INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE PULMONARY ARTERY undergoes major structural reorganization during sustained elevation of pulmonary arterial pressure. As the pulmonary arterial wall enlarges during remodeling, vascular cells must migrate, requiring the breakdown of extracellular matrix barriers. Increased activity of serine proteases in pulmonary arteries has been observed during early remodeling in monocrotaline-induced pulmonary hypertension (40), suggesting serine proteases are involved in early remodeling. After removal of the hypertensive stimulus, the excess extracellular matrix proteins that have accumulated during exposure to hypoxia decrease toward normal levels (20). The decrement in content of highly insoluble collagen and elastin molecules in a relatively brief time suggests that matrix metalloproteinases (MMPs) play a role in the resorption of vascular collagen.

MMPs are a family of degradative enzymes expressed during periods of active remodeling, such as embryonic development, wound healing, and involution of tissues (38). Members of this family are secreted in a latent form and are cleaved by other proteases to lower-molecular-weight active forms. In addition, latent MMPs can be activated by nonproteolytic processes such as organomercurial compounds, plasmin, and reactive oxygen species. The counterregulatory tissue inhibitor of metalloproteinase (TIMP) modulates MMP activities, and the balance between MMP and TIMP is thought to determine overall turnover of matrix proteins.

There is considerable evidence that MMPs are important in the maintenance of the integrity and stability of systemic blood vessels (8). For example, mutations in TIMP-3 genes are thought to be the cause of a rare genetic form of macular degeneration (Sorsby's fundus dystrophy) in humans, and a stromelysin promoter variant is associated with progression of atherosclerosis (8). In atherosclerosis, MMPs colocalize with foamy macrophages within regions of atherosclerotic plaques, and cytokines, such as tumor necrosis factor-, that induce production of MMPs may be involved in initiating a cascade of events leading to plaque rupture and vascular occlusion. Mechanical injury to the wall of blood vessels in animal models induces MMPs, and administration of an MMP inhibitor substantially reduces the rate of migration of the smooth muscle cells, suggesting that MMPs play a role in restenosis. These observations suggest that inflammation plays a key role in modulating MMP expression in systemic vessels.

The aim of this study was to determine whether expression of interstitial collagenase, stromelysin-1, and gelatinases was increased during recovery from hypoxic pulmonary hypertension in the rat. We also analyzed the level of TIMP-1 expression to determine whether an imbalance between MMPs and their principal inhibitor occurs during regression of pulmonary artery remodeling. In the rat hypoxic model of pulmonary hypertension, little if any inflammation is demonstrable by ultrastructural examination of blood vessels (18). This model, therefore, provides an opportunity to study MMP expression in remodeling in the absence of vascular injury or inflammatory disease.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Outbred 6-wk-old male rats (n = 224) and adult male Hartley-strain guinea pigs (n = 4; Charles River Breeding Laboratories) were used. Animals were retained in the animal quarters for 1 wk and were fed standard animal chow and water ad libitum. Guinea pigs were killed by a lethal intraperitoneal injection of pentobarbital sodium, and dermal collagen was extracted and used as a substrate for collagenase. Institutional guidelines for animal use were followed.

Antibodies. Sheep anti-rabbit stromelysin-1 polyclonal antibody was provided by M. Lark (Merck Sharpe and Dohme Research Laboratories, Rahway, NJ), goat anti-mouse TIMP-1 polyclonal antibody by D. Denhardt (Rutgers University, Piscataway, NJ) (7), and rabbit anti-human ₅-subunit polyclonal antibody by S. Johansson (Uppsala University, Uppsala, Sweden) (4).

Nucleic acid probes. The probes were a 1,300-bp EcoR I-BamH I fragment of rat transin (the rat homologue of stromelysin-1) (pKSTREB) (17), a 700-bp EcoR I fragment of rat stromelysin-2 (5), a 1,250-bp EcoR I cDNA fragment of rat uterine smooth muscle collagenase 54 (21), a 700-bp Hind III-EcoR I fragment of mouse TIMP-1 cDNA (P TIMP-F) (10), and a 1,500-bp EcoR I fragment of human pro-₁(I) collagen cDNA (Hf677) (2).

