Expression of Stromelysin-3 in Atherosclerotic Lesions: Regulation via CD40–CD40 Ligand Signaling In Vitro and In Vivo

Human recombinant IL-1β, TNFα, and IFNγ were obtained from Endogen. Phorbol-12 myristate 13-acetate (PMA) and polymyxin B were purchased from Sigma Chemical Co. Human recombinant CD40L (rCD40L) was generated as described previously (44) and the mouse anti–human stromelysin-3 antibody 5ST-4A9 was produced in a program sponsored by Bristol Myers Squibb and is subject of an issued U.S. utility patent number 5484726 (45). Experiments employing rCD40L were performed in the presence of polymyxin B. Anti-CD40L, a rat mAb IgG2 antibody raised against mouse CD40L was prepared as described (46) and provided by Immunex Corp. Rat IgG salt-free crystalline powder (Sigma Chemical Co.), reconstituted in pyrogen-free normal saline and sterile filtered, was obtained from Immunex Corp. Both rat anti–mouse CD40L antibody and rat IgG contained <2 pg/μl of endotoxin. Anti–human CD40 as well as control IgG1 mAb (FITC conjugated) used for immunohistochemistry were obtained from PharMingen. Low density lipoprotein receptor–deficient (LDLR^−/−) mice (B6/129LDRr-tm1Her) were obtained from Jackson Laboratory.

Human vascular EC were isolated from saphenous veins by collagenase treatment (1 mg/ml; Worthington Biochemicals) and cultured in dishes coated with fibronectin (1.5 μg/cm²; New York Blood Center Reagents). Cells were maintained in medium 199 (M199; BioWhittaker) supplemented with 1% penicillin/streptomycin (BioWhittaker), 5% fetal bovine serum (FBS) (Atlanta Biologicals), 50 μg/ml heparin (Sigma Chemical Co.), and endothelial cell growth factor (Pel-Freez Biological). SMC were isolated from human saphenous veins by explant outgrowth (47) and cultured in DMEM (BioWhittaker) supplemented with 1% l-glutamine (BioWhittaker), 1% penicillin/streptomycin, and 10% FBS. Both cell types were subcultured following trypsinization (0.5% trypsin [Worthington Biochemicals]/0.2% EDTA [EM Science]) in 75 cm² culture flasks (Becton Dickinson) and used throughout passages two to four. Culture media and FBS contained <40 pg endotoxin/ml as determined by chromogenic Limulus amoebocyte assay analysis (QLC-1000; BioWhittaker). EC and SMC were characterized by immunostaining with anti von Willebrand factor and anti SMC α-actin antibody (Dako), respectively. Both cell types were cultured 24 h before the experiment in media lacking FBS: vascular EC were cultured in M199 supplemented with 0.1% human serum albumin and vascular SMC cultured in insulin/transferrin medium as described previously (48).

Mononuclear phagocytes were isolated by density gradient centrifugation (49), using lymphocyte separation medium (Organon-Teknika), and subsequent counterflow elutriation from freshly prepared human PBMCs obtained from leukopacs of healthy donors (provided by Steve K. Clinton, Dana-Farber Cancer Institute, Boston, MA). Mononuclear phagocytes were used directly (monocytes) for the experiments or cultured for 1, 3, or 9 d (MØ) in RPMI 1640 containing 2% human serum (Sigma Chemical Co.). The purity of monocytes/MØ was ≥96%, as determined by FACS^® analysis (anti–human CD68 mAb FITC; PharMingen). For certain studies, MØ were stimulated in RPMI 1640 lacking serum.

Freshly isolated human CD4⁺ T cells were a gift from Dr. F.W. Luscinskas (Brigham and Women's Hospital, Boston, MA). The purity of CD4⁺ T cells was ≥98%, as determined by FACS^® analysis (anti–human CD4 mAb FITC; Calbiochem). Human CD4⁺ T cells were activated with PMA (50 ng/ml, 12 h) and CD40L cell surface expression confirmed by FACS^® analysis using FITC-labeled anti–human CD40L mAb (Calbiochem). Finally, activated CD4⁺ T cell membranes were prepared as described previously (42) and used for stimulation at a ratio of T cells to SMC/EC/MØ of 10:1.

