c-jun NH2-terminal kinase pathway is involved in constitutive matrix metalloproteinase-1 expression in a hepatocellular carcinoma-derived cell line

Authors


Abstract

Transcription factor c-Jun serves for cellular proliferation, survival, differentiation and transformation and is recognized as an important factor in cancer development, including hepatocellular carcinoma (HCC). The purpose of present study is to determine the involvement of c-Jun in matrix metalloproteinase-1 (MMP-1) expression, which is previously reported by us to be expressed only in the early stage of human HCC showing stromal invasion. Of 5 human HCC cell lines examined, only HLE cells revealed mRNA and protein expression as well as enzymatic activity of MMP-1. Transient transfection of an MMP-1 promoter/luciferase construct (including 4.4 kb full promoter region) into HLE and HCC-T cells (MMP-1 nonproducer) showed that high promoter activity was observed only in HLE cells without inducers, and that this promoter activity was still observed when a shorter 0.6 kb proximal promoter construct was transfected. The 0.6 kb promoter region contained 3 AP-1 sites, and c-jun mRNA was constitutively expressed in HLE cells without inducers. Furthermore, phosphorylated c-Jun and c-Jun NH2-terminal kinase (JNK) were detected in HLE cells. Promoter activity of the 0.6 kb construct was suppressed with SP600125, a potent inhibitor of JNK, but not with PD98059 and SB203580, potent inhibitors of MEK1/2 and p38, respectively. The inhibitory effect of SP600125 was also observed at protein expression level and in enzymatic activity of MMP-1. Taken together, this study suggests that the JNK pathway is involved in the expression of MMP-1 in HCC cells and may represent a new functional role of c-Jun for HCC development. © 2004 Wiley-Liss, Inc.

Invasion and metastasis are the major causes of treatment failure in patients with cancer. Enzymatic degradation of different macromolecular components of the extracellular matrix (ECM), which is an essential step in the process of invasion and metastasis, plays a key role in the dissemination of cancer cells.1, 2 Of several proteolytic enzymes, it has been clarified that metalloproteinases (MMPs) are responsible for ECM destruction, and that they participate in cancer cell invasion and metastasis.3, 4

We previously showed that only well-differentiated cancer cells of early hepatocellular carcinoma (HCC), smaller than 2 cm in diameter, express matrix metalloproteinase-1 (MMP-1) by in situ hybridization and immunohistochemistry.5 Early HCC is usually described as well-differentiated carcinoma6 and stromal invasion of cancer cells is a common finding in early HCC.7 As HCC is usually associated with liver fibrosis/cirrhosis,8, 9, 10 where type 1 collagen is mainly deposited, it is quite likely that early HCC cells invade surrounding fibrous tissue by secreting MMP-1. MMP-1 expression is only detected in the early stage of HCC, which is coincident with clinical feature of HCC. In advanced stage of HCC, they are usually encapsulated with fibrous tissue and invade surrounding fibrous tissues no longer.

A recent study revealed that MMP-2 and membrane type 1-matrix metalloproteinase (MT1-MMP) can cause the cells to proliferate11 besides their well-known original function, that is, ECM resolution. Moreover, MT1-MMP was reported to increase cell mobility.12 MMP-1 was also supposed to cause the proliferation of hepatocytes in the rat fibrotic liver infected with recombinant adenovirus harboring human MMP-1 gene.13

c-Jun and activated c-Jun by a phosphorylation cascade of mitogen-activated protein kinase (MAPK) families have been shown to play an important role in embryonic cell differentiation, apoptosis and proliferation as well as carcinogenesis of hepatocytes.14 The phosphorylation of c-Jun is conducted by c-Jun NH2- terminal kinases (JNK) among 4 distinctly regulated groups of MAPK pathways; the other 3 groups are extracellular signal-related kinases (ERK)-1/2, p38 proteins and ERK5.15, 16, 17, 18

Several recent studies describing involvement of c-Jun in the early stage of HCC development14 or in liver regeneration after partial hepatectomy19 remind us that JNK pathway has a key role in the MMP-1 gene expression in HCC. MMP-1 gene expression is truly known to be regulated by c-Jun, but the regulation of MMP-1 gene expression by JNK pathway in HCC cells remains unknown. The present study has shown for the first time that one HCC cell line, which constitutively expresses MMP-1 without any stimulators such as phorbor ester, is under control of transcriptional regulation of the MMP-1 gene transcription via the activation of c-Jun through JNK pathway. These results may indicate an additional functional role of c-Jun in HCC development.