Hypoxic exposure and hemodynamic measurements. Rats were exposed to 10% O₂-90% N₂ (hypoxia) at ambient pressure (15). Groups of animals were studied before exposure to hypoxia (day 0), and after exposure to hypoxia for 10 days (day 10) or to hypoxia for 10 days followed by recovery in air for 1, 3, 7, or 14 days (days 11, 13, 17, and 24, respectively). In addition, one group (n = 3) was exposed to hypoxia for 3 days and used to measure steady-state mRNA levels for pro-₁(I) collagen as a positive control for Northern blot analysis. Age-matched control animals breathed room air. Hypoxic animals were fed standard rat chow and given water ad libitum, and the amount of food was weighed. Control animals were fed the same amount as hypoxic animals to ensure similar final body weights. Mean right ventricular pressure (RVP) was measured in anesthetized rats injected intraperitoneally with 50 mg/kg of pentobarbital sodium; the ratio of weights of cardiac ventricles [right ventricle/(left ventricle + septum), RV/(LV+S)] and hematocrit were measured as previously described (15).

Light microscopy and thickness of pulmonary arterioles. Photomicrography of hematoxylin and eosin-stained tissue sections and measurements of pulmonary artery wall thickness (vessels with external diameters 110-180 µm) were performed as previously described (20).

Preparation of tissues. The main pulmonary artery, the entire left extrapulmonary artery, and the proximal 3 mm of the right extrapulmonary artery were excised en bloc. Individual tissues were used for all assays except Northern blot analysis and SDS-PAGE analysis, for which tissues were pooled (n = 5-6). For immunohistochemistry, lungs were perfused via the trachea with a 1:1 solution of OCT (Miles Scientific) and PBS, frozen in liquid nitrogen, and stored at 70°C.

Hydroxyproline and protein measurements. Pulmonary artery segments were homogenized and then hydrolyzed in 6 N HCl at 116°C for 48 h, and total protein and hydroxyproline contents were assayed as previously described (20).

Total protease activity. Degradation of [¹⁴C]carboxymethyl transferrin by pulmonary artery homogenates was used to assay for noncollagenolytic activity (35). The substrate was prepared by reaction of reduced bovine transferrin with [¹⁴C]acetic anhydride (108 mCi/mM; Amersham). Pulmonary arteries were homogenized (model PT 10/35, Brinkman Instruments) in transferrin buffer (0.1 M Tris · HCl-0.15 M NaCl, pH 7.5) and centrifuged at 9,000 g for 15 min at 4°C. The pellet was resuspended at 10 mg/ml in transferrin buffer and stored at 70°C. Tissue homogenates (20 µl) were incubated with 25 µl of [¹⁴C]carboxymethyl transferrin and 530 µl of buffer (0.1 M Tris · HCl, 0.15 M NaCl, 10 mM CaCl₂, and 5 mM MgCl₂, pH 7.5). The mixture was incubated at 37°C for 1 h, and 200 µl of 10% TCA were added, followed by centrifugation at 10,000 g for 15 min at 4°C. Radioactivity in the supernatant was counted by liquid scintillation spectrometry (model LS 6000IC, Beckman Instruments). To determine the classes of enzymes, homogenates from day 13 animals were incubated with EDTA (100 µM), 1,10-phenanthroline (100 µM), phenylmethylsulfonyl fluoride (PMSF; 100 µM), and 3,4-dichloroisocoumarin (100 µM; Sigma) (3).

