| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-smooth muscle actin expression and collagen gel contraction: implications in tissue remodeling
* CRRET Laboratory, CNRS UMR 7149, University of Paris XII, Créteil, France;
INSERM U697 and
INSERM U716, Laboratoire de Pharmacologie, Hôpital Saint-Louis, Paris, France;
Department of Ophthalmology, University of California at Irvine, Irvine, California, USA; and
|| Department of Ophthalmology at Fondation Ophtalmologique A. de Rothschild and Bichat Hospital, AP-HP, Paris, France
1Correspondence: CRRET Laboratory, CNRS UMR 7149, University Paris XII, 61 Av du Général de Gaulle, 94010 Créteil Cedex, France. E-mail: menashi{at}univ-paris12.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-smooth muscle actin (SMA) is considered a marker of this differentiation process, we analyzed the effect of EMMPRIN on its expression in corneal and skin fibroblasts by Western blots, immunocytochemistry, and a functional assay of collagen lattice contraction. Increasing EMMPRIN expression by cDNA transfection or by treatment with exogenously added recombinant EMMPRIN resulted in an up-regulation of SMA expression. EMMPRIN also increased the contractile properties of the treated fibroblasts as demonstrated by the immunohistochemical appearance of stress fibers and by the accelerated contraction of fibroblast-embedded collagen lattices. Blocking EMMPRIN expression by small interfering RNA inhibited SMA and collagen gel contraction induced not only by EMMPRIN but also by transforming growth factor-β, a major mediator of myofibroblast differentiation that also regulated EMMPRIN expression. These findings, combined with the fact that EMMPRIN and SMA colocalized to the same cells in the stroma of pathological corneas, expand on the mechanism by which EMMPRIN remodels extracellular matrix during wound healing and cancer.—Huet, E., Vallée, B., Szul, D., Verrecchia, F., Mourah, S., Jester, J. V., Hoang-Xuan, T., Menashi, S., Gabison, E. E. Extracellular matrix metalloproteinase inducer/CD147 promotes myofibroblast differentiation by inducing -smooth muscle actin expression and collagen gel contraction: implications in tissue remodeling.
Key Words: epithelial-stromal interactions • wound healing • MMPs • TGFβ
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-smooth muscle actin (SMA), which characterizes the myofibroblasts both in vitro and in vivo. The presence of SMA stress fibers is generally related to the contractile activities of the myofibroblasts, because a direct correlation was demonstrated between the level of SMA and myofibroblast contraction (1)
. The cytokine transforming growth factor-β (TGFβ) is considered to be the major actor in the differentiation of fibroblasts into myofibroblasts because it is capable of up-regulating SMA in fibroblasts (2
3
4)
.
In addition to stimulating SMA expression, TGFβ is known to be a potent stimulator of connective tissue formation as it controls both the expression of components of the ECM network, such as fibrillar collagens and fibronectin, and the expression of protease inhibitors, including plasminogen activator inhibitor-1 or tissue inhibitors of metalloproteinases (TIMPs) (5)
. By contrast, the turnover of the laid down matrix is regulated by the activity of matrix metalloproteinases (MMPs), a family of enzymes capable of degrading almost all ECM proteins. Therefore, the balance between matrix formation and deposition by cytokines like TGFβ and their degradation by MMPs seemingly dictates the outcome of the healing process, whether normal, hypertrophic/fibrotic or thinning/ulceration. In addition to its effect on the stimulation of ECM synthesis and proteases inhibitors, TGFβ can itself regulate the expression of MMPs, as it was shown to down-regulate the interstitial collagenase (MMP-1) and stromelysin (MMP-3) but to increase MMP-2 and MMP-9 production in several cell types (6
7
8
9)
. Certain MMPs such as MMP-2, MMP-9, and MT1-MMP can in turn activate latent TGFβ (10
, 11)
, thus demonstrating the complex interrelations that may exist in regulating matrix synthesis and degradation. One would expect therefore that the myofibroblast, considered the key cell for the connective tissue remodeling that takes place during wound healing, would be equipped with an elaborate regulatory system necessary for a fine control of matrix turnover.
