The role of ADAM 15 in glomerular mesangial cell migration.

Mesangial cells (MC) occupy the core of the renal glomerulus and are surrounded by a mesangial matrix. In certain diseases, MC migrate through this matrix into the pericapillary space. The mechanisms involved, however, are poorly understood. Members of the ADAM (A Disintegrin And Metalloproteinase) family of membrane proteins have the potential to be key modulators of cell-matrix interactions through the activities of their constituent domains. We have studied the possible role of ADAM 15 in human (H) MC migration in vitro. HMC ADAM 15 was expressed at low levels in serum-free medium but was increased during migration. Antibodies to the individual domains of ADAM 15 and the incorporation of antisense ADAM 15, (but not control oligonucleotide) inhibited this migration. Furthermore, inhibition of migration by the broad spectrum metalloproteinase inhibitor BB3103, demonstrated that metalloproteinase activity was essential for migration. ADAM 15, extracted from HMC membranes, was an active metalloproteinase, which degraded both type IV collagen and gelatin prepared from fibrillar collagen. Activity was inhibited by EDTA but not by phenylmethylsulfonyl fluoride. This is the first report of the potential of ADAM 15 for involvement in the restructuring of the mesangial matrix and in the migration of MC in disease.

Mesangial cell migration through the mesangial matrix and into the pericapillary space is a feature of a number of renal diseases, including mesangiocapillary glomerulonephritis (1). In addition, it appears that extracellular matrix components such as fibronectin (2), thrombospondin (3), and heparin-like glycosaminoglycans (4) can modulate this migration. This movement of cells must involve disengagement from and remodeling of the surrounding extracellular matrix. Possible mediators for this remodeling include a variety of serine proteinases and matrix metalloproteinases (MMPs) 1 but also the newly described ADAM family of molecules.
The ADAMs are a family of cell surface molecules that possess both disintegrin and MMP domains. The disintegrin domain of these molecules resembles the sequence of the snake venom disintegrins and binds to integrins on the cell surface (5). This displaces the integrin from its matrix target, thus freeing the cell from its substrate. Moreover the binding of a disintegrin to integrins on its own cell membrane in addition to those on adjacent cells is a possibility, freeing both cells from their substratum. Furthermore, because the ADAM molecules are proteolytically processed (6), cleavage of the ADAM may also take place, releasing the bound disintegrin from the cell and freeing the cell from the matrix. In addition, the binding of the disintegrin domain to integrins on adjacent cells may direct the MMP to the site of integrin/matrix interaction resulting in matrix degradation and facilitating cell migration.
The deduced structure of the MMP catalytic domain appears functional in many of the 30 or more ADAM family members (that is, there is an active site consensus sequence) (7), whereas in some it lacks the necessary zinc-binding site. The third extracellular region of the ADAM molecule is a cysteine-rich domain, which may promote membrane fusion (8,9). The transmembrane domain is adjacent to a cytoplasmic domain containing putative binding sites for several intracellular signaling molecules (6), suggesting that binding of the ADAM molecule, through either the disintegrin or MMP domain, may mediate signal transduction and affect cell function.
The structure of the ADAM molecules suggests that they are centrally involved in the remodeling of tissue that occurs following damage and/or disease. At present, however, the specific role of many of the ADAMs is limited. ADAMs 1-6 are believed to be involved in the binding and fusion of sperm to egg (10,11). Others are involved in the "shedding" of membrane-bound receptors and cytokines (12,13) or the activation of cytokines, for example TNF␣ by TACE (TNF␣-converting enzyme) (14).
We have recently demonstrated that the interaction of matrix proteins with their integrin receptors leads to an alteration in the production of metalloproteinases by glomerular cells (15). Because the ADAM molecules bind to the same receptors, they possess the potential for inducing the same intracellular effects as the matrix molecules themselves. The ADAM molecules may, therefore, affect the interaction of the cell with the extracellular matrix at many levels, and an increase in ADAM expression is likely to be centrally involved in the response of cells following injury.