Surgical specimens of human carotid atheroma and aorta were obtained by protocols approved by the Human Investigation Review Committee at the Brigham and Women's Hospital. Serial cryostat sections (5 μm) were cut, air dried onto microscope slides (Fisher Scientific), and fixed in acetone at −20°C for 5 min. Sections were preincubated with PBS containing 0.3% hydrogen peroxidase activity. The sections were then incubated (90 min) with primary or control (mouse myeloma protein MOPC-21; Sigma Chemical Co.) antibody diluted in PBS supplemented with 5% appropriate serum. After washing three times in PBS, sections were incubated with the respective biotinylated secondary antibody (45 min; Vector) followed by avidin–biotin–peroxidase complex (Vectastain ABC kit; Vector), and antibody binding was visualized with 3-amino-9-ethyl carbazole (Vector) according to the recommendations provided by the supplier. For colocalization of stromelysin-3 with CD40 or the respective cell type, double immunofluorescence staining was performed. The anti–human stromelysin-3 mAb (1:200) was applied for 60 min followed by biotinylated anti–mouse secondary antibody for 45 min and Texas red–conjugated streptavidin (Amersham). Subsequent to application of the avidin–biotin blocking kit (Vector), rabbit–anti human CD40 antibody (1:250; Santa Cruz), anti-muscle actin mAb for SMC (Enzo Diagnostics), anti-CD31 mAb for EC (1:400, Dako), or anti-CD68 mAb for MØ (1:600; Dako) were added and sections incubated overnight at 4°C. Subsequently, the appropriate secondary antibodies were applied for 30 min followed by Streptavidin–FITC (Amersham). To detect stromelysin-3 in mouse atherosclerotic lesions, the stromelysin-3 mAb and a specific anti– mouse absorbed rat biotinylated secondary antibody (1:500), followed by avidin–biotin–peroxidase complex, were applied.

Frozen tissue from five nonatherosclerotic arteries and seven atheromatous carotid plaques were homogenized (IKA-Labortechnik, Ultra-turrax T 25) and lysed (0.3 mg tissue/ml lysis buffer) as described previously (5). The lysates were clarified (16,000 g, 15 min) and the protein concentration for each tissue extract as well as for the cell culture samples was determined using a bicinchoninic acid protein assay according to the instructions of the manufacturer (Pierce). 50 μg total protein were applied to Western blot analysis.

In situ hybridization was performed according to the instructions of the manufacturer (Hyb-Probe™; Shandon/Lipshaw). Frozen tissue sections, obtained as described above, were fixed in cold acetone, air-dried, and incubated with a mixture of FITC labeled stromelysin-3–specific (5′-GGTACCGTCAACCAGGTCCTCGTCCACG-3′; 5′-CTCAGAGTCGGGTCTACTGACCGTCC-3′; 5′-CCTACTGGTCCCGTGTCTGGACGACGTCCA-3′; 5′-ACGGTCCGGTGCTTATAGTCC-GATCTCCTGG-3′), CD40L-specific (5′-TTATGGGTGTCAA-GGCGGTTTGGAACGCCCGTT-3′; 5′-TGAAAAACGACA-CATAGAAGTATCTTCCAA-3′; 5′-TACTAGCTTTGTATGTTGGTTTGAAGAGGGGCT-3′; 5′-TGCAGGAAACCGAATGAGTTTGAGACTTGT-3′), or random oligomers in hybridization buffer (30% formamide, 0.6 M NaCl₂, 10% dextran sulfate, 50 mM Tris [pH 7.5], 0.1% sodiumpyrophosphate, 0.2% Ficoll, 5 mM EDTA) for 10 min at 65°C and subsequently for 2 h at 37°C in a moist chamber. Thereafter, slides were immersed in TBS/ Triton (50 mM Tris, 150 mM NaCl, pH 7.6/0.1% Triton X-100) to allow coverslips to float off and were washed (TBS/ Triton) three times for 3 min at 37°C. The slides were forwarded to the immunological reaction by incubation with blocking solution (10 min, rt) and subsequent addition of the alkaline phosphatase–conjugated rabbit Fab′ anti FITC (30 min, rt, moist chamber) as the primary detection reagent. Finally, the slides were washed twice in TBS (3 min, rt), covered with alkaline phosphatase substrate buffer (5 min, rt), and developed using the NBT/BCIP (nitro-blue-tetrazolium/5-bromo-4-chloro-3-indolyl phosphate) chromogen solution (1–2 h, rt, moist chamber).

