| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Medicine, Division of Digestive Diseases, Emory University School of Medicine, Atlanta, Georgia, USA;
Inserm, U539, Nantes, France
1Correspondence: Deptartment of Medicine, Division of Digestive Diseases, Emory University, 615 Michael St., Atlanta, GA 30322, USA. E-mail: lcharri{at}emory.edu
 |
ABSTRACT |
|---|
Key Words: metargidin (MDC-15) • cell-cell interactions • cell-matrix interactions • inflammatory diseases • proteases
|
IMPORTANT MECHANISMS INVOLVED IN INFLAMMATION |
|---|
 TOPABSTRACT IMPORTANT MECHANISMS INVOLVED IN... ADAM-15: AN RGD INTEGRIN-BINDING...ADAM-15 AND ATHEROSCLEROSISADAM-15 AND ANGIOGENESISADAM-15 AND RAADAM-15 AND IBDOTHER POTENTIAL ROLES OF...CONCLUSIONSREFERENCES |
|---|
, such cell-cell interactions include adhesion of leukocytes to endothelial cells and cell-cell interactions between lymphocytes and antigen-presenting cells. Cell-cell interactions of leukocytes and other cell types such as intestinal epithelial cells (IECs) are also thought to participate in the recruitment of leukocytes to inflammation sites (1)
. Cell-substratum interactions are required for immune cells to migrate to inflammation sites and include endothelial and eventually epithelial transmigration. These migration processes require remodeling of ECM through proteolytic degradation.
|
There also has been evidence of a relation between angiogenesis and leukocyte infiltration in chronic inflammation. The two phenomena appear to be dependent on each other since inflammatory mediators released by immune cells promote angiogenesis, which, in turn, contributes to inflammation by increasing the recruitment of immune cells to inflammatory sites (2)
. The importance of angiogenesis in the pathogenesis of rheumatoid arthritis (RA) is now well established (2)
, and there has been increasing evidence for a role of angiogenesis in inflammatory bowel disease (IBD) (3)
. Adhesion molecules have been found to be important in angiogenesis by regulating endothelial cell migration (2)
.
Therefore, cell-cell and cell-substratum interactions appear of utmost importance in the mechanisms underlying inflammation, and identifying the mediators of cell migration and cell adhesion events that are involved in inflammatory diseases could help determine therapeutic strategies. Interestingly, the discovery of the disintegrin and metalloprotease (ADAM) family of proteins (4)
, also called MDC (metalloprotease, disintegrin, Cys-rich) proteins, raises the issue of what role these and, more particularly, the RGD motif-containing member, ADAM-15 (5)
, play as potential factors involved in inflammatory diseases.
|
ADAM-15: AN RGD INTEGRIN-BINDING SEQUENCE-CONTAINING ADAM PROTEIN |
|---|
|
TOPABSTRACTIMPORTANT MECHANISMS INVOLVED IN...ADAM-15: AN RGD INTEGRIN-BINDING...ADAM-15 AND ATHEROSCLEROSISADAM-15 AND ANGIOGENESISADAM-15 AND RAADAM-15 AND IBDOTHER POTENTIAL ROLES OF...CONCLUSIONSREFERENCES |
|---|
. ADAM proteins are very complex molecules that possess adhesive and proteolytic activities. They are synthesized as proproteins containing a prodomain that inhibits the catalytic site of the metalloprotease domain through a Cys-switch mechanism. During maturation processes that take place in the endoplasmic reticulum, the prodomain is cleaved by furin or furin-like proprotein convertases, and the resulting protein constitutes the mature form with an active metalloprotease domain. The diversity of domains present in ADAM proteins accounts for the multitude of biological activities they exert (6)
. ADAM proteins indeed have been described as involved in processes as diverse as cell fusion, growth factor and cytokine shedding, cell differentiation, cell migration, and cell adhesion (6)
.
Among the 34 ADAM proteins hitherto described, including 19 human ADAMs, human ADAM-15/MDC-15 is the only ADAM that possesses an RGD integrin-binding sequence in the disintegrin loop. For this reason, ADAM-15 also is named metargidin for metalloprotease-RGD-disintegrin (5)
. The presence of this RGD sequence suggests a specific role of ADAM-15 in integrin binding and therefore in cell-cell interactions. The RGD motif is involved in the binding of ADAM-15 to 
vβ3 and 5β1 integrins (7
, 8)
; however, ADAM-15 also is able to bind to 9β3 integrin in an RGD-independent manner (9)
. As a consequence of such molecular interactions, ADAM-15 overexpression has been found to enhance cell-cell interactions in the fibroblastic cell line NIH3T3 (10)
. Another major function of ADAM-15 is cell migration, which is thought to be the result of ECM remodeling through its metalloprotease domain (11)
. ADAM-15 is an ADAM in which the metalloprotease domain is conserved in the mature form. ADAM-15 metalloprotease domain possesses the consensus Zn2+-binding site: HexxHxxGxxH, and in vitro studies showed that ADAM-15 is able to cleave type IV collagen and gelatin (11)
. This metalloprotease activity is thought to be involved in ADAM-15-mediated enhancement of glomerular mesangial cell migration (11)
. In addition, ADAM-15 cytoplasmic tail presents motifs involved in cell signaling (5
, 6
, 12
, 13)
. The intracellular domain of ADAM-15 possesses proline-rich sequences, suggesting possible interaction with Src homology (SH) 3 domain-containing proteins. Poghosyan et al. (12)
showed that the cytoplasmic domain of ADAM-15 can form specific, phosphorylation-dependent interactions with Src family protein-tyrosine kinases and with the Grb2 adaptor protein in hematopoietic cell lines. The amino acid sequence of ADAM-15 cytoplasmic tail also contains potential tyrosine phosphorylation sites, suggesting that ADAM-15 could interact with SH2 domain-containing proteins via phosphotyrosines (12)
, as well as potential serine and threonine phosphorylation sites (13)
. Collectively, these observations suggest that ADAM-15 could play a role in cell signaling events.