To date little is known about the expression or potential role of ADAMs in the kidney. ADAM TS1, a novel ADAM protein that is released extracellularly and possesses a thrombospondin-binding site, is induced in several tissues (including the kidney) in mice following the systemic administration of lipopolysaccharide or phorbol ester (16). In addition, the expression of ADAM 12 (also known as MDC9) has been described in glomerular and tubular epithelial cells, and a role for this molecule has been suggested in interactions with the basal lamina (17). ADAM 15,MDC 15 (6), or Metargidin (18) has been shown to be present in cultured human aortic smooth muscle cells, although it is not seen in normal vessels in vivo. It is, however, up-regulated in atherosclerotic lesions (6). Our present report examines the expression of ADAM 15 by human mesangial cells in vitro and its involvement in the migration of these cells following injury.

EXPERIMENTAL PROCEDURES
Human Mesangial Cell (HMC) Culture and Identification-Human glomeruli were obtained by the serial sieving of normal human kidney cortex recovered at nephrectomy. HMC were maintained in RPMI 1640 containing 20% v/v FCS. The cells were confirmed as mesangial cells by morphology and by the use of immunohistochemistry as previously described (19). Mesangial cells showed positive staining for intracellular myosin fibrils and were negative for factor VIII and cytokeratin. All experiments were repeated using cells from at least three different sources. Before experimental procedures HMC were growth-arrested for 48h by culture in medium containing 0.2% (w/v) lactalbumin hydrolysate in the absence of serum (20).
Fusion Protein Source and Synthesis-ADAM 15 cytotail/GST fusion protein plasmid was kindly provided by Dr. Carl Blobel, Sloane-Kettering Institute, Memorial Sloane-Kettering Cancer Centre, New York. This protein was purified by the use of glutathione beads, characterized, and used to raise antiserum by standard techniques.
Preparation of anti-ADAM 15 Antibodies-Antibodies were raised in rabbits to each of the individual domains (cytoplasmic, metalloproteinase, and disintegrin) of the ADAM 15 molecule. Synthetic peptides corresponding to these domains (6) were synthesized (Enzyme Research Laboratories Ltd., Swansea, UK) and used to immunize rabbits by standard protocols. Serum was collected at monthly intervals, and the specificity and titer of the antisera confirmed by Western blot analysis on cell lysate preparations from HMC. Each anti-ADAM 15 antibody was affinity-purified by binding to the appropriate synthetic peptide bound to peptide affinity columns (SulfoLink, Pierce) followed by elution with 100 mM glycine buffer, pH 2.5. The specificity and purity of the antibodies was monitored by Western blot analysis and by silver stain of the purified immunoglobulin. Each was immunoreactive against the intact molecule, in addition to the specific domain against which it was raised. None was cross-reactive with any of the other peptides that had been synthesized Purification of ADAM 15-HMC lysate was applied to an anti-ADAM 15 cytoplasmic domain antibody-linked affinity column. The column was washed with phosphate-buffered saline and bound ADAM 15 was eluted in 1-ml fractions with 100 mM glycine buffer, pH 2.5, and neutralized with 50 l of 1 M Tris, pH 9.5. The purity of the preparation was monitored by SDS-PAGE and silver staining, and the identity of the eluted protein was assessed by Western blotting with the antibodies to the other ADAM 15 domains. A protein was eluted that ran with the anticipated molecular masses of around 135 kDa as previously described by Blobel and co-workers (21) when visualized by silver staining in polyacrylamide gels or by Western blotting (see Fig. 1).
Western Blotting-Western blotting was carried out as described by Herron et al. (6). Cells were lysed with 0.5% Nonidet P-40 in TBS, pH 7.5, containing 5 mM EDTA, 1 mM PMSF, 10 g/ml soya bean tripsin inhibitor, 2 g/ml leupeptin, 4 g/ml aprotinin, and 2 g/ml pepstatin A. Equal amounts of protein were loaded in each well and then sepa-rated by 12% SDS-PAGE and transferred onto nitrocellulose membrane. ADAM 15 was then detected with purified antibody and visualized by ECL (Amersham Biosciences).