Collagenase activity. Type I collagen extracted from dermis (12) was labeled with [¹⁴C]acetic anhydride (108 mCi/mM; Amersham) (11), digested with pepsin (25), and used as substrate [specific activity, 6.7 × 10⁴ counts · min^-1 (cpm) · mg¹]. Tissues were homogenized in 0.15 M NaCl and centrifuged at 6,000 g for 20 min at 4°C, and the pellet was washed in 0.15 M NaCl. The pellet was resuspended at 10 mg/ml in a 0.04 M Tris buffer with 0.15 M NaCl, 0.01 M CaCl₂, 250 µg/ml of streptomycin, and 200 U/ml of penicillin G, pH 7.5, frozen in liquid nitrogen, and stored at 70°C. Tissue homogenate (50 µl) was incubated with 50 µl of [¹⁴C]collagen, 2 mM 4-aminophenylmercuric acetate (Sigma), and 350 µl of buffer (0.04 M Tris, 0.15 M NaCl, 10 mM CaCl₂, and 5 mM MgCl₂, pH 7.5) for 48 h at 37°C. The mixture was centrifuged at 10,000 g for 10 min, 0.04 M phosphotungstic acid and 50 µl of 2 N HCl were added to the supernatant, and radioactivity was counted.

Gelatinolytic activity. The substrate was prepared by denaturing [¹⁴C]collagen in boiling water for 15 min (19). Tissue homogenate (50 µl), prepared as described in Preparation of tissues, was incubated with 50 µl of [¹⁴C]gelatin and 350 µl of buffer (0.04 M Tris, 0.15 M NaCl, 10 mM CaCl₂, and 5 mM MgCl₂, pH 7.5) for 48 h at 37°C. After precipitation with 12% (wt/vol) TCA, the radioactivity in the supernatant was counted.

Gelatin zymography. Tissues were homogenized in NP-40 lysis buffer (0.1 M Tris, 1% NP-40, and 0.15 M NaCl, pH 8.0) containing 5 mM EDTA and 2 mM PMSF. Homogenates were incubated for 1 h at 4°C, centrifuged at 27,000 g for 20 min, and stored at 70°C. Extracts were diluted with 2% SDS, 0.15% glycerol, 0.25 M Tris · HCl (pH 6.8), and 0.1% bromphenol blue (19), and protein content was assayed (BCA protein assay kit, Pierce Chemical). Samples were loaded without boiling onto a 12% polyacrylamide gel copolymerized with 0.1% gelatin and separated by electrophoresis at 4°C (13). The gels, incubated for 60 min in 2.5% Triton X-100 and then for 24 h at 37°C in 50 mM Tris · HCl (pH 8.0) containing 5 mM CaCl₂ to allow for gelatin digestion, were stained in 0.25% Coomassie blue containing 50% methanol-10% acetic acid followed by destaining in 10% methanol-10% acetic acid. Enzymatic activity in the gel was visualized as negative staining, and enzyme sizes were referenced to molecular-weight markers.

Western blot analysis. Pulmonary artery extracts, prepared as described for gelatin zymography, were denatured in Laemmli's buffer (16) containing -mercaptoethanol for 10-15 min and subjected to SDS-PAGE with a 12.5% polyacrylamide gel. The protein was blotted onto polyvinylidene fluoride membranes (Millipore) with a semidry transfer unit (Hoefer Scientific Instruments) and buffer (48 mM Tris, 39 mM glycine, 0.375 mM SDS, and 20 mM methanol) for 60 min at 80 mA. Equal loading was based on comparison with ₅-integrin (4), a protein that did not change during pulmonary hypertension. The membranes were blocked for 1 h with a 4% solution of dry milk powder in 0.2% Tween 20-PBS. The polyvinylidene fluoride membranes were incubated overnight at 4°C with nonimmune sera or immune sera: anti-rabbit stromelysin-1 or anti-mouse TIMP-1 antibodies. Membranes were washed, incubated with ¹²⁵I-protein A (0.1 µCi/ml) or ¹²⁵I-protein G (0.1 µCi/ml; ICN Biomedicals) for 1 h, washed extensively, air-dried, and exposed to X-AR-5 photographic film (Kodak) for 24 h. Molecular weights were compared with size markers. Density of bands was scanned with an imaging detector (model GS-670, Bio-Rad Laboratories).