These ADAM-15 properties raise this protein as a target of choice in the identification of mediators of inflammatory diseases, and it has been the subject of several studies aimed at determining its roles in different inflammatory diseases. Here we review and discuss these recent discoveries pointing at ADAM-15 as a mediator of mechanisms underlying inflammation and as a possible therapeutic target for prevention of inflammatory diseases.
|
ADAM-15 AND ATHEROSCLEROSIS |
|---|
. Soon after that, ADAM-15 was cloned from human umbilical vein endothelial cells (HUVECs) and cultured human aortic smooth muscle cells (SMCs) (14)
, and ADAM-15 protein expression was found in HUVECs (14
15
16
17
18)
. These findings led authors to investigate the expression of ADAM-15 during atherosclerosis, because this disease is initiated and promoted by the recruitment of inflammatory cells into the vascular wall (19)
and involves tissue morphogenesis and remodeling. ADAM-15 expression was assessed in sections from normal and hypercholesterolemic nonhuman primate (Macaca nemestrina) aorta and was not found to be detectable in normal aorta (14)
. However, ADAM-15 expression was detected in lesions of atherosclerosis in macrophage-rich and intimal regions (14)
. This work demonstrating ADAM-15 up-regulation during atherosclerosis suggested a potential involvement of ADAM-15 in this pathology.
Another interesting finding was the up-regulation in atherosclerosis of 5β1 and vβ3 integrins (16)
, which are ADAM-15 binding partners (7
, 8)
and which have been shown to be involved in the development and progression of atherosclerosis (20
, 21)
. Like 5β1 and vβ3 integrins, ADAM-15 was found to be increased at both the mRNA and protein levels in human atherosclerotic carotid samples, in comparison with normal samples where ADAM-15 was only weakly expressed (16)
. Immunohistochemistry showed a coexpression of ADAM-15 and its binding partners, 5β1 and vβ3 integrins in SMCs, present in the neointima of atherosclerotic arteries (16)
. More particularly, ADAM-15 was expressed prominently in foam cells of smooth muscle origin and SMCs surrounding lipid cores in the neointima (16)
. In vitro studies determined that treatment of human arterial or venous vascular SMCs with platelet-derived growth factor (PDGF) induces an up-regulation of 5β1 and vβ3 integrins, followed by an up-regulation of ADAM-15 (16)
. It is thought that ADAM-15 might act as an integrin regulator (13, 14, 16, 22–24) and that its up-regulation might occur in response to increased integrin expression to modulate integrin-matrix interactions (16)
. ADAM-15 could either favor or inhibit SMC migration. Since it possesses a functional metalloprotease domain, ADAM-15 could cleave cell-matrix bonds and therefore support cell migration, as has been suggested for glomerular mesangial cells (11)
. Binding of ADAM-15 to 5β1 integrin also could favor cell migration through competition with 5β1 integrin/fibronectin interactions, which suppress cell migration (25)
. On the other hand, increased ADAM-15 expression could inhibit SMC migration through enhancement of cell-cell interactions, as demonstrated in NIH3T3 cells (10)
. Moreover, it is reasonable to speculate that ADAM-15 binding to its partners 5β1 and vβ3 integrins could down-regulate these integrin-dependent migration and proliferation in SMCs (21)
. However, further studies, discussed later in this review, show an involvement of ADAM-15 in angiogenesis (17
, 26)
, and support stimulation of SMC migration by ADAM-15. The mechanisms underlying ADAM-15 up-regulation subsequent to increased integrin expression remain to be elucidated.
Other studies (18
, 27)
have pointed out that ADAM-15 is also involved in atherosclerosis through platelets that appear to play a critical role in the initiation of atherosclerotic lesions (28)
. ADAM-15 has been shown to interact with platelet IIbβ3 integrin (GPIIb-IIIa) (18
, 27)
. Langer et al. (18)
have indeed demonstrated that platelets adhere to immobilized recombinant ADAM-15 under static and high shear rate conditions and that this adhesion is as strong as the established adhesion to fibrinogen. Such interactions occur only when IIbβ3 integrin is activated (18)
. Moreover, recombinant disintegrin domain of ADAM-15 was shown to inhibit IIbβ3 integrin/fibrinogen interactions and bind to platelet surface (27)
. Interestingly, overexpression of ADAM-15 in endothelium-like cells increases platelet adhesion under physiological flow conditions, and treatment of HUVECs with a combination of proinflammatory cytokines (50 ng/ml TNF- and 20 ng/ml IFN
) slightly but significantly increases ADAM-15 expression (18)
. These findings suggest that ADAM-15 expressed on endothelial cells could act as a receptor for platelets or as a receptor involved in the recruitment of immune cells during inflammation. In addition, adhesion of platelets on ADAM-15-overexpressing endothelium-like cells leads to attraction and binding of additional platelets, resulting in the formation of microthrombus (18)
. Platelet adhesion is known to result in the secretion of inflammatory mediators that stimulate endothelial cells and recruit additional platelets to the thrombus (28
29
30)
. Interestingly, incubation of platelets with recombinant ADAM-15 induces platelet secretion of CD62P/P-selectin as well as CD40L (18)
, a finding of utmost importance because CD40L triggers inflammatory processes in endothelial cells (28
29
30)
. CD40L indeed initiates proteolytic activity and induces chemokine secretion and adhesion molecule expression in endothelial cells, thereby generating signals for the recruitment of leukocytes at the site of injury (29
, 30)
. It is therefore reasonable to speculate that binding of platelet IIbβ3 integrin to endothelial ADAM-15 could result in activation of platelets and subsequent secretion of inflammatory mediators that would, in turn, stimulate endothelial cells to produce inflammatory mediators, leading to the recruitment of additional platelets and formation of thrombus as well as recruitment of inflammatory cells. The role of platelets in atherosclerosis has been well documented (28)
, and, therefore, platelets could constitute a link between ADAM-15 and this pathology.
|
ADAM-15 AND ANGIOGENESIS |
|---|
, 3)
. Several findings point to ADAM-15 as a mediator of angiogenesis. ADAM-15 is expressed in endothelial cells (14
15
16
17
18)
and is up-regulated on the angiogenic endothelial cell surface (14)
. Moreover, ADAM-15 has been shown to be involved in cell migration (11)
and is able to digest gelatin and type IV collagen (11)
, which are required for endothelial cell sprouting and migration. In addition, ADAM-15 interacts with 5β1 and vβ3 integrins (7
, 8)
, which are known to promote cell proliferation and migration of endothelial cells and SMCs (21)
. Because 5β1 is a proangiogenic adhesion integrin and vβ3 an integrin preferentially expressed on angiogenic blood vessel (31
, 32)
, the functions of these integrins and ADAM-15 may be interdependent. These findings have led to studies aimed at investigating the role of ADAM-15 in angiogenesis (17
, 26)
.
