Histometric and Serial Section Observations of the Intrahepatic Bile Ducts in Primary Biliary Cirrhosis

GASTROENTEROLOGY 761326-1332, 1979 Histometric and Serial Section Observations of the Intrahepatic Bile Ducts in Primary Biliary Cirrhosis YASUNI NAKANUMA and GOROKU OHTA Second Department of Pathology, Kanazawa University, School of Medicine, Kanazawa, Japan Histometric examinations, based on the assumption that hepatic arterial branches and bile ducts run parallel within the portal tracts, suggest that in primary biliary cirrhosis bile ducts with a lumen (the smallest diameter between the subepithelial basal membranes) below 70-80pm are destroyed The smaller the ducts, the more they are destroyed Extensive destruction of the ducts was seen more frequently in the nonfbrotic stage of primary biliary cirrhosis than in later stages Serial sections of the intrahepatic bile ducts in primary biliary cirrhosis revealed three types of periductal lesions preceding the disappearance of bile ducts (A) periductal cellular reaction including features of chronic nonsuppurative destructive cholangitis, (B) periductal edema, and (C) periductal fibrosis In the nonfibrotic stage, types A and C were frequent, whereas in the fibrotic stage types A and B were increased, and type C was predominant in the cirrhotic stage In primary biliary cirrhosis (PBC), destructive cholangitis, followed by disappearance of the interlobular bile ducts, is considered to be the primary lesion of unknown etiology - The present investigations, using histometric methods, were undertaken to determine the size and distribution of the destroyed bile ducts, and furthermore, with the help of serial sections, to detect the character of periductal alterations preceding bile duct destruction and the distribution in different stages of PBC Received April 7, 1978 Accepted January 6, 1979 Address requests for reprints to Y Nakanuma, MD, Second Department of Pathology, Kanazawa University School of Medi- cine, Takaramachi 13-1, Kanazawa 920, Japan This study was supported by a Grant from the Japanese Education Ministry, and Research grants for specific diseases from the Japanese Health and Welfare Ministry We are very grateful to Dr Hans Popper in New York for revision of this paper and the chairmen of many departments in Japan for permission to examine specimens of livers with primary biliary cirrhosis 0 1979 by the American Gastroenterological Association OOlS-5085/79/061326-07$0200 Materials and Methods Twenty-seven livers from patients with clinical and histologic diagnosis of PBC (20 autopsy and 7 wedge biopsy specimens) were obtained from many institutions in Japan, including our department All patients were Japanese (31-66 yr of age), and 23 were female Jaundice was present in all except 2 asymptomatic cases, and extrahepatic bile duct obstruction was ruled out by laparotomy and/or autopsy in all patients Cholestasis was seen histologically in 23 cases Mitochondrial antibodies (AMA) were found in 14 of 17 patients (824%) Patients in whom AMA was neither detected nor examined were diagnosed as PBC on the basis of clinical, laboratory, and morphologic findings The livers showed either chronic nonsuppurative destructive cholangitis (CNSDC) and/or marked reduction in the number of interlobular bile ducts? All liver specimens were fixed in 10% formalin and embedded in paraffin Primary biliary cirrhosis was staged on the basis of histologic changes Livers showing no periportal fibrosis and only destructive cholangitis were designated as the nonfibrotic stage; slight to severe fibrosis with ductular proliferation and destruction as the fibrotic stage; and regenerative nodules throughout the liver as the cirrhotic stage The nonfibrotic stage largely corresponds to Scheuer s stage 1, but also includes some cases with focal ductular proliferations and without septa formation, and the fibrotic stage corresponds to Scheuer s stage 3, but also includes some cases of Scheuer s stage 2 with periportal fibrosis The cirrhotic stage corresponds to Scheuer s stage 4R 1 Histometric studies were performed on liver sections of 26 cases of PBC (6 wedge biopsy and 20 autopsy specimens) and 35 control autopsy livers The 26 cases of PBC included 4 of the nonfibrotic stage (wedge biopsy specimens), 12 of the fibrotic stage (2 wedge biopsy and 10 autopsy specimens), and 10 of the cirrhotic stage (autopsy specimens) The controls had an age distribution similar to the PBC patients and consisted of 20 cases with normal liver, 10 with chronic hepatitis, and 5 with cirrhosis showing mild or no proliferation of bile ducts or ductules For histometric analysis, three sections obtained from the right lobe, the left lobe, and the hilar portion were used in 15 autdpsy cases of PBC and in all

