Increased expression of leukotriene C₄ synthase and predominant formation of cysteinyl-leukotrienes in human abdominal aortic aneurysm

Abstract

Leukotrienes (LTs) are arachidonic acid-derived lipid mediators involved in the pathogenesis and progression of diverse inflammatory disorders. The cysteinyl-leukotrienes LTC₄, LTD₄, and LTE₄ are important mediators of asthma, and LTB₄ has recently been implicated in atherosclerosis. Here we report that mRNA levels for the three key enzymes/proteins in the biosynthesis of cysteinyl-leukotrienes, 5-lipoxygenase (5-LO), 5-LO-activating protein (FLAP), and LTC₄ synthase (LTC₄S), are significantly increased in the wall of human abdominal aortic aneurysms (AAAs). In contrast, mRNA levels of LTA₄ hydrolase, the enzyme responsible for the biosynthesis of LTB₄, are not increased. Immunohistochemical staining of AAA wall revealed focal expression of 5-LO, FLAP, and LTC₄S proteins in the media and adventitia, localized in areas rich in inflammatory cells, including macrophages, neutrophils, and mast cells. Human AAA wall tissue converts arachidonic acid and the unstable epoxide LTA₄ into significant amounts of cysteinyl-leukotrienes and to a lesser extent LTB₄. Furthermore, challenge of AAA wall tissue with exogenous LTD₄ increases the release of matrix metalloproteinase (MMP) 2 and 9, and selective inhibition of the CysLT1 receptor by montelukast blocks this effect. The increased expression of LTC₄S, together with the predominant formation of cysteinyl-leukotrienes and effects on MMPs production, suggests a mechanism by which LTs may promote matrix degradation in the AAA wall and identify the components of the cysteinyl-leukotriene pathway as potential targets for prevention and treatment of AAA.

Leukotrienes (LTs) are powerful lipid mediators released by inflammatory cells, such as macrophages, neutrophils, eosinophils, and mast cells (1, 2). In cellular biosynthesis of LTs, arachidonic acid is oxygenated by 5-lipoxygenase (5-LO) into the unstable epoxide LTA₄ with the aid of the accessory 5-LO activating protein (FLAP). Here, the pathway branches. The LTA₄ hydrolase (LTA₄H) can metabolize LTA₄ to LTB₄, which acts as a potent chemoattractant. Alternatively, LTA₄ can be conjugated with glutathione by LTC₄ synthase (LTC₄S) to form cysteinyl-leukotrienes LTC₄, LTD₄, and LTE₄ (cysLTs), which promote airway smooth-muscle constriction and increased vascular permeability. LTB₄ and cysLTs exert biological effects through two pairs of G protein-coupled receptors denoted BLT1/BLT2 and CysLT1/CysLT2, respectively (3–6). LTs and their receptors exert diverse functions in immune and inflammatory responses and have been associated with many diseases, in particular asthma (7).

Abdominal aortic aneurysm (AAA) is a cardiovascular disease associated with aging, whereby chronic inflammation in the aortic wall and proteinase-mediated degradation of structural matrix proteins are believed to contribute to its development and rupture (8, 9). Inflammatory cells such as macrophages, neutrophils, and mast cells have been shown to be sources of proteolytic enzymes in the AAA wall (10). Moreover, several studies have pointed to the intraluminal thrombus as a site of proteolytic activity (11, 12) and source of chemotactic factors (13, 14), potentially influencing the structural and cellular composition of the AAA wall (15). LTs have been the subject of recent investigations indicating an important role for those lipid mediators in cardiovascular diseases. Thus, genetic, morphological, biochemical, pharmacological, and functional evidence collected from human and animal studies have implicated LTs, in particular LTB₄, in the development and progression of atherosclerosis (16, 17) and as risk factors in human myocardial infarction (18). Less is known regarding LTs and AAA, and studies using animal models have yielded conflicting results about the involvement of the 5-LO pathway in aneurysmal disease (19–21). In the present study we examined the expression of 5-LO cascade-related enzymes in human AAA. Biochemical analysis revealed an overexpression of 5-LO, FLAP, and LTC₄S, but not of LTA₄H, in the AAA wall. Using immunohistochemistry, we also found that these three enzymes, required for cysLT biosynthesis, are expressed in the media and adventitia layers of the AAA wall and colocalize with the presence of inflammatory cells, in particular macrophages and mast cells. Moreover, this enzymatic machinery is catalytically active and capable of converting arachidonic acid into LTs. Finally we show that LTD₄ induces metalloproteinases in a CysLT receptor-dependent manner, suggesting a mechanism by which these mediators may promote arterial wall degradation and development of AAA.

