Downregulation of ceramide synthase-6 during epithelial-to-mesenchymal transition reduces plasma membrane fluidity and cancer cell motility.

Epithelial-to-mesenchymal transition (EMT) promotes cell motility, which is important for the metastasis of malignant cells, and blocks CD95-mediated apoptotic signaling triggered by immune cells and chemotherapeutic regimens. CD95L, the cognate ligand of CD95, can be cleaved by metalloproteases and released as a soluble molecule (cl-CD95L). Unlike transmembrane CD95L, cl-CD95L does not induce apoptosis but triggers cell motility. Electron paramagnetic resonance was used to show that EMT and cl-CD95L treatment both led to augmentation of plasma membrane fluidity that was instrumental in inducing cell migration. Compaction of the plasma membrane is modulated, among other factors, by the ratio of certain lipids such as sphingolipids in the membrane. An integrative analysis of gene expression in NCI tumor cell lines revealed that expression of ceramide synthase-6 (CerS6) decreased during EMT. Furthermore, pharmacological and genetic approaches established that modulation of CerS6 expression/activity in cancer cells altered the level of C16-ceramide, which in turn influenced plasma membrane fluidity and cell motility. Therefore, this study identifies CerS6 as a novel EMT-regulated gene that has a pivotal role in the regulation of cell migration.

Increased extracellular matrix remodeling is associated with tumor progression in human hepatocellular carcinomas.

Matrix metalloproteinase-2 (MMP2) is a key enzyme in the process of extracellular matrix remodeling involved in tumor invasion and metastasis. The activation of MMP2 involves interplay with the membrane type-matrix metalloproteinase-1 (MT1-MMP) and the tissue inhibitor of metalloproteinase-2 (TIMP2). In vitro, activated hepatic stellate cells are a main source of MMP2 and collagen I induces MMP2 activation. The steady-state mRNA levels of MMP2, MT1-MMP, TIMP2, collagen I, collagen IV, and laminin gamma1 were compared with MMP2 activity in 55 hepatocellular carcinomas, 47 matching nontumor biopsies and 19 histologically normal livers. In hepatocellular carcinomas, increased collagen I mRNA levels were strongly associated with those of MMP2 (Spearman R =.74, P <.001), MT1-MMP (R =.65, P <.001) and TIMP2 (R = 0.61, P <.001). MMP2 activity was correlated with the mRNA expression of collagen I (R =.45 P <.01), collagen IV (R =.40, P <.01) and laminin gamma1 (R =.33, P <.05). Unlike collagen IV and laminin gamma1 mRNAs, MMP2, MT1-MMP, TIMP2, collagen I mRNA levels were increased in nonencapsulated compared with encapsulated tumors (P <.05). In addition, MMP2 activity was fourfold higher (P <.01) in tumors arising in cirrhotic livers than in those arising in noncirrhotic livers. Moreover, tumor recurrence was associated with 4.6- and 2.8-fold (P <.05) higher collagen I and MMP2 mRNA levels, respectively, in hepatocellular carcinomas arising in cirrhotic livers. Thus, a high extracellular matrix remodeling favors tumor progression in hepatocellular carcinomas.

Repeated endotoxin exposure induces interstitial fibrosis associated with enhanced gelatinase (MMP-2 and MMP-9) activity.

BACKGROUND: Dysregulation of matrix metalloproteinases (MMPs) has been implicated in lung injury associated with inflammatory disorders and several lung diseases such as pulmonary fibrosis. OBJECTIVE: We studied a murine model of lipopolysaccharide (LPS)-induced chronic inflammation in order to analyse the relationship between MMP activity in bronchoalveolar lavage fluid and collagen deposition in lung tissue. BP2 mice were exposed to repeated aerosols of LPS of E. coli for 8 months. RESULTS: The inflammatory reaction induced by LPS increased throughout the time of exposure and was associated after 10 weeks with collagen deposition in the alveolar walls. Meantime, we observed in BAL fluid from LPS-exposed mice an early induction of MMP-9 correlated with neutrophil recruitment. MMP-2 increased during the early inflammatory phase, and also during the development of the fibrotic phase. CONCLUSION: Repeated exposure of mice to an aerosol of LPS can lead to pulmonary interstitial fibrosis and MMPs seem to be associated with this process.