Stroma cells with differentiated myofibroblast features also are found in different primary and metastatic epithelial tumors and are thought to play a central role in tumor-associated tissue remodeling (12)
. These fibroblastic cells adjacent to neoplastic cell nests, which were shown to express important amounts of SMA, are normally referred to as peritumoral fibroblasts or reactive stroma. Their presence is considered to be a response of the host cells to inductive stimuli exerted by tumor cells, which in turn are able to participate actively in tumor progression by secretion of proteolytic enzymes, thus promoting tumor invasion and metastasis. Indeed, it is now generally admitted that an important proportion of such enzymes is produced by stroma myofibroblasts as a host response to tumor (13
, 14)
. Therefore, tumor cells appear to signal neighboring fibroblasts on the one hand to differentiate to myofibroblasts and, on the other, to secrete more MMPs.
The factor that induces the peritumoral fibroblasts to increase MMP production has been identified as extracellular matrix metalloproteinase inducer (EMMPRIN)/CD147, a membrane-spanning molecule highly expressed in tumor cells. Its most characterized role is the induction of MMP production in neighboring cells (15
, 16)
and its overexpression in vivo results in enhanced tumor growth (17)
. EMMPRIN also was shown to modulate MMP expression during normal tissue remodeling and differentiation as well as in nontumoral pathological conditions such as skin or corneal ulcerations (18
, 19)
, rheumatoid arthritis (20)
, and lung and liver fibrosis (21
, 22)
. As all these situations also are associated with the presence of myofibroblasts, the possibility that EMMPRIN may be involved in signaling fibroblasts not only to induce MMPs production but also to change phenotype to that of myofibroblasts has been addressed in this study using a corneal wound-healing model.
The cornea, with its simple cellular organization providing a particularly useful model for studying wound-healing events, is a highly organized transparent tissue that needs to remain transparent to refract light properly. Penetrating incisions or ablation injury to the corneal stroma stimulates a typical fibrotic repair response involving hypercellularity, expression of SMA, and deposition of disorganized ECM. This fibrotic repair response can be harmful to the cornea because it may cause opacity, and the contraction associated with remodeling alters corneal curvature (23)
. Chronic wound healing may be associated also with excessive proteolysis and result in corneal thinning or ulceration with the risk of perforation (24)
, but the reason for the different outcomes, fibrosis or thinning, following chronic wound healing is yet to be determined.
We recently identified EMMPRIN in healthy and diseased corneas and demonstrated its ability to induce MMP expression following epithelial-stromal interaction (19)
. In this study, we addressed the question of whether this MMP induction by EMMPRIN represents only part of a broader process of fibroblast activation that involves the differentiation of fibroblast to myofibroblast. We therefore examined the effect of EMMPRIN on SMA in corneal and skin fibroblasts as well as on the contractile potential of these cells. The relationship between EMMPRIN and TGFβ in this differentiation process also was explored.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Telomerase-immortalized corneal fibroblasts (HTK cell line) (25)
were cultured in Dulbecco modified Eagle medium (DMEM) (Invitrogen, Cergy Pontoise, France) supplemented with 10% fetal calf serum (FCS; PAA Laboratories, Les Mureaux, France).
Primary fibroblasts were obtained from skin biopsies of patients undergoing breast plastic surgery, following informed consent. Fibroblasts were propagated in DMEM supplemented with 10% FCS. Fibroblasts were used at passages 4–6.
The antibodies: anti-EMMPRIN mAb used for immunoblots and immunohistochemistry (anti-CD147, clone HIM6, Becton Dickinson France, Le Pont de Claix, France), anti-EMMPRIN pAb used for immunohistochemistry (anti-EMMPRIN, clone K-20, Santa Cruz Biotechnology, Tebu-bio, Le Perray en Yvelines, France), anti- SMA (clone 1A4, Sigma-Aldrich, St. Quentin Fallavier, France), pan-TGFβ blocking antibody (clone 1D11, R&D Systems, Lille, France), anti-MMP-1 mAb (clone 41–1E5, Calbiochem, VWR International, Strasbourg, France), anti-GAPDH (clone 6C5, Applied Biosystems, Courtabouef, France), horseradish-peroxidase [HRP]-conjugated anti-mouse and anti-rabbit antibodies (Jackson ImmunoResearch, Immunotech, Marseille, France), and anti-mouse IgG Alexa 488 and anti-goat IgG Alexa 594 (Molecular Probes Inc., Eugene, OR, USA; Invitrogen). TIMP-1 pAb was kindly donated by Dr. R. Fridman (Wayne State University, Detroit, MI, USA).
