Sequential analysis of multistage hepatocarcinogenesis reveals that miR-100 and PLK1 dysregulation is an early event maintained along tumor progression.

MicroRNAs (miRNAs) have an important role in a wide range of physiological and pathological processes, and their dysregulation has been reported to affect the development and progression of cancers, including hepatocellular carcinoma (HCC). However, in the plethora of dysregulated miRNAs, it is largely unknown which of them have a causative role in the hepatocarcinogenic process. In the present study, we first aimed to determine changes in the expression profile of miRNAs in human HCCs and to compare them with liver tumors generated in a rat model of chemically induced HCC. We found that members of the miR-100 family (miR-100, miR-99a) were downregulated in human HCCs; a similar downregulation was also observed in rat HCCs. Their reduction was paralleled by an increased expression of polo like kinase 1 (PLK1), a target of these miRNAs. The introduction of miR-100 in HCC cells impaired their growth ability and their capability to form colonies in soft agar. Next, we aimed at investigating, in the same animal model, if dysregulation of miR-100 and PLK1 is an early or late event along the multistep process of hepatocarcinogenesis. The obtained results showed that miR-100 downregulation (i) is already evident in very early preneoplastic lesions generated 9 weeks after carcinogenic treatment; (ii) is also observed in adenomas and early HCCs; and (iii) is not simply a marker of proliferating hepatocytes. To our knowledge, this is the first work unveiling the role of a miRNA family along HCC progression.

RNA interference against urokinase in hepatocellular carcinoma xenografts in nude mice.

The serine protease urokinase-type plasminogen activator (u-PA) is overexpressed in hepatocellular carcinoma (HCC) and its expression level is inversely correlated with the patients' survival. The purpose of this study was to examine the effects of vector-based RNA interference (RNAi) of u-PA on the growth of human HCC xenografts in nude mice in order to investigate the role of u-PA in human HCC. Our results showed that the subcutaneous injection of small interfering RNAs (siRNA) u-PA SKHep1C3 stable transfected cells (pS siRNA u-PA) led to a growth delay in xenograft development, compared to those generated from empty vector; the molecular characterization of nodules (carried out by PCR, RT-PCR and immunohistochemical analysis) revealed the presence of plasmid DNA, the u-PA gene expression knockdown, at both mRNA and protein levels, giving evidence of a long-term and target-specific inhibition by vector-based RNAi 11 weeks after cell inoculation. We further studied the effects of u-PA down modulation on extracellular matrix (ECM) proteins evaluating the expression and organization of fibronectin (FN; one of the main ECM proteins). Immunohistochemical and immunofluorescence analysis of FN revealed FN fibrils in pS siRNA u-PA xenografts and in pS siRNA u-PA cells, thus identifying the FN fibril organization as a downstream effect of u-PA knockdown in this system.

Androgen receptor mRNA under-expression in poorly differentiated human hepatocellular carcinoma.

Many studies suggest that hepatocellular carcinoma (HCC) is an androgen-dependent tumor with an incidence five times higher in males, but few data are available on the androgen receptor (AR) mRNA levels in different physiological classes of human liver specimens. In this study 108 human hepatic samples have been analyzed for AR mRNA expression by a comparative RT-PCR assay. These consisted of 35 non-tumoral hepatic samples (3 normal parenchymas, 4 steatosis, 10 hepatitis, 18 cirrhosis), 38 tumoral specimens derived from uninodular and multinodular HCCs and 35 peritumoral hepatic tissues. Normalized AR mRNA levels in tumoral and peritumoral liver tissues spanned from 0 to 146% and from 7 to 125% respectively. Only in a relatively small percentage of HCCs, the levels of expression of AR mRNA were higher than in the corresponding peritumoral tissues (16% of total HCCs). Although extremely variable, the AR mRNA levels were related to histological tumoral differentiation and proved to be lower in the highly dedifferentiated HCCs as compared to the well differentiated ones. Therefore, the evaluation of AR expression in HCC patients might be relevant for the planning of clinical studies on anti-androgen therapies, which might be useful only in the cases in which a high level of AR mRNA is detected, considering the high heterogeneity of AR mRNA levels which characterizes HCC samples. It is likely that the HCCs, expressing low or undetectable levels of AR mRNA, would not benefit by the anti-androgen therapy.

u-PA and c-MET mRNA expression is co-ordinately enhanced while hepatocyte growth factor mRNA is down-regulated in human hepatocellular carcinoma.