Results

Expression of 5-LO, FLAP, and LTC₄S mRNA in Human AAA.

In human AAA wall we found increased transcript levels of 5-LO, FLAP, and LTC₄S compared with control aortas, corresponding to 2.5-fold (2.46 ± 0.26, P = 0.002), 1.5-fold (1.51 ± 0.09, P = 0.0012), and 1.5-fold (1.81 ± 0.21, P = 0.0068) increases, respectively (Fig. 1). In contrast, the levels of LTA₄H mRNA in AAA wall were not significantly different when compared with control (Fig. 1).

Fig. 1.

Expression of 5-LO, FLAP, LTA₄H, and LTC₄S mRNA in human AAA wall (n = 28) was determined by real-time quantitative RT-PCR and compared with human control aortas (CTRL; n = 8). Data are presented as mean ± SEM. **P < 0.01; ***P < 0.001.

Expression of 5-LO, FLAP, and LTC₄S Protein in Human AAA.

To check the presence of leukotriene cascade-related enzymes at the protein level, we performed immunostaining in serial sections of AAA wall. The tunica media expressed focal accumulations of 5-LO, FLAP, and LTC₄S immunostaining that colocalize with the presence of CD68- and CD66b-positive cells, reflecting the presence of macrophages and neutrophils, respectively (Fig. 2A). In the same area it was also possible to see the presence of T lymphocytes, as shown by CD3-positive staining (Fig. 2A).

Fig. 2.

Coexpression of 5-LO, FLAP, and LTC₄S protein with inflammatory cell markers in human AAA. Serial sections of human AAA (n = 3) were stained for 5-LO, FLAP, LTC₄S, and CD68 (macrophage), CD66b (neutrophil), CD3 (T lymphocyte), and tryptase (mast cell) markers, revealing similar distribution in media (A) and adventitia (B) layers. (Magnification: 20×. Scale bars, 50 μm.)

In the adventitia, once again, there were also plenty of 5-LO- and LTC₄S-positive cells that colocalize with CD68 (macrophages) and CD66b (neutrophils) (Fig. 2B). Moreover, in the adventitia, it was also possible to detect the presence of mast cells in the same area, as determined by tryptase staining (Fig. 2B).

We also carried out immunohistochemistry of AAA thrombus. The abluminal side did not show the presence of cells, confirming previous observations that this part of thrombus is mainly composed of densely packed fibrin (22). In contrast, the luminal side of the thrombus showed the presence of focal accumulations of inflammatory cells, in particular macrophages and neutrophils, that colocalized with cells positive for 5-LO and LTC₄S (Fig. 3).

Fig. 3.

Expression of 5-LO and LTC₄S in the luminal part of intraluminal thrombus of human AAA. Serial sections of the luminal side of human AAA (n = 3) were stained for 5-LO, LTC₄S, and CD68 (macrophage) and CD66b (neutrophil) markers, showing coexpression in that area. L, lumen. (Magnification: 40×. Scale bar, 100 μm.)

Immunohistochemical staining of LTA₄H in AAA tissue showed the presence of LTA₄H in the media layer (Fig. S1A), adventitia layer (Fig. S1B), and luminal thrombus (Fig. S1C), indicating that the branch of 5-LO pathway responsible for LTB₄ biosynthesis is also present.

Biosynthesis of LTs.

Homogenates of human aortic tissue from AAA surgery were used to study the metabolism of arachidonic acid. Incubations of AAA wall tissue with arachidonic acid produced significant amounts of cysLTs (2.62 ± 0.61 ng/μg of protein) when compared with LTB₄ production (1.25 ± 0.51 ng/μg of protein) (Fig. 4), as assessed by reverse-phase HPLC coupled to enzyme immunoassay (EIA). Similarly, more cysLTs (2.15 ± 0.97 ng/μg of protein) than LTB₄ (1.02 ± 0.50 ng/μg of protein) were produced in the thrombus, although the differences did not reach statistical significance (Fig. 4).

Fig. 4.

Synthesis of cysLTs and LTB₄ in human AAA. Homogenates of human AAA wall and thrombus (n = 9) were incubated with arachidonic acid. CysLTs and LTB₄ were measured by HPLC/EIA, and data are presented as mean ± SEM. *P < 0.05.

