The Nature and Origin of Proliferated Bile Ductules in Alcoholic Liver Disease

Transcription

1 The Nature and Origin of Proliferated Bile Ductules in Alcoholic Liver Disease TOSHIKAZU UCHIDA, M.D. AND ROBERT L. PETERS, M.D. Proliferation of bile ductules is commonly seen in expanded portal tracts and periportal areas in many conditions, including advanced alcoholic liver disease. Such ductules are usually tortuous and irregular and composed of cuboidal cells. They frequently have poorly defined lumens. The epithelial cells are similar to those of normal bile ductules and small bile ducts; however, cells that appear intermediate between duct epithelial cells and hepatocytes are frequently identified by light and electron microscopy. The origin of the ductular cells from hepatocytes may be confirmed by the demonstration of the markers of hepatocytes, such as glycogen and glucose-6-phosphatase activity in some proliferated bile ductules. In addition, alcoholic hyalin is occasionally recognizable in the epithelial cells of bile ductules. The majority of periportal bile ductules appears to have been derived from transformation of hepatic cords rather than multiplication of preexisting bile ducts. The proliferated bile ductules seem to communicate between the bile canaliculi and the interlobular bile ducts. (Key words: Bile ductules, Cholangioles; Ductular transformation; Alcoholic liver disease) Am J Clin Pathol 1983; 79: THE BILE DUCTULES (cholangioles) are small biliary channels lined by epithelial cells and connecting the bile canaliculi of hepatocytes with the bile ducts. 910 They consist of two to four cuboidal epithelial cells on cross section and do not accompany terminal branches of portal veins and hepatic arteries. The proliferation of ductular structures commonly is seen in various liver diseases with hepatic injury, including alcoholic liver disease. The proliferated ductules often are characterized by an elongated or anastomosing configuration with poorly defined lumens; hence, the terms "atypical bile ductules," "pseudocholangioles," or "neocholangioles" are sometimes applied. Bile ductules are distributed irregularly in the fibrosis extending into the parenchyma, often associated with neutrophilic reaction. Since the middle of the nineteenth century, there has been a considerable controversy about the origin and function of proliferated bile ductules. 2,6,7 While some investigators 1 ' 4 suggested that bile ductules might originate from ductular transformation (metaplasia, modulation) of the liver cell cords; others 5 ' 913 thought they must be derived from the lengthening and tortuosity of preexisting interlobular bile ducts and ductules. The Received May 14, 1982; received revised manuscript and accepted for publication August 9, Address reprint requests to Dr. Uchida: USC Liver Unit, 7750 Golondrinas 1200 Building, Downey, California University of Southern California School of Medicine, Department of Pathology Liver Unit, Rancho Los Amigos Hospital, Downey, California proliferated ductules have been thought to function as an interposing link between the parenchyma and interlobular bile ducts, or alternatively, as an early form in the regeneration of new hepatocytes. The ultrastructural features of proliferated ductules were studied in detail by Schaffner and his colleagues, 8 ' 9 who concluded that ductules might be derived from proliferation of the preexisting ducts. In the present study, we attempt to clarify the nature and origin of proliferated ductules by enzyme histochemical staining in addition to light and electron microscopic studies. Liver tissue from patients with alcoholic liver disease was studied, because the proliferation of bile ductules is a frequent feature of that condition. Materials and Methods One liver biopsy specimen was obtained from each of five patients with an established history of chronic alcoholism and the clinical and laboratory findings of severe alcoholic decompensation. The patients were admitted in 1980 and 1981 to the University of Southern California Liver Unit at Rancho Los Amigos Hospital with features of hepatic decompensation. The summary of clinical data is given in Table 1. The patients had no clinical evidence or history of biliary obstructive disease or viral hepatitis. Percutaneous liver biopsies were performed under local anesthesia using Klatskin or Tru-cut needles. The liver tissues obtained were divided into three portions. The first portion was fixed in B5 (10% formalin in buffered mercuric chloride solution) embedded in araldite, sectioned 2 /im thick, and processed for conventional staining, including hematoxylin and eosin (H & E), Masson's trichrome, and periodic acid-schiff reaction both with (DiPAS) or without (PAS) prior diastase reaction; the second was immediately fixed in cold 1% osmium tetroxide in Milloning's solution (ph = 7.5) for two hours with or without prior fixation in cold Karnovsky's solution for two hours. This was embedded in Epon 812 after dehydration. After orientation by toluidine blue and PAS staining of 1-^m sections, adjacent /83/0300/0326 $01.20 American Society of Clinical Pathologists 326