Other materials: TGFβ1, epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), and TNF- were from R&D Systems. (TGFβ1 was used throughout this study and was referred to as TGFβ.) Marimastat was obtained from British Biotechnology (Oxford, UK). BM chemiluminescence Western blotting substrate (POD) and DAPI were from Roche Diagnostics (Meylan, France). Protease Inhibitor Cocktail Set V EDTA-free was from Calbiochem. Recombinant EMMPRIN (recEMMPRIN) used in the study was an affinity-purified Flag-tag protein consisting of the extracellular domain of human EMMPRIN expressed in mammalian cells (19)
. EMMPRIN cDNA construct, human EMMPRIN full-length cDNA in pcDNA3 (pcDNA3 hEMMPRIN) (17)
, was kindly provided by Dr. Stanley Zucker (State University of New York, Stony Brook, NY, USA).
cDNA transfection
HTK corneal fibroblasts at 40–50% confluence were transiently transfected for 24 h with the pcDNA3 hEMMPRIN or with pcDNA3 alone using FuGENE-6 (Roche Diagnostics) in DMEM/10% serum according to the manufacturer’s protocol. Western blotting analysis confirmed increased EMMPRIN expression (mean 2.5x), which persisted over 3–4 days.
Small interfering RNA (siRNA) transfection
Two different EMMPRIN siRNA oligos, siEMMPRIN1 and siEMMPRIN2 (Ambion, siRNA IDs: 10372 and 215973, respectively, Applied Biosystems) or scramble siRNA oligos (BLOCK-iT fluorescent oligo, Invitrogen), siScramble, (33 nmol/L) were transfected into the cells using oligofectamine transfection reagent in Opti-MEM I (Invitrogen) in the absence of FCS and antibiotics. After 6 h incubation, an equal volume of DMEM/20% FCS was added to the transfection mixture, which was then cultured for a further 18 h. The cells then were washed and cultured in the presence of DMEM/10% FCS prior to analysis of EMMPRIN inhibition by Western blot and further experiments. Density measurements of Western blot bands have shown a mean of 75% inhibition in EMMPRIN levels (4 different experiments), without significantly affecting TIMP-1 levels.
Three-dimensional collagen gel culture
Type I collagen was extracted from rat tail tendons as described previously (26)
. Fibroblast-seeded collagen lattices were prepared by mixing on ice 5 ml collagen (2 mg/ml) in 0.1% acetic acid with 1 ml of 10x DMEM and 1 ml 0.1 N NaOH to neutralize the pH before adding 3 ml of cell suspension, harvested by trypsinization from subconfluent cultures, to give a final cell density of 1.5 x 105/ml of lattice mixture. Collagen gels were cast in 60 mm non-cell culture-treated dishes (4 ml/gel). After polymerization (30 min at 37°C), gels were detached and covered with 6 ml of DMEM with 1% FCS, with or without added TGFβ (10 ng/ml) or EMMPRIN (2 µg/ml) as indicated. At each time point, the lattices were photographed and rates of gel contraction were calculated by determining the remaining surface area by computer-based analysis and expressed as either percentage of initial area or percentage of control area. Data are presented graphically or in tabular form as mean ± SD. Lattice area was statistically compared with Student’s t test. Significant differences were determined at P < 0.001, P < 0.01, and P < 0.05.
For biochemical analysis on the embedded cells, gels were washed by 30 min incubation with several changes of serum-free medium, centrifuged at 10,000 g for 2 min, and the pellets were homogenized using micropotters in radio-immuno-precipitation assay (RIPA) buffer (50 mM Tris pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.1% SDS, 0.5% Nonidet P-40, 0.5% sodium deoxycholate), supplemented with 1/100 Protease Inhibitor Cocktail Set V. After protein determination using the Pierce BCA Protein Assay kit (Perbio Science, Brebières, France), extracts were analyzed by Western blots.
For collagen lattice populated with cDNA- or siRNA-transfected cells, 24 h after transfection, cells were recovered, counted, and embedded into collagen lattice at 1.5 x 105 cells/ml as described above.
