A cytokeratin-immunohistochemical study of focal nodular hyperplasia of the liver: further evidence that ductular metaplasia of hepatocytes contributes to ductular "proliferation"

A cytokeratin-immunohistochemical study was performed on eight specimens of focal nodular hyperplasia of the liver in order to determine whether hepatocytes in this lesion can express "bile duct type" cytokeratins. Serially cut cryostat sections were reacted with a panel of six monoclonal antibodies specifically reactive with cytokeratin polypeptides nr. 7, 8, 18 and 19, using a 3-step immunoperoxidase procedure. Hepatocytes were positive for cytokeratins nr. 8 and nr. 18, whereas ductules contained in addition to cytokeratins nr. 8 and nr. 18 also polypeptides nr. 7 and nr. 19. In all cases, a variable number of hepatocytes close to fibrous septa or inside the nodules expressed cytokeratin nr. 7. In four cases, a small number of hepatocytes were also immunoreactive for cytokeratin nr. 19. Our data demonstrate that hepatocytes in focal nodular hyperplasia can express one or even two "bile duct type" cytokeratins, supporting the concept that ductular metaplasia of hepatocytes contributes to the ductular "proliferation" observed in this tumor-like lesion.     

A cytokeratin-immunohistochemical study of focal nodular hyperplasia of the liver: further evidence that ductular metaplasia of hepatocytes contributes to ductular “proliferation”

ABSTRACT— A cytokeratin-immunohistochemical study was performed on eight specimens of focal nodular hyperplasia of the liver in order to determine whether hepatocytes in this lesion can express “bile duct type” cytokeratins. Serially cut cryostat sections were reacted with a panel of six monoclonal antibodies specifically reactive with cytokeratin polypeptides nr. 7, 8, 18 and 19, using a 3-step immunoperoxidase procedure. Hepatocytes were positive for cytokeratins nr. 8 and nr. 18, whereas ductules contained in addition to cytokeratins nr. 8 and nr. 18 also polypeptides nr. 7 and nr. 19. In all cases, a variable number of hepatocytes close to fibrous septa or inside the nodules expressed cytokeratin nr. 7. In four cases, a small number of hepatocytes were also immunoreactive for cytokeratin nr. 19. Our data demonstrate that hepatocytes in focal nodular hyperplasia can express one or even two “bile duct type” cytokeratins, supporting the concept that ductular metaplasia of hepatocytes contributes to the ductular “proliferation” observed in this tumor-like lesion.     

Hematopoietic stem cell markers are expressed by ductal plate and bile duct cells in developing human liver

The identification of ductal plate cells as likely progenitors for bile duct epithelium and hepatocytes and their possible reappearance as oval cells in the regenerating liver have generated much interest in their pluripotential capacities. We have examined the distribution of three hematopoietic stem cell markers, c-kit, CD34, and CD33 in addition to laminin, the standard cytokeratin markers CAM 5.2, CK 18, and CK 7 and the oval cell marker OV-6 in fetal liver during various stages of development. Hematopoietic stem cell markers were expressed in ductal plate cells in a pattern similar to the early cytokeratin markers CAM 5.2 and CK 18. Cells stained strongly for these early cytokeratin markers until 22 weeks. Thereafter, the expression of these markers decreased while positivity for CK 7 increased. Bile duct cells showed a distribution of hematopoietic and cytokeratin markers resembling that of ductal plate cells. Both ductal plate cells and bile duct cells expressed OV6 strongly throughout development. This study showed similarity between hepatic and bile duct precursors and bone marrow stem cells. The comparable distribution of markers in bile duct epithelium and ductal plate cells may imply fewer transitional stages between ductal plate cells and bile duct epithelium than between the putative stem cells and hepatocytes.     

Expression of cytokeratin 19 during human liver organogenesis

ABSTRACT— Immunohistochemistry with monoclonal anti-cytokeratin antibodies has revealed the presence of cytokeratin 19 in embryonic and early fetal hepatocytes. With the differentiation of bile ducts at about the 10th week, cytokeratin 19 disappears from liver cells but remains in bile duct cells. This marks an important step in the organogenesis of the liver.     

