J. Pathol. 186: 429–433 (1998)

APOPTOSIS AND CELL PROLIFERATION INBILIARY ATRESIA

1,2, 1*, 1, 2, 2, 2 1

1Department of Pathology, Tohoku University School of Medicine, Sendai, Japan2Department of Pediatric Surgery, Tohoku University School of Medicine, Sendai, Japan

SUMMARY

Biliary atresia (BA), which is thought to result from progressive destruction of the bile ducts by a necroinflammatory process, is themost common cause of obstructive jaundice in infancy. Abnormalities in the cell turnover of remodelling ductal plates are considered oneof the important aetiological factors in this disorder, but little work has been done on this topic. Programmed cell death or apoptosis wastherefore examined by TdT-mediated dUTP biotin nick end labelling (TUNEL) and cell proliferation by Ki67 immunostaining in 34cases of BA. The results were compared with normal control liver (five cases) and congenital dilatation of the bile ducts (CDB, five cases)in order to study the cell turnover or tissue dynamics of BA. The TUNEL labelling index (LI) in bile ducts (48·9&13·2 per cent) wassignificantly higher than that of the control normal liver (3·6&2·8 per cent) and of CDB (2·5&5·1 per cent). The Ki67 LI in the bileducts of BA (15·0&5·57 per cent) was also significantly higher than that of CDB (8·6&5·4 per cent). No significant differences of theTUNEL and Ki67 LIs in hepatocytes were, however, observed between BA, CDB, and normal liver. The TUNEL LI was significantlyhigher than the Ki67 LI in the bile ducts of BA. BA is therefore associated with increased and disorganized cell turnover of the bile ducts,which is related to malformation of the ductal plate or abnormal bile duct development. ? 1998 John Wiley & Sons, Ltd.

KEY WORDS—biliary atresia; liver; apoptosis; cell proliferation; TUNEL; Ki67

INTRODUCTION

Biliary atresia (BA), the most common cause ofobstructive jaundice in infancy,1,2 is thought to resultfrom progressive destruction of the bile ducts through anecroinflammatory process.1,3 This affects not only partor all of the extrahepatic ducts, but also the intrahepaticbranches of the biliary tree, which subsequently resultsin paucity of interlobular ducts, or ductopenia.1,2 Theintrahepatic bile ducts have been demonstrated to origi-nate from the ‘ductal plates’, which have been proposedto be derived from periportal hepatocytes through trans-formation or metaplasia of hepatocytes into bileductules.2–5 The aetiology of BA still remains unknown,but ‘the ductal plate malformation’, or insufficientremodelling of a persistent ductal plate, has been con-sidered to play an important role in the development ofBA.1–5

Apoptosis, or programmed cell death, is important incell turnover in both physiological and pathologicalconditions of many organs.6–8 In particular, the home-ostasis between apoptosis and cell proliferation shapesorgans during mammalian morphogenesis and disordersin this process result in various congenital abnor-malities including atrophy, hyperplasia, regression, andothers.7,8 Terada and Nakanuma recently demonstratedthat remodelling during human intrahepatic bile ductdevelopment occurs by means of apoptosis.9 It thereforebecomes very important to examine cell turnover in theliver with BA, in order to study the possible involve-

CCC 0022–3417/98/120429–05$17.50? 1998 John Wiley & Sons, Ltd.

ment of impaired regulation of apoptosis and cell pro-liferation in the development of this congenital biliarydisorder.

In this study, we examined cell proliferation by Ki67immunostaining and programmed cell death or apop-tosis by the 3*-OH nick end labelling technique, orthe TdT-mediated dUTP-biotin nick end labelling(TUNEL) method,7 in 34 cases of BA and compared theresults with those of normal infant liver and congenitaldilatation of the bile ducts (CDB), in which dilatation ofthe extrahepatic biliary duct was present, but the intra-hepatic branches were not affected, except for reactiveductal proliferation.10

MATERIALS AND METHODS

Liver specimens

Surgical pathology specimens of BA (n=34, 14 maleand 20 female) were retrieved from the pathology files ofTohoku University Hospital, Sendai, Japan. All biopsyspecimens were obtained from the right anterior segmentof the liver during hepaticojejunostomy or Kasai’s pro-cedure.11 The ages of the patients ranged from 12 to 293days after birth (mean age 64·6&44·4 days). Five surgi-cal pathology specimens of CDB [two male and threefemale, mean age 108·0&91·4 days (60–240 days afterbirth)] were also retrieved from the pathology files asabove. They were also biopsy specimens, obtained fromthe right anterior segment of the liver during curativesurgery.

