EGF receptor (EGFR) promotes intestinal epithelial restitution, an important early process in the reepithelialization of ulcers. During epithelial restitution, the mechanism of EGFR activation is not known. We evaluated the role of TNF-converting enzyme (TACE), a metalloprotease disintegrin that proteolytically processes plasma membrane-anchored EGFR ligand precursors into their mature active forms, in wound-induced EGFR activation and epithelial restitution. With the use of scrape-wounded rat intestinal epithelial-1 (RIE-1) cell monolayers to model epithelial ulceration and restitution, we observed the rapid wound-dependent release of EGFR ligands into culture medium. RIE-1 cells express TACE, and treatment with phorbol ester, an established TACE stimulus, triggered the extracellular release of an EGFR ligand, transforming growth factor-α. Blockade of TACE using TNF processing inhibitor (TAPI-1), a specific hydroxamate inhibitor of metalloprotease disintegrins, prevented release of EGFR ligands from wounded RIE-1 cell monolayers. The restitution of wounded RIE-1 cell monolayers was also dose-dependently inhibited by TAPI-1, establishing the role of metalloprotease disintegrins in this process. These results have established a mechanism of EGFR activation in wounded intestinal epithelium and show an important functional role for metalloprotease disintegrin-mediated ectodomain shedding during intestinal epithelial restitution. Therefore, activation of the TACE-EGFR system might promote the healing of intestinal tract ulcers in patients.
- RIE-1 cells
ulceration of the intestinal tract is a hallmark of active inflammatory diseases, such as inflammatory bowel disease (41). Areas of epithelial denudation erode the normal physiological barrier between the host and the intestinal lumen, allowing access of luminal contents, such as microbes and food particles, to the intestinal mucosa and submucosa, which incites inflammation. Ulcers also contribute to blood loss and diarrhea in inflammatory bowel disease patients, and the extent and depth of ileocolonic ulcers are strongly associated with disease progression and complications in Crohn's disease, one form of inflammatory bowel disease (1). The importance of ulcers in the causation of symptoms of inflammatory bowel disease is also highlighted by the ability of the most effective therapies for colitis to heal ulcers (14).
The healing of intestinal tract ulcers is initiated by migration of ulcer-edge epithelial cells into the area denuded of epithelium (2, 22, 23, 29). This process, termed epithelial restitution, results in rapid reepithelialization of the ulcer and is followed by a phase of epithelial cell proliferation and remodeling of the epithelium into the normal, highly convoluted architecture. Prior investigations have unraveled some of the mechanisms of regulation of epithelial restitution in the intestinal tract. Peptide growth factors, such as fibroblast growth factor, keratinocyte growth factor, and EGF have been shown to stimulate intestinal epithelial restitution by increasing the migration of cells from the wound edge into the ulcer via a mechanism that depends on transforming growth factor-β (TGF)-β (4, 11, 16, 18). Certain cytokines, such as interleukin-1 and TNF-α, at low concentrations also promote intestinal epithelial restitution (13, 17), as do lysophosphatidic acid (42) and intestinal trefoil factor (15).
Growth factors act on target cells in the intestine by binding to and inducing dimerization of receptor tyrosine kinases, notably the EGF receptor (EGFR). EGFR is a member of the ErbB family of growth factor receptors and constitutes a homodimer of ErbB1 receptors (reviewed in Ref. 48). EGFR is the receptor for several related growth factors including EGF, heparin-binding EGF-like growth factor, amphiregulin, and TGF-α. This receptor signals to downstream intracellular pathways ultimately leading not only to increased cell movement, but also to survival and proliferation, among many other effects. Previously, we (20) and others (38) demonstrated, in an in vitro model of intestinal epithelial wounding followed by healing, that the EGFR is activated very early (within minutes) in the restitution process without the addition of any exogenous ligands for this receptor. Moreover, experiments with inhibitors of the EGFR have shown that activation of this receptor strongly promotes intestinal epithelial restitution (38). Thus although the importance of the EGFR in intestinal epithelial restitution is established, the mechanism of its activation has remained obscure.