In selected experiments we directly assessed the LTA₄H and LTC₄S activities by incubations of subcellular fractions with the unstable epoxide LTA₄. To do that, we isolated (by ultracentrifugation) cytosolic (rich in LTA₄H) and membrane (rich in LTC₄S) fractions from AAA wall and incubated them with LTA₄ (10 μM). HPLC quantification showed a 2-fold higher formation of LTC₄ (2.14 pmol/μg protein; Fig. S2A) compared with LTB₄ (1.10 pmol/μg protein; Fig. S2B) in the wall samples. Moreover, the tandem EIA analysis of HPLC fractions showed a stronger LTC₄S activity as compared with LTA₄H activity (Fig. 5) to convert the same substrate both in the wall (LTC₄ vs. LTB₄: 21.45 ± 7.37 vs. 10.25 ± 0.84 ng/μg protein) and in thrombus (LTC₄ vs. LTB₄: 13.22 ± 3.75 vs. 3.34 ± 0.71 ng/μg protein), confirming the data obtained from incubations with arachidonic acid.

Fig. 5.

Synthesis of LTC₄ and LTB₄ in human AAA. The AAA wall and the thrombus (n = 8) were homogenized at 100,000 × g for 1 h at 4 °C. The cytosolic fraction (supernatant) and the membrane fraction (pellet) were incubated with 10 μM of LTA₄. LTC₄ and LTB₄ were measured by tandem HPLC/EIA, and data are presented as media ± SEM. *P < 0.05.

LTD₄-Induced Release of Matrix Metalloproteinases.

It has been shown that cysLTs can induce the release of matrix metalloproteinases (MMPs) from macrophages (19, 23), such as MMP-9 and MMP-2, and the presence of these proteinases has been associated with degradation of connective tissue (9). Therefore, we examined the potential role of LTD₄ as a stimulus for the release of MMPs from AAA tissue.

Gelatin gel zymography showed an increased level of MMP-2 and its proform when AAA wall was incubated with LTD₄ (10 μM) as compared with the control group (Fig. 6 A and B). Moreover, pretreatment of AAA tissue with the selective CysLT1 antagonist montelukast (1 μM) reduced the MMP signals to control levels (Fig. 6 A and B).

Fig. 6.

Release of LTD₄-mediated MMP-2 in human AAA and inhibition by montelukast. Minced vascular wall derived from AAA lesions (n = 4) was preincubated for 30 min in the absence or presence of montelukast (1 μM), followed by incubation with LTD₄ (10 μM) for 24 h at 37 °C. The conditioned media were analyzed by gelatin gel zymography, and bands relative to proMMP-9, proMMP-2, and MMP-2 were analyzed by Image J software. (A) Representative gelatin gel; (B) densitometric analysis of gelatin gel zymography (n = 4). Data are expressed as mean ± SEM. *P < 0.05.

Discussion

LTs are paracrine lipid mediators that have potent proinflammatory biological activities. One class of LTs, the cysLTs, are well recognized as important signaling molecules in human asthma, and leukotriene-modifying drugs have been produced for this application (24). However, more recent studies have strongly implicated LTs as mediators of vascular inflammation, and in most cases LTB₄ was identified as the culprit. Consequently, a role for the LTB₄ arm of the 5-LO pathway, including FLAP, LTA₄H, and BLT1, in cardiovascular diseases is currently in the focus of intense research efforts, both within academia and industry (16, 25).

AAA is a vascular pathology characterized by weakening of the medial elastic lamina layer and infiltration of media and adventitia by immune cells, such as macrophages and T cells, which release a variety of chemokines and other mediators that can enhance degradation of the extracellular matrix (10). The most known and studied mediators involved in aortic wall degeneration are MMPs, and several investigations have indicated a pivotal role for MMPs in the pathological process leading to AAA growth and rupture (26). Furthermore, it has been suggested that the presence of intraluminal thrombus could affect the structural and cellular composition of the underlying vessel wall (15) and sustain the proteolytic activity by releasing MMPs (11).

Thus far, only a limited number of animal studies have addressed the potential role of LTs in AAA, yielding different results. Funk and coworkers found a protective effect of 5-LO gene deletion against AAA in atherosclerosis-prone, ApoE^−/− mice (19). In a later study, the same authors failed to detect a role for 5-LO in the angiotensin II-induced AAA model in the ApoE-deficient mouse (21). On the other hand, other investigators found that genetic deletion or pharmacological inhibition of the BLT1 receptor afforded significant protection against AngII-induced AAA (20, 27). In line with the latter observations, Houard et al. (14) recently proposed a role for LTB₄, derived from neutrophils within the intraluminal thrombus, as a chemotactic factor in AAA.