Five autopsy specimens of infant livers withouthepatobiliary disease [two male, three female, mean age

*Correspondence to: Hironobu Sasano, MD, Department ofPathology, Tohoku University School of Medicine, 2-1 Seiryou-machi,Sendai, 980, Japan. E-mail: hsasano@patholo2.med.tohoku.ac.jp

Received 27 November 1997Revised 24 February 1998

Accepted 6 July 1998

430 N. FUNAKI ET AL.

33·60&65·2 days (0–150 days after birth)] were alsoretrieved from the pathology files of Tohoku UniversityHospital. The specimens were all fixed in 10 per centneutral formalin for 24–48 h at room temperature andembedded in paraffin wax.

Immunohistochemistry

Details of the immunohistochemical procedures havebeen previously reported.8,12 After deparaffinization,autoclave treatment at 105)C for 5 min was employedfor antigen retrieval. The primary antibody used in thisstudy was the monoclonal antibody MIB-1 (Immuno-tech, Marseille, France). Immunostaining was per-formed by the Strept-avidin–biotin method using aHistofine Kit (Nichirei, Tokyo, Japan). The negativecontrol of immunostaining was 0·01 PBS or normalmouse IgG instead of the primary antibodies. Immuno-reactivity was not observed in these tissue sections.

TUNEL

Double-strand DNA breaks were detected byTUNEL according to the method of Gavrieli et al.7 andSasano et al.,13 with some modifications. We used arecently developed Apoptosis in situ Detection Kit(Wako, Osaka, Japan) for TUNEL. Deparaffinizedsections were washed with distilled water and treatedwith Protein Digestion Enzyme for 5 min at 37)C. Afterwashing with three changes of 0·01 PBS, sections weretreated with TdT solution, incubated with 3 per centhydrogen peroxide for 5 min to block endogenous per-oxidase activity, and then treated with peroxidase-conjugated antibody for 10 min at room temperature.After washing in 0·01 PBS, nick end labelling wasvisualized by immersing reacted sections in 3*,3*-diaminobenzidine solution with 0·006 per cent hydro-gen peroxide and counterstaining with methyl green. Asa negative control, the tissue sections were incubatedwith TdT buffer that did not contain the enzyme. Forthe positive control, tissue sections were treated withDNase I prior to treatment with TdT.

Counting and statistical analysis

The images were directly captured through a digitalCCD camera (ProgRes 3012 PPC, Kronton ElectronikCo. Ltd., Encring, Germany) with a PRI-Macintoshinterface board attached to an operating light micro-scope (Carl Zeiss Co./Ltd., Jena, Germany). Resolutionwas set at 998#774 pixel. The images were subsequentlytransferred to a Power Macintosh 9500/120 personalcomputer-controlled operating system and processedwith Macintosh software Adobe Photoshop 3.OJ.

Screen images for the examination were selectedrandomly by two of us (NF and HS). Ki67-immunopositive cells and TUNEL-positive cells wereevaluated for each section in the screen images. At least200 cells were independently counted in each specimenby NF and HS. Intra- and inter-observer differenceswere less than 5 per cent and disconcordant cases were

? 1998 John Wiley & Sons, Ltd.

re-evaluated simultaneously by the two observers. Thepercentage of positive cells among the counted cells wasdesignated as the labelling index (LI) and the resultswere presented as mean&standard error. In the caseswith normal liver, CDB, and BA, Ki67 and TUNEL LIswere obtained separately in hepatocytes and bile ducts.Ki67 and TUNEL LIs were separately evaluated in thesethree structures when discernible. Overall comparisonwas performed with the one-way analysis of variance bythe Kruskal–Wallis rank-sum test. The Bonferronimethod was applied for the significance of simultaneousmultiple comparisons.

RESULTS

The results are summarized in Table I and Fig. 1.

TUNEL

Staining was confined to the nucleus (Fig. 2A). Therewere no significant differences in TUNEL LI in hepato-cytes between normal liver and cases of CDB and BA,but the TUNEL LI in intrahepatic bile ducts(48·9&13·2 per cent) was significantly higher than thatof the normal liver (3·6&2·8 per cent) and of CDB(2·5&3·1 per cent) (p<0·0001) (Fig. 1A).