Hepatocyte growth factor/scatter factor (HGF/SF) is one of the most important humoral mediators of liver regeneration. It is potentially related to molecular mechanisms of hepatocarcinogenesis via a paracrine system involving its cellular receptor, c-met. In this study, the expression patterns of HGF and c-met were evidenced by multiplex RT-PCR in different specimens of human hepatic tissues (n = 71). A significant increase of c-met mRNA expression was detected in hepatitis (P = 0.001), cirrhosis (P = 0.006), and hepatocellular carcinoma (HCC) tissue (P = 0.003) compared with normal parenchyma and steatosis. HGF mRNA expression was significantly higher only in hepatitis (P = 0.01). Over-expression of c-met mRNA and under-expression of HGF mRNA were detected in the HCCs compared with the corresponding peri-tumoral tissues. Neither HGF nor c-met expression was related to age, sex, tumor size, grading, presence of pseudocapsula, and proliferative activity of the malignant hepatocytes. A significant inverse correlation was found between c-met mRNA expression level and survival (in months) of patients (P = 0.007), as previously shown for urokinase-type plasminogen activator (u-PA) mRNA (P = 0.027). In addition, c-met mRNA expression was strictly associated with u-PA mRNA level in HCC samples (P = 0.001). These data show that a loss of balance concerning HGF, c-met, and u-PA mRNA expression occurs during hepatocarcinogenesis. Particularly, up-regulation of c-met and u-PA mRNA transcription appears to be coordinately regulated, and their levels of expression are inversely correlated with survival; they must therefore play an important role in the development and progression of human HCC and may also be relevant prognostic markers.

Gene response of human skin fibroblasts to urokinase- and tissue-type plasminogen activators.

In a previous work we have reported evidences on the mitogenic activity of urokinase-type and tissue-type plasminogen activator (u-PA, t-PA) on serum-deprived human dermal fibroblasts. In this work we have studied the transcription-dependent changes of some cell-cycle related genes associated with the biological activity of PAs, as well as the possible involvement of protein tyr kinases (PTK) and/or protein kinase C (PKC) in the mitogenic signal transduction. The data obtained demonstrate that the growth factor activity of PAs is associated with: - a rapid transient activation of early response genes, c-fos, c-jun and c-myc; - the subsequent coordinated down-regulation of p53 and p21CIP1; - the constant expression of the MEK1 mRNA in every phase of the cell cycle. Quiescent (G0) cells did not express c-fos, c-jun, c-myc and cyclin A, but upon stimulation with mitogens (fetal calf serum (FCS), u-PA, t-PA) the cyclin A mRNA expression was observed in concomitance with the activation of DNA synthesis. Therefore u-PA, t-PA and FCS similarly modulate the expression of c-fos, c-jun, c-myc, p53, p21CIP1 and cyclin A with only slight differences likely related to the time required for activation of DNA synthesis. The PAs mitogenic stimulation of serum-starved cells was associated with the internalization of their molecules, as revealed by immunostaining. The biological activity of u-PA, t-PA, as well as that of limiting concentration of FCS (1%), was mediated by PTK and PKC. Conversely, PTK, but not PKC, was involved in the activation of the proliferative response of basic fibroblast growth factor in the same experimental conditions. In conclusion, u-PA and t-PA can utilize two different pathways, one depending on PTK and the other on PKC in a way similar to the mitogenic activity induced by low concentration of FCS (1%).

Quantitative in situ hybridization for the evaluation of gene expression in asynchronous and synchronized cell cultures and in tissue sections.