EGFR ligands are synthesized as precursor molecules, such as pro-TGF-α, which are expressed in a plasma membrane-anchored form (48). Recent studies (8) have shown that cell- surface retention of pro-TGF-α prevents its biological activity as a ligand for EGFR and that release of ligands such as TGF-α from cells into the extracellular milieu is an important step in EGFR activation. The release of TGF-α and other EGFR ligands from the plasma membrane requires the proteolytic processing of precursor forms into active growth factors by the zinc-dependent metalloprotease disintegrin TNF-converting enzyme (TACE), also known as a disintegrin and metalloprotease-17 (8, 37, 43). This process has been termed ectodomain shedding and is required not only for the extracellular release of certain growth factors, but also of some cytokines, cell surface receptors, and a variety of other proteins (40). We investigated a potential role for metalloprotease disintegrin-mediated ectodomain shedding of EGFR ligands in the activation of the EGFR in wounded intestinal epithelium. Our results show that intestinal epithelial cells express TACE. Metalloprotease disintegrin activity is required for the shedding of EGFR ligands from wounded monolayers of rat intestinal epithelial-1 (RIE-1) cells, leading to the autocrine or paracrine activation of this receptor. Importantly, functional studies using a specific pharmacological inhibitor of metalloprotease disintegrin activity demonstrate a role for ectodomain shedding of EGFR ligands in epithelial restitution in the intestinal tract.
MATERIALS AND METHODS
Cell lines and reagents.
RIE-1 cells (6, 7) (originally provided by K. Brown, Babraham Institute, Cambridge, UK) were maintained in DMEM (Invitrogen, Carlsbad, CA) containing 2 mM l-glutamine and 5% fetal bovine serum (Omega Scientific, Tarzana, CA) at 37°C in 95% air-5% CO2. HT-29 cells (American Type Culture Collection, Manassas, VA) were grown under the same conditions as RIE-1 cells but with 10% fetal bovine serum. Unless otherwise stated, reagents were obtained from Sigma (St. Louis, MO).
Wounding of RIE-1 cell monolayers.
To prepare whole cell protein extracts, cells were grown to confluence in 10-cm tissue culture dishes. Before wounding, cells were exposed to medium containing 0.5% fetal bovine serum for 30 min. The monolayer was cross-scraped in each of four directions for a total of 25, 50, or 100 scrapes with a 200-μl pipette tip to maximize the length of the wound edge. After 2.5 min, the monolayer was washed with cold PBS and prepared for immunoprecipitation and Western blot analysis.
Preparation of wound-conditioned medium.
Confluent monolayers were wounded as described above. After 2.5 min, wound-conditioned medium was collected, and floating cells were removed by spinning in a 4°C microcentrifuge at 3,000 rpm for 5 min. The wound-conditioned medium was then added to fresh monolayers of RIE-1 or HT-29 cells, and after 2.5 min, the cells were washed with cold PBS and prepared for immunoprecipitation and Western blot analysis as described below. Similar results were obtained when the duration of exposure to wound-conditioned medium was 5 min.
Immunoprecipitation and Western blot analysis.
Cells were scraped into lysis buffer containing (in mM) 50 Tris·HCl, pH 7.4, 150 NaCl, 1 EDTA, 40 β-glycerophosphate, 1 orthovanadate, and 1 sodium fluoride, plus 1% Nonidet P-40, 0.25% sodium deoxycholate, and 1:200 protease inhibitor cocktail set III (Calbiochem, San Diego, CA). Lysate containing 500 μg protein was incubated with anti-EGFR antibody at 4°C overnight on a rocking platform. Protein A/G Plus agarose (Santa Cruz Biotechnology, Santa Cruz, CA) was added for 1 h, after which washed immunoprecipitates were separated by SDS-PAGE on 7.5% acrylamide/Tris·HCl gels. Proteins were transferred to nitrocellulose and blotted with anti-phosphotyrosine antibody (clone 4G10; Upstate Biotechnology, Charlottesville, VA), and were stripped and reblotted with anti-EGFR antibody. Bound primary antibodies were detected by using peroxidase-conjugated secondary antibodies and ECL reagents (Amersham Biosciences, Piscataway, NJ).
The concentration of TGF-α secreted into culture medium was quantified by using a commercial radioimmunoassay, according to the manufacturer's instructions (Peninsula Laboratories, San Carlos, CA). Briefly, RIE-1 cells grown in six-well plates or in 10-cm dishes were pretreated with TNF processing inhibitor (TAPI-1) or DMSO vehicle for 30 min before wounding. After varying times, culture medium was collected and centrifuged in a microcentrifuge at 3,000 rpm for 5 min to remove cells. One milliliter of cell-free medium was concentrated by using Mircocon YM-3 cartridges, nominal molecular mass cut-off at 3 kDa (Millipore, Billerica, MA). TGF-α was quantified in concentrated medium using a radioimmunoassay with 125I-labeled TGF-α as tracer. Unknown concentrations of TGF-α were determined by comparison with a standard curve generated by using recombinant TGF-α, which was processed in parallel with the experimental samples in the same matrix (tissue culture medium).