Expression of cytokeratin 19 during human liver organogenesis.

Immunohistochemistry with monoclonal anti-cytokeratin antibodies has revealed the presence of cytokeratin 19 in embryonic and early fetal hepatocytes. With the differentiation of bile ducts at about the 10th week, cytokeratin 19 disappears from liver cells but remains in bile duct cells. This marks an important step in the organogenesis of the liver.     

Morphological and immunohistochemical assessment of intrahepatic bile duct development in the rat

The hypotheses that intrahepatic bile ducts are derived either by a transformation of periportal liver cells or by dichotomous branching of the extrahepatic bile ducts were investigated in fetal and postnatal rat livers by histological and immunohistochemical methods using an antiserum to prekeratin which, in the liver, binds to biliary epithelial cells (BEC). In conventionally stained sections, bile duct development was observed to begin in the 19 day fetus around the larger branches of the portal vein, with the formation of lumina surrounded by cuboidal or elongated hepatoblast-like cells on the portal aspect and readily distinguished hepatoblasts on the lobular aspect. At 21 days, these structures had developed into canals of Hering lined jointly by recognizable liver cells and BEC. The number of canals of Hering per portal tract peaked at 22 days' gestation and diminished in number at birth and over the ensuing 56 h, with a concomitant increase in fully formed ducts. Bile ducts lined completely by BEC were first found at 20 days. Immunohistochemically, prekeratin antigens were first detected at 20 days in duct-like structures not only in phenotypic BEC but also in adjacent cells with an hepatoblast phenotype. Such intermediate cells were present until birth. These findings support the view that intrahepatic bile ducts develop by a reorganization and modulation of the periportal hepatoblasts to BEC.     

New clues for the developing human biliary system at the porta hepatis

The human biliary system is formed from the hepatic diverticulum, a structure which develops from the embryonic foregut in the fourth week of gestation. The cephalic portion of the hepatic diverticulum lies within the septum transversum, and gives rise to entodermal cells which become the primitive hepatocytes. The caudal part of the hepatic diverticulum is molded by mesenchyme to form the gallbladder, cystic duct, and extrahepatic bile duct. The gallbladder is initially tubular in shape, and undergoes morphological changes to become saccular during the 11th week of gestation. The extrahepatic bile duct elongates and widens as gestation progresses, and intramural mucus glands develop. There is no solid stage during the development of the extrahepatic bile duct. The extrahepatic bile duct is a well‐defined tubular structure by the 6th week of gestation, whereas the intrahepatic biliary system during this period of gestation is represented by the primitive ductal plate. The ductal plate undergoes structural changes from the 11th week of gestation, beginning at the porta hepatis and progressing through gestation to the periphery of the liver. This remodeling process shapes the ductal plate from a flat sheath of biliary epithelium surrounding the portal vein branches into a network of interconnecting tubular structures. Mesenchyme plays an important role in ductal plate remodeling. The intrahepatic biliary system is in luminal continuity with the extrahepatic bile duct throughout gestation at the porta hepatis. The major bile ducts at the porta hepatis are fully formed by the 16th week of gestation.     

The correlation between portal myofibroblasts and development of intrahepatic bile ducts and arterial branches in human liver

BACKGROUND/AIMS: The development of the intrahepatic bile ducts most likely requires interactions between epithelial and mesenchymal cells. In view of the epithelial-mesenchymal interactions between portal myofibroblasts (pMFs) and biliary epithelial cells in adult diseases of the bile ducts, we investigated the presence and function of pMFs during the development of intrahepatic bile ducts, as well as the development of intrahepatic branches of the hepatic artery. METHODS: We performed haematoxylin-eosin-stainings and immunohistochemistry for alpha-smooth-muscle actin, cytokeratin 19 and vimentin on serial sections of 45 fetal and postnatal liver biopsies. RESULTS: The mesenchyme of portal tracts in the ductal plate stage devoid of a hepatic artery branch, contained numerous and diffusely scattered pMFs. Portal tracts with a hepatic artery branch were always larger than those without and showed a decreasing number of pMFs. In the remodeling stage, all portal tracts contained a hepatic artery branch, and pMFs were restricted to the periductal mesenchyme. These periductal pMFs disappeared after full incorporation of the bile duct. CONCLUSION: Our findings strongly suggest interactions between pMFs and epithelial cells of the developing bile ducts. The development of the intrahepatic arterial branches always precedes the incorporation of the tubular segments of the ductal plate.     