RNA analysis. Total RNA was extracted by the guanidine isothiocyanate method, and Northern blot analysis was performed (20) with 20 µg or more of total RNA. Nitrocellulose filters (Schleicher & Schuell) were hybridized to nick translated cDNA probes (24) labeled with [³²P]dCTP (3,000 Ci/mM; ICN Biomedicals) to a specific activity of ~1 × 10⁸ cpm/µg. As a positive control, a probe for pro-₁(I) collagen was used (2), which has been shown to increase at 3 days of hypoxia compared with control pulmonary arteries (20). Southern blot analysis (27) was performed with the ³²P-labeled cDNA probes mentioned above.

RT-PCR. Total RNA (50 µg) was used to synthesize single-stranded cDNA using 750 units of Moloney mouse leukemia virus reverse transcriptase (GIBCO) and 50-100 pM of oligo(dT)_12-18 (Pharmacia LKB Biotechnology). Single-stranded cDNA was used for RT-PCR and primed with two oligonucleotide primers, 5'-CCGTCCAGAAGATCGATGCA-3' and 5'-CCATCTACACAGAGACAGTT-3', complementary to the rat transin cDNA (15), and 5'-ACCACCTTATACCAGCGTTA-3' and 5'-AAACAGGGAAACACTGTGCA-3', complementary to mouse TIMP-1 cDNA (10). Amplification was performed in a 100-µl reaction mixture (50 mM NaCl, 10 mM KCl, 10 mM Tris · HCl, 1.5 mM MgCl₂, 3 mM dithiothreitol, gelatin, and 200 µM each nucleotide 5'-triphosphate, pH 8.8) containing 2.5 U of Taq DNA polymerase (Perkin-Elmer). The reaction was carried out with a DNA thermal cycler (model 480, Perkin-Elmer Cetus) for 30 cycles as follows: denaturation at 94°C, 1.5 min; annealing at 54°C, 1 min; extension at 75°C, 1.5 min; and final extension, 10 min. The RT-PCR products were electrophoresed on a 1.5% agarose gel, DNA fragments were blotted onto nitrocellulose, and Southern blot analysis was performed.

Immunohistochemistry. Frozen sections of peripheral lung tissue (6 µm thick) were cut at 17°C on a cryostat microtome (model 855, American Optical Reichert Scientific Institute) and mounted on gelatinized slides that were air-dried and stored desiccated at 70°C. Tissue sections were warmed to room temperature, hydrated in a humidified chamber for 2 h at 4°C, OCT was removed, and tissues were fixed in 75% ethanol for 5 min. After PBS rinses, 1% BSA in PBS with 0.05% azide as a blocking agent was applied for 2 h at room temperature or overnight at 4°C. The blocking agent was removed, and sheep anti-rabbit stromelysin-1 (1:50 to 1:5,000 dilutions) was applied overnight at 4°C. The primary antibodies were removed by aspiration, and the slides were thoroughly washed with PBS. The secondary antibody, rhodamine-conjugated rabbit anti-sheep IgG (1:1,000 dilution; Cappel) with PBS, was applied for 2 h at room temperature or overnight at 4°C. Tissues were washed thoroughly in PBS, mounted in gelvatol (Air Products and Chemicals) with N-propyl galate, and photographed with T-Max 400 film (Kodak) on a microscope (Labophot Fluorescence Microscope, Nikon) equipped with an epifluorescence attachment.

To evaluate the specificity of the primary antibody, five experiments were performed on pulmonary artery tissue sections obtained from day 13 animals. First, rabbit stromelysin-1 antibody (1:50) was incubated with rat stromelysin-1 in PBS for 30 min at 37°C, held overnight at 4°C, and centrifuged at 1,500 g for 10 min. Tissues incubated with the supernatants containing the absorbed antibody showed weak, nonspecific fluorescence, indicating that the antibody was immunologically specific for rat stromelysin-1. Second, tissues were stained with anti-mouse desmin, and immunoreactivity was observed only in vascular smooth muscle cells, demonstrating specificity of unrelated antibodies. Third, tissues were evaluated for autofluorescence before their reaction with secondary antibody, and none was observed. Fourth, fluorescence with the conjugated secondary antibody was examined. Fifth, tissues were reacted with an appropriate nonimmune serum followed by conjugated secondary antibody. Weak, nonspecific fluorescence was observed in the latter two experiments.

