- Email: [email protected]
Differential expression of intestinal trefoil factor in biliary epithelial cells of primary biliary cirrhosis
Differential Expression of Intestinal Trefoil Factor in Biliary Epithelial Cells of Primary Biliary Cirrhosis Yasuhiko Kimura,1,2 Patrick S. C. Leung,1 Thomas P. Kenny,1 Judy Van De Water,1 Mikio Nishioka,2 Andrew S. Giraud,3 James Neuberger,4 Gordon Benson,5 Rashmi Kaul,6 Aftab A. Ansari,7 Ross L. Coppel,8 and M. Eric Gershwin1 Intestinal trefoil factor (ITF) promotes epithelial cell migration and mucosal restitution during inflammation. We used real-time quantitative PCR, in situ nucleic acid hybridization, and immunohistochemistry to study the expression of the ITF gene and protein expression in the liver of primary biliary cirrhosis (PBC) and controls. There were significantly higher levels of ITF messenger RNA (mRNA) in PBC liver compared with primary sclerosing cholangitis (PSC) (P < .05) or normal controls (P < .001) and also higher in hepatitis C virus (HCV) liver (P < .05) and cryptogenic cirrhosis (P < .01) compared with normal controls. However, only in PBC was there a significant difference between small (interlobular and bile ductules) and large (intrahepatic and septal) bile ducts. Using in situ hybridization, the highest levels of ITF gene expression were localized to the large bile ducts in PBC. This differential expression of ITF was also noted at the protein level. Thus, in PBC, although 92% of large bile ducts expressed the ITF protein, only 2% of small bile ducts (P < .0001) expressed ITF. In contrast, in control livers, 34% of large bile ducts and 13% of small bile ducts expressed ITF. ITF protein is absent in small bile ducts in all stages of PBC. In conclusion, the expression of ITF may play an important role in bile duct damage. In small bile ducts, ITF production in response to damage is absent, making such cells vulnerable to damage and providing a thesis for the selective loss of small, but not large, bile ducts in PBC. (HEPATOLOGY 2002;36:1227-1235.)
P
rimary biliary cirrhosis (PBC) is a chronic and progressive autoimmune liver disease characterized by the presence of antimitochondrial antibodies, immune destruction of intrahepatic biliary epithelial cells (BECs), and eventually cirrhosis.1-3 Histologically, interlobular bile ducts are the primary site of destruction, and Abbreviations: PBC, primary biliary cirrhosis; BECs, biliary epithelial cells; PDC-E2, E2 subunit of pyruvate dehydrogenase; ITF, intestinal trefoil factor; SP, spasmolytic polypeptide; TFF, trefoil factor family; mRNA, messenger RNA; PCR, polymerase chain reaction; PSC, primary sclerosing cholangitis; HCV, hepatitis C virus; cDNA, complementary DNA. From the 1Division of Rheumatology, Allergy and Clinical Immunology, University of California, Davis, CA; 2Third Department of Internal Medicine, Kagawa Medical University, Kagawa, Japan; 3Department of Medicine, University of Melbourne at Western Hospital, Footscray, Australia; 4Queen Elizabeth Hospital, Liver Unit, Edgbaston, Birmingham, United Kingdom; 5UMDNJ Robert Wood Johnson Medical School, Camden, NJ; 6Department of Pediatrics, University of Minnesota Medical School, Minneapolis, MN; 7Department of Pathology, Emory University, Atlanta, GA; and 8Department of Microbiology, Monash University, Victoria, Australia. Received February 4, 2002; accepted July 22, 2002. Address reprint requests to: M. Eric Gershwin, M.D., Division of Rheumatology, Allergy and Clinical Immunology, University of California at Davis, TB 192, School of Medicine, Davis, CA 95616. E-mail: [email protected]; fax: 530-752-4669. Copyright © 2002 by the American Association for the Study of Liver Diseases. 0270-9139/02/3605-0027$35.00/0 doi:10.1053/jhep.2002.36157
progressive disappearance of interlobular bile ducts in PBC is irreversible1,2 Furthermore, intrahepatic large bile ducts are not affected, and septal bile ducts do not disappear despite the presence of chronic nonsuppurative cholangitis. Although the immunodominant autoantigen, the E2 subunit of pyruvate dehydrogenase (PDCE2), is present in all nucleated cells including the large bile ducts, the reason why the interlobular bile ducts and ductules selectively become the specific target of immune destruction is unknown. One theory is that there is subtle but distinct heterogeneity among the BECs that line the different levels of bile ducts. Intestinal trefoil factor (ITF, also called TFF3), spasmolytic polypeptide (SP; SP/TFF2), and pS2/TFF1 belong to the trefoil factor family (TFF). The TFF proteins are mucin-associated protease-resistant peptides that contain 1 or more trefoil motifs composed of 6 cysteine residues and are thought to influence the rheologic properties of mucus gels and thus promote protection to the mucosal surface.4,5 In the gastrointestinal mucosa of inflammatory bowel diseases and in gastric ulceration, trefoil factors are up-regulated and appear to promote epithelial migration and wound healing.6 SP was originally identified as a human homologue7 of a peptide isolated from porcine pan1227
1228
KIMURA ET AL.
creas,8 and pS2 was discovered during studies of estrogeninducible messenger RNA (mRNA) in breast carcinoma cell lines.9 ITF was first cloned from rat intestine and is predominantly produced by goblet cells of the small and large intestine.10,11 ITF-deficient mice lack epithelial restitution following colitis induced by dextran sulfate sodium.12 Recent data also suggest that trefoil peptides enhance epithelial motility and spreading, possibly via extracellular matrix alteration, and accelerate mucosal epithelial recovery after injury.13 In addition, ITF is involved in phosphorylation of -catenin and the epithelial growth factor receptor,12,14 which is important for the normal development of intrahepatic bile ducts. Although the precise molecular mechanisms by which ITF mediates its biologic effect remain to be fully defined, one of the potential mechanisms by which its protective effect is mediated is via epithelial cell apoptosis. Epithelial cells are protected from cell death by maintenance of cellcell and cell-substratum contacts and disruption of such interaction leading to detachment, resulting in apoptosis. This thesis is further strengthened by the fact that ITF expression in a human epithelial cell line confers resistance of this cell line to undergo apoptosis, induced by serum starvation and ceramide,15 and attributed to the ability of ITF to induce distinct intracellular signaling pathways that influence both cell migration and apoptosis.16 The functional roles of the trefoil proteins TFF1, TFF2, and TFF3 prompted us to examine their expression at both the RNA and protein level in liver tissues from PBC patients. Results of our preliminary studies showed that, whereas no detectable differences were noted in the expression of TFF1 and TFF2 in liver specimens from PBC patients and controls, significant differences were noted in the expression of ITF. The studies reported herein, therefore, were focused on defining the expression of TFF3 or ITF in more detail at both the mRNA and protein level in liver.
Patients and Methods Liver Tissue Liver tissue specimens from 60 patients with PBC (20 frozen, 40 paraffin embedded) and 129 controls (50 frozen, 79 paraffin embedded) were studied. Frozen liver specimens were used for real-time quantitative polymerase chain reaction (PCR) and in situ nucleic acid hybridization, and paraffin-embedded liver specimens were used for immunohistochemistry. The PBC patients included 27 patients (2 frozen, 25 paraffin embedded) that were classified as stage I or II and 33 patients (18 frozen, 15 paraffin embedded) classified as stages III and IV. The controls included healthy liver (10 frozen, 25 paraffin embedded), primary sclerosing cholangitis (PSC) (16 fro-
HEPATOLOGY, November 2002
zen, 20 paraffin embedded), alcoholic cirrhosis (8 frozen, 11 paraffin embedded), hepatitis C virus (HCV)-related liver disease (8 frozen, 13 paraffin embedded), and cryptogenic cirrhosis (8 frozen, 10 paraffin embedded). The tissues were obtained either from patients undergoing orthotopic liver transplantation or from liver biopsy specimens from the University of California at Davis, CA; Kagawa Medical University, Japan; Queen Elizabeth Hospital, United Kingdom; the Department of Medicine, Robert Wood Johnson Medical School, Camden, NJ; and the Department of Pathology, University of Minnesota, Minneapolis, MN. The diagnosis of all patients was based on established criteria, including histologic, clinical, and laboratory analysis.1 RNA Isolation and PCR Detection RNA Extraction. Total RNA was isolated from approximately 100 mg of frozen liver tissue by the guanidinium thiocyanate-phenol-chloroform method, using Trizol reagent (Invitrogen, Carlsbad, CA). Liver tissue was homogenized in 1 mL Trizol reagent; 200 L chloroform was added and the mixture shaken vigorously and then left at room temperature for 5 minutes. After centrifugation at 12,000g (4°C) for 15 minutes, the aqueous phase was transferred to a fresh tube, and 500 L isopropanol was added. The mixture was left at room temperature for 10 minutes, followed by centrifugation at 12,000g (4°C) for 10 minutes. The RNA precipitate was finally rinsed with 75% ethanol and dried. Finally, the RNA was dissolved in 50 L of distilled water and quantitated, using a spectrophotometer, at OD260. First-Strand cDNA Synthesis. Total RNA (1 g) was used for the synthesis of single-strand complementary DNA (cDNA) with reverse transcriptase. First, total RNA was incubated for 10 minutes at 70°C with 1 L Oligo (dT) and 500 nmol/L dNTPs, quick-chilled on ice, and further incubated for 50 minutes at 42°C with 200 units reverse transcriptase (SuperScript II, Gibco-BRL, Gaithersburg, MD), 10 nmol/L dithiothreitol, and 5⫻ firststrand buffer in a final volume of 20 L. Reactions were then heat inactivated at 70°C for 15 minutes. This cDNA pool was used for standard PCR in addition to quantitation by real-time PCR. PCR Detection. As a positive control, the primers for the housekeeping gene GAPDH were used. Amplification of ITF-specific DNA fragments was accomplished by adding 1 L of the cDNA pool to a PCR mixture that contained 0.2 mmol/L dNTPs, 0.5 mol/L each of sense and antisense-specific primers (Table 1), 10⫻ PCR buffer, and 2.5 units Taq polymerase (Promega, Madison, WI) in a total volume of 50 L. We used the touchdown method for amplification, an initial denaturation at
HEPATOLOGY, Vol. 36, No. 5, 2002
KIMURA ET AL.