Bile duct cell apoptosis is a rare event in primary biliary cirrhosis

BACKGROUND: The frequency of apoptosis in bile duct cells of primary biliary cirrhosis is still unclear spanning from rare to 50% in the various reports. AIM: To study bile duct cell apoptosis in stage I primary biliary cirrhosis lesions. PATIENTS: Nine stage I-II biopsies with a total number of 26 bile ducts of different sizes, selected from a larger series on the basis of the expression on serial frozen sections of HLA-DR and Fas antigens. METHODS: Apoptosis was evaluated by a DNA fragmentation assay on frozen sections, according to the manufacturer's protocol and by expression of apoptosis related cytokeratin neoepitopes. Bile duct cell proliferation was assessed by MIB1 (Ki-67) expression. RESULTS: Apoptosis was frequently found in inflammatory cells of portal tracts and sinusoids. Apoptosis of hepatocytes was also systematically observed. Only 4 positive bile duct cells were found in 3 bile ducts from 3 biopsies. Quantitative evaluation was not attempted. Cholangiocyte proliferation was observed in the same ducts and occasionally in other biopsies. CONCLUSIONS: These data suggest that cholangiocyte death by apoptosis at the level of typical primary biliary cirrhosis lesions is a rare event, at least in early stages of the disease. The observed rate of proliferation was consistent with the rate of apoptosis.     

Differential expression of MHC class II subregion products on bile duct epithelial cells and hepatocytes in patients with primary biliary cirrhosis

To study the expression of MHC Class II subregion gene products on biliary epithelial cells in primary biliary cirrhosis, frozen sections from liver biopsies of 15 patients with primary biliary cirrhosis were studied immunohistochemically using HLA-D subregion specific monoclonal antibodies L243 (HLA-DR), Leu10 (HLA-DQ) and B7/21 (HLA-DP). Patients with early stages of primary biliary cirrhosis showed expression of HLA-DP, HLA-DR and HLA-DQ subregion gene products on bile duct epithelial cells. In advanced stages of disease, no MHC Class II antigens or only HLA-DR and HLA-DP were expressed on bile duct cells. While normal hepatocytes did not express detectable amounts of MHC Class II antigens, hepatocytes from liver biopsies of four patients with primary biliary cirrhosis showed a distinct staining exclusively with monoclonal antibodies specific for HLA-DR. The expression of MHC Class II antigens on parenchymal cells was independent of a lymphocytic infiltration into the tissue. This study demonstrates that bile ductular cells, but not hepatocytes, express a full set of MHC Class II molecules at least during the early stages of primary biliary cirrhosis. We propose, therefore, that the expression of both HLA-DR and HLA-DQ subregion products on bile duct epithelial cells may be a necessary, although not sufficient, condition for the initiation of an autoimmune process leading to the destruction of intrahepatic bile ducts in primary biliary cirrhosis.     

Serum vitamin A deficiency and increased intrahepatic expression of cytokeratin antigen in alcoholic liver disease