Statistical analysis. ² Analysis with Yates' correction was used to assess nonparametric data (28). Data were analyzed with one-way ANOVA (28) followed by Duncan's post hoc test (9). A P value of <0.05 was considered significant.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Animal, hemodynamic, and biochemical results. Animal survival was 100% (33/33) in control group and was 98% (188/191) in the hypoxic group [² = 1.34, not significant (NS)]. Body weights were not significantly different between hypoxic animals and their age-matched controls on any day. Mean RVP increased twofold on day 10 compared with day 0, decreased by day 13 compared with day 10, and was not different from that of control levels on day 24 (Fig. 1A). The RV/(LV+S) increased by day 10, was decreased on day 13 compared with day 10, but was above control levels on day 24 (Fig. 1B). Hematocrit increased during hypoxia, decreased during recovery, but remained elevated on day 24 (Fig. 1C). The medial walls of muscular pulmonary arteries were thickened 2.5-fold on day 10, decreased on day 13 compared with day 10, and were at the control level on day 24 (Fig. 1D). Hydroxyproline and protein contents of main pulmonary arteries were ~75 and 150%, respectively, greater than control levels on day 10; both decreased on day 13 compared with day 10, and both were at control levels on day 17 (Fig. 1, E and F).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Measurements of pulmonary hypertension and vascular remodeling. Rats were exposed to 10% O2 (hypoxia) at ambient pressure for 10 days or were exposed to hypoxia for 10 days and allowed to recover in air for 1, 3, 7, or 14 days (days 11, 13, 17, and 24, respectively). A: mean right ventricular pressure (RVP). B: ratio of weight of right ventricle to that of left ventricle plus septum [RV/(LV+S)]. C: hematocrit. D: medial wall thickness of muscular pulmonary arteries. E: hydroxyproline content. F: protein content. Data are means ± SE; n = 4-6 animals except for D where n = 18 and 4-5 arteries/animal were measured. * P < 0.05 compared with day 0.  P < 0.05 compared with day 10.

Proteolytic activity. Degradation of [¹⁴C]carboxymethyl transferrin was at control level on days 10 and 11, was elevated 3.5-fold on day 13, and was at control level on day 24 (Fig. 2A). Protease activity on day 13 was inhibited ~64% by 100 µM EDTA and ~47% by 100 µM 1,10-phenanthroline, both MMP inhibitors (3), and ~15% by 100 µM PMSF and ~15% by 100 µM 3,4-dichloroisocoumarin, both serine protease inhibitors (3) (all P < 0.05, n = 5-6 pulmonary arteries). These results indicate that MMPs are the predominant proteases in day 13 pulmonary arteries. Collagenolytic activity was at control levels on day 10, increased fivefold at day 13 (P < 0.05, n = 5-6 pulmonary arteries), decreased on day 17 compared with day 10, and was at control levels on day 24 (Fig. 2B). Gelatinolytic activity was at control levels on day 10 compared with day 0, increased on days 11, 13, and 17, and was at control level on day 24 (Fig. 2C). Collagenolytic and gelatinolytic activities were completely inhibited by 100 µM EDTA (data not shown). On zymographic analysis, two bands of gelatinolytic activities were observed at ~68 kDa and faintly at ~60 kDa on day 0 (Fig. 3). Three bands of ~72, 68, and 60 kDa were observed on days 10, 11, 13, and 17, with the greatest density of bands on day 11 (Fig. 3). The density of the bands appeared to decrease on day 17 compared with day 11.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Proteolytic activity in pulmonary arteries. Conditions are as described in Fig. 1. A: total proteolytic activity. B: collagenolytic activity. C: gelatinolytic activity. Collagenolytic activity was not measured on day 11. Data are means ± SE; n = 5-6. * P < 0.05 compared with day 0.  P < 0.05 compared with day 13.