Tumor hepatocytes and basement membrane-Producing cells specifically express two different forms of the endostatin precursor, collagen XVIII, in human liver cancers.

Endostatin is an endogenous inhibitor of angiogenesis and tumor growth in mice, which may be generated by proteolytic cleavage of collagen XVIII. In normal tissues, 2 variants of the endostatin precursor, namely the SHORT and LONG forms, regulate tissue specificity. We analyzed 53 human liver biopsies (18 hepatocellular carcinomas, 16 metastases of colorectal cancer, 3 cholangiocarcinomas, and 16 controls) by RNA dot blots, double-labeling immunohistochemistry, and in situ hybridization, using common and variant-specific probes. Tumor hepatocytes expressed the LONG form, whereas cholangiocarcinoma cells expressed the SHORT form, which was deposited in tumor basement membranes. Metastatic colorectal carcinoma cells did not express collagen XVIII. In the stromal compartment of primary and metastatic cancers, myofibroblasts and vascular endothelial cells expressed the SHORT form. Both basement membrane components, collagen IV and the SHORT collagen XVIII form, were codistributed and their mRNA levels strongly correlated (R =.75, P <.001). In addition, freshly isolated human hepatocytes expressed the LONG form and culture-activated stellate cells the SHORT form. Moreover, the full-length LONG form is a plasma protein. Thus, the LONG form is a hepatocyte-specific variant, and the SHORT form is a major component of the tumor extracellular matrix in primary and metastatic liver cancers. In the clinical context, the global expression of the endogenous endostatin precursor, collagen XVIII, in liver cancer results from the combined expression profiles of tumor cells, stromal cells, and nontumor hepatocytes at the advancing edge of the tumor, particular to each type of cancer.

Mechanism in the sequential control of cell morphology and S phase entry by epidermal growth factor involves distinct MEK/ERK activations.

Cell shape plays a role in cell growth, differentiation, and death. Herein, we used the hepatocyte, a normal, highly differentiated cell characterized by a long G1 phase, to understand the mechanisms that link cell shape to growth. First, evidence was provided that the mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) cascade is a key transduction pathway controlling the hepatocyte morphology. MEK2/ERK2 activation in early G1 phase did not lead to cell proliferation but induced cell shape spreading and demonstration was provided that this MAPK-dependent spreading was required for reaching G1/S transition and DNA replication. Moreover, epidermal growth factor (EGF) was found to control this morphogenic signal in addition to its mitogenic effect. Thus, blockade of cell spreading by cytochalasin D or PD98059 treatment resulted in inhibition of EGF-dependent DNA replication. Our data led us to assess the first third of G1, is exclusively devoted to the growth factor-dependent morphogenic events, whereas the mitogenic signal occurred at only approximately mid-G1 phase. Moreover, these two growth factor-related sequential signaling events involved successively activation of MEK2-ERK2 and then MEK1/2-ERK1/2 isoforms. In addition, we demonstrated that inhibition of extracellular matrix receptor, such as integrin beta1 subunit, leads to cell arrest in G1, whereas EGF was found to up-regulated integrin beta1 and fibronectin in a MEK-ERK-dependent manner. This process in relation to cytoskeletal reorganization could induce hepatocyte spreading, making them permissive for DNA replication. Our results provide new insight into the mechanisms by which a growth factor can temporally control dual morphogenic and mitogenic signals during the G1 phase.

Tumor progression is associated with a significant decrease in the expression of the endostatin precursor collagen XVIII in human hepatocellular carcinomas.