2 Vol. 79 No. 3 DUCTULAR TRANSFORMATION 327 ultrathin sections were obtained, double-stained with uranyl acetate and lead citrate, and observed through an Hitachi HU-12 electron microscope. More than two periportal areas were examined per biopsy specimen and each periportal area must have at least three bile ductules. The third piece of liver tissue (except for case 5) was snap-frozen in isopentene that was cooled to a slush in liquid nitrogen for enzyme histochemistry. Five-micron thick, cryostat sections were stained for glucose-6- phosphatase (G6P) activity without fixation. Other cryostat sections were fixed in cold acetone for 1 hour and stained for succinic dehydrogenase (SD), diphosphopyridine nucleotide diaphorase (DPND), alkaline phosphatase (ALP), and gamma-glutamyl transpeptidase (GGTP) activities. Routine staining for H & E and PAS also was performed on acetone-fixed cryostat sections. The staining procedure of G6P, ALP, and GGTP activities has been described previously. 12 The staining of SD and DPND activities was done according to Hunt. 3 Negative control sections for enzyme activities were prepared in identical fashion but incubated without specific substrate. Four liver biopsy specimens with no pathologic changes were used for normal controls: one was used for electron microscopic study and three were used for enzyme histochemical staining. The number of proliferated bile ductules was calculated for two expanded portal and periportal areas in each specimen. The number of the nuclei of epithelial cells was counted to calculate the number of cells present. To examine for the presence of alcoholic hyalin in the bile ductules in alcoholic liver disease, a survey of 135 consecutive liver biopsy specimens from patients with alcoholic liver disease with hyalin was done. The biopsies were performed between January 1, 1980 and September 20, 1981 at Rancho Los Amigos Hospital. Results Conventional Light Microscopy Liver biopsy specimens in four cases showed extensive portal, periportal, perivenular, and sinusoidal fibrosis without cirrhosis; the remaining specimen (case 3) showed early cirrhosis. A few alcoholic hyaline bodies were present in hepatocytes of cases 2, 3, and 5. The bile ductules were markedly proliferated in all cases. The proliferated bile ductules were confined to expanded portal tracts and periportal areas, which were undergoing arachnoid fibrosis. The periportal areas had previously contained hepatocytes (Fig. 1). The perivenular areas had extensive fibrosis with widespread hepatocellular dropout but had only a few bile ductules, these were trapped and distributed randomly within the fi- Table 1. Clinical Summary of Five Patients Age (yr) Sex M F M M M Race Mexican White White White Black Primary Complaints Jaundice and ascites Jaundice and ascites Jaundice, ascites, and hematemesis Jaundice and anorexia Ascites and leg edema brosis, without a close association with blood vessels. The proliferated ductules were especially prominent in the area between fibrotic periportal areas and parenchyma, where direct continuity of ductules with hepatocellular cords frequently was seen (Fig. 1). Sometimes, the proliferated ductules had a sprinkling of neutrophils around them and often oriented towards ductules (cases 1 and 2). Many macrophages containing DiPAS-positive granules were present among the ductules. These DiPASpositive granules were seen sometimes in both hepatocytes and ductular epithelial cells. Each bile ductule had a mean of 11.2 nuclei (range, 2-58) in the total of 448 ductules observed in the noncirrhotic livers. Some portal tracts and their contiguous periportal area had more than 59 bile ductules. Generally, the more extensive thefibrosisand hepatocellular dropout in the periportal area, the more ductular structures were present. The epithelial cells forming ductules usually were cuboidal, 10 ^m in diameter, each with a single, centrally located, ovoid nucleus, which had densely and uniformly packed chromatin, a small nucleolus, and glassy eosinophilic cytoplasm (Figs. 1 and 2). The epithelial cells were similar to those of normal small duct and ductular cells, although sometimes dark cells were found. Among the ductular epithelial cells, there were larger cells characterized by a large nucleus and relatively abundant eosinophilic granular cytoplasm, which were apparently isolated in the fibrosis (Fig. 2). Some of the larger cells closely resembled hepatocytes (Fig. 2). Ductular structures situated closer to the hepatocellular parenchyma had larger numbers of such cells. Sporadically placed ductular epithelial cells showed strong positive PAS staining for glycogen, the remaining cells giving a negative reaction (Fig. 3). This staining was abolished by prior diastase digestion. The liver specimen in Figure 2 (case 3) had alcoholic hyalin in the epithelial cells of ductules. In the survey of 135 alcoholic liver biopsy specimens containing variable numbers of alcoholic hyalin in the parenchyma, 12 (8.9%) had alcoholic hyalin sporadically in the ductular epithelial cells. The presence of alcoholic hyalin in the bile ductular cells previously has been reported using electron microscopy. 14