View larger version (96K): [in this window] [in a new window]   Fig. 3.   Gelatinase activity in pulmonary artery tissue. Homogenates of pulmonary artery tissues (15 µg of protein) were separated on an acrylamide gel copolymerized with gelatin and examined for gelatinolytic activity. Equal loading was checked by measuring protein levels before gel was loaded. Lanes indicate days of exposure to hypoxia or recovery from hypoxia. Presence of ~72-, 68-, and 60-kDa band was faint on day 0 and was increased at 10 days of hypoxia and 1, 3, and 7 days of recovery.

Stromelysin and TIMP-1. An ~57-kDa stromelysin-1 protein, expressed at all times, was increased on day 13 compared with day 10 and decreased by day 24 (Fig. 4A). Scanning densitometer readings indicated a twofold increase in stromelysin-1 protein levels on day 13 compared with all other days (P < 0.05; n = 5 pulmonary arteries). A constitutively expressed ~28.5-kDa TIMP-1 protein was faintly visible on day 0. The signal appeared greater on days 10, 13, and 17 than on day 0 and appeared to decrease by day 24 (Fig. 4B). Equal loading of samples was demonstrated by unchanged levels of ₅-integrin protein throughout the hypoxic and recovery periods (Fig. 4C). Densitometry readings for TIMP-1 showed a threefold increase in TIMP-1 levels on day 13 compared with day 10 (P < 0.05; n = 6 pulmonary arteries). Densitometry readings were consistent for stromelysin-1 (n = 4 tissues) and TIMP-1 (n = 6 tissues). Hybridization signals for rat stromelysin-1 (transin), stromelysin-2, and TIMP-1 mRNAs were not detectable under low-stringency conditions by Northern blot analysis (Fig. 5). This was not an artifact because pro-₁(I) collagen mRNA was detected on day 0 and increased on day 3 of hypoxia with the same method (Fig. 5). No stromelysin-1 or TIMP-1 products were detected by RT-PCR analysis after 30 cycles (not shown).

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Autoradiograms of immunoblots for antibodies to stromelysin-1 and tissue inhibitor of metalloprotease (TIMP)-1. Homogenates of pulmonary artery tissue (20 µg of protein) were examined during hypoxia and recovery from hypoxia. A: stromelysin-1. Left: ~57-kDa stromelysin-1 protein (arrow). Right: corresponding densitometer measurement of stromelysin-1 protein indicating expression at all times, with increased expression on day 13. B: TIMP-1. Left: ~28.5-kDa TIMP-1 protein (arrow) was expressed at all times. Right: corresponding densitometer measurements of TIMP-1 protein indicating apparent increased expression on days 10 and 13 and reduction toward control by day 24. C: 150-kDa 5-integrin protein control showed no change during hypoxia and recovery. Immunoblots are representative of n = 4-5 blots for each protein, showing same results.

View larger version (76K): [in this window] [in a new window]   Fig. 5.   Autoradiogram of Northern hybridizations to rat RNA. Total RNA (10 µg) was applied to filters and hybridized with radiolabeled probes. Lane 1, day 0 (control) and lane 2, 3-day hypoxic rats showing 5.3- and 4.7-kb bands for pro-1(I) collagen (arrows). Lane 3, day 13 rat transin (rat homologue of stromelysin-1). Lane 4, day 13 rat stromelysin-2. Lane 5, day 13 TIMP-1. No hybridization signals are seen in lanes 3-5.

Immunohistochemistry. Immunoreactive stromelysin-1 was detected predominantly in the luminal portion of the vessel wall of muscular pulmonary arteries on day 0 (Fig. 6A). The intensity of the stain appeared slightly increased on days 10, 11, and 13 compared with days 0 and 17 (Fig. 6, B-E). An accompanying hematoxylin and eosin-stained section of a day 17 muscular pulmonary artery is shown (Fig. 6F) for orientation. In separate control experiments, tissues were incubated with the following: rabbit stromelysin-1 antibody absorbed with rat stromelysin-1, PBS in place of the primary antibody followed by incubation with the conjugated secondary antibody, and nonimmune serum followed by the conjugated secondary antibody. For all control experiments, weak, nonspecific fluorescence was observed on days 0, 10, and 13 (Fig. 6, G-I).

