Increased Sp1-dependent transactivation of the LAMgamma 1 promoter in hepatic stellate cells co-cultured with HepG2 cells overexpressing cytochrome P450 2E1.

Laminin is a basement-membrane protein that increases in liver fibrosis. To study the role of oxidative stress on laminin expression, hepatic stellate cells (HSC) were co-cultured with HepG2 cells that do or do not express (E47 or C34 cells, respectively) CYP2E1, a potent generator of oxygen radicals. Co-incubation of HSC with E47 cells increased laminin beta1 and gamma1 proteins compared with co-incubation with C34 cells; this increase was prevented by antioxidants and CYP2E1 inhibitors. Similar results were observed in co-culture with primary hepatocytes from saline- or pyrazole-treated (with high levels of CYP2E1) rats. Laminin alpha1 chain was not detectable in the HSC in any of the systems; however, laminin alpha2 chain increased in HSC co-cultured with E47 cells. Synthesis but not turnover of laminin beta1 and gamma1 proteins was increased in HSC in the E47 co-culture. Laminin beta1 and gamma1 mRNAs were up-regulated in HSC in the E47 co-culture because of transcriptional activation of both genes. Transfection experiments in HSC with reporter constructs driven by the laminin gamma1 promoter showed maximal responsiveness with the -230/+106 and the -1400/+106 constructs in the E47 system. Gel-shift assays demonstrated an increase in Sp1 binding to the laminin gamma1 promoter in HSC when co-incubated with E47 cells, which was blocked by an anti-Sp1 antibody. Co-transfection of a Sp1 expression vector further increased the responsiveness of the -330LAMgamma1-CAT reporter vector in HSC in the HSC/E47 system. These results show that diffusable CYP2E1-derived oxidative-stress mediators induce synthesis of laminins by a transcriptional mechanism in HSC. Such interactions between hepatocytes and HSC may be important during liver fibrosis.

The laminins are a family of adhesive glycoproteins composed of high molecular weight disulfide-bonded heterotrimers composed of ␣␤␥ chains. To date, five ␣, three ␤, and two ␥ chains have been described, forming at least 12 trimeric laminin isoforms (1). Although laminins are distinguished by both subunit composition and tissue distribution, the overall domain structure is well conserved (1). Laminin 1 (EHS laminin) is an 820-kDa heterotrimer consisting of one each of ␣1 (400-kDa), ␤1 (220-kDa), and ␥1 (200-kDa) chains, and laminin 2 (Merosin) is an 800-kDa heterotrimer formed of ␣2 (380-kDa), ␤1 (220-kDa), and ␥1 (200-kDa) chains. Both covalent and noncovalent interactions contribute to the basic structure of most laminins, a cruciform with three short arms and one long arm (2). The individual laminin subunits share a common structural motif composed of repeated epidermal growth factor-like domains interrupted by globular domains.
A number of biological activities have been attributed to laminin, including cell attachment, differentiation, and migration, along with interactions with other matrix components (3). Laminin has also been implicated in the process of tumor invasion and metastasis (4). Different isoforms of the laminin molecule may vary with respect to tissue distribution and developmental expression. The ␤1 and the ␥1 chains are expressed in most tissues that produce basement membranes, but their ratios vary considerably (5).
In the liver, increased deposition of laminins, the main noncollagenous glycoproteins in all basement membranes, was demonstrated in rats with chemically induced carcinoma (6) and within the lobule of livers from patients with malignancies (7). In contrast, in normal adult livers, laminins are located mainly in the portal tracts and are only sparsely deposited in the space of Disse (8). In human liver, the laminin isoforms have not been clearly identified so far, and only information from rodents is available. Rat hepatic stellate cells (HSC), 1 the main source of extracellular matrix components in both normal and fibrotic livers, express ␤1, ␥1, and an ␣ chain of 380 kDa that probably corresponds to unprocessed form of the ␣2 chain (9). Normal adult hepatocytes do not express laminin 1 in vivo, but synthesize both ␤1 and ␥1 chain mRNAs after a few hours in culture (6,7). Several hepatoma cell lines of either human or rat origin express both ␤1 and ␥1 chains at high levels and a 380-kDa polypeptide that is genetically different from the ␣1 chain but stained by polyclonal anti-laminin antibodies (10). Although most studies have focused on laminins polymerized into a basement membrane, much laminin is found freely soluble and diffusible. Furthermore, although the polymerization of laminin in vitro occurs by a self-assembly mechanism, basement membranes are formed at discrete sites close to the cell membrane (11).
The 5Ј-flanking region of the mouse laminin ␥1 (LAM␥1) gene has been cloned and characterized (12). The LAM␥1 gene appears to contain two transcription start sites (Ϫ169 and Ϫ234) and it does not contain a TATA or CAAT box, however, it has several interesting features, including the presence of ten GC boxes, which act as putative binding sites for the redox-sensitive transcription factor Sp1 (two in tandem and the rest monomeric), and a stretch of nine nearly identical repeats of 11 nucleotides between Ϫ200 and Ϫ450 with the sequence 5Ј-CCC(G/T)CCC(A/T)CCT-3Ј. The consensus sequence for cAMP responsiveness is also present in the LAM␥1 promoter and is similarly found in the promoters of other extracellularmatrix proteins (12). These motifs act as transcription activators in several extracellular-matrix genes. The integrity of the CTC-rich region is required to promote LAM␥1 activity. Consequently, it could be hypothesized that there is interplay between CTC and GC elements, GC boxes being predominant in the activation of truncated CTC-less LAM␥1 promoter. The region from Ϫ830 to Ϫ224 appears to contain a negative regulatory element, which decreases the promoter activity of the LAM␥1 gene. Deletion of Ϫ94 to Ϫ61 nucleotides reduced the promoter activity by severalfold in HepG2 cells, and deletion of 20 bp from Ϫ41 to Ϫ21 completely abolished the promoter activity in HepG2 cells (13). On the other hand, the LAM␥1 promoter had a relatively high level of activity in NIH-3T3 cells, which synthesize little laminin. This result suggests that the 830-bp promoter segment may lack a negative regulatory element, which inhibits transcription in certain cells (12).
Ethanol and several other low molecular weight agents induce cytochrome P450 2E1 (CYP2E1), which is of special interest because it metabolizes and activates hepatic toxins as well as carcinogens and fatty acids (14). CYP2E1 is a loosely coupled enzyme that generates high amounts of reactive oxygen species (ROS) such as superoxide radical and H 2 O 2 (15,16). Oxidative stress plays an important role in mechanisms by which ethanol damages the liver (17). Our laboratory has carried out studies to evaluate whether hepatocytes with high levels of CYP2E1 can interact with nonparenchymal liver cells such as HSC via release of diffusable mediators (18,19). In the present study we evaluated whether CYP2E1-derived oxidative stress could modulate the expression of basement-membrane proteins such as laminin, which are elevated during liver fibrosis, and the role of oxidative stress in this inductive up-regulation of laminin.

