A Three-dimensional Collagen Lattice Activates NF-κB in Human Fibroblasts: Role in Integrin α₂ Gene Expression and Tissue Remodeling

Human fibroblast cultures established by outgrowth from healthy human skin biopsies were provided by M. Simon (SUNY, Stony Brook, NY). The cells were maintained in DME (GIBCO-BRL, Gaithersburg, MD), supplemented with 10% FCS (Hyclone, Logan, UT), 100 U/ml penicillin, 100 U/ml streptomycin (GIBCO-BRL), and grown in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. Cells between population doubling levels 15 and 20 (the 6th and 10th passage) were used for the experiments. Bisindolylmaleimide GF 109203X (BIM) was purchased from CalBiochem-Novabiochem Corp. (La Jolla, CA). SN50 was obtained from BIOMOL (Plymouth Meetings, PA). The control peptide of SN50, SM, was synthesized by Research Genetics Inc. (Huntsville, AL) based on published sequences (Lin et al., 1995b). Polyclonal antibodies against NF-κB p65, p50, c-Rel, RelB, Iκ-Bα, and monoclonal antibody against Sp1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies against β-tubulin and human integrin α₂ subunit were purchased from Chemicon International, Inc. (Temecula, CA). [¹⁴C]Chloramphenical, [α-³²P]dCTP, and [γ-³²P]ATP were obtained from DuPont-NEN (Boston, MA).

Plasmids basicCAT and cytomegalovirus-CAT have been described previously (Zutter et al., 1994). Plasmids RSV–β-galactosidase, human α₂ cDNA probe, wild-type IκB-α, and constitutive IκB-α^{32, 36} were provided by L. Taichman (SUNY), Y. Takada (The Scripps Institute, La Jolla, CA) (Takada and Hemler, 1989), W. Greene (The Gladstone Institute of Virology and Immunology), respectively. Human MMP-1 cDNA and α₅ cDNA were purchased from American Type Culture Collection (Rockville, MD) and GIBCO-BRL, respectively. An oligonucleotide complementary to 28S ribosomal RNA was purchased from CLONTECH (Palo Alto, CA). NF-κB and Sp1 enhancer element consensus sequences 5′-AGT TGA GGG GAC TTT CCC AGG C-3′ and 5′-ATT CGA TCG GGG CGG GGC GAG C-3′, respectively, were purchased from Promega Corp. (Madison, WI). The four repeats of NF-κB consensus element (GGGACTTTCC) as well as four repeats of its mutated sequence (ATCACTTTCC; mutated bases are underlined) were synthesized and inserted, respectively, into the BamHI site of a promoter-CAT vector purchased from Promega Corp. The construction of a series of α₂CAT plasmids has been described previously (Zutter et al., 1994). For the opposite orientation of an upstream sequence of α₂ promoter (−549 through −351), the DNA fragment was digested with restriction enzymes (BglII and SmaI), filled in at 3′ end, and ligated back into the vector. Deletion mutant, pα₂549.dis, was made by restriction enzyme digestion of fragments (−351 through −122 by SmaI and SacII) from pα₂549CAT. A second deletion mutant, pα₂549.del, was made by removing a fragment (−549 though −351) from pα₂776CAT by BglII and SmaI followed by filling in and ligation. For SV-40 promoter–CAT constructs, the α₂ promoter fragment (−549 through −351) in both orientations was inserted into the BamHI site of the vector. The correctness of the recombinant DNA work was confirmed through restriction digest analysis. The PCR protocol for generating promoter fragments used in DNA–protein binding reaction was as follows: 20 ng of DNA template was amplified at 94°C for 2 min, followed by 30 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 30 s in a programmable thermal controller (PTC-100, Model 60; MJ research, Watertown, MA). The sequences of primers used are: pair 1 (5′- ⁻⁵⁹⁰GTATTGCTTAAATATCA⁻⁵⁷³-3′; 5′-⁻⁵¹⁴AGGTTAGAAACTAACTC⁻⁵³¹-3′); pair 2 (5′- ⁻⁵¹⁶CTGGTCATTCTGCGCTTA⁻⁴⁹⁸-3′; 5′- ⁻⁴¹³CACTAGCCCTAAACCACA⁻⁴³¹-3′); and pair 3 (5′- ⁻⁴¹⁵GTGCCCTCGGACCCCGCT⁻³⁹⁷-3′; 5′- ⁻³⁴²AGTCCCCGGGAGAACGTG⁻³⁶⁰-3′).