Endostatin inhibits angiogenesis and tumor growth in mice. The role of its endogenous precursor collagen XVIII in human cancer is unknown. In normal tissues, two variants of collagen XVIII, namely, the short and long forms regulate tissue specificity, the long form being almost exclusively expressed by hepatocytes in the liver. We analyzed RNA arrays from 57 hepatocellular carcinomas (HCCs) with common and variant-specific probes and investigated the relationships between collagen XVIII expression and angiogenesis by measuring the CD34-positive microvessel density. Low collagen XVIII expression by tumor hepatocytes was associated with large tumor size (r, -0.63; P < 0.001) and replacement of trabeculae with pseudoglandular-solid architecture (chi2, 28; P < 0.001), which indicate tumor progression. Tumors expressing the highest collagen XVIII levels were smaller and had lower microvessel density (P = 0.01) than those expressing moderate levels; and HCCs with the lowest collagen XVIII levels approached a plateau of microvessel density, which indicated that a decrease in collagen XVIII expression is associated with angiogenesis in primary liver cancer. HCCs recurring within 2 years of resection showed 2.2-fold lower collagen XVIII mRNA than nonrecurring ones (P = 0.02). The findings relied on the hepatocyte-specific long form. Thus, the endogenous expression of the endostatin precursor decreases along with tumor progression in HCCs.

The promoter of the long variant of collagen XVIII, the precursor of endostatin, contains liver-specific regulatory elements.

The endostatin precursor collagen XVIII is expressed at high levels in human livers, the main source being hepatocytes. We have studied the regulatory elements in the promoter 2 of the Col18a1 gene that directs the transcription of the NC1-517 variant of collagen alpha1(XVIII), which is the main form expressed in the liver. The 5'-flanking region of Col18a1 gene was cloned, and a series of 5'-deletions from -3286 bp to +285 bp were linked to the luciferase reporter gene. Transfection experiments in HepG2 cells allowed to identify a silencer-like element containing putative HNF1 and HNF3 sites and activator elements containing stretches of GC-rich sequences. Another putative HNF3 site in close apposition to a NF1/CTF site was localized upstream of the silencer-like element. Cotransfection experiments showed that the Col18a1 promoter 2 was transactivated by Sp1 and HNF3alpha. Gel-shift analyses showed that HNF3, NF1/CTF, and Sp1-like sites specifically recognized nuclear factors. Super-shift experiments indicated that HNF3beta was the major form of HNF3 interacting with the HNF3/NF1 site. The well-differentiated hepatoma cell line mhATFS315 transfected with a truncated form of HNF3beta, which competitively blocks HNF3 transactivating activity, expressed the Col18a1gene at a very low level. Taken together, these data strongly suggest that Col18a1 is a liver-specific gene. Furthermore, gel-shift analyses performed with nuclear factors prepared from well-differentiated hepatocellular carcinomas showed increased HNF3/NF1 binding activity compared with normal livers. Consequently, the precursor of endostatin might be differently expressed according to the differentiated and/or transformed state of hepatocytes.

Assessing matrix metalloproteinase expression and activity in hepatocellular carcinomas.

The matrix metalloproteinases (MMPs) constitute a large family of zincand calcium-dependent endopeptidases that cleave extracellular matrix components (1). Hence, MMPs are classified according to their substrate specificities: interstitial collagenases, stromelysins, gelatinases, membrane-type matrix metalloproteinases (MT-MMPs), and elastase (Table 1). The regulation of MMP activity involves gene expression, proteolytic processing of the propeptides to active forms, and inhibition by specific tissue inhibitors of matrix metalloproteinase (TIMPs). MMPs are involved in situations that require extracellular matrix remodeling, including wound healing, development, inflammation, fibrosis angiogenesis, and tumor invasion (2-5). Several complementary methods have provided an insightful description of the expression levels of MMPs and their pathological correlates. These include immunohistochemistry and Northern, Western, and dot blots. Additionally, the activity of MMPs is evaluated by gel substrate analysis. This approach has demonstrated that an increase in the expression of MMP2 (6,7), MT1-MMP (7), TIMP1, and TIMP2 (8-10) is associated with liver fibrosis. Similarly, in hepatocellular carcinomas, a high expression of MMP2, MMP9, MT1-MMP, and matrilysin is related to tumor aggressiveness (11-14). Consistently, by gel substrate analysis, MMP2 activity is increased in primary and secondary liver cancers (13,15,16). By in situ hybridization, the sources of.