In the present study we used a collection of fresh AAA specimens obtained from patients undergoing elective surgery for repair of infrarenal aneurysms, to assess the expression profile and activity of enzymes in the 5-LO pathway. Surprisingly, we found that cysLTs, rather than LTB₄, may be involved in human AAA.

5-LO, FLAP, and LTC₄S mRNA Levels Are Increased in Human AAA Wall.

Analysis of 28 patients revealed a striking increase in the expression of 5-LO mRNA. In addition, we found significant increases in the mRNA levels of FLAP and LTC₄S (Fig. 1), thus completing the triplet of key proteins required for cysLTs biosynthesis. In contrast, LTA₄H did not display significantly elevated levels of mRNA, demonstrating an important difference between human AAA and human atherosclerosis, in which LTA₄H is up-regulated and has been shown to correlate with symptoms of plaque instability (17).

LT-Synthesizing Enzymes Are Present in Media and Adventitia Layers at Sites Rich in Inflammatory Cells.

By immunostaining it was possible to detect the presence of proteins of the whole LT enzymatic pathway in several segments of the AAA wall. Thus, the media layer presented focal accumulations of 5-LO-, FLAP-, LTA₄H-, and LTC₄S-positive cells. The same areas also stained positively for CD68, indicating that macrophages could be the main source of LTs and confirming previous results obtained in animal models (19, 28). Moreover, in agreement with previous studies demonstrating the presence of different types of immune cells in AAA (10, 13, 15), we could detect CD66b- and CD3-positive cells, markers of human neutrophils and T lymphocytes, respectively, and these cells clustered together with macrophages.

Going deeper into the AAA wall, through the adventitia layer, immunostaining showed a pattern similar to the one observed in the media, with presence of focal accumulations of inflammatory cells, mostly macrophages and neutrophils, which are positive for 5-LO, FLAP, LTC₄S, and LTA₄H. However, in the adventitia, cells staining strongly for tryptase were also detected, indicating the presence of mast cells that are known to be involved in the early and late phases of inflammatory reactions by release of chemokines, cytokines, and lipid mediators, such as cysLTs (29).

AAA Wall Converts Arachidonic Acid Primarily into cysLTs.

Next, we assessed whether our mRNA and morphological data correlated with differences in enzyme activities. Thus, we incubated homogenized human AAA wall tissue with arachidonic acid and analyzed the formation of LTs to probe the entire 5-LO pathway machinery. Tandem HPLC/EIA analysis revealed a significantly increased ability of the wall to convert arachidonic acid into cysLTs and lesser amounts of LTB₄, in accord with the results obtained by real-time PCR (Fig. 1). These data indicated a greater LTC₄S activity as compared with LTA₄H activity. To directly probe the LTC₄S and LTA₄H activities, we also carried out incubations with the unstable epoxide intermediate LTA₄, the immediate upstream substrate for both of these enzymes. Because soluble GSTs could potentially interfere with the LTC₄S assay, we subjected the AAA wall homogenates to ultracentrifugation and collected a cytosolic fraction containing LTA₄H, as well as a membrane fraction enriched in membrane-bound LTC₄S and devoid of soluble GSTs. Analysis of these incubates clearly demonstrated a higher activity of LTC₄S as compared with LTA₄H, yielding significantly more LTC₄ than LTB₄ (Fig. 5 and Fig. S2). Although large amounts of LTs could be detected in certain experiments (Fig. S2), the samples varied significantly with respect to enzyme activities. Comparison of data in Figs. 4 and 5 also shows that a lot more LTs were formed in incubations with LTA₄. This is probably explained by the fact that we used homogenized tissues in our activity measurements, in which the membrane-associated assembly of 5-LO and FLAP has been disrupted. In incubations with LTA₄, this step is circumvented, allowing direct assessment of downstream LTC₄S and LTA₄H.

Intraluminal Thrombus Is also a Potential Source of cysLTs.