Collagen gels were prepared according to a procedure previously described (Xu and Clark, 1996). Pepsin-solubilized bovine dermal collagen dissolved in 0.012 M HCl was 99.9% pure containing 95–98% type I collagen and 2–5% type III collagen (Vitrogen 100; Celltrix Laboratories, Palo Alto, CA). Collagen for cultures was prepared by mixing 2.0 mg/ml of type I collagen, 100 U/ml penicillin, 100 U/ml streptomycin, and 1% FCS in DME, pH 7.0–7.4. Human dermal fibroblasts from subconfluent cultures starved in 1% FCS for 24 h were mixed with collagen solution for a final concentration of 5 × 10⁵ cells/ml. The collagen cell suspension (4 ml) was immediately placed onto 2% BSA-coated (ICN Biomedicals, Aurora, OH) between, 60-mm petri dishes (Falcon, Becton Dickinson Labware, Lincoln Park, NJ) and incubated at 37°C for 2 h or for the time periods described in figure legends before the addition of 5 ml of 1% FCS/DME to each dish. In most experiments described here, cells used as control are cultured in 1% FCS/DME on conventional tissue plastic plates.

After incubation at 37°C in 95% air, 5% CO₂, and 100% humidity for the indicated time, cultures were carefully washed twice in DME, and then processed for various analysis. In experiments where inhibitors were used, the levels of lactate dehydrogenase activity released were measured (LD Diagnostic kit; Sigma Chemical Co., St. Louis, MO) and were similar between cells cultured in the presence or absence of inhibitors. When gel contraction experiments were performed, the contraction process was observed and photographed at indicated time points. The surface areas of the gels were measured from prints. Data are presented as relative value and represent 10 individual experiments.

For monolayer collagen coating of plastic dishes, the collagen used for lattices was diluted to a final concentration of 50 μg/ml with PBS. This solution was added to plastic dishes at a final concentration of 6.4 μg/cm² and incubated overnight at 4°C. Coated dishes were blocked with 2% BSA for 2 h at room temperature and rinsed with PBS twice before use.

Total RNA was isolated from cell monolayers and collagen gel cultures using a modification of guanidinium thiocyanate method (Chromczynski and Sacchi, 1987). After centrifugation at 14,000 g to remove culture medium, collagen gels were dissolved in 4 M guanidinium isothiocyanate and repeatedly passed through a 20.5-gauge needle. For Northern blot hybridization, 3–5 μg of total RNA was treated with glyoxal/DMSO, separated by electrophoresis on an 1% agarose gel in 10 mM phosphate buffer, pH 7.0, and then transferred to Hybond⁺ nylon membranes (Amersham Corp., Arlington Heights, IL). Ethidium bromide (0.5 μg/ml) was included in the gel to monitor equal loading by the quantity of 18S and 28S ribosomal RNA present. cDNA probes were labeled with [α-³²P]dCTP by the random primer procedure (DuPont-NEN, Boston, MA). Oligonucleotide probes were end labeled with [γ-³²P]ATP in the presence of polynucleotide kinase (Boehringer Mannheim, Indianapolis, IN). The filters were hybridized to the labeled probes in QuickHyb solution (Stratagene, La Jolla, CA) for 3 h at 68°C and washed according to manufacturer's protocol. The signals were detected by autoradiography (X-Omat AR; Eastman Kodak, Rochester, NY) at −80°C for optimal exposure. All results shown are representative of at least two independent experiments.

4–7 μg of proteins from cytoplasmic and nuclear fractions prepared as previously described (Xu and Clark, 1997) were separated on SDS–polyacrylmide gel and transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). The membranes were incubated with a blocking solution containing 2% BSA, 2% horse serum, 50 mM Tris, pH 7.5, 150 mM NaCl, and 0.05% Tween 20 for 1 h at room temperature, and then incubated overnight at 4°C with various primary antibodies: p50, Iκ-Bα, β-tubulin, Sp1 and integrin α₂ subunit. For detection with enhanced chemiluminescence (Amersham Corp.), the blot was incubated with HRP-conjugated goat anti–rabbit antibody (1:1,000 dilution; Amersham Corp.) in 50 mM Tris, pH 7.5, 150 mM NaCl, and 0.05% Tween 20 for 1 h at room temperature. For detection with alkaline phosphatase, the membrane was sequentially incubated with biotinylated anti–rabbit goat IgG (H+L) (Vector Labs., Inc., Burlington, CA) at 6 μg/ml (1:250 dilution), and alkaline phosphatase at 1:600 dilution, and then visualized by NBT/ BCIP.

