Tel: +66-043-348-387, Fax: +66-043-202-475
Infectious Causes of Cancer
Involvement of MMP-9 in peribiliary fibrosis and cholangiocarcinogenesis via Rac1-dependent DNA damage in a hamster model
Article first published online: 16 FEB 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 127, Issue 11, pages 2576–2587, 1 December 2010
How to Cite
Prakobwong, S., Yongvanit, P., Hiraku, Y., Pairojkul, C., Sithithaworn, P., Pinlaor, P. and Pinlaor, S. (2010), Involvement of MMP-9 in peribiliary fibrosis and cholangiocarcinogenesis via Rac1-dependent DNA damage in a hamster model. Int. J. Cancer, 127: 2576–2587. doi: 10.1002/ijc.25266
- Issue published online: 16 FEB 2010
- Article first published online: 16 FEB 2010
- Manuscript Accepted: 8 FEB 2010
- Manuscript Received: 30 SEP 2009
- The Office of the Higher Education Commission
- Thailand (Strategic Scholarships for Frontier Research Network for the Ph.D. Program Thai Doctoral degree)
- Research Team Strengthening Grant
- National Center for Genetic engineering and Biotechnology
- National Science and Technology Development Agency, Thailand
- Faculty of Medicine, Khon Kaen University, Khon Kaen, 40002, Thailand. Grant Number: I52101
- DNA damage;
- Opisthorchis viverrini;
Peribiliary fibrosis caused by chronic infection with Opisthorchis viverrini (OV) is a risk factor of cholangiocarcinoma (CCA) in northeastern Thailand. Matrix metalloproteinases (MMPs) are enzymes capable of degrading and remodeling the extracellular matrix in the process of fibrosis and carcinogenesis. We examined MMPs expression and their role in fibrogenesis and cholangiocarcinogenesis in hamsters treated with OV and N-nitrosodimethylamine (NDMA). We assessed the time profiles of MMPs, inducible nitric oxide synthase (iNOS), Rac1, α-smooth muscle actin (α-SMA) and DNA lesions (8-nitroguanine and 8-oxo-7,8-dihydro-2′-deoxyguanosine, 8-oxodG) in relation to fibrosis and CCA development. Histopathology revealed OV and NDMA synergistically induced peribiliary fibrosis time-dependently, and CCA occurred at 3 months, whereas OV or NDMA alone induced less fibrosis. Hydroxyproline levels in the liver and plasma were positively associated with the expression of collagen I and α-SMA. MMP-9 expression was significantly increased and correlated with the accumulation of myofibroblast, fibrosis levels and cholangiocarcinogenesis. MMP-9 activity was correlated with iNOS, and immunocolocalization was observed in inflammed tissues, early and invasive CCA. OV and NDMA synergistically induced MMP-9 expression in association to Rac1. In addition, Rac1 was colocalized with iNOS, and 8-nitroguanine, in inflammed tissues and CCA. Formation of 8-nitroguanine and 8-oxodG increased with tumor progression. The results suggest that MMP-9 expression is associated with the accumulation of peribiliary fibrosis in conjunction to the induction of iNOS and Rac1 that may potentiate DNA damage and cholangiocarcinogenesis.
Cholangiocarcinoma (CCA) caused by chronic infection with Opisthorchis viverrini (OV) is a major health problem in northeastern Thailand,1 where at least 4 million people are estimated to be infected.2 In addition, people in this area traditionally eat fermented food containing high concentration of potent human carcinogens nitrosamines, including N-nitrosodimethylamine (NDMA). Therefore, the combination of OV infection plus nitrosamine intake may greatly contribute to CCA development in endemic areas.1
The accumulation of fibrotic lesions significantly increases the risk of various cancers, but the molecular mechanism is still unclear. Fibrosis is the result of cellular process in tissue repairs during injury, and myofibroblasts secrete extracellular matrix (ECM) and synthesize matrix metalloproteinases (MMPs) and other ECM-degrading enzymes.3 Dysfunction of tissue repair and imbalance of synthesis and degradation of fibrotic tissue leads to the accumulation of fibrosis.4 This aberrant ECM interferes normal cell function, leading to cell proliferation and development of myofibroblast-induced inflammation and angiogenesis that facilitate carcinogenesis.5 Degradation of ECM mediated by MMPs has been associated primarily with tumor invasion and metastasis. It is now becoming clear that MMPs also play a key role in tumor development and growth.6 Overexpression of MMP-9 has been observed in CCA patients.7 Several lines of evidence indicate that MMPs regulate cell growth and survival. Moreover, MMP-1 increases the levels of cellular reactive oxygen species (ROS), leading to DNA damage by inducing the expression of Rac1 (Ras-related C3 botulinum toxin substrate 1), an essential component of ROS-producing NADPH oxidase.8 MMPs directly participate in the generation of signals that induce the proliferation of tumor cells by activating the cell surface growth factor precursors, releasing and activating latent growth factors sequestered in the ECM, and altering the structure of essential ECM components.6 These imply that the accumulation of fibrosis and continuous MMPs activation potentiates not only ECM remodeling, but also tumor development.