We describe an image analysis (IA) system that has been applied for the quantitative evaluation of mRNAs evidenced by in situ hybridization (ISH) with radiolabelled probes in cultured cells and in tissue sections. The ISH-IA method was used for the evaluation of cultured cell morphological parameters such as cell and nucleous area (CA and NA, respectively) in parallel with the levels of mRNAs detected as hybridization grains areas (GA). The evaluation of these parameters, together with the analysis of the levels of mRNAs (c-jun, cyclin A) specific for given cell cycle phases (i.e. G1 and S/G2), allowed the identification, in asynchronous cultures of human skin fibroblasts, of cells in G1 and S/G2 phases. The mRNA levels measured by ISH-AI were comparable with those detected by RT-PCR. This method was also applied for the analysis of fibronectin (FN) gene expression in control skin fibroblasts in relationship with the different phases of the cell cycle and in comparison with a tumor cell line (Sk-Hep1), heterogeneous either for morphometric parameters or for the levels of this transcript. Finally, the ISH-AI was applied for the semiquantitative evaluation of the expression, localization and alternative splicing pattern of FN mRNA in normal liver and in hepatocellular carcinoma (HCC) tissue sections.

Expression of urokinase-type plasminogen activator (u-PA), u-PA receptor, and tissue-type PA messenger RNAs in human hepatocellular carcinoma.

Expression of plasminogen activators (PAs) and urokinase-type PA receptor (u-PAR) is associated with tumor growth and invasion. For in vivo human tumor tissues, there is no information on gene expression of PAs in hepatocellular carcinoma (HCC) or other hepatic pathophysiological conditions. In this study we examined the relative levels of u-PA, tissue-type PA (t-PA), and u-PAR mRNA expression in human HCC by reverse transcription-PCR compared with those expressed in peritumoral hepatic tissues. Twenty-five of 25 HCCs expressed u-PA mRNA, as well as 16 of 25 hepatic peritumoral tissues. However, none of the 14 cases of nontumorous liver samples (i.e., normal parenchyma, steatosis, and nonspecific reactive and chronic hepatitis) showed detectable levels of u-PA mRNA. The same samples analyzed for uPAR and t-PA mRNAs exhibited higher levels of these mRNAs in the malignant tissues compared with nontumorous ones. A strong correlation was found between the relative levels of u-PA and t-PA mRNAs detected in the tumor and in the corresponding peritumoral tissues (P < 0.001 for u-PA; P < 0.02 for t-PA). However, there was no correlation between the expression of u-PA and t-PA in HCC (P = 0.565). Furthermore, a significant inverse correlation was found between survival months of male patients and the relative level of u-PA mRNA (P < 0.05) detected at the time of biopsy, whereas no correlation was found in the case of t-PA mRNA. These results are in line with the possible differential biological role of u-PA and t-PA in the tumor etiopathogenesis and suggest that the detection of relative levels of u-PA mRNA may be a useful prognostic factor for male HCC patients.

Phosphorylation of human plasminogen activators and plasminogen.

Plasminogen (PG), urokinase-type plasminogen activator (u-PA) and tissue-type PA (t-PA) are the main molecules involved in fibrinolysis and in many other physiological and pathological processes. In the present study we report that human t-PA, purified from human melanoma cells, and PG, purified from human plasma, both contain P-Tyr residues, as revealed by immunoblotting analyses with monoclonal anti-P-Tyr antibodies. In addition HPLC amino acid analysis of acid-hydrolyzed t-PA, PG and u-PA, shows that: (i) P-Ser and P-Tyr residues are present in t-PA; (ii) P-Thr and P-Tyr are present in PG; (iii) P-Ser, P-Thr and P-Tyr are present in u-PA. The utilization of monoclonal anti-P-Ser and anti-P-Thr antibodies in immunoblotting experiments has confirmed these data which indicate that phosphorylation is a common feature of PAs and of PG.

Plasminogen activators in nude mice xenotransplanted with human tumorigenic cells.