MATERIALS AND METHODS
Cell Culture-The model used in most of the experiments described below is based on the co-culture of HepG2 cell lines that express (E47 cells) or do not express (C34 cells) the human CYP2E1 (20, 21) with primary HSC or with an immortalized rat stellate cell line (HSC-T6). Primary HSC were isolated from male Sprague-Dawley rats (600 Ϯ 25 g) (Charles River Breeding Laboratories, MA) by in situ liver perfusion with bacterial collagenase and pronase, followed by density-gradient centrifugation with Nycodenz according to published protocols (22). Results of Figs. 2-5 were with the primary HSC whereas other results were carried out with the HSC-T6 cells. Cell viability (95%) was assessed by the trypan blue exclusion method. Purity of the HSC fraction (97%) was determined as described previously (23). To extend results with the HepG2 cells to intact primary hepatocytes, selected experiments ( Fig. 4) were carried out using hepatocytes isolated from saline control rats or from pyrazole-treated rats whose CYP2E1 levels are elevated about 3-to 4-fold (24).
Details on the co-culture model are shown in Fig. 1. Cells were co-incubated using cell-culture inserts of 3-m pore size to separate both cell populations; this allows study of the effect of diffusable CYP2E1-derived mediators from the hepatocyte on HSC functions, which resembles the physiological situation in the liver. The HSC were plated on the bottom and the HepG2 cells or the hepatocytes on the filter to create a gradient of the released mediators. The ratio of HepG2 or hepatocytes to HSC was 5:1, similar to that of parenchymal:nonparenchymal cells in liver. After overnight incubation of HSC alone in minimum essential medium (MEM) supplemented with 10% fetal bovine serum and essential amino acids, the HSC medium was discarded, the cells were washed 5 times in MEM, and the cell-culture inserts containing the HepG2 cells or the primary hepatocytes were transferred along with their overnight culture medium. At this time all additions were made (t ϭ 0 h).
General Methodology-Plasmid DNA preparation as well as transfection procedures, Northern and Western blot analysis, nuclear runon, the activities of antioxidant enzymes and glutathione (GSH) levels were carried out as previously described (23,25,26). Mouse laminin ␣1, ␤1, and ␥1 cDNA probes used for Northern blot and nuclear run-on assays were kindly provided by Dr. Yoshihiko Yamada (National Institutes of Health). The glyceraldehyde-3-phosphate dehydrogenase cDNA clone was purchased from the American Type Culture Collection. The Sp1 expression vector was a gift from Dr. Robert Tjian (University of California, Berkeley). Western blots were carried out routinely under denaturing conditions on cell lysates or with incubation media using anti-laminin 1 antibody (1/5000), which recognizes all ␣1, ␤1, and ␥1 subunits (Sigma) or anti-laminin ␣2 chain antibody (1/2000), which recognizes only the N-terminal moiety of ϳ300 kDa and not the Cterminal moiety of ϳ75-80 kDa of the ␣2 chain (Santa Cruz Biotechnologies). The anti-Sp1 (1/2000), anti-tubulin (1/5000), and anti-fibrinogen (1/5000) antibodies were from Santa Cruz Biotechnologies. Goat anti-rabbit IgG conjugated to horseradish peroxidase was used as secondary antibody (1:5000; Chemicon). In most of the Western blots, the laminin ␤1 and ␥1 chains run very close and were difficult to separate during the electrophoresis (11), therefore bands have been labeled as laminin ␤1 and ␥1, except for the Western blot in Fig. 2A in which both bands were clearly separated because of a longer run. Fig. 3, C and D show Western blots run in nondenaturing conditions.
Laminin Synthesis and Turnover-To assay the synthesis of laminin ␤1 and ␥1 subunits, the HSC-T6 cells or HSC-T6 co-cultured with the C34 or E47 cells were plated separately in 10% fetal bovine serum-MEM; after 12 h, the medium from the HSC was removed and the inserts containing the C34 or E47 cells together with their culture FIG. 1. Scheme of the co-culture model. HSC were seeded on the bottom plate at a density of 2.5 ϫ 10 5 cells in 3 ml of culture medium. CYP2E1 activity was monitored by the p-nitrophenol oxidation method, and 1.25 ϫ 10 6 C34 or E47 cells (in some cases primary hepatocytes) were plated on the insert in 3 ml of culture medium. After overnight incubation (typically 16 h), the medium from the HSC was discarded, the cells washed 5 times in MEM, and the inserts were transferred together with the incubation medium from the C34, E47, or primary hepatocytes onto the HSC. New medium was added to HSC plated with empty inserts; these were considered as non-co-cultured controls. Various additions were made at this time (t ϭ 0 h), and samples of HSC were collected at selected time points. medium were transferred onto the plates containing the HSC; fresh medium was also added to the HSC that were plated with an empty insert as a reference group (not to be co-cultured with HepG2 cells). 22 h later, the media from the complete co-culture systems were replaced with methionine-cysteine-free MEM plus 10% dialyzed fetal bovine serum, the cells were incubated for 2 h, after which they were pulselabeled with 150 Ci of EasyTag TM EXPRE 35 S 35 S Protein Labeling Mix (PerkinElmer Life Sciences) for 0, 2, 4, 8, and 12 h to study the synthesis of laminin ␤1 and ␥1. A set of samples was pulsed for 2 h in the presence of 40 M cycloheximide, an inhibitor of protein synthesis, as a control for the analysis of laminin ␤1 and ␥1 synthesis. The cells were washed in 1ϫ phosphate-buffered saline and lysed at the indicated time points with 150 l of 10 mM Tris-HCl buffer, pH 7.4, 0.5% Triton X-100, 1 mM EDTA, 150 mM NaCl, 0.5% sodium deoxycholate, 1% SDS, and 1 mM phenylmethylsulfonyl fluoride.
To assay the turnover of laminin ␤1 and ␥1, defined as the loss of [ 35 S]methionine-labeled intracellular laminin ␤1 and ␥1 when HSC-T6 cells were chased with cold methionine, plus secretion of [ 35 S]methionine-labeled laminin ␤1 and ␥1 into the medium, the co-cultures were treated as above but pulse-labeled with the EXPRE 35 S 35 S mix for 24 h. The cells were then washed three times and chased with complete MEM supplemented with 300 g/ml of cold methionine. Cells were washed in 1ϫ phosphate-buffered saline and lysed at 0, 1, 2, 4, 8, and 12 h in the same lysis buffer. In all cases, laminin ␤1 and ␥1 was immunoprecipitated with anti-laminin 1 IgG-protein G-agarose as follows: 40 g of protein of cell lysates and 1 mg of protein from the culture medium were first incubated with 10 l of preimmune rabbit serum for 15 min, followed by the addition of 50 l of a 50% (v/v) suspension of protein G-agarose. After centrifugation for 2 min at 13,000 rpm, the supernatant was incubated with anti-laminin 1 IgG by rocking overnight at 4°C followed by addition of 50 l of a suspension of protein G-agarose. Samples were centrifuged for 1 min at 13,000 rpm, the pellets were washed three times with lysis buffer, once with lysis buffer plus 2% SDS, and three times with 0.1 M Tris-HCl buffer, pH 6.8. Laminins ␤1 and ␥1 were eluted by boiling for 5 min in Laemmli's buffer, samples were centrifuged for 2 min at 13,000 rpm to remove the protein Gagarose, resolved on a 5% SDS-PAGE, and dried. The intensity of the radioactive signal was quantified using PhosphorImager (Molecular Dynamics) and the ImageQuant software.