Ki67

Immunoreactivity was detected only in the nucleus(Fig. 2B). There were no significant differences of Ki67LI in hepatocytes between normal liver and cases ofCDB and BA. The Ki67 LI in biliary ducts of BA(15·0&5·57 per cent) was significantly higher than thatof CDB (8·6&5·4 per cent) (p<0·05) (Fig. 1B). The Ki67LI in BA tended to be higher than that in the normalcontrol liver (10·1&9·8 per cent), but the differencesdid not reach statistical significance (Fig. 1B). In thebile ducts of BA, the TUNEL LI was significantlyhigher (p<0·01) than the Ki67 LI. There were no signifi-cant differences of the Ki67 and TUNEL LI in the

Table I—Summary of the results

Ki67 LI (%) TUNEL LI (%)

Normal liver (n=5)Ducts 10·1&9·8 3·6&2·8Hepatocytes 3·4&2·4 2·5&2·3

Congenital dilatation ofthe bile ducts (n=5)

Ducts 8·6&5·4 2·5&3·1Hepatocytes 4·2&1·8 1·9&1·9

Biliary atresia (n=34)Ducts 15·0&5·57 48·9&13·2Hepatocytes 4·3&2·6 2·2&2·5

Ki67 LI (labelling index): percentage of Ki67-positive cells.TUNEL LI (labelling index): percentage of TUNEL-positive cells.

J. Pathol. 186: 429–433 (1998)

431CELL TURNOVER IN BILIARY ATRESIA

hepatocytes of the three groups, or in the ductal cells ofnormal liver or of CDB.

DISCUSSION

The importance of cell loss through apoptosis hasbeen increasingly recognized in both normal and patho-logical development and the balance between cell pro-liferation and cell elimination is now considered veryimportant in regulating this process. Balanced cell pro-liferation and apoptosis have recently been shown to beinvolved in the normal development of intrahepatic bileducts and hepatocytes in the human.9 However, cellproliferation and apoptosis have not been examined inthe liver with BA, in which impairment of normaldifferentiation into definitive adult bile ducts is believedto be involved,14 except for an immunohistochemical

? 1998 John Wiley & Sons, Ltd.

study of proliferating cell nuclear antigen (PCNA), anauxiliary protein for DNA polymerase ä.15,16

In this study, we examined cell proliferation usingKi67 immunostaining and programmed cell death by the3*-OH nick end labelling or TUNEL technique. We usedKi67 immunostaining because Hall et al. have shownthat the overexpression of PCNA in cells is not neces-sarily associated with actual cell proliferation.17 DNAfragmentation studied by TUNEL can be detected inmodes of cell death other than apoptosis and TUNELpositivity is not necessarily a specific marker of apop-tosis, but TUNEL does simplify the identification ofapoptotic cells in tissue sections.7,13 Two of us indepen-dently evaluated Ki67- and TUNEL-positive cells in the

Fig. 1—(A) Percentage of TUNEL-positive cells (TUNEL LI) in bileducts in cases of biliary atresia (BA), congenital dilatation of the bileducts (CDB), and normal control liver (NCL). The TUNEL LI in BAwas significantly higher than that of CBD and normal control liver. (B)Percentage of Ki67-positive cells (Ki67 LI) in bile ducts in cases ofbiliary atresia (BA), congenital dilatation of the bile ducts (CDB), andnormal control liver (NCL). The Ki67 LI in BA was significantlyhigher than in CDB

Fig. 2—(A) TUNEL stain in the liver with BA. Positive cells weredetected predominantly in the ductal plate and ducts (#150). (B) Ki67immunohistochemistry in the liver with BA. Ki67 immunoreactivitywas detected predominantly in the nuclei of ductal cells (#150)

J. Pathol. 186: 429–433 (1998)

432 N. FUNAKI ET AL.

fields selected randomly in this study; the fields were,however, selected by the observers and the possibilitythat this may have influenced the results cannot becompletely ruled out.