Western blot analysis
Cells were lysed in RIPA buffer, incubated on ice for 5 min, and then scraped with a rubber policeman. Some The lysates were clarified by centrifugation at 13,000 g for 5 min at 4°C, and 20 µg samples were analyzed by Laemmli SDS-PAGE (Bio-Rad, Marne la Coquette, France) on 10% gels and then transferred to Immobilon-P PVDF membrane (Millipore, Guyancourt, France). Membranes were immunoblotted with either anti-EMMPRIN mAb or anti-SMA mAb overnight at 4°C, or with GAPDH mAb for 1 h at room temperature, followed by 1 h incubation with HRP-conjugated anti-mouse antibody and visualized with BM chemiluminescence substrate. Protein loading was verified by ponceau red staining and by comparing with the intensity of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) bands.
Northern blotting
Total RNA was obtained using RNeasy (Qiagen, Courtaboeuf, France) and analyzed by Northern hybridization (20 µg/lane) with 32P-labeled cDNA probes for EMMPRIN and GAPDH as described previously (5)
. The EMMPRIN probe was excised from pcDNA3 vector with EcoRI. Quantifications were performed with a Storm 840 PhosphorImager (Amersham Biosciences, GE Healthcare, Orsay, France).
Confocal immunohistochemistry and cytochemistry
Cryostat sections (8 µm) were prepared from frozen cornea and were immunostained as described previously (27)
. Briefly, the sections were fixed in chilled acetone for 10 min, then rehydrated in PBS and incubated in blocking solution (3% BSA in PBS) for 30 min. Sections were incubated for 1 h with either anti-EMMPRIN mAb, anti-EMMPRIN pAb, or anti-SMA mAb, then for 30 min with conjugated affinity-purified donkey anti-mouse IgG and/or anti-goat IgG (Alexa 488 and 594, respectively).
For immunocytochemistry, cells were seeded on type I collagen (0.1 mg/ml)-coated glass slides, cultured for 2–3 days in the presence of 10% FCS. The cells were then treated in one of the following two ways: 1) the cells first were serum starved for 48 h, after which they were treated with 2 µg/ml recEMMPRIN for a further 48 h; 2) the cultured cells first were transfected with EMMPRIN siRNA for 6 h, after which serum was added for a 18 h incubation as described above. The cells then were serum starved for 48 h and treated with 10 ng/ml TGFβ for a further 48 h in serum-free medium.
Slides then were washed and fixed in 4% paraformaldehyde/DMEM (v/v) and processed as above using either anti-EMMPRIN mAb, anti-EMMPRIN pAb, or anti- SMA mAb, followed by conjugated affinity-purified donkey anti-mouse and/or anti-goat IgG (Alexa 488 and 594, respectively). The slides were mounted and examined with a laser scanning confocal microscope (Leica Lasertechnik, Heidelberg, Germany). Negative controls were prepared using the same procedure, but PBS was substituted for the primary antibody.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
confirms the low but distinct staining of EMMPRIN in the stroma of normal A) human cornea, and B) human skin sections. However, EMMPRIN expression was increased greatly in the stroma of injured human cornea with a developed ulceration, as is shown by the much stronger and broader staining of EMMPRIN in the affected area compared to a distant nonulcerated region of the same cornea (Fig. 1C
). This increase was more apparent at the subepithelial stroma where interactions between epithelial and stromal cells are thought to take place. Examination of healed corneal lesions shows that this up-regulation of EMMPRIN in the ulcerated stroma is maintained even after complete epithelial closure (Fig. 1D
). Hence, EMMPRIN induction localizes with areas of stromal remodeling in the healing cornea.
|
TGFβ regulates EMMPRIN expression in fibroblasts
To investigate the possibility that cytokines regulate EMMPRIN expression, we examined the effect of TGFβ, TNF-, bFGF, and EGF on EMMPRIN expression in corneal fibroblast (HTK) cell culture. Figure 2
shows that the addition of TGFβ had the greatest effect, increasing EMMPRIN expression at both the protein and RNA levels. As illustrated (Fig. 2B
), the increase of EMMPRIN RNA expression is observed as early as 2 h after TGFβ addition and reached a maximum at 6 h. Conversely, reducing endogenously produced TGFβ by incubating the cells with TGFβ-blocking antibodies inhibited EMMPRIN levels (Fig. 2C
), showing that EMMPRIN is constitutively under TGFβ regulation. This inhibition by TGFβ-blocking antibodies was observed in both sparse and confluent cultures. Note that the EMMPRIN level was much higher in sparse cultures, which are known to express more TGFβ receptor (28)
compared to confluent cells, further supporting the notion that TGFβ signaling regulates EMMPRIN production. The regulation of EMMPRIN by TGFβ also has been confirmed with primary skin fibroblasts (not shown).