Table 1. Primers Sequence
TFF3-sense (12F) TFF3-antisense (79R) TFF3-antisense (234R) GAPDH-sense GAPDH-antisense
5⬘-TGGTCCTGGCCTTGCTGT-3⬘ 5⬘-GGCACACTGGTTTGCAGACA-3⬘ 5⬘-GAGGTGCCTCAGAAGGTGC-3⬘ 5⬘-AAGGTCGGAGTCAACGGATTT-3⬘ 5⬘-ACCAGAGTTAAAAGCAGCCCTG-3⬘
NOTE. 12F and 79R were used for real-time PCR amplification of ITF mRNA; 12F and 234R were used for PCR amplification of ITF mRNA to make the specific probes of in situ hybridization.
94°C for 10 minutes, and 1 cycle each with a 20-second denaturation at 94°C, a 1-minute annealing at 58°C, 56°C, 54°C, or 52°C, a 30-second extension at 72°C, followed by 30 cycles with a 20-second denaturation at 94°C, a 1-minute annealing at 50°C, and a 30-second extension at 72°C. After PCR, 2 L aliquots of the products were subjected to 2% agarose gel electrophoresis and stained with ethidium bromide. Observation of a band of predicted size on gel electrophoresis indicated the presence of mRNA in the original tissue sample. Quantitative Real-Time PCR. The expression of subtracted messages were verified by quantitative realtime PCR, using GeneAmp 5700 Sequence Detection System (PE Applied Biosystems). We employed the SYBR green I dye method of detection. Briefly, PCR reactions were performed using 96-well trays with 2.5 L 10⫻ diluted cDNA, together with 0.3 mol/L each of sense and antisense primers (Table 1), 1 unit AmpliTaq Gold (PE Applied Biosystems), 0.2 mmol/L dNTPs, 2.5 L 10⫻ PCR buffer, 1.5 L MgCl solution, and 2.5 L SYBR green dye solution (1:10,000 in DMSO, Molecular Probes) in a total volume of 25 L. The reaction mixture was preheated at 95°C for 10 minutes, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. Standard curves for ITF and GAPDH (housekeeping) were generated by using serial dilutions of cDNA, which were run at the same time, respectively, and all reaction wells were run in duplicate. Data from each well were analyzed using GeneAmp 5700 analysis software, and the transcription value of the target gene was obtained by plotting on a standard curve. The value of each amplified gene was normalized by the amount of GAPDH quantified from the same sample. In Situ Hybridization Preparation of RNA Probe. We amplified the specific sequence of ITF (219 bp) using the PCR, and the amplified products were subcloned into the PCR IITOPO vector using the TA cloning kit (Invitrogen, Carlsbad, CA). The subcloned DNA sequence was confirmed utilizing the DBS Automated DNA Sequencing
1229
Facility at the University of California at Davis. After linearization of the plasmid with appropriate restriction enzymes, single-strand RNA probes (antisense and sense) were obtained by in vitro transcription with sp6- and T7polymerases, and these probes were labeled with digoxigen. Efficient labeling of RNA probes was confirmed by dot blotting of ascending dilutions of probe nylon membranes. Preparation of Liver Section. The liver specimens were snap-frozen in OCT compound and stored at ⫺70°C until use. Frozen 8-m sections were cut using a cryostat and mounted on poly-L-lysine-coated slides. The sections were fixed with 4% paraformaldehyde for 30 minutes at room temperature and incubated with 10 g/mL proteinase K for 10 minutes at 37°C. After postfixation in 4% paraformaldehyde, the slides were immersed in 0.2 N HCl for 20 minutes and acetylated with acetic anhydride for 10 minutes at room temperature. The slides were dehydrated with 100% ethanol for 5 minutes and air-dried. Hybridization. Pretreated sections were incubated with hybridization solution (50% formamide, 5⫻ SSPE, 0.1% sodium dodecyl sulfate, 1⫻ Denhardt’s solution, 5% dextran sulfate, 50 mmol/L dithiothreitol, 100 g/mL yeast transfer RNA, and 100 g/mL herring sperm DNA) containing digoxigen-labeled antisense or sense RNA probe in a moist chamber at 52°C overnight. After hybridization, the sections were washed twice with 2⫻ SSC, then incubated in 10 g/mL RNase A for 30 minutes at 37°C. The slides were washed with 2⫻ SSC twice, 0.1⫻ SSC for 30 minutes at 67°C, and 0.1⫻ SSC for 10 minutes at room temperature. Hybridized sections were washed with washing buffer (0.1 mol/L maleic acid, 0.15 mol/L NaCl, pH 9.5) briefly, blocked with 1% blocking solution for 30 minutes, followed by incubation with alkaline phosphatase-conjugated sheep anti-digoxigen antibody (1:500) for 1 hour. After washing with washing buffer twice, the sections were incubated with NBT/BCIP substrate diluted with detection buffer (100 mmol/L Tris-HCl, 100 mmol/L NaCl, 50 mmol/L MgCl, pH 9.5) for 2 or 3 days in a dark room. The color reaction was stopped with TE buffer (10 mmol/L TrisHCl, pH 8.0, 1 mmol/L EDTA). Finally, the sections were counterstained with Nuclear Red, mounted with crystal mount, and then observed by light microscopy. Immunohistochemical Staining for ITF Immunohistochemical staining of ITF was performed on formalin-fixed, paraffin-embedded sections, using a standard avidin-biotin complex method. Tissue sections were first deparaffinized in xylene and rehydrated. The sections were then blocked with TBS (20 mmol/L Tris-
1230
KIMURA ET AL.
Fig. 1. ITF gene expression in PBC, PSC, and normal liver. Total RNA isolated from liver specimens was reverse transcribed and amplified by PCR with specific primers. As a positive control, the housekeeping gene GAPDH was also amplified.
HCl, pH 8.0, 150 mmol/L NaCl) containing 10% goat serum for 20 minutes at room temperature, then incubated with ITF antibodies diluted at a previously defined optimal dilution of 1:1,000 for 1 hour at room temperature.17 After incubation with the primary antibody, the slides were washed for 5 minutes with TBS and incubated with corresponding biotinylated secondary antibody for 30 minutes. Slides were washed for 5 minutes with TBS and incubated with the alkaline phosphatase-conjugated ABC reagent (Vector Laboratories, Inc., Burlingame, CA) for 30 minutes. After washing with TBS for 5 minutes, the slides were incubated with Vector Red substrate containing levamisole for 30 minutes for color development. Finally, the sections were counterstained with methyl green, mounted with crystal mount, and observed by light microscopy. Negative controls were performed by omitting the primary or secondary antibody.
HEPATOLOGY, November 2002
PBC, PSC, and normal hepatic specimens. The expression of GAPDH was also performed in parallel in efforts to confirm that the tissue sections contained undegraded mRNA in all liver samples used for the quantitative realtime PCR. As seen in Fig. 1, although ITF mRNA was readily detectable (although to varying levels) in each of the PBC and PSC RNA preparations, minimal to undetectable levels were noted in the control specimens. In efforts to derive quantitative data, real-time PCRs were performed using RNA preparations from 20 PBC, 16 PSC, 10 normal, 8 alcoholic cirrhosis, 8 HCV-related liver disease, and 8 cryptogenic cirrhosis hepatic frozen specimens. A standard curve, with good correlation, was obtained (Fig. 2); the melting curves of the PCR products demonstrated that there were no nonspecific products or primer dimers. The relative levels of the expression of ITF (mean ⫾ SEM) are shown in Fig. 3. In each liver disease, ITF gene expression level was higher than that in normal controls. Interestingly, the expression level of ITF mRNA in PBC livers (1.56 ⫾ 0.32) was significantly higher compared with normal control livers (0.15 ⫾ 0.02) (P ⬍ .001) and PSC (0.72 ⫾ 0.13) (P ⬍ .05). The expression level of ITF mRNA in HCV-related liver disease (1.33 ⫾ 0.25) (P ⬍ .05) or in cryptogenic cirrhosis (1.45 ⫾ 0.21) (P ⬍ .01) was also significantly higher compared with normal control livers (0.15 ⫾ 0.02). In Situ Hybridization for ITF mRNA Twenty PBC, 16 PSC, 10 normal, 8 alcoholic cirrhosis, 8 HCV-related liver disease, and 8 cryptogenic cirrho-
Statistical Analysis StatView 4.0 was used to ascertain the significance of results obtained by real-time PCR. The frequency of cells positive by in situ nucleic acid hybridization and immunohistochemical staining was analyzed using the Fisher’s exact probability test.