Gel mobility shift assay was performed as previously described (Xu and Clark, 1997). NF-κB and Sp1 recognition sequences 5′-AGT TGA GGG GAC TTT CCC AGG C-3′ and 5′-ATT CGA TCG GGG CGG GGC GAG C-3′, respectively, were purchased (Promega, Madison, WI). These oligonucleotides as well as PCR-generated DNA fragments were end labled by [γ-³²P]ATP. The nuclear extracts (3–5 μg) were incubated with 1 μg poly (dI/dC) (Boehringer Mannheim) and 2 μg BSA (GIBCO-BRL) in a binding buffer (10 mM Tris, pH 7.9, 5 mM MgCl₂, 50 mM KCl, 10% glycerol, and 1–5 × 10⁴ cpm end-labeled oligonucleotides) for 20 min at room temperature. The samples were separated on a 5% native polyacrylamide gel in 0.5× TBE buffer (Tris-borate-EDTA). For supershift assays, 2 μl antibodies were added to the reaction mixture containing end-labeled oligonucleotides, and then incubated for additional 30 min at room temperature. The samples were separated on a 3.5% native polyacrylamide gel.

DNA transfection was performed as previously described with a few modifications (Xu et al., 1996). Cells were passaged at 5–7 × 10⁵ per 10-cm plate. Transfection was performed 22–24 h after passage. Cotransfection was performed with either two plasmids, pα₂-CAT and pRSV-βgal control DNA (15 μg each plasmid), or three plasmids, wild-type or constitutive IκB-α^{32, 36}, pα₂–CAT, and pRSV–βgal control DNA (10 μg each), in 1 ml 0.25 M CaCl₂ were added dropwise to 1 ml 2× HEBS (50 mM Hepes, pH 7.05, 280 mM NaCl, 1.5 mM Na₂HPO₄) to form a precipitate. Cells in plates were rinsed twice with DME. Precipitate was added to the plate and incubated for 20 min at 37°C in 5% CO₂. DME containing 10% FBS and 50 mM chloroquine was added to 10 ml. After 4 h, medium was replaced with 0.5% FCS. After 40 h, the transfected cells were trypsinized, washed twice with PBS, and then subcultured into collagen gel or onto either tissue plastic or collagen-coated surface for 18–24 h in 0.5% FCS. Protein extracts were prepared after cells were released from 3D COL (Xu and Clark, 1997), and CAT enzyme activity was analyzed as described previously (Xu et al., 1996).

Previously we observed that 3D COL induces integrin α₂ mRNA expression by activating protein kinase C (PKC)-ζ (Xu and Clark, 1997). PKC-ζ has been shown to play an essential role in activating NF-κB by dominant-negative PKC-ζ blocking TNF-α–stimulated NF-κB activity and constitutively active PKC-ζ activating NF-κB in NIH 3T3 cells (Diaz-Meco et al., 1993, 1994; Dominguez et al., 1993; Folgueira et al., 1996). We also obtained evidence that the DNA binding activity of NF-κB, along with the activation of PKC-ζ and integrin α₂ mRNA expression, is induced by 3D COL (Xu and Clark, 1997). Here we examined whether 3D COL can signal the induction of transactivating activity of NF-κB in human dermal fibroblasts. A κB-CAT reporter plasmid was constructed by inserting four repeats of NF-κB consensus element into an SV-40 promoter CAT reporter plasmid vector. Fibroblasts were transfected with the plasmid followed by subculture in 3D COL. As shown in Fig. 1,A, 3D COL induced transactivating activity of NF-κB, consistent with the observation made in DNA binding activity (Xu and Clark, 1997). A control plasmid containing the mutant NF-κB recognition sequences did not demonstrate the inducibility by 3D COL (Fig. 1,A). The composition of the NF-κB DNA-binding complex was examined next by gel mobility supershift assay. The p50 and p65(RelA) of NF-κB family were found present in the binding complex (Fig. 1 B). The composition of the DNA-binding complex remained unchanged during full time-course (24 h) of the induction examined (data not shown).