3 328 UCHIDA PETERS A.J.C.P. March 1983 FIG. 1 (upper, left). Case 3. Proliferated bile ductules trapped in periportal fibrosis. They are composed of cuboidal cells and assume an anastomosing, elongated, and tortuous configuration. The lumens are ill-defined. Arrows indicate a continuous transition of two hepatocyte columns to bile ductules. Hematoxylin and Eosin (X330). FIG. 2 (upper, right). Case 3. There is alcoholic hyalin (arrows) in the epithelial cells of this bile ductule. Some epithelial cells have conspicuous nucleoli and relatively abundant cytoplasm resembling the hepatocyte. Many inflammatory cells are seen surrounding ductular cells. Hematoxylin and Eosin (X660). FIG. 3 (lower, left). Case 4. Some epithelial cells (arrows) in proliferated bile ductules stain positively for glycogen. PAS (X660).

4 vol. 79. No. 3 DUCTULAR TRANSFORMATION Electron Microscopy The majority of the epithelial cells in the proliferated bile ductules had very similar features to those of normal bile ductules and small ducts (Fig. 4).8~" The nucleus was round or ovoid, with uniformly and densely packed chromatin, and the cytoplasm had scanty, small organ- elles, including mitochondria and endoplasmic reticulum. The Golgi apparatus and tonohlaments were welldeveloped and pinocytotic vesicles often were seen next to the luminal surface. There were many microvilli projecting into the irregularly shaped lumens with occasional bleb formation. The intercellular connection was straight and tight with distinct junctional complexes in FIG. 4. Case 4. Electron micrograph of proliferated bile ductule. The epithelial cells constituting this ductule resemble those of normal duct. They have relatively scanty cytoplasm, decreased numbers of organelles, and well-developed tonofilaments. The intercellular connection is straight in the luminal half and interdigitated in the basilar half. The basal lamina surrounds the ductule. However, there are four bile canaliculi-like structures (arrows), the same as found in hepatocytes (X8,770).

5 330 UCHIDA AND PETERS the luminal half and was interdigitated or widened, with lacunae formation in the basilar half. There were basal lamina surrounding the ductules. Intermingled with these biliary-appearing epithelial cells were large cells with characteristics similar to those of hepatocytes (Figs. 5 and 6). Each had a large nucleus with a conspicuous nucleolus and peripherally condensed chromatin material. The relatively abundant cy- A.J.C.P. March 1983 toplasm had many large mitochondria, glycogen granules, and occasional fat droplets (Fig. 5). Endoplasmic reticulum was present but poorly developed. The junction between ductular cells was often straight, but others were irregular and frequently contained bile canaliculilike structures (Fig. 4). It was sometimes difficult to find distinct lumens between ductular cells, apart from the structures that resembled canaliculi. A basal lamina, FIG. 5. Case 1. Electron micrograph of proliferated bile ductule. One epithelial cell has a large nucleus with distinct nucleolus, fat droplets (F), and glycogen. This cell is surrounded by incomplete basal lamina. The other cells contain well-developed tonohlaments {arrows) (X 10,500).