Results Detection of ITF mRNA by Real-Time PCR As a preliminary experiment for detection of ITF gene usage in the liver, standard PCR was performed using
Fig. 2. (A) Amplification plots of ITF gene. Four-fold serial dilutions of cDNA, including mRNA of ITF, were PCR amplified using the GeneAmp 5700 Sequence Detection System. (B) Standard curve for ITF, generated by plotting CT values (threshold cycle). Slope: ⫺3.066, Intercept: 29.338, Correlation: 0.997.
HEPATOLOGY, Vol. 36, No. 5, 2002
Fig. 3. Quantitative analysis of liver ITF mRNA expression (mean ⫾ SEM). PBC, normal, PSC, alcoholic cirrhosis, HCV-related liver disease, and cryptogenic cirrhosis were analyzed by real-time PCR. Expression levels of ITF mRNA were amplified and compared between each of the 2 groups. PBC vs. PSC (P ⬍ .05); PBC vs. normal (P ⬍ .001); HCV vs. normal (P ⬍ .05); and cryptogenic cirrhosis vs. normal (P ⬍ .01).
Fig. 4. Light micrographs of in situ hybridization of ITF mRNA. (A and B) PBC: The large bile duct BECs of PBC intensely express ITF mRNA (A; arrowheads). In contrast, in the small bile duct, BECs of PBC have very weak signals (B). (C) Normal: There are no positive signals in the large bile duct BECs. (D) PSC: Positive signals (arrowheads) are shown in large bile duct BECs. (E) Cryptogenic cirrhosis: Positive signals (arrowheads) are shown in small bile duct BECs. (F) Alcoholic cirrhosis: There are no positive signals in the small bile duct of BECs (arrows). (Original magnification ⫻400.)
KIMURA ET AL.
1231
sis hepatic frozen specimens were analyzed for localization and differential expression of ITF mRNA by in situ nucleic acid hybridization. Using an antisense probe against ITF mRNA, specific signals were localized, as expected, in bile duct BECs (Fig. 4). A total of 20 PBC liver sections were examined for in situ nucleic acid hybridization with ITF, and large, including septal, bile duct BECs were present in 16 of 20, small (interlobular and bile ductule) bile duct BECs were present in 18 of 20. Finally, large bile ducts were positive in 13 of 16 cases (81%), and small bile ducts were positive in 6 of 18 cases (33%) in PBC patients (Table 2). The frequency of positive staining in large bile ducts of PBC was significantly higher than that seen in small bile ducts of PBC (P ⬍ .005). In contrast, this differential expression was not shown in either normal or liver diseased controls. In control livers, 41% of large bile ducts and 26% of small bile ducts expressed ITF mRNA.
1232
KIMURA ET AL.
HEPATOLOGY, November 2002
Table 2. Expression of ITF mRNA in Biliary Epithelial Cells by In Situ Hybridization Positive/Total (%)
PBC Controls Normal PSC Alcoholic HCV Crypotogenic cirrhosis
Small Bile Ducts* (%)
Large Bile Ducts† (%)
P Value‡
6/18 (33) 13/50 (26) 1/10 (10) 7/16 (43) 2/8 (25) 1/8 (12) 2/8 (25)
13/16 (81)§ 15/36 (41) 3/10 (30) 4/9 (44) 2/5 (40) 3/6 (50) 3/6 (50)
⬍.005 .12 .26 .97 .56 .12 .33
*Small: interlobular bile duct and bile ductule. †Large: intrahepatic large and septal bile ducts. ‡P value: small versus large bile ducts, using Fisher’s exact probability test. §Note also that P ⬍ .001 when comparing large bile ducts of PBC with controls.
The frequency of positive staining in large bile ducts of PBC (81%) was significantly higher than that seen in large bile ducts of controls (41%) (P ⬍ .001). In addition, of particular interest, 4 of 16 (25%) PBC large bile ducts demonstrated markedly increased levels of expression of ITF mRNA (Fig. 4A), whereas this increased level of expression was not noted in either the larger bile duct BECs of controls or the smaller bile duct BECs of PBC or normal controls. Importantly, specificity for the in situ hybridization technique was demonstrated by the observation that liver tissue sections run in parallel and showed no detectable specific signal when hybridized with the sense probe or without the probe. The Expression and Localization of ITF Protein by Immunohistochemistry The localization and expression of ITF protein were further characterized by immunohistochemistry using PBC and control paraffin-embedded hepatic sections (Table 3). A total of 40 PBC liver sections were examined for immunohistochemistry with ITF, and large bile duct BECs were present in 13 of 40, small bile duct BECs were present in 36 of 40. The reduced number of large bile ducts was due to the use of some needle biopsy specimens. Large bile ducts were positive in 12 of 13 cases (92%) (Fig. 5A), and small bile ducts were positive in only 1 of 36 cases (2%) in PBC. In control livers, 32% of large bile ducts and 18% of small bile ducts expressed ITF protein. The frequency of positive staining in large bile of PBC (92%) was significantly higher than that seen in large bile ducts of controls (32%) (P ⬍ .0001). In agreement with the results of in situ nucleic acid hybridization, in PBC livers, the frequency of positive staining in large bile ducts (92%) was significantly higher than that in small bile ducts (2%) (P ⬍ .0001). Interestingly, by immunohisto-
chemistry, this marked differential expression of ITF between large and small bile ducts was shown only in PBC patients. We also studied the distribution by immunohistochemistry of ITF in stage I, II, compared with stage III, IV, PBC liver. Essentially, there was no difference. Thus, for example, 0% (0/24) of small bile ducts of early stage PBC liver stained for ITF compared with 8% (1/12) of patients with stage III, IV. Similarly, there was no difference in ITF staining of large bile ducts: 100% (4/4) staining in early PBC and 92% (8/9) in advanced PBC.
Discussion ITF gene expression has shown to be induced at the ulcerative and inflammatory margins of the gastrointestinal tracts in inflammatory bowel disease.6 The ulcer-associated cell lineage occurs at the sites of chronic ulcerative conditions of the gut, such as Crohn’s disease, and is associated with the expression of all 3 trefoil peptides including ITF. Because trefoil proteins are commonly associated with protected and regenerative events in epithelial cells, the data herein suggest that such high expression in the liver is indicative of the potential induction of signal within the cellular elements of the liver of these patients to exert or respond to “danger” signals and to protect the liver from injury. Clearly, this has also been noted in salivary gland epithelial cells in Sjogren’s syndrome.18 Indeed, ITF expression has also been demonstrated in salivary glands,19 respiratory tract,20,21 and uterus.22,23 Interestingly, we found that the ITF mRNA expression level in PBC liver was significantly higher than that in PSC. Both PBC and PSC are organ-specific autoimmune diseases with chronic progressive cholestasis, belonging to the so-called bile duct loss syndromes, leading to cholestatic cirrhosis. PSC is histologically characterized by periductal fibrosis and ductal loss in severely affected ducts24 Table 3. Immunohistochemical Analysis of ITF in PBC and Control Livers Positive/Total (%)
PBC Controls Normal PSC Alcoholic HCV Cryptogenic cirrhosis
Small Bile Ducts* (%)
Large Bile Ducts† (%)
P Value‡
1/36 (2) 14/79 (18) 4/25 (16) 3/20 (15) 3/11 (27) 2/13 (15) 2/10 (20)
12/13 (92)§ 21/66 (32) 9/24 (37) 6/17 (35) 2/9 (22) 2/8 (25) 2/8 (25)
.0001 .048 .16 .15 .79 .79 .79
*Small: interlobular bile duct and bile ductule. †Large: intrahepatic large and septal bile ducts. ‡P value: small versus large bile ducts, using Fisher’s exact probability test. §Note also that P ⬍ .001 when comparing large bile ducts of PBC with controls.
HEPATOLOGY, Vol. 36, No. 5, 2002
KIMURA ET AL.
1233
Fig. 5. Expression of ITF protein. (A and B) PBC: In large bile duct BECs of PBC, there is strong expression (arrow) of ITF (A). In contrast, the small bile duct BECs of PBC do not express (arrow) ITF (B). (C) Normal: The small bile duct BECs of normal liver do not demonstrate positive staining (arrow) for ITF. (D) PSC: The small bile duct BECs of PSC demonstrate diffuse staining (arrow). (E) Cryptogenic cirrhosis: Small bile duct BECs demonstrate strong staining (arrowhead) for ITF. (F) HCV-related liver disease: A portion of the large bile duct BECs of HCV-related liver disease demonstrate strong staining (arrowheads). Diffuse staining is noted throughout the rest of BECs. (Original magnification ⫻400.)
but not by primary destruction of BECs. The periductal environment, including bile duct ischemia through the peribiliary vascular plexus, is thought to be important for the pathogenesis of PSC.25 On the other hand, ductal lesions in PBC are characterized by immune-mediated destruction of BECs. The unique immunologic damage of BECs in PBC may induce ITF, resulting in higher expression compared with PSC. Recently, it has been reported that Cox-2 is expressed at higher levels in bile duct BECs in PSC compared with PBC,26 and expression of HSP70 and ubiquitin is reported to be significantly increased in BECs of PBC compared with PSC.27 Thus, this data also suggest that there may be distinct heterogeneity among bile duct BECs of PBC and PSC in response to mucosal damage. We report herein the differential expression of ITF between small and large bile ducts in PBC patients. This
preferential production of ITF protein was seen in the large bile ducts, including intrahepatic large and septal bile ducts of PBC patients. In the gastrointestinal tract, ITF is selectively expressed in high concentrations by mucus-producing goblet cells of the small and large intestine28 and is induced along with mucins after damage in the foregut.29 Such expression also implies important interactions between the trefoil peptides and mucin glycoproteins.30 In liver, mucin is most abundant in large bile ducts.31 These observations are compatible with our results. Mucin is composed of carbohydrates and 9 apomucins (MUC 1, 2, 3, 4, 5AC, 5B, 6, 7, 8), and colocalization of ITF and MUC2 has been demonstrated throughout the small and large bowel mucosa.32 Moreover, MUC2 expression is elevated in hepatolithiasis, with chronic proliferating cholangitis of large bile ducts.33 On the other hand, ITF expression in small bile ducts, such as
1234
KIMURA ET AL.