The clinical and histologic significance of cytokeratin antigen expression in various intrahepatic locations was assessed in 57 patients with alcoholic liver disease as part of a large Veterans Administration Cooperative Study of Alcoholic Hepatitis. Cytokeratin antigen was demonstrated in fixed, paraffin-embedded liver tissue by an avidin-biotin peroxidase method using a mixture of two different monoclonal antibodies, AE1 (acidic; 48, 50 and 56.5 kD) and AE3 (basic; 52, 56, 58 and 65 to 67 kD). In contrast to the normal liver, in which only bile duct epithelium was positive, this antibody mixture stained both bile ducts and hepatocytes in pathologic livers. Serum levels of vitamin A showed a significant inverse correlation with the amount of cytokeratin antigen (scale: 0 to 3) in hepatocytes without Mallory bodies (p = 0.001), in Mallory body-containing hepatocytes (p less than 0.0001) and in bile ducts (p = 0.0074). Increased amount of cytokeratin antigen in each of these locations, in turn, correlated directly with the histologic severity of the liver disease. Histologic severity (fibrosis, parenchymal degeneration/necrosis, hepatocyte regeneration and inflammation) was significantly higher in patients when either Mallory bodies (p less than 0.0001) or cytokeratin antigen (p = 0.0021) was present in hepatocytes. Demonstration of cytokeratin antigen in hepatocytes which contained Mallory bodies correlated positively (p = 0.03) with clinical severity of the liver disease as determined by high serum bilirubin and prolonged prothrombin time (Maddrey's discriminant function).(ABSTRACT TRUNCATED AT 250 WORDS)     

Distribution of HLA-DR, HLA-DP, HLA-DQ antigens in liver tissue from patients with primary sclerosing cholangitis

This study describes the distribution of the major histocompatibility class II antigens HLA-DR, HLA-DP, and HLA-DQ in liver tissue from eight patients with primary sclerosing cholangitis, by means of the immunoperoxidase technique on cut cryostat sections. HLA-DR was expressed on the bile duct epithelium and vascular endothelium from all patients. HLA-DP showed an expression much like HLA-DR on the bile duct epithelium, but the vascular endothelium mostly did not stain for HLA-DP. HLA-DQ was generally more weakly expressed, and only three of the patients expressed HLA-DQ on the bile ducts. None of the hepatocytes expressed any of the MHC class II antigens.     

Surgical anatomy of the medial segment (S4) of the liver with special reference to bile ducts and vessels

BACKGROUND/AIMS: Resection of the inferior area of the medial segment (S4a) plus S5 with preservation of the superior area of the medial segment (S4b) is being performed to manage hilar bile duct carcinoma and pT2 type gallbladder carcinoma, and thus, attention has been focused on the surgical anatomy of the medial segment of the liver to identify the specific vessels and bile ducts of the areas of that segment to be resected and to be preserved. METHODOLOGY: Anatomical study of the bile duct, portal vein, middle hepatic vein, and middle hepatic artery to the medial segment branches of the liver (S4) was performed in a total of 171 specimens comprised of 71 adult cadavers, and 100 liver casts. RESULTS: 1) Two main types of bile duct branches of the medial segment (B4) were recognized. Type I included the branches which joined to the left hepatic duct on the hilar duct side (35.5%), and type II included the branches that joined on the peripheral side (54.6%). Several subtypes were also found in both types. The B2-B3 confluence was mostly on the left (41.7%) or posterior (42.7%) to the umbilical portion (UP) of the portal vein, and to the right of the UP (hilar side) in only 15.6%. 2) The portal vein of the medial segment branches (P4): P4a branched from the right angle and upper right border of the UP in every specimen. The most common morphology was 1 large and 2-3 small branches (41%). P4b was almost always found to branch posterior to the UP and lower than P4a, and the most common morphology was 1 large and 0-1 small branches (57.8%). 3) The middle hepatic vein: In 83.2% a common trunk was observed at the confluence with the inferior vena cava, and 8 types of the middle hepatic vein were recognized. 4) The middle hepatic artery: It arose from the left hepatic artery in 61.5%, from of the right hepatic artery in 27.5%, from the proper hepatic artery in 5.5%, and from both the left and the right hepatic artery in 5.5%. CONCLUSIONS: The detailed vascular and bile duct anatomy of S4 is described. This study should be helpful in identifying the specific vessels and bile ducts of the areas of the medial segment to be resected and to be preserved, thereby facilitating resection of the medial segment.     