View larger version (96K): [in this window] [in a new window]   Fig. 6.   Immunolocalization of stromelysin-1 in muscular pulmonary arterioles. A: control; B: 10-day hypoxia; C: day 11; D: day 13; E: day 17; F: hematoxylin and eosin-stained section as reference. Arrow, muscular pulmonary vessel. G-I: immunofluorescence of stromelysin-1 after incubation of tissues with PBS in place of primary antibody followed by secondary antibody. G: day 0; H: day 10; I: day 13. Pulmonary arteries were incubated with PBS followed by rhodamine-conjugated rabbit anti-sheep IgG. Note weak, nonspecific staining at these times. d, Day; av, alveolus; aw, airway; bv, blood vessel. Bar in E, 20 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We observed a rapid decrease in collagen content of main pulmonary arteries of rats during the first several days after removal from a hypoxic environment. Transient increases in total proteolytic, collagenolytic, and gelatinolytic activities and increased expression of stromelysin protein and the ~68-kDa gelatinase were noted 1-3 days after return to normoxia. A temporal correlation between the expression of MMPs and rapid decrease in vascular collagen content suggests an association between collagen resorption and MMP activity. The relative proportion of collagen to noncollagen proteins remained constant throughout remodeling and regression. This observation suggests that the rates of turnover of cells and matrix proteins occurred proportionately, consistent with "physiological" resorption of collagen. The processes controlling removal of excess collagen in remodeled arteries may be analogous to the involutional loss of collagen in the postpartum uterus, which is mediated by marked increases in levels of MMP activities (38).

We studied in greatest detail proteases expressed on day 13, the time of peak total proteolytic and collagenolytic activities as well as of stromelysin protein expression. Total proteolytic activity on day 13 was reduced 47-64% by MMP inhibitors, and ~15% of total activity was blocked by serine protease inhibitors, suggesting that MMPs are the predominant proteases expressed on day 13. Interstitial collagenase was observed to increase fivefold on day 13. In contrast, gelatinolytic activity and expression of the ~72-, 68-, and 60-kDa gelatinases by zymography peaked on day 11. The reason for the earlier expression of gelatinase is not known, but we speculate that degradation of basement membranes could be occurring before breakdown of interstitial collagens. Gelatinase degrades basement membrane components (collagen types IV, V, VII, and X and fibronectin) as well as partially degraded collagen (38). The stimulus for basement membrane degradation could be changes in tissue structure, perhaps caused by a decrease in smooth muscle cell hypertrophy or cell death during involutional loss of cell mass during regression of pulmonary hypertension. As discussed below, stromal expression of MMPs may be a consequence of changes in cell shape or stimulation by oxygen-derived free radicals produced in blood vessels during the transition from hypoxia to normoxia. Further work is needed to determine whether breakdown of basement membranes or other structural components occurs before degradation of interstitial collagens to explain the early appearance of gelatinase.

Stromelysin-1, an MMP with broad substrate specificity (38), was studied using molecular, biochemical, and immunohistochemical methods. In homogenates prepared from main pulmonary arteries, we observed constitutive expression of stromelysin-1 by immunoblot and a transient twofold increased level above control on day 13. The location of immunoreactive stromelysin-1 appeared to be mainly on the luminal portion of the muscular pulmonary arteries, but the resolution of the photomicrographs was not sufficient to determine the cellular source(s) of stromelysin-1. Procollagenase, in contrast, is predominantly localized in the media and adventitia within connective tissue-type mast cells, which are abundant in remodeled pulmonary arteries of rats (33). We observed differences in the temporal expression of stromelysin-1 by immunoblot, which showed an increase only on day 13, and by immunohistochemistry, which showed increased staining on days 10, 11, and 13. It is possible that central and distal pulmonary arteries express stromelysin-1 differently in response to a hypertensive stimulus. It is known that differences in the pattern of structural remodeling in central and peripheral arteries occur, which is probably attributed to the composition of extracellular matrix proteins, populations of cells in the vessel walls, and differences in cellular and molecular responses to the hypertensive stimulus (14, 31). It is possible, for example, that a subset of smooth muscle cells capable of expressing stromelysin-1 is present during active remodeling in small branches. Alternatively, different hemodynamic stresses might signal the expression of stromelysin-1 in the two types of vessels. A systematic study contrasting MMP expression in small and large arteries is needed to determine the significance of these observations.