interlobular and bile ductules, was absent; interlobular bile ducts are the primary site of destruction in PBC. It is of interest to note that the transcription of ITF message is highly regulated and that cell-specific expression of ITF via the silencer region in the ITF gene, including activation of a silencer inhibitor, has been demonstrated in intestinal goblet cell lines.34 Ineffective expression and/or mutations within this gene may be one of the mechanisms for the regulation of ITF by PBC patients. Further studies should also be directed on whether there is preferential production of ITF protein in different calibers of large bile ducts. For example, intrahepatic large bile ducts are normally spared in PBC. In contrast, the large septal bile ducts may sometimes become involved. Additional studies aimed at a more careful analysis of the expression of ITF and its regulated gene mutation using DNA prepared from various levels of the biliary system in PBC patients are needed to address this issue. In this respect, we also note recent data that apoptotic cholangiocytes are a potential source of immunogenic PDC-E235,36; the caspase family of proteolytic enzymes may have the potential to generate immunogenic fragments.37 These observations are important because ITF may play a role in the regulation of intestinal apoptosis.15,16,38 The absence or lack of response of ITF may lead to small bile duct imbalance of homeostasis caused by apoptosis. Further studies are needed to understand the molecular mechanism in regulating ITF expression and the role of ITF with mucins and other epithelial cell protective mechanisms.
References 1. Scheuer PJ. Primary biliary cirrhosis: diagnosis, pathology, and pathogenesis [Review]. Postgrad Med J 1983;4:106-115. 2. Rubin E, Schaffner F, Popper H. Primary biliary cirrhosis. Am J Pathol 1965;46:387-407. 3. Kaplan MM, Gandolfo JV, Quaroni EG. An enzyme-linked immunosorbent assay (ELISA) for detecting antimitochondrial antibody. HEPATOLOGY 1984;4:727-730. 4. Gum JR Jr, Hicks JW, Gillespie AM, Carlson EJ, Komuves L, Karnik S, Hong JC, et al. Goblet cell-specific expression mediated by the MUC2 mucin gene promoter in the intestine of transgenic mice. Am J Physiol 1999;276:G666-G676. 5. Tomasetto C, Masson R, Linares JL, Wendling C, Lefebvre O, Chenard MP, Rio MC. pS2/TFF1 interacts directly with the VWFC cysteine-rich domains of mucins. Gastroenterology 2000;118:70-80. 6. Wright NA, Poulsom R, Stamp G, Van Noorden S, Sarraf C, Elia G, Ahnen D, et al. Trefoil peptide gene expression in gastrointestinal epithelial cells in inflammatory bowel disease. Gastroenterology 1993;104:1220. 7. Tomasetto C, Rio MC, Gautier C, Wolf C, Hareuveni M, Chambon P, Lathe R. hSP, the domain-duplicated homolog of pS2 protein, is co-expressed with pS2 in stomach but not in breast carcinoma. EMBO J 1990; 9:407-414. 8. Jorgensen KH, Thim L, Jacobsen HE. Pancreatic spasmolytic polypeptide (PSP): I. Preparation and initial chemical characterization of a new polypeptide from porcine pancreas. Regul Pept 1982;3:207-219.
HEPATOLOGY, November 2002
9. Masiakowski P, Breathnach R, Bloch J, Gannon F, Krust A, Chambon P. Cloning of cDNA sequences of hormone-regulated genes from the MCF-7 human breast cancer cell line. Nucleic Acids Res 1982;10:7895-7903. 10. Suemori S, Lynch-Devaney K, Podolsky DK. Identification and characterization of rat intestinal trefoil factor: tissue- and cell-specific member of the trefoil protein family. Proc Natl Acad Sci U S A 1991;88:11017-11021. 11. Podolsky DK, Lynch-Devaney K, Stow JL, Oates P, Murgue B, De-Beaumont M, Sands BE, et al. Identification of human intestinal trefoil factor. Goblet cell-specific expression of a peptide targeted for apical secretion. J Biol Chem 1993;268:12230. 12. Mashimo H, Wu DC, Podolsky DK, Fishman MC. Impaired defense of intestinal mucosa in mice lacking intestinal trefoil factor. Science 1996; 274:262-265. 13. Kanai M, Mullen C, Podolsky DK. Intestinal trefoil factor induces inactivation of extracellular signal-regulated protein kinase in intestinal epithelial cells. Proc Natl Acad Sci U S A 1998;95:178-182. 14. Liu D, el-Hariry I, Karayiannakis AJ, Wilding J, Chinery R, Kmiot W, McCrea PD, et al. Phosphorylation of -catenin and epidermal growth factor receptor by intestinal trefoil factor. Lab Invest 1997;77:557-563. 15. Taupin DR, Kinoshita K, Podolsky DK. Intestinal trefoil factor confers colonic epithelial resistance to apoptosis. Proc Natl Acad Sci U S A 2000; 97:799-804. 16. Kinoshita K, Taupin DR, Itoh H, Podolsky DK. Distinct pathways of cell migration and antiapoptotic response to epithelial injury: structure-function analysis of human intestinal trefoil factor. Mol Cell Biol 2000;20: 4680-4690. 17. Taupin DR, Pang KC, Green SP, Giraud AS. The trefoil peptides spasmolytic polypeptide and intestinal trefoil factor are major secretory products of the rat gut. Peptides 1995;16:1001-1005. 18. Tapinos NI, Polihronis M, Tzioufas AG, Skopouli FN. Immunopathology of Sjogren’s syndrome. Ann Med Interne (Paris) 1998;149:17-24. 19. Devine DA, High AS, Owen PJ, Poulsom R, Bonass WA. Trefoil factor expression in normal and diseased human salivary glands. Hum Pathol 2000;31:509-515. 20. Wiede A, Jagla W, Welte T, Kohnlein T, Busk H, Hoffmann W. Localization of TFF3, a new mucus-associated peptide of the human respiratory tract. Am J Respir Crit Care Med 1999;159:1330-1335. 21. dos Santos Silva E, Ulrich M, Doring G, Botzenhart K, Gott P. Trefoil factor family domain peptides in the human respiratory tract. J Pathol 2000;190:133-142. 22. Wiede A, Hinz M, Canzler E, Franke K, Quednow C, Hoffmann W. Synthesis and localization of the mucin-associated TFF-peptides in the human uterus. Cell Tissue Res 2001;303:109-115. 23. Hauser F, Poulsom R, Chinery R, Rogers LA, Hanby AM, Wright NA, Hoffmann W. hP1.B, a human P-domain peptide homologous with rat intestinal trefoil factor, is expressed also in the ulcer-associated cell lineage and the uterus. Proc Natl Acad Sci U S A 1993;90:6961-6965. 24. Lee YM, Kaplan MM. Primary sclerosing cholangitis. N Engl J Med 1995; 332:924-933. 25. Kobayashi S, Nakanuma Y, Matsui O. Intrahepatic peribiliary vascular plexus in various hepatobiliary diseases: a histological survey. Hum Pathol 1994;25:940-946. 26. Hayashi N, Yamamoto H, Hiraoka N, Dono K, Ito Y, Okami J, Kondo M, et al. Differential expression of cyclooxygenase-2 (COX-2) in human bile duct epithelial cells and bile duct neoplasm. HEPATOLOGY 2001;34:638650. 27. Sakisaka S, Koga H, Sasatomi K, Ohishi M, Kawaguchi T, Harada M, Taniguchi E, et al. Ursodeoxycholic acid reduces expression of heat shock proteins in primary biliary cirrhosis. Liver 2000;20:78-87. 28. Sands BE, Podolsky DK. The trefoil peptide family. Annu Rev Physiol 1996;58:253-273. 29. Taupin D, Pedersen J, Familari M, Cook G, Yeomans N, Giraud AS. Augmented intestinal trefoil factor (TFF3) and loss of pS2 (TFF1) expression precedes metaplastic differentiation of gastric epithelium. Lab Invest 2001;81:397-408. 30. Kindon H, Pothoulakis C, Thim L, Lynch-Devaney K, Podolsky DK. Trefoil peptide protection of intestinal epithelial barrier function: cooper-
HEPATOLOGY, Vol. 36, No. 5, 2002
31.
32.
33.
34.
ative interaction with mucin glycoprotein. Gastroenterology 1995;109: 516-523. Sasaki M, Nakanuma Y, Terada T, Kim YS. Biliary epithelial expression of MUC1, MUC2, MUC3, and MUC5/6 apomucins during intrahepatic bile duct development and maturation. An immunohistochemical study. Am J Pathol 1995;147:574-579. Longman RJ, Douthwaite J, Sylvester PA, Poulsom R, Corfield AP, Thomas MG, Wright NA. Coordinated localisation of mucins and trefoil peptides in the ulcer associated cell lineage and the gastrointestinal mucosa. Gut 2000;47:792-800. Sasaki M, Nakanuma Y, Kim YS. Expression of apomucins in the intrahepatic biliary tree in hepatolithiasis differs from that in normal liver and extrahepatic biliary obstruction. HEPATOLOGY 1998;27:54-61. Iwakiri D, Podolsky DK. A silencer inhibitor confers specific expression of
KIMURA ET AL.