6 Vol. 79 No. 3 DUCTULAR TRANSFORMATION Enzyme Histochemistry In the normal liver, the small bile ducts and ductules have negative reaction for glycogen and for G6P and ALP activities.12 GGTP activity was present on the inner surface and in the apical cytoplasm. The cytoplasm of duct epithelial cells had diffuse positive staining for SD and DPND activities: the former activity was weaker and the latter stronger than in hepatocytes. continuous with that of the neighboring, more typical, biliary-appearing cells, was present but was often poorly developed. In addition to the large cells, frequently there were biliary-appearing cells, having ultrastriictural features similar to those of hepatocytes. The hepatocytic cords and the bile ductules sometimes shared the same lumen (Fig. 6). There were a few epithelial cells of ductules containing alcoholic hyalin. ^ ''-- v^.-r^v"'."-^' 331 'X-A<- v 'OS - - v - \, '-6/..,* ; - rj ' r^ ^.-"A. FIG. 6. Case 5. Electron micrograph showing continuous transition of a cord of hepatocytes to a bile ductule. The cells share the same lumen and form a structure similar to the canal of Hering. One cell (I) has some hepatocellular characteristics, including prominent nucleolus, large mitochondria, and bile canaliculus. This cell is separated by a more typical ductular cell from the hepatocyte. Note the basal lamina (arrows) extends from the ductule to hepatocyte column (X7.000).

7 332 UCHIDA AND PETERS A.J.C.P. March 1983 w 'I m FIG. 7. Bile ductules trapped in densefibrosishave glucose-6-phosphatase activity as indicated by lead precipitates in the cytoplasm (arrows). Lead phosphate (X 660) (case 3). While most epithelial cells of proliferated bile ductules had negative staining for G6P activity, some had positive staining, though fainter than in hepatocytes (Fig. 7). All cells forming ductules gave a positive, discrete staining reaction for SD and DPND activities throughout the whole cytoplasm but of variable intensity. SD activity apparently was much stronger in many epithelial cells of ductules than in normal small bile ducts. ALP activity rarely was found and then was only faintly positive on the luminal surface of proliferated ductules. Areas of fibrosis, blood vessels, and some inflammatory cells stained positively. GGTP activity was positive extensively and strongly on the inner surface but only weakly in the cytoplasm of epithelial cells of proliferated ductules. The staining was somewhat stronger on the inner surface of the proliferated ductules than in the normal ducts and the bile canaliculi of hepatocytes. Discussion We have shown that it is reasonable to assume the majority of the proliferated ductules appear to have been derived from transformation of hepatocellular cords for the following reasons: transitional or intermediate forms of hepatocytes of variable degrees were recognized by light and electron microscopy; markers of hepatocytes, such as glycogen and G6P activity, often were demonstrated; and alcoholic hyalin occasionally was recognized in the epithelial cells of bile ductules. The elongated shape of proliferated ductules, having an average of 11.2 nuclei, may reflect transformation of hepatic cords. Further evidence of ductular transformation of hepatic cords is the observation that ductular cells often contain alpha-1 antitrypsin globules in alpha-1 antitrypsin deficiency disease (unpublished data). It seems that ductular transformation starts via the canal of Hering, and epithelial cells go through three stages of evolution, i.e., hepatocytes, intermediate cells, and established biliary cells. Sometimes, a single ductule contained all three kinds of cells. Many proliferated ductules were in a transitional stage to the more typical ductules. In the course of transformation, the ductular epithelial cells seem to gradually lose the original hepatocellular markers and characteristics. Some bile ductular cells could have originated from the preexisting bile ducts. The stimulus for transformation of hepatic cords to bile ductules is not understood well, but the most active and extensive proliferation of bile ductules almost always is associated with necrotizing and fibrotic changes involving dropout of hepatocytes, accumulation of clusters of macrophages with DiPAS-positive granules, fresh collagen deposition, and some inflammatory reaction. The new channels produced by the transformation seem to be necessary to allow communication between the bile canaliculi of remaining, or regenerating, hepatocytes and the interlobular bile ducts in the portal tracts, when the periportal hepatocytes drop out extensively. With further decrease of the necrosis and progressive fibrosis of the periportal area, the proliferated ductules appear to decrease in number, because the communicating channels may be rearranged. Bhathal and Christie 1 demonstrated that the proliferated bile ductules connected the bile canaliculi with interlobular bile ducts, using Indian ink injection and intravital fluorescent methods in 2-naphthyl isothiocyanate administration in guinea pigs. The functional activity of absorption and secretion of the bile ductules was stressed by Sasaki and associates 8 and Sternlieb.'' We believe that the proliferated ductules also have this function, as suggested by the pinocytotic vesicles, well-developed Golgi apparatus, and numerous microvilli on the luminal surfaces. The intense staining of GGTP activity on the luminal surface is further supportive evidence as GGTP seems to play an important role on amino acid transport through the cell membrane. Similar bile ductules frequently are found in the livers of patients with primary biliary cirrhosis and viral hepatitis, again associated with periportal hepatocellular necrosis, inflammation, and collagen. However, the proliferated bile ductules may be multiplication of the interlobular bile ducts, particularly if they are confined to within the original portal tracts. The early phase of mechanic biliary obstruction also may be due to multipli-