MATERIAL AND METHODS

HCC cell lines

Human HCC cell lines HLE,20 PLC/PRF/521 and Huh-722 were obtained from the Japanese Cancer Research Resources Bank (Osaka, Japan). HCC-M and HCC-T were previously established by us.23, 24 HCC-M, HLE, PLC/PRF/5 and Huh-7 were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), nonessential amino acids and antibiotics. HCC-T was cultured with RPMI-1640 containing 10% FBS and antibiotics.

Reagents

c-Jun NH2-terminal kinase inhibitor SP600125 was purchased from Calbiochem-Novabiochem (San Diego, CA). MEK/ERK inhibitor PD98059 and p38 inhibitor SB203580 were obtained from Sigma Chemical (St. Louis, MO).

RNA isolation and RT-PCR

Isolation of total RNA and RT-PCR were performed as described previously.25 To amplify each gene, a pair of sense and antisense primers (Table I) were chosen with the help of a computer program (Oligo 5.0, Primer analysis software, National Bioscience, Plymouth, MN). Each target gene was amplified in the GeneAmp PCR System 9600 (Perkin Elmer, Norwalk, CT). The initial denaturation was at 94°C for 2 min, followed by each cycle of reaction at 94°C for 30 sec, at the described annealing temperature for 30 sec (Table I), and at 72°C for 30 sec, and followed by postextension at 72°C for 7 min. Glyceraldehyde 3-phosphate dehydrogenase (G3PDH), which was used as an internal control, was amplified for 23 cycles as previously described.25 The PCR products were separated by electrophoresis on a 1.5% agarose gel, stained with ethidium bromide.

Table I. Primers for RT-PCR Amplification
Gene Nucleotide sequenceAnnealing temperature (°C)CycleProduct (bp)
AlbuminSense5′-CCCCGGAACTCCTTTTCTTTG-3′5640675
 Antisense5′-CATCGAACACTTTGGCATAGCA-3′   
HNF-4Sense5-′CTGCTCGGAGCCACAAAGAGATCCATG-3′5730371
 Antisense5-′ATCATCTGCCACGTGATGCTCTGCA-3′   
AFPSense5′-CGCTGGAACGTGGTCAATGTA-3′5640706
 Antisense5-′CACCCTGAGCTTGGCACAGA-3′   
MMP-1Sense5′-GGTGCCCAGTGGTTGAAAAAT-3′5735716
 Antisense5′-CATCACTTCTCCCCGAATCGT-3′   
MMP-2Sense5′-TCTTCCCTCGCAAGCCCAAGT-3′5735685
 Antisense5′-ACAGTGGACATGGCGGTCTCAG-3′   
MMP-9Sense5′-TGGGCTACGTGACCTATGAC-3′5935200
 Antisense5′-CAAAGGTGAGAAGAGAGGGC-3′   
TIMP-1Sense5′-TTCTGCAATTCCGACCTCGTC-3′5830385
 Antisense5′-GCAGTTTGCAGGGGATGGATA-3′   
c-junSense5′-CCTGTTGCGGCCCCGAAACT-3′6230495
 Antisense5′-ACCATGCCTGCCCCGTTGAC-3′   
c-fosSense5′-TTTGCCTAACCGCCACGATGAT-3′6230500
 Antisense5′-TTGCCGCTTTCTGCCACCTC-3′   
G3PDHSense5′-ACCACAGTCCATGCCATCAC-3′6223452
 Antisense5′-TCCACCACCCTGTTGCTGTA-3′   

Zymography

To examine the gelatinolytic activity, gelatin zymography was performed as described before25 with a slight modification. After cells were cultured in serum-free media for 48 hr, aliquots of the culture media from HCC cells were mixed with 4 × gel loading buffer (10% SDS, 4% sucrose, 0.25 M Tris-HCl, pH 6.8, 0.1% bromphenol blue) without boiling, then electrophoresed on a 10% polyacrylamide gel containing 1 mg/mL of gelatin. The volume of media loaded was adjusted according to the cell number. Gels were scanned in a digital scanner and densitometric measurement was performed with NIH Image software (version 1.55).