|
Differentiation of fibroblast to myofibroblast in vitro is associated with an increase in EMMPRIN levels
Because TGFβ induces the differentiation of fibroblasts to a myofibroblast phenotype, we questioned the possibility that the up-regulation of EMMPRIN by TGFβ may be related to the differentiation process. The expression of EMMPRIN therefore was measured in different experimental systems known to initiate myofibroblast differentiation, which were characterized by increased SMA expression. Treatment of fibroblasts with TGFβ induced the expression of both EMMPRIN and SMA (Fig. 3
A) in a time-dependent manner. The coexpression of EMMPRIN and SMA also was observed in the presence of serum (Fig. 3B
) as well as when the cells were embedded in tridimensional collagen lattice (Fig. 3C
), a culture system known also to induce myofibroblasts differentiation (29
, 30)
. Time-course analysis of gel contraction reveals that both EMMPRIN and SMA levels are highest at the point of maximal contraction. Hence, EMMPRIN levels are increased during myofibroblast differentiation induced by either cytokines or by mechanical stress. These results have been reproduced with primary skin fibroblasts (not shown).
|
EMMPRIN induces SMA production and accelerates collagen gel contraction
To determine if EMMPRIN is an active player in the differentiation process, we investigated the direct effect of EMMPRIN on SMA production and collagen gel contraction.
Increased EMMPRIN level obtained by the transient transfection of the corneal fibroblasts with EMMPRIN cDNA caused an increase in SMA expression and was associated with accelerated gel contraction, as shown in Fig. 4
A. Similarly, the treatment of corneal fibroblasts with exogenously added recEMMPRIN increased SMA protein expression and collagen gel contraction (Fig. 4B
). It should be noted that EMMPRIN expression also was increased by this treatment, confirming the previously observed reported autocrine regulation of EMMPRIN by itself (19
, 31)
. The results in Fig. 4B
with the recEMMPRIN also have been obtained using primary skin fibroblasts (not shown).
|
Confocal immunohistochemistry experiments also were performed in order to look for phenotypic changes in the fibroblasts as well as the expression of SMA following recEMMPRIN treatment. Untreated fibroblasts displayed elongated spindle-shaped morphology with only weak staining of SMA and rarely seen stress fibers. In the recEMMPRIN-treated fibroblasts, a significant subpopulation appeared larger with a well-spread appearance and expressing SMA stress fibers, indicating their transition into myofibroblasts (Fig. 4C
). The recEMMPRIN also increased EMMPRIN staining, which may result from the recEMMPRIN adhering to the cell surface or to its increased synthesis, as shown also in the biochemical analysis (Fig. 4B
).
To determine whether the SMA induction was related to the MMP-inducing activity of EMMPRIN, we examined the effect of a broad-spectrum MMP inhibitor, marimastat, on the expression of SMA in the presence of EMMPRIN. RecEMMPRIN, at the concentration used (2 µg/ml), greatly increased the levels of secreted MMP-1 (Fig. 4D
). However, the presence of 5µM marimastat did not inhibit SMA induction by EMMPRIN (Fig. 4D
), indicating an MMP-independent mechanism.
