrained control; **P≤0.05 vs. 5% strain. All gels are representative.

The adaptor protein Shc may play an important role in linking activated PTKs to downstream targets [45,46]. Indeed, Shc has previously been shown to activate the ERK pathway via tyrosine kinase Grb2/Sos signaling complexes [47]. To test whether Shc mediates strain-induced changes in MMP-2, BAECs were transfected with Shc-SH2 to ablate endogenous Shc activity [47] (over-expression confirmed by Western blot using anti-Shc monoclonal IgG from Transduction Laboratories, Lexington, KY). This led to a significant reduction in strain-induced increases in phospho-ERK-1 activity, and even more profoundly with phospho-ERK-2 activity (Fig. 7c, lower gel, 88±11% reduction in combined levels of pp-ERK-1/2). These reductions were concomitant with significant decreases in both basal and strain-induced increases in pro-MMP-2 activity in media (Fig. 7c, upper gel). A similar inhibition by Shc-SH2 of strain-induced increases in MMP-2 protein expression (as monitored in media) was also observed (data not shown).

4. Discussion

Structural adaptation of the vasculature occurs in response to both physiological and pathological changes in blood pressure and flow [1,2]. ECM components must be degraded and resynthesized to facilitate remodeling [3–5]. Of key interest therefore is whether changes in the vessel hemodynamic environment can regulate expression and function of matrix-degrading endothelial MMPs. We demonstrate that cyclic strain increases BAEC MMP-2 activity and expression in a force- and time-dependent manner, mediated through both Gβγ/p38- and PTK/Shc/MAPK-dependent pathways.

Unlike vascular SMCs, only a few studies have reported on the effects of cyclic strain on vascular ECs. Cyclic strain-mediated regulation of membrane type-1 MMP (MT1-MMP) in ECs has been reported [22,48]. A recent study by Wang et al. [49] also focuses closely on this issue. Unlike the present paper, this study employs a sinusoidal waveform cell-stretch paradigm ranging from 0% to 20% strain. These workers report time- and stretch-dependent increases for pro-MMP-2 activity and MMP-2 protein expression (observed only at the upper strain threshold) using human umbilical vein endothelial cells (HUVECs), which were of relatively similar magnitude to those seen here. This may reflect differences between the cell types (BAEC vs. HUVEC) and/or stretching paradigm (cardiac vs. sinusoidal) used.

A complex control of endothelial MT1-MMP expression in response to distinct hemodynamic stimuli, comprising modulation of specific transcription factor expression (early growth response gene product-1, Egr-1) and post-translational modification (serine phosphorylation of Sp1) suggests highly coordinated hemodynamic regulation of MMP expression [22,23,48]. Moreover, arteries incubated under pressure displayed significant increases in MMP-2 (but not MMP-9) expression and activity while simultaneously degrading elastin, suggesting that both MMP-2 expression and matrix degradation are locally enhanced at sites of higher transmural pressures [21,50]. Our in vitro study demonstrates that cyclic strain promotes MMP-2 production in vascular ECs, reinforcing the importance of local hemodynamic environment in controlling MMP expression and function in vivo.

The in vivo mechanisms that stimulate vascular cell production/activation of MMPs remain unclear. Cytokines, intimal injury, hemodynamic forces, and other factors leading remodeling are thought most responsible [3,4]. In the current study, under serum-depleted conditions, the predominant gelatinase isoform released from BAECs following strain was MMP-2 (although under serum conditions, MMP-9 was also strain-dependently increased; unpublished observation). To elucidate the signaling mechanisms involved in strain-mediated modulation of MMP-2, we initially focussed on Gi-protein pathways. Previous studies have demonstrated rapid, dose-dependent Gi-protein activation in HUVECs by cyclic uniaxial strain [51]. Moreover, PTX-sensitive G-proteins have been implicated in MMP activation and function in several cell types [52]. However, in the current study, selective ablation of individual Giα-protein subunits with dominant negative mutants or blockade of Giα-signaling with PTX/NF023 had no effect on strain-induced increases in pro-MMP-2/MMP-2, suggesting lack of any significant involvement of Giα-signaling in transduction of these events. Interestingly, strain-mediated increases in pro-MMP-2/MMP-2 were significantly attenuated by βArk-ct, implicating Gβγ-pathways in mechanotransduction of these events. Recent studies suggest that shear-induced Ras activation is mediated through Gβγ-signaling in human vascular ECs [53], implying that Gβγ activation of the Ras/ERK pathway could putatively mediate the strain-induced MMP-2 responses observed here. However, this seems unlikely as the inhibitory effect of βArk-ct was independent of changes in downstream ERK activity. In contrast, cyclic strain increased BAEC phospho-p38 activity, and selective inhibition of p38 activity with PD169316 significantly reduced strain-induced changes in pro-MMP-2 and MMP-2. Since inhibition of Gβγ functionality with βArk-ct inhibited strain-induced phospho-p38, this suggests that the involvement of Gβγ-signaling in strain-mediated regulation of MMP-2, occurs in part, via a p38-dependent pathway.