To evaluate total protein synthesis by HSC, cells were treated as above and incubated with the EasyTag TM EXPRE 35 S 35 S labeling mix for 0, 2, 4, 8, and 12 h, followed by addition of 30% trichloroacetic acid to stop the reaction. The cells were washed in 1x phosphate-buffered saline, and the trichloroacetic acid-precipitable counts were determined in a scintillation counter after resuspension of the pellets in scintillation liquid. To evaluate loss of [ 35 S]methionine labeled HSC total protein, the co-cultures were pulse-labeled for 24 h as above, followed by chasing for 0, 1, 2, 4, 8, and 12 h. The HSC were treated with 30% trichloroacetic acid and washed, and the trichloroacetic acid-precipitated counts were determined as described above.
Quantitative comparison of the intensity of the signal scanned in the PhosphorImager was performed using ImageQuant software.
Construction of the Laminin ␥1-CAT Expression Vectors-These constructs were kindly provided by Dr. Yoshihiko Yamada (National Institutes of Health) and were generated as described in Ref. 12: a recombinant plasmid, pKH 130, containing the promoter region of the laminin ␥1 chain gene (Ϫ833 to ϩ106) fused to the structural part of the gene encoding CAT as a readout for transcriptional activity was constructed as follows: a 6-kb HindIII fragment containing the first exon and a single NcoI site located at ϩ300 bp was subcloned into PUC19. This plasmid, pKH102, was linearized with NcoI and then partially digested with the exonuclease Bal31. HindIII linkers were attached, and the plasmid was self-ligated. One of the plasmids, pKH121, containing 106 bp of the 5Ј-untranslated region was selected. The 1-kb BamHI-HindIII fragment of pKH121 whose BamHI site was converted to NdeI by filling the end with Klenow fragment and attaching NdeI linker was inserted into the NdeI-HindIII site of pSVOCAT (27). The Ϫ1400LAM␥1-CAT construct contains a 1.4-kb laminin ␥1 promoter segment (ϩ106 to Ϫ1300) cloned into the HindIII site of pSVOCAT. To generate the Ϫ2500LAM␥1-CAT construct, the SmaI segment from the intron 1 was cloned into the NdeI site of pKH130 (12).
For electrophoretic mobility-shift assays, synthetic oligonucleotides for Sp1, AP1, and NFB (Promega, Madison, WI) were end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase (Promega, Madison, WI). Binding reactions were carried out in a total volume of 10 l with 5 g or 0.5 g (for AP1) of the nuclear protein extract from HSC, 1 l of 10ϫ binding buffer containing 1 g/l of poly(dI-dC), and 20,000 cpm of labeled oligonucleotides at room temperature for 30 min. For competition studies, 1000-fold cold oligonucleotides were added along with labeled oligonucleotides. Antibodies raised against Sp1, AP1 (c-Jun/c-Fos), and NFB (p50/p65) (Santa Cruz Biotechnologies) were used for supershifting oligonucleotides-nuclear protein complexes. Polyacrylamide gel electrophoresis (6%) was performed at 150 volts for 2 h in 0.5ϫ TBE (45 mM Tris borate/1 mM EDTA, pH 8).
Southwestern Blot Analysis-Southwestern (DNA-protein) blotting was performed by the method of Singh et al. (29). Briefly, 25 g of nuclear protein extract from HSC were electrophoresed on a SDS exponential 5-20% gradient polyacrylamide gel. After electrophoresis, the gel was soaked in 25 mM Tris, 190 mM glycine pH 8.3, and 20% methanol for 1 h. Nuclear proteins were electrotransferred onto 0.2-m nitrocellulose membranes using the same buffer containing 0.1% SDS. After overnight transfer at 4°C, the membranes were blocked in Blotto containing 10% nonfat dry milk in TNE buffer (50 mM Tris pH 7.5, 40 mM NaCl, 1 mM EDTA). DNA binding was carried out for 3 h with TNE buffer containing 5 g/ml poly(dI-dC) and 2 ϫ 10 5 cpm/ml [␣-32 P]dCTP multiprime-labeled 80-bp fragment (Ϫ230 to Ϫ150) derived from the Ϫ2500LAM␥1-CAT construct by PCR amplification. Membranes were washed three times for 5 min each with TNE at room temperature and exposed in the PhosphorImager screen.
Nomenclature-The designation HSC refers to data found in the HSC incubated with an empty insert; HSC/C34 refers to results obtained in the HSC after co-culture with the C34 cells, whereas HSC/E47 refers to results obtained in the HSC after co-culture with the E47 cells.
Statistics-Results are expressed as mean Ϯ S.E. Most of the experiments were repeated at least three times except for the nuclear run-on, which is from one experiment only. Statistical evaluation was carried out using Student's t test.