Because the intraluminal thrombus has been suggested as an important source of chemotactic LTB₄ (14), we also analyzed this component of the aneurysm for presence of 5-LO pathway proteins and related enzyme activities. Immunohistochemical analysis of thrombus revealed the presence of inflammatory cells only at the luminal side. Staining for cell markers CD68 and CD66b showed the presence of macrophages and neutrophils, but no T cells. These cells also stained for 5-LO, LTC₄S, and LTA₄H, indicating that cells in the thrombus could potentially produce and release cysLTs, as well as the chemotactic factor LTB₄. In fact, further activity measurements, including incubations of thrombus with both arachidonic acid and LTA₄, revealed a predominant production of cysLTs that in the latter case reached statistical significance (Fig. 5).

Challenge with LTD₄ Increases Release of MMP-2 from AAA Wall Tissue.

Production of cysLTs in the vascular wall may have several effects that potentially could promote inflammation and aneurysm formation. For instance, MMPs have been implicated in the pathogenesis of aortic aneurysm (8, 9), and a recent study showed that reduced formation of AAA in 5-LO knockout mice is associated with a decrease in aortic MMP-2 activity (19). Along these lines, we show that challenge of human AAA wall tissue with LTD₄ induces a significant increase of proMMP-2 and MMP-2 levels that potentially could contribute to matrix degradation in the vessel wall. This effect seems to be mediated via CysLT1 receptors, because montelukast, a selective CysLT1 antagonist, could block the increased release of MMP-2 (Fig. 6).

In addition to increased MMP activity, previous studies have shown that cysLTs and their receptors, in particular CysLT1, are involved in de novo expression and secretion of several proinflammatory cytokines, such as IL-5, macrophage inflammatory protein-1α and -β, as well as TNF-α from macrophages and mast cells (30–32). These activities of cysLTs are believed to be involved in the amplification of innate as well as adaptative immune responses.

Anti-LTs as Potential Drugs Against AAA.

Assuming that production of cysLTs occurs during the course of AAA development, it should be possible to pharmacologically block its synthesis and actions with already-existing anti-LT drugs. Inhibition of 5-LO or FLAP would block synthesis of all LTs, as well as antiinflammatory lipoxins. Selective inhibition of LTC₄S, on the other hand, could block synthesis of cysLTs, while potentially sparing lipoxin biosynthesis, as previously demonstrated for a selective inhibitor of LTA₄H (33, 34). This approach may also be advantageous, given the steadily increasing number and complex cross-regulations reported for cysLT receptors (35–37). Finally, on the basis of the results of the present study, an obvious approach would be to use established antiasthma drugs, such as montelukast or similar CysLT1 antagonists, for treatment of AAA, a possibility that should be straightforward to test in a clinical trial.

Materials and Methods

Sample Collection.

Patients (n = 28) who were about to undergo elective surgery for infrarenal AAA were selected for the study. Tissue was taken from the anterior or lateral wall, because tissue sampling from the posterior wall was considered to add to the risk of surgery. Control ascending aorta samples for RNA studies were obtained from eight organ donors without clinical or macroscopic signs of aortic atherosclerosis at University Central Hospital, Helsinki, Finland. The use of organ donor tissues was approved by The National Authority for Medicolegal Affairs of Finland.

AAA samples were immediately fixed in 4% formaldehyde for light microscopy or snap-frozen in liquid nitrogen for RNA isolation. All of the procedures were approved by the local ethics committees, and all of the patients gave written informed consent.

Real-Time PCR.

Total RNA from frozen samples was isolated with TRIzol (BRL-Life Technologies) and an RNEasy Mini Kit (Qiagen) after homogenization using FastPrep (MP Biomedicals) according to the manufacturer's instructions. The quality of RNA was analyzed with an Agilent 2100 bioanalyzer (Agilent Technologies) and quantity assessed by a NanoDrop. cDNA was synthesized from total RNA by using SuperScript II reverse transcriptase (Invitrogen), and 2 μl of cDNA were amplified by real-time PCR performed in a TaqMan 7300 instrument (Applied Biosystems). Normalization was made to cyclophilin A. The following primer/probe pairs were obtained by Assay-on-Demand (Applied Biosystems): 5-LO, FLAP, LTA₄H, and LTC₄S, with assay IDs Hs00386528_m1, Hs00233463_m1, Hs00168505_m1, and Hs00168529_m1, respectively. The PCR amplification was related to a standard curve.

Immunohistochemistry.