Nucleotide sequence analysis for single nucleotide polymorphism at −1607 bp

To detect the known 1G/2G polymorphism at −1607 bp in the MMP-1 promoter region, PCR amplification was performed with a pair of primers, M-F (5′-ACATGTTATGCCACTTAGAT-3′; −1654/−1635) and M-R (5′-TCCCCTTATATGGATTCCTGTT-3′; −1536/−1517), followed by nucleotide sequence analysis.26

Plasmid constructs

A fragment encompassing the essential sequence for transcriptional activity of MMP-1 promoter was amplified from genomic DNA isolated from HCC-T cells using the following primers: sense, 5′-TTTCAAATCCATCTCAAATTCACA-3′ (−4363/−4340); antisense, 5′-ACTGGCCTTTGTCTTCTTTCTCAG-3′ (+49/+72). BglII digestion of the resulting 4,429 bp PCR product (−4363/+72; 4.4 kb construct) yielded a 1.2 kb fragment (−1196/+72), which was cloned into the pGL3 Basic vector (Promega, Tokyo, Japan) that had been treated to have the same termini. A 5′ deletion promoter 0.6 kb construct (−522/+72) was generated from this construct (−1196/+72) by using convenient restriction site (KpnI at −517 bp). Those chimeric constructs were sequenced and found to be identical to the previously reported sequence27 except that our sequence from HCC-T had a “T” at site −320 nucleotide, where the published sequence contained a “C” at that site. Expression plasmid pCMV-jun was the kind gift of Dr. Tom Curran.

Transient transfection and assessment of promoter activity

The MMP-1 promoter/luciferase constructs were transfected into HLE and HCC-T cells using the calcium phosphate-DNA coprecipitation method. In some experiments, human MMP-1 promoter/luciferase reporter gene construct (−522/+72; 0.6 kb) was cotransfected with pCMV-jun expression vector into HCC-T cells. Total amount of tranfected DNA was adjusted with pCMV empty vector. Transfection efficiency was normalized by using pRL-CMV vector (Promega) as an internal control. Five hours after transfection, the cells were treated with 15% glycerol for 105 sec, then incubated for 48 hr. Firefly and Renilla luciferase activities were measured using Dual-Luciferase Reporter Assay System (Promega). Cell transfections and luciferase assays were repeated independently more than 3 times, each performed in duplicate.

Western blot analysis for MMP-1 and phosphorylated c-Jun

For the detection of MMP-1, culture media described above were also used for Western blot analysis. The concentration of protein was measured with DC protein assay Kit (Bio-Rad, Richmond, CA). Samples were mixed with 4 × SDS sample buffer (0.25 M Tris-HCl, pH 6.8, 8% SDS, 20% glycerol, 5% β-mercaptoethanol) and equal amount of protein per lane was run on a 10% SDS-PAGE and transferred onto a PVDF membrane (Amersham Biosciences, Buckinghamshire, U.K.). Blots were incubated with anti-MMP-1 antibody (Daiichi Fine Chemical, Takaoka, Japan) at a dilution of 1:1,000 for 2 hr, followed by incubation with rabbit antimouse IgG second antibody (Dako, Glostrup, Denmark) at a dilution of 1:2,000 for 1 hr at room temperature. The proteins were visualized by chemiluminescence by using ECL Plus detection kit (Amersham Biosciences) according to the manufacturer's instruction. For the detection of phosphorylated c-Jun, PhosphoPlus c-Jun (Ser63) II and c-Jun (Ser73) Antibody Kit (Cell Signaling Technology, Beverly, MA) was used according to the manufacturer's instruction. For the confirmation of equal loading, blots were reprobed with anti-β-actin antibody (Sigma Chemical).

In vitrokinase assay

Phosphorylated JNK and total-JNK were detected with Fast Activated Cell-Based ELISA (FACE) JNK Kit (Active Motif North America, Carlsbag, CA) following manufacturer's instructions. Briefly, about 1,000 cells per well were seeded in 2 96-well plates as replicates and incubated for 48 hr and the cells were fixed. One plate was treated with the antiphospho-JNK antibody, while the other plate was treated with anti-JNK antibody. The relative number of cells in each well was then determined through use of the Crystal Violet reagent. Once the phospho-JNK and total JNK signals were normalized for cell number, a comparison of the ratio of phosphorylated JNK to total JNK for each of the cell growth conditions was determined.