RESULTS
Laminin ␤1, ␥1, and ␣2 Protein Levels Increase in Primary HSC Co-cultured with E47 Cells-Primary HSC were co-cultured with either the C34 or E47 cells. HSC lysates and aliquots of the incubation medium were collected at 1, 2, 3, 4, and 5 d and analyzed by Western blot for the expression of different laminin chains. A time-dependent increase in laminin ␤1 (220 kDa) and ␥1 (200 kDa) production was observed in both systems but was higher in the E47 co-culture when compared with the C34 co-culture ( Fig. 2A). At 3 d of co-culture there was a 2to 4-fold increase in the HSC content of laminin ␤1 and ␥1 chains, as well as a 3-fold increase in secretion of laminin ␤1 and ␥1 to the medium of the E47 co-culture as assessed by Western blot (Fig. 2, B and C). ␤-Tubulin levels, as a control for loading, were identical in both co-cultures. This difference in the amount of laminin ␤1 and ␥1 subunits in the culture medium in the E47 cell co-culture did not come from laminin ␤1 and ␥1 produced by the E47 cells because the medium from C34 and E47 cells cultured alone did not show any significant differences in laminin ␤1 and ␥1 content (Fig. 2D). High molecular weight bands in the 400-kDa region where laminin ␣1 would appear could not be detected in any of the blots in Fig. 2, When the same samples were blotted and incubated with anti-laminin ␣2 chain antibody, laminin ␣2 chain expression was observed, and higher expression of laminin ␣2 was de-tected intracellularly at 3 d in HSC co-cultured with E47 cells as compared with the HSC/C34 co-culture ( Fig. 3A) as well as in the culture medium (Fig. 3B). When the same cell extracts were electrophoresed under nondenaturing and nonreducing conditions, a dimer/s of about 400 kDa was observed after 3 d of culture, whereas a heterotrimer of about 800 kDa was observed after 5 d but not 3 d (Fig. 3, C and D) of culture. There was increased formation of the dimer/s at 3 d and the heterotrimer at 5 d in the HSC/E47 co-culture compared with the HSC/C34 co-culture, in agreement with the increase in the individual ␤1, ␥1, and ␣2 laminin chains. Of note is the observation that the 800-kDa heterotrimer band was not detectable until day 5 of co-incubation with the HepG2 cells, which suggests that as-sembly of the ␣2 subunit (whose expression is also increased under higher oxidative stress conditions) with the ␤1 and ␥1 subunits appears to be delayed with respect to the increase in levels of expression of the individual ␣2, ␤1, and ␥1 laminin chains or the 400-kDa dimer/s. The laminin heterotrimer is believed to play the major role in laminin's action as a basement-membrane protein.  2. Laminin ␤1 and ␥1 protein levels increase in primary HSC co-cultured with E47 cells. Primary HSC were co-cultured with either C34 or E47 cells for 1, 2, 3, 4, and 5 d and cell lysates collected and analyzed by Western blot for laminin ␤1 and ␥1 expression (A). A representative Western blot of HSC lysates (B) and culture medium (C) collected at 3 d is shown. Incubation medium from C34 and E47 cells cultured alone did not show any differences in laminin ␤1 and ␥1 secreted (D). Arbitrary units under the blots refer to the intensity of the laminin ␤1 and ␥1/␤-tubulin ratio or the laminin ␤1 and ␥1/fibrinogen ratio. blot. Fig. 4A shows that both catalase and vitamin E prevented the increase in laminin ␤1 and ␥1 subunits in HSC co-incubated with E47 cells; these antioxidants also decreased laminin ␤1 and ␥1 in the HSC/C34 co-culture. Catalase and vitamin E had no effect on the ␤-tubulin loading control levels. Thus, laminin ␤1 and ␥1 levels appear to be sensitive to induction by ROS. To validate that the increase in laminin ␤1 and ␥1 levels found with the HSC/E47 co-culture was indeed because of CYP2E1, the co-cultures were incubated in the presence of CYP2E1 inhibitors such as 5 mM diallylsulfide, 2 mM 4-methyl pyrazole, 0.1 mM sodium diethyldithiocarbamate, and 10 M phenylisothiocyanate. The increase in laminin ␤1 and ␥1 production by the E47 co-culture was significantly lowered by the CYP2E1 inhibitors (Fig. 4B). These inhibitors produced only minor changes in the C34 system (which lacks CYP2E1 expression), validating their specificity for CYP2E1. Moreover, transfection with antisense CYP2E1 cDNA blocked the increase of the E47 co-culture on laminin ␤1 and ␥1 levels in HSC without any effect in the HSC/C34 co-culture, whereas transfection of C34 or E47 cells with sense CYP2E1 further increased laminin ␤1 and ␥1 production in HSC about 2-fold (Fig. 4B). Transfection with empty plasmid (pCI-neo) had no effect. To confirm that these CYP2E1 inhibitors and transfection experiments modulated CYP2E1 levels and activity, CYP2E1 expression was determined by Western blot analysis and its activity measured by the catalytic oxidation of p-nitrophenol to p-nitrocatechol. The CYP2E1 inhibitors lowered CYP2E1 activity by 70 -80%, transfection with antisense CYP2E1 decreased CYP2E1 activity by 80%, whereas transfection with sense CYP2E1 elevated CYP2E1 activity 3-fold (data not shown). These results indicate that the increase in laminin ␤1 and ␥1 content in the HSC/E47 co-culture is mediated via CYP2E1 acting through a ROS-dependent mechanism.
Hepatocytes from Pyrazole-treated Rats Increase Laminin ␤1 and ␥1 Protein in HSC-To validate the results obtained with the HepG2 cell lines with primary hepatocytes, freshly isolated primary HSC were co-incubated with primary hepatocytes from either control or pyrazole-treated rats. Pyrazole induces CYP2E1 protein expression in hepatocytes about 3-to 4-fold over the levels present in saline control hepatocytes and stabilizes the protein against degradation (24). Laminin ␤1 and ␥1 protein levels increased 3-fold and more than 6-fold in HSC cultured with saline-and pyrazole-hepatocytes, respectively, when compared with HSC cultured alone (Fig. 5, basal conditions, lanes marked as "Ϫ"). This increase by both co-cultures was prevented by added catalase and vitamin E, indicating the involvement of ROS. In addition, the increase in laminin ␤1 and ␥1 proteins by both co-cultures was decreased by CYP2E1 inhibitors: 5 mM diallylsulfide, 0.1 mM sodium diethyldithiocarbamate, and 10 M phenyl isothiocyanate (Fig. 5). These results suggest that CYP2E1-derived ROS contribute to the increase in laminin ␤1 and ␥1 protein content of HSC in the saline-hepatocyte/HSC co-culture, and to the further increase produced by the pyrazole-hepatocytes/HSC co-culture.
CYP2E1-derived Diffusible Mediators Increase the Synthesis of Laminin ␤1 and ␥1 Proteins by HSC but Do Not Affect the Turnover of Newly Synthesized Laminin ␤1 and ␥1-The results described above indicate that co-culture of HSC with E47 cells increases laminin ␤1 and ␥1 protein levels. This effect could involve transcriptional activation of the LAM␤1 and LAM␥1 genes with elevated mRNA synthesis and/or mRNA stability, increased translational efficiency, or decreased turnover of newly synthesized laminin ␤1 and ␥1 protein, with subsequent accumulation in the HSC. To address these possibilities, direct analysis of laminin ␤1 and ␥1 protein synthesis, turnover, and mRNA levels by the co-cultures was performed. For the subsequent experiments, because of the need for large amounts of HSC and because transfection experiments with reporter constructs were to be used, a HSC-T6 cell line with a higher efficiency for transfection was used rather than primary HSC.
Laminin ␤1 and ␥1 protein synthesis was assessed by labeling with [ 35 S]methionine for varying times, up to 12 h, followed by immunoprecipitation with anti-laminin 1 IgG as described under "Material and Methods." There was a strong increase in the incorporation of [ 35 S]methionine into laminin ␤1 and ␥1 in the HSC/E47 co-culture compared with HSC cultured alone or in the presence of the C34 cells (Fig. 6, A and B). Cycloheximide blocked laminin ␤1 and ␥1 synthesis in all systems, validating that the increase in the laminins ␤1 and ␥1 [ 35 S]methionine signal was caused by a protein synthesis-dependent reaction. Total protein synthesis by HSC was not altered by co-culturing with E47 cells (Fig. 6C). These results suggest that an increase in laminin ␤1 and ␥1 synthesis is associated with the induction of both proteins in the presence of CYP2E1-mediated oxidative stress. Although synthesis of the laminin ␣2 chain was not determined in these experiments, Western blot analysis confirmed that laminin ␣2 chain levels were elevated after 12 h of culture with the HSC/E47 co-culture (Fig. 6D), analogous to the increase found with primary HSC co-cultured with E47 cells (Fig. 3A).