Although the precise mechanisms underlying the NF-κB induction remain unclear, it is known that the dissociation of NF-κB from IκB, with which it is sequestered as an inactive precursor, and the translocation of NF-κB from cytoplasm to the nucleus are prerequisite for the subsequent nuclear activity of NF-κB. To assess whether nuclear activity of NF-κB observed in Fig. 1 is secondary to its translocation, subcellular fractions from cells cultured in 3D COL were prepared and immunoblotted for p50, one of two subunits that were found in the DNA-binding complex of NF-κB in response to 3D COL (Fig. 1,B). The nuclei isolated from cells incubated in 3D COL from 30 min to 24 h showed the presence of p50 in contrast to those from cells grown on tissue culture (Fig. 2,A). As a control, Sp1, the transcription factor that binds to its recognition site independent of 3D COL induction (Xu and Clark, 1997), showed the unchanged nuclear protein level during the time-course of incubation in 3D COL (Fig. 2,A). The kinetics of protein levels of NF-κB inhibitory protein, IκB, was also examined. Whereas Iκ-Bβ is not detectable in adult dermal fibroblasts (data not shown), Iκ-Bα was detected in the cytoplasmic fractions of quiescent cells (Fig. 2,B, lane 1). The accumulation of Iκ-Bα markedly decreased 1 h after cells were cultured in 3D COL and then resynthesized after 4 h (Fig. 2,B). This observation is in accord with the autoregulation of Iκ-Bα by signals that stimulate NF-κB activation via Iκ-Bα degradation (Sun et al., 1993). The detection of β-tubulin in the same sample preparations confirmed the specificity of Iκ-Bα decrease (Fig. 2 B). Thus, a correlation was observed between 3D COL stimulation and NF-κB activation, probably by regulating the cellular level of Iκ-Bα.

Next we examined the role of NF-κB in the induction of integrin α₂ mRNA expression by 3D COL. Since p50 responded to a 3D COL signal by undergoing nuclear translocation (Fig. 2,A) and forming NF-κB DNA-binding complexes (Xu and Clark, 1997), a cell-permeable synthetic peptide inhibitor, SN50—which carries the hydrophobic region of the signal peptide sequence from Kaposi's fibroblast growth factor (K-FGF) linked to the nuclear localization sequence (NLS) of p50 (Lin et al., 1995b)—was used to inhibit NF-κB activity in fibroblasts. SN50 has been reported to specifically inhibit the nuclear translocation of NF-κB in intact cells such as murine endothelial LE-II cells and human monocytic THP-1 cells stimulated by TNF-α and lysophosphatidic acid (LPA) (Lin et al., 1995b). The effectiveness of SN50 in inhibiting NF-κB activity stimulated by 3D COL in adult human dermal fibroblasts was examined. As shown in Fig. 3,A, NF-κB DNA binding activity was inhibited by the presence of SN50 in a concentration-dependent manner. In contrast, one control peptide, SM, which contains the same signal peptide sequence from K-FGF linked to a random amino acid sequence instead of p50 NLS (Lin et al., 1995b), did not show measurable inhibitory effects on the DNA binding of NF-κB stimulated by 3D COL (Fig. 3,A). To rule out the possibility that the inhibition of NF-κB DNA binding by SN50 is a nonspecific effect we examined the DNA binding ability of the SN50-treated nuclear extracts to Sp1 recognition sequence. The treatment of fibroblasts with SN50 did not prevent Sp1 from binding to its specific sequence (Fig. 3 A).

We next measured integrin α₂ mRNA level from cells treated with SN50. The mRNA levels of MMP-1 and integrin α₅ were measured in parallel. The 28S ribosomal RNA was measured as control for gel loading. SN50 inhibited the mRNA levels of integrin α₂ and MMP-1, but not integrin α₅ (Fig. 3 B). These results suggest that NF-κB plays an important role in the regulation of integrin α₂ and MMP-1 expression by 3D COL.

