저작자표시 - 비영리 - 변경금지 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물을영리목적으로이용할

저작자표시 - 비영리 - 변경금지 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물을영리목적으로이용할수없습니다. 변경금지. 귀하는이저작물을개작, 변형또는가공할수없습니다. 귀하는, 이저작물의재이용이나배포의경우, 이저작물에적용된이용허락조건을명확하게나타내어야합니다. 저작권자로부터별도의허가를받으면이러한조건들은적용되지않습니다. 저작권법에따른이용자의권리는위의내용에의하여영향을받지않습니다. 이것은이용허락규약 (Legal Code) 을이해하기쉽게요약한것입니다. Disclaimer

Clinicopathological characteristics of cholangiocarcinoma: Comparison between cholangiolar differentiation and bile ductal differentiation Jung Eun Ko Department of Medical Science The Graduate School, Yonsei University

Clinicopathological characteristics of cholangiocarcinoma: Comparison between cholangiolar differentiation and bile ductal differentiation Directed by Professor Young Nyun Park The Master's Thesis submitted to the Department of Medical Science, the Graduate School of Yonsei University in partial fulfillment of the requirements for the degree of Master of Medical Science Jung Eun Ko June 2015

This certifies that the Master's Thesis of Jung Eun Ko is approved. The Graduate School Yonsei University June 2015

ACKNOWLEGEMENTS 설렘과두려움을안고시작했던학위과정, 이제비로소모든과정을마치며지난시간을되돌아봅니다. 처음연세대학교에인턴으로지냈던 2012년 8월부터오늘까지약 3년의시간은저에게학문의길뿐만아니라성장의시간이었고감사한삶이었습니다. 이렇게작지만소중한결실을맺기까지부족한저에게격려와지도를해주신박영년교수님께진심으로감사드립니다. 그리고저의논문심사를맡아주시고, 소중한조언을해주셨던안상훈교수님, 우현구교수님깊은감사를드립니다. 제석사학위논문을무사히마칠수있도록항상성심성의껏지도해주시고, 많은조언을해주신이형진선생님, 유정은선생님감사드립니다. 언제나나의엔돌핀이되어주었던동기이자언니같은동생지산아고마워. 함께한시간은많지않지만아낌없는조언과곁에서묵묵히함께해주신김영주선생님, 전영식선생님감사합니다. 제가연구를잘마무리할수있도록끝까지많은도움을주신병리학교실방근배선생님, 차종훈선생님감사합니다. 같은길을걷고있는사람이있다는것, 그것이제게얼마나큰힘이되는지모릅니다. 끝까지함께기쁨을나눌수있어감사해. 졸업을축하해지은아, 찬우야. 그리고어렵고힘들때마다함께힘이되어준은희언니, 지희언니, 경진아고마워! 항상저를믿고응원해주시는사랑하는부모님께감사드립니다. 이해해주시고믿어주셔서제가포기하지않고무사히여기까지올수있었습니다. 열심히. 그리고잘하여서부모님께서해주신은혜에보답할수있는자랑스러운딸이되도록하겠습니다. 멀리서늘기도로함께해주시는사랑하는할머니, 할아버지. 감사하고, 사랑합니다. 항상변함없는마음으로같은자리를지켜주는오빠, 그리고사랑하는나의친구들언제나내편에서서끝까지응원해줘서정말고마워. 늘감사하는마음으로보답하며살도록하겠습니다. 감사합니다. 고정은

TABLE OF CONTENTS ABSTRACT 1 I. INTRODUCTION 4 II. MATERIALS AND METHODS 6 1. Case selection and histopathological examination 7 2. Tissue microarray construction 7 3. Immunohistochemistry 7 4. Total RNA extraction, cdna synthesis, and quantitative real-time reverse-transcriptase PCR (qrt-pcr) 9 5. Statistical analysis 10 III. RESULTS 11 1. Histological evaluation for ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 11 2. Comparison of clinicopathological features between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 11 3. Comparison of CRP, N-cadherin, NCAM, CLDN18 and EMT-related marker expression between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 16 4. Clinicopathological features according to cholangiolar or bile ductal differentiation markers 20 5. Comparison of prognosis between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 25

IV. DISCUSSION 28 V. CONCLUSION 31 REFERENCES 32 ABSTRACT (IN KOREAN) 35

LIST OF FIGURES Figure 1. Comparison of the expression of CRP and CLDN18 between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 17 Figure 2. Comparison of the expression of N-cadherin, NCAM and vimentin between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 18 Figure 3. Comparison of the expression of EMT related molecules between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation 19 Figure 4. Kaplan Meier s plot analysis for overall and disease-free survival in ICCs 26 Figure 5. Kaplan Meier curves for survival rates of ICC patients 27

LIST OF TABLES Table 1. List of antibodies used for immunohistochemistry 9 Table 2. Comparison of clinicopathologic features of ICCs between c h o l a n g i o l a r d i f f e r e n t i a t i o n a n d b i l e d u c t a l differentiation 13 Table 3. Comparison of clinicopathologic features of ICCs between mass-forming ICCs with cholangiolar differentiation and mass-forming ICCs with bile ductal differentiation 15 Table 4. Comparison of clinicopathologic features between CRP positive ICCs and CRP negative ICCs 22 Table 5. Comparison of clinicopathologic features between N- cadherin positive ICCs and N-cadherin negative ICCs 23 Table 6. Comparison of clinicopathologic features between CLDN18 positive ICCs and CLDN18 negative ICCs 24