Cell extracts (25 μg total protein/lane) and culture supernatants were separated by standard SDS-PAGE under reducing conditions and blotted to polyvinylidene difluoride membranes (Bio-Rad) using a semidry blotting apparatus (0.8 mA/cm², 30 min; Bio-Rad). Blots were blocked and first and second monoclonal antibodies were diluted in 5% defatted dry milk/PBS/0.1% Tween 20. After 1 h of incubation with the respective primary antibody, blots were washed three times (PBS/0.1% Tween 20) and the secondary, peroxidase–conjugated, goat anti–mouse antibody (Jackson ImmunoResearch) was added for another hour. Finally, the blots were washed (20 min, PBS/0.1% Tween 20) and immunoreactive proteins were visualized using the Western blot chemiluminescence system (NEN™).

For radioimmunoprecipitation experiments, cells were washed and further incubated with unlabeled medium lacking methionine/cysteine. Subsequently, medium containing 10% FBS and 50 mCi/ml [³⁵S]methionine/cysteine (NEN™) was added to the cells for 24 h. Supernatants were harvested and concentrated (10×) using Centricon 3 devices (Amicon). Immunoprecipitation buffer (50 mM Tris-HCl, 0.1% SDS, 0.1% sodium deoxycholate, 1% NP-40, 150 mM NaCl, 5 mM EDTA, 20 μg/ml soybean trypsin inhibitor, 0.1% mM PMSF, 0.2 U/ml aprotinin, 0.025% sodium azide) was added to the cultures and cells were harvested by scraping. Subsequently, non immune mouse serum (Vector) was added (24 h, 4°C) to preclear the samples. Antigens in supernatants and cell extracts were immunoprecipitated with the specific anti stromelysin-3 antibody 5ST-4A9 (2 h, 4°C) and pelleted by subsequent addition of rabbit anti–mouse IgG (18 h, 4°C) as well as protein A–Sepharose beads (2 h, 4°C). The beads were washed four times in 50 mM Tris-HCl and finally 50 μl SDS-PAGE loading buffer (200 mmol/liter Tris, 5% glycerol, 0.1% SDS, 3% β-mercaptoethanol, 0.1 mg/ml bromophenol blue) was added. After heating the samples 5 min at 95°C, supernatants were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. Blots were dried and exposed to x-ray film for detection of the immunoprecipitated antigen.

Total RNA from unstimulated or stimulated (24 h) EC, SMC, or MØ was isolated by a one-step preparation according to the method of Chomczynski and Sacchi (50). The cDNA was prepared by reverse transcription (RT) of total RNA (1 μg) with oligo(dT) using superscript reverse transcriptase (GIBCO BRL). The RT products were diluted in 480 μl twice distilled water and 10 μl of these cDNA preparations were mixed on ice with 10 μl primers (20 μM), 80 μl reaction mix (including 10 ml of PCR buffer, 2.5 mM MgCl), 4 μl dNTPs (200 μM; final concentration), and 0.5 μl Taq-polymerase (2.5 U; all GIBCO BRL). PCR was performed for 35 cycles at 95°C (120 s), 62°C (120 s), and 72°C (180 s, 2 s prolongation per cycle) after hot start. The sequences of primers for human stromelysin-3 were 5′-CTGCAGTCATCTGGGCTGAGACTC-3′ and 5′-CCATGGCAGTTGGTGCAGGAGCAG-3′. The primers were obtained from Integrated DNA Technologies. Aliquots of the PCR products were run on 1.3% agarose gels and visualized by UV transillumination.