MMP2 activation by collagen I and concanavalin A in cultured human hepatic stellate cells.

Fibrosis occurs in most chronic liver injuries and results from changes in the balance between synthesis and degradation of extracellular matrix components. In fibrotic livers, there is a markedly increased activity of matrix metalloproteinase 2 (MMP2), a major enzyme involved in extracellular matrix remodeling. We have previously shown that hepatic stellate cells secrete latent MMP2 and that MMP2 activation occurs in coculture of hepatic stellate cells and hepatocytes concomitantly with matrix deposition. In the present work we investigated the effects of various extracellular matrix components and concanavalin A, an inducer of immune-mediated liver injuries, on MMP2 activation in cultured human hepatic stellate cells. Collagen I induced a dose-dependent MMP2 activation, which was not blocked by both actinomycin and cycloheximide. Collagen VI, laminin, and a reconstituted basement membrane (matrigel) were ineffective in inducing activation. Specific antibodies against the subunits of alpha2beta1 integrins, the major collagen I receptor, induced partial inhibition of MMP2 activation. Treatment of cells with concanavalin A resulted in a marked activation of MMP2 that correlated with the proteolytic processing of MT1-MMP, the MMP2 activator, from a Mr=60 kd toward a Mr=43 kd polypeptide. Actinomycin and cycloheximide inhibited the MMP2 activation induced by concanavalin A. Recombinant tissue inhibitor of metalloproteinase-2 and the MMP inhibitor BB-3103, but not PMSF, blocked MMP2 activation induced by collagen I or concanavalin A, and MT1-MMP processing to its Mr-43 kd form. These results suggest that the accumulation of collagen I may specifically contribute to the remodeling of extracellular matrix in fibrotic livers by inducing MMP2 activation.

Comparative effects of betamethasone, cyclosporin and nedocromil sodium in acute pulmonary inflammation and metalloproteinase activities in bronchoalveolar lavage fluid from mice exposed to lipopolysaccharide.

Matrix metalloproteinases (MMPs) are particularly potent in degrading basement membrane collagen associated with lung injury in inflammatory processes. We have investigated the effects of betamethasone, cyclosporin, and nedocromil on MMP2 and MMP9 activities, on TNF-alpha and IL-10 release, as well as on the recruitment of inflammatory cells in the bronchoalveolar lavage (BAL) fluid after aerosol administration of lipopolysaccharide (LPS) in mice. When mice were pretreated with betamethasone (5 mg/kg, po), MMP2 and MMP9 activities, TNF-alpha in BAL fluids, and the enhanced neutrophil number of LPS-exposed mice were reduced, whereas the level of IL-10 was increased. Pretreatment of mice with cyclosporin (10 mg/kg, po) did not significantly reduce MMP activities, but cyclosporin inhibited neutrophil recruitment, inhibited increase TNF- alpha and inhibited IL-10 decrease. Nedocromil sodium (30 mg/kg, ip) had no influence on the LPS-induced MMP activities, on neutrophil recruitment, or on IL-10 level, but this drug elicited a significant inhibition of TNF- alpha level. These results showed that treatment with the antiinflammatory drugs cyclosporin and nedocromil sodium did not lead to reduction of MMP release. However, since betamethasone reduced the LPS-induced pulmonary inflammation and production of MMPs, these results suggest that corticosteroids may decrease tissue remodelling associated with acute lung injury.

Differential expression and origin of membrane-type 1 and 2 matrix metalloproteinases (MT-MMPs) in association with MMP2 activation in injured human livers.