ABSTRACT Clinicopathological characteristics of cholangiocarcinoma: Comparison between cholangiolar differentiation and bile ductal differentiation Jung Eun Ko Department of Medical Science The Graduate School, Yonsei University (Directed by Professor Young Nyun Park) Recently intrahepatic cholangiocarcinoma (ICC) has been subclassified into cholangiolar differentiation and bile ductal differentiation; however their clinicopathological and molecular characteristics have not been fully understood. We studied 142 human ICC cases of Severance hospital from 1997 to 2013, and there were 20 cases (14.1%) of ICC with cholangiolar differentiation, and 122 cases (85.9%) of ICC with bile ductal differentiation. The expression of c-reactive protein (CRP), claudin 18 (CLDN18), N-cadnerin, Neural cell adhesion molecule (NCAM), vimentin, and epithelial-mesenchymal transition (EMT)-related markers (ZEB1, ZEB2, TWIST, SNAIL and loss of E-cadherin) were evaluated by immunohistochemistry or real-time PCR. The expression levels of these markers and clinicopathological features were compared between two groups. ICC patients with cholangiolar differentiation revealed higher incidence of female and viral hepatitis, and less incidence of hepatolithiasis, ductal epithelial dysplasia compared 1

to those with the ICC with bile ductal differentiation (P <0.05, for all). The massforming gross type was found in all of ICCs with cholangiolar differentiation in contrast that it was detected in 72 cases (59%) of ICCs with bile ductal differentiation (P = 0.005). The ICCs with cholangiolar differentiation showed less perineural invasion compared to ICCs with bile ductal differentiation (P = 0.013). The protein expression of CRP, N-cadherin and NCAM was more frequently found in ICCs with cholangiolar differentiation compared to those with bile ductal differentiation (P < 0.05, for all). The protein expression of CLDN18 and ZEB1 was more frequently detected in ICCs with bile ductal differentiation compared to those with cholangiolar differentiation (P < 0.05, for all). The protein expression of TWIST and E-cadherin loss showed no significant difference between two groups. The mrna expression levels of SNAIL and ZEB1 were lower in ICCs with cholangiolar differentiation compared to ICCs with bile ductal differentiation (P <0.05, for both), whereas that of ZEB2 showed no significant difference between two groups. ICCs with cholangiolar differentiation showed better overall survival compared to ICCs with bile ductal differentiation (P = 0.021). ICCs with CRP expression or N-cadherin expression revealed better prognosis compared those without (P <0.05, for all). In conclusion, ICC with cholangiolar differentiation and ICC with bile ductal differentiation are suggested to be distinct based on clinicopathological characteristics. ICC with cholangiolar differentiation is considered to be less aggressive type of ICC with better prognosis compared to ICC with bile ductal differentiation. CRP and N-cadherin are suggested to be good markers for cholangiolar differentiation. 2

Key Words: intrahepatic cholangiocarcinoma, cholangiolar differentiation, bile ductal differentiation, c-reactive protein, N-cadherin, epithelial mesenchymal transition 3

Clinicopathological characteristics of cholangiocarcinoma: Comparison between cholangiolar differentiation and bile ductal differentiation Jung Eun Ko Department of Medical Science The Graduate School, Yonsei University (Directed by Professor Young Nyun Park) I. INTRODUCTION Cholangiocarcinoma (CC) is a malignancy of biliary epithelium, arises in any portion of biliary tree, including intrahepatic, perihilar, or distal extraheptic bile duct. 1-3 The biologic behaviors, clinical characteristics of CC vary dependent on their anatomic location of origin. 4-6 Especially, intrahepatic cholangiocarcinoma (ICC), which arises from the liver periphery, has distinct characteristics compared to the CC originated from large bile ducts; hilar or extrahepatic CC. ICC is often mass forming-type, more frequently associated with chronic liver parenchymal disease such as viral hepatitis and shows less perineural, lymphatic invasion compared to hilar CC. 5,7-11 Recently, ICC has been further classified as two categories based on its histological features: cholangiolar differentiation and bile ductal differentiation. 12-14 The 4

cholangiolar differentiation is composed of cuboidal to low columnar tumor cells, resembling small bile duct of the liver, while bile ductal differentiation is composed of tall columnar tumor cells, similar to large bile duct. Interestingly, cholangiolar differentiation is associated with viral hepatitis while the bile ductal differentiation is associated with hepatolithiasis. ICC with bile ductal differentiation commonly expressed pancreatic cancer markers such as TFF1, AGR2 and S100P, and shows worse prognosis compared to cholanigolar differentiation. Taken together, according to the relevant morphologies, etiologies and molecular patterns, the ICC with cholangiolar differentiation is likely to originate from hepatic progenitor cells, while the ICC with bile ductal differentiation is similar to extrahepatic bile duct or pancreatic adenocarcinoma. 15,16 Through this, ICC is heterogeneous group of tumor possessing various cellular origin and different processes for carcinogenesis. 10 Furthermore, several transcriptomic studies reported molecular subclasses for ICCs. 17,18 CCs were largely grouped distinct classes with distinct gene expression profile and mutations; good and poor prognosis classes, 17 and proliferation and inflammatory classes. 18 The more aggressive classes, poor prognosis class and proliferation class, were associated with activation of oncogenic signaling such as EGF, MET, RAS, AKT and poor clinical outcome. In contrast, the less aggressive class, inflammatory class was characterized by activation of inflammatory signaling pathways, and good clinical outcome. Therefore, ICC is heterogeneous in its cellular origin, etiology, histologic feature, and molecular profile. However, this heterogeneity of ICC is not well understood. The biological background of the molecular classification and the relationship with histological subgroup of ICC is unknown; furthermore, clinicopathological characteristics in relation to microscopic findings have not been fully understood. 5

Therefore, we integratively analyzed the molecular signatures, clinicopathologic characteristics, and clinical outcomes according to the histological subgroup of ICC. 6

II. MATERIALS AND METHODS 1. Case selection and histopathological examination We enrolled consecutive intrahepatic cholangiocarcinoma (ICC) patients who had undergone surgical resection from 1997 to 2013 in our institution. The cases without appropriate paraffin-embedded tissue or the cases that pretreated with any kind of preoperative treatment were excluded. The representative blocks of formalin-fixed and paraffin-embedded tissue were sectioned and stained with hematoxylin-eosin (H&E). ICCs were grouped in to cholangiolar differentiation and bile ductal differentiation according to the histologic features. 12,13 This study was approved by the institutional review board of Severance Hospital (4-2014-0865) and the requirement for informed consent was waived. 2. Tissue microarray construction Core tissue biopsies were taken from individual paraffin embedded cholangiocarcinoma donor blocks and arranged in recipient tissue-array blocks using a trephine apparatus (Beecher Instruments,Silver Springs, FL, USA). At least 2 cores were sampled from each tumor, with the number of cores depending on the degree of heterogeneity present on histologic examination. 3. Immunohistochemistry Four-micron thick tissue sections were deparaffinized with xylene and rehydrated with graded alcohols. After washing in distilled water, sections were immersed in 3% hydrogen peroxide to block endogenous peroxidase. Information on antibodies used and antigen-retrieval conditions are described in Table 1. Immunohistochemical 7