Evaluation of TACE expression.
Expression of TACE was evaluated by indirect immunofluorescence in RIE-1 cells cultured on glass coverslips. Cells were washed with PBS, fixed with 4% paraformaldehyde for 10 min at 4°C, and washed again three times. Blocking and permeabilization were carried out with PBS containing 0.2% Triton X-100 and 5% normal goat serum for 10 min at room temperature. Cells were then incubated overnight at 4°C with equal amounts of normal rabbit IgG or rabbit anti-TACE IgG (diluted 1:400) in PBS containing 5% normal goat serum. After being washed three times with PBS, cells were incubated with 1:500-diluted Cy3-conjugated goat anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA) for 45 min, washed three times, and mounted on 50% glycerol/PBS. Staining was visualized with an epifluorescence microscope fitted with appropriate filters. Images were captured digitally with a Zeiss Axiocam.
Neutralizing anti-EGFR antibody.
HT-29 cell monolayers were exposed to 0, 2.5, 5, or 10 μg/ml normal mouse IgG or neutralizing mouse anti-human EGFR IgG clone LA1 (Upstate Biotechnology) for 30 min in medium containing 0.5% fetal bovine serum. Wound-conditioned medium from RIE-1 monolayers was supplemented with the same concentrations of mouse IgG or neutralizing mouse anti-human EGFR IgG and was added to the HT-29 monolayers. After 2.5 min, the HT-29 cells were washed with cold PBS and prepared for immunoprecipitation and Western blot analysis as described in Immunoprecipitation and Western blot analysis.
Inhibition with TAPI-1.
RIE-1 monolayers were incubated with equal volumes of DMSO or 1–125 μM TAPI-1 (Calbiochem, San Diego, CA) for 30 min before wounding or before stimulation with recombinant human EGF 5 ng/ml (Biosource, Camarillo, CA) in medium containing 0.5% fetal bovine serum.
Quantification of epithelial restitution.
To evaluate the effects of TAPI-1 on epithelial restitution, we utilized the model of RIE-1 cell migration after wounding (20). We adapted this model to allow direct measurement of the area of circular wounds set in confluent RIE-1 cell monolayers by image analysis. RIE-1 cells were grown on 18-mm glass coverslips placed in 12-well tissue culture plates. Thirty minutes before wounding, medium was changed to fresh medium containing 0.5% fetal bovine serum supplemented with varying amounts of TAPI-1, AG-1478 (Calbiochem), or an equal volume of DMSO vehicle. Wounds were set by using a flat steel scraper mounted in a commercial drill press that was lowered onto the RIE-1 monolayers and rotated by hand for three revolutions. After wounding, fresh medium was replaced. At various time points after wounding, the cells were washed with PBS and fixed with 4% paraformaldehyde. The monolayer surface area denuded of cells was quantified by computer analysis of digital images acquired by using a ×10 phase-1 objective fitted to a Zeiss Axiovert microscope with an Axiocam digital camera. Calibration was performed with a stage micrometer. In each experiment, %closure was determined from the ratio of residual wound area after 16 h to the average initial size of the wounds. The average initial surface area of wounds was 889,226 μm2, and the coefficient of variation of wound size was 4.2% (n = 8). Typically, six wounds were set for each experimental condition. Preliminary studies revealed that circular wounds closed in a linear relationship with time (not shown).
In all cases, experiments were performed on at least three occasions, and representative results from a single experiment are shown. ANOVA with post hoc Tukey's tests or Student’s unpaired two-tailed t-tests were carried out as appropriate. P values of <0.05 were considered significant.
Wounded RIE-1 monolayers release EGFR ligands.
