Matrix metalloproteinase 13 mediates nitric oxide activation of endothelial cell migration

Abstract

To explore the mechanisms by which NO elicits endothelial cell (EC) migration we used murine and bovine aortic ECs in an in vitro wound-healing model. We found that exogenous or endogenous NO stimulated EC migration. Moreover, migration was significantly delayed in ECs derived from endothelial NO synthase-deficient mice compared with WT murine aortic EC. To assess the contribution of matrix metalloproteinase (MMP)-13 to NO-mediated EC migration, we used RNA interference to silence MMP-13 expression in ECs. Migration was delayed in cells in which MMP-13 was silenced. In untreated cells MMP-13 was localized to caveolae, forming a complex with caveolin-1. Stimulation with NO disrupted this complex and significantly increased extracellular MMP-13 abundance, leading to collagen breakdown. Our findings show that MMP-13 is an important effector of NO-activated endothelial migration.

Fig. 1.

NO induces EC movement in a wound-healing model. (A and B) BAEC (A) or MAEC (B) monolayers were injured and treated with 10^-6 M bradykinin, 500 μM l-NAME, bradykinin/l-NAME, or 100 μM DEA-NO, and cell movement was monitored by microscopy (n = 3 by triplicate; mean ± SD; *, P < 0.05 vs. control; #, P < 0.05 vs. bradykinin). (C) BAEC monolayers were seeded on 8-μm porous transwell filters and treated with 100 μM DEA-NO, 10^-6 M bradykinin, 0.5 mM l-NAME, or the combination bradykinin/l-NAME. After 6 h of plating, transmigration was evaluated by confocal microscopy (see Methods for details, n = 2 by triplicate; *, P < 0.05 vs. control). (D) ECs were injured and treated with 100 μM DEA-NO or 10^-6 M bradykinin. Total cell number was evaluated at the indicated times (n = 3 by triplicate; mean ± SD). (E) EC monolayers were treated with 100 μM DEA-NO or 500 μM l-NAME. MMP activities were measured by fluorimetry from culture supernatants collected at regular time points (n = 3 by quadruplicate; mean ± SD; *, P < 0.05 vs. control). (F) EC monolayers (a–c) were injured (d–n, see Methods for details) and treated with 100 μM DEA-NO (b, c, g, and h), 100 μM DEA (m and n), 10^-6 M bradykinin (i and j), or 10^-6 M bradykinin/500 μM l-NAME (k and l). MMP-13 was visualized over time (0, 1, and 2 has indicated) with immunohistochemical staining (MMP-13, FITC, green; nuclei Hoechst, blue) by confocal microscopy (magnification: ×40, n = 3).

Fig. 2.

ECs from eNOS-deficient mice migrate slower and show lower MMP-13 levels when compared with their eNOS WT counterparts. (A) Aortas and hearts from a pool of eNOS-deficient mice and eNOS WT mice were homogenized and used to evaluate eNOS, MMP-13, MT1-MMP, ICAM-2, and GAPDH expression (n = 3 animals by triplicate). (B) Aortic rings were isolated from eNOS-deficient mice and eNOS WT mice, and MMP-13 was visualized by immunohistochemistry and immunohistofluorescence using an anti-MMP-13 antibody. MMP-13 was visualized by peroxidase staining (Upper, magnification ×20) and FITC (Lower, magnification ×60). Aortic nuclei were stained with Hoechst. ICAM-2 expression was visualized in red (n = 5). (C) MAECs from eNOS WT or eNOS-deficient mice were isolated, and lysates were immunobloted to detect the expression of MMP-13 and eNOS (Left). In addition, a wound-healing assay was performed in cells treated with 100 μM DEA-NO, 10^-6 M bradykinin, 500 μM l-NAME, and bradykinin/l-NAME (n = 3 by triplicate; mean ± SD; *, P < 0.01 vs. control).

Fig. 3.