stain for NCAM, ZEB1, TWIST, and E-cadherin was performed using automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA). CRP, CLDN18, N-cadherin and vimentin were performned using the DAKO Envision Kit (Dako) according to the manufacturer s instructions. All slides were counterstained with hematoxylin. Brown membranous and/or cytoplasmic staining was counted as positive for CRP, CLDN18, NCAM, N-cadherin, vimentin, E-cadherin, and nuclear and/or cytoplasmic staining for ZEB1, TWIST was counted as positive. For all antibodies studied, except for NCAM, CRP, TWIST and E-cadherin, the immunohistochemical stain results were interpreted in a semiquantitative manner and given a score, from 0 to 3, as follows: 0: staining in <5% of tumor cells; 1: weak or moderate staining in >5%; 2: moderate or strong staining in 5%; and 3: moderate of strong staining in 50% of tumor cells. Positive staining was defined as staining scores of 2 and 3 whereas o and 1 were regarded as negative. For NCAM, positivity was defined as membranous expression in 1% of tumor cells with moderate or strong intensity. For CRP, positivity was defined as membranous expression in 50% of tumor cells with moderate or strong intensity. For TWIST, positivity was defined as nuclear expression in 3% of tumor cells with moderate or strong intensity. For E-cadherin, immunohistochemical scoring was performed as follows: 0: loss of membranous expression in 5%; 1: loss of membranous E- cadherin expression in >3%. 8

Table 1. List of antibodies used for the immunohistochemistry Antibody Source Dilution Antigen retrieval NCAM (mouse mab clone Microwave, citrate Zymed (San Francisco, CA) 1:100 123C3) (ph 6.0) N-cadherin (mouse mab Microwave, citrate Zymed (San Francisco, CA) 1:300 clone 3B9) (ph 6.0) ZEB1 (rabbit mab clone Cell signaling (Danvers, Microwave, citrate 1:100 D80D3) MA, USA) (ph 6.0) E-cadherin (mouse mab Microwave, citrate Dako (Glostrup, Denmark) 1:100 clone NCH-28) (ph 6.0) Santa Cruz Biotechnology Microwave, citrate TWIST (rabbit pab) 1:50 (Santa Cruz, CA) (ph 6.0) vimentin (mouse mab clone Microwave, citrate Dako (Glostrup, Denmark) 1:200 Vim3B4) (ph 6.0) Abcam (Cambridge, MA, Microwave, citrate CRP (rabbit pab) 1:1000 USA) (ph 6.0) Microwave, citrate CLDN18 (rabbit pab) Sigma (St. Louis,MO, USA) 1:100 (ph 6.0) Abbreviations: mab, monoclonal antibody; pab, polyclonal antibody; CRP, c-reactive protein; CLDN18, claudin 18 4. Total RNA extraction, cdna synthesis, and quantitative real-time reverse transcriptase PCR (qrt-pcr) Quantitative real-time RT-PCR was performed using fresh frozen tissues, which were available in 60 cases of ICC. Total RNA was isolated using Trizol reagent (Life Technologies, Gaithersburg, MD, USA) according to the manufacturer s protocol. RNA pellet was dried and eluted using RNase-free water and purity was validated using gel electrophoresis and quantified with a spectrophotometer NanoDrop (Thermo Scientific, Wilmington, DE, USA). Complentary DNA synthesis was performed with TOPscript cdna Synthesis kit (Enzynomics, Daejeon, Korea). Briefly, the reaction master mix containing 2 RT Buffer, 20 Enzyme Mix, and nuclease-free water was mixed with 1µg of each total RNA sample. The mixtures were incubated for 60 minutes at 37 C, 5 minutes at 95 C, and then kept at 4 C. Real-time quantitative RT-PCR was carried out using the Applied Biosystems Step-One plus Real-Time PCR System. All reagents for quantitative RT-PCR were purchased from Applied Biosystems. The TaqMan 2x 9

universal PCR Master mix, 20x TaqMan assay, and RT products in a 20μl reaction volume were processed as follows: 95 C for 10 minutes, 40 cycles of 95 C for 15 seconds, and then 60 C for 60 seconds. The signal was collected at the endpoint of every cycle. The mean values of the Ct, obtained in triplicate, were used for data analysis. The Assay IDs of the primers were as follows: SNAIL (Hs00950344_a1), ZEB1 (Hs00232783_ml), ZEB2 (Hs00207691_ml) and GAPDH (Hs_99999905_m1). 5. Statistical analyses Statistical analysis was performed using the IBM SPSS 20.0.0.1 (IBM Corporation, NY, USA).We assessed the immunohistochemical stain results using the Chisquare test, and a Mann-Whitney U test was used to compare the results of the realtime quantitative RT-PCR. Survival analyses for disease-free survival and overall survival were carried out with Kaplan-Meier s method and log-rank tests. A p-value of less than 0.05 was considered statistically significant for all analysis. 10

III. RESULTS 1. Histological evaluation for ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation Histological evaluation for ICCs with cholangiolar differentiation and those with bile ductal differentiation were performed according to the previous report. 12,13 Briefly, cell morphology of the cholangiolar differentiation is cuboidal, with eosinophilic or amphophilic cytoplasm while retaining glandular, micropapillary, solid or cribriform pattern, while the features for bile ductal differentiation are long shaped and mucinous cytoplasm, and desmoplastic stroma. All of ICCs demonstrated mixed cholangiolar and bile ductal differentiation. ICC showing more than 10% of tumor area with cholangiolar component was defined as ICC with cholangiolar differentiation, and the other case was defined as ICCs with bile ductal differentiation 2. Comparison of clinicoparhological features between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation Approximately 14.1% (20/142) ICCs were grouped as ICCs with cholangiolar differentiation, and the remaining 85.9% (122/142) ICCs were ICCs with bile ductal differentiation, and the clinicopathological features were compared between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation (Table 2). The ICCs with cholangiolar differentiation were composed of higher proportions of female cases compared to the ICCs with bile ductal differentiation (P=0.028). ICCs with cholangiolar differentiation were frequently associated with viral hepatitis (HBV or HCV, defined by serological test, P=0.001), while ICCs with bile ductal differentiation were associated with hepatolithiasis (P=0.043), ductal epithelial 11

dysplasia (P=0.004). In serologic test, the ICCs with cholangiolar differentiation demonstrated lower carbohydrate antigen 19-9 (CA19-9) levels, compared to ICCs with bile ductal differentiation (P=0.002). However, the levels of carcinogembryonic antigen (CEA), alpha-fetoprotein (AFP), protein induced by vitamin K absence/antagonist-ii (PIVKA-II) were not significantly different in two groups. The gross morphology was different according to the histologic subgroup. The mass-forming gross type was found in all of ICCs with cholangiolar differentiation in contrast that it was detected in 72 cases (59%) of ICCs with bile ductal differentiation (P=0.005). ICCs with cholangiolar differentiation showed less frequent perineural invasion, (P=0.013) and more frequent fibrous capsule formation (P=0.019). 12