To study the apparent turnover of laminin ␤1 and ␥1 chains, the HSC were first labeled with synthesized laminin ␤1 and ␥1 subunits and total proteins was calculated from the semilogarithmic plot of counts incorporated per minute versus time. Prior to the initiation of the chase (0 h), there was an increase in [ 35 S]methionine-labeled laminin ␤1 and ␥1 in the HSC/E47 co-culture (as described above) (Fig. 7,  A and B). Pulse-chase experiments revealed that the turnover of newly synthesized laminin ␤1 and ␥1 (reflected as time for 50% loss of intracellular [ 35 S]methionine counts) was 5.6 h for stellate cells cultured alone, 5.7 h for stellate cells co-incubated with C34 cells, and 6.0 h for stellate cells co-cultured with E47 cells (Fig. 7, A and B).
Loss of intracellular radioactivity incorporated into laminin from the HSC labeled with methionine could reflect turnover of the labeled laminin ␤1 and ␥1 proteins because of intracellular degradation or secretion into the medium or both. To evaluate the latter, samples of media from the pulse-chase experiment were collected at the same time points, and immunoprecipitation of laminins ␤1 and ␥1 chains was carried out as described above (Fig. 8). Very low rates of secretion of the newly synthesized laminin were observed over the 12-h chase under these conditions. In fact, 25 times more protein had to be immunoprecipitated to observe the bands shown in Fig. 8A than for the results of Fig. 7A. The loss of intracellular cpm as a function of time cannot be accounted for by the small secretion of newly synthesized laminin ␤1 and ␥1 during the 12-h chase as the increase in total medium cpm was just a few percent of the decline in total cellular cpm. The amount of laminin secreted to the culture medium was minimal in both co-culture systems. Yurchenco et al. (11) expressed the ␣, ␤, and ␥ chains subunits of laminin 1 in all combinations in a near-null background, and showed that in the absence of its normal partners, the ␣ chain is secreted as intact protein and protein that had been cleaved in the coiled-coil domain. In contrast, the ␤ and ␥ chains, expressed separately or together, remain intracellular with formation of ␤␤ or ␤␥, but not ␥␥, disulfide-linked dimers. Secretion of the ␤ and ␥ chains required simultaneous expression of all three chains and their assembly into ␣␤␥ heterotrimers. They concluded that the ␣ chain can be delivered to the extracellular environment as a single subunit, whereas the ␤ and ␥ chains cannot, and that the ␣ chain drives the secretion of the trimeric molecule. Such an ␣ chain-dependent mechanism could allow for the regulation of laminin export into a nascent basement membrane, and might serve an important role in controlling basement-membrane formation. As mentioned above, the laminin ␣1 chain was not observed in the HSC, whereas the ␣2 chain was detected both in cell lysates FIG. 7. Turnover of intracellular laminin ␤1 and ␥1 proteins made by the HSC in the co-cultures. A, turnover of laminin ␤1 and ␥1 was studied in HSC cultured alone or with C34 or E47 cells after pulsing with [ 35 S]methionine for 24 h, followed by chasing with complete MEM supplemented with cold methionine for 0, 1, 2, 4, 8, and 12 h. Immunoprecipitation of 40 g of HSC protein with antilaminin 1 antibody, washing, and fluorography was carried out as described under "Materials and Methods" and in Ref. 19. Laminin ␤1 and ␥1 levels were determined from PhosphorImager analysis of fluorographs shown in A, whereas total protein turnover was determined from the decline in trichloroacetic acid-precipitable counts. B and C, turnover of intracellular laminin ␤1 and ␥1 and total protein, respectively. and in the culture medium, suggesting that the laminin ␣2 chain was driving secretion of the ␤1 and ␥1 subunits into the medium. Importantly, the experiments described above suggest that turnover of laminin ␤1 and ␥1 proteins was similar in both co-culture systems and not likely to explain the increase in laminin ␤1 and ␥1 proteins by the HSC/E47 co-culture. Turnover of total HSC proteins were also similar (about 3.6 to 3.9 h) for HSC cultured alone, or co-incubated with C34 or E47 cells (Fig. 6C). There were no differences in the secretion of total proteins in both co-culture systems (Fig. 7C).
Induction of Laminin ␤1 and ␥1 Proteins in the E47 Coculture Involves Transcriptional Regulation-To determine why laminin ␤1 and ␥1 synthesis was elevated, total RNA was isolated from HSC-T6 cultured alone or with C34 or E47 cells and was analyzed by Northern blot for laminin ␣1, ␤1, and ␥1 chains mRNAs. Laminin ␣1 mRNA was not detected in any of the cell models. The HSC/C34 co-culture resulted in an increase in laminin ␤1, but not ␥1 mRNA levels over mRNA levels in the HSC cultured alone (Fig. 9A). The HSC/E47 coculture produced an increase in laminin ␥1 mRNA as well as further induction of laminin ␤1 mRNA. Overall, there was a 2to 3-fold increase in ␤1 and ␥1 laminin mRNA levels in HSC co-cultured with E47 cells compared with the HSC/C34 coculture (Fig. 9A). Nuclear in vitro transcription assays were performed to study whether the CYP2E1-mediated effect on laminin ␤1 and ␥1 mRNA levels in HSC is regulated at the transcriptional level. As shown in Fig. 9B, enhanced laminins ␤1 and ␥1 expression occurs through a transcriptional mechanism with both co-culture systems, however, the newly transcribed laminin ␤1 and ␥1 mRNAs were increased in HSC co-incubated with E47 cells when compared with HSC co-incubated with C34 cells. It is not understood why laminin ␥1 transcription appears to be up-regulated in the HSC/C34 coculture compared with HSC alone, whereas the mRNA levels are similar.