EC migration depends on MMP-13 expression. (A) Immunoblot detection of MMP-13, MT1-MMP, and MMP-2 (Left) and GAPDH, MMP-13, and MT1-MMP (Right) in MMP-13 (Left) and GAPDH (Right) silenced cells. Shown is one representative experiment from a total of three. (B) ECs were injured and treated with 100 μM DEA-NO (n = 3 by quadruplicate; mean ± SD; *, P < 0.05 vs. control; #, P < 0.05 vs. DEA-NO). (C) ECs were subject or not to MMP-13 silencing, injured, and treated with 10^-6 M bradykinin (n = 3 by triplicate; mean ± SD; *, P < 0.05 vs. nonstimulated/nonsilenced; #, P < 0.05 vs. stimulated with bradykinin/nonsilenced).

Fig. 4.

MMP-13 colocalizes with caveolin-1 in endothelial plasma membranes. (A) Confocal microscopy analysis of BAECs showing the expression of MMP-13 in nonpermeabilized and permeabilized BAECs. (B) Confocal microscopy analysis showing the expression of caveolin-1 (green) and MMP-13 (red) in BAECs. Merged panel shows colocalization of MMP-13 and caveolin-1 (yellow) (n = 5). (C) Caveolae-enriched fractions were isolated and identified by immunoblot with anti-caveolin-1. MMP-13 was also evaluated in the same fractions (n = 3). (D) Caveolin-1-positive (fractions 9 and 10) and -negative (fraction 12) fractions from ECs were subjected to cross-coimmunoprecipitation with anti-MMP-13 (Left) and anti-caveolin-1 (Right). Caveolin-1 or MMP-13 were detected by immunoblot (n = 4). (E) Cell lysates were subjected to cross-coimmunoprecipitation as indicated, in octyl-glucoside (OG), Nonidet P-40, and RIPA buffers (n = 3).

Fig. 5.

NO disrupts the MMP-13/caveolin-1 complex in BAEC. (A) Immunoblot of MMP-13 from cell lysates or culture media collected after 2 h of treatment of ECs with 100 μM DEA-NO or 10^-6 M bradykinin. The graph represents the densitometric analysis of data from three independent experiments (mean ± SD; *, P < 0.05). (B) MMP-13 activity assay in cell media from ECs treated for 2 h with 100 μM DEA-NO (n = 3; mean ± SD; P < 0.05). (C) Immunoprecipitation of MMP-13 from caveolin-1-enriched fractions of vehicle and ECs treated with 100 μM DEA-NO. Caveolin-1 was detected by immunoblot (n = 4 by triplicate; mean ± SD; *, P < 0.01). (D) MMP-13 was expressed in E. coli and purified by affinity chromatography. IP, immunoprecipitation; WB, Western blot. (Left) Coomassie staining of a 12% SDS/PAGE. M, molecular weight marker; L, bacterial lysate from cells treated with 1 mM isopropyl β-d-thiogalactoside (IPTG); P, purified MMP-13. (Right) Immunoblots from purified MMP-13, with anti-6XHis and anti-MMP-13 antibodies. (E) Crossed-coimmunoprecipitation experiments from purified MMP-13 incubated with cell lysates of BAECs for 16 h at 4°C and treated with DEA-NO or vehicle for 2 h (n = 3; mean ± SD; *, P < 0.05).

Fig. 6.

NO disrupts the in vitro MMP-13/caveolin-1 complex. (A) Recombinant 6XHis–MMP-13 and GST–caveolin-1 were purified and subject to immunoprecipitation (IP) experiments in the presence or absence of 100 μM DEA-NO to visualize the binding to caveolin-1 and MMP-13, respectively (n = 3; mean ± SD; *, P < 0.05). IB, immunoblot. (B) Recombinant 6XHis–MMP-13 or GST–caveolin-1 were pretreated with 100 μM DEA-NO for 1 h, bound to GST–caveolin or 6XHis–MMP-13, respectively (nontreated with DEA-NO), and subjected to immunoprecipitation (IP) experiments (n = 3; mean ± SD; *, P < 0.05). IB, immunoblot.