Table 2. Comparison of clinicopathologic features of ICCs between cholangiolar differentiation and bile ductal differentiation Clinicopathologic features Cholangiolar differentiation (n=20) Bile ductal differentiation (n=122) P value* Age (years, median, IQR) 58 (55-69) 64 (57-69) 0.274 Gender (Male, Female, %) 7 (35), 13 (65) 76 (62), 46 (38) 0.028 Serum markers CA19-9 (U/mL, median, IQR) 8.8 (1.9-37.6) 30.6 (8.2-283) 0.002 CEA (ng/ml, median, IQR) 2.4 (1.5-4) 2.1 (1.4-383) 0.614 Alpha-fetoprotein (IU/mL,median, IQR) 3.0 (1.6-5) 2.8 (2.1-5.1) 0.697 PIVKA-II (mau/ml, median, IQR) 37.5 (24-40) 27.0 (19-35) 0.065 Tumoral pathology Tumor size (cm, median, IQR) 5 (3.7-6.8) 5.0 (2.8-6) 0.186 Gross morphology (%) 0.005 Mass forming 20 (100) 72 (59) Periductal infiltrating 0 8 (6) Intraductal growth 0 19 (16) Mixed 0 23 (19) Differentiation (%) 0.071 Well differentiation 10 (56) 30 (26) Moderate differentiation 7 (39) 62 (54) Poor differentiation 1 (5) 20 (17) Undifferentiation 0 3 (3) Fibrous capsule formation (present, %) 2 (10) 0 0.019 Microvessel invasion (present, %) 12 (60) 80 (66) 0.623 Bile duct invasion (present, %) 5 (25) 45 (37.2) 0.326 Serosal invasion (present, %) 18 (90) 82 (67) 0.061 Perineural invasion (present, %) 3 (15) 53 (46) 0.013 Non-tumoral pathology Viral hepatitis (present, %) 10 (53) 17 (16) 0.001 Hepatolithiasis (present, %) 0 22 (18) 0.043 Ductal epithelial dysplasia (present, %) 0 36 (30) 0.004 Abbreviations: IQR, interquartile range. *p-values were calculated by Fisher s exact test, Pearson chi-square and Mann-Whitney U test. 13

On the basis of gross morphology, ICC is classified into three subtypes: massforming type, periductal infiltrating type, and intraductal type. 19 The periductal infiltrating and intraductal type tumor cells grow longitudinally along large bile ducts, while mass-forming type tumor cells grow along small bile duct in liver. Because mass-forming type tumor cells composed of cholangiolar and bile ductal components, we further analyzed the clinicopathologic features of mass-forming ICCs according to the histologic subgroup (Table 3). Similar to the result of whole ICC cases, mass-forming ICCs with cholangiolar differentiation were associated with female gender, background liver parenchymal disease, lower CA19-9 levels, less frequent preneural invasion, compared to mass-forming ICCs with bile ductal differentiation. (P<0.05 at all) In addition, mass-forming ICCs with cholangiolar differentiation demonstrated the better tumor differentiation, more fibrous capsule formation compared to mass forming ICCs with bile ductal differentiation (P<0.001, P=0.045, respectively). 14

Table 3. Comparison of clinicopathologic features of ICCs between massforming ICCs with cholangiolar differentiation and mass-forming ICCs with bile ductal differentiation Clinicopathologic features Cholangiolar differentiation (n=20) Bile ductal differentiation (n=72) P value* Age (years, median, IQR) 58 (55-69) 63 (55-69) 0.526 Gender (Male, Female, %) 7 (35), 13 (65) 48 (67), 24 (33) 0.019 Serum markers CA19-9 (U/mL, median, IQR) 8.8 (1.9-37.6) 60.3 (14-950) 0.004 CEA (ng/ml, median, IQR) 2.4 (1.5-4) 3.6 (1.6-11.9) 0.294 Alpha-fetoprotein (IU/mL,median, IQR) 3.0 (1.6-5) 2.9 (1.9-5.2) 0.852 PIVKA-II (mau/ml, median, IQR) 37.5 (24-40) 27.0 (21-31) 0.076 Tumoral pathology Tumor size (cm, median, IQR) 5 (3.7-6.8) 5 (3.3-6.5) 0.748 Differentiation (%) <0.001 Well differentiation 10 (56) 3 (5) Moderate differentiation 7 (39) 44 (65) Poor differentiation 1 (5) 17 (25) Undifferentiation 0 3 (5) Fibrous capsule formation (present, %) 2 (10) 0 0.049 Microvessel invasion (present, %) 12 (60) 58 (81) 0.076 Bile duct invasion (present, %) 5 (25) 19 (26) 1.000 Serosal invasion (present, %) 18 (90) 59 (82) 0.509 Perineural invasion (present, %) 3 (15) 21 (34) 0.009 Non-tumoral pathology Viral hepatitis (present, %) 10 (53) 14 (23) 0.020 Hepatolithiasis (present, %) 0 4 (6) 0.573 Ductal epithelial dysplasia (present, %) 0 13 (18) 0.063 Abbreviations: IQR, interquartile range. *p-values were calculated by Fisher s exact test, Pearson chi-square and Mann-Whitney U test. 15