LDLR^−/− at a minimum age of 8 wk were fed a high cholesterol diet (1.25% cholesterol, 0% cholate; Research Diets) and treated for a period of 12 wk with either rat IgG (control) or rat anti– mouse CD40L (M158) antibody (both at 250 μg per mouse twice a week intraperitoneal; both provided by Immunex Corp.) (51). Thereafter, the mice (10 per group) were killed, the aortic arches embedded, and tissue sections examined by immunohistochemistry as described below. The extent of lesion formation and the expression of vascular adhesion molecule-1 in these animals has been reported separately (51). No difference in the serum lipid profiles (total cholesterol, very low density lipoprotein, low density lipoprotein, and high density lipoprotein [analyzed by fast protein liquid chromatography] or triglycerides) or in circulating leukocytes, hematocrit, or body weight was observed among the groups.

Immunohistochemical analysis of the expression of stromelysin-3 in normal aortic specimens (n = 5; Fig. 1 A, left) and human atherosclerotic fatty streaks (n = 5; data not shown) revealed little or no expression of the enzyme. In contrast, well-developed human carotid atherosclerotic lesions (n = 7) consistently showed strong stromelysin-3 immunoreactivity, most prominently at the luminal border and in the shoulder region of the plaque (Fig. 1 A, right). Western Blot analysis, performed on protein extracts of the surgical specimens and using the identical antibody used for the immunohistochemistry studies, revealed barely detectable immunoreactive stromelysin-3 in control specimens but markedly increased levels of the proteinase in atherosclerotic tissue (Fig. 1 B). The immunoreactive bands detected had molecular masses of ∼64, 48, 35, and 28 kD, corresponding to the zymogen, intermediate, and active forms of stromelysin-3 (9, 10, 22, 23, 52, and 53). Higher magnifications of the immunohistochemical analysis, as well as immunofluorescent double staining with respective cell-selective antibodies, localized stromelysin-3 within EC, SMC, and MØ of the plaque (Fig. 2). Tissues showed no staining with the respective control IgG1 antibody (data not shown). Because we recently localized CD40 and CD40L in human atherosclerotic plaques and have shown that CD40 ligation induces interstitial collagenases and gelatinases in atheroma-associated cells (41–43), we investigated the possible colocalization of stromelysin-3 with CD40. Indeed, cells expressing stromelysin-3 also bear CD40 (Fig. 3). Furthermore, we analyzed the cellular localization of stromelysin-3 transcripts by in situ hybridization (Fig. 4). Human atheroma (Fig. 4, C–E), but not normal arteries (Fig. 4, A and B), contained stromelysin-3 mRNA. Within the atherosclerotic lesion, stromelysin-3 transcripts localized most prominently at the luminal border and the shoulder region of the plaques, areas described above as positive for the immunoreactive protein. The staining for the transcripts colocalized with smooth muscle cell- and macrophage-like cells (Fig. 4, D and E) as well as the endothelium (Fig. 4 E). Furthermore, transcripts for the immune mediator CD40L showed a similar distribution on adjacent sections (Fig. 4, G and H). In situ hybridization with negative control probes did not yield any signal (Fig. 4 F).