Stromelysin-1 mRNA was not detected in pulmonary artery tissue by Northern blot analysis under low-stringency conditions. Stromelysin-1 RT-PCR products were not visualized after 30 cycles. These observations suggest that posttranscriptional mechanisms are more likely to contribute to high levels of stromelysin-1 protein observed during the posthypoxic period. These posttranscriptional mechanisms may involve increased translation of stromelysin-1 protein or stromelysin-1 derived from cells migrating into the artery during remodeling.

We examined the changes in TIMP-1 protein in regression of pulmonary artery remodeling. TIMP-1 was expressed at low levels in normotensive pulmonary arteries but appeared to increase during the late hypertensive (day 10) and recovery phases compared with control pulmonary arteries. The level of TIMP-1 appeared to be greatest on day 13, the time of peak proteolytic activities and greatest expression of MMPs. It is difficult to draw conclusions about these changes in protein levels on TIMP-1-MMP balance in tissue because of sequestration of enzyme, kinetics of enzyme-inhibitor binding, and turnover rates of both components.

Several possible mechanisms may be responsible for activation of MMPs in pulmonary arteries. First, the change in tissue oxygen tension during the transition from hypoxia to normoxia may have activated latent MMPs and induced proteolytic activity. Various reactive oxygen species, including hypochlorous acid (HOCl), hydrogen peroxide (H₂O₂), and hydroxyl radical generated by hypoxanthine/xanthine oxidase (X/XO), activate isolated latent procollagenase, and their activation potentials are comparable to other nonproteolytic activators (26). Neutrophil procollagenase and progelatinase are activated by HOCl (36), an effect that is enhanced by addition of cathepsin G (6). Incubation of human vascular smooth muscle cell culture medium containing latent progelatinase with X/XO resulted in activation of progelatinase (22). Noninflammatory cells, such as UMR-106 osteosarcoma cells (23) and Walker 256 carcinosarcoma cells (29), secrete latent MMPs that are activated by addition of HOCl and H₂O₂, respectively. It is possible that oxygen free radicals were generated during the increase in tissue oxygen tension that occurred after the rats were removed from the hypoxic environment, analogous to the generation of oxygen free radicals during ischemia-reperfusion. It has been speculated that the mechanism of nonproteolytic activation of procollagenase involves a conformational change in the enzyme molecule that disrupts a cysteine-zinc atom interaction and frees the zinc atom to participate in proteolytic reactions (30). It is possible that oxygen free radical generation is an initiating event in induction of MMP activity during the posthypoxic period.

A second possible mechanism modulating MMP expression may involve changes in mechanical forces on cells after the reduction in blood pressure in the posthypoxic period. Werb and colleagues (1, 34, 37) observed that alterations in cell shape, actin cytoskeleton, and cellular interactions with integrins can induce production of stromelysin and interstitial collagenase. These results suggest that expression of MMPs may be under the control of physical changes in tissue architecture. In tissues, proteolysis is stimulated by removing a distending force from the body of a gravid rat uterus (39). Experiments from our laboratory (32) suggest that release of static mechanical tension from isolated pulmonary arteries induces proteolysis and expression of MMPs. In main pulmonary arteries, alterations of MMP expression may be a consequence of physical forces when blood pressure is reduced after return to normoxia.

In conclusion, our results show increased proteolysis in main pulmonary arteries of rats that is restricted to the period of rapid reduction in collagen content after reversal of hypoxic pulmonary hypertension. The observation that temporally controlled proteolysis occurs in vessels in the absence of major inflammatory changes suggests that noninflammatory mechanisms may control endogenous MMP expression in pulmonary arteries. Future experiments are needed to establish whether a causal relationship exists between collagen resorption and MMP activity and whether proteolysis in pulmonary arteries plays a role in regulation of pulmonary hemodynamics.