Nude mice have been subcutaneously inoculated with human tumorigenic fibrosarcoma cells (HT-1080) producing urokinase-type plasminogen activator (u-PA) or with human tumorigenic melanoma cells (G-361) producing tissue-type plasminogen activator (t-PA). Human u-PA (hu-PA) and t-PA (ht-PA) were found in the plasma and in the tumors of mice injected with HT-1080 or G-361 cells, respectively. Metastases containing ht-PA were observed in different organs of mice transplanted with G-361 cells, while mice injected with HT-1080 cells did not develop metastases. These data would suggest a relationship between the metastatic potential of G-361 cells and t-PA. The parallel increase of the levels of endogenous murine PAs (m-PA) activities might play a crucial role in the early stages of tumor growth and metastasis, since the biological effects of the PAs produced by the transplanted tumor cells can not be dissociated from those of the PAs induced in the host.

Urokinase-type and tissue-type plasminogen activators as growth factors of human fibroblasts.

In this study we have verified the mitogenic effect of urokinase-type (u-PA) and tissue-type plasminogen activators (t-PA) on human normal fibroblasts. We report that both PAs can induce DNA replication and cell division in serum-deprived cultured human skin fibroblasts. The activity of u-PA and t-PA is, respectively, three- and twofold more potent than that exerted by epidermal growth factor (EGF) with an activity slightly lower (50-60%) than that induced by basic fibroblast growth factor (bFGF). The u-PA and t-PA, but not plasmin, induced DNA synthesis, which could be neutralized by anti-u-PA and anti-t-PA antibodies, respectively, but was insensitive to aprotinin treatment. The addition of anti-u-PA-receptor (u-PAR) monoclonal antibodies to the assays selectively suppressed the mitogenic effect exerted by u-PA, but not that of t-PA, and the amino-terminal fragment of u-PA, containing the EGF-like domain and the kringle module, did not elicit any mitogenic activity. Anti-bFGF antibodies completely suppressed the mitogenic activity of bFGF, but did not have any effect on that of u-PA and t-PA; the activity of both PAs was inhibited by anti-fibronectin IgG concentrations ineffective on bFGF. These results indicate that PAs may be considered growth factors of human fibroblasts.

RT-PCR detection of fibronectin EDA+ and EDB+ mRNA isoforms: molecular markers for hepatocellular carcinoma.

Alternative splicing of fibronectin pre-mRNA has been shown to be independently regulated at the EDA and EDB regions in a tissue and developmental stage-specific manner. In this study, RT-PCR approaches were developed for the detection of EDA and EDB FN mRNA isoforms in hepatocarcinoma cells (SK-Hep-I) grown in vitro and in human liver biopsies. While EDA+ and EDB+ isoforms were not present in control adult liver, they were detectable in the hepatocarcinoma cells and in fetal liver. The RT-PCR analysis, extended to biopsies of malignant and non-malignant hepatic tissues, showed that FN mRNAs containing the EDA and EDB sequences were present in the 14 hepatocellular carcinomas (HCCs) tested but absent in the non-tumorous liver tissues (i.e., normal parenchyma, non-specific reactive and chronic hepatitis, steatosis). The EDB+ FN mRNA isoforms were also detected in 3 cases of benign neoplasm (hepatocellular adenoma, HCA, I; nodular focal hyperplasia, NFH, 2), while the EDA+ was only detectable in I of the 2 cases of NFH. In addition, both EDA+ and EDB+ isoforms were expressed in 5 out of 9 cirrhotic livers surrounding the tumors. This molecular analysis, which can also be performed on small liver biopsies (2 mg), may therefore be a useful additional tool in the diagnosis of HCC.

Tyrosine phosphorylation of human urokinase-type plasminogen activator.

Immunoblotting analysis of purified human urokinase plasminogen activator (u-PA), gives a positive signal when reacted with anti-phosphotyrosine monoclonal antibodies (MoAb anti-P-Tyr); competition with o-phospho-DL-tyrosine (P-Tyr) but not o-phospho-DL-threonine or serine (P-Treo, P-Ser) completely suppresses this signal. Either the 55 kDa u-PA form and the lower Mw form (33 kDa) derived from the 55 kDa u-PA are Tyr-phosphorylated also the u-PA secreted in the culture media of human fibrosarcoma cells (HT-1080) is phosphorylated in tyrosine as well as u-PA present in tissue extracts of tumors induced in nude mice by HT-1080 cells. These data show that urine purified human u-PA and u-PA produced by human fibrosarcoma cells, in vitro and in vivo, are phosphorylated in tyrosine; furthermore our data show that u-PA is the major Tyr-phosphorylated protein present in these human tumor cells.