Serum levels of tissue polypeptide specific antigen are correlated with hepatocyte cytokeratin expression in alcoholic liver disease

BACKGROUND: Serum levels of the tumor marker tissue polypeptide specific antigen (TPS, cytokeratin 18 fragments) are increased in patients with alcoholic liver disease, particularly in cases of alcoholic hepatitis. Mallory bodies, characteristic of alcoholic hepatitis, are cytokeratin 8 and 18 aggregates. The study was aimed at investigating the possible relationship of serum TPS levels with hepatocyte cytokeratin expression in patients with alcoholic liver disease. METHODS: Twenty-four patients with alcoholic liver disease were studied. Immunohistochemical staining for cytokeratins 8 and 18 was performed in liver specimens by means of CAM 5.2 monoclonal antibody. The number of hepatocytes containing CAM 5.2-reactive cytokeratin inclusions was compared with serum TPS levels. MAIN RESULTS AND CONCLUSIONS: The vast majority of alcoholics (95%) showed increased (>100 units/liter) serum TPS levels. Serum TPS levels were significantly correlated with the number of hepatocyte cytokeratin inclusions. Serum TPS levels can predict hepatocyte cytokeratin expression in patients with alcoholic liver disease.     

A scanning electron microscopic study of postnatal development of rat peribiliary plexus

Hepatic microcirculation in the developing stages is not fully clarified. This study aimed at clarifying the development of hepatic microcirculation, especially peribiliary vascular plexuses, in neonatal rats by corrosion cast procedures and scanning electron microscopy. Peribiliary vascular plexuses of large bile ducts at the hilus of 1-day-old rats showed a simple capillary network that directly poured into the portal vein. The hepatic artery promptly tapered down into many small branches and ended in sinusoids or capillary plexuses around portal vein branches near the hilus. Neither apparent hepatic artery branches nor peribiliary vascular plexuses were found in the peripheral areas of the liver. On the seventh day, very loose peribiliary vascular plexuses and small hepatic artery branches appeared up to the peripheral portal tracts, and the peribiliary vascular plexuses of large bile ducts were still capillary networks. After the seventh day, peribiliary vascular plexuses of large and medium-sized portal tracts often poured into side branches of the portal vein rather than directly into the portal vein as shown in the rats from the first day mentioned above. Peribiliary vascular plexuses of the large bile duct in 2-week-old rats showed a double layer, that is, an outer layer composed of arteries and veins and an inner layer of capillary vessels, that resembled peribiliary vascular plexuses of adult livers. After nearly 4 wk, hepatic microcirculations had almost fully developed into those of adult rats. It was concluded that the gradual postnatal development of the hepatic arterial system and the peribiliary vascular plexuses occurred in parallel with the maturation of the intrahepatic biliary trees.     

Alterations in bile ducts and peribiliary microcirculation in rats after common bile duct ligation

The chronological changes in ductular epithelium and peribiliary capiliary plexuses (PBPs) after bile duct ligation are not well understood. Therefore, we examined alterations in both intrahepatic bile ductules and peribiliary microcirculation in rats after ligation of the common bile duct using immunohistochemistry, transmission electron microscopes (TEM), and scanning electron microscopes (SEM). Ductular proliferations appeared first in the peripheral areas of the portal spaces and then gradually advanced along with increasing jaundice. Distorted configuration of the hepatic parenchyma with interconnecting stroma rich in irregular ductules developed 4 weeks after the ligation. Numerous biliary cell types and cells of types intermediate between hepatocytes and biliary cells together with poorly fenestrated capillary-type vessels appeared in the periportal parenchyma, in association with an increased number of canalicular-ductular junctions on TEM. These biliary cells were often found within the lobules and apart from the stroma. SEM examination of hepatic microvascular casts using methacrylated resin showed formation of irregular portal and periportal capillary networks, partly derived from coarsened sinusoids. Direct connections between the newly formed capillary networks and the pre-existing PBPs or sinusoids were numerous, although there were few direct connections between the capillary networks and the hepatic arterial branches. Thus, these proliferated ductules and newly formed complicated capillary networks might play an important role in the effective transport of biliary materials between hepatocytes and native bile ducts or proliferated ductules through the altered microcirculation after bile duct ligation.     