Matrix metalloproteinase-2 (MMP2) activation is associated with basement membrane remodeling that occurs in injured tissues and during tumor invasion. The newly described membrane-type MMPs (MT-MMPs) form a family of potential MMP2 activators. We investigated the localization and steady-state levels of MT1-MMP and MT2-MMP mRNA, compared with those of MMP2 and tissue inhibitor of MMP-2 in 22 hepatocellular carcinomas, 12 liver metastases from colonic adenocarcinomas, 13 nontumoral samples from livers with metastases, 10 benign tumors, and 6 normal livers. MMP2 activation was analyzed by zymography in the same series. The expression of MT1-MMP mRNA and the activation of MMP-2 were increased in hepatocellular carcinomas, metastases, and cholestatic nontumoral samples. MT2-MMP mRNA was rather stable in the different groups. MT1-MMP mRNA levels, but not MT2-MMP mRNA, correlated with MMP-2 and tissue inhibitor of MMP-2 mRNA levels and with MMP2 activation. In situ hybridization showed that MT1-MMP mRNA was expressed in stromal cells, and MT2-MMP mRNA was principally located in both hepatocytes and biliary epithelial cells. Consistently, freshly isolated hepatocytes expressed only MT2-MMP mRNA, and culture-activated hepatic stellate cells showed high levels of MT1-MMP mRNA. These results indicate that in injured livers, MMP2 activation is related to a coordinated high expression of MMP2, tissue inhibitor of MMP-2, and MT1-MMP. Furthermore, the finding of a preferential expression of MT2-MMP in hepatocytes, together with our previous demonstration that the activation of stellate cell-derived MMP2 in co-culture requires interactions with hepatocytes (Am J Pathol 1997, 150:51-58), suggests that parenchymal cells might play a pivotal role in the MMP2 activation process.

Collagen XVIII is localized in sinusoids and basement membrane zones and expressed by hepatocytes and activated stellate cells in fibrotic human liver.

Type XVIII collagen is a recently discovered nonfibrillar collagen associated with basement membranes in mice and expressed at high levels in human liver. We studied the origin, distribution, and RNA levels of type XVIII collagen in normal and fibrotic human livers by in situ hybridization, immunohistochemistry, and Northern and dot blots and compared procollagen alpha1(XVIII) RNA levels with those of procollagen alpha1(IV) and laminin gamma1, the two major components of liver basement membranes. In normal liver, type XVIII collagen was heavily deposited in perisinusoidal spaces and basement membrane zones. The major source of type XVIII collagen was hepatocytes and, to a lesser extent, endothelial, biliary epithelial, and vascular smooth muscle cells and peripheral nerves. In cirrhosis, type XVIII collagen formed a thick deposit along capillarized sinusoids. Grain counts after in situ hybridization showed myofibroblasts to increase their expression 13-fold in active and twofold in quiescent fibrosis, whereas hepatocytes increased their expression only twofold in both active and quiescent fibrosis. Activated stellate cells in vitro expressed type XVIII collagen at high levels. These data indicate that type XVIII collagen is a component of the perisinusoidal space and is associated with basement membrane remodeling. Hepatocytes and activated stellate cells are important sources of type XVIII collagen in normal and fibrotic liver respectively, which suggests tissue-specific regulation of its expression.

Laminin isoforms in non-tumoral and tumoral human livers. Expression of alpha1, alpha2, beta1, beta2 and gamma1 chain mRNA and an alpha chain homologous to the alpha2 chain.