3. Comparison of CRP, N-cadherin, NCAM, CLDN18 and EMT-related marker expression between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation We compared the expression of the differentiation markers that we found in gene expression profiles (CRP, CLDN18), and previously reported cholangiolar/ductular differentiation markers (N-cadherin, NCAM, and vimentin). 7,13 The CRP protein expression was more frequently observed in ICCs with cholangiolar differentiation, compared to ICCs with bile ductal differentiation (P<0.001, Figure 1A, B). In contrast, CLDN18 protein expression was more frequently found in ICCs with bile ductal differentiation compared to ICCs with cholangiolar differentiation (P=0.006, Figure 1A, C). As previously reported, the positive expression of N-cadherin and Neural cell adhesion molecule (NCAM) was associated with ICCs with cholangiolar differentiation (P<0.001 and P=0.018, respectively). However, the expression of vimentin was more prevalent in ICCs with bile ductal differentiation than ICCs with cholangiolar differentiation, although not statistically significant (P=0.124, Figure 2A-D). Because ICC with bile ductal differentiation were associated with the phenotype of tumor invasiveness (perineural invasion; Table 2), we also analyzed the expression of epithelial-mesenchymal transition (EMT) related genes. The protein expression level of zinc finger E-box binding homeobox 1 (ZEB1) was more prevalent in ICCs with bile ductal differentiation than in ICCs with cholangiolar differentiation (P=0.044, Figure 3A, B). TWIST was more frequently observed in ICCs with bile ductal differentiation than ICCs with cholangiolar differentiation, although not statistically significantly (P=0.308, Figure 3A, C). The differential expression of EMT-related genes was further confirmed by mrna levels. The mrna levels of SNAIL, ZEB1 were also significantly higher in ICCs with bile ductal differentiation 16

than ICCs with cholangiolar differentiation (P<0.001, for both) (Figure 3E, F). There was no significant difference in E-cadherin loss on immunostaining and mrna level of zinc finger E-box binding homeobox 2 (ZEB2) according to cholangiolar differentiation (P=1.000, P=0.119, Figure 3D, G). Figure 1. Comparison of the expression of CRP and CLDN18 between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation. A) ICCs with cholangiolar differentiation showing CRP expression without CLDN18 expression. In contrast, ICCs with bile ductal differentiation showing CLDN18 expression without CRP expression. B) Comparison of CRP expression between ICCs with cholangiolar differentiation and those with bile ductal differentiation. C) Comparison of CLDN18 expression between ICCs with cholangiolar differentiation and those with bile ductal differentiation (Original magnification, x200). 17

Figure 2. Comparison of the expression of N-cadherin, NCAM and vimentin between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation. A) ICCs with cholangiolar differentiation showing strong N- cadherin and NCAM expression. In contrast, ICCs with bile ductal differentiation showing strong vimentin expression. Comparison of B) N-cadherin, C) NCAM, and D) vimentin expression between ICCs with cholangiolar differentiation and those with bile ductal differentiation (Original magnification, x200). 18

Figure 3. Comparison of the expression of EMT related molecules between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation. A) Representative features of protein expression of ZEB1, TWIST, and E-cadherin in ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation. Comparison of B) ZEB1, C) TWIST, and D) E-cadherin protein 19

expression between ICCs with cholangiolar differentiation and those with bile ductal differentiation. Box plot graphs demonstrating comparisons of E) SNAIL, F) ZEB1, G) ZEB2 mrna levels between two groups (Original magnification, x200). 4. Clinicopathological features according to cholangiolar or bile ductal differentiation markers ICCs were divided into two groups according to CRP protein expression status, and clinicopathological features were compared between CRP-positive and CRPnegative groups (Table 4). CPR-positive ICCs demonstrated more frequent cholangiolar differentiation (P > 0.001). CRP-positive ICCs were associated with viral hepatitis (P=0.002), and less associated with ductal epithelial dysplasia (P=0.026), compared to CRP-negative ICCs. CRP-positive ICCs were lower CA19-9 levels, compared to CRP-negative ICCs (P=0.002). Perinueral invasion was less frequent in CRP-positive ICCs than in CRP-negative ICCs (P=0.002). Next, we divided into two groups according to N-cadherin protein expression status, and clinicopathological features were compared N-cadherin positive and N-cadherin negative groups (Table 5). N-cadherin positive ICCs showed more frequent cholangiolar differentiation (P > 0.001). N-cadherin positive ICCs were less associated with hepatolithiasis and ductal epithelial dysplasia compared to N- cadherin negative ICCs (P=0.046, P=0.005, respectively). Perinueral invasion was less frequent in N-cadherin positive ICCs than in N-cadherin negative ICCs (P=0.011). Furthermore, we divided into two groups according to CLDN18 protein expression status, and clinicopathological features were compared CLDN18- positive and CLDN18-negative groups (Table 6). CLDN18-negative ICCs showed more frequent cholangiolar differentiation (P = 0.003) and more associated with 20

viral hepatitis compared to CLDN18- positive ICCs (P=0.004). 21

Table 4. Comparison of clinicopathologic features between CRP positive ICCs and CRP negative ICCs Clinicopathologic features CRP positive (n=19) CRP negative (n=121) P value* Age (years, median, IQR) 63 (55-70) 64 (56-69) 0.274 Gender (Male, Female, %) 10 (53), 9 (47) 71 (59), 50 (41) 0.627 Cholangiolar differentiation (present, %) 13 (68) 7 (6) <0.001 Serum markers CA19-9 (U/mL, median, IQR) 10 (5.6-76.2) 28.5 (8.1-337) 0.002 CEA (ng/ml, median, IQR) 2.1 (1.9-76.2) 2.2 (1.4-6.1) 0.614 Alpha-fetoprotein (IU/mL,median, IQR) 1.7 (1.4-4.9) 2.9 (2.2-5) 0.697 PIVKA-II (mau/ml, median, IQR) 38 (36-40) 25.5 (19-32) 0.065 Tumoral pathology Tumor size (cm, median, IQR) 4.3 (3.6-5.1) 4.9 (2.9-6.5) 0.186 Gross morphology (%) 0.032 Mass forming 18 (95) 73 (60) Periductal infiltrating 0 8 (7) Intraductal growth 1 (5) 18 (15) Mixed 0 22 (18) Differentiation (%) 0.837 Well differentiation 6(35) 33 (29) Moderate differentiation 9(53) 59 (52) Poor differentiation 2 (12) 19 (17) Undifferentiation 0 3 (2) Fibrous capsule formation (present, %) 1 (5) 1 (1) 0.256 Microvessel invasion (present, %) 13 (68) 77 (64) 0.800 Bile duct invasion (present, %) 4 (21) 46 (38) 0.200 Serosal invasion (present, %) 15 (79) 84 (69) 0588 Perineural invasion (present, %) 2 (11) 54 (47) 0.002 Non-tumoral pathology Viral hepatitis (present, %) 9 (53) 18 (17) 0.002 Hepatolithiasis (present, %) 1 (5) 21 (17) 0.308 Ductal epithelial dysplasia (present, %) 1 (5) 35 (29) 0.026 Abbreviations: IQR, interquartile range. *p-values were calculated by Fisher s exact test, Pearson chi-square and Mann-Whitney U test. 22