Identification of the Sequences of the LAM␥1 Promoter in HSC Required for CYP2E1-mediated Responsiveness-Transient-transfection experiments in HSC-T6 with chimeric constructs harboring progressive 5Ј deletions of the LAM␥1 promoter linked to the CAT reporter gene were performed to identify the regions of the LAM␥1 promoter required for CYP2E1dependent activation. HSC cells were transfected with the constructs shown in Fig. 10. The percentage of acetylation of chloramphenicol in the HSC co-cultured with E47 cells and transfected with the Ϫ330LAM␥1-CAT and the Ϫ1400LAM␥1-CAT was significantly higher (about 46-fold) than in HSC cultured alone or with C34 cells (Fig. 10, A and B). These data are consistent with previous findings (Fig. 9, A and B) that CYP2E1-FIG. 8. Secretion of laminin ␤1 and ␥1 during the pulse-chase in the cocultures. A, secretion of laminin ␤1 and ␥1 into the medium was determined at the same time-points indicated for the pulse-chase (Fig. 7), immunoprecipitating 1 mg of total protein with the anti-laminin 1 antibody. Laminin ␤1 and ␥1 levels were determined from PhosphorImager analysis of fluorographs shown in A, whereas total protein secretion was determined from the trichloroacetic acid-precipitable counts. B and C, secretion of laminin ␤1 and ␥1 and total protein, respectively. The cpm refers to total accumulation in the medium at the indicated time point as the medium was not changed during the 12 h incubation. dependent LAM␥1 activation is exerted, at least in part, at the transcriptional level. Interestingly, the activity of the Ϫ330LAM␥1-CAT and Ϫ1400LAM␥1-CAT constructs were very similar (about 45% acetylation of chloramphenicol). On the other hand, the activity of the Ϫ580LAM␥1-CAT (and Ϫ2500LAM␥1-CAT) constructs was significantly lower (4% of acetylation of chloramphenicol) and very close to the basal levels of promoter activity by the shorter constructs (Ϫ200LAM␥1-CAT and Ϫ250LAM␥1-CAT). These data suggest that the Ϫ230 to Ϫ480 region (and perhaps the Ϫ1400 to Ϫ2500 region) of the LAM␥1 gene may contain a silencer-like element which reduces promoter activity. Reporter activity was similar for HSC incubated alone compared with the HSC/C34 co-culture for the various constructs, which is in agreement with the similar laminin ␥1 mRNA levels found in the HSC compared with the HSC/C34 co-culture (Fig. 9A). In general, the pattern of promoter expression was similar for all three systems (HSC alone, HSC/C34, and HSC/E47 co-culture), i.e. all three systems showed positive response to the Ϫ330LAM␥1-CAT and Ϫ1400LAM␥1-CAT constructs, and negative responsiveness to the Ϫ580LAM␥1-CAT and Ϫ2500LAM␥1-CAT constructs. However, the E47 co-culture clearly showed the most robust responses to the Ϫ330LAM␥1-CAT and Ϫ1400LAM␥1-CAT constructs, likely a reflection of the presence of redoxsensitive sites in these regions. Thus, the promoter responses are not unique for CYP2E1 effects but are enhanced by CYP2E1-derived diffusable factors.

CYP2E1-derived Oxidative Stress Transactivates the LAM␥1 Promoter in HSC Co-incubated with E47 Cells through a Sp1dependent Mechanism-The LAM␥1
promoter possesses several putative binding sites for redox-sensitive transcription factors including Sp1, AP-1, and NFB (30). To determine whether any of these transcription factors could, in response to the increase in ROS produced by the E47 co-culture, transactivate the LAM␥1 promoter, electrophoretic mobility-shift assays were carried out with nuclear extracts from HSC incubated alone or cultured with either C34 or E47 cells (Fig. 11A  and inset). HSC co-cultured with E47 cells showed increased binding of Sp1 to an oligonucleotide containing its putative binding site GGGCGG when compared with HSC cultured alone or with C34 cells. There were no changes in the binding activity of NFB and AP-1 (The inset shows an electrophoretic mobility-shift assay loading only 0.5 g of protein from the same samples). Competition studies with a 1000-fold excess of cold Sp1 oligonucleotide blunted the binding of the radiolabeled Sp1 oligonucleotide. The complex of Sp1 protein-DNA was supershifted with an anti-Sp1 antibody, demonstrating specificity of the complex (Fig. 11A, last lane). Southwestern analysis carried out with a double-stranded oligonucleotide obtained by PCR amplification of the Ϫ230 to Ϫ150 region of the LAM␥1 promoter (where strong reporter activity was noted, Fig. 10) and nuclear proteins from HSC cultured alone, with C34 cells, or with E47 cells, showed a single band of about 100 kDa, which corresponds to the molecular mass of Sp1 (95-106 kDa) for all samples (Fig. 11B). The binding activity was comparable between the HSC and the HSC/C34 co-culture but was 2-fold higher in HSC incubated with E47 cells. The increase in binding activity for the E47 co-culture was prevented by the CYP2E1 inhibitor diallylsulfide and by the free-radical-scavenging agent tempol (Fig. 11B). To verify that the band was indeed Sp1, a Southwestern analysis was carried out in which the membrane was incubated with an excess of anti-Sp1 antibody before hybridization with the double-stranded DNA oligonucleotide. No signal was obtained (data not shown), indicating that the detected band was Sp1, and suggesting that the protein mediating the CYP2E1 effects on the LAM␥1 promoter may be Sp1. To eliminate the possibility that increased binding and transactivation of the LAM␥1 promoter could be caused by increased levels of Sp1 in the HSC co-cultured with E47 cells, a Western blot analysis of total Sp1 protein was carried out with nuclear protein extracts. Results in Fig. 11C show that no differences in Sp1 protein content were observed in HSC cultured alone or with C34 or E47 cells. Finally, co-transfection of HSC with an Sp1 expression vector plus the Ϫ330LAM␥1-CAT reporter construct, which contains seven putative binding sites for Sp1 and whose binding activity was induced about 8-fold in the HSC/E47 co-culture compared with the C34 system (Fig.  10B), were carried out (Fig. 11D). Although the co-transfection with the Sp1 expression vector enhanced reporter activity in all three systems (HSC alone, HSC/C34, and HSC/E47), HSC transfected with the Sp1 expression vector and co-cultured with E47 cells showed higher CAT activity for the Ϫ330LAM␥1-CAT reporter construct than did HSC cultured alone or with C34 cells (about 5-fold) (Fig. 11D). A Western blot analysis was also carried out to validate that Sp1 levels were elevated after transfection and levels were comparable in the three HSC systems (Fig. 11D, lower panel). DISCUSSION Basement membranes are cell-associated heteropolymers that are essential for tissue development and maintenance. The functions of these extracellular matrices are both architec- FIG. 9. Induction of laminin ␤1 and ␥1 in the E47 co-culture involves transcriptional regulation. A, total RNA was isolated from HSC cultured alone or with C34 and E47 cells and analyzed by Northern blot for laminin ␣1, ␤1, and ␥1. Laminin ␣1 mRNA was not detected. Numbers below the blot refer to the ratio of arbitrary densitometric units of laminin ␤1 or ␥1/GAPDH. B, nuclear in vitro transcription assays for laminin ␤1 and ␥1 were carried out with newly transcribed mRNA from HSC cultured alone or with C34 or E47 cells. The signals of newly transcribed GAPDH and S14 mRNAs were used as housekeeping genes. There was a slight increase in both the GAPDH and S14 signals by the C34 or E47 cell co-cultures perhaps reflecting a general effect by factors released from hepatocytes on stellate cell transcription. However, the increase was the same for both co-cultures, and results were normalized to account for the slight increase in the loading controls. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
tural and informational, with basement membranes acting as substrata, filters, and solid-phase agonists (31)(32)(33). In the liver, laminin is present during maturation of the organ and accumulates in the adult during the capillarization process that occurs in alcoholic cirrhosis and hepatocarcinogenesis (8,34). Laminin ␥1 mRNA is abundant in HSC, the major site of matrix formation, and it is also expressed at high levels in transformed hepatoma cell lines and during chemically induced hepatocarcinogenesis in the rat (6,35), but it is not present in normal hepatocytes (8). The presence of laminin ␥1 chain is a prerequisite for basement-membrane formation, with its absence causing an early embryonic lethality (36). Besides self-assembling, laminins also interact with other laminin isoforms as well as with nonlaminin extracellular matrix molecules such as perlecan, nidogen, and collagen to form a polymeric matrix; such associations appear necessary for proper basement membrane assembly (1).