Horiuchi et al. (the team of Carl P. Blobel; ref. 17
) established ADAM-15 knockout (ADAM-15–/–) mice, which do not display any obvious defects or pathological phenotypes. Because ADAM-15 expression is high in vascular cells of wild-type mice and vasculature does not present any evident alterations in ADAM-15–/– mice, Horiuchi et al. (17)
investigated the role of ADAM-15 in pathological neovascularization in two experimental models for neovascularization: retinopathy of prematurity and melanoma progression. The retinopathy of prematurity model consists in placing young mice in high oxygen (75%) for several days, after which mice are returned to room air, where the drop in oxygen concentration triggers a strong angiogenic stimulus and results in pathological neovascularization in the retina. In this model, ADAM-15–/– mice were found to have 64% lower angiogenic response than wild-type mice (17)
. Angiogenesis observed in wild-type mice that underwent oxygen concentration change was correlated with transient increased ADAM-15 protein expression in the retina in comparison with untreated control age-matched wild-type mice (17)
. Immunohistochemistry showed that ADAM-15 expression in retinal capillaries was enhanced particularly in new blood vessels (17)
. Altogether, these observations strongly suggest ADAM-15 as a mediator of neovascularization.
The second model of neovascularization investigated by Horiuchi et al. (17)
relies on the fact that tumors must recruit host vessels in order to survive (33)
. In this model, melanoma cells were implanted in the flanks of ADAM-15–/– and wild-type mice, and tumor progression and associated neovascularization were assessed (17)
. ADAM-15–/– mice displayed a significantly decreased tumor progression compared to wild-type mice (17)
; however, the difference in tumor progression between ADAM-15–/– and wild-type mice could not be directly correlated to a defect in angiogenesis because tumors exhibited similar morphological appearance and density of blood vessels in both genotypes (17)
. If this latter model does not give direct proof of an involvement of ADAM-15 in pathological neovascularization, the model of retinopathy of prematurity certainly does. The mechanisms by which ADAM-15 exerts this effect as well as the mechanisms underlying angiogenesis-associated ADAM-15 up-regulation were not identified in this study.
The finding made by Horiuchi et al. (17)
that ADAM-15 is involved in angiogenesis and in tumor progression has been further correlated by Trochon-Joseph et al. (26)
, who assessed the effects of human ADAM-15 recombinant disintegrin domain (RDD) in angiogenesis, tumor growth, and metastatic spreading as well as the underlying mechanisms. The authors found, using an in vitro approach, that RDD inhibits capillary-like structure formation of an endothelial cell line (26)
. This effect appears to be mediated through inhibition of cell proliferation and cell migration (26)
. Cell adhesion assays showed that RDD inhibits, by 
30%, endothelial cell adhesion to immobilized adhesive proteins (26)
. To assess the in vivo relevance of RDD-mediated inhibition of angiogenesis, the authors employed a very interesting approach using in vivo-induced synthesis and secretion of RDD by skeletal muscle with a Tet-ON/Tet-OFF system activated by doxycycline in nude mice (26)
. Using this system, RDD was found to significantly (78%) inhibit tumor growth of inoculated breast cancer cells (26)
. Interestingly, RDD-treated tumors exhibited 53% fewer blood vessels than nontreated control tumors (26)
. This correlation between tumor growth and tumor vascularization strongly suggests that inhibition of tumor growth by RDD results from an inhibition of angiogenesis. The fact that ADAM-15 RDD inhibits angiogenesis and tumor growth (26)
, which require ADAM-15, indicates that ADAM-15 regulates these biological processes and suggests that RDD and ADAM-15 have different mechanisms of action. It has to be noted that human ADAM-15 RDD possesses an RGD motif while the disintegrin domain of murine ADAM-15 possesses a TDD sequence instead (5)
. It is therefore reasonable to think that these two proteins have different endothelial binding partners. Because inhibition of endothelial cell adhesion to immobilized adhesive proteins by RDD is similar whether cells adhere on vitronectin, fibronectin, or fibrinogen, it is suggested that RDD does not inhibit integrin-mediated cell adhesion (26)
. The fact that ADAM-15 mediates angiogenesis while its RDD inhibits this phenomenon also suggests that ADAM-15-mediated neoangiogenesis requires other domains of ADAM-15 than the disintegrin domain. It is therefore very tempting to speculate that the role of ADAM-15 in angiogenesis is mediated via its metalloprotease activity and subsequent degradation of ECM components (11)
, leading to migration of endothelial cells and SMCs.
Further in vitro studies performed by Horiuchi et al. (17)
using HUVECs treated with vascular endothelial growth factor (VEGF) or basic fibroblast growth factor (bFGF), which are two potent angiogenic factors, for 3 and 9 h did not indicate that these two growth factors are involved in the regulation of ADAM-15 in endothelial cells and, therefore, in its angiogenic activity (17)
. However, Komiya et al. (34)
found that treatment of HUVECs with the isoform 165 of VEGF (VEGF165) enhances ADAM-15 mRNA expression by 2.2-fold. This result suggests that VEGF165 is responsible for angiogenesis-associated up-regulation of ADAM-15 expression (17)
, and that VEGF165–mediated angiogenesis could occur via ADAM-15. Interestingly, angiogenic factors such as VEGF, which has been involved in angiogenesis in different chronic inflammatory processes (35)
, are known to require proteolytic cleavage to play their role. Matrix metalloprotease-9 (MMP-9) is known to release sequestered VEGF from the matrix, allowing it to stimulate angiogenesis (36)
. In this respect, mast cell chymase also possesses angiogenic activities by indirect release of ECM-sequestered VEGF through cleavage of an unidentified substrate (36)
. Therefore, in addition to the potential roles of ADAM-15 in angiogenesis through interactions with its ligands or via ECM degradation, ADAM-15 could directly or indirectly be involved in the release of endothelial growth factors, as suggested by Böhm et al. (37)
.