BACKGROUND/AIMS: Laminins, the major non-collagenous basement membrane components, are involved in various biological processes. Laminin isoforms have never been characterized in human livers. The expression of five laminin mRNA was investigated in livers with or without cancer and in hepatoma cells and, by comparison, in both rat hepatoma and hepatic stellate cells. METHODS: Laminin alpha1, alpha2, beta1, beta2 and gamma1 mRNA was detected by northern blot and/or RT-PCR in livers without chronic disease (n=5), in both tumoral and non-tumoral areas of livers with hepatocellular carcinomas (n=13) or metastases (n=18), in human HBGC2 and rat Faza-567 hepatoma cell lines, and in 6-day-old rat hepatic stellate cell cultures. RESULTS: Laminin alpha1, alpha2 and beta1 mRNA were found in 25-33% and gamma1 mRNA in 58% of the livers, the signal for laminin beta2 mRNA being faint in all the samples. Laminin alpha2, beta1, beta2 and gamma1 mRNA were expressed in hepatoma and stellate cells. The laminin alpha2 cDNA probe recognized a 3.5 kb mRNA different from the expected 9 kb mRNA. Using degenerated oligonucleotides, RT-PCR products from both rat hepatoma and stellate cells revealed 90% identity with the alpha2 chain sequence. Antibodies against peptide deduced from the conserved C-terminal domain of both alpha1 and alpha2 chains recognized polypeptides corresponding to the degradation products of alpha2 chain in liver extracts and both media and cell layers from hepatoma and stellate cells. In addition, a Mr=130000 polypeptide was revealed by these antibodies in liver extracts and cell layers, which was consistent with the expected size deduced from the 3.5 kb mRNA. CONCLUSIONS: This first report on laminin isoforms in human livers indicates that laminin 1 (alpha1-beta1-gamma1), 2 (alpha2-beta1-gamma1), 3 (alpha1-beta2-gamma1) and 4 (alpha2-beta2-gamma1) mRNA and a polypeptide homologous to the alpha2 isoform, which could correspond to a truncated form of this chain, are usually expressed in non-tumoral and/or tumoral livers.

Sp1-mediated transactivation of LamC1 promoter and coordinated expression of laminin-gamma1 and Sp1 in human hepatocellular carcinomas.

The laminin-gamma1 chain is present in most basement membranes and is involved in various physiological and pathological processes, including carcinogenesis in the liver. We have investigated the role of the transcription factor Sp1 in the activation of the LamC1 gene, which encodes laminin-gamma1, both in hepatocytes and in human hepatocellular carcinomas. DNAse I hypersensitive sites were mapped in the murine LamC1 promoter using early hepatocyte primary cultures in which LamC1 becomes activated. Three hypersensitive sites were found in enhancer-like elements that contain GC-rich regions. Gel-shift analyses showed that specific complexes were resolved using GC-containing oligonucleotides and Faza 567 hepatoma cells, which constitutively express laminin-gamma1 at a high level. Increased GC-binding activity was observed using nuclear extracts from early hepatocyte cultures versus normal liver. Sp1 overexpression in normal hepatocytes transfected with an Sp1 expression vector induced a marked increased of laminin-gamma1 mRNA content and co-transfection of promoter fragments in Drosophila melanogaster SL2 cells demonstrated that Sp1 transactivates LamC1. In human hepatocellular carcinomas, Sp1 and laminin-gamma1 mRNA were simultaneously expressed at high levels, and gel-shift experiments demonstrated a higher GC-binding activity to Sp1 compared with control livers. In situ hybridization indicated that cells exhibiting a high content of laminin-gamma1 mRNA were also strongly positive for Sp1 mRNA, including both cancer cells at the invasion front and stromal cells. These results show that Sp1 is involved in the activation of LamC1 that occurs in human hepatocellular carcinomas.

Activation of the envelope proteins by a metalloproteinase enables attachment and entry of the hepatitis B virus into T-lymphocyte.

Previously, we identified an HBV binding factor (HBV-BF), a 50-kDa serum glycoprotein which interacts with HBV envelope proteins and which is also located in the membrane of normal human hepatocyte (A. Budkowska et al. (1993) J. Virol. 67, 4316). Here we show that HBV-BF is a neutral metalloproteinase which shares substrate specificity and properties with a newly described family of membrane type matrix metalloproteinases. HBV-BF treatment of the HBV resulted in the cleavage of the N-terminal part of the middle HBV envelope protein at the pre-S2(136-141) amino acid sequence VRGLYF/L (containing a single arginine cleavage site). HBV-BF affected the reactivity of the large HBV protein with pre-S1-specific MAbs, probably inducing the conformational change of the pre-S1 domain. The HBV-BF-digested virus remained morphologically intact with unchanged S antigenic determinants. The structural modifications of the viral envelope proteins induced by HBV-BF enabled cell membrane attachment and viral entry into the T-lymphocyte. Both processes were blocked by the metalloproteinase inhibitor 1,10 phenanthroline. Thus, the host-dependent proteolytic activation of the envelope proteins seems to be essential for the HBV entry into the cell. HBV-BF under a membrane bound or a secreted form could be (one of) the molecule(s) responsible for the HBV proteolytic activation.