The present study revealed that the frequency of bothapoptotic and proliferative cells was higher in the ductsof BA than in CDB and normal control liver, althoughthe control cases were different in age from the atresiacases and are therefore not true age-matched controls.On the other hand, the LIs in hepatocytes were notsignificantly different between the cases with BA or CDBand normal controls. The results of our study areconsistent with those of Cocjin et al.,15 who demon-strated a higher PCNA labelling index in the bile ductsof BA than in normal controls, although the differencesof Ki67 LI in ducts between BA and the normal controlliver did not reach statistical significance in our study.These results demonstrate that the bile ducts in the liverwith BA are associated with increased cell turnover.Enhanced proliferation of primitive bile ducts in theliver in BA may represent a compensatory response toprogressive biliary obstruction. On the other hand, it iswell recognized that the intrahepatic bile ducts in BA areabnormal even in neonates as young as 3 days of age18

and that the abnormalities of intrahepatic bile ducts arepresent in cases of BA in which the extrahepatic bileduct is patent.19 In addition, mild but ongoing bile ductdestruction persists even after good biliary drainage isachieved by Kasai’s procedure.1 Increased cell prolifer-ation and apoptosis are not observed in the liver withCDB, in which extrahepatic obstruction of the bile ductis considered to cause compensatory proliferation ofintrahepatic bile ducts. In CDB, no intrinsic abnormali-ties are detected in intrahepatic bile duct developmentand the long-term clinical prognosis following surgery isexcellent, compared with that of BA.10 The increasedcell turnover of bile ducts observed in BA is thereforeconsidered not to merely represent a compensatoryreaction to biliary obstruction, but to be related tointrinsic abnormalities of the biliary ductal systemassociated with BA.

Tan et al. reported14 that the patterns of cytokeratinimmunolocalization in the ducts of BA patients are verysimilar to those of fetal ductal plate-derived primitivebile ducts. In normal fetal liver development, certainbile ducts are selected by the remodelling process, andthe redundant ductal plate structures are deleted byapoptosis9,14 but more than several rows of primitivebile ducts are usually detected in the liver with BA.14

These findings suggest the absence of an organizedselection–deletion process in the BA liver. Increased cellturnover in the bile ducts of the liver with BA, especiallythe increased rate of apoptosis, is considered to reflectan on-going abnormal process of development in the‘ductal plate’ of the liver.

In the bile ducts of BA, apoptotic cells greatly out-numbered proliferative cells, which is considered to beone cause of paucity of interlobular ducts, or ducto-penia. The onset of biliary atresia may be caused by adevelopmental aberration due to an inherent abnor-mality of one or multiple epithelial–mesenchymal inter-active signals fundamental to the orderly progression of

? 1998 John Wiley & Sons, Ltd.

bile duct differentiation.14–16,18–20 However, a severe andprotracted inflammatory reaction, possibly caused byextravasated bile in the adjacent periductal tissue, mayalso be involved in the destruction of bile ducts.1,2,14,20

In addition, marked infiltration of cytotoxic T lym-phocytes is well documented around bile ducts in theliver in BA and the aberrant expression of major histo-compatibility complex class II antigens or HLA-DR inbiliary epithelium is also considered to play an import-ant role in the progression of biliary tract damagemediated by these cytotoxic T cells.21,22 Variouscytokines secreted from cytotoxic lymphocytes or othercells are thought to generate apoptosis by activatingcytoplasmic proteinases such as interleukin-1â-converting enzyme23 or granzyme â.24 Increased apop-tosis in bile ducts in the liver with BA may therefore alsobe partly caused by the secretion of various cytokinesfrom these infiltrating cytotoxic T lymphocytes.

REFERENCES1. Desmet VJ. Congenital diseases of intrahepatic bile ducts: variations on the

theme ‘ductal plate malformation’. Hepatology 1992; 16: 1069–1083.2. Nio M, Ohi R, Chiba T. Morphology of intrahepatic bile ducts in

jaundice-free patient with biliary atresia. Biliary Atresia 1991; 1: 7–10.3. Desmet VJ, Callea F. Cholestatic syndromes of infancy and childhood. In:

Zakim D, Boyer TD, eds. Hepatology: a Text Book of Liver Disease. Vol 2.2nd edn. Philadelphia: W. B. Saunders, 1990; 1355–1395..