Altogether, these observations point to another aspect of regulation of inflammation by ADAM-15. This role of ADAM-15 is very important because angiogenesis is involved in various diseases lacking therapeutic strategies, such as cancer and chronic inflammation, including RA and IBD (2
, 3)
.
|
ADAM-15 AND RA |
|---|
. Articular cartilage chondrocytes have been shown to possess a metalloprotease activity and are believed to play an important role in the cartilage destruction that occurs during degenerative joint diseases (39)
. After the discovery of ADAM proteins, several studies aimed at investigating the expression and potential role of these metalloproteases in articular cartilage chondrocytes as well as in arthritic tissues. As early as 1997, after ADAM-15 was cloned (5)
, ADAM-15 expression was found in cultured articular cartilage chondrocytes (40
, 41)
, therefore raising it as a potential metalloprotease involved in the previously described aggrecanase activity of chondrocytes and in degenerative joint disease. The interest for ADAM-15 in degenerative joint disease studies is further justified by the fact that ADAM-15 has been shown to cleave ECM components such as gelatin and type IV collagen (11)
. Interestingly, treatment of human chondrocyte monolayer cultures with interleukin (IL) -1 or retinoic acid was found to increase ADAM-15 mRNA levels (41)
. Since retinoic acid and IL-1 have been shown to induce aggrecanase activity in cultured cartilage explants or chondrocytes (42)
, this latter finding of ADAM-15 up-regulation suggests that ADAM-15 could be responsible for chondrocyte aggrecanase activity and/or for cytokine-induced cartilage remodeling. However, the target structures of such proteolytic activity still need to be elucidated.
Interestingly, increased ADAM-15 expression was found in rheumatoid synovial tissue compared with normal synovial tissue or even osteoarthritic synovial tissue (34
, 37)
. While ADAM-15 was detected only in synoviocytes in the lining layer in normal synovial tissue, in RA synovium it was expressed strongly in the highly hyperplastic lining, as well as in plasma cell infiltrates present in the subintimal lining, in endothelial cells, in the marginal zones of follicular lymphocytes infiltrates, around germinal center-like structures, and next to blood vessels (34
, 37)
. The expression of ADAM-15 in endothelial cells and immune cell infiltrates suggest a role for ADAM-15 in RA via immune cell recruitment. In the hyperplastic lining of RA samples, ADAM-15 was expressed by fibroblast-like and macrophage-like synoviocytes, the latter cell type being the major source of ADAM-15 in RA synovial lining, accounting for increased ADAM-15 expression in comparison with normal lining (37)
. In RA synovial tissue, ADAM-15 also was highly expressed in multinucleated giant cells present at the border of cartilage destruction in some of the tissue samples (37)
. It is therefore very tempting to speculate that ADAM-15 also could degrade ECM, either directly through its metalloprotease activity (11)
or indirectly through proteolytic activation of matrix metalloproteases (MMPs). It was suggested that the latter could occur via plasma cells (37)
because plasma cells express MMP-3 and MMP-9 (43)
. In RA, several proteases have been identified to play a role in tissue remodeling. For example, cathepsins B, L, and K; MMPs; and aggrecanases produced by synovial fibroblasts, upon proinflammatory stimulation, leads to collagen II and aggrecan digestion and subsequent cartilage degradation, whereas proinflammatory stimulation of osteoclasts results in the production of cathepsin K and MMP-13, leading to collagen I degradation and subsequent bone remodeling (44)
. Moreover, among the proteases released by neutrophils, long known to play an important role in RA (45)
, neutrophil elastase degrades collagen, elastin, laminin, fibronectin, and proteoglycan (46)
. Therefore ADAM-15 could possibly activate such proteases through proteolytic cleavage.
Angiogenesis is highly associated with the pathogenesis of RA (2
, 47)
. In RA synovium, angiogenesis begins at the early stage of the disease and is crucial for progression of the synovitis (47)
. Both ADAM-15 and its ligand vβ3 integrin (8
, 7)
are strongly expressed in RA synovium (37
, 48)
and are involved in angiogenesis (17
, 26
, 32)
. It is therefore reasonable to speculate that ADAM-15 could participate in the development and/or maintenance of RA by mediating angiogenesis, as discussed earlier in this review. Interestingly, a linear correlation was found between ADAM-15 expression levels and vascular density in RA and osteoarthritic synovial tissues (34)
. VEGF, which is involved in angiogenesis in different chronic inflammatory processes (35)
and is highly expressed in RA synovial tissue, is known to play a key role in angiogenesis in RA synovium (47
, 49)
. Therefore, ADAM-15 possibly plays a role in RA via secretion of VEGF or other angiogenic factors (37)
. Interestingly, as previously mentioned, a link between VEGF and ADAM-15 was demonstrated by Komiya et al. (34)
, who showed that treatment of HUVECs with VEGF165 increases ADAM-15 mRNA expression. VEGF165-mediated increase of ADAM-15 expression appears to be dependent on VEGF receptor 2 (VEGFR-2) -mediated signaling because it occurs only after VEGFR-2 expression induced by TNF- in RA synovial fibroblasts and because it was blocked with a specific inhibitor of VEGFR-2 (34)
.
Another potential role of ADAM-15 in RA lies in the fact that it has been reported to produce soluble forms of CD23 through ectodomain shedding of inflammatory cell membrane-bound CD23/Fc
RII, the low affinity IgE receptor (50)
that is expressed on B cells, monocytes, macrophages, and eosinophils (51)
and is cleaved from the cell surface to generate a number of soluble forms. Soluble forms of CD23 play an important role in the up-regulation of IgE synthesis by interaction with B cells (52)
. Interestingly, both membrane-bound and soluble forms of CD23 have been shown to be elevated in a number of diseases, including RA and IBD (53)
. ADAM-15 therefore could account for increased level of soluble CD23 in synovial fluid and sera of RA patients (54)
. Because CD23 is known to stimulate immune cells (53)
, ADAM-15 could play a role by amplifying inflammation of RA synovitis. Despite studies from Weskamp et al. (55)
pointing at ADAM-10 as the main ADAM involved in CD23 shedding in vivo, ADAM-15 cannot be excluded from assuming this role after up-regulated expression in pathological conditions.