Therefore, we hypothesize that OV infection increases MMPs expression and peribiliary fibrosis via inducible nitric oxide synthase (iNOS) and Rac1-dependent DNA damage potentiates CCA formation and progression. Recently, we have demonstrated that OV induces chronic inflammation with a concomitant increase in periductal fibrosis in hamsters.9 We have also reported that overexpression of iNOS leads to the formation of 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG), an oxidative DNA lesion, and 8-nitroguanine, an nitrative DNA lesion, in hamsters infected with OV,10, 11 and CCA patients.12 Nitric oxide (NO) amplifies the effects of the proinflammatory cytokines on various cell types and causes damage to DNA.13 In addition, reactive nitrogen intermediates accelerate ECM destruction by MMP-9 activation.14 These mechanisms may be involved in OV-induced cholangiocarcinogenesis.
To clarify the role of MMPs in fibrosis and CCA, we examined the time profile of CCA development and fibrosis in hamsters induced by OV infection and NDMA administration. Hepatic mRNA expression of collagens, MMPs, cytokines and tissue inhibitor of MMPs (TIMPs) was assessed by quantitative RT-PCR. MMP-2, MMP-9 and urokinase plasminogen activator (uPA) activities were examined by zymography. Accumulation of collagen was assessed by Masson's Trichrome and hydroxyproline assay. The expression of Rac1, MMP-9, iNOS and α-smooth muscle actin (α-SMA) and DNA damage were examined by double immunofluorescence technique.
Material and Methods
OV metacercariae were isolated from naturally infected fish by pepsin digestion as described previously.11 Cyprinid fish obtained from Khon Kaen province, Thailand, were digested in 0.25% pepsin-HCl and OV metacercariae were isolated and counted, and viable cysts were used for hamster infection.
The Animal Ethics Committee of Khon Kaen University (AEKKU 17/2552) approved this study. Four- to six-week-old male golden hamsters were housed under conventional conditions and fed a stock diet and given water ad libitum. Animals were divided into 4 groups; normal control, either treated with NDMA or OV alone, and treated with OV + NDMA. Hamsters were treated with 50 OV metacercariae by oral inoculation and/or given with 12.5 ppm of NDMA in water ad libitum for 2 months and withdrawn thereafter. Seven hamsters were included in an experimental group and sacrificed at 21 days, and 1, 3, 4, 5 and 6 months post-treatment.
Primers for quantitative RT-PCR analysis
The oligonucleotide-specific primer pairs for identification of hamster relative genes expression of collagens I and III, MMPs-2 and -9, TIMPs-1 and -2, TNF-α, TGF-β, and GAPDH were used as described previously.9 In addition, specific primers for Rac1 and uPA identification were designed based on Genbank Nos. NM134366.1 and NM008873.2. The primers were: Rac1 (sense: 5′-GTACATCCCCACCGTCTTTG-3′; antisense: 5′-GCAGGACTCACAAGGGAAAA-3′); uPA (sense: 5′-AGGTGGAGCAGCTCATCTTG-3′; antisense: 5′-GTCTGAACCAAACGGAGCAT-3′). GAPDH was used as an endogenous control. All primer sets had a calculated annealing temperature of 55°C. The confirmation of the identity of the PCR fragment was followed by sequencing using the respective gene-specific primers with the Cy5-labeled primer as described previously.9 Sequence of uPA fragment size 173 base pair was 5′GTCTGAACCAAACGGAGCATCACCAAACCTTGGAGGCAGGCAGATGGTCTGTATGGACCTGGATGGCTGAGCACAATGACCCGTGCTGGTACGTATCTTCAGCAAGGCAATATCATTATGGTGGGCCAGGCTATCAGCGTTATAGCCTTCATGCAAGATGAGCTGCTCCACCT3′. Percentage of the nucleotide sequence identities to rat, mouse and human were 89 (Genbank No. NM013085), 90 (Genbank No. NM008873.2) and 76 (Genbank No. NM002658.3), respectively. Percentage of the amino acid sequence, deduced from the nucleotide sequence, identities to