4. Sadler TW. Langman’s Medical Embryology. 5th edn. Baltimore: Williams& Wilkins, 1985; 224–226..

5. Jørgensen MJ. The ductal plate malformation. A study of the intrahepaticbile duct lesion in infantile polycystic disease and congenital hepatic fibrosis.Acta Pathol Microbiol Scand 1977; 257: 1–88(Suppl).

6. Cummings MC, Winterford CM, Walker NI. Apoptosis. Am J Surg Pathol1997; 21: 88–101.

7. Gavrieli Y, Shermen Y, Ben-Sasson SA. Identification of programmed celldeath in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol1992; 119: 493–501.

8. Yabuki N, Sasano H, Tobita M, et al. Analysis of cell damage andproliferation in Helicobacter pylori-infected human gastric mucosa frompatient with gastric adenocarcinoma. Am J Pathol 1997; 151: 821–829.

9. Terada T, Nakanuma Y. Detection of apoptosis and expression ofapoptosis-related proteins during human intrahepatic bile duct develop-ment. Am J Pathol 1995; 146: 67–74.

10. Landing BH. Considerations of the pathologenesis of the neonatal hepatitis,biliary atresia and choledochal cyst: the concept of infantile obstructivecholangiopathy. Prog Pediatr Surg 1974; 6: 113–139.

11. Kasai M. Advances in treatment of biliary atresia. Jpn J Surg 1983; 13:265–276.

12. Sasano H, Frost AR, Saitoh R, et al. Immunolocalization of cyclins D andE and cyclin dependent kinase (cdk) 2 and 4 in human breast carcinoma.Anticancer Res 1997; 17: 3685–3690.

13. Sasano H, Imatani A, Shizawa S, Suzuki T, Nagura H. Cell proliferationand apoptosis in normal and pathologic human adrenal. Mod Pathol 1995;8: 11–17.

14. Tan CEL, Driver M, Howard ER, Moscoso GJ. Extrahepatic biliaryatresia: a first-trimester event? Clues from light microscopy and immuno-histochemistry. J Pediatr Surg 1994; 29: 808–814.

15. Cocjin J, Rosenthal P, Bulson V, et al. Bile ductule formation in fetal,neonatal, and infant livers compared with extrahepatic biliary atresia.Hepatology 1996; 24: 568–574.

16. Hossain M, Murahashi O, Ando H, Iio K, Kaneko K, Ito T. Immuno-histochemical study of proliferating cell nuclear antigen in hepatocytes ofbiliary atresia: a parameter to predict clinical outcome. J Pediatr Surg 1995;30: 1297–1301.

17. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear antigen(PCNA) immunolocalization in paraffin sections: an index of cell prolifer-ation with evidence of deregulated expression in some neoplasms. J Pathol1990; 162: 285–295.

18. Grenholz SK, Lilly JR, Shikes R. Biliary atresia in the newborn. J PediatrSurg 1986; 21: 1147–1148.

19. Hashimoto T, Yura J. Percutaneous transhepatic cholangiography (PTC) inbiliary atresia with special reference to the structure of the intrahepatic bileducts. J Pediatr Surg 1981; 16: 22–25.

20. Desmet VJ. What is congenital hepatic fibrosis? Histopathology 1992; 20:465–477.

J. Pathol. 186: 429–433 (1998)

433CELL TURNOVER IN BILIARY ATRESIA

21. Kobayashi H, Puri P, Sean D, Surana R, Miyano T. Hepatic overexpressionof MHC class II antigens and macrophage-associated antigens (CD68) inpatients with biliary atresia of poor prognosis. J Pediatr Surg 1997; 32:590–593.

22. Saidman SL, Duquesnoy RJ, Zeevi A. Recognition of major histocompat-ibility complex antigens on cultured human biliary epithelial cells byalloreactive lymphocytes. Hepatology 1991; 13: 239–246.

? 1998 John Wiley & Sons, Ltd.

23. Miura M, Zhu H, Rotello R, Hartweig EA, Yuan J. Induction of apoptosisin fibroblasts by IL-1b-converting enzyme, a mammalian homolog of the C.elegans cell death gene ced-3. Cell 1993; 75: 653–660.

24. Shresta S, Heusel JW, Macivor DM, Wesselschmidt RL, Russell JH, LeyTJ. Granzyme B plays a critical role in cytotoxic lymphocyte-inducedapoptosis. Immunol Rev 1995; 146: 211–221.

J. Pathol. 186: 429–433 (1998)