35.
36.
37.
38.
1235
intestinal trefoil factor in gobletlike cell lines. Am J Physiol Gastrointest Liver Physiol 2001;280:G1114-G1123. Koga H, Sakisaka S, Ohishi M, Sata M, Tanikawa K. Nuclear DNA fragmentation and expression of Bcl-2 in primary biliary cirrhosis. HEPATOLOGY 1997;25:1077-1084. Odin JA, Huebert RC, Casciola-Rosen L, LaRusso NF, Rosen A. Bcl-2dependent oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis. J Clin Invest 2001;108:223-232. Matsumura S, Van De Water J, Kita H, Coppel RL, Tsuji T, Yamamoto K, Ansari AA, et al. Contribution to antimitochondrial antibody production: cleavage of pyruvate dehydrogenase complex-E2 by apoptosis-related proteases. HEPATOLOGY 2002;35:14-22. Potten CS. Epithelial cell growth and differentiation. II. Intestinal apoptosis. Am J Physiol 1997;273:G253-G257.
P
rimary biliary cirrhosis (PBC) is a chronic and progressive autoimmune liver disease characterized by the presence of antimitochondrial antibodies, immune destruction of intrahepatic biliary epithelial cells (BECs), and eventually cirrhosis.1-3 Histologically, interlobular bile ducts are the primary site of destruction, and Abbreviations: PBC, primary biliary cirrhosis; BECs, biliary epithelial cells; PDC-E2, E2 subunit of pyruvate dehydrogenase; ITF, intestinal trefoil factor; SP, spasmolytic polypeptide; TFF, trefoil factor family; mRNA, messenger RNA; PCR, polymerase chain reaction; PSC, primary sclerosing cholangitis; HCV, hepatitis C virus; cDNA, complementary DNA. From the 1Division of Rheumatology, Allergy and Clinical Immunology, University of California, Davis, CA; 2Third Department of Internal Medicine, Kagawa Medical University, Kagawa, Japan; 3Department of Medicine, University of Melbourne at Western Hospital, Footscray, Australia; 4Queen Elizabeth Hospital, Liver Unit, Edgbaston, Birmingham, United Kingdom; 5UMDNJ Robert Wood Johnson Medical School, Camden, NJ; 6Department of Pediatrics, University of Minnesota Medical School, Minneapolis, MN; 7Department of Pathology, Emory University, Atlanta, GA; and 8Department of Microbiology, Monash University, Victoria, Australia. Received February 4, 2002; accepted July 22, 2002. Address reprint requests to: M. Eric Gershwin, M.D., Division of Rheumatology, Allergy and Clinical Immunology, University of California at Davis, TB 192, School of Medicine, Davis, CA 95616. E-mail: [email protected]; fax: 530-752-4669. Copyright © 2002 by the American Association for the Study of Liver Diseases. 0270-9139/02/3605-0027$35.00/0 doi:10.1053/jhep.2002.36157
progressive disappearance of interlobular bile ducts in PBC is irreversible1,2 Furthermore, intrahepatic large bile ducts are not affected, and septal bile ducts do not disappear despite the presence of chronic nonsuppurative cholangitis. Although the immunodominant autoantigen, the E2 subunit of pyruvate dehydrogenase (PDCE2), is present in all nucleated cells including the large bile ducts, the reason why the interlobular bile ducts and ductules selectively become the specific target of immune destruction is unknown. One theory is that there is subtle but distinct heterogeneity among the BECs that line the different levels of bile ducts. Intestinal trefoil factor (ITF, also called TFF3), spasmolytic polypeptide (SP; SP/TFF2), and pS2/TFF1 belong to the trefoil factor family (TFF). The TFF proteins are mucin-associated protease-resistant peptides that contain 1 or more trefoil motifs composed of 6 cysteine residues and are thought to influence the rheologic properties of mucus gels and thus promote protection to the mucosal surface.4,5 In the gastrointestinal mucosa of inflammatory bowel diseases and in gastric ulceration, trefoil factors are up-regulated and appear to promote epithelial migration and wound healing.6 SP was originally identified as a human homologue7 of a peptide isolated from porcine pan1227
1228
KIMURA ET AL.
creas,8 and pS2 was discovered during studies of estrogeninducible messenger RNA (mRNA) in breast carcinoma cell lines.9 ITF was first cloned from rat intestine and is predominantly produced by goblet cells of the small and large intestine.10,11 ITF-deficient mice lack epithelial restitution following colitis induced by dextran sulfate sodium.12 Recent data also suggest that trefoil peptides enhance epithelial motility and spreading, possibly via extracellular matrix alteration, and accelerate mucosal epithelial recovery after injury.13 In addition, ITF is involved in phosphorylation of -catenin and the epithelial growth factor receptor,12,14 which is important for the normal development of intrahepatic bile ducts. Although the precise molecular mechanisms by which ITF mediates its biologic effect remain to be fully defined, one of the potential mechanisms by which its protective effect is mediated is via epithelial cell apoptosis. Epithelial cells are protected from cell death by maintenance of cellcell and cell-substratum contacts and disruption of such interaction leading to detachment, resulting in apoptosis. This thesis is further strengthened by the fact that ITF expression in a human epithelial cell line confers resistance of this cell line to undergo apoptosis, induced by serum starvation and ceramide,15 and attributed to the ability of ITF to induce distinct intracellular signaling pathways that influence both cell migration and apoptosis.16 The functional roles of the trefoil proteins TFF1, TFF2, and TFF3 prompted us to examine their expression at both the RNA and protein level in liver tissues from PBC patients. Results of our preliminary studies showed that, whereas no detectable differences were noted in the expression of TFF1 and TFF2 in liver specimens from PBC patients and controls, significant differences were noted in the expression of ITF. The studies reported herein, therefore, were focused on defining the expression of TFF3 or ITF in more detail at both the mRNA and protein level in liver.
Patients and Methods Liver Tissue Liver tissue specimens from 60 patients with PBC (20 frozen, 40 paraffin embedded) and 129 controls (50 frozen, 79 paraffin embedded) were studied. Frozen liver specimens were used for real-time quantitative polymerase chain reaction (PCR) and in situ nucleic acid hybridization, and paraffin-embedded liver specimens were used for immunohistochemistry. The PBC patients included 27 patients (2 frozen, 25 paraffin embedded) that were classified as stage I or II and 33 patients (18 frozen, 15 paraffin embedded) classified as stages III and IV. The controls included healthy liver (10 frozen, 25 paraffin embedded), primary sclerosing cholangitis (PSC) (16 fro-
HEPATOLOGY, November 2002
zen, 20 paraffin embedded), alcoholic cirrhosis (8 frozen, 11 paraffin embedded), hepatitis C virus (HCV)-related liver disease (8 frozen, 13 paraffin embedded), and cryptogenic cirrhosis (8 frozen, 10 paraffin embedded). The tissues were obtained either from patients undergoing orthotopic liver transplantation or from liver biopsy specimens from the University of California at Davis, CA; Kagawa Medical University, Japan; Queen Elizabeth Hospital, United Kingdom; the Department of Medicine, Robert Wood Johnson Medical School, Camden, NJ; and the Department of Pathology, University of Minnesota, Minneapolis, MN. The diagnosis of all patients was based on established criteria, including histologic, clinical, and laboratory analysis.1 RNA Isolation and PCR Detection RNA Extraction. Total RNA was isolated from approximately 100 mg of frozen liver tissue by the guanidinium thiocyanate-phenol-chloroform method, using Trizol reagent (Invitrogen, Carlsbad, CA). Liver tissue was homogenized in 1 mL Trizol reagent; 200 L chloroform was added and the mixture shaken vigorously and then left at room temperature for 5 minutes. After centrifugation at 12,000g (4°C) for 15 minutes, the aqueous phase was transferred to a fresh tube, and 500 L isopropanol was added. The mixture was left at room temperature for 10 minutes, followed by centrifugation at 12,000g (4°C) for 10 minutes. The RNA precipitate was finally rinsed with 75% ethanol and dried. Finally, the RNA was dissolved in 50 L of distilled water and quantitated, using a spectrophotometer, at OD260. First-Strand cDNA Synthesis. Total RNA (1 g) was used for the synthesis of single-strand complementary DNA (cDNA) with reverse transcriptase. First, total RNA was incubated for 10 minutes at 70°C with 1 L Oligo (dT) and 500 nmol/L dNTPs, quick-chilled on ice, and further incubated for 50 minutes at 42°C with 200 units reverse transcriptase (SuperScript II, Gibco-BRL, Gaithersburg, MD), 10 nmol/L dithiothreitol, and 5⫻ firststrand buffer in a final volume of 20 L. Reactions were then heat inactivated at 70°C for 15 minutes. This cDNA pool was used for standard PCR in addition to quantitation by real-time PCR. PCR Detection. As a positive control, the primers for the housekeeping gene GAPDH were used. Amplification of ITF-specific DNA fragments was accomplished by adding 1 L of the cDNA pool to a PCR mixture that contained 0.2 mmol/L dNTPs, 0.5 mol/L each of sense and antisense-specific primers (Table 1), 10⫻ PCR buffer, and 2.5 units Taq polymerase (Promega, Madison, WI) in a total volume of 50 L. We used the touchdown method for amplification, an initial denaturation at
HEPATOLOGY, Vol. 36, No. 5, 2002
KIMURA ET AL.