rat, mouse and human were 84 (Genbank No. NP037217.3), 89 (Genbank No. NP032899.1) and 69 (Genbank No. NP002649.1), respectively. Sequence of Rac1 fragment (177 base pairs) was 5′-GCAGGACTCACAAGGGAAAAGCAAATTAAGAACACATCTGTGTTGCGGATAGGAGAGGGGACGCAATCTGTCATAATCTTCTTGTCCAGCTGTATCCCATAAGCCCCAGATTCACTGGCTTTCCATCTACCATAAGCATTGGCAGAAGTAGTTGTCGAAGGACGGTGGGAGATGTAC-3′. Percentage of the nucleotide sequence identities to rat, mouse and human were 93 (Genbank No. NM134366), 94 (Genbank No. NM009007) and 93 (Genbank No. NM018890), respectively. Percentage of the amino acid sequence, deduced from the nucleotide sequence, identities to rat, mouse and human were 81 (Genbank No. NP599193.1), 95 (Genbank No. NP599193.1) and 85 (Genbank No. NP061485.1), respectively. In addition, sequence of other genes and percentage of the nucleotide and amino acid sequence identities to rat, mouse and human are shown as described previously.9
Isolation of mRNA and cDNA production
Approximately 150 mg of the hamster liver, from the hilar region and including second-order bile duct, was obtained for total RNA isolation using TRIzol reagent (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Total RNA was treated with 5 units of DNase (Promega, Madison, WI). Total RNA (3 μg) was reverse-transcribed into cDNA using Oligo(dT) 15 primers (Promega) following the protocol for transcription by M-MLV reverse transcriptase (Promega).
Quantitative RT-PCR analysis
Relative mRNA expression was performed by an ABI7500 thermal cycler using a SYBR Green assay. All data were analyzed using Rotor Gene 5 software (Corbett, Australia) with a cycle threshold (Ct) in the linear range of amplification and then processed by the 2−ΔΔCt method as described previously.9
The total collagen content in liver and plasma was determined by estimating hydroxyproline content using base hydrolysis for the dissolution of tissue or plasma as described previously.9 The absorbance was read at 540 nm in comparison to the cis-4-hydroxy-L-proline standard (Sigma, St. Louis, MO).
Gelatin zymography and plasminogen activator zymography
Gelatinolytic activity of liver homogenates and plasma was examined by gelatin zymography as described previously.9 The recombinant human MMP-2 (Calbiochem-Novabiochem Corporation, San Diego, CA) and MMP-9 (Sigma) were included in each gel as standard protein.
Casein zymographical analysis was employed in the presence or absence of plasminogen. The electrophoresis was performed on 10% SDS-PAGE containing 1 mg/ml casein (Sigma) in the presence or absence of 13 μg/ml pure human plasminogen (Sigma), followed by washing with 2.5% Triton X-100 solution, and incubation in activation buffer with or without 10 mM phenylmethylsulfonyl fluoride (PMSF, Sigma).
Western blot analysis
Twenty micrograms of liver homogenate was separated in 12% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Amersham Bioscience, Piscataway, NJ). Membranes were incubated with mouse monoclonal anti-Rac1 antibody (1:1,000; Abcam, Cambridge, MA), rabbit polyclonal anti-α-SMA antibody (1:1,000; Abcam), or mouse monoclonal anti-iNOS antibody (1:1,000; Santa Cruz biotechnology, CA). β-Actin was detected by incubation with mouse polyclonal anti-β-actin antibody (1:5,000; Amersham Bioscience), and used as an internal control for standardization of the expression levels of other proteins. After washing, membranes were incubated with horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit IgG antibody (Amersham Bioscience). The immunoreactive material was visualized by enhanced chemiluminescence (ECL Plus Western blotting kit, GE Healthcare Biosciences Corp., Piscataway, NJ). To verify the consistency of band intensity in Western blotting, 20-μg protein of pooled tumor homogenate was included in each run.