Cell Proliferation-Cell proliferation was assessed in two ways. (i) Cells were grown either in the presence of serum or in serum-free conditions for 48 h, and BrdUrd (bromodeoxyuridine) uptake during the following 24 h was measured. BrdUrd (Sigma) was added to the cells at a concentration of 0.1 mM in medium, and the incubation continued at 37°C. The amount of BrdUrd incorporated into cell DNA was determined by the addition of mouse anti-BrdUrd antibodies (Amersham Biosciences), followed by two cycles of anti-mouse immunoglobulin secondary antibodies and the application of APAAP complex (DAKO Ltd., Cambridgeshire, UK), and finally visualized by staining with Fast Red. (ii) The MTT assay was used as previously described (20) to determine the relative numbers of HMC following treatment.
Northern Blot Analysis-Total RNA (up to 40 g) was run on a denaturing gel and transferred by vacuum blotting onto a nylon membrane (Amersham Biosciences, Hybond Nϩ). ADAM 15 mRNA was detected by hybridization with a 32 P-labeled probe prepared from PCR products and detected as previously described (15) Cell Migration Assay-HMC were sub-cultured into gridded (2-mm square) 35-mm Petri dishes (Nunc International) and grown to confluence. Cells were growth-arrested and then were carefully removed from half of the dish, up to the edge of a line of grids, using a cell scraper. The culture was then continued for a further 96 h in the absence of FCS or for 48 h in the presence of FCS at which time the cells were fixed with paraformaldehyde (3% v/v) and stained with a solution of crystal violet (0.5% w/v) for 3 mins. The cells were washed thoroughly with phosphate-buffered saline, and the number of cells that had migrated into the squares adjacent to the line of origin was determined.
To examine the effect of MMP inhibition on mesangial cell migration, the hydroxamine inhibitor BB-3103 (British Biotech Pharmaceuticals Ltd, Oxford, UK), prepared in Me 2 SO, was initially added to the cells at a range of concentrations from 100 nM to 10 M. For the subsequent inhibition experiments, the inhibitor was used at a concentration of 1 M, a concentration at which there were no toxic effects on the cells. To determine the effect of the anti-ADAM antibodies on HMC migration, the affinity-purified antibodies were added at a concentration of 2 g/ml to the dishes following the removal of cells, and the migration was determined as before. To collect RNA and cellular protein from the maximum number of migrating cells, this cell migration protocol was subsequently modified to allow the infliction of multiple scratches using a 10-tooth comb on confluent layers of HMC in 35-mm Petri dishes as described by Kinsella and Wight (22).
In a second migration system, HMC were seeded into 8-well chamber slides (Permanox Slides, Lab-Tek) and grown to confluence. Cells were growth-arrested as before and an area of each well was then carefully denuded of cells using a plastic pipette tip. The migration of the cells into the denuded area was then followed by time-lapse photography, and the area covered by the migrating cells was measured using an "Openlab" software package. (Improvision, UK) A third method of studying cell migration was used in experiments to analyze the effect of antisense ADAM 15 oligonucleotides. This migration assay was performed as described by Belien et al. (23) using a "cell sedimentation manifold" (CSM Inc., Phoenix, AZ). Ten-well Teflon printed microscope slides were precoated with 1% bovine serum albumin in phosphate-buffered saline for 30 min at room temperature before use and stored at 4°C. Medium (50 l) was added to each well, and the cell sedimentation manifold was placed on the slide. Two thousand cells were added in a volume of 1 l to each chamber and allowed to sediment and adhere to the slide for 1 h on ice then overnight at 37°C. The manifold was then removed, and the medium was aspirated and replaced with 40 l of fresh medium in the presence or absence of the test reagents. The area covered by cells was recorded at regular intervals, and the ratio of each area to that at the time of addition of the test materials was quantified by image analysis as above.