Since NF-κB is a transcription factor, the requirement of its activity in 3D COL stimulation of integrin α₂ mRNA expression indicates that 3D COL may control the transcription of integrin α₂ gene. To assess this possibility, we examined whether 3D COL regulates the promoter activity of integrin α₂ gene. The promoter region of the integrin α₂ gene was cloned and a series of deletion mutants of 5′ flanking sequence were inserted into a CAT reporter vector (Fig. 4,A) (Zutter et al., 1994, 1995b). Fibroblasts were transiently transfected with these CAT reporter constructs followed by incubation in 3D COL or on two-dimensional surfaces. As shown in Fig. 4,B, the reporter gene directed by sequences upstream of −92 bp of integrin α₂ promoter (pα₂122CAT, pα₂351CAT, pα₂549CAT, and pα₂776CAT) showed inhibited basal expression when compared to the activity of pα₂92CAT (Fig. 4,B), indicating the presence of potential silencer element(s). Upon stimuilation by 3D COL, the upstream sequences up to −351 bp of integrin α₂ promoter (pα₂92CAT, pα₂122CAT, and pα₂351CAT) did not show a positive response. In contrast, the plasmids containing either −549 or −776 bp of upstream sequences (pα₂549CAT and pα₂776CAT) demonstrated 3D COL inducibility. Therefore a region located between −92 and −122 bp probably has negative regulatory sequences for basal promoter activity, whereas the sequences located between −549 and −351 bp of integrin α₂ promoter appear to modulate positive response to 3D COL. For the convenience of this report, we designate this region α₂^549–351. To understand whether the positive response to 3D COL is elicited by collagen signals resident in collagen molecules of any form or a special collagen structure, we cultured fibroblasts transfected with pα₂549CAT on tissue plastic plates, on monolayer collagen-coated plates, and in 3D COL. Cells grown on collagen-coated plates failed to induce reporter gene activity when compared to cells grown on plastic surface (Fig. 4 C), supporting our previous observations of α₂ mRNA (Xu and Clark, 1997). Therefore signals from collagen in a particular 3D structure seemed responsible for the positive response of α₂ promoter modulated by α₂^549–351.

To further confirm that the positive response conferred by α₂^549–351 is an enhancer-like response, we examined α₂^549–351 based on the classic definition of an enhancer element: orientation- and distance-independent activity. A second set of CAT reporter plasmids were constructed that include inverted orientation of α₂^549–351 (pα₂549.revCAT), the altered distance of α₂^549–351 relative to the transcription initiation site by deletion of the sequence between −351 and −122 bp from pα₂549CAT (pα₂549. disCAT), and the deletion of α₂^549–351 from pα₂776CAT (pα₂549.delCAT) (Fig. 5,A). Fibroblasts transfected with this set of plasmids were compared to those with pα₂549 CAT, the plasmid containing natural orientation of α₂^549–351. 3D COL induced CAT level directed from integrin α₂ promoter region (pα₂549) was mediated by α₂^549–351 since the deletion of this region from pα₂776 (pα₂549.del) abrogated 3D COL inducibility of the reporter plasmid (Fig. 5,A), confirming the inability of pα₂351 to respond to 3D COL stimulation (Fig. 4,B). The 3D COL response mediated by α₂^549–351 was also independent of either its orientation or its distance to the transcription initiation site since both reporter plasmids (pα₂549.revCAT and pα₂549.disCAT) showed the similar 3D COL response compared to natural parental reporter plasmid pα₂549CAT (Fig. 5,A). Furthermore, the deletion of a region between −351 and −122 in pα₂549.disCAT did not restore the basal promoter activity (Fig. 5,A), indicating the importance of the region between −92 and −122 in its negative regulation as demonstrated by pα₂122 (Fig. 4 B).

To test whether α₂^549–351 alone is sufficient for 3D COL induction, we examined the function of this region in a different promoter context. The α₂^549–351 was inserted in either natural or inverted orientation into an SV-40 promoter-CAT vector lacking of enhancer elements (Fig. 5,B). 3D COL induced CAT activity directed from the SV-40 promoter in the presence of α₂^549–351 in either orientation, but failed to do so in its absence (Fig. 5 B). Taken together, these results indicate that 3D COL regulated integrin α₂ gene expression at transcription level and that a promoter region between −549 and −351 bp was necessary and sufficient for the transcriptional stimulation by 3D COL to occur.

The requirement of the NF-κB for 3D COL induction of α₂ promoter activity was investigated. Fibroblasts transfected with α₂^549–351-containing plasmid (pα₂549CAT) were treated with SN50 or its peptide control, SM. SN50 but not SM, inhibited CAT activity directed by −549 bp of integrin α₂ promoter stimulated by 3D COL (Fig. 6,A), consistent with the observation made at mRNA steady-state level (Fig. 3,B). A PKC inhibitor, BIM, also inhibited the promoter activity, supporting our previous finding that PKC is required for integrin α₂ mRNA expression induced by 3D COL (Xu and Clark, 1997). To further confirm the inhibitory effects of SN50 on α₂ promoter activity induced by 3D COL, we used a second approach by cotransfection of cells with α₂ reporter plasmids and wild-type IκB-α or a stable mutant IκB-α^{32, 36}. Since 3D COL led to IκB-α degradation (Fig. 2,B), the IκB-α^{32, 36} mutant, which is resistant to induced degradation (Traenckner et al., 1995), could serve as a potent and specific inhibitor for NF-κB activity (Wang et al., 1996). As shown in Fig. 6,B, whereas IκB-α^{32, 36} (IκBDN)-transfected cells did not alter their basal promoter activity as compared with wild-type IκB-α (mock) transfected cells, they demonstrated a drastic reduction in 3D COL–induced pα₂549CAT promoter activity. In contrast, the mutant IκB-α did not affect the promoter activity of either pα₂92CAT or pα₂351CAT when the transfected cells were cultured in 3D COL (Fig. 6 B). Therefore it appears that NF-κB transactivating activity was required for the α₂^549–351-mediated the 3D COL inducibility of integrin α₂ promoter.