lune 1979 BILE DUCT DESTRUCTION IN PRIMARY BILIARY CIRRHOSIS 1327 control livers; two sections from the right and left lobe in 4 autopsy cases of PBC; and one section from the right lobe in one autopsy case of PBC; three sections selected at random from many serial sections from the available wedge biopsies of PBC The area of each section was about 5 cm2 in autopsy and 2 cm2 in wedge biopsy specimens The sections were stained with H & E In control livers, the bile duct and the artery of approximately equal size usually run parallel in the portal tract (Figure 1) By contrast, bile ducts were often absent in the vicinity of the arteries in PBC (Figure 2) since the bile ducts had apparently been destroyed and replaced by lymphocytic infiltration, granulation tissue, or fibrous tissue It was, therefore, assumed that the level and extent of bile duct destruction of livers with PBC could be expressed objectively by estimating the size and number of the arteries as a marker for histometric examination of destroyed bile ducts in the portal tracts The level and extent of bile duct destruction might be expressed quantitatively by calculation of the ratio of the number of arteries accompanying the corresponding bile ducts to the number of all arteries found in PBC livers, and by comparison of the ratio with that of controls Since a portal tract occasionally contained two or three bile ducts and arteries respectively, it was necessary to establish criteria for selection of the bile duct and artery that constituted a pair in the portal tract The largest artery and the largest bile duct in the portal tract were regarded as a pair, when the duct was located within a distance three times the external diameter of the artery measured from the outside of its media Preliminary observations suggested that such a distance was suitable for selection of a pair A pair of smaller sized arteries and ducts, if present in the portal tract, was selected similarly The rate of parallelism, that is, the proportion of the number of arteries (lumen > 35 pm) having a corresponding bile duct (lumen > 25 pm) to all arteries (lumen > 35 pm) present in all liver sections in each case was calculated The results reflected the degree and extent of bile duct destruction in PBC livers The shortest diameter between the outsides Figure 2 There are no bile ducts in the vicinity of arteries Au- topsy case of PBC (cirrhotic stage) (H & E, x 130) of the tunica media was taken as the parameter of the lumen of the artery, and the shortest diameter between the subepithelial basal membrane as the parameter of the bile duct Where data were subjected to statistical tests, standard deviation and P-values were determined Furthermore, the ratio of the lumen of the bile duct to that of the associated artery selected by the above mentioned criteria was calculated in several sections from each case and the mean value of the ratios for each case is shown in Figure 3 The ratios of both structures of any size in controls were relatively constant (07-08) An artery of any size was largely associated with a bile duct with a lumen 07-08 times that of the artery in every portal tract of control livers Therefore, the lumen of arteries not associated with a corresponding bile duct was thought to represent the approximate size of destroyed bile ducts in livers with PBC 2 Bile duct destruction was searched for in serial sections 10 09 t Figure 1 A portal tract in normal liver Note the presence of a bile duct and artery next to each other (H & E, x 260) NI CH 1,c External caliber ratio of bile duct/artery in control livers NL = normal liver (20 cases); CH = chronic hepatitis (10 cases) LC = liver cirrhosis (5 cases)