Paraffin sections were cleared in xylene, rehydrated in graded ethanol (100–70%), and subjected to antigen retrieval by boiling in DIVA buffer (Biocare Medical) for 20 min. After cooling to room temperature in water, sections were incubated in 3% hydrogen peroxide (H₂O₂) for 5 min to inhibit endogenous peroxidase activity. The specimens were then rinsed three times for 5 min each in PBS, blocked for 30 min in PBS with normal goat serum (1:5), and incubated overnight at 4 °C in a humidified chamber with and without primary antibodies against human 5-LO (a gift from Olof Rådmark, Karolinska Institutet, Stockholm, Sweden), human FLAP (a gift from Jilly Evans, Amira Pharmaceuticals, San Diego, CA), human LTA₄H and LTC₄S (prepared in house), human CD68 (Dako), human CD66b (Fitzgerald Industries), human CD3 (Santa Cruz Biotechnology), and Tryptase (Dako). The samples were then rinsed in PBS and incubated with biotinylated secondary antibody (1:1,500) for 1 h at room temperature followed by avidin–biotin amplification (ABC Elite) for 30 min, and developed with 3,3′-diamino-benzidine (Sigma). Sections were counterstained with Mayer’s hematoxylin for 1 min and mounted. Negative controls were obtained by substituting the primary antibodies with PBS.

Reverse-Phase HPLC Coupled to EIA.

Human AAA (n = 9) were collected immediately after surgery, dissected, weighed, homogenized in 0.1 M PBS (pH 7.8) containing proteinase inhibitor mixture (Roche) and EDTA (1 mM), and sonicated three times for 10 s on ice. Homogenates were incubated with 1 mM ATP, 2 mM Ca²⁺, and 40 μM arachidonic acid for 10 min at room temperature. The reaction was stopped by adding two volumes of cold methanol. Samples were centrifuged at 10,000 × g for 10 min at 4 °C, and the resulting supernatants were acidified to pH 3–4 and 5.6 for LTB₄ or cysLTs, respectively. Samples were purified by solid-phase extraction (Supelco LC18) and analyzed by reverse-phase HPLC. The column (Nova-Pak C18; Waters) was eluted with acetonitrile/methanol/water/acetic acid (30:35:35:0.01 by volume) at 0.8 mL/min, and absorbance was monitored at 270 nm or 280 nm for LTB₄ and cysLTs, respectively. Fractions corresponding to the retention times of LTB₄, LTC₄, LTD₄, and LTE₄ standards (Cayman Chemical) were collected, dried under nitrogen, and resuspended with EIA buffer. Levels of LTB₄ or cysLTs were determined in duplicate assays with an LTB₄ or cysLT EIA kit (Cayman Chemical) by using dilutions within the linear portion of the standard curves.

In selected experiments (n = 8), AAA wall tissue was homogenized in PBS (pH 7.4) 1:3/wt/vol and then centrifuged at 100,000 × g for 60 min at 4 °C to prepare cytosolic (supernatant) and membrane (pellet) fractions. The cytosolic fraction was incubated with 10 μM LTA₄ (Cayman Chemical) and 0.2% BSA for 10 min at room temperature, and the pellet was resuspended in PBS (initial volume) and incubated with 10 μM LTA₄, 5 mM glutathione, and 0.2% BSA for 10 min at room temperature. The reaction was stopped with two volumes of cold methanol. Samples were purified, extracted, and analyzed by HPLC as described above.

MMP Assays and Gel Zymography.

The aneurysm walls from selected surgeries (n = 4) were cut into pieces of ≈1 mm³ and divided equally into three separate samples. They were incubated in RPMI medium 1640 (Sigma) with LTD₄ (10 μM) with or without pretreatment with montelukast (Cayman Chemical; 1 μM, 30 min at 37 °C) for 24 h at 37 °C (2 mL/g wet tissue), as previously described (12, 14). The conditioned media containing the released material were collected and stored at −80 °C until use.

Zymography supplies were purchased from Invitrogen. ProMMP-2, MMP-2, proMMP-9, and active MMP-9 distributions were determined from conditioned media. Gelatin substrate zymograms were prepared using precast 10% SDS-polyacrylamide gels containing 1 mg/mL of gelatin. Equal volumes of experimental media samples were diluted into 2× Tris-glycine SDS sample buffer and electrophoretically separated under nonreducing conditions. Proteins were incubated in renaturating buffer (Invitrogen) for 30 min at room temperature. The gels were incubated overnight at 37 °C in developing buffer (Invitrogen). After 1 h staining with Coomassie brilliant blue and destaining for 2 d with 10% acetic acid and 40% methanol in water, gelatinase activity was evident by clear bands against a dark blue background. Quantification of the bands was performed using digital camera Fujifilm LAS-1000 and densitometry software ImageJ.