Table 1. Primers Sequence
TFF3-sense (12F) TFF3-antisense (79R) TFF3-antisense (234R) GAPDH-sense GAPDH-antisense
5⬘-TGGTCCTGGCCTTGCTGT-3⬘ 5⬘-GGCACACTGGTTTGCAGACA-3⬘ 5⬘-GAGGTGCCTCAGAAGGTGC-3⬘ 5⬘-AAGGTCGGAGTCAACGGATTT-3⬘ 5⬘-ACCAGAGTTAAAAGCAGCCCTG-3⬘
NOTE. 12F and 79R were used for real-time PCR amplification of ITF mRNA; 12F and 234R were used for PCR amplification of ITF mRNA to make the specific probes of in situ hybridization.
94°C for 10 minutes, and 1 cycle each with a 20-second denaturation at 94°C, a 1-minute annealing at 58°C, 56°C, 54°C, or 52°C, a 30-second extension at 72°C, followed by 30 cycles with a 20-second denaturation at 94°C, a 1-minute annealing at 50°C, and a 30-second extension at 72°C. After PCR, 2 L aliquots of the products were subjected to 2% agarose gel electrophoresis and stained with ethidium bromide. Observation of a band of predicted size on gel electrophoresis indicated the presence of mRNA in the original tissue sample. Quantitative Real-Time PCR. The expression of subtracted messages were verified by quantitative realtime PCR, using GeneAmp 5700 Sequence Detection System (PE Applied Biosystems). We employed the SYBR green I dye method of detection. Briefly, PCR reactions were performed using 96-well trays with 2.5 L 10⫻ diluted cDNA, together with 0.3 mol/L each of sense and antisense primers (Table 1), 1 unit AmpliTaq Gold (PE Applied Biosystems), 0.2 mmol/L dNTPs, 2.5 L 10⫻ PCR buffer, 1.5 L MgCl solution, and 2.5 L SYBR green dye solution (1:10,000 in DMSO, Molecular Probes) in a total volume of 25 L. The reaction mixture was preheated at 95°C for 10 minutes, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. Standard curves for ITF and GAPDH (housekeeping) were generated by using serial dilutions of cDNA, which were run at the same time, respectively, and all reaction wells were run in duplicate. Data from each well were analyzed using GeneAmp 5700 analysis software, and the transcription value of the target gene was obtained by plotting on a standard curve. The value of each amplified gene was normalized by the amount of GAPDH quantified from the same sample. In Situ Hybridization Preparation of RNA Probe. We amplified the specific sequence of ITF (219 bp) using the PCR, and the amplified products were subcloned into the PCR IITOPO vector using the TA cloning kit (Invitrogen, Carlsbad, CA). The subcloned DNA sequence was confirmed utilizing the DBS Automated DNA Sequencing
1229
Facility at the University of California at Davis. After linearization of the plasmid with appropriate restriction enzymes, single-strand RNA probes (antisense and sense) were obtained by in vitro transcription with sp6- and T7polymerases, and these probes were labeled with digoxigen. Efficient labeling of RNA probes was confirmed by dot blotting of ascending dilutions of probe nylon membranes. Preparation of Liver Section. The liver specimens were snap-frozen in OCT compound and stored at ⫺70°C until use. Frozen 8-m sections were cut using a cryostat and mounted on poly-L-lysine-coated slides. The sections were fixed with 4% paraformaldehyde for 30 minutes at room temperature and incubated with 10 g/mL proteinase K for 10 minutes at 37°C. After postfixation in 4% paraformaldehyde, the slides were immersed in 0.2 N HCl for 20 minutes and acetylated with acetic anhydride for 10 minutes at room temperature. The slides were dehydrated with 100% ethanol for 5 minutes and air-dried. Hybridization. Pretreated sections were incubated with hybridization solution (50% formamide, 5⫻ SSPE, 0.1% sodium dodecyl sulfate, 1⫻ Denhardt’s solution, 5% dextran sulfate, 50 mmol/L dithiothreitol, 100 g/mL yeast transfer RNA, and 100 g/mL herring sperm DNA) containing digoxigen-labeled antisense or sense RNA probe in a moist chamber at 52°C overnight. After hybridization, the sections were washed twice with 2⫻ SSC, then incubated in 10 g/mL RNase A for 30 minutes at 37°C. The slides were washed with 2⫻ SSC twice, 0.1⫻ SSC for 30 minutes at 67°C, and 0.1⫻ SSC for 10 minutes at room temperature. Hybridized sections were washed with washing buffer (0.1 mol/L maleic acid, 0.15 mol/L NaCl, pH 9.5) briefly, blocked with 1% blocking solution for 30 minutes, followed by incubation with alkaline phosphatase-conjugated sheep anti-digoxigen antibody (1:500) for 1 hour. After washing with washing buffer twice, the sections were incubated with NBT/BCIP substrate diluted with detection buffer (100 mmol/L Tris-HCl, 100 mmol/L NaCl, 50 mmol/L MgCl, pH 9.5) for 2 or 3 days in a dark room. The color reaction was stopped with TE buffer (10 mmol/L TrisHCl, pH 8.0, 1 mmol/L EDTA). Finally, the sections were counterstained with Nuclear Red, mounted with crystal mount, and then observed by light microscopy. Immunohistochemical Staining for ITF Immunohistochemical staining of ITF was performed on formalin-fixed, paraffin-embedded sections, using a standard avidin-biotin complex method. Tissue sections were first deparaffinized in xylene and rehydrated. The sections were then blocked with TBS (20 mmol/L Tris-
1230
KIMURA ET AL.
Fig. 1. ITF gene expression in PBC, PSC, and normal liver. Total RNA isolated from liver specimens was reverse transcribed and amplified by PCR with specific primers. As a positive control, the housekeeping gene GAPDH was also amplified.
HCl, pH 8.0, 150 mmol/L NaCl) containing 10% goat serum for 20 minutes at room temperature, then incubated with ITF antibodies diluted at a previously defined optimal dilution of 1:1,000 for 1 hour at room temperature.17 After incubation with the primary antibody, the slides were washed for 5 minutes with TBS and incubated with corresponding biotinylated secondary antibody for 30 minutes. Slides were washed for 5 minutes with TBS and incubated with the alkaline phosphatase-conjugated ABC reagent (Vector Laboratories, Inc., Burlingame, CA) for 30 minutes. After washing with TBS for 5 minutes, the slides were incubated with Vector Red substrate containing levamisole for 30 minutes for color development. Finally, the sections were counterstained with methyl green, mounted with crystal mount, and observed by light microscopy. Negative controls were performed by omitting the primary or secondary antibody.
HEPATOLOGY, November 2002
PBC, PSC, and normal hepatic specimens. The expression of GAPDH was also performed in parallel in efforts to confirm that the tissue sections contained undegraded mRNA in all liver samples used for the quantitative realtime PCR. As seen in Fig. 1, although ITF mRNA was readily detectable (although to varying levels) in each of the PBC and PSC RNA preparations, minimal to undetectable levels were noted in the control specimens. In efforts to derive quantitative data, real-time PCRs were performed using RNA preparations from 20 PBC, 16 PSC, 10 normal, 8 alcoholic cirrhosis, 8 HCV-related liver disease, and 8 cryptogenic cirrhosis hepatic frozen specimens. A standard curve, with good correlation, was obtained (Fig. 2); the melting curves of the PCR products demonstrated that there were no nonspecific products or primer dimers. The relative levels of the expression of ITF (mean ⫾ SEM) are shown in Fig. 3. In each liver disease, ITF gene expression level was higher than that in normal controls. Interestingly, the expression level of ITF mRNA in PBC livers (1.56 ⫾ 0.32) was significantly higher compared with normal control livers (0.15 ⫾ 0.02) (P ⬍ .001) and PSC (0.72 ⫾ 0.13) (P ⬍ .05). The expression level of ITF mRNA in HCV-related liver disease (1.33 ⫾ 0.25) (P ⬍ .05) or in cryptogenic cirrhosis (1.45 ⫾ 0.21) (P ⬍ .01) was also significantly higher compared with normal control livers (0.15 ⫾ 0.02). In Situ Hybridization for ITF mRNA Twenty PBC, 16 PSC, 10 normal, 8 alcoholic cirrhosis, 8 HCV-related liver disease, and 8 cryptogenic cirrho-
Statistical Analysis StatView 4.0 was used to ascertain the significance of results obtained by real-time PCR. The frequency of cells positive by in situ nucleic acid hybridization and immunohistochemical staining was analyzed using the Fisher’s exact probability test.
Results Detection of ITF mRNA by Real-Time PCR As a preliminary experiment for detection of ITF gene usage in the liver, standard PCR was performed using
Fig. 2. (A) Amplification plots of ITF gene. Four-fold serial dilutions of cDNA, including mRNA of ITF, were PCR amplified using the GeneAmp 5700 Sequence Detection System. (B) Standard curve for ITF, generated by plotting CT values (threshold cycle). Slope: ⫺3.066, Intercept: 29.338, Correlation: 0.997.
HEPATOLOGY, Vol. 36, No. 5, 2002
Fig. 3. Quantitative analysis of liver ITF mRNA expression (mean ⫾ SEM). PBC, normal, PSC, alcoholic cirrhosis, HCV-related liver disease, and cryptogenic cirrhosis were analyzed by real-time PCR. Expression levels of ITF mRNA were amplified and compared between each of the 2 groups. PBC vs. PSC (P ⬍ .05); PBC vs. normal (P ⬍ .001); HCV vs. normal (P ⬍ .05); and cryptogenic cirrhosis vs. normal (P ⬍ .01).