In addition to the potential role of ADAM-15 in RA through ECM degradation, RA synovium angiogenesis, and CD23 shedding, other roles of ADAM-15 in RA remain to be defined. Since RA synovial lining cells express both ADAM-15 (34
, 37)
and 5β1 and vβ3 integrins (48
, 56)
, ADAM-15 could act as an adhesion molecule to modulate cell-cell and/or cell-matrix interactions. A more recent study (57)
shows that aging ADAM-15–/– mice present accelerated development of osteoarthritic lesions compared with wild-type mice, suggesting that ADAM-15 expressed by chondrocytes has a homeostatic, rather than a destructive, role in cartilage. It is possible that ADAM-15 plays both homeostatic and pathological roles depending on the cells and the context in which it is expressed. Similarly, we cannot exclude the possibility that ADAM-15 could present both pro- and anti-inflammatory activities. A lot of known mediators of inflammation have been described to have anti-inflammatory roles. For example, proteases can inactivate inflammatory mediators through proteolytic degradation. Prolonged exposure of active IL-1β to MMP-2, MMP-3, or MMP-9 results in its degradation (58)
. Neutrophil elastase also has been shown to regulate inflammation by cleavage of immunoglobulins, complement components, and complement receptor type 1 on neutrophils, resulting in inhibition of neutrophil stimulation and subsequent release of elastase (59)
. Proteases also can have indirect antiangiogenic activities. For example, plasminogen cleavage by MMP-7 and MMP-9 and collagen XVIII cleavage by elastases release angiostatin and endostatin, respectively, which are antiangiogenic proteins (46)
.
Further studies, therefore, are necessary to elucidate the role of ADAM-15 in OA and RA as well as in inflammation in general.
|
ADAM-15 AND IBD |
|---|
, Western blot (13)
, and immunohistochemistry (24)
. While Western blot performed on samples from human normal small intestine showed the expression of both the proform and the mature form of ADAM-15 (13)
, samples from human normal colon presented only one form, with an intermediate size between the proform and the mature form (13
, 24)
. This protein could be a different isoform of ADAM-15 that would be specific to the colon. The physiological significance of this differential expression of ADAM-15 between the small intestine and the colon remains to be determined. Immunohistochemistry experiments showed that, within the intestine, several cell types express ADAM-15 (24)
. ADAM-15 indeed was found in epithelial cells lining the crypt and the surface, as well as in endothelial cells, pericryptic myofibroblasts, SMCs, and in the rare resident mononuclear inflammatory cells that are present in the lamina propria; however, myenteric and submucosae plexus did not show any expression of ADAM-15. The colocalization of ADAM-15 and its binding partners or substrates in the intestine observed by these authors is summarized in Table 1
.
|
The human small intestine-like cell line, Caco2-BBE, commonly used as a model of human enterocytes, also was shown to express ADAM-15 in both its proform and mature form (13)
. Moreover, reverse transcriptase-polymerase chain reaction (RT-PCR) experiments showed that at least 2 isoforms of ADAM-15 that differ in their cytoplasmic tail are present in Caco2-BBE cells. This finding correlates with the description of different variants of ADAM-15. A total of 6 isoforms of ADAM-15 have been described (accessions NM_207191, NM_003815, NM_207194, NM_207195, NM_207196, NM_207197). It is important to note that the different variants encoding ADAM-15 proteins with different cytoplasmic tails present different capacities of interacting with Src family proteins (60
, 61)
. Thus, differential expression of ADAM-15 isoforms could result in different cell signaling events.
Studies aimed also at determining the status of intestinal ADAM-15 expression during IBD. Real time RT-PCR experiments showed a drastic increase of ADAM-15 mRNA in samples from patients with ulcerative colitis or Crohn’s disease in comparison with samples from normal tissue (24)
. This up-regulation of ADAM-15 during IBD was further confirmed at the protein level by immunoblot experiments (24)
. These experiments did not determine whether there is a differential expression of ADAM-15 variants during IBD vs. normal colon because the primers used for the real time RT-PCR experiments were designed to amplify a region of ADAM-15 cDNA localized in the extracellular domain. Such studies would be interesting since differential expression of ADAM-15 variants could lead to different intracellular signals (60
, 61)
. The finding that ADAM-15 is up-regulated during IBD suggests that ADAM-15 could be involved in the mechanisms underlying intestinal inflammation.
Some roles of ADAM-15 have been investigated in vitro, in Caco2-BBE cells. Wound-healing experiments showed that overexpression of ADAM-15 inhibits the mechanisms of wound repair in Caco2-BBE cells (13
, 62)
. Therefore, the role of ADAM-15 in cell migration appears to vary in different cell types because ADAM-15 enhances glomerular mesangial cell migration (11)
. However, the finding that ADAM-15 inhibits wound healing in Caco2-BBE cells is in accordance with the studies by Herren et al. (10)
showing that ADAM-15 inhibits wound -healing mechanisms in the fibroblastic cell line NIH3T3. In NIH3T3 cells, ADAM-15 inhibits cell migration by enhancing cell-cell interactions (10)
, which also could be the mechanism involved in Caco2-BBE cells. It is therefore reasonable to speculate that ADAM-15 could be involved in IEC homotypic cell-cell interactions involved in the formation of the epithelial barrier. Interestingly, in IBD, the localization of ADAM-15 and its binding partners (24)
suggests that ADAM-15 could mediate cell-cell interactions required for leukocyte homing that occurs during intestinal inflammation (Table 1)
. Observations of a colocalization of ADAM-15-positive endothelial cells with 5β1 integrin-positive leukocytes in tissues from patients with ulcerative colitis or Crohn’s disease (24)
strongly suggest a role for ADAM-15 in cell-cell interactions between leukocytes and endothelial cells and therefore in the transmigration of leukocytes across endothelium. Similarly, the colocalization of ADAM-15-positive IECs with 5β1 integrin-positive leukocytes (24)
strongly suggests a role of ADAM-15 in leukocyte adhesion to and transmigration across intestinal epithelium.
Our very recent work aimed at determining whether ADAM-15 mediates heterotypic cell-cell interactions of IECs with T lymphocytes (62)
. We found that an antibody raised against ADAM-15 ectodomain decreased Jurkat cell attachment to Caco2-BBE monolayers (62)
. Moreover, ADAM-15 overexpression by Caco2-BBE cells enhanced Jurkat cell adhesion to Caco2-BBE monolayers (62)
. The establishment of such cell-cell interactions was dependent on both the RGD integrin-binding sequence and the cytoplasmic tail because the mutation of the RGD motif or the deletion of the intracellular domain inhibited exogenous ADAM-15-mediated increase in Jurkat cell attachment to Caco2-BBE monolayers (62)
. These findings demonstrated that intestinal epithelium-derived ADAM-15 mediates heterotypic cell-cell interactions of IECs with T lymphocytes. The requirement for the RGD motif in these cell-cell interactions indicates that ADAM-15 directly participates in such adhesion events by interactions with its ligands. Epithelial ADAM-15 ligands on T lymphocytes could be vβ3 and 5β1 (8)
. The requirement for ADAM-15 cytoplasmic tail indicates that signal transduction mediated through ADAM-15 intracellular domain might also be involved in the establishment of these cell-cell interactions. ADAM-15-mediated signal transduction already has been strongly suggested (12)
. ADAM-15 binding to its ligands could possibly regulate adhesion mechanisms via up-regulation of other adhesion molecules involved in these heterotypic cell-cell interactions. Such adhesion molecules could be lymphocyte LFA-1, Eβ7 integrin, N-cadherin or epithelial ICAM-1, E-cadherin, or CD1d (63
64
65
66
67
68)
.