A rapid and highly sensitive solid-phase enzyme immunoassay specific for human fibronectin using a characterized monoclonal antibody.

A solid-phase enzyme immunoassay (EIA) was developed for the quantitation of human fibronectin in body fluids and cell culture media. In the assay a human fibronectin-specific murine monoclonal IgG1 (f-33) was used as capture antibody and polyclonal rabbit anti-fibronectin as detector antibody. The antibody showed no reactivity to purified monkey, dog, rabbit, horse, sheep, mouse, bovine or chicken fibronectins. The determinant of the monoclonal antibody was mapped to the cell-binding region of the fibronectin molecule. This localization was based on the use of purified fragments of fibronectin, immunoblotting, EIA and inhibition of fibroblast adhesion and spreading by the antibody. The detection limit of the fibronectin assay was 2 ng/ml. The assay was used for the quantitation of fibronectin in human plasma, urine and cerebrospinal fluid specimens and culture media of human cells.

Tissue type plasminogen activator, but not urokinase, exerts transformation-enhancing activity.

Previous studies have established that plasma cryoprecipitates of tumor patients, culture media of transformed cells and defined proteolytic fragments of fibronectin enhance the morphological cell transformation ( TEF activity) in cultures of chicken embryo fibroblasts infected with temperature-sensitive mutants of Rous sarcoma virus. We now report that purified human tissue type plasminogen activator (t-PA), but not urokinase (u-PA), has a similar TEF activity, at doses as low as 2 ng/ml (30 pM). Specific antibodies effectively neutralized the activity. No significant contamination (less than or equal to 1%) between the preparations of t-PA and fibronectin (FN) or its fragments ( FNdp ) was detected. The results suggest that t-PA may have a direct role in the process of morphological cell transformation in vitro.

Monoclonal antibody against the N-terminal end of human plasma fibronectin.

Purified human plasma fibronectin was digested with cathepsin G and the degradation products were tested for reactivity towards a monoclonal antibody. In an immunoblotting assay, after sodium dodecyl sulphate/polyacrylamide-gel electrophoresis of the digestion products, the 85 000-Mr and 72 000-Mr gelatin- and heparin-binding fragments as well as the N-terminal 30 000-Mr heparin-binding fragment reacted with the antibody, whereas the 64 000-Mr gelatin- and heparin-binding fragment did not. In enzyme immunoassay the antibody reacted with intact fibronectin and the 30 000-Mr fragment but not with a 40 000-Mr gelatin-binding fragment. The alignment of the binding domains in these fragments and in the intact molecule [Vartio (1982) Eur. J. Biochem. 123, 223-233] localizes the antigenic determinant to the 21 000 Da N-terminal Staphylococcus aureus-binding region of fibronectin.

Evidence for preferential proteolytic cleavage of one of the two fibronectin subunits and for immunological localization of a site distinguishing them.