Previous work in our laboratory has been aimed at characterizing the intercellular communication between the hepatocyte and the stellate cell to better understand the mechanisms by which oxidative stress and other mediator molecules perpetuate the fibrogenic response in stellate cells (18,19). A co-culture model containing HSC with a HepG2 cell line that overexpresses cytochrome P450 2E1 (E47 cells) or a control HepG2 cell line (C34 cells) was developed. The interest in CYP2E1 was due to its activation of numerous hepatotoxins, generation of oxidative stress, and possible role in contributing to alcohol-induced liver injury (14,17,(37)(38)(39)(40)(41)(42). This co-culture system, with constant generation of ROS, revealed increased translation of collagen mRNA by a ROS-dependent mechanism. Because of its importance as an abundant extracellular matrix component whose levels are elevated during liver injury, we extended our co-culture studies to the possible regulation of laminin production under oxidative stress conditions generated by CYP2E1.
Kleinman et al. (43) have shown that normal adult liver contains both laminin ␤1 and ␥1 chains but lacks the ␣1 chain. Initial results shown in Figs. 2 and 3 revealed a time-dependent increase in both intra-and extracellular laminin ␤1, ␥1, and ␣2 proteins (after 3 days of culture) which was enhanced in HSC co-cultured with E47 cells compared with the HSC/C34 co-culture. The laminin ␣1 chain could not be detected, confirming its lack in normal liver (43). To show interactions between the ␤1, ␥1, and ␣2 laminin chains, Western blots of HSC lysates collected after 3 or 5 d of co-culture were carried out under nondenaturing conditions. The laminin 1 antibody used recognizes the ␣1, ␤1, and ␥1 chains of laminin. Two bands of about 400 and 800 kDa were found after 5 d of culture, whereas the 400-kDa band was found after 3 d of culture. The 400-kDa band may likely be the dimers ␤1/␥1, ␤1/␤1, and/or ␥1/␥1, because the molecular mass of the ␤1 chain is about 220 kDa and that of the ␥1 chain is about 200 kDa. The 800-kDa band may likely be the heterotrimer ␣2␤1␥1 because the molecular mass of the ␣2 chain under native conditions would be of about 380 kDa plus 220 kDa of the ␤1 and 200 kDa of the ␥1. Although other ␣ chains (␣3, ␣4, and ␣5) have been described, none of them have been detected in the liver except for the ␣2 chain (9). Of note is the fact that although there seems to be a time-dependent increase in laminins ␣2, ␤1, and ␥1 with higher levels of expression in the E47 co-culture, the potential assembly of the three subunits to the 800-kDa heterotrimer is delayed, relative to the increase in levels of the individual chains, or to the assembly into dimers of ␤1 and ␥1 (400 kDa).
The amount of H 2 O 2 and lipid peroxidation end products in HSC or in the medium was previously shown to be increased by the E47 cell co-culture (18,19). The antioxidant defense in HSC did not change with any of the culture conditions (data not shown). Evidence for oxidative stress involvement in the increase in laminin ␤1 and ␥1 proteins by the E47 co-culture is based on the prevention of these effects by addition of antioxidants such as catalase or vitamin E to the incubation medium and by CYP2E1 inhibitors such as diallylsulfide, 4-methyl pyr- FIG. 11. CYP2E1-derived oxidative stress transactivates the LAM␥1 promoter in HSC co-incubated with E47 cells through a Sp1-dependent mechanism. A, electrophoretic mobility-shift assays were carried out with nuclear extracts (5 g of protein) from HSC cultured alone or with C34 or E47 cells using radiolabeled oligonucleotides containing the consensus sequence for either Sp1, AP-1, or NFB as described under "Material and Methods." The Sp-1/DNA complex was competed with a 1000-fold of cold probe and supershifted with an anti-Sp1 antibody. The inset shows an electrophoretic mobility-shift assay for AP-1 carried out with 0.5 g of nuclear extract. B, Southwestern analysis was carried out using an [␣-32 P]dCTP double-stranded oligonucleotide obtained by PCR amplification of the Ϫ230 to Ϫ150 region of the LAM␥1 promoter, and nuclear protein extracts from HSC cultured alone or with C34 cells, or with E47 cells that were incubated in the absence or presence of 5 mM diallylsulfide (DAS), a CYP2E1 inhibitor, or 10 M tempol, a spin trap agent. C, Western blot analysis showing no differences in the expression of Sp1 protein in HSC in the three cell culture systems. D, CAT assay with cell extracts from HSC co-transfected with a Sp1-expression vector plus the Ϫ330LAM␥1-CAT deletion construct. A Western blot analysis of the Sp1 levels after transfection is shown at the bottom of the figure. azole, sodium diethyldithiocarbamate, and phenylisothiocyanate. Transfection of E47 cells with an antisense CYP2E1 construct lowered laminin ␤1 and ␥1 protein expression to basal levels, whereas transfection with a sense CYP2E1 plasmid into E47 or C34 cells further increased laminin ␤1 and ␥1 expression. The relevance of these findings was further extended to co-cultures of HSC with primary hepatocytes from pyrazole-treated rats, with high CYP2E1 content, when compared with co-cultures with hepatocytes from saline-treated rats.
Experiments were carried out to determine the mechanism(s) responsible for the increase in laminin ␤1 and ␥1 levels by the E47 co-culture. The synthesis of laminin ␤1 and ␥1 chains was similar between HSC cultured alone or with C34 cells, but an increase in synthesis was found in HSC cultured with E47 cells. This effect was blocked by cycloheximide. The turnover of laminin ␤1 and ␥1 proteins was similar in HSC cultured alone or with C34 or E47 cells, and the export of laminin ␤1 and ␥1 chains to the culture medium was very low during the 12-h chase; thus, changes in laminin ␤1 and ␥1 degradation and/or secretion do not account for the increase in laminin ␤1 and ␥1 proteins. These results indicate that increased synthesis of laminin protein is one mechanism of regulation for the induction of laminin ␤1 and ␥1 protein by the E47 co-culture. We next analyzed whether elevated mRNA levels could account for the increase in synthesis of laminin protein found in the E47 system. Northern blot analysis revealed elevated laminin ␤1 and ␥1 mRNAs in the E47 compared with the C34 co-culture. Nuclear run-on experiments documented increased synthesis of both mRNAs. Thus, enhanced laminin ␤1 and ␥1 expression in the E47 system results from transcriptional activation of the LAM␤1 and LAM␥1 genes. There is specificity in the ability of CYP2E1-derived mediators to interact with the HSC and stimulate the synthesis of certain proteins, e.g. collagen (19) and laminin, whereas total protein synthesis is not altered, nor is the synthesis of other HSC proteins such as catalase, tissue inhibitor of metalloproteinase 1, or metalloproteinase 13 altered (19).