Analysis of DNA sequence in this region revealed that a site located between −457 and −447 bp (GGGACGCACC) shares sequence homology but for one base (underlined) to the NF-κB consensus sequence GGGRNNYYCC (R indicates purine; N is any base; Y is pyrimidine) present in a set of inducible genes expressed by human monocytic and endothelial cells (Parry and Mackman, 1994). The oligonulceotides synthesized based on this sequence, however, mediated neither DNA–protein complex formation as judged by gel mobility shift assay nor transactivation when inserted into an SV-40 promoter-CAT reporter vector as judged by CAT analysis of transfected cells cultured in 3D COL (data not shown). The observation raised a possibility that α₂ promoter context is required for the detection of NF-κB binding activity to the region. Nuclear protein binding of the entire α₂^549–351 region was thus examined. Three DNA fragments were generated by PCR: F1, −590 and −514 bp; F2, −516 and −413 bp; and F3, −415 and −342 bp (Fig. 7,A). Gel mobility shift assay showed the formation of three specific F2– protein complexes, which were moderately increased by 3D COL (Fig. 7,B) whereas F1 and F3 did not demonstrate the specific nuclear protein binding (data not shown). Three approaches were taken to determine whether NF-κB was directly involved in the F2-protein complexes. First, the DNA binding activity of 3D COL–induced nuclear proteins to F2 region of α₂ promoter was reduced to approximately basal level by the incubation of cells with SN50, the p50 nuclear translocation inhibitor (Fig. 7 B), suggesting NF-κB activity is required for the binding complex formation. However, the competition with unlabeled NF-κB consensus sequence did not affect the DNA complex formation on F2 (data not shown), suggesting the absence of direct contact of NF-κB with this DNA fragment. Third, supershift assays with antibodies against various proteins of NF-κB family did not yield band shifts (data not shown), supporting the failure of NF-κB to bind this promoter region. Therefore, NF-κB appears to play a critical role in integrin α₂ promoter activation by 3D COL through α₂^549–351 without direct binding to the region.

It has been reported by various laboratories that integrin α₂β₁ mediates reorganization and contraction of collagen gels by human cells including fibroblasts (Schiro et al., 1991), cutaneous squamous carcinoma cells (Fujii et al., 1995), retinal pigment epithelial cells (Kupper and Ferguson, 1993), and transformed osteosarcoma cells (Riikonen et al., 1995). It was proposed that stimulants such as EGF (Fujii et al., 1995) and TGF-β (Riikonen et al., 1995b) induce 3D COL contraction by increasing integrin α₂β₁ expression. The transfection of integrin α₂ cDNA into a cell line RD cells that expresses β₁ chain but possess a very low level of α₂β₁ integrin restores the ability of RD cells to contract collagen gels (Schiro et al., 1991). We hypothesized that since NF-κB mediated α₂ gene expression, it may be in the regulatory pathway leading to α₂-mediated collagen gel contraction. The effects of NF-κB on 3D COL contraction by fibroblasts were examined. Treatment of cells with SN50, but not SM, significantly slowed the contraction process during 72 h examined (Fig. 8, A and B). The expression of integrin α₂ protein by fibroblasts cultured in 3D COL was similarly reduced by SN50 but not SM (Fig. 8,C). The amount of a nonspecific band of low molecular weight, serving as an internal control, was similar in all conditions (Fig. 8 C), confirming the specificity of the inhibtion. Therefore, it further indicates that NF-κB was involved in the integrin α₂ gene expression stimulated by 3D COL and tissue reorganization possibly by maintaining cellular level of integrin α₂.