To determine the mechanisms involved in stromelysin-3 expression in atheroma-associated cells, we analyzed the expression of the enzyme in human vascular EC and SMC as well as in human mononuclear phagocytes in vitro. Human vascular EC, SMC, and MØ released immunoreactive stromelysin-3 constitutively in moderate amounts. Stimulation of the cells with the classical mediators of MMP regulation, IL-1 (1 and 10 ng/ml), TNF-α (5 and 50 ng/ml), or IFNγ (100 and 1,000 U/ml), did not affect the expression of stromelysin-3, as illustrated here for IL-1 and SMC (Fig. 5). However, the immunohistochemical studies presented above indicated possible involvement of the CD40– CD40L signaling pathway. Stimulation with both T cell– derived as well as recombinant human CD40L induced stromelysin-3 expression in all three cell types (Figs. 5 and 6). Besides an increased intensity of the higher (∼64 kD) molecular mass band, which corresponds to the molecular mass of the stromelysin-3 zymogen, CD40 ligation induced the expression of immunoreactive proteins of lower molecular mass, corresponding to the molecular mass of the active cleavage products of the zymogen (Fig. 5). In control experiments, addition of blocking anti-CD40L antibody during stimulation of the cells with native or recombinant ligand inhibited induction of stromelysin-3 expression.

Among the cell types analyzed, the immunoreactive forms of stromelysin-3 detected after CD40 ligation showed slight differences in the intensity of the respective bands but were identical in molecular mass. In vascular SMC, CD40 ligation concentration and time dependently elevated expression of the 64-kD protein, as well as induced immunoreactive proteins with molecular masses of ∼48 and ∼28 kD. The induction of the lower molecular mass forms of stromelysin-3 in cultures of SMC required stimulation with 3–10 μg/ml rCD40L (Fig. 6 A). An increase in the 64-kD protein, as well as in the ∼48-kD immunoreactive protein, occurred after 1 h of stimulation, whereas detection of the 28-kD band required at least 6–12 h of exposure to CD40L, as determined by Western blot analysis (data not shown) and radioimmunoprecipitation experiments (Fig. 7). The patterns of immunoreactive proteins detected in cultures of vascular smooth muscle cells resembled those found in fibroblasts, an established source of stromelysin-3 (data not shown). The pattern of immunoreactive bands observed with supernatants of CD40L-stimulated EC (Fig. 6 B) resembled that obtained with cultures of SMC, except that the 64-kD form was much less abundant in EC.

Macrophages, derived from monocytes cultured for 9 d, showed an additional immunoreactive protein at ∼35 kD. Stimulation with ≥0.3 μg/ml rCD40L increased levels of immunoreactive proteins with apparent molecular masses of 64, 48, 35, and 28 kD (Fig. 6 C). Induction of these immunoreactive bands required at least 6 h of stimulation (data not shown). Specificity of the antibody used was confirmed by Western blot analysis performed with the antistromelysin-3 antibody preincubated with recombinant stromelysin-3 (Fig. 6, blocking). In contrast to monocyte-derived MØ cultured for 9 d, freshly isolated peripheral blood monocytes expressed stromelysin-3 neither constitutively nor when stimulated with rCD40L (Fig. 6 D). Responsiveness of monocyte-derived cells to CD40 ligation required a minimum of three days of in vitro culture (data not shown).

The stromelysin-3 expression induced by CD40 ligation in human vascular EC, SMC, and MØ resulted from de novo synthesis of the protein. Metabolic labeling and immunoprecipitation experiments yielded autoradiographic bands resembling the patterns of immunoreactive proteins observed by Western blot analysis as shown here for vascular SMC (Fig. 7). In accordance with our protein analysis, RT-PCR experiments showed increased product corresponding to stromelysin-3 transcript after CD40 ligation in EC and SMC as well as MØ (data not shown).