Previously, two distinct, complementary signaling mechanisms mediating MMP induction have been reported: activating protein-1 (AP-1)-dependent transcriptional activation via the ERK-1/2 pathway and AP-1-independent enhancement via p38α MAPK by mRNA stabilization [54]. As p38 inhibition attenuates strain-induced MMP-2 responses in BAECs, it is possible that Gβγ-signaling impacts on stress activated protein kinase/c-jun N-terminal kinase (SAPKs/JNKs) and p38 pathways in these cells. The recent findings of Wang et al. [49] are noteworthy in this regard as they demonstrate that JNK inhibition attenuates strain-induced MMP-2 expression in HUVECs. While SAPKs/JNKs and p38 are also activated by Gβγ-subunits in a pathway (in myofibroblasts) involving RhoA, Rac1 and Cdc42, Rho kinase inhibition did not attenuate MMP-2 activity or expression [55]. In addition, since PTKs are important regulators of MMPs [56] and are regulated by Gβγ-subunits [57], it is possible that Gβγ-signaling also affects BAEC PTK activity. Moreover, an effect of Gβγ-subunits on MT1-MMP and TIMP-2, both involved in MMP-2 activation [3–5], cannot be ruled out in these events. Further work will be required to delineate the precise mechanism of p38/Gβγ-dependent activation of pro-MMP-2 and MMP-2 in BAECs following cyclic strain.

MMP activity and expression is consistently attenuated following MEK inhibition suggesting that ERK may regulate strain-induced changes in MMP-2 in BAECs [3–5,58]. In this regard, pharmacological inhibition with specific ERK inhibitors significantly attenuated the strain-induced changes in pro-MMP-2/MMP-2 expression and activity observed. Integrins may serve as mechanosensors in vascular ECs, subsequently leading to ERK or p38 activation. Indeed, shear stress causes integrin-Shc association and recruitment of the Grb2/Sos signaling complex that then leads to ERK activation [45,47]. In the current study, strain-induced changes in pro-MMP-2 activity were unaffected following integrin blockade with either linear or cyclic RGD peptides. This rules out the possibility that RGD-dependent integrins mediate these events, although it is noteworthy that RGD peptides were found to inhibit strain-induced EC tube formation and migration (unpublished observation). As cyclic strain has been shown to effect the expression/localization of several different integrin subtypes [59], it is possible that an RGD-independent integrin(s) may mediate these strain-induced events through association with the Shc signaling pathway.

Tyrosine kinases have been implicated in hemodynamically induced changes in EC function. Shear stress has been shown to induce a rapid, transient tyrosine phosphorylation of Flk-1 and its concomitant association with Shc [45,47], while cyclic strain has been shown to activate proline-rich tyrosine kinase 2 [61] and epidermal growth factor receptor [62]. Moreover, Shc is implicated in signaling via many different types of receptors, including integrins and tyrosine kinases [45,47,60] and its involvement in ERK activation has been established [63]. Our findings are in agreement with studies demonstrating that stretch-induced increases in ERK activity were attenuated by PTK inhibition [64]. Our data further demonstrate that selective inhibition of Shc signaling significantly attenuated strain-induced pro-MMP-2 and MMP-2 increases in BAECs, while concomitantly inhibiting phospho-ERK activity. As ERK inhibition attenuates strain-induced MMP-2 responses, it is likely that PTK/Shc is responsible, in part, for mediating these events. These data confirm the potential importance of PTKs in force-mediated regulation of MMP-2 in BAECs via PTK/Shc/ERK association.

In conclusion, our data demonstrate that cyclic strain stimulates MMP-2 activity and expression in BAECs. Moreover, we show for the first time that strain-induced MMP-2 responses are independent of Giα-protein activation but dependent on Gβγ/p38- and PTK/Shc/ERK interactions in these cells. It is subsequently tempting to speculate that strain-induced changes in endothelial MAPK signaling may functionally regulate endothelial phenotype in vivo by modulating MMP production. In this regard, our future goal is to extend these investigations to in vivo and ex vivo animal models of cyclic strain, as cultured cells are not necessarily the most representative model of the physiological situation.


This research was supported by grants from the Wellcome Trust (PAC, YAB), Health Research Board of Ireland (PAC, PMC), National Institutes of Health (DK09227, HL59696 and AA-12610, EMR), Dublin City University Educational Trust (PMC) and the American Heart Association (JPC). We thank Drs. Dermot Walls and Brendan O'Connor for critical comments.


  • Time for primary review 32 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
View Abstract