Statistical analysis

The data were expressed as mean ± standard deviation. Statistical analysis was performed using Mann-Whitney test. p-values less than 0.05 were considered statistically significant.

RESULTS

MMP-1 expression and cell marker characteristics in HCC cell lines

Among the HCC cell lines examined, definite expression of MMP-1, MMP-2 and MMP-9 mRNAs was found only in HLE cells, while these 3 MMPs were not detected in HCC-M and Huh-7 cells (Fig. 1a). HCC-T and PLC/PRF/5 cells showed a slight positive transcription of MMP-2. TIMP-1 mRNA was expressed in all cell lines in the present study (Fig. 1a). Gelatin zymography revealed definite bands at 76, 68 and 53 kDa only in HLE cells, which corresponded to enzymatic activity of MMP-9, MMP-2 and MMP-1, respectively (Fig. 1b). The other 4 cell lines did not show any bands.

Figure 1.

MMPs and TIMP-1 expression in HCC cell lines. (a) MMP-1, MMP-2, MMP-9 and TIMP-1 expressions in 5 HCC cell lines were detected by RT-PCR analysis. Total RNA was extracted from HCC cells and 1 μg of total RNA was used as a template as described in text. The strong bands of MMP-1, MMP-2 and MMP-9 were observed only in HLE cell line. HCC-T and PLC/PRF/5 cells showed a slight positive transcription of MMP-2. The expression of TIMP-1 was seen in every cell line. G3PDH (bottom) was used as an internal control. (b) Gelatin zymography of conditioned media from cultured 5 HCC cell lines. Semiconfluent HCC cells grown in 60 mm tissue culture plates were replaced with serum-free media and cells were cultured for a further 48 hr. Bands of negative staining indicated zones of enzyme activities. Gelatinolytic activity of MMP-1, MMP-2 and MMP-9, which corresponded to 53, 68 and 76 kDa bands, respectively, were seen only in HLE cells. (c) Cell marker characteristics in HCC cell lines. The expressions of albumin, HNF-4 and α-fetoprotein (AFP) were also assayed by RT-PCR analysis. The distinct albumin and HNF-4 mRNA was observed in PLC/PRF/5 and Huh-7 cells. The strong band of AFP was found in Huh-7 cells.

Five cell lines were established from patients with different differentiation stages of characteristics. Both PLC/PRF/5 and Huh-7 cells showed positive mRNAs of both albumin and HNF-4, while HCC-M, HCC-T and HLE cells did not show these mRNAs (Fig. 1c). These results are coincident with the original characteristics of those cell lines, indicating PLC/PRF/5 and Huh-7 cells as well-differentiated HCC cells and HCC-M, HCC-T and HLE cells as less differentiated HCC cells.

Analysis of −1607 nucleotide polymorphism in MMP-1 promoter region

A single nucleotide polymorphism at −1607 nucleotide in the MMP-1 promoter region creates a PEA-3-binding site (5′-GGAA-3′) as shown in Figure 2 and affects the transcriptional level of MMP-1.28 HCC-M, HCC-T and Huh-7 cells possessed 2G/2G genotype, while HLE cells showed 1G/2G genotype (Table II). There was no correlation between the genotypes and MMP-1 expression levels, indicating that another mechanism may regulate gene expression of MMP-1 in HCC cells.

Figure 2.

Schematic representation of human MMP-1 promoter constructs. Three convenient restriction sites, HindIII, BglII and KpnI, were used to generate the 4.4, 1.2 and 0.6 kb promoter constructs, respectively, which were transfected into HLE and HCC-T cells to determine the promoter activity. Three AP-1 sites, 2 PEA-3 sites and TATA box are included in the 0.6 kb promoter construct. The “Y” indicates functional polymorphism in MMP-1 promoter. The bent arrow represents the beginning of transcription. Modified from Benbow and Brinckerhoff.18

Table II. Relationshionship Between Polymorphism in MMP-1 Promoter and Expression
GenotypeHCC-M (2G/2G)HCC-T (2G/2G)HLE (1G/2G)PLC/PRF/5 (1G/2G)Huh-7 (2G/2G)
MMP-1 expression by RT-PCR+
Enzymatic activity of MMP-1 by zymography+