Effect of Antisense ADAM 15 Oligonucleotides on Migration-Antisense ADAM 15 oligonucleotides together with matched controls were purchased from Biognostic (TCS Biologicals, Buckingham, UK) and added at a concentration of 2 M to confluent HMC in 8-well chamber slides. The efficiency of transfection was determined using an fluorescein isothiocyanate-labeled oligonucleotide. This demonstrated Ͼ90% uptake of the oligonucleotide within 8 h. For experimental procedures the medium containing the oligonucleotide was therefore aspirated after 8 h and an area of cells was removed as before. The remaining cells were then washed, and the oligonucleotide-containing medium was replaced in the well. The effect of the antisense oligonucleotide compared with a matched control oligonucleotide was determined by timelapse photography as described. In additional experiments, the effect of the antisense ADAM 15 oligonucleotide on migration was studied in the cell sedimentation manifold system as described, using the same protocol.
Collagen Degradation Assay-Affinity-purified ADAM 15 (at concentrations up to 1.111 g) was incubated with Collagen IV (10 g) (Sigma) in vitro at 37°C for up to 96 h. The reaction was carried out in a final volume of 25 l, containing 5 l of assay buffer (400 mM Tris, 10 mM Ca 2ϩ , pH 8.0) in the presence or absence of the inhibitors EDTA (10 mM), PMSF (10 g/ml), or leupeptin (2 g/ml). At the end of the incubation period the reaction was terminated by addition of reducing PAGE buffer and boiled and separated on a 7.5% polyacrylamide gel. The degradation of collagen was assessed by Western blotting of the gel using rabbit anti-human Type IV collagen antibody (ICN, Basingstoke, UK) and visualized by ECL (Amersham Biosciences, Bucks, UK).
Zymography-Zymography of ADAM 15 was carried out as previously described (15) using 7.5% polyacrylamide gels incorporating gelatin at a concentration of 1 mg/ml. In some experiments APMA (paminophenylmercuric acetate) (Sigma) was preincubated with the

Western Blot and PCR Analysis of ADAM 15 in HMC-
Following extraction, ADAM 15 mRNA was shown by RT-PCR to be present in growth-arrested HMC and was up-regulated in the presence of serum ( Fig. 2A). ADAM 15 protein was barely detectable by Western blotting in growth-arrested HMC lysates. Its expression was increased, however, by incubating the cells with FCS (Fig. 2B).

ADAM 15 Expression in a Multiscratch
Model of Cell Injury-Following the induction of multiscratch wounds in confluent HMC monolayers, ADAM 15 mRNA levels increased rapidly to a maximum at around 4 -6 h and remained elevated for up to 24 h (Fig. 3A). The RT-PCR data was confirmed by Northern blotting of total cellular RNA collected 3 h following the infliction of multiscratches (Fig. 3B). In addition, ADAM 15 protein expression was shown by Western blotting to be increased at 24 h (Fig. 4).
ADAM 15 Mediates HMC Migration-The potential role of ADAM 15 in cell migration was investigated using the domainspecific antibodies. In the presence of serum, the antibodies to the disintegrin and metalloproteinase domains significantly decreased HMC migration (Fig. 5).
To establish that the observed changes were not due to an effect on cell proliferation, HMC were stained with BrdUrd to identify the number of proliferating cells. The anti ADAM 15 antibodies had no effect on HMC proliferation. This finding was confirmed by assaying cell number using the MTT assay. These results suggested that ADAM 15 expression was essential for HMC migration. To confirm this, the effect of transfecting the cells with the antisense ADAM 15 oligonucleotide was determined. Initially migration was monitored in 8-well chamber slides by time-lapse photography. The antisense ADAM 15 oligonucleotide decreased the rate of migration relative to controls and to the matched control oligonucleotide, (which itself had no significant effect on migration) (Fig. 6). Migration was also inhibited after cell sedimentation, where the area of the expanding circle of cells was measured over time. There was a significant effect of the antisense ADAM 15 oligonucleotide relative to controls after 72 h of incubation, whereas the matched control oligonucleotide had no significant effect (Fig. 7).