As a first step toward determining the mechanism of EGFR activation during intestinal epithelial wounding and restitution, we evaluated whether wounding of RIE-1 monolayers induces the release of EGFR ligands. For these experiments, confluent monolayers of RIE-1 cells were left unwounded or varying numbers of wounds were set, and after 2.5 min, wound-conditioned culture medium was collected, cleared of cells, and transferred to fresh unwounded RIE-1 cell monolayers. After a further 2.5 min, the conditioned medium-stimulated RIE-1 monolayers were lysed and EGFR activation was evaluated by antiphosphotyrosine immunoblotting of EGFR immunoprecipitates. Wound-conditioned medium induced a dose-dependent phosphorylation of EGFR tyrosine residues, indicating that wounded RIE-1 monolayers rapidly release EGFR ligands into culture supernatant (Fig. 1A). Some prior reports (47, 48) have indicated that the EGFR can be phosphorylated in a ligand-independent manner, for example by transactivation through intracellular intermediate kinases, such as c-Src. To exclude ligand-independent mechanisms as potential explanations for the EGFR phosphorylation induced by wound-conditioned medium, we evaluated the effect of a neutralizing antibody against the EGFR in this model. Because neutralizing anti-EGFR antibodies are available only against human but not rodent EGFR, we transferred wound-conditioned medium from RIE-1 monolayers onto unwounded monolayers of human colon cancer HT-29 cells. HT-29 cells were pretreated with neutralizing anti-EGFR antibody or control mouse IgG for 30 min before and during stimulation with wound-conditioned medium. We observed inhibition of EGFR phosphorylation by neutralizing anti-EGFR antibody, but not by control IgG (Fig. 1B). We noted slightly greater recovery of EGFR from samples treated with either control or neutralizing anti-EGFR IgG for reasons that are unclear. These results confirm that EGFR phosphorylation in this model depends on the presence of EGFR ligands in wound-conditioned medium and is not due to an indirect ligand-independent mechanism.
The release of EGFR ligands from wounded RIE-1 monolayers was further characterized by measuring the concentration of one EGFR ligand, TGF-α, in the supernatant of control and wounded RIE-1 monolayers. Baseline TGF-α concentration of medium exposed to unwounded monolayers for 30 min was not significantly different than that of fresh medium. The concentration of TGF-α in cell culture supernatant was increased by approximately twofold after wounding, compared with unwounded monolayers (Fig. 1C). Compared with the capacity of wound-conditioned medium to phosphorylate the EGFR, the increase in TGF-α that we measured in this medium was low, perhaps indicating that TGF-α is not the predominant EGFR ligand released by RIE-1 cells. Furthermore, the concentration of TGF-α that is measured in culture medium reflects the amount remaining that does not bind EGFRs. Together, these results indicate that wounded RIE-1 monolayers release TGF-α, and probably other EGFR ligands, into culture supernatant.
RIE-1 cells express TACE.
Prior reports have revealed a role for the metalloprotease disintegrin TACE in the ectodomain shedding of EGFR ligands (8, 37, 43). We therefore evaluated TACE expression in RIE-1 cells. Indirect immunofluorescence of TACE revealed predominant plasma membrane pattern of staining in RIE-1 cells (Fig. 2). Isotype control IgG did not produce any staining of RIE-1 cells. We also stimulated RIE-1 cells with phorbol ester, a well-characterized inducer of TACE activity (9, 34, 44), to evaluate whether intestinal epithelial cells are capable of inducible shedding of EGFR ligands. Phorbol ester caused rapid release of TGF-α into tissue culture supernatant, which was inhibited by preincubation of cells with TAPI-1, a selective hydroxamate inhibitor of metalloprotease disintegrins (26, 30, 46), which supports the functional importance of TACE in RIE-1 cells (not shown).
EGFR ligands are released from wounded RIE-1 monolayers by ectodomain shedding.
The rapid release of EGFR ligands from wounded RIE-1 monolayers suggested strongly that the growth factors existed in a presynthesized form ready for immediate release from the wounded cells. To evaluate a potential role for metalloprotease disintegrin-mediated ectodomain shedding in the release of EGFR ligands from wounded RIE-1 cell monolayers, we preincubated monolayers with TACE inhibitor TAPI-1 or vehicle before wounding and collected wound-conditioned medium. EGFR phosphorylation was dose-dependently inhibited in fresh monolayers stimulated with conditioned medium prepared from 2.5-min wound-conditioned medium of TAPI-1-treated monolayers (Fig. 3A). When EGFR phosphorylation was directly evaluated in wounded RIE-1 cell monolayers pretreated with TAPI-1, inhibition of wound-induced EGFR activation was also observed (Fig. 3B). In contrast, TAPI-1 did not inhibit EGF-induced EGFR activation, indicating the selectivity of this inhibitor. We also quantified the concentration of TGF-α in culture medium supernatant from vehicle or TAPI-1-treated RIE-1 cell monolayers after wounding and observed dose-dependent inhibition of TGF-α release by TAPI-1 (Fig. 3C). These results indicate that wounding of RIE-1 cell monolayers results in metalloprotease disintegrin-mediated ectodomain shedding of EGFR ligands.