Transcriptional activity of MMP-1 promoter in HLE and HCC-T cells

In order to clarify the contribution of transcriptional factors in HLE cells, 4.4 kb MMP-1 promoter construct, which covered the entire promoter region, was transfected into HLE and HCC-T cells. The MMP-1 promoter activity of HLE cells was much higher than that of HCC-T cells (Fig. 3a). These data indicate that some transcriptional factors were involved in constitutive MMP-1 expression in HLE cells. Then, to determine the region responsible for the difference in the promoter activity between HLE and HCC-T cells, 2 5′ deletion constructs (1.2 or 0.6 kb) were prepared and transfected into HLE and HCC-T cells. The promoter activity in HLE cells was higher than that in HCC-T cells (about 4- or 5-fold) not only when transfected with the 1.2 kb construct but also with the minimal 0.6 kb construct (Fig. 3b). Since 3 AP-1 sites were present within the 0.6 kb promoter construct (Fig. 2), transcription factor AP-1 may be responsible for MMP-1 expression in HLE cells.

Figure 3.

Basal transcriptional activity of MMP-1 promoter in HLE and HCC-T cells. As described in text, HLE and HCC-T cell lines were transfected with human MMP-1 promoter/luciferase reporter gene construct (4.4 kb; a) and 2 5′ deletion constructs (1.2 or 0.6 kb; b) together with pRL-CMV. After transfection, cells were incubated for 48 hr. Relative luciferase activities (mean ± SD) were normalized for pRL-CMV activity and calculated as percent of the HLE promoter activity, which was transfected with 1.2 kb construct. Transfections and assays were performed independently 4 to 6 times, each run in duplicate.

Different c-jun and c-fos expression in HCC cells

Next we examined the gene expression of c-jun and c-fos in HLE and HCC-T cells to examine the contribution of AP-1 protein to MMP-1 expression in HLE cells. The definite band of c-jun was seen in HLE cells, while no band was detected in HCC-T cells (Fig. 4). c-fos gene expression was not noted in these cell lines. In order to determine the contributions of c-Jun to MMP-1 promoter activity, we transfected c-Jun expression vector (pCMV-jun vector) into HCC-T cells and measured the MMP-1 promoter activity. When c-Jun was overexpressed in HCC-T cells, 3.3-fold transcriptional activation of MMP-1 promoter was detected (Fig. 5).

Figure 4.

c-jun and c-fos expression in HLE and HCC-T cells. c-jun and c-fos expression was detected with RT-PCR analysis. HLE cells, but not HCC-T cells, expressed c-jun mRNA without exogenous stimulation. c-fos mRNA was not observed in either HLE or HCC-T cells.

Figure 5.

Effect of c-Jun overexpression on HCC-T promoter activity. Human MMP-1 promoter/luciferase reporter gene construct (0.6 kb; 2.5 μg) was cotransfected with pCMV-jun vector (2.5 μg) in HCC-T cells and the effect of c-Jun on the activation of MMP-1 transcription was examined. Total amount of transfected DNA was adjusted with pCMV (vehicle) to 7.5 μg. After transfection, cells were incubated for 48 hr. Relative luciferase activities (mean ± SD) were shown after normalization for pRL-CMV activity. Transfections and assays were performed independently 4 times, each run in duplicate.

JNK activity and c-Jun phosphorylation in HCC cells

To confirm the constitutive activation of JNK in HLE cells, phosphorylated and total JNK were measured with the FACE Kit. The amount of phosphorylated JNK protein in HLE cells was more than twice as much as that in HCC-T cells (Fig. 6), though the level of total JNK was almost the same in the 2 cell lines.

Figure 6.

Measurement of phosphorylated and total JNK. HLE and HCC-T cells were cultured in 96-well plates for 48 hr and the cells were fixed. Total and phospho-JNK were each assayed in triplicate using antiphosphorylated and anti-JNK antibodies from the FACE JNK Kit. Data were plotted (mean ± SD) after correction for cell number (performed through use of Crystal Violet). Note that the level of total JNK was almost the same with HLE and HCC-T cells.

Then, phosphorylated c-Jun was analyzed by Western blot analysis. Phosphorylated c-Jun was detected with both phospho-c-Jun antibody (Ser63) and phospho-c-Jun antibody (Ser73) only in HLE cells, but not in HCC-T cells (Fig. 7). These data suggested that the JNK pathway is constitutively activated in HLE cells, resulting in the high MMP-1 expression in the cells.

Figure 7.