Fig. 4. Light micrographs of in situ hybridization of ITF mRNA. (A and B) PBC: The large bile duct BECs of PBC intensely express ITF mRNA (A; arrowheads). In contrast, in the small bile duct, BECs of PBC have very weak signals (B). (C) Normal: There are no positive signals in the large bile duct BECs. (D) PSC: Positive signals (arrowheads) are shown in large bile duct BECs. (E) Cryptogenic cirrhosis: Positive signals (arrowheads) are shown in small bile duct BECs. (F) Alcoholic cirrhosis: There are no positive signals in the small bile duct of BECs (arrows). (Original magnification ⫻400.)
KIMURA ET AL.
1231
sis hepatic frozen specimens were analyzed for localization and differential expression of ITF mRNA by in situ nucleic acid hybridization. Using an antisense probe against ITF mRNA, specific signals were localized, as expected, in bile duct BECs (Fig. 4). A total of 20 PBC liver sections were examined for in situ nucleic acid hybridization with ITF, and large, including septal, bile duct BECs were present in 16 of 20, small (interlobular and bile ductule) bile duct BECs were present in 18 of 20. Finally, large bile ducts were positive in 13 of 16 cases (81%), and small bile ducts were positive in 6 of 18 cases (33%) in PBC patients (Table 2). The frequency of positive staining in large bile ducts of PBC was significantly higher than that seen in small bile ducts of PBC (P ⬍ .005). In contrast, this differential expression was not shown in either normal or liver diseased controls. In control livers, 41% of large bile ducts and 26% of small bile ducts expressed ITF mRNA.
1232
KIMURA ET AL.
HEPATOLOGY, November 2002
Table 2. Expression of ITF mRNA in Biliary Epithelial Cells by In Situ Hybridization Positive/Total (%)
PBC Controls Normal PSC Alcoholic HCV Crypotogenic cirrhosis
Small Bile Ducts* (%)
Large Bile Ducts† (%)
P Value‡
6/18 (33) 13/50 (26) 1/10 (10) 7/16 (43) 2/8 (25) 1/8 (12) 2/8 (25)
13/16 (81)§ 15/36 (41) 3/10 (30) 4/9 (44) 2/5 (40) 3/6 (50) 3/6 (50)
⬍.005 .12 .26 .97 .56 .12 .33
*Small: interlobular bile duct and bile ductule. †Large: intrahepatic large and septal bile ducts. ‡P value: small versus large bile ducts, using Fisher’s exact probability test. §Note also that P ⬍ .001 when comparing large bile ducts of PBC with controls.
The frequency of positive staining in large bile ducts of PBC (81%) was significantly higher than that seen in large bile ducts of controls (41%) (P ⬍ .001). In addition, of particular interest, 4 of 16 (25%) PBC large bile ducts demonstrated markedly increased levels of expression of ITF mRNA (Fig. 4A), whereas this increased level of expression was not noted in either the larger bile duct BECs of controls or the smaller bile duct BECs of PBC or normal controls. Importantly, specificity for the in situ hybridization technique was demonstrated by the observation that liver tissue sections run in parallel and showed no detectable specific signal when hybridized with the sense probe or without the probe. The Expression and Localization of ITF Protein by Immunohistochemistry The localization and expression of ITF protein were further characterized by immunohistochemistry using PBC and control paraffin-embedded hepatic sections (Table 3). A total of 40 PBC liver sections were examined for immunohistochemistry with ITF, and large bile duct BECs were present in 13 of 40, small bile duct BECs were present in 36 of 40. The reduced number of large bile ducts was due to the use of some needle biopsy specimens. Large bile ducts were positive in 12 of 13 cases (92%) (Fig. 5A), and small bile ducts were positive in only 1 of 36 cases (2%) in PBC. In control livers, 32% of large bile ducts and 18% of small bile ducts expressed ITF protein. The frequency of positive staining in large bile of PBC (92%) was significantly higher than that seen in large bile ducts of controls (32%) (P ⬍ .0001). In agreement with the results of in situ nucleic acid hybridization, in PBC livers, the frequency of positive staining in large bile ducts (92%) was significantly higher than that in small bile ducts (2%) (P ⬍ .0001). Interestingly, by immunohisto-
chemistry, this marked differential expression of ITF between large and small bile ducts was shown only in PBC patients. We also studied the distribution by immunohistochemistry of ITF in stage I, II, compared with stage III, IV, PBC liver. Essentially, there was no difference. Thus, for example, 0% (0/24) of small bile ducts of early stage PBC liver stained for ITF compared with 8% (1/12) of patients with stage III, IV. Similarly, there was no difference in ITF staining of large bile ducts: 100% (4/4) staining in early PBC and 92% (8/9) in advanced PBC.
Discussion ITF gene expression has shown to be induced at the ulcerative and inflammatory margins of the gastrointestinal tracts in inflammatory bowel disease.6 The ulcer-associated cell lineage occurs at the sites of chronic ulcerative conditions of the gut, such as Crohn’s disease, and is associated with the expression of all 3 trefoil peptides including ITF. Because trefoil proteins are commonly associated with protected and regenerative events in epithelial cells, the data herein suggest that such high expression in the liver is indicative of the potential induction of signal within the cellular elements of the liver of these patients to exert or respond to “danger” signals and to protect the liver from injury. Clearly, this has also been noted in salivary gland epithelial cells in Sjogren’s syndrome.18 Indeed, ITF expression has also been demonstrated in salivary glands,19 respiratory tract,20,21 and uterus.22,23 Interestingly, we found that the ITF mRNA expression level in PBC liver was significantly higher than that in PSC. Both PBC and PSC are organ-specific autoimmune diseases with chronic progressive cholestasis, belonging to the so-called bile duct loss syndromes, leading to cholestatic cirrhosis. PSC is histologically characterized by periductal fibrosis and ductal loss in severely affected ducts24 Table 3. Immunohistochemical Analysis of ITF in PBC and Control Livers Positive/Total (%)
PBC Controls Normal PSC Alcoholic HCV Cryptogenic cirrhosis
Small Bile Ducts* (%)
Large Bile Ducts† (%)
P Value‡
1/36 (2) 14/79 (18) 4/25 (16) 3/20 (15) 3/11 (27) 2/13 (15) 2/10 (20)
12/13 (92)§ 21/66 (32) 9/24 (37) 6/17 (35) 2/9 (22) 2/8 (25) 2/8 (25)
.0001 .048 .16 .15 .79 .79 .79
*Small: interlobular bile duct and bile ductule. †Large: intrahepatic large and septal bile ducts. ‡P value: small versus large bile ducts, using Fisher’s exact probability test. §Note also that P ⬍ .001 when comparing large bile ducts of PBC with controls.
HEPATOLOGY, Vol. 36, No. 5, 2002
KIMURA ET AL.
1233
Fig. 5. Expression of ITF protein. (A and B) PBC: In large bile duct BECs of PBC, there is strong expression (arrow) of ITF (A). In contrast, the small bile duct BECs of PBC do not express (arrow) ITF (B). (C) Normal: The small bile duct BECs of normal liver do not demonstrate positive staining (arrow) for ITF. (D) PSC: The small bile duct BECs of PSC demonstrate diffuse staining (arrow). (E) Cryptogenic cirrhosis: Small bile duct BECs demonstrate strong staining (arrowhead) for ITF. (F) HCV-related liver disease: A portion of the large bile duct BECs of HCV-related liver disease demonstrate strong staining (arrowheads). Diffuse staining is noted throughout the rest of BECs. (Original magnification ⫻400.)
but not by primary destruction of BECs. The periductal environment, including bile duct ischemia through the peribiliary vascular plexus, is thought to be important for the pathogenesis of PSC.25 On the other hand, ductal lesions in PBC are characterized by immune-mediated destruction of BECs. The unique immunologic damage of BECs in PBC may induce ITF, resulting in higher expression compared with PSC. Recently, it has been reported that Cox-2 is expressed at higher levels in bile duct BECs in PSC compared with PBC,26 and expression of HSP70 and ubiquitin is reported to be significantly increased in BECs of PBC compared with PSC.27 Thus, this data also suggest that there may be distinct heterogeneity among bile duct BECs of PBC and PSC in response to mucosal damage. We report herein the differential expression of ITF between small and large bile ducts in PBC patients. This
preferential production of ITF protein was seen in the large bile ducts, including intrahepatic large and septal bile ducts of PBC patients. In the gastrointestinal tract, ITF is selectively expressed in high concentrations by mucus-producing goblet cells of the small and large intestine28 and is induced along with mucins after damage in the foregut.29 Such expression also implies important interactions between the trefoil peptides and mucin glycoproteins.30 In liver, mucin is most abundant in large bile ducts.31 These observations are compatible with our results. Mucin is composed of carbohydrates and 9 apomucins (MUC 1, 2, 3, 4, 5AC, 5B, 6, 7, 8), and colocalization of ITF and MUC2 has been demonstrated throughout the small and large bowel mucosa.32 Moreover, MUC2 expression is elevated in hepatolithiasis, with chronic proliferating cholangitis of large bile ducts.33 On the other hand, ITF expression in small bile ducts, such as
1234
KIMURA ET AL.
interlobular and bile ductules, was absent; interlobular bile ducts are the primary site of destruction in PBC. It is of interest to note that the transcription of ITF message is highly regulated and that cell-specific expression of ITF via the silencer region in the ITF gene, including activation of a silencer inhibitor, has been demonstrated in intestinal goblet cell lines.34 Ineffective expression and/or mutations within this gene may be one of the mechanisms for the regulation of ITF by PBC patients. Further studies should also be directed on whether there is preferential production of ITF protein in different calibers of large bile ducts. For example, intrahepatic large bile ducts are normally spared in PBC. In contrast, the large septal bile ducts may sometimes become involved. Additional studies aimed at a more careful analysis of the expression of ITF and its regulated gene mutation using DNA prepared from various levels of the biliary system in PBC patients are needed to address this issue. In this respect, we also note recent data that apoptotic cholangiocytes are a potential source of immunogenic PDC-E235,36; the caspase family of proteolytic enzymes may have the potential to generate immunogenic fragments.37 These observations are important because ITF may play a role in the regulation of intestinal apoptosis.15,16,38 The absence or lack of response of ITF may lead to small bile duct imbalance of homeostasis caused by apoptosis. Further studies are needed to understand the molecular mechanism in regulating ITF expression and the role of ITF with mucins and other epithelial cell protective mechanisms.