Immunoperoxidase localization of ursodeoxycholic acid in rat biliary epithelial cells. Evidence for a cholehepatic circulation.

To explain the hypercholeretic effect of ursodeoxycholic acid, a cholehepatic shunt circulation has been postulated. This pathway includes secretion by the hepatocyte into bile and absorption by the biliary epithelial cells. To test this possibility, we have attempted to localize ursodeoxycholic acid in hepatocytes and portal bile duct cells by an indirect immunoperoxidase technique using specific polyclonal antibodies against ursodeoxycholic acid and cholic acid conjugates. Rat livers were fixed with paraformaldehyde, incubated with antibodies directed against bile acids and examined by electron microscopy. After infusion of ursodeoxycholic acid and incubation with antibodies against this bile acid, an electron-dense staining was observed on the apical (luminal) membrane of the portal bile ductal cells. In contrast, no staining was observed when no ursodeoxycholic acid was infused, or after infusion with taurocholate and incubation with antibodies against cholic acid conjugates. These observations are consistent with absorption of ursodeoxycholic acid by the portal bile ducts and a cholehepatic circulation of this bile acid.     

Immunoperoxidase localization of ursodeoxycholic acid in rat biliary epithelial cells. Evidence for a cholehepatic circulation

ABSTRACT— To explain the hypercholeretic effect of ursodeoxycholic acid, a cholehepatic shunt circulation has been postulated. This pathway includes secretion by the hepatocyte into bile and absorption by the biliary epithelial cells. To test this possibility, we have attempted to localize ursodeoxycholic acid in hepatocytes and portal bile duct cells by an indirect immunoperoxidase technique using specific polyclonal antibodies against ursodeoxycholic acid and cholic acid conjugates. Rat livers were fixed with paraformaldehyde, incubated with antibodies directed against bile acids and examined by electron microscopy. After infusion of ursodeoxycholic acid and incubation with antibodies against this bile acid, an electron-dense staining was observed on the apical (luminal) membrane of the portal bile ductal cells. In contrast, no staining was observed when no ursodeoxycholic acid was infused, or after infusion with taurocholate and incubation with antibodies against cholic acid conjugates. These observations are consistent with absorption of ursodeoxycholic acid by the portal bile ducts and a cholehepatic circulation of this bile acid.     

Treatment strategy for hilar cholangiocarcinoma,with special reference to the limits of ductal resection in right-sided hepatectomies

The surgical anatomy of the hepatic hilar region is characterized by the three-dimensional formation of the branches of the bile duct, portal vein, and hepatic artery. The limit of ductal resection in hepatectomy for hilar cholangiocarcinoma is the most peripheral point where the hepatic ducts can be separated from the vasculature. The limit is different for each type of hepatectomy because the portal vein branches that should be preserved or divided vary with the extent of the hepatectomy, and therefore the limit of separation of the hepatic ducts differs. Surgeons are required to understand the surgical anatomy and to identify the precise area of cancer spread on a preoperative cholangiogram so as to choose the appropriate type of hepatectomy, and to ensure that the remnant ductal margin is cancer-negative.     

Rat liver alkaline phosphatases

Using biochemical and electron microscopic histochemical techniques, we studied membrane-bound alkaline phosphatase activities of rat hepatocytes and portal triads. Activity in portal triads was localized to capillaries surrounding bile ducts (peribiliary plexus) and arterioles. Despite the reputation of alkaline phosphatase as a "biliary enzyme," activity was not observed in bile ducts. Livers were separated into hepatocyte and portal triad fractions with collagenase. Enzyme from hepatocytes migrated faster during electrophoresis and eluted later during anion-exchange chromatography than that from portal triads. Thus, hepatocyte enzyme is more negatively charged (and also possibly smaller) than portal triad enzyme. Twelve hours after bile duct obstruction, new activity appeared on lateral and sinusoidal membranes of hepatocytes; appearance of portal triads did not change with obstruction. Electrophoretic mobilities of the two forms were not altered by obstruction. We conclude that two distinct liver alkaline phosphatases exist, one in hepatocytes, the other in portal triad blood vessels.