We then used siRNA silencing to evaluate the effect of EMMPRIN inhibition on both SMA expression and gel contraction. EMMPRIN siRNA transfection of corneal fibroblasts markedly down-regulated EMMPRIN expression (75%), and this was associated with a clear decrease in the protein levels of SMA (55%) as compared with their levels in the scrambled control siRNA-transfected cells (Fig. 5
A). Similar results were obtained using two different EMMPRIN siRNA. EMMPRIN siRNA also inhibited collagen gel contraction, which already was noticeable 1 day after gel formation, and the inhibition persisted for up to 7 days (Fig. 5B
). The EMMPRIN inhibition by siRNA previously was found to last for at least 5 days after transfection (not shown).
|
EMMPRIN inhibition decreases TGFβ-mediated SMA expression and myofibroblast contraction
The possibility that some of the TGFβ effects may be mediated by EMMPRIN has been examined by silencing EMMPRIN expression. Corneal fibroblasts were transfected with EMMPRIN siRNA and then subjected to TGFβ treatment. EMMPRIN reduction in siRNA-transfected cells was associated with an inhibition of SMA expression compared to control scrambled siRNA-transfected cells. In addition, SMA response to TGFβ also was attenuated by the EMMPRIN siRNA treatment (Fig. 6
A), with an inhibition ranging from 50–80% in the different experiments. This inhibition was transcribed by a parallel decrease in collagen gel contraction (Fig. 6B
). However, siRNA treatment was without effect on TIMP-1, not normally regulated by EMMPRIN (32)
, or on its known stimulation by TGFβ (33)
. Immunoconfocal analysis of corneal fibroblasts demonstrates increased EMMPRIN staining following TGFβ treatment, which was associated with the appearance of SMA stress fibers. It is interesting to note that SMA induction often was localized in the subpopulation of EMMPRIN overexpressing cells (Fig. 6C
, arrow). Treatment with EMMPRIN siRNA reduced EMMPRIN staining and inhibited SMA stress fibers’ appearance in TGFβ-treated cells (Fig. 6C
), consistent with the biochemical analysis. These biochemical, functional, and morphological data confirm a role of EMMPRIN in mediating, at least in part, TGFβ effects on myofibroblast differentiation.
|
The induction of EMMPRIN in the fibroblasts of remodeling stroma in vivo is associated with SMA expression
In this study we present evidence that EMMPRIN induces SMA expression, whether EMMPRIN is exogenously added, endogenously increased by transfection, or induced by TGFβ treatment of fibroblasts. We sought to examine whether the EMMPRIN overexpressing remodeling stroma in vivo (Fig. 1)
was associated with a myofibroblast phenotype. Double-labeled confocal microscopy analysis of EMMPRIN (red) and SMA (green) was performed on human cornea with chronic ulceration. The staining in Fig. 7
shows that whereas most of the EMMPRIN-positive fibroblasts also expressed SMA, a more detailed analysis reveals the presence of a subpopulation of fibroblasts localized in the subepithelial region adjacent to the ulceration that express particularly high levels of both EMMPRIN and SMA. The presence of these heterogeneous populations in terms of the relative expression of EMMPRIN and SMA may suggest the existence of different induction pathways for SMA in vivo.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, implicating EMMPRIN in the MMP-mediated stromal thinning and ulceration. EMMPRIN itself is greatly increased in the stroma of such pathological wounds, in particular at the level of the subepithelium, implicating direct interactions between the epithelium and the stroma in its up-regulation. These epithelial-mesenchymal interactions play a critical role in development, cancer progression, and wound healing and were shown to modify the phenotype of both epithelial cells and fibroblasts (34
, 35)
. Known key mediators of these interactions include members of the TGFβ family, released primarily from the epithelium to target stromal cells after disruption of the basement membrane (36)
. In this study, we demonstrate that TGFβ up-regulates EMMPRIN expression in fibroblasts and therefore is likely to contribute to the increased EMMPRIN seen in the ulcerated stroma. The inhibitory effect of TGFβ-blocking antibody suggests that EMMPRIN is in fact constitutively under the regulation of TGFβ, but its levels can be further increased in response to raised extracellular levels of this cytokine. This result concords with our findings that EMMPRIN is up-regulated in low-density cultures, where TGFβ and its receptors are known to be greatly increased (28)
. However, factors other than TGFβ also may up-regulate EMMPRIN in this system. EGF, which was shown to regulate EMMPRIN in epithelial tumor cells (37)
, had no such effect in fibroblasts, but the intervention of other cytokines cannot be excluded. EMMPRIN itself was shown to be able to induce its own expression in neighboring cells (Fig. 4
; 19
, 31
), and because it is particularly enriched on the surface of epithelial cells (32
, 38)
, it has the potential, following injury involving basement membrane rupture, to induce increased expression in stromal cells. Because EMMPRIN can be shed from the cell surface (39)
, it also can act as a diffusible factor, reaching other cells in the vicinity. Therefore, likely both epithelial-derived EMMPRIN and TGFβ contribute to the increased EMMPRIN expression observed in the injured stroma.