We report here that NF-κB activity mediated the expression of integrin α₂ gene induced by 3D COL and the contraction of 3D COL populated by adult human dermal fibroblasts. The conclusion is supported by four lines of evidence described in this report. First, 3D COL induced nuclear translocation of p50, an NF-κB subunit, and the degradation and resynthesis of IκB-α (Fig. 2, A and B), an inhibitory protein of NF-κB and an NF-κB-responsive gene product (LeBail et al., 1993; Chiao et al., 1994). Second, two inhibitors of NF-κB activity, a nuclear translocation inhibitor, SN50, and a stable IκB-α mutant, IκB-α^{32, 36}, both reduced α₂ promoter activity that resides in the upstream region between −549 and −351 bp (Fig. 6, A and B). Third, SN50 weakened the protein complex formation in an α₂ promoter fragment from −516 to −413 bp (Fig. 7,B). Fourth, SN50 reduced both α₂ mRNA (Fig. 3,B) and protein levels (Fig. 8,C), and slowed down the collagen gel contraction process (Fig. 8 B). The observation that 3D COL signaled induction of NF-κB activity is in line of evidence from other laboratories that cell–ECM interactions are associated with activation of NF-κB (Qwarnstrom et al., 1994; Lin et al., 1995a; Lofquist et al., 1995) or liver transcription factors (Liu et al., 1991).

The modulation in Iκ-Bα kinetics presents a potentially critical link between cytosolic regulatory events and nuclear transcription in response to 3D COL induction. Like cytokine, phorbol ester, and lipopolysaccharide, which stimulate myeloid, epithelial, and fibroblast cells (Beg and Baldwin, 1993; Brown et al., 1993; Cordle et al., 1993; Henkel et al., 1993; Sun et al., 1993), 3D COL induces the degradation and subsequence resynthesis of Iκ-Bα (Fig. 2,B). Thus 3D COL may be listed as another extracellular signal that elicits an array of intracellular signal transmitters leading to posttranslational modification of Iκ-Bα with an eventual consequence of NF-κB activation. Iκ-Bα has been known as a target of many intracellular signals such as tyrosine phosphorylation events (Imbert et al., 1996), the Ras-Raf pathway (Finco and Baldwin, 1993; Li and Sedivy, 1993), PKC-ζ (Diaz-Meco et al., 1994), double-stranded RNA–dependent kinase (Yang et al., 1995), mitogen-activated protein kinase/extracellular response kinase (ERK)-1 (Lee et al., 1997), MAP3K-related kinase (Malinin et al., 1997), and IκB kinase (Zandi et al., 1997). PKC-ζ was reported to be associated with Iκ-Bα phosphorylation and subsequent NF-κB activation (Diaz-Meco et al., 1993; Dominguez et al., 1993; Diaz-Meco et al., 1994; Folgueira, 1996). Along with the evidence in this report that 3D COL modulated cellular level of Iκ-Bα (Fig. 2 B), is our previous finding that PKC-ζ activity is induced under the same condition (Xu and Clark, 1997). Therefore, an attractive hypothesis would be that signals from 3D COL are transmitted into the nucleus in part through a PKC-ζ/Iκ-Bα pathway.

