The recently identified claudins are dominant components of tight junctions, responsible for cell adhesion, polarity and paracellular permeability. Certain claudins have been shown to have relevance in tumor development, with some of them, especially claudin-4, even suggested as future therapeutic target. The aim of the present study was to analyze the expression of claudin-4 in the biliary tree, biliary tract cancers and hepatocellular carcinomas. A total of 107 cases were studied: 53 biliary tract cancers, 50 hepatocellular carcinomas, 10 normal liver and 10 normal extrahepatic biliary duct samples. Immunohistochemical analysis was performed on conventional specimens and on tissue microarrays as well. Claudin-4 was further investigated by Western blot analysis and real-time RT-PCR. Intense membranous immunolabeling was found for claudin-4 in all biliary tract cancers unrelated to the primary site of origin, namely intrahepatic, extrahepatic or gallbladder cancers. Normal biliary epithelium showed weak positivity for claudin-4. In contrast, normal hepatocytes and tumor cells of hepatocellular carcinomas did not express claudin-4. The results of Western immunoblot analysis and real-time RT-PCR were in correlation with the immunohistochemical findings. Cytokeratins, as CK7 (92%) and CK19 (83%) were mostly positive in biliary tract cancers, however, one-third of hepatocellular carcinomas also expressed CK7 (34%). HSA antibody (HepPar1) reacted with the majority of hepatocellular carcinomas (86%), while being positive in a low percentage of the biliary tract cancers (8%). In conclusion, this is the first report of a significantly increased claudin-4 expression in biliary tract cancers, which represents a novel feature of tumors of biliary tract origin. Claudin-4 expression seems to be a useful marker in differentiating biliary tract cancers from hepatocellular carcinomas and could well become a potential diagnostic tool.

Agrin is a multifunctional heparan sulfate proteoglycan originally discovered in the neuromuscular junctions and later observed in numerous other localizations. The presence of agrin in the liver, either healthy or diseased, has formerly not been reported. We detected agrin in minor amounts in the basement membranes of blood vessels and bile ducts in the healthy liver. The proliferation of bile ductules and the formation of new septal blood vessels in liver cirrhosis, as well as neoangiogenesis in the hepatocellular carcinoma (HCC) result in a dramatic increase in the quantity of agrin. Vascular and peribiliary basement membranes were strongly immunopositive for agrin in 29/29 human liver specimens with cirrhosis and HCC. However, sinusoidal walls of regenerative nodules in the cirrhotic liver consistently remained negative. Given the selectivity of agrin for tumor microvessels, agrin immunohistochemistry may prove helpful in recognizing malignant transformation in cirrhotic livers. Similar immunohistochemical observations were made on the liver of rats exposed to a combined cirrhosis/HCC induction treatment. In both human and rats, agrin probably originates from activated myofibroblasts, vascular smooth muscle cells and biliary epithelial cells. Increased agrin expression in human specimens, in the liver of 4/4 treated rats, as well as in isolated rat liver mesenchymal cells was verified by quantitative RT-PCR. Considering that agrin binds various growth factors, and it directly interacts with cell membrane receptors such as alphav-integrins, we hypothesize a stimulatory role for agrin in neoangiogenic processes such as tumor vascularization, and a supportive role in bile ductule proliferation.

No abstract

Agrin is a recently identified proteoglycan component of vascular and bile duct basement membranes in the liver. The selective deposition of agrin in hepatocellular carcinoma (HCC) microvessels versus sinusoidal walls prompted us to investigate the utility of agrin immunohistochemistry (IHC) in detecting malignant hepatocellular lesions. We focused on the differential diagnostic problems often presented by hepatocellular adenomas (HCAs) and dysplastic nodules. IHC for agrin was performed on 138 formalin-fixed, paraffin-embedded surgical specimens from 93 patients, including cirrhotic liver tissues (25), focal nodular hyperplasia (10), large regenerative nodules (8), low-grade (23) and high-grade (7) dysplastic nodules, small HCC (8), HCC (27), and HCA (30). Agrin immunostaining was compared with that of CD34 and, in selected cases, to glypican-3. The combination of agrin and CD34 sensitively (0.94) and specifically (0.93) identified lesions judged previously as malignant by histology. The majority of benign lesions were clearly agrin-negative, whereas the strength and extent of agrin IHC faithfully reflected dysplasia in "atypical" HCAs and in high-grade dysplastic nodules. Malignant lesions were uniformly positive. In conclusion, as agrin is highly selective for tumor blood vessels, IHC for agrin facilitates the discrimination of benign and malignant hepatocellular lesions. Moreover, whereas glypican-3 in some HCCs may appear in few scattered cells only, agrin is diffusely deposited in virtually all malignant lesions, which may prove advantageous in the evaluation of small specimens such as core biopsies.

No abstract

No abstract

The recently described matrilin protein family is part of the extracellular matrix, their pathophysiological role as well as distribution in liver diseases, however, have not yet been studied. Considering that matrilins have been found to play role in cell growth and tissue remodeling, their possible involvement in carcinogenesis has been raised. The main objective of this study was to investigate the changes in matrilin-2 expression which is one of the main components of basement membranes. Thirty-five cases of surgically resected hepatocellular carcinomas, 35 corresponding surrounding liver tissues and 10 normal liver samples were used for the study. In 15 of 35 cases the tumor developed on the basis of cirrhosis. Matrilin-2 protein expression was detected in normal liver around bile ducts, portal blood vessels, while sinusoids were negative by immunohistochemistry. Cirrhotic surrounding tissue showed intensive matrilin-2 staining along the sinusoids. Tumorous neovasculature was found strongly positive by immunohistochemistry. No differences, however, were detected by morphometry regarding the amount of protein expression based on the grade of hepatocellular carcinomas. Real-time RT-PCR did not show significant differences in matrilin-2 mRNA expression between normal, cirrhotic and tumor samples. This suggests posttranslational modification of matrilin-2 manifesting in altered distribution in liver fibrosis. Our data indicate that matrilin-2 is a novel basement membrane component in the liver, which is synthetised during sinusoidal "capillarization" in cirrhosis and in hepatocellular carcinoma. This is the first report to describe the expression and distribution of matrilin-2 in human normal and cirrhotic liver as well as in hepatocellular carcinoma.

No abstract