Table 5. Comparison of clinicopathologic features between N-cadherin positive ICCs and N-cadherin negative ICCs Clinicopathologic features N-cadherin positive (n=31) N-cadherin negative (n=109) P value* Age (years, median, IQR) 59 (55-66) 65 (57-70) 0.101 Gender (Male, Female, %) 16 (52), 15 (48) 67 (62), 42 (38) 0.408 Cholangiolar differentiation (present, %) 11 (36) 8 (7) <0.001 Serum markers CA19-9 (U/mL, median, IQR) 34 (7.6-383) 27 (8-291) 0.118 CEA (ng/ml, median, IQR) 2.9 (1.7-4.9) 2.1 (1.3-5.8) 0.958 Alpha-fetoprotein (IU/mL,median, IQR) 1.6 (1.3-6) 3 (2.2-4.5) 0.905 PIVKA-II (mau/ml, median, IQR) 36 (25-42) 27 (20-34) 0.091 Tumoral pathology Tumor size (cm, median, IQR) 5 (3.9-6.3) 4.5 (2.7-6) 0.318 Gross morphology (%) 0.009 Mass forming 28 (90) 63 (58) Periductal infiltrating 0 8 (7) Intraductal growth 1 (3) 18 (17) Mixed 2 (7) 20 (18) Differentiation (%) 0.430 Well differentiation 9 (31) 30 (29) Moderate differentiation 13 (45) 55 (54) Poor differentiation 7 (24) 14 (14) Undifferentiation 0 3 (3) Fibrous capsule formation (present, %) 1 (3) 1 (1) 0.398 Microvessel invasion (present, %) 20 (65) 71 (65) 1.000 Bile duct invasion (present, %) 6 (19) 43 (40) 0.054 Serosal invasion (present, %) 25 (81) 73 (67) 0.184 Perineural invasion (present, %) 6 (20) 49 (48) 0.011 Non-tumoral pathology Viral hepatitis (present, %) 7 (26) 20 (21) 0.601 Hepatolithiasis (present, %) 1 (3) 21 (19) 0.046 Ductal epithelial dysplasia (present, %) 2 (7) 34 (31) 0.005 Abbreviations: IQR, interquartile range. *p-values were calculated by Fisher s exact test, Pearson chi-square and Mann-Whitney U test. 23

Table 6. Comparison of clinicopathologic features between CLDN18 positive ICCs and CLDN18 negative ICCs Clinicopathologic features CLDN18 positive (n=56) CLDN18 negative (n=84) P value* Age (years, median, IQR) 66 (58-70) 63 (55-68) 0.130 Gender (Male, Female, %) 33 (59), 23 (41) 49 (58), 35 (42) 1.000 Cholangiolar differentiation (present, %) 2 (4) 18 (21) 0.003 Serum markers CA19-9 (U/mL, median, IQR) 22.5 (3.7-130) 37.6 (10-950) 0.875 CEA (ng/ml, median, IQR) 1.8 (1.2-3.5) 3 (1.9-7.5) 0.365 Alpha-fetoprotein (IU/mL,median, IQR) 3 (2.2-4.5) 2.8 (1.7-5.2) 0.018 PIVKA-II (mau/ml, median, IQR) 27 (18-42) 28 (21-35) 0.524 Tumoral pathology Tumor size (cm, median, IQR) 4.3 (2.7-6.3) 4.8 (3.2-6.0) 0.496 Gross morphology (%) 0.179 Mass forming 31 (55) 61 (73) Periductal infiltrating 4 (7) 4 (5) Intraductal growth 11 (20) 8 (9) Mixed 10 (18) 11 (13) Differentiation (%) 0.255 Well differentiation 17 (32) 23 (30) Moderate differentiation 31 (57) 37 (47) Poor differentiation 6 (11) 15 (19) Undifferentiation 0 3 (4) Fibrous capsule formation (present, %) 0 2 (2) 0.515 Microvessel invasion (present, %) 35 (63) 55 (66) 0.723 Bile duct invasion (present, %) 22 (39) 27 (33) 0.471 Serosal invasion (present, %) 35 (63) 63 (75) 0.134 Perineural invasion (present, %) 25 (48) 22 (32) 0.428 Non-tumoral pathology Viral hepatitis (present, %) 4 (8) 23 (30) 0.004 Hepatolithiasis (present, %) 7 (13) 14 (17) 0.631 Ductal epithelial dysplasia (present, %) 16 (29) 20 (24) 0.561 Abbreviations: IQR, interquartile range. *p-values were calculated by Fisher s exact test, Pearson chi-square and Mann-Whitney U test. 24

5. Comparison of prognosis between ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation We evaluated the prognostic significance of the histologic subgroup, and their differentiation markers (CRP, N-cadherin and CLDN18). The ICCs with cholangiolar differentiation demonstrated significantly better overall survival when compared to ICCs with bile ductal differentiation (P=0.021). The CRP- or N- cadherin positive ICCs, also showed significantly better survival when compared to those negative ones (P=0.011, P=0.041, respectively. Figure 4A). The cholangiolar differentiation, CRP and N-cadherin were not the significant prognostic factor for disease free survival in ICC patients (Figure 5A). Regarding the mass-forming ICC subgroup, the positive for ICCs with cholangiolar differentiation, CRP, or N- cadherin were good prognostic factor for overall survival (P<0.05, for all, Figure 4B) and N-cadherin was prognostic factor for favorable disease-free survival (P=0.018, Figure 4C). However, CLDN18 was not significant prognostic factor for disease free-survival and overall survival in both ICC and mass-forming ICC subgroup (Figure 5B, C). 25

Figure 4. Kaplan Meier s plot analysis for overall and disease-free survival in ICCs. A) Survival curves showed better overall survival in ICCs with cholangiolar differentiation, CRP-positive, and N-cadherin positive expression. B) Kaplan- Meier s plot analysis showed better overall survival in ICCs with cholangiolar differentiation, CRP-positive, and N-cadherin positive patients with mass-forming gross morphology. C) Kaplan-Meier s plot analysis showed better disease-free survival in ICCs N-cadherin positive expression patients with mass-forming gross morphology. 26