Toward investigating the role of NF-κB in α₂ gene expression, we first questioned whether 3D COL induced α₂ transcription. Although several groups have observed 3D COL induction of α₂ mRNA and/or protein steady-state levels (Klein et al., 1991; Langholz et al., 1995; Xu and Clark, 1996), it was not known whether the regulation occurs at transcriptional level. In this report an upstream region between −549 and −351 bp was identified to confer the 3D COL indicibility of α₂ promoter (Figs. 4 and 5). We confirmed the presence of a general silencer between −92 and −122 bp that suppresses the basal integrin α₂ promoter activity as proposed previously (Zutter et al., 1995b). The stimulated expression by 3D COL was mediated by sequences upstream of this region, suggesting that 3D COL could regulate α₂ gene transcription in a DNA sequence–dependent manner, probably by releasing the negative impact of the silencer present between −92 and −122 bp on the promoter activity. A further study of the DNA binding pattern in the 3D COL–responsive region from −549 to −351 bp revealed that there were three DNA sequence–specific protein complexes (Fig. 7). Although the regulation of promoter activity by 3D ECM has not been studied as extensively as those by soluble factors, a similar study on β-casein gene promoter activity conducted by Schmidhauser et al. (1992) identified a 160-bp region in the promoter responsible for its activation by 3D matrigel. In another study, the promoter of TGF-β1 was found to be downregulated by 3D matrigel (Streuli et al., 1993). The involvement of NF-κB activity in α₂ promoter activity was investigated using two inhibitors, SN50, a peptide inhibitor for NF-κB nuclear translocation (Lin et al., 1995b), and IκB-α^{32, 36}, a stable IκB-α that serves to inhibit NF-κB. Interestingly, whereas NF-κB activity was consistently found to be required for full activation of integrin α₂ expression at levels of mRNA steady-state (Fig. 3,B), protein (Fig. 8,C), and transcription (Fig. 6), as well as for α₂ promoter–DNA protein complex formation (Fig. 7 B), there was no evidence for direct physical involvement of NF-κB in α₂ promoter; this observation is based on competition experiments with unlabeled NF-κB consensus sequence and supershift assay with antibodies against NF-κB proteins (unpublished laboratory data). The observation was further supported by our failure to find any functional activity of a near-perfect NF-κB consensus site located in the region between −549 and −351 bp of α₂ promoter as judged by two sets of experiments: gel mobility shift assay with the synthetic oligonucleotides based on the α₂ NF-κB–like sequence, and functional assay with the site inserted into a reported gene vector (unpublished laboratory data). Taken together, these observations suggest an indirect regulatory mechanism by which NF-κB directs the synthesis of another transcription factor that in turn activates the integrin α₂ promoter. Among many genes known to be mediated by NF-κB are these transcription factors, c-myc (Duyao et al., 1990; LaRosa et al., 1994), interferon regulatory factor 1 (Fujita et al., 1989; Harada et al., 1994), RelA (Ueberla et al., 1993), p50 (Ten et al., 1992), and α-1 acid glycoprotein/enhancer-binding protein (Lee et al., 1996). In fact, we found that 3D COL–induced integrin α₂ mRNA expression requires newly synthesized protein (Xu and Clark, 1997), an observation in accord with the synthesis of an intermediary transcription factor induced by NF-κB. Further studies will be required to address the identity of these three DNA–protein complexes, their regulation by 3D COL, and their functional roles in integrin α₂ gene expression.

The physiological role of NF-κB family proteins is under intensive study. NF-κB has been implicated in biological and pathological processes such as cell death (Grimm et al., 1996; Wu et al., 1996), angiogenesis (Shono et al., 1996), rheumatoid arthritis (Yang et al., 1995), and human cutaneous T lymphoma formation (Thakur et al., 1994). The appropriate regulation of NF-κB activity seems critical for proper biological function. For example, NF-κB activity is required for the H₂O₂-induced tube formation in human microvascular endothelial cells grown on 3D COL (Shono et al., 1996) and antitumor properties of antineoplastic agent taxol in macrophages (Hwang and Ding, 1995). On the other hand, the constitutive NF-κB activation in IκBα^–/– transgenic mice causes skin defects (Beg et al., 1995) and severe widespread dermatitis (Klement et al., 1996). The important role of NF-κB in mediating reorganization and contraction of 3D COL as reported here adds another potential biological function of NF-κB. The inhibition of NF-κB activity, which led to inhibited expression of integrin α₂ after 3D COL stimulation, correlated with the inability of cells to contract collagen gel (Fig. 8). NF-κB, therefore, may regulate 3D COL contraction through cell surface expression of α₂β₁ level. However, it must be noted that other possibilities cannot be ruled out. For example, we found that the presence of α₂β₁ alone, although neccessary, was not sufficient for the reorganization and contraction of collagen matrices (unpublished laboratory data), suggesting that the availability of α₂β₁ receptor and the α₂β₁-elicited second mesenger pathway probably present two separate control levels for 3D COL contraction. Furthermore, the inhibition of NF-κB by SN50 also decreased the 3D COL– induced expression of MMP-1 (Fig. 3 B), a protein involved in tissue remodeling. Additionally, NF-κB is strongly implicated in the transcriptional regulation of several growth factors and cytokine genes including interferon-β, IL-1β, IL-2, IL-6, TNF-α and TGF-β1 (for review see Baldwin, 1996), which might offer another interpretation on the mechanism whereby NF-κB modulates tissue remodeling. The essential role for NF-κB in reorganization and contraction of COLs as reported here may present a potentially important nuclear regulatory site for tissue remodeling.

We thank Dr. M. Simon for adult human dermal fibroblasts, Dr. Y. Takada for human α₂ cDNA probe, and Dr. W. Greene for IκB-α plasmids.

BIM

bisindolylmaleimide GF 109203X

COL

collagen lattice

ECM

extracellular matrix

IκB

inhibitor for NF-κB

IL

interleukin

MMP-1

type I matrix metalloproteinase

NF-κB

nuclear factor κB

PKC

protein kinase C

3D

three-dimensional

TGF-β

transforming growth factor-β

TNF

tumor necrosis factor