The in situ observations of stromelysin-3 expression in human atherosclerotic plaques, in combination with the in vitro findings that CD40 ligation selectively mediates the expression of stromelysin-3, suggested an in vivo evaluation of the importance of CD40–CD40L signaling for the expression of this enzyme within atherosclerotic plaques. For this purpose, an established animal model of arteriosclerosis was used: LDLR^−/− mice were fed a high cholesterol (1.25%) diet to develop atherosclerotic lesions (51). Immunohistochemical analysis of lesions within the aortic arch as well as the thoracic portion of the aorta revealed a stromelysin-3 expression pattern similar to that found in human atherosclerotic plaques. All three vascular cell types—EC, SMC and MØ—within the lesions stained for the proteinase. Treatment of the mice with rat IgG (n = 8) did not affect the stromelysin-3 expression (Fig. 8 A) compared to controls. In contrast, mice treated with the anti–mouse CD40L antibody (n = 8) showed substantially reduced immunoreactivity for stromelysin-3 within the atherosclerotic lesions (Fig. 8 B). Since the anti-CD40L antibody treatment also resulted in a decrease of total plaque number and area (51), lesions of similar sizes within the treatment groups were compared. No immunoreactivity was observed in tissues stained with the control IgG1 antibody (data not shown).

This study establishes the expression of stromelysin-3 in advanced human atheroma and furthermore provides evidence for (a) colocalization of this matrix metalloproteinase with CD40 on lesional EC, SMC, and MØ in situ and (b) regulation of de novo expression of stromelysin-3 by CD40 ligation, rather than by the classical soluble mediators of MMPs such as IL-1, TNF-α, or IFN-γ. We also demonstrated in vivo that the interruption of CD40– CD40L interaction markedly reduced the expression of stromelysin-3 in mouse atheroma. The finding that EC, SMC, and MØ express stromelysin-3 establishes atheroma-associated cells as novel sources of stromelysin-3. Thus, this report provides evidence that stromelysin-3, whose expression correlates with the invasiveness of malignancies (3, 11, 12, 16, 17, 54), might also participate in the pathogenesis of another common human disease, atherosclerosis. In view of its unusual substrate specificity, stromelysin-3 might participate in several pathways. The pattern of stromelysin-3 expression within tissues that undergo extensive remodeling in pathological (carcinomas; 11, 12, 16, 17, 54) as well as physiological (placenta and uterus; 11, 18) events, indicates a role for stromelysin-3 in these processes. Although related by sequence to the MMP family, stromelysin-3 does not hydrolyze many of the extracellular matrix components that are substrates for other MMPs such as fibronectin, laminin, elastin, or collagen type I and type IV (23). From this perspective, the enzyme has been considered a nonmatrix–degrading metalloproteinase. Instead, stromelysin-3 appears to act as a predominant regulator of serpin function. Interestingly, a recent report postulated that extracellular matrix degradation might depend not only on an imbalance between matrix metalloproteinases and their inhibitors, but might accelerate in face of decreased levels of certain serpins such as α₂-M, α₂-AP, or α₁-PI (33), preferred substrates of stromelysin-3 (23). Pathways by which this enzyme might regulate matrix degradation include (a) inactivation of matrix-degrading, enzyme-inhibiting serpins as described for cathepsins (32) or for elastolytic activity by α₁-PI, shown to simultaneously augment the proliferative and invasive activity of cells (23); (b) activation of other members of the MMP family (17) copiously expressed in atheroma, such as interstitial collagenase, gelatinases A and B, and stromelysin-1 (5) (at sites of chronic inflammation, such as atherosclerosis, constitutively active stromelysin-3 [9] might act proximally to promote conversion of the zymogen forms of these classical MMPs to their active forms); and (c) degradation of matrix molecules not cleaved by the classical MMPs such as extracellular proteins containing amino acids with unusual long side chains, including those generated in vivo by certain posttranslational modifications (55). Thus, matrix degradation, a critical step in the progression from the stable atheroma to one prone to rupture and capable of causing thrombotic complications, might indeed involve the action of stromelysin-3. The serpin-degrading function of stromelysin-3 may thus have significance beyond tumor invasion and metastasis.