Overexpression of matrix metalloproteinase-2 and tissue inhibitor of matrix metalloproteinase-2 in liver from patients with gastrointestinal adenocarcinoma and no detectable metastasis.

Degradation of basement membranes is a key step in tumoral invasion, mainly mediated by matrix metalloproteinases (MMPs) and their inhibitors (TIMPs). Since the liver is a main target for metastases from gastrointestinal adenocarcinoma, we have investigated MMP2 and TIMP2 expression by RT-PCR, in situ hybridization and zymography in the liver of patients with gastrointestinal adenocarcinomas and no detectable hepatic metastasis (n = 12), in tumoral and nontumoral liver from patients with hepatic metastasis (n = 9) and in control liver (n = 4). MMP2 and TIMP2 mRNA levels were increased in liver from patients with gastrointestinal adenocarcinomas and no detectable metastasis, compared with those of either control liver (5-fold and 3.2-fold, respectively) or nontumoral areas of liver from patients with metastasis (7.8-fold and 3-fold, respectively). MMP2 and TIMP2 transcripts were located in mesenchymal cells of portal tracts and sinusoids. MMP2 was mainly in its latent form. In liver from patients with hepatic metastasis, the tumoral/nontumoral ratios for MMP2 and TIMP2 mRNA were 6.2 +/- 4 and 1.5 +/- 0.4, respectively. Both transcripts were localized in the stromal cells of liver metastases, and the active form of MMP2 was found only in the tumoral areas. In the matching nontumoral areas the signals for MMP2 and TIMP2 mRNA were restricted to mesenchymal cells in portal tracts and sinusoidal cells. Our data show that liver stromal cells express high levels of MMP2 and TIMP2 in patients with colonic carcinoma without liver metastasis, suggesting the distant induction of these transcripts by the primary tumor.

In situ detection of matrix metalloproteinase-2 (MMP2) and the metalloproteinase inhibitor TIMP2 transcripts in human primary hepatocellular carcinoma and in liver metastasis.

BACKGROUND/AIMS: Metalloproteinase (MMP)-2 and the metalloproteinase inhibitor TIMP2, play a critical role in tumor invasion. We have investigated the cellular sources of MMP2 and TIMP2 in primary and secondary human liver cancers. METHODS: Using in situ hybridization and zymography, we analyzed surgical biopsies from matching pairs of tumoral and non-tumoral liver from six hepatocellular carcinomas and seven liver metastases and from four liver donors. The cellular sources of MMP2 and TIMP2 were further characterized using an anti-alpha-smooth muscle actin antibody on contiguous sections. RESULTS: In hepatocellular carcinoma and liver metastases, in situ hybridization showed that MMP2 and TIMP2 mRNA were expressed by anti-alpha-smooth muscle actin-positive cells at the invasive front. Slender fibroblasts embedded in a denser matrix were MMP2(+)/TIMP2(+)/anti-alpha-smooth muscle actin(+). Intratumor microvessels showed a strong labeling for MMP2 but weak for TIMP2 mRNA. In contrast, the endothelial lining of the central veins was MMP2(+)/TIMP2(+) in non-tumoral areas with signs of blood-flow obstruction. In control livers, MMP2 and TIMP2 mRNA distribution was restricted to fibroblasts and endothelial cells within portal tracts and scattered sinusoidal cells. Direct zymography of samples comprising the invasive front revealed variable amounts of both proMMP2 and its active form in hepatocellular carcinoma, whereas strong bands corresponding to both active and latent forms of MMP2 were detected in liver metastases. CONCLUSIONS: The striking density of MMP2(+)/TIMP2(+)/anti-alphaSM(+) stellate-shaped cells in the perisinusoidal space adjacent to liver tumors suggests that hepatic stellate cells, upon differentiation to myofibroblasts, may contribute to the dissemination of liver metastases through the sinusoidal network.