Metalloprotease disintegrin-mediated ectodomain shedding of EGFR ligands promotes intestinal epithelial restitution.
Prior studies (13, 20, 38) demonstrated that autocrine or paracrine EGFR activation promoted intestinal epithelial restitution. Because we found that EGFR activation in this model depends on ectodomain shedding of ligands for this receptor, we reasoned that ectodomain shedding would regulate intestinal epithelial restitution. We therefore tested the effect of blockade of metalloprotease disintegrin activity with TAPI-1 on epithelial restitution. Monolayers of RIE-1 cells were preincubated with vehicle or varying concentrations of TAPI-1, and circular wounds were set in the monolayer. After 16 h, the closure of wounds was quantified by comparing the residual area of wound uncovered by cells to the original area that was denuded of cells (Fig. 4A). TAPI-1 dose-dependently decreased the closure of intestinal epithelial wounds, indicating that metalloprotease disintegrin-mediated ectodomain shedding is involved in intestinal epithelial restitution (Fig. 4B). Consistent with prior reports (20, 38), inhibition of EGFR autophosphorylation with AG-1478 also inhibited closure of wounds, confirming the importance of EGFR in this model.
To further establish selectivity of TAPI-1 for the inhibition of intestinal epithelial restitution mediated by ectodomain shedding of EGFR ligands, wound-conditioned medium or EGF was added to RIE-1 cell monolayers that were pretreated with TAPI-1 before circular wounds were set. Wound-conditioned medium or EGF, but not regular medium, significantly reversed the effects of TAPI-1 on wound closure, indicating that the inhibitory effect of TAPI-1 on intestinal epithelial restitution is due to blockade of ectodomain shedding of EGFR ligands, rather than another nonspecific effect (Fig. 4C). In monolayers not treated with TAPI-1, wound-conditioned medium or EGF both caused a slight, nonsignificant increase in restitution.
Results of our experiments have revealed that ectodomain shedding, a fundamental regulatory process that controls the activity of numerous proteins, is also an important contributor to the physiology of intestinal epithelial restitution. The metalloprotease disintegrin TACE is expressed in intestinal epithelial cells. When activated in wounded intestinal epithelial cell monolayers, metalloprotease disintegrin activity releases EGFR ligands from cell surface attachment, which leads to activation of the EGFR. Importantly, inhibition of metalloprotease disintegrin activity impairs intestinal epithelial restitution, a process that is important for the healing of ulcers in the intestinal epithelium. These observations have therefore identified a previously unrecognized physiological function for the metalloprotease disintegrin TACE, which has implications for the therapy of inflammatory bowel disease and other ulcerating diseases of the intestine.
The important role of the EGFR in promoting the cell migration, which forms the basis for epithelial restitution, has been well recognized (13, 20, 38). In wounded epithelia, the EGFR transmits extracellular signals to downstream mediators of cell migration, which includes phospholipase C (38) and NF-κB (20). However, the mechanism of activation of this receptor during epithelial restitution remained obscure. It was previously known that general metalloprotease inhibitors could decrease the release of EGFR ligands from stimulated cells, which suggested that those ligands are synthesized in precursor forms that are subject to proteolytic cleavage (3, 19). The individual metalloprotease that was required for release of EGFR ligands was identified as TACE, the same metalloprotease that is required for the shedding of TNF-α from the cell surface (37). TACE is a member of the disintegrin and metalloprotease family of proteins (40). Mice genetically engineered to lack TACE activity exhibit an abnormal phenotype that is very similar to that of TGF-α or EGFR knockout mice, and cells isolated from these mice have a dramatically reduced capacity to secrete TGF-α (37). TACE mutant mice are born at less-than-expected Mendelian ratios, and many die shortly after birth, with gross defects in eye and hair development. The lethality of these mice probably stems from dysmorphogenesis of epithelial organs, notably the lungs, placenta, and intestines. The importance of TACE for the normal developmental roles of TGF-α and EGFR is further highlighted by the exacerbation of the open-eye phenotype of Egfrwa−2 mice that harbor hypomorphic Egfr alleles (31) by the presence of mutant Tace alleles (43). Although spontaneous ulcers have not been reported in TACE mutant mice, it has not been possible to determine whether experimentally induced ulcers heal normally in these mice because of their early lethality and developmental defects. Interestingly, recent reports have indicated that additional proteases other than TACE may participate in the ectodomain shedding of TGF-α but not of other EGFR ligands (33). However, the severe phenotype observed in TACE-deficient mice strongly supports the importance of TACE for TGF-α availability, at least during development and early life.