Furthermore, the action of stromelysin-3 on serpins also might affect atherosclerosis by pathways other than extracellular matrix remodeling. Serpins regulate multiple functions associated with the disease, including (a) blood pressure (e.g., the serpin α₁–antitrypsin inhibits renin and thus renin–angiotensinogen interaction, as well as angiotensinogen itself [30]), (b) fibrinolysis (e.g., the serpin α₁–antiplasmin targets plasmin, a key effector of fibrinolysis [23]), (c) blood coagulation (e.g., the serpins anti-thrombin III and heparin cofactor II rapidly interact with thrombin in the presence of heparin [31]), or (d) lipoprotein uptake (cleavage of the serpins α₁-antitrypsin and α₂-macroglobulin is associated with increased low density lipoprotein uptake into cells, indicating that those cleaved serpins disturb the intracellular cholesterol homeostasis [27–29]). Aside from these effects, stromelysin-3 might further affect atherosclerosis via the insulin-like growth factor–insulin-like growth factor binding protein (IGF–IGFBP) system. Human atheroma contain IGF-1 and IGFBPs (56). These mediators are associated with cardiovascular pathophysiology particularly via their critical role in vascular growth (57). IGF-1 directly accelerates arteriosclerosis in rat aorta allografts via myointimal proliferation and intimal thickening (58). The level of free, biologically active IGF-1 depends on the degree of complex formation with the respective binding proteins. A recent study identified IGFBP-1 as a potential physiological substrate for human stromelysin-3 (59). Finally, IGF also acts as a survival factor for human vascular SMC derived from normal vessels as well as coronary atherosclerotic plaques (60), suggesting stromelysin-3 may prevent apoptosis of vascular SMC by augmenting IGF-1 levels.

The potential relevance of serpin degradation for the pathogenesis of atherosclerosis in humans is supported by reports showing that levels of several serpins, including AT-III (36), α₂-M, and α₂-AP (35), decrease in this prevalent disease. In addition, recent in vivo studies demonstrated that serpins inhibit coronary restenosis in atherosclerotic swine (37) and atherosclerotic plaque development in rabbits post injury (38). Our findings that (a) differentiated atherosclerotic lesions but not their precursor, the fatty streak, bear stromelysin-3; (b) the immunoreactive bands detected within the atheromatous plaques resembled mostly the pattern obtained with MØ in vitro; and (c) only differentiated MØ, but not peripheral blood monocytes, exhibit stromelysin-3 induction; support a role of this enzyme in the late rather than the early states of atherosclerosis. Interestingly, the Rotterdam study showed increased AT-III levels in moderate peripheral arterial atherosclerosis and decreased levels in more severe atherosclerosis (61).

Future studies will have to establish whether these various functions apply to human atherosclerosis and whether stromelysin-3 is the crucial mediator in reduction of all or only certain atheroma-associated serpin activities. However, based on the recently burgeoning background information on stromelysin-3, our findings suggest functions of this proteinase in regulation of plaque progression and (in)stability by mechanisms distinct from those ascribed to members of the MMP family, including augmented extracellular matrix degradation, promotion of blood coagulation, decreased fibrinolytic activity, and dysregulation of blood pressure as well as lipoprotein catabolism. The colocalization of stromelysin-3 with CD40 and CD40L within the human atherosclerotic lesion, as well as the demonstration that CD40L, rather than the classical mediators of MMPs, induce stromelysin-3 expression in atheroma-associated cells demonstrate CD40–CD40L signaling as a new, and, as indicated by the in vivo studies, probably crucial pathway of stromelysin-3 induction.

We thank Maria Muszynski, Eugenia Shvartz, and Elissa Simon-Morrissey (Brigham and Women's Hospital) for their skillful technical assistance.

This work was supported in part by grants from the National Heart, Lung, and Blood Institute to Dr. Peter Libby (HL-43364) and the Swiss National Research Fund to Dr. François Mach and was performed during the tenure of the Paul Dudley White fellowship of the American Heart Association by Dr. Uwe Schönbeck.