1328 NAKANUMA AND OHTA GASTROENTEROLOGY Vol 76, No 6 Fig 4-A Artery (35-54~) Fig 4-B Artery (55-74~) 100 % 100 7 % 50 - y 3 7 L 50 _ y * * i I * * er,e 8% NL CH l,c PBC PBC PBC PBC (total)(n) (F) (C) NL 4 f * CH LC PBC PBC PBC PBC (total) (N) (F) (C) Fig 4-C Artery (75-94p) Fig 4-l) Artery ( >95p) 100 50-50 i _ x + 1 ; * NL CH LC PBC PBC PBC PBC (total)(n) (F) (Cl NL CH LC PBC *PBC PBC PBC (total) (N) (F) (C) Figure 4 Rate of parallelism of arteries with associated bile ducts in each group of arterial size (N = nonfibrotic stage; F = fibrotic stage; C = cirrhotic stage) NL = normal liver (20 cases) CH = chronic hepatitis (10 cases); LC = liver cirrhosis (5 cases) There are no arteries above 95 pm diameter in the studied cases of the nonfibrotic stage PBC in 12 cases of PBC 4 cases of the nonfibrotic stage (wedge biopsies), 4 cases of the fibrotic stage (2 wedge biopsies and 2 autopsy livers), 4 cases of the cirrhotic stage (autopsy livers), and in 1 case of a normal autopsy liver The specimens were obtained from the right lobe, the left lobe, and hilar portion, respectively, in all au- topsy cases; 2 specimens from the right lobe and the left lobe of a surgical biopsy and a single specimen in five surgical biopsy specimens About 200 serial sections were cut from each paraffin block; a total number of 600 serial sections were made in each autopsy case (about 5 pm in thickness) and every fourth section was

tune 1!379 BILE DUCT DESTRUCTION IN PRIMARY BILIARY CIRRHOSIS 1329 Figure 5 A Bile duct destruction with periductal cellular reac- Figure 6 A Bile duct destruction with periductal edema A hytion A hyperplastic, and degenerating bile duct (ar- perplastic bile (arrow) is seen in the edematous conrow) is surrounded by a Iymphoid aggregate It dis- nective tissue showing mild infiltration of Iymphoid appears within the lesion in B and C The distance cells The duct disappears as seen in B and C The dishetwecn A and B is 20 pm, and B and C is 20 Wm rc- tance between A and B is 20 pm, and B and C is 20 am, spectivcly Case 6, Wedge biopsy H & E, x 260 respectively Case 5 Wedge biopsy H & E, x 260 stained with H & E The histologic changes associated of the portal tracts and had a lumen of > 20 pm Septal with bile duct disappearance were determined by bile ducts were larger than the interlobular bile ducts, sketching and tracing all bile ducts seen in the sections and were only seen in the central part of portal tracts Tlhe intrahepatic biliary tree was classified as follows It is sometimes difficult to distinguish morphologically Bile ductules were considered tubular structures in the interlobular from septal bile ducts Therefore, both peripheral zone of the portal tract with lumen of c 20 types of ducts were called bile duct and classified acpm; interlobular bile ducts were in the peripheral zone cording to their diameter

1330 NAKANUMA AND OHTA GASTROENTEROLOGY Vol 76, No 6 Results The rates of parallelism of arteries with their corresponding bile ducts are presented in four groups on the basis of the size of the lumen of the arteries (Figure 4) In control liver specimens, the rate of parallelism is rather constant, in that approximately 70-80% of the arteries are accompanied by bile ducts The rate is appreciably lower in liver specimens with PBC, but it increased with the size of the arteries Only 74 & 91% (mean f standard deviation) of arteries with diameters of 35-54 pm were accompanied by bile ducts (Figure 4A); of arteries 55-74 pm in diameter, 219 f 171% (Figure 48); of arteries 75-94 pm in diameter, 336 f 263% (Figure 4C); and of arteries larger than 95 pm in diameter, 753 k 150% (Figure 40) The values for the first three groups of arteries in PBC were significantly lower than that of control liver specimens (P < OOl), but that of the last group (>95 pm) was almost equal to that of control liver specimens (P > 005) Bile ducts which originally had run parallel to the arteries with a lumen below 95 pm were extensively destroyed, and their presumed original lumen was calculated to be below 70-80 pm (Figure 3) The smaller the ducts, the more were destroyed Although there was a significant difference in the rate of parallelism of bile ducts to the arteries of both the 35-54 pm and the 55-74 pm groups between the nonfibrotic stage and fibrotic stage (P < OOl), the numbers of cases in the nonfibrotic stage are too small to be certain of this difference There was no difference in the rate of