Purified plasma fibronectin was digested sequentially by thrombin and cathepsin G or by cathepsin G alone and the degradation products and their gelatin-binding and heparin-binding fractions were analyzed in NaDodSO4-polyacrylamide gel electrophoresis followed by immunoblotting with a defined monoclonal anti-fibronectin antibody. In early cathepsin G digests, several gelatin-binding fragments were detected: a few large (Mr greater than or equal to 150 000) polypeptides and fragments of Mr = 85 000, 72 000, 64 000 and 40 000. The 85 000-Mr and 64 000-Mr fragments appeared as closely spaced doublets and reacted with the antibody while the 72 000-Mr and 40 000-Mr fragments did not. Therefore the 64 000-Mr fragments are likely to be derived from the 85 000-Mr fragments. Three large fragments that bound to heparin, but not to gelatin were detected: Mr = 145 000, 135 000 and 120 000. Of these only the 135 000-Mr peptide reacted with the antibody. When fibronectin was digested with thrombin, polypeptides of Mr = 180 000-200 000 and a 30 000-Mr NH2-terminal fragment were produced. Cathepsin G added to this mixture further cleaved the fragments to a digestion pattern resembling that obtained from intact fibronectin except that the 85 000-Mr and 64 000-Mr fragments appeared as single bands and the amount of the 72 000-Mr fragment was reduced. The results suggest that thrombin cleaves the 30 000-Mr fragment preferentially from the NH2-terminal end of one of the two subunits of fibronectin and that the 85 000-Mr, 72 000-Mr and 64 000-Mr fragments obtained by the additional cathepsin G digestion were derived from the other chain. The results are consistent with the model that the antigenic determinant resides 72 000-85 000 Da from the NH2-terminus and is cleaved by cathepsin G alternatively at one of its sides. Thus, the components of the 85 000-Mr and 64 000-Mr doublets are derived from different subunits and the region located by the antibody may be responsible for the difference in their migration in the polyacrylamide gel.

Fibronectin and its proteolytic fragments. Potential as cancer markers.

Since the discovery of fibronectin as a transformation-sensitive 'cell surface' protein, it has been the focus of intensive studies. Fibronectin has multiple interactions and functions and is composed of distinguishing properties of malignant cells, properties. Invasiveness and metastasis, distinguishing properties of malignant cells, involve penetration through components of the extracellular matrix. Enzymatic degradation of matrix components is involved in these phenomena. Proteolytic targets of the pericellular matrices of cells in culture include fibronectin that has been found to be even selectively susceptible to proteinases. Defined fragments of fibronectin have transformation-promoting activity in experimental conditions and such fragments, detected in tumor patient body fluids, may serve as markers for tumor progression.

Transformation-enhancing activity of gelatin-binding fragments of fibronectin.

Studies have established that cryoprecipitates of the plasma of tumor patients contain a biological activity enhancing morphological cell transformation (transformation-enhancing factor; TEF) in cultures of chicken embryo fibroblasts infected with temperature-sensitive mutants of Rous sarcoma virus. We report here that similar TEF activity is effected by defined fragments of human plasma fibronectin obtained by limited digestion with major humoral or tissue proteinases. TEF activity was obtained from plasminolytic fragments of fibronectin and from cathepsin G-treated fibronectin. No activity was recorded from intact dimeric fibronectin or its reduced and alkylated subunits, from fibrinogen or its plasminolytic fragments, or from plasmin (EC 3.4.21.7) or cathepsin G (EC 3.4.21.20) treated or untreated with proteinase inhibitors. All of the TEF activity of the proteolytic fragments of fibronectin was located on the gelatin-binding peptides. The minimum effective doses in the TEF assay were 750 ng/ml of plasmin-treated fibronectin, 100 ng/ml of gelatin-binding plasminolytic fibronectin (enriched in Mr 180,000--190,000 polypeptides), and 100 ng/ml of gelatin-binding fragments of cathepsin G-treated fibronectin (enriched in a Mr 30,000 fragment). TEF activity of proteinase-treated fibronectin was inhibited by gelatin and by intact dimeric fibronectin. The potent TEF activity of proteolytic fragments of fibronectin raises the possibility that they may have a role in malignant transformation.

Inhibition of Rous sarcoma-virus-induced transformation by proteins involved in blood coagulation.

Plasma, serum and the main proteins involved in the final steps of blood coagulation, fibrinogen, thrombin and plasma transglutaminase, have been tested for their ability to inhibit cell transformation and/or to induce reversion of the transformed to the normal phenotype in RSV-infected chicken embryo fibroblasts. The results obtained show that plasma, and not serum, fibrinogen, when converted to fibrin by thrombin, and activated transglutaminase alone prevent focus appearance and cause reversion of transformed to normal cell morphology.