In wound healing, TGFβ is known to exert some its actions through the differentiation of fibroblasts into myofibroblasts, SMA-containing contractile cells that are essential for ECM remodeling during normal and pathological wound healing. The fact that TGFβ also can induce fibroblasts to up-regulate their EMMPRIN expression prompted us to examine the effect of the increased EMMPRIN on myofibroblast differentiation and phenotype. Our results show that EMMPRIN actively participates in the process of myofibroblast differentiation by inducing SMA expression and collagen-matrix contraction. These data expand on the known functions of EMMPRIN and suggest that, in addition to regulating the degradative potential of the cell, it also can influence the contractile phenotype by triggering reorganization of the actin cytoskeleton to form contractile stress fibers in an MMP-independent manner. This conclusion is further supported by the observed inhibitory effect of EMMPRIN siRNA on TGFβ-induced SMA expression and collagen gel contraction, showing that EMMPRIN also plays a role in the TGFβ-mediated matrix contraction by fibroblasts. This result suggests also that EMMPRIN can act as a downstream mediator of TGFβ in the process of fibroblast-to-myofibroblast differentiation. Indeed, microarray studies conducted on skin fibroblasts (40)
identified EMMPRIN as part of a gene cluster activated rapidly by TGFβ, suggesting that it may be a direct target of the TGFβ-signaling pathway through Smads. This conclusion also is supported by our results showing EMMPRIN mRNA induction as early as 2 h after TGFβ addition (Fig. 2B
). It is important to note that our results obtained with corneal fibroblasts also were reproduced using primary skin fibroblasts, showing that the observed regulation of EMMPRIN by TGFβ and by the induction of SMA is general to the fibroblastic cell type and is not restricted to corneal cells.
The regulation of EMMPRIN by TGFβ is more difficult to reconcile in view of the known opposing effects on certain MMPs. While EMMPRIN up-regulates MMP-1 in fibroblasts (32
, 41)
, this proteinase is known to be inhibited by TGFβ (6)
. However, the fact that collagen gel contraction in vitro is accompanied by an important increase in MMP-1 (42
, 43)
suggests a more complex protease regulation in myofibroblasts that necessitates a comprehensive examination of the signaling pathways involved. Because TGFβ is not the only factor regulating and controlling EMMPRIN levels, the overall resulting effect on the proteolytic potential of the cell hypothetically would depend on the relative levels of TGFβ and EMMPRIN.
Although EMMPRIN is known mainly for its MMP-inducing ability, it is becoming increasingly clear that it has other additional functions. EMMPRIN was found, for example, to increase the synthesis of hyluronan (44)
and vascular endothelial growth factor (45)
in tumor cells and thus was implicated in other aspects of tumor progression, such as tumor survival and angiogenesis. A role for EMMPRIN in differentiation also was evoked based on its localization in the basal layers, but not the superficial layers, of the epithelia of dermis (38
, 46)
and cornea (19)
, suggesting that EMMPRIN is expressed in actively differentiating cells of normal epithelia. In this study, we further demonstrate a role for EMMPRIN in the differentiation of myofibroblasts through the up-regulation of their contractile properties. A role for EMMPRIN in controlling cell shape already has been suggested, as the transfection of the drosophila homologue of EMMPRIN was found to affect the intracellular architecture of different cell types in which it is expressed (47)
. Our results show for the first time an effect of EMMPRIN on cytoskeleton composition leading to a modification of cell shape and tensile properties.
EMMPRIN is a transmembrane-spanning protein that can act as a receptor. It was identified as the signaling receptor for extracellular cyclophilins in situations related to chemotaxis and adhesion of immune cells (48)
. However, EMMPRIN's receptor on fibroblast cells responsible for the stimulation of MMP production remains elusive (41
, 49)
. It has been suggested that EMMPRIN may serve as its own counter-receptor in cancer cells via homophilic interactions, implying that it can act also as a ligand (50)
. In this study, we compared the effects of exogenously added soluble EMMPRIN lacking the transmembrane domain with those of increased endogenous expression of the entire molecule by cDNA transfection. In both cases, there was an increase in SMA expression. However, it is difficult to speculate on the nature of EMMPRIN, whether receptor or ligand, involved in these two situations because, when soluble EMMPRIN is added as a ligand, it induces endogenous EMMPRIN expression in a positive autocrine feedback loop (19
, 31
; Fig. 4B
). Thus, while the added recombinant molecule may act as a ligand, the overexpressed endogenous EMMPRIN on the same cell potentially can act as a receptor.