References 1. Scheuer PJ. Primary biliary cirrhosis: diagnosis, pathology, and pathogenesis [Review]. Postgrad Med J 1983;4:106-115. 2. Rubin E, Schaffner F, Popper H. Primary biliary cirrhosis. Am J Pathol 1965;46:387-407. 3. Kaplan MM, Gandolfo JV, Quaroni EG. An enzyme-linked immunosorbent assay (ELISA) for detecting antimitochondrial antibody. HEPATOLOGY 1984;4:727-730. 4. Gum JR Jr, Hicks JW, Gillespie AM, Carlson EJ, Komuves L, Karnik S, Hong JC, et al. Goblet cell-specific expression mediated by the MUC2 mucin gene promoter in the intestine of transgenic mice. Am J Physiol 1999;276:G666-G676. 5. Tomasetto C, Masson R, Linares JL, Wendling C, Lefebvre O, Chenard MP, Rio MC. pS2/TFF1 interacts directly with the VWFC cysteine-rich domains of mucins. Gastroenterology 2000;118:70-80. 6. Wright NA, Poulsom R, Stamp G, Van Noorden S, Sarraf C, Elia G, Ahnen D, et al. Trefoil peptide gene expression in gastrointestinal epithelial cells in inflammatory bowel disease. Gastroenterology 1993;104:1220. 7. Tomasetto C, Rio MC, Gautier C, Wolf C, Hareuveni M, Chambon P, Lathe R. hSP, the domain-duplicated homolog of pS2 protein, is co-expressed with pS2 in stomach but not in breast carcinoma. EMBO J 1990; 9:407-414. 8. Jorgensen KH, Thim L, Jacobsen HE. Pancreatic spasmolytic polypeptide (PSP): I. Preparation and initial chemical characterization of a new polypeptide from porcine pancreas. Regul Pept 1982;3:207-219.
HEPATOLOGY, November 2002
9. Masiakowski P, Breathnach R, Bloch J, Gannon F, Krust A, Chambon P. Cloning of cDNA sequences of hormone-regulated genes from the MCF-7 human breast cancer cell line. Nucleic Acids Res 1982;10:7895-7903. 10. Suemori S, Lynch-Devaney K, Podolsky DK. Identification and characterization of rat intestinal trefoil factor: tissue- and cell-specific member of the trefoil protein family. Proc Natl Acad Sci U S A 1991;88:11017-11021. 11. Podolsky DK, Lynch-Devaney K, Stow JL, Oates P, Murgue B, De-Beaumont M, Sands BE, et al. Identification of human intestinal trefoil factor. Goblet cell-specific expression of a peptide targeted for apical secretion. J Biol Chem 1993;268:12230. 12. Mashimo H, Wu DC, Podolsky DK, Fishman MC. Impaired defense of intestinal mucosa in mice lacking intestinal trefoil factor. Science 1996; 274:262-265. 13. Kanai M, Mullen C, Podolsky DK. Intestinal trefoil factor induces inactivation of extracellular signal-regulated protein kinase in intestinal epithelial cells. Proc Natl Acad Sci U S A 1998;95:178-182. 14. Liu D, el-Hariry I, Karayiannakis AJ, Wilding J, Chinery R, Kmiot W, McCrea PD, et al. Phosphorylation of -catenin and epidermal growth factor receptor by intestinal trefoil factor. Lab Invest 1997;77:557-563. 15. Taupin DR, Kinoshita K, Podolsky DK. Intestinal trefoil factor confers colonic epithelial resistance to apoptosis. Proc Natl Acad Sci U S A 2000; 97:799-804. 16. Kinoshita K, Taupin DR, Itoh H, Podolsky DK. Distinct pathways of cell migration and antiapoptotic response to epithelial injury: structure-function analysis of human intestinal trefoil factor. Mol Cell Biol 2000;20: 4680-4690. 17. Taupin DR, Pang KC, Green SP, Giraud AS. The trefoil peptides spasmolytic polypeptide and intestinal trefoil factor are major secretory products of the rat gut. Peptides 1995;16:1001-1005. 18. Tapinos NI, Polihronis M, Tzioufas AG, Skopouli FN. Immunopathology of Sjogren’s syndrome. Ann Med Interne (Paris) 1998;149:17-24. 19. Devine DA, High AS, Owen PJ, Poulsom R, Bonass WA. Trefoil factor expression in normal and diseased human salivary glands. Hum Pathol 2000;31:509-515. 20. Wiede A, Jagla W, Welte T, Kohnlein T, Busk H, Hoffmann W. Localization of TFF3, a new mucus-associated peptide of the human respiratory tract. Am J Respir Crit Care Med 1999;159:1330-1335. 21. dos Santos Silva E, Ulrich M, Doring G, Botzenhart K, Gott P. Trefoil factor family domain peptides in the human respiratory tract. J Pathol 2000;190:133-142. 22. Wiede A, Hinz M, Canzler E, Franke K, Quednow C, Hoffmann W. Synthesis and localization of the mucin-associated TFF-peptides in the human uterus. Cell Tissue Res 2001;303:109-115. 23. Hauser F, Poulsom R, Chinery R, Rogers LA, Hanby AM, Wright NA, Hoffmann W. hP1.B, a human P-domain peptide homologous with rat intestinal trefoil factor, is expressed also in the ulcer-associated cell lineage and the uterus. Proc Natl Acad Sci U S A 1993;90:6961-6965. 24. Lee YM, Kaplan MM. Primary sclerosing cholangitis. N Engl J Med 1995; 332:924-933. 25. Kobayashi S, Nakanuma Y, Matsui O. Intrahepatic peribiliary vascular plexus in various hepatobiliary diseases: a histological survey. Hum Pathol 1994;25:940-946. 26. Hayashi N, Yamamoto H, Hiraoka N, Dono K, Ito Y, Okami J, Kondo M, et al. Differential expression of cyclooxygenase-2 (COX-2) in human bile duct epithelial cells and bile duct neoplasm. HEPATOLOGY 2001;34:638650. 27. Sakisaka S, Koga H, Sasatomi K, Ohishi M, Kawaguchi T, Harada M, Taniguchi E, et al. Ursodeoxycholic acid reduces expression of heat shock proteins in primary biliary cirrhosis. Liver 2000;20:78-87. 28. Sands BE, Podolsky DK. The trefoil peptide family. Annu Rev Physiol 1996;58:253-273. 29. Taupin D, Pedersen J, Familari M, Cook G, Yeomans N, Giraud AS. Augmented intestinal trefoil factor (TFF3) and loss of pS2 (TFF1) expression precedes metaplastic differentiation of gastric epithelium. Lab Invest 2001;81:397-408. 30. Kindon H, Pothoulakis C, Thim L, Lynch-Devaney K, Podolsky DK. Trefoil peptide protection of intestinal epithelial barrier function: cooper-
HEPATOLOGY, Vol. 36, No. 5, 2002
31.
32.
33.
34.
ative interaction with mucin glycoprotein. Gastroenterology 1995;109: 516-523. Sasaki M, Nakanuma Y, Terada T, Kim YS. Biliary epithelial expression of MUC1, MUC2, MUC3, and MUC5/6 apomucins during intrahepatic bile duct development and maturation. An immunohistochemical study. Am J Pathol 1995;147:574-579. Longman RJ, Douthwaite J, Sylvester PA, Poulsom R, Corfield AP, Thomas MG, Wright NA. Coordinated localisation of mucins and trefoil peptides in the ulcer associated cell lineage and the gastrointestinal mucosa. Gut 2000;47:792-800. Sasaki M, Nakanuma Y, Kim YS. Expression of apomucins in the intrahepatic biliary tree in hepatolithiasis differs from that in normal liver and extrahepatic biliary obstruction. HEPATOLOGY 1998;27:54-61. Iwakiri D, Podolsky DK. A silencer inhibitor confers specific expression of
KIMURA ET AL.
35.
36.
37.
38.
1235
intestinal trefoil factor in gobletlike cell lines. Am J Physiol Gastrointest Liver Physiol 2001;280:G1114-G1123. Koga H, Sakisaka S, Ohishi M, Sata M, Tanikawa K. Nuclear DNA fragmentation and expression of Bcl-2 in primary biliary cirrhosis. HEPATOLOGY 1997;25:1077-1084. Odin JA, Huebert RC, Casciola-Rosen L, LaRusso NF, Rosen A. Bcl-2dependent oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis. J Clin Invest 2001;108:223-232. Matsumura S, Van De Water J, Kita H, Coppel RL, Tsuji T, Yamamoto K, Ansari AA, et al. Contribution to antimitochondrial antibody production: cleavage of pyruvate dehydrogenase complex-E2 by apoptosis-related proteases. HEPATOLOGY 2002;35:14-22. Potten CS. Epithelial cell growth and differentiation. II. Intestinal apoptosis. Am J Physiol 1997;273:G253-G257.