RT-PCR for ADAM 15 in cells exposed to the antisense ADAM 15 oligonucleotide demonstrated a reduction in the amount of ADAM 15 mRNA relative to control cells (Fig. 8A). In addition, antisense ADAM 15 decreased the induction of ADAM 15 following multiscratching (Fig. 8B).
Metalloproteinase Activity Is Integral to HMC Migration-MMP activity is involved in migration in many cell systems. The effect on HMC migration, of inhibiting MMP activity, was examined using the hydroxamine inhibitor BB-3103. Preliminary experiments determined that the optimal effective nontoxic concentration of BB-3103 was 1 M. At this concentration BB-3103 inhibited by 36% the number of cells migrating into 2-mm squares (Fig. 9). This supported previous reports of MMP involvement in migration (24), and the degree of inhibition observed was similar to the maximum achieved by incubation with the domain-specific antibodies to ADAM 15.
ADAM 15 Is a Metalloproteinase with Gelatinolytic Activity-The data presented above demonstrate a role for ADAM 15 in HMC migration and together with the BB3103 inhibition data suggest that the MMP domain of the molecule may be actively involved. The possibility that this domain is proteolytic, however, has not previously been addressed in any system. The first possibility examined was that ADAM 15 might act in a similar manner to the membrane type MMPs (MMP14 -17) and convert other latent MMPs to an active form.
Affinity-purified ADAM 15 was, therefore, incubated for up to 72 h with HMC-conditioned medium containing latent MMP2 (gelatinase A). The incubation mixture was then analyzed by gelatin zymography. No shift in the zone of lysis corresponding to the activation of latent MMP2 was observed. There was, however, an obvious zone of lysis in the non-reduced gel at a higher molecular size, which was present only in those lanes that contained the purified ADAM 15 (Fig. 10). This activity was not increased by prior incubation with APMA.
To confirm this proteolytic activity, affinity-purified ADAM 15 was incubated with Collagen IV in vitro at 37°C for up to 96 h. There was a dose-dependent degradation of this basement membrane protein (Fig. 11), which was inhibited by the addition of EDTA (10 mM) but was not affected by the serine proteinase inhibitors, PMSF and leupeptin (Fig. 12).
Inhibition of ADAM 15 Proteolytic Activity by Anti-ADAM 15 Antibodies-Incubation of purified ADAM 15 with the fluorogenic substrate Dnp-Pro-␤-cyclohexyl-Ala-Gly-Cys(Me)-His-Ala-Lys(N-Me-Abz)-NH 2 resulted in substrate cleavage, measured by strength of fluorescent emission at 460 nm. This cleavage was inhibited both by the antibody to the metalloproteinase domain of the ADAM 15 molecule and by the antibody to the disintegrin domain (Fig 13). No inhibition of proteolytic activity was seen when non-immune IgG was substituted for the antibodies in the incubations (not shown).

DISCUSSION
The role of the ADAM molecules in tissue homeostasis is poorly understood. These molecules have separate domains that would enable the cell to both interfere with integrin binding to the extracellular matrix and to degrade that matrix, suggesting a role in cell migration. However, their ability to influence cell migration has not previously been demonstrated. We have shown here for the first time that the ADAM 15 molecule may be involved in the migration of human mesangial cells following the initiation of migration in vitro. Low levels of ADAM 15 mRNA were detected by RT-PCR in HMC and were increased following exposure to serum. Subsequently, the levels of ADAM 15 mRNA and protein were both shown to increase following the initiation of migration, suggesting a degree of transcriptional up-regulation. The fact that the antisense oligonucleotide and antibodies raised against the separate extracellular domains of the ADAM 15 molecule decreased the migration of HMC in vitro further suggested that there was a causal relationship between the level of ADAM 15 and the rate of movement of the HMC.