TACE is widely expressed throughout the body (5, 35). In the intestinal tract, TACE expression has been shown in colonic crypt epithelial cells and in lamina propria mononuclear cells (10, 27). Immunohistochemical studies we performed on normal and ulcerated human colon have confirmed these prior findings and additionally identified the presence of TACE immunoreactivity in a variety of other intestinal cell types, including smooth muscle and vascular endothelium (unpublished observations). The diverse cell types that express TACE point toward the importance of ectodomain shedding in many physiological settings and potentially also in disease. EGFRs are also expressed by intestinal epithelial cells (20, 28, 32). Colocalization of TACE and EGFR expression on intestinal epithelial cells supports the likely physiological relevance of metalloprotease disintegrin-mediated ectodomain shedding of EGFR ligands in ulcer healing in vivo.
The shedding of transmembrane-anchored growth factors can be induced by a number of stimuli, including phorbol ester, hyperosmolarity, UV radiation (34), G protein-coupled receptors (39), prostaglandins (36), and, as we show here, scrape-wounding. Prior studies have characterized some of the intracellular signaling molecules used by these stimuli to achieve TACE activation and growth factor release. Inhibitors of PKC or extracellular signal-regulated kinase block phorbol ester-induced growth factor shedding (24, 34), whereas p38 MAPK and extracellular signal-regulated kinase appeared to play a role in growth factor shedding induced by ultraviolet radiation and hyperosmolarity (34). Prior reports indicated that scrape-wounding intestinal epithelial cell monolayers activated both extracellular signal-regulated kinase and c-Jun NH2-terminal kinase (21, 25), which suggested that these pathways might participate in growth factor shedding after this stimulus. However, we found that EGFR ligand release from wounded RIE-1 cell monolayers was insensitive to inhibitors of extracellular signal-regulated kinase, c-Jun NH2-terminal kinase, p38 MAPK, and PKC (not shown), suggesting the existence of a novel unidentified mechanism of TACE activation by this stimulus.
TACE controls the shedding of a large and diverse group of proteins, which, in addition to EGFR ligands, includes TNF-α, TNF receptors, selectins, β-amyloid protein, and others (reviewed in Ref. 40). In the case of TGF-α, shedding increases the activity of this growth factor for EGFR signaling. However, for other proteins, induced shedding by TACE might inactivate the protein, as is the case with TNF receptors, which would have the effect of decreasing the biological activity of TNF-α. If and how signals that activate TACE achieve substrate specificity is poorly understood. The initial recognition of TACE as a regulator of TNF-α secretion led to trials of TACE inhibitors for the treatment of inflammatory diseases in which TNF-α is believed to play a pathogenic role. In one study of experimental arthritis, TACE inhibition was associated with worse inflammation (45). In a study (12) of rodent colitis induced by rectal instillation of the hapten trinitrobenzene sulfonic acid, TACE inhibitors reduced overall inflammation score. Thus it remains unclear whether therapeutic targeting of metalloprotease disintegrin activity remains a viable strategy for the treatment of TNF-α-mediated inflammatory diseases. Certain cancers, such as colon cancer and breast cancer, show dependence on EGFR ligands for growth and viability. One recent study (8) showed that blockade of TACE activity inhibited tumorigenesis in a nude mouse model, suggesting that TACE might possess oncogenic functions.
Our findings have illustrated a previously unrecognized physiological function for TACE, promotion of intestinal epithelial restitution. This suggests that therapeutic blockade of TACE, as has been proposed for the treatment of inflammatory diseases and cancer, might have the untoward effect of impairing wound healing in the intestine and possibly other epithelial surfaces. Conversely, our results suggest that selective activation of TACE-mediated ectodomain shedding of EGFR ligands might prove a useful therapeutic strategy for the healing of intestinal ulceration. The therapeutic potential of targeting metalloprotease disintegrins in inflammatory diseases or cancer will be difficult to realize until greater selectivity can be achieved, because these enzymes can cleave such a variety of cell surface substrate proteins.
This work was supported by grants from the Crohn's and Colitis Foundation of America (First Award) and the National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-60792 to (L. J. Egan).
The authors acknowledge the expert assistance of James Tarara of the Mayo Clinic Optical Morphology Core Facility and Linda Murphy of the Mayo Clinic Cancer Center Core Research Facility.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society