Figure 5. Kaplan Meier curves for survival rates of ICC patients. A) Kaplan- Meier curves for disease-free survival of patients with ICC showed according to cholangiolar differentiation, CRP-, and N-cadherin protein expression status. B) Kaplan Meier curves for overall survival of patients with ICC and mass-forming type ICC demonstrated according to the CLDN18 protein expression status. C) Kaplan Meier curves for disease-free survival of patients with ICC and massforming type ICC demonstrated according to the CLDN18 protein expression status. 27

IV. DISCUSSION Cholangiocarcinoma is very heterongenous tumor in the points of etiology, morphology, cell of-origin, and clinical features. Pathologically, ICC has various morphologies and is thus generally subclassified as two distinct groups; ICCs with cholangiolar differentiation and ICCs with bile ductal differentiation. These two groups have different characteristics not only in morphological features but in etiological and clinical features and immunophenotye. In the ICCs with cholangiolar differentiation, the morphology of the cholangiolar differentiation is similar to small bile duct or hepatic progenitor cells, 20 while the ICCs with bile ductal differentiation is similar to large bile duct or pancreatic duct. In addition, the underlying liver disease is dependent upon subgroup, as the patient of viral hepatitis is commonly associated with, like HCC, the ICC with cholangiolar differentiation while the ICC with bile ductal differentiation is associated with hepatolithiasis. Also, pancreatic cancer markers, such as TFF1, AGR2 and S100P, were significantly expressed in ICCs with bile ductal differentiation, and these tumors showed significantly poor overall survival. 12,13 This suggests that ICC shows the tumor heterogeneity in terms of embryological development as well as pathological features. In the present study, ICC with cholangiolar differentiation (14.1%) was frequently associated with clinicopathologic features, including less frequent perinueral invasion, and good differentiation. Fibrous capsule formation and lack of ductal epitherial dysplasia were more frequently observed. EMT-related proteins, such as ZEB1 were significantly less expressed in ICC with cholangiolar differentiation, and these tumors showed good prognosis. Therefore, ICC with cholangiolar differentiation was more closely related to less aggressive behavior. 28

Embryologically, at embryonic day (E)9.5, biliary trees and pancreas are originated from ventral endoderm. 21 Together with extrahepatic biliary tree, pancreas arises from the ventral endoderm of the foregut at almost same time, whereas small intrahepatic biliary tree is originated from the hepatic stem cells. Hepatic stem cells in the canals of Hering differentiated into hepatoblasts and to hepatocytes or intrahepatic cholangiocytes. 20,22,23 This suggests that extrahepatic bile duct and pancreatic duct have similar cellular origin. Contrary to these two types, small bile duct is thought to have a smilar origin to those of hepatoblast and hepatocyte, indicating that embryological origin of duct varies. By gene expression pattern, ICC can also be classified into inflammation group and proliferation group, associated with the former representing good prognosis and the latter representing poor. In this study, according to the result of microarray data, inflammation or good prognosis group is associated with cholangiolar differentiation, whereas proliferation or poor prognosis group with bile ductal differentiation, indicating that genomic and genetic characterization of ICC is highly associated with classification of histological subgroup. 17,18 With patient outcome for ICC with cholangiolar differentiation (liver-like CC) is better than that of ICC with bile ductal differentiation (pancreas cancer-like CC), 6 patient prognosis clearly can be divided based on the subclassification. According to the recent report, cholangiolocellular carcinoma (CLC) is a type of combined HCC-CC largely containing cells shaped similar to cholangiolar differentiation and features better prognosis than ICC with less lymph node metastasis and perineural invasion, which are well-known prognostic factors of CC. 24,25 Because this subgroup of ICCs has been reported to show less aggressive behaviors, compared to ICC with bile duct differentiation, it is important that a suitable marker is developed 29

to facilitate its diagnosis. Indeed, it is thought that the expression of CRP and N- cadherin may serve as a good prognostic marker. 30

V. CONCLUSION In conclusion, ICC with cholangiolar differentiation and ICC with bile ductal differentiation are suggested to be distinct based on clinicopathological characteristics. ICC with cholangiolar differentiation is considered to be less aggressive type of ICC with better prognosis compared to ICC with bile ductal differentiation. CRP and N-cadherin are suggested to be good markers for cholangiolar differentiation. 31

REFERENCES 1. Blechacz BR, Gores GJ. Cholangiocarcinoma. Clin Liver Dis 2008;12:131-50, ix. 2. Patel T. Cholangiocarcinoma. Nat Clin Pract Gastroenterol Hepatol 2006;3:33-42. 3. Razumilava N, Gores GJ. Classification, diagnosis, and management of cholangiocarcinoma. Clin Gastroenterol Hepatol 2013;11:13-21 e1; quiz e3-4. 4. Malhi H, Gores GJ. Cholangiocarcinoma: modern advances in understanding a deadly old disease. J Hepatol 2006;45:856-67. 5. Aishima S, Kuroda Y, Nishihara Y, Iguchi T, Taguchi K, Taketomi A, et al. Proposal of progression model for intrahepatic cholangiocarcinoma: clinicopathologic differences between hilar type and peripheral type. Am J Surg Pathol 2007;31:1059-67. 6. Nakanuma Y, Harada K, Ishikawa A, Zen Y, Sasaki M. Anatomic and molecular pathology of intrahepatic cholangiocarcinoma. J Hepatobiliary Pancreat Surg 2003;10:265-81. 7. Nakanuma Y, Sasaki M, Ikeda H, Sato Y, Zen Y, Kosaka K, et al. Pathology of peripheral intrahepatic cholangiocarcinoma with reference to tumorigenesis. Hepatology Research 2008;38:325-34. 8. Kozaka K, Sasaki M, Fujii T, Harada K, Zen Y, Sato Y, et al. A subgroup of intrahepatic cholangiocarcinoma with an infiltrating replacement growth pattern and a resemblance to reactive proliferating bile ductules: 'bile ductular carcinoma'. Histopathology 2007;51:390-400. 9. Nakanuma Y, Sato Y, Harada K, Sasaki M, Xu J, Ikeda H. Pathological classification of intrahepatic cholangiocarcinoma based on a new concept. World J Hepatol 2010;2:419-27. 10. Aishima S, Oda Y. Pathogenesis and classification of intrahepatic cholangiocarcinoma: different characters of perihilar large duct type versus 32