Western blot analysis for phosphorylated c-Jun. Proteins were extracted from HLE and HCC-T cells and assessed by Western blot analysis using 2 antibodies to phosphorylated c-Jun (Ser63 and Ser73). In HLE cells, c-Jun was phosphorylated at Ser63 and Ser73. Blots were reprobed for β-actin to confirm the equal protein loading.

Effects of MAP kinase inhibitors on MMP-1 promoter activity

HLE cells had the ability to produce MMP-1 without exogenous stimuli and expressed c-jun (Fig. 4). As c-Jun activation is achieved by MAP kinase families, we used 3 MAP kinase inhibitors to determine the responsible enzyme of MAP kinase families for constitutive MMP-1 gene expression in HLE cell line. It was clearly noted that JNK inhibitor SP600125 (50 μM) reduced the promoter activity of 0.6 kb construct to approximately 40% (Fig. 8a), while either MEK/ERK inhibitor PD98059 (20 and 50 μM) or p38 inhibitor SB203580 (5 μM and 20 μM) did not reduce the promoter activity of MMP-1 (Fig. 8b and c), further indicating that c-Jun activation through JNK pathway may participate in constitutive MMP-1 expression in HLE cells.

Figure 8.

Effect of MAP kinase inhibitors on MMP-1 promoter activity. Inhibition of MMP-1 promoter activities by 3 MAP kinase inhibitors was investigated in constitutively MMP-1-expressing HLE cells. Treatment of HLE cells with SP600125 (a) showed a significantly decreased activity in a dose-dependent fashion, but treatment with PD98059 (b) or SB203580 (c) did not reduce the MMP-1 promoter activity. Details were described in text. Before transfection, HLE cells were cultured for 24 hr in the presence of SP600125 (20 and 50 μM; JNK inhibitor), PD98059 (20 and 50 μM; MEK/ERK inhibitor), or SB203580 (5 and 20 μM; p38 inhibitor). After transfection with human MMP-1 promoter/luciferase reporter gene construct (0.6 kb) together with pRL-CMV, HLE cells were incubated for 48 hr in the presence of the same concentrations of each inhibitor. Transfections and assays were performed independently 4 to 6 times, each run in duplicate. Statistical significance was defined as p < 0.01 (double asterisk).

Effect of JNK inhibitor on MMP-1 protein expression and gelatinolytic activity

To examine the inhibitory effect of SP600125 on MMP-1 expression directly, we performed Western blot analysis and zymography after treatment of HLE cells with SP600125. Both MMP-1 protein expression and gelatinolytic activity were actually reduced 48 hr after the addition of SP600125 (Fig. 9). In contract, the effect of SP600125 on MMP-2 expression was less effective. This result may reflect the number of AP-1-binding sites among the promoter regions of MMP-1 and MMP-2.

Figure 9.

Effect of JNK inhibitor on MMP-1 expression and gelatinolytic activity in HLE cells. HLE cells were cultured in serum-free media for 48 hr in the presence of 50 μM JNK inhibitor SP600125 and its effects on the expression of MMP-1 was analyzed by gelatin zymography (a) and Western blot analysis (b). Representative films are shown from 3 independent experiments. The enzymatic activity and protein level of MMP-1 were reduced to 62.7% ± 9.8% (p < 0.05) and 37.4% ± 5.8% (p < 0.01), respectively. The enzymatic activity of MMP-2 was slightly reduced to 92% ± 8.2%, but the statistical significance was not detected.

DISCUSSION

Clarification of the regulatory mechanism of MMP-1 gene expression in HCC cells is quite important because MMP-1 expression was observed only in an early stage of HCC5 and this phenomenon is accountable to the histologic feature suggesting stromal invasion at the early stage of HCC. In advanced stage of HCC, tumor is encapsulated and does not express MMP-1. In general, HCC tumor cells show very slight atypia in the early stage and they are sometimes indistinguishable from normal hepatocytes. As the tumor grows, dedifferentiation occurs in HCC cells, where well-differentiated cancer cells are replaced by less differentiated cancer cells.6, 29 Since HCC usually arises from a rigid cirrhotic liver in which interstitial collagen is deposited predominantly,8, 10, 30 tumor cells have to dissolve surrounding fibrous tissue for their expansion by producing MMP-1.