Moreover, we found that ADAM-15-mediated attachment of Jurkat cells enhanced Caco2-BBE wound repair (62)
. This finding is of great physiological relevance for IBD since intestinal inflammation is characterized by epithelial disruption followed by epithelial restitution. Possible mechanisms to explain this latter finding include: 1) decreased ADAM-15-mediated homotypic cell-cell interactions of IECs resulting from Jurkat cell binding and therefore increased motility of IECs, and 2) T lymphocyte activation upon adhesion to IECs through ADAM-15 and subsequent production of factors that increase IEC wound-healing, such as IL-2 (69)
and IL-22, which is increased in active Crohn’s disease (70)
.
In our recent work, we also investigated the role of T lymphocyte-derived ADAM-15 in cell-cell interactions (62)
. After overexpressing ADAM-15 in Jurkat cells, we found that ADAM-15 mediates homotypic cell aggregation of T lymphocytes (62)
. The RGD sequence appeared to be crucial for these T cell homotypic interactions because the mutation of this motif completely abolished the formation of Jurkat cell aggregates (62)
. ADAM-15-mediated formation of T cell aggregates can be explained by the fact that lymphocytes express both ADAM-15 and at least two of its ligands, i.e., vβ3 and 5β1 integrins (5
, 7
8
9
, 12
, 71)
. However, the cytoplasmic tail of ADAM-15 was found to play a secondary role in T cell aggregation because the deletion of ADAM-15 intracellular domain only partially inhibited or delayed the formation of these clusters (62)
. We found also that ADAM-15-mediated formation of T cell aggregates induced tumor necrosis factor (TNF-) mRNA expression. These findings strongly suggest that signal transduction mediated through ADAM-15 is required not only for the establishment of T cell homotypic interactions but also for the induction of TNF- mRNA expression. This latter speculation is supported by the finding that the cytoplasmic tail of ADAM-15 induces signaling transduction in immune cells (12)
. As discussed for IEC/T lymphocyte heterotypic interactions, ADAM-15-mediated aggregation of T lymphocytes could also be the result of up-regulation of other adhesion molecules involved in these homotypic cell-cell interactions, such as N-cadherin and LFA-1 (72
, 73)
.
These findings are of a particular relevance for the pathogenesis of IBD because most descriptions of early Crohn’s disease lesions include mucosal lymphoid aggregates (74
, 75)
that are mainly composed of T cells and are believed to contribute to the pathogenesis of Crohn’s disease (74)
. Moreover, TNF- is well established as a cytokine that plays a major role in inflammation (75)
.
Since membrane-bound and soluble forms of CD23 have been shown to be elevated in IBD (53)
, ADAM-15 roles in the pathogenesis of IBD also could be partly mediated via CD23 processing (50)
, as previously discussed.
|
OTHER POTENTIAL ROLES OF ADAM-15 IN INFLAMMATION |
|---|
Maturation/release of cytokines and growth factors
ADAM-15 could be responsible for the maturation and release of cytokines. Pro-TNF-, which is expressed as a membrane-bound protein, is proteolytically released by ADAM-17, thus also named TNF--converting enzyme (TACE), resulting in the release of the TNF- biologically active form from the cell surface (76)
. The maturation of IL-1β occurs in the intracellular compartment where IL-1β-converting enzyme (ICE/caspase-1) cleaves pro-IL-1β, resulting in its mature active form (77)
. Proteolytic cleavage of substrates is not restricted to only one particular protease. For example, neutrophil proteinase-3 also can process pro-TNF- to its active soluble form and directly activate IL-1β (78
, 79)
. Similarly, MMP-2, MMP-3, and MMP-9 also have been described to process IL-1β (58)
. While ADAM-15 has been shown to induce the shedding of membrane-bound proteins (50)
, ADAM-15-mediated cleavage of intracellular substrates was suggested by the studies from Lum et al. (80)
but has not been demonstrated. In this respect, intracellular activity of ADAM-15 expressed in immune cells also could include roles in immune response, such as antigen processing and presentation, or proteolytic activation of granular proteases such as granzymes, in which cathepsins have been involved (44)
. ADAM-15 itself is processed in endoplasmic reticulum by a furin-like proprotein convertase (80)
; moreover, some ADAMs were demonstrated capable of autocatalysis (6)
.
Other mediators of inflammation, such as growth factors, also require proteolytic cleavage in order to play their role. For example, proteolytic activation of transforming growth factor-β (TGF-β) by MMP-2 and MMP-9 promotes angiogenesis (81)
. Another example is VEGF, which has already been discussed in this review.
Shedding of receptors for ILs
The shedding of IL receptors results in secreted soluble forms of receptors that can mediate events at a distance from cells of origin. ADAM-15 has been speculated to be responsible for such ectodomain shedding activity (37)
. Two other members of the ADAM family, ADAM-10 and ADAM-17, have been shown to catalyze the shedding of IL-6 receptor (IL-6R) (82)
, leading to the secretion soluble IL-6R that is known to account for the pathophysiological effects of IL-6 (83)
. Interestingly, IL-6 and its soluble receptors have been shown to enhance proteoglycan degradation in articular cartilage (84)
.
Regulation of protease systems involved in inflammation
Different protease systems have been described as major components of inflammation. Consisting in proteolytic cascades triggered by immunological or inflammatory factors and resulting in final and/or intermediate products with proinflammatory properties, they include the kinin, coagulation, plasminogen activator/plasmin, and complement systems.