parallelism between the fibrotic and the cirrhotic stages Three types of lesions of the bile ducts preceded their disappearance, as indicated by the serial sections Bile duct destruction with predominant cellular reaction was designated type A (Figure 5A-C) The bile duct was destroyed within an aggregation of small round cells occasionally associated with epithelioid granulomata The changes were those of CNSDC described by Rubin et al Bile ducts disappearing within an epithelioid granuloma were also regarded as belonging to this type In type B, the destruction of bile ducts was associated with periductal edema The bile ducts disappeared within edematous connective tissue, usually accompanied by a slight degree of fibrosis or round cell infiltration (Figure 6A-C) In type C, the destruction of ducts was accompanied by fibrosis without inflammatory changes (Figure 7A-C) The epithelium of the bile ducts frequently showed proliferative changes (papillary intraluminal projections or stratification) and regressive changes (acidophilic or focal necrotic changes or rupture) above or below the area of destruction These alterations were found more often in types A and B than in type C Figure 7 A Bile duct destruction with periductal fibrosis Bile duct (arrow) in fibrous connective tissue B Degeneration of epithelium C The bile duct is no longer seen The distance between A and B is 20 pm, and B and C is 20 pm, respectively Case 2, Wedge biopsy H & E, x 260 The total number of destroyed bile ducts and the frequency of each of the three types of bile duct destruction are shown in Table 1 All three types of bile duct lesions were seen in almost all cases, except in two cases of the fibrotic stage In the nonfibrotic stage, type C was predominant (mean value of four cases 478%); type A was second in fre-

lune 1979 BILE DUCT DESTRUCTION IN PRIMARY BILIARY CIRRHOSIS 1331 quency In the fibrotic stage, type A (mean value of 4 cases 551%) and type B (306%) were commonly found, and in 2 cases (wedge biopsies) type C was absent In the cirrhotic stage, type C (mean value of 4 cases 765%) was again the most prominent lesion Discussion In order to establish the level and extent of the bile duct destruction in livers with PBC, a method for histometric examination of the intrahepatic bile ducts was applied It indicated a considerable variation in the extent of destruction of medium-sized bile duct in all stages of PBC, but less variation in that of smaller bile ducts These observations suggest that the degree of destruction of medium-sized bile ducts differs with the portion of the liver examined, while the destruction of smaller bile duct is equally extensive throughout the liver in almost all cases There is no significant difference between the cirrhotic and fibrotic stages in the parallelism of the arteries with the associated bile ducts However, a statistical difference between the nonfibrotic and fibrotic stages was found, possibly because of the small number of cases of the nonfibrotic stages available Nevertheless, active and widespread destruction of bile ducts reflecting progression of the disease can be postulated to occur in the nonfibrotic stage, with or without clinical manifestations Therefore, the degree of bile duct destruction seems unlikely to parallel hepatocytic destruction in PBC If this is the case, the development of cirrhosis in PBC cannot be explained only by the mechanism which causes the destruction of intrahepatic bile ducts, but may require additional factors, such as autoimmune destruction of hepatocytes, ll copper deposition, chronic cholestasis4 and intrahepatic circulatory disturbance - Serial sections showed that destruction and disappearance of bile ducts are associated with three types of change a cellular reaction (type A), edema (type B), and mild fibrosis (type C) These observations confirmed that chronic nonsuppurative destructive cholangitis, similar to our type A, is a characteristic feature of PBC, but this lesion is not the only type of bile duct destruction, even in the nonfibrotic stage of PBC The three types of destruction were found in every stage of PBC, and no cholangitic lesion specific for the stages of PBC is identified, at least in our classification The simultaneous presence of two or three types of bile duct destruction