SMA generally has been used as a marker of fibrosis because SMA-containing myofibroblasts have been observed in most fibrotic tissues, and TGFβ, the cytokine principally associated with fibrosis, increases both the expression of SMA and that of ECM components. At the same time, accumulated evidence indicates that the main role of the smooth-muscle-like features of myofibroblasts is in the force generation which results in wound contraction (30)
. In view of our results showing that EMMPRIN, an MMP inducer, can itself induce SMA, it is tempting to speculate on the existence of several types of myofibroblasts with different and even opposing effects on the ECM: those with ECM synthetic phenotype present in most fibrotic tissues and driven more by TGFβ, and those found in ulcerated tissues having an increased degradative potential due to a greater EMMPRIN effect. Indeed, SMA-expressing myofibroblasts, expressing high levels of matrix-degrading proteinases such as MMP-2 and MT1-MMP, also were localized in the stroma surrounding many cancer tissues, with the potential to promote tumor invasion through tissue degradation (13
, 14)
. Thus, the cellular pathways leading to the expression of SMA may be different from those of tissue remodeling, and the acquisition of myofibroblasts phenotype, characterized by the expression of SMA, may be more representative of the contractile role of these cells. The effect of myofibroblasts on matrix remodeling may then depend, among other factors, on the levels of TGFβ and EMMPRIN, tilting the balance toward synthesis or degradation, respectively.
The interplay and relative role of TGFβ and EMMPRIN in myofibroblast differentiation, SMA expression, and matrix remodeling in normal and pathological wound healing is under investigation in our laboratory. Understanding the underlying mechanisms controlling the myofibroblast phenotype has important implications for the control of pathological wound-healing responses such as excessive matrix contraction, scarring, or ulceration.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 30, 2007. Accepted for publication September 27, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-smooth muscle actin expression upregulates fibroblast contractile activity. Mol. Biol. Cell 12,2730-2741-smooth muscle actin expression in granulation tissue myofibroblasts and in quiescent and growing cultured fibroblasts. J. Cell Biol. 122,103-111-smooth muscle actin expression and myofibroblast transformation in cultured corneal keratocytes. Cornea 15,505-516[Medline] and TGF-β. J. Cell Sci. 118,2143-21531 β1 and 2 β1 integrins. J. Cell Biol. 131,1903-1915This article has been cited by other articles:
|
|
 |
|
  A. Margulis, K. H. Nocka, N. L. Wood, S. F. Wolf, S. J. Goldman, and M. T. Kasaian MMP dependence of fibroblast contraction and collagen production induced by human mast cell activation in a three-dimensional collagen lattice Am J Physiol Lung Cell Mol Physiol, February 1, 2009; 296(2): L236 - L247. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= distance from ulceration.
70%) of HTK cells were serum starved for 18 h prior to cytokine treatment in serum-free medium for 24 h, then 20 µg of cell extracts were analyzed with either anti-EMMPRIN or anti-GAPDH monoclonal antibodies. B) Northern blot analysis of EMMPRIN regulation by TGFβ. Subconfluent fibroblast cultures were treated with TGFβ (10 ng/ml) for 2, 4, 6, and 24 h in medium containing 1% serum. EMMPRIN mRNA levels were detected and compared to that of GAPDH. C) Effect of TGFβ inhibition on EMMPRIN expression. Sparse and confluent HTK cultures were obtained by seeding 4 x 104 cells/cm2 (confluent) and 5 x 103 cells/cm2 (sparse). The cells were allowed to attach in complete DMEM for 24 h and then incubated in absence (Ctrl) or in the presence of 0.5 µg/ml pan-TGFβ-blocking antibody (anti-TGFβ) for 24 h before EMMPRIN expression analysis by Western blot. All blots are representative of at least 3 different experiments.