An important finding that may indicate the mode of action of ADAM 15 in HMC migration was its ability to degrade the extracellular matrix protein collagen IV and denatured collagen I (gelatin). This proteolytic activity in vivo would not only facilitate cell migration by allowing the degradation of the mesangial matrix and cell detachment but could also promote the movement of the cell through its surrounding ECM barrier. These findings with ADAM 15 are similar to those of other members of the ADAM family that have been shown to possess proteolytic activity, such as ADAM 12 (Meltrin ␣) (28) and ADAMTS-1 (29). Both of these molecules have been shown to be active metalloproteinases through their binding to ␣-2 macroglobulin. In addition, ADAM 10 (MADM) has been reported to possess type IV collagenase activity in vitro (30). Although the mechanism of inhibition by antisense was through a direct effect on ADAM 15 expression we hypothesized that the antibodies inhibited HMC migration through an effect on the tertiary structure of the ADAM 15 molecule, which may have inhibited its metalloproteinase activity. This was supported using a sensitive fluorometric assay to investigate the effect of the antibodies on the proteolytic activity of purified ADAM 15. Both the antibody raised against the metalloproteinase domain and that raised against the disintegrin domain inhibited the cleavage of the substrate, whereas control IgG had no effect. This result, taken together with the effect of the BB-3103 proteinase inhibitor on migration, imply that at least part of the inhibitory activity of both the anti-metalloproteinase and anti-disintegrin antibodies on the migration of the HMC in vitro is due to their inhibition of the enzymatic activity of ADAM 15. Neither the antibodies to the extracellular domains of the ADAM 15 molecule nor the metalloproteinase inhibitor BB-3103 affected HMC proliferation as measured by labeling with BrdUrd. This was somewhat unexpected, as it has been previously reported that the metalloproteinase activity of MMP-2 is involved in the proliferation of mesangial cells (25). Our find- , purified from HMC lysates were incubated with 10 g of collagen IV for 96 h. The mixture was then separated on a 7.5% polyacrylamide gel and transferred to nitrocellulose, and the collagen was visualized by immunoblotting with anti-collagen IV antibody as described.
FIG. 12. ADAM 15 is a metalloproteinase. Collagen IV (10 g) was incubated at 37°C for 72 h alone (a) or with ADAM 15 (1 g) purified from HMC as above in the absence of proteinase inhibitors (b), in the presence of 10 mM EDTA (c), in the presence of 250 g/ml leupeptin (d), and in the presence of 1 mM PMSF (e). The mixture was then separated on a 7.5% polyacrylamide gel, and the collagen was visualized by immunoblotting with anti-collagen IV antibody as described. ings were confirmed by MTT assay where there were no changes in cell number or viability following incubation with either the antibodies or the metalloproteinase inhibitor. These results confirm a separate series of studies suggesting that factors affecting proliferation in MC are separate from those that effect migration (26,27).
The RGD integrin-binding sequence of ADAM 15 (18) is reported to facilitate the interaction of this molecule with both the ␣ V ␤ 3 and ␣ V ␤ 1 integrin (32,33). Considered together with its proteolytic activity, this gives ADAM 15 the potential for involvement in cell migration at several levels. The migration of the mesangial cell from the glomerulus into the pericapillary space in diseases such as mesangiocapillary glomerulonephritis is a major feature of this disease. Little is known about the mechanisms controlling the migration of the mesangial cells, although it has been reported that MC stimulated with a combination of TNF-␣, IL-1␤, and IFN-␥ but not unstimulated MC migrated toward a RANTES gradient (34). It has also been reported that heparin (4), platelet secretory products, and platelet-derived growth factor (2,35) can stimulate their migration in vitro.
The work presented here describes a possible mechanism by which HMC migration is mediated. We have demonstrated for the first time that ADAM 15 is proteolytically active and that this activity is associated with the migration of HMC. Because inhibition of migration was only partially complete in our system, however, it is likely that there are additional molecules, possibly including other members of the ADAM family, that also contribute. Thus, in both mesangial and other cell types, a degree of redundancy may exist. A recent report (31) that ADAM 9 can also modulate cell migration, suggests that this is an important role of this family of molecules. The inhibition of cell migration by selective matrix metalloproteinase inhibitors may provide an anti-inflammatory role for these compounds. This may allow a novel approach to therapeutic strategies with targeted inhibition of such molecules preventing chronic progression of disease.