8 Vol. 79 No. 3 DUCTULAR TRANSFORMATION 333 cation of bile ducts, since the proliferated ducts in obstruction usually have distinct lumena and mostly are confined to within the portal tract. However, the ductular transformation of hepatic cords must be looked for in each disease before it can be concluded that the results of these studies are applicable universally to explain bile ductular proliferation. The proliferated bile ductules in alcoholic liver diseases may have accompanying inflammatory exudation, including neutrophils. The neutrophils are sprinkled in the areas with proliferated ductules and are oriented towards them, even in the ductular lumens as seen in cases 1 and 2. Clinical and laboratory data did not support complicating biliary obstructive disease, and the lack of edema in the area of ductular proliferation seems to exclude this possibility. Since neutrophilic exudation appears to be more frequent in the more recently formed ductules, it could be either a complicating nonspecific reaction or an initiating component of ductular transformation. References 1. Bhathal PS, Christie GS: A fluorescence microscopic study of bile duct proliferation induced in guinea pigs by alpha-naphthyl isothiocyanate. Lab Invest 1969; 20: Gall EA: Posthepatic, postnecrotic, and nutritional cirrhosis. A pathologic analysis. Am J Pathol 1960; 36: Hunt RD: Microscopic histochemical methods for the demonstration of enzymes. Selected histochemical and histopathological methods. Edited by S. W. Thompson, Springfield, Charles C. Thomas Publisher, 1966, pp Leduc EH: Cell modulation in liver pathology. J Histochem Cytochem 1955; 7: Masuko K, Rubin E, Popper H: Proliferation of bile ducts in cirrhosis. Arch Pathol 1964; 78: Milne LS: The histology of liver tissue regeneration. J Pathol Bacterid 1909; 13: Roman B: Acute yellow atrophy of the liver. Arch Pathol 1927; 4: Sasaki H, Schaffner F, Popper H: Bile ductules in cholestasis: Morphologic evidence for secretion and absorption in man. Lab Invest 1967; 16: Schaffner F, Popper H: Electron microscopic studies of normal and proliferated, bile ductules. Am J Pathol 1961; 38: Steiner JW, Carruthers JS: Studies on the fine structure of the terminal braches of the biliary tree. I. The morphology of normal bile canaliculi, bile pre-ductules (Ducts of Hering) and bile ductules. Am J Pathol 1961; 38: Sternlieb I: Functional implications of human portal and bile ductular ultrastructure. Gastroenterology, 1972; 63: Uchida T, Miyata H, Shikata T: Human hepatocellular carcinoma and putative precancerous disorders. Their enzyme histochemical study. Arch Pathol Lab Med 1981; 105: Ungar H, Goldberg MG: The nature of bile duct proliferations in the liver following ingestion of DL-ethionine. Lab Invest 1959;8: Yokoo H, Minick OT, Batti F, Kent G: Morphologic variants of alcoholic hyalin. Am J Pathol 1972; 69:25-40