peripheral small duct type. J Hepatobiliary Pancreat Sci 2014. 11. Nakanuma Y, Sato Y. Hilar cholangiocarcinoma is pathologically similar to pancreatic duct adenocarcinoma: suggestions of similar background and development. J Hepatobiliary Pancreat Sci 2014;21:441-7. 12. Liau JY, Tsai JH, Yuan RH, Chang CN, Lee HJ, Jeng YM. Morphological subclassification of intrahepatic cholangiocarcinoma: etiological, clinicopathological, and molecular features. Mod Pathol 2014;27:1163-73. 13. Yu TH, Yuan RH, Chen YL, Yang WC, Hsu HC, Jeng YM. Viral hepatitis is associated with intrahepatic cholangiocarcinoma with cholangiolar differentiation and N-cadherin expression. Mod Pathol 2011;24:810-9. 14. Tsai JH, Huang WC, Kuo KT, Yuan RH, Chen YL, Jeng YM. S100P immunostaining identifies a subset of peripheral-type intrahepatic cholangiocarcinomas with morphological and molecular features similar to those of perihilar and extrahepatic cholangiocarcinomas. Histopathology 2012;61:1106-16. 15. Cardinale V, Carpino G, Reid L, Gaudio E, Alvaro D. Multiple cells of origin in cholangiocarcinoma underlie biological, epidemiological and clinical heterogeneity. World J Gastrointest Oncol 2012;4:94-102. 16. Roskams T. Liver stem cells and their implication in hepatocellular and cholangiocarcinoma. Oncogene 2006;25:3818-22. 17. Andersen JB, Spee B, Blechacz BR, Avital I, Komuta M, Barbour A, et al. Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology 2012;142:1021-31 e15. 18. Sia D, Hoshida Y, Villanueva A, Roayaie S, Ferrer J, Tabak B, et al. Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes. Gastroenterology 2013;144:829-40. 19. Blechacz B, Komuta M, Roskams T, Gores GJ. Clinical diagnosis and staging of cholangiocarcinoma. Nat Rev Gastroenterol Hepatol 2011;8:512-33

22. 20. Zaret KS, Grompe M. Generation and regeneration of cells of the liver and pancreas. Science 2008;322:1490-4. 21. Zaret KS. Regulatory phases of early liver development: paradigms of organogenesis. Nat Rev Genet 2002;3:499-512. 22. Cardinale V, Wang Y, Carpino G, Mendel G, Alpini G, Gaudio E, et al. The biliary tree--a reservoir of multipotent stem cells. Nat Rev Gastroenterol Hepatol 2012;9:231-40. 23. Deutsch G, Jung J, Zheng M, Lora J, Zaret KS. A bipotential precursor population for pancreas and liver within the embryonic endoderm. Development 2001;128:871-81. 24. Komuta M, Spee B, Vander Borght S, De Vos R, Verslype C, Aerts R, et al. Clinicopathological study on cholangiolocellular carcinoma suggesting hepatic progenitor cell origin. Hepatology 2008;47:1544-56. 25. Ariizumi S, Kotera Y, Katagiri S, Nakano M, Nakanuma Y, Saito A, et al. Long-term Survival of Patients with Cholangiolocellular Carcinoma After Curative Hepatectomy. Annals of Surgical Oncology 2014;21:451-8. 34

ABSTRACT (IN KOREAN) 담세관분화와담관분화를보이는 담관상피암종의임상병리학적특성비교 < 지도교수박영년 > 연세대학교대학원의과학과 고정은 최근간내담관암에는담세관분화및담관분화를보이는것이있다고보고되었으나, 그임상병리학적및분자병리학적특성에대해서는아직밝혀지지않았다. 본연구에서는 1997년부터 2013년까지의세브란스병원에서수술받은간내담관암환자 142명을선별하여연구를진행하였다. 병리조직학적검색소견상담세관분화는 20 (14.1%) 예에서담관분화는 122 (85.9%) 예에서관찰되었다. 면역조직화학염색및실시간중합효소연쇄반응을이용하여 c-reactive protein (CRP), claudin (CLDN18), N-cadherin, Neural cell adhesion molecule (NCAM), vimentin 그리고상피간엽이행마커로잘알려진 zinc finger E-box binding homeobox1 (ZEB1), zinc finger E-box binding homeobox2 (ZEB2), TWIST, SNAIL 그리고 E-cadherin에대한발현과임상병리학적특성을담세관분화및담관분화를보이는두그룹에서비교하였다. 담세관분화를보이는간내담관암은담관분화가있는간내담관암보다여성의발생이높았고, B형또는 C형만성간염과연광성이높았던반면, 간내담석증과담관상피이형성비율은낮았다. (P < 0.05). 육안소견상담세관분화를보이는간내 35

담관암은모두 (20/20, 10%) 종괴형성형의소견을보이는반면, 담관분화가있는간내담관암은 72예 (72/122, 59%) 가종괴형성형이였다 (P = 0.005). 또한담세관분화가있는간내담관암에서신경주위침범이담관분화가있는간내담관암보다더많이관찰되었다 (P = 0.013). CRP, N-cadherin그리고 NCAM의단백질발현은담세관분화가있는간내담관암에서높았고, CLDN18과 ZEB1의단백질발현은담관분화가있는간내담관암에서높았다 (P < 0.05). 반면, TWIST와 E-cadherin의단백질발현은두군간에차이가없었다. SNAIL과 ZEB1의 mrna발현은담관분화가있는간내담관암보다담세관분화가있는간내담관암에서더낮게발현되었지만 (P < 0.05), ZEB2의 mrna발현은두군간에차이가없었다. 환자추적관찰분석결과담세관분화가있는환자군이담관분화가있는환자군보다예후가더좋았으며, CRP와 N-cadherin의단백질이발현되는환자군이그렇지않은환자군보다예후가더좋았다 (P < 0.05). 이상의소견으로담세관분화가있는간내담관암과담관분화가있는간내담관암은서로다른임상병리학적및분자병리학적특성을가지며, 담세관분화가있는간내담관암이담관분화가있는간내담관암보다종양의생물학적악성도가적으며, 환자의예후도더좋았다. 또한, CRP, N-cadherin이담세관분화를보이는간내담관암의좋은마커로생각한다. 핵심되는말 : 담관상피암종, 담세관분화, 담관분화, c-reactive protein, N-cadherin, 상피간엽이행 36