Activation of matrix metalloproteinase-2 from hepatic stellate cells requires interactions with hepatocytes.

Activation of matrix metalloproteinase (MMP)-2, the 72-kd collagenase IV/gelatinase A, is involved in extracellular matrix remodeling. It has been suggested that a membrane-type MMP (MT-MMP-1) and the tissue inhibitor of metalloproteinase (TIMP)-2 are involved in MMP-2 processing, but the exact mechanism(s) of its activation remains unclear. We have investigated the role of cell-cell cooperation in the activation of pro-MMP-2 in the liver, using pure cultures and co-cultures of hepatocytes and hepatic stellate cells (HSCs). Northern blot analysis and in situ hybridization showed that, in both pure and co-cultures, HSCs, but not hepatocytes, expressed MMP-2, TIMP-2, and MT-MMP-1 mRNA. Zymography analyses revealed the latent form of MMP-2 in medium from 2-day-old pure HSC cultures with higher amounts in medium from hepatocyte/HSC co-cultures. When hepatocytes were added to 10-day-old HSC cultures, the activated form of MMP-2 was detected, concomitantly with the deposition of an abundant extracellular matrix. Incubation of plasma membrane-enriched fractions from hepatocytes with conditioned medium from pure HSC cultures generated the activated species of MMP-2 (62 and 59 kd). Activation of pro-MMP-2 by hepatocyte membranes was inhibited by EDTA, heat, and trypsin but not by serine proteinase inhibitors. These data show that the co-expression of TIMP-2, MMP-2, and MT-MMP-1 by HSCs does not lead to secretion of the activated form of MMP-2. Hepatocytes, which do not express MMP-2, TIMP-2, or MT-MMP-1, induce MMP-2 activation through a plasma membrane-dependent mechanism(s), thus suggesting that cell-cell interactions are involved in this process in vivo.

Expression of laminin gamma 1 cultured hepatocytes involves repeated CTC and GC elements in the LAMC1 promoter.

Laminin gamma 1 chain is present in all basement membranes and is expressed at high levels in various diseases, such as hepatic fibrosis. We have identified cis- and trans-acting elements involved in the regulation of this gene in normal rat liver, as well as in hepatocyte primary cultures and hepatoma cell lines. Northern-blot analyses showed that laminin gamma 1 mRNA was barely detectable in freshly isolated hepatocytes and expressed at high levels in hepatocyte primary cultures, as early as 4 h after liver dissociation. Actinomycin D and cycloheximide treatment in vivo and in vitro indicated that laminin gamma 1 overexpression in cultured hepatocytes was under the control of transcriptional mechanisms. Transfection of deletion mutants of the 5' flanking region of murine LAMC1 gene in hepatoma cells that constitutively express laminin gamma 1 indicated that regulatory elements were located between -594 bp and -94 bp. This segment included GC- and CTC-containing motifs. Gel-shift analyses showed that two complexes were resolved with different affinity for the CTC sequence depending on the location of the GC box. The pattern of complex formation with nuclear factors from freshly isolated and cultured hepatocytes was different from that obtained with total liver and similar to that with hepatoma cells. Southwestern analysis indicated that several polypeptides bound the CTC-rich sequence. Affinity chromatography demonstrated that A M(r) 60,000 polypeptide was a major protein binding to the CTC motif. This polypeptide is probably involved in the transcriptional activation of various proto-oncogenes and extracellular matrix genes that are expressed at high levels in both hepatoma cells and early hepatocyte cultures.