First we examined the expression of MMP-1 in cultured human HCC cell lines, which were derived from various stages of differentiation, and found that MMP-1 gene and protein expressions as well as enzymatic activity were observed only in HLE cells without any stimulation. It is true that HLE cells do not produce albumin or α-fetoprotein, but they may lie in a process of dedifferentiation of HCC cells and be considered as special cells to maintain early property of HCC cells. In any event, investigation of the mechanism of MMP-1 gene expression in HLE cells should provide precious clue to clarify the regulatory mechanism of MMP-1 expression in HCC. We thus used HLE cells and nonproducing HCC-T cells as a control for further analysis to explore the regulatory mechanisms of MMP-1 expression.

Transient transfection assay with the 0.6 kb construct (−522/+72) showed about 5-fold higher activity in HLE cells than that in HCC-T cells. Within this promoter region, there are 3 AP-1 sites at −436, −181 and −72 bp (Genbank accession number AF023338), 1 PEA-3 site (polyomavirus enhancer A-binding protein-3), 1 reversed PEA-3 site, CACCC box, TTCA motif and TATA box.27 We could not detect any mutations in these transcription factor-binding elements in HCC cell line. Moreover, RT-PCR analysis showed that the expression of c-jun, but not c-fos, was detected without exogenous stimulation in HLE cells. Therefore, we hypothesized that in HLE cells, MMP-1 transcription was constitutively activated through the c-Jun-binding to the AP-1-binding sites. The AP-1 transcription factor itself can be formed by either the dimerization of Jun family members or the formation of Jun/Fos heterodimers.31, 32, 33 Increased MMP-1 promoter activity observed in HCC-T cells tranfected with c-jun expression vector indicates that only c-Jun is enough for MMP-1 expression via AP-1-binding sites.

JNKs phosphorylate specific sites (ser63 and ser73) of c-Jun and enhance the transcriptional activity of AP-1, whose phosphorylation is induced after exposure to ultraviolet irradiation, growth factors, or cytokines.34, 35 To confirm activation of the JNK pathway in HLE cells, we next examined phosphorylation of JNK and c-Jun with Western blot analysis. Although we could not find any difference in the amount of total JNK between HLE and HCC-T cells, phosphorylated JNK in HLE cells was twice as much as that in HCC-T. Moreover, a strong band of phosphorylated c-Jun was detected in HLE cells, while no band or very faint band, if any, was found in HCC-T cells.

Transient transfection assays with 3 independent MAP kinase inhibitors36, 37 revealed that JNK- specific inhibitor SP600125 (50 μM) reduced the promoter activity of MMP-1 to approximately 40% in HLE cell line, while neither MEK/ERK-specific inhibitor PD98059 (20 and 50 μM) nor p38-specific inhibitor SB203580 (5 and 20 μM) reduced the promoter activity of MMP-1. SP600125 also reduced the gelatinolytic activity and MMP-1 protein expression in HLE cells. Taken together, JNK pathway is always activated in HLE cells, which results in constitutive expression of MMP-1.

Benbow et al.38 reported that in melanoma cells, constitutive MMP-1 expression in the absence of the 2G single nucleotide polymorphism was mediated by p38 and ERK1/2 MAP kinases. Our data also support that, in the absence of 2G allele, HLE cells utilize the alternative activation mechanism to achieve high levels of MMP-1 expression.

Eferl et al.14 have demonstrated that the requirement for c-Jun was restricted to the early stages of tumor (HCC) development by antagonizing p53 activity, resulting in suppression of apoptosis. JNK activation may be required for the early stages of HCC cells, because the activation leads the cells to be more invasive (through MMP-1 activation) and proliferative (via inactivation of p53). This activation seems to explain our previous findings showing that MMP-1 expression was restricted to the early stage of HCC. In addition, it is also reported that MMP-1 itself has an ability to induce hepatocyte proliferation.13 Thus, these findings indicate the significance of not only c-Jun activation but also MMP-1 expression for the invasion and proliferation of HCC cells.

In summary, the present study indeed confirms the relationship between the JNK activation and MMP-1 expression in HCC cells and indicates the new aspect of c-Jun activation in the early stage of HCC development. Further investigations to clarify the difference in the regulatory mechanism of MMP-1 gene expression between regenerating hepatocyte and HCC cells will help us to reverse liver cirrhosis and to prevent HCC development by modifying MMP-1 expression.

Acknowledgements

The authors thank Mr. Hideo Tsukamoto for his excellent technical support.

Ancillary