Consistent with these results, transient transfection of HSC co-cultured with E47 cells with chimeric constructs driven by different sequences of the LAM␥1 promoter indicated the presence of two redox-sensitive enhancer elements located in the Ϫ230 to Ϫ150 and Ϫ1300 to Ϫ480 regions that were not stimulated in HSC cultured alone or with C34 cells. Little data are available on the molecular mechanisms involved in the transcriptional regulation of basement-membrane genes in liver fibrosis. Laminin is highly expressed in hepatic fibrosis (44). The 5Ј-untranslated region of LAM␥1 contains a stem-loop structure spanning from ϩ76 to ϩ106. Deletion of 47 bp within the 5Ј-untranslated region (ϩ59 to ϩ106) of the LAM␥1 completely blocked promoter activity in astrocytes, confirming that this downstream region could be one of the major points of transcriptional regulation (30). The chimeric constructs used for transient-transfection studies in the HSC systems contained the 5Ј-untranslated region of the first exon, which is GC-rich and has a stem loop structure; whether these could be redox-sensitive and operate coordinately with other factors released by the E47 system is not known.
The regulation of the expression of laminin ␥1 mRNA in hepatoma cells involves several regions within the 2-kb promoter, and transfection of LAM␥1 promoter fragments in these cells indicated that regulatory elements are located between Ϫ594 and Ϫ94 bp (35). The Ϫ230 to Ϫ150 region of the promoter contains several monomeric Sp1-binding sites and a cAMP-responsive element. The LAM␥1 gene promoter contains multiple cognate sites for Sp1 binding which have the ability to recruit other transcription factors to initiate transcription from TATA-less promoters (45). TATA-less promoters typically have multiple transcription-initiation sites that are located very close or within the regions that contain Sp1-binding sites. The multiple Sp1-binding sites in these classes of genes suggest that they could be redox-sensitive promoters. Overexpression of Sp1 in normal hepatocytes increases endogenous LAM␥1 gene expression and co-transfected LAM␥1 promoter (46). High-binding activity was observed in Sp1-transfected nuclear extracts. Sp1 and laminin ␥1 mRNA are both highly expressed in human hepatocarcinomas, particularly at the invasive front (47).
In view of the above, we evaluated Sp1-binding activity in the different HSC systems. Electrophoretic mobility-shift assays showed increased Sp1-binding activity in nuclear extracts from HSC co-incubated with E47 cells compared with that of HSC cultured alone or with C34 cells. The DNA-protein complex was shifted by an anti-Sp1 antibody and competed by a 1000-fold excess of cold oligonucleotide containing the Sp1binding site. Two other well known redox-sensitive transcription factors, AP-1 and NFB, showed the same binding activity in all three systems. The AP-1 site located at Ϫ650 bp is not involved in LAM␥1 transcription in hepatoma cells, whereas several sequences in the Ϫ480-bp to Ϫ175-bp region have been identified and may bind specific regulatory factors, including Sp1 and immediate early gene products coded by the erg-family (48). The results described above suggest that the increased promoter activity in transient-transfection studies with the Ϫ230 to Ϫ150 reporter construct could be mediated by increased binding of the redox-sensitive transcription factor Sp1 to this region.
To verify this, we performed Southwestern analysis to determine the binding capacity of the Ϫ230 to Ϫ150 region of the promoter to nuclear proteins from HSC cultured alone or with C34 or E47 cells. This region of the promoter binds to a protein of about 100 kDa, and the binding is increased 2-to 3-fold in HSC co-cultured with E47 cells. A role for a CYP2E1-mediated effect and for ROS was validated by addition of diallylsulfide, a CYP2E1 inhibitor, and of 4-hydroxy-tempo (tempol), a widespectrum free radical scavenger, both of which were able to prevent the enhanced Sp1 binding of the E47 co-culture. Sp1 is a dimer of molecular masses 95 and 106 kDa. To determine whether this binding protein was Sp1, the same samples were analyzed by Southwestern blot incubating the membrane with anti-Sp1 antibody before hybridization with the Ϫ230 to Ϫ150 double-stranded oligonucleotide; no binding was detected after incubation with the antibody. To ensure that the increase in Sp1 binding observed was not a result of increased Sp1 synthesis, a Western blot analysis was carried out with nuclear proteins from HSC cultured alone or with C34 or E47 cells, but no differences were observed among the three systems. These results suggest that the increased LAM␥1 promoter activity found with the HSC/E47 co-culture transfected with the Ϫ230 to Ϫ150 construct could be due to increased Sp1 binding. Furthermore, in experiments in which a Sp1 expression vector was co-transfected along with the Ϫ330LAM␥1-CAT reporter construct, a 5-fold increase in CAT activity was detected in the HSC/E47 co-culture compared with HSC cultured alone or with C34 cells.
The LAM␥1 gene is transcriptionally up-regulated by interleukin-1␤ due to an increased binding of NFB to a B consensus sequence on the LAM␥1 promoter (49). Both interleukin-1␤ and TGF-␤ transiently increase laminin ␥1 mRNA due to enhanced binding of nuclear proteins on the GC-rich bcn-1 motif in the promoter. The cooperative induction of the LAM␥1 promoter and the endogenous LAM␥1 gene by TFE3 and Smad3 is augmented by the TGF-␤ signaling pathway (50). In our reporter assays, there was decreased activity in all systems (HSC alone, HSC/C34, and especially HSC/E47 co-culture) with the Ϫ2500LAM␥1-CAT construct, suggesting the presence of a silencer-like element between the Ϫ1300 and Ϫ2400 region. Not much is known about the sequence of the LAM␥1 promoter upstream of Ϫ1000 bp. We have not analyzed possible redoxsensitive elements further upstream that could be responsible for the transactivation of the Ϫ1400 to Ϫ480 region of the promoter. Whether this could be mediated directly by ROS or involve other factors such as cytokines, most of which are redox-sensitive molecules, still remains to be elucidated.
In summary, these results suggest that CYP2E1, present in the hepatocyte, can release diffusible mediators, most likely stable ROS such as H 2 O 2 and lipid peroxidation metabolites, which can up-regulate the LAM␥1 gene, with a subsequent increase in synthesis of laminin ␤1 and ␥1 proteins. Up-regulation of the LAM␥1 gene is due, in part, to enhanced Sp1 binding to the Ϫ230to Ϫ150-bp promoter region, which contains several redox active binding sites. The enhanced Sp1 binding appears to be due to redox activation of Sp1 via CYP2E1-derived diffusible ROS. Such up-regulation of important matrix-synthesizing genes may play a role in liver injury produced by hepatotoxins activated by CYP2E1, e.g. CCl 4 , benzene, acetaminophen, nitrosamines, and perhaps in mechanisms of alcohol-induced liver injury.