The kinin system consists of kinins such as bradykinin (BK) and kallidin, which are generated after proteolytic cleavage of their respective precursors, high-molecular-weight kininogen and low-molecular-weight kininogen, by plasma serine proteases named kallikreins (85
, 86)
. After their release in plasma, BK and kallidin undergo rapid proteolytic cleavage by enzymes generically called kininases, leading to the formation of the active metabolites, des-Arg-BK9 and des-Arg10-kallidin (85
, 86)
. Kinins and their active metabolites exert their effects through binding to bradykinin B2 and bradykinin B1 receptors, respectively. Activation of these receptors leads to release of prostaglandins and leukotrienes; mast cell mediators; various cytokines, including IL-1; chemotactic peptides; increased vascular permeability; and leukocyte recruitment. The role of the kinin system in RA has been well documented (85
86
87)
.
The coagulation system and the plasminogen activator/plasmin system are two interdependent protease systems known to be involved in the mediation of inflammation in RA. Extravascular coagulation that arises in RA is thought to be due to the concomitant local extravasation of plasma fibrinogen resulting from increased permeability of inflamed synovial microvasculature and the induction of the membrane glycoprotein tissue factor (TF) expression by immunological and proinflammatory stimuli (38)
. The TF-mediated coagulation pathway consists of a cascade of proteolytic events involving the serine proteases factor VIIa, factor Xa, and thrombin and resulting in the degradation of fibrinogen into fibrin by thrombin (38)
. Potential deleterious effects of fibrin deposition include hypoxia and acidosis in synovial fluid; proinflammatory effects of fibrin and its degradation products, such as increased expression of IL-1β and IL-6 in monocytes; and increased production of IL-8 by endothelial cells (38)
. Fibrin degradation is the result of the serine protease plasmin via the plasminogen activator/plasmin system. Plasmin derives from proteolytic cleavage of plasminogen by two types of plasminogen activators (PAs), urokinase (uPA) and tissue-type PA (tPA), which are secreted upon immunological or proinflammatory stimuli (38)
. While tPA is secreted and binds to fibrin, uPA, as well as its inactive proform, prouPA, binds to a specific cell surface receptor, uPAR. Deleterious effects of activation of this system include not only proinflammatory effects of fibrin degradation products but also lie in the fact that plasmin substrates also include ECM components such as fibronectin, laminin, collagen, cartilage proteoglycan, as well as pro-MMPs, leading to their activation. Thus, in addition to its role in fibrinolysis, the PA/plasmin system appears to be involved in cell migration and remodeling (38)
. Proinflammatory effects of plasmin also include activation of latent cytokines such as TGFβ, induction of proinflammatory cytokine expression, and TF (38)
. Activation of plasminogen activator/plasmin system also has indirect proinflammatory effects by mediating angiogenesis. In this respect, components of this system have been shown to be induced by VEGF and bFGF (88
, 89)
. All these systems appear able to interact with each other because kallikrein as well as coagulation factors XIa and XIIa also are able to activate plasminogen (90)
.
The complement system has long been known as an important mediator of inflammation (87)
. After being triggered by stimuli such as immune complexes or tissue antigens reacting with antibody or by microbial antigens, interactions and activation of complement proteins present in the serum proceed in a cascade, leading to the formation of a complex (C5b6789), which causes foreign cell lysis. The sequential interaction of the complement proteins also results in a variety of proinflammatory and cytolytic factors (87)
. This activation cascade involves proteases such as C3 and C5 convertases and leads to: 1) histamine secretion from mast cells and subsequent increased vascular permeability and SMCs contraction; 2) leukocyte chemotaxis, adhesion, aggregation, activation and subsequent release of IL-1, lysosomal enzymes, formation of eicosanoids such as leukotriene B4, and formation of reactive oxygen species (ROS); 3) platelet activation and aggregation and subsequent release of serotonin and thromboxanes; and 4) enhancement of phagocytosis (87)
. Interestingly, immune complexes and complements are thought to trigger the pathogenesis of RA (87
, 91)
.
It is reasonable to speculate that ADAM-15 could be implicated in the proteolytic regulation of such cascades of events or in the adhesion/aggregation process involved.
Shedding of immunoglobulin Fc receptors
Proinflammatory effects of immune complexes in RA are mediated not only by complement activation but also by binding to receptors for Fc fragments of immunoglobulins on neutrophils, such as CD16 or FcRIII, the low affinity IgG receptor that induces the release of cytokines (IL-1 and TNF-) and cytolytic enzymes, the formation of ROS, and the production of eicosanoids (91)
. However, Fc receptors also can undergo proteolytic cleavage resulting in their secretion. The capacity of ADAM-15 to catalyze CD23/FcRII ectodomain shedding has previously been discussed in this review. However, other Fc receptors undergo proteolytic cleavage and therefore could be ADAM-15 substrates. For example, soluble forms of FcRIII (sFcRIII) are released by neutrophil elastase (92)
upon neutrophil activation. Interestingly, sFc-RIII concentration has been shown to increase at inflammatory sites in RA and IBD (93
, 94)
. In addition, proteolytic cleavage of Fc receptor also can increase their binding affinity, as demonstrated for monocyte FcRII (95)
.
Initiation of intracellular signaling
In addition to the potential direct role of ADAM-15 in intracellular signaling through its cytoplasmic tail, ADAM-15 also could play an indirect role. For example, intracellular signals for vascular cell migration and proliferation are induced by the binding of uPA to its receptor uPAr, or indirectly by protease-mediated modifications of cell-cell or cell-matrix interactions such as laminin-5 cleavage by MMP-2 or E-cadherin release by MMP-3 and MMP-7 (46)
. It has been demonstrated that human bradikynin B2 receptor can directly be activated by kallikreins and some other proteases, independently of bradykinin release (86)
. Similarly, proteolytic activation of protease-activated receptors (PARs) initiates intracellular signals that lead to inflammatory processes (96)
. Proteases thought to be involved in PAR activation include proteases released by inflammatory cells, such as mast cell tryptase and neutrophil granzyme A and proteinase 3 (96)
. Neutrophil cathepsin G is known to cleave PAR4 on platelets, which induces their aggregation (97)
. Therefore, the interaction between endothelial ADAM-15 and platelet IIbβ3 integrin (18
, 27)
could possibly serve as an anchorage point for ADAM-15 to cleave such platelet receptor, leading to microthrombus formation (18)
. Interestingly, studies have shown that PAR1 and PAR2 contribute to the development of RA (96)
. While thrombin, which is elevated in RA synovium, is suspected as the protease that activates PAR1, the protease responsible for PAR2 activation still has not been identified (96)
. It is therefore reasonable to speculate that ADAM-15 could be responsible for PAR2 proteolytic activation in RA. Moreover, integrins also could initiate intracellular signals upon binding to ADAM-15 (98)
.