indicates that all three may be caused by the same etiologic factor, or factors, and may only represent different phases of bile duct damage The frequency of type C, periductal fibrosis, in all cases of the nonfibrotic stage and the relative increase in the number of type A and B in the fibrotic stage suggests Table 1 Frequency of Each Type of Bile Duct Destruction in the Studied Case of PBC Among all types of destruction in each case (%) Type A Type B Type C Nonfibrotic stage Case 1 (WB) 370 152 478 Case 2 (WB) 231 154 615 Case 3 (WB) 429 143 429 Case 4 (WB) 478 130 391 Mean of 4 cases 377 145 478 Total numbers of destroyed bile duct in each case 46 13 14 23 Fibrotic stage Case 5 (WB) 333 667 0 16 Case 6 (WB) 870 120 0 8 Case 7 (A) 629 219 152 151 Case 8 (A) 373 216 412 51 Mean of 4 cases 551 305 141 Cirrhotic stage Case 9 (A) Case 10 (A) Case 11 (A) Case 12 (A) Mean of 4 cases 178 164 656 67 56 176 769 108 96 87 817 104 72 112 816 125 101 135 765 Stagxng of PBC is according to Ohta and Nakanuma WB wedge biopsy, A autopsy, type A bile duct destruction with cellular reaction, type B bile duct destruction with edema, and type C bile (duct destruction with mild fibrosis

1332 NAKANUMA AND OHTA GASTROENTEROLOGY Vol 76, No 6 that the initial lesion of the bile ducts leading to destruction is a disruption of the duct epithelium with mild periductal fibrosis The cellular inflammatory and/or edematous changes of type A and B might be secondary phenomena The type B lesion, periductal edema, may represent biliary fluid imbibition after initial bile duct injury, and type A may result from an autoimmune mechanism after initial destruction of the duct epithelium or from an unknown autoantigen presumably derived from biliary fluid A transition from type B to A, or vice versa, is possible, too, but requires further investigations The predominance of type C in the cirrhotic stage, however, suggests another possibility, namely that the inflammatory or edematous changes around the damaged bile ducts in types A and B regress completely during the long course of the disease, but produce periductal fibrosis, the type C lesion The morphogenesis of the type C lesion in the cirrhotic stage of PBC possibly differs from that of this lesion in the noncirrhotic stage The pathogenesis of all types of described bile duct lesions, remains unknown 1 2 References Sherlock S Primary biliary cirrhosis In Progress in Liver Diseases Edited by H Popper, F Schaffner Volume V New York and London, Grime & Stratton, Inc, 1976, p 559-574 Schaffner F Primary biliary cirrhosis Clin Gastroenterol 4351-366, 1975 3 4 5 6 7 8 9 10 11 12 13 14 15 Ohta G, Nakanuma Y Primary biliary cirrhosis with emphasis on the intrahepatic bile duct Intern Med (Jpn) 39405-412, 1976 Baggenstoss AH, Foulk WT, Butt HR, et al The pathology of primary biliary cirrhosis with emphasis on histogenesis Am J Path01 42259-276, 1964 Rubin E, Schaffner F, Popper H Primary biliary cirrhosis Am J Path01 46387-407, 1964 Hadziyannis S, Scheuer PJ, Feizi T, et al Immunological and histological studies in primary biliary cirrhosis J Clin Path01 2395-98, 1970 Ichida F Primary biliary cirrhosis and related hepatic diseases Jap J Gastroenterol 721428-1430, 1975 Scheuer PJ Liver biopsy interpretation Baltimore, Williams & Wilkins Co, 1973, p 33-38 Thomas HC, Potter BJ, Sherlock S Is primary biliary cirrhosis an immune complex disease? Lancet 21261-1263, 1977 Dudley FJ, Fox RA, Sherlock S Cellular immunity and hepatitis-associated Australia antigen liver disease Lancet 1723-726, 1972 Ohta G, Nonomura A Nishimura 1 Leucocyte migration inhibition with inner and outer membranes of mitochondrial and insoluble hepatocyte membrane prepared from rat liver in patients with chronic hepatitis and cirrhosis Clin Exp lmmunol 26491-504, 1976 Deerling TB, Dickson ER, Flemming CR, et al Effect of d-penicillamine on copper retention in patients with primary biliary cirrhosis Gastroenterology 721208-1212, 1977 Okudaira M Alcoholism and the liver with regards to the morphological changes of the liver Acta Hapat Jap 7296-299 1966 Popper H, Rubin E, Krus S, ct al Postnecrotic cirrhosis in alcoholics Gastroenterology 39679-685, 1960 Popper H, Elias H Histogenesis of hepatic cirrhosis studied by the three-dimentional approach Am J Path01 31405-442, 1955