Because of its multiple potential substrates and ability to mediate cell adhesion, ADAM-15 is likely to play different roles at different stages and on different mediators of inflammation. Further studies are needed in order to establish the exact roles of ADAM-15 in inflammatory diseases. However, the findings cited in this review raise the possibility of ADAM-15 as a potential target in the therapy against inflammation.
|
CONCLUSIONS |
|---|
. Some of the therapies that have been elaborated against inflammatory diseases therefore consist in inhibition of cell adhesion molecules or inhibition of protease activity. Inhibition of protease activity raises the problem of the specificity of these inhibitors. No specific inhibitor of ADAM-15 has yet been identified. While inhibitors of metalloproteases, including ADAM proteins, have been available, their use is very delicate because proteases can play both pro- and anti-inflammatory roles, as has been discussed. For this reason, clinical application of MMP inhibitors has been limited.
Inhibition of ADAM-15-mediated adhesion seems a good option because inhibition of adhesion molecules has proven effective in inflammatory diseases. For example, antibodies raised against 4 or 4β1 integrins inhibit lymphocyte 4β1 integrin interactions with vascular cell adhesion molecule 1 (VCAM-1) and/or mucosal addressin cell adhesion molecule 1 (MadCAM-1), and present beneficial effects in patients with IBD (99)
. It is therefore reasonable to think that antibodies raised against ADAM-15 could be used for therapy against inflammatory diseases. The ability of anti-ADAM-15 antibodies to inhibit ADAM-15 functions already has been demonstrated. For example, we have shown that an antibody raised against ADAM-15 ectodomain was able to reduce T cell attachment to intestinal epithelial monolayers (62)
. Similarly, antibodies against ADAM-15 binding partners also could have antiadhesion effects. Indeed, an antibody raised against IIbβ3 integrin has been shown to inhibit platelet adhesion to ADAM-15 by 55–60% (18)
. The use of anti-ADAM-15 antibodies presents the advantage that not only the adhesion function but also the metallopotease activity of ADAM-15 could be inhibited. Martin et al. (11)
have shown that either a broad-spectrum MMP inhibitor or anti-ADAM-15 antibody inhibits mesangial cell migration, which suggests that anti-ADAM-15 antibody can inhibit ADAM-15 metalloprotease activity.
Soluble adhesion molecules also constitute a good alternative for inhibition of cell adhesion. Soluble adhesion molecules have been shown to inhibit interactions of inflammatory cells with endothelial cells (100)
. Concerning ADAM-15, ADAM-15 RDD has been demonstrated to inhibit angiogenesis (26)
, and to bind to the platelet surface, leading to inhibition of platelet IIbβ3 integrin/fibrinogen interactions (27)
.
An effective way to abolish ADAM-15 activity could be through inhibition of its expression. Such inhibition of expression can be achieved by using antisense oligonucleotides. This strategy already has been used for endothelial adhesion molecules such as ICAM-1, VCAM-1, or E-selectin and was demonstrated to reduce HL-60 cell adhesion to activated endothelial cells (100)
. ICAM antisense oligonucleotides have proven effective in vivo as inhibitors of ICAM-1 up-regulation and leukocyte recruitment in animal models of inflammatory diseases (100)
, presenting beneficial effects in patients with Crohn’s disease (99)
. ADAM-15 constitutes a target of choice for such a strategy since ADAM-15 possesses both adhesion and metalloprotease properties. Inhibition of ADAM-15 expression therefore would lead to inhibition of these two activities. Martin et al. (11)
have shown that the use of ADAM-15 antisense oligonucleotides inhibits mesangial cell migration.
Targeting proinflammatory mediators, such as cytokines or growth factors, that are responsible for ADAM-15 up-regulation during inflammation would be another way, although indirect, to inhibit ADAM-15 expression. Hitherto, studies have shown that ADAM-15 is up-regulated by PDGF, VEGF165, and the combination of TNF- and IFN (16
, 18
, 34)
. However, since these factors, in particular cytokines, are highly redundant and can have both stimulatory and inhibiting effects on immune system, this strategy might not be the most appropriate.
ADAM-15 appears to be a target of choice for therapeutic purposes since the inhibition of one protein could lead to the inhibition of two major events involved in inflammatory diseases. However, as previously discussed, the same protein or factor can have both pro- and anti-inflammatory effects. Therefore, the exact roles of ADAM-15 in inflammation should be elucidated before undertaking such therapeutic approaches.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 20, 2007. Accepted for publication August 30, 2007.
|
REFERENCES |
|---|
vβ3. J. Biol. Chem. 273,7345-7350vβ3 and 5β1 integrins on different haemopoietic cells. J. Cell Sci. 112,579-587[Abstract]9β1 to the ADAM-12 and -15 disintegrin domains mediates cell-cell interaction. J. Biol. Chem. 275,34922-349305β1and vβ3 in atherosclerosis. J. Cell. Biochem. 89,808-823[CrossRef][Medline]5β1and vβ3 integrins on smooth muscle cell spreading and migration in fibrin gels. Thromb. Haemost. 84,701-705[Medline] and Notch. Cell 90,589-592[CrossRef][Medline]6β1integrin, leading to a marked induction of fibroblast cell motility. J. Cell Sci. 113,2319-2328[Abstract]IIb β3) on platelets upregulates CD40L and triggers CD40L-dependent matrix degradation by endothelial cells. Circulation 106,2111-21175 βbeta1 with the central cell-binding domain of fibronectin. Am. J. Pathol. 156,1345-1362v integrins during angiogenesis. Cancer J. Sci. Am. 6,245-249v and β3 integrin subunits in cytokine stimulated fibroblast-like cells derived from synovial tissue of rheumatoid arthritis and osteoarthritis in vitro. Ann. Rheum. Dis. 56,729-7361, 3 and 5 chains of β1 integrins. Virchows Arch. 425,171-180[Medline]E β7 integrin. Nature 372,190-193[CrossRef][Medline] from cells. Nature 385,729-733[CrossRef][Medline]. J. Biol. Chem. 270,23688-23692 and IL-1β from a stimulated human monocytic cell line in the presence of activated neutrophils or purified proteinase 3. Proc. Natl. Acad. Sci. U. S. A. 96,6261-6266
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