Immunohistochemical and histopathological study
Double immunofluorescence technique was performed in formalin-fixed and paraffin-embedded tissue as described previously.11 Deparaffined sections (5-μm thickness) were incubated with the primary antibodies; mouse monoclonal anti-Rac1 (1:100), goat monoclonal anti-MMP-9 (1:100; R&D Systems, Minneapolis, MN), rabbit polyclonal anti-iNOS (1:100), rabbit polyclonal anti-α-SMA (1:100), rabbit polyclonal anti-8-nitroguanine,10 or mouse monoclonal anti-8-oxodG (JAICA, Japan) antibody for overnight at room temperature. Then, the sections were incubated with an appropriate secondary antibodies, including Alexa Fluor® 488 goat anti-mouse IgG, Alexa Fluor 594 goat anti-rabbit IgG, Alexa Fluor 488 anti-goat IgG and Alexa Fluor 594 anti-mouse IgG (1:400 each, Invitrogen) for 3 hr at room temperature. The intensity of immunoreactivity was examined under the fluorescence microscope (Nikon, Japan).
To examine the localization of biliary cytokeratin (CK)-19, a marker of bile duct proliferation, α-SMA and MMP-9 in tissue sections, we performed immunohistochemical technique. The sections were incubated with rabbit polyclonal anti-CK-19 antibody (1:200, Abcam), rabbit polyclonal anti-α-SMA antibody (1:100), or goat monoclonal anti-MMP-9 antibody (1:100) overnight at room temperature. The sections were treated with horseradish peroxidase-conjugated rabbit anti-goat IgG antibody (1:200) or goat anti-rabbit IgG antibody (1:200, Zymed Laboratory, San Francisco, CA), and visualized with 3,3-diaminobenzidine tetrahydrochloride as a chromogen followed by nuclear staining with hematoxylin. CK-19 immunostaining was assessed morphometrically for percentage area of the liver occupied by proliferating bile duct epithelial cells. The number of α-SMA-positive cells was assessed in 10 randomly selected fields (×200) per section by 2 blinded observers from the authors as described previously.15
Histological studies were performed by conventional H & E staining, and the collagen tissue was stained by Masson's Trichrome.9 To distinguish fibrotic bands from perisinusoidal fibrosis, the sections were scored blindly by an established scoring system, where +4 is established fibrosis, +3 is advanced fibrosis, +2 is early bridging fibrosis, +1 is early pericellular fibrotic changes and 0 is no fibrosis.16
Data are presented as mean ± SD. One-way ANOVA was used to compare among groups. The percentage of cells staining positive for α-SMA was compared using the Mann–Whitney test. Pearson's correlation coefficient was used to analyze correlations for parametric data. Regression analysis was performed if data had significant relationships. Statistical analyses were performed using SPSS version 11.5. p values less than 0.05 were considered statistically significant.
Accumulation of peribiliary fibrosis and CK-19 expression related to CCA development in hamster liver
Gross appearance and histopathological changes in OV + NDMA group is shown in Figure 1a. White granules and small tumors (average size <0.5 cm) were observed at 3 and 4 months, respectively. Large tumors (average size >2 cm) with invasion to the neighboring organ through the diaphragm were observed at 5–6 months. Histopathological finding revealed that inflammation and bile duct hyperplasia were observed at 21 days and 1 month post-treatment. By 3 months, small bile ducts, lymphoid follicle and early stage of tumors were observed, and tumor progression occurred at 4–6 months. The percentage of CCA in OV + NDMA group was 71.42% (5/7) for 3 months, 85.71% (6/7), for 4 and 5 months and 100% (7/7) for 6 months. In hamsters treated with OV or NDMA alone, similar histopathological feature was observed as described previously.9, 17
The intense expression of CK-19 was found in all of bile ductule hyperplasia and neoplastic glands, and progressively increased in biliary carcinoma (Fig. 1a). The morphometric measure revealed the proliferation of bile duct epithelial cells increased in the order of hyperplasia < dysplastic tissue < invasive CCA. CK-19 expression in OV + NDMA group increased time-dependently as tumor progression and the percentage of CK-19-positive cells was 7–11% from 21 days to 1 month, 27% for early CCA at 3 months and 54, 58 and 64% for progressive CCA at 4, 5 and 6 months, respectively.
The collagen distribution in the liver tissue was evaluated by Masson's Trichrome (Fig. 1b). Accumulation of fibrotic tissue was observed at the inflammation area and the stroma of tumor in the order of OV + NDMA > OV > NDMA groups, and increased in relation to tumor progression. The grading score of fibrotic tissue in OV + NDMA group was +1 at Day 21 to 1 month, +2 at 3 months, +3 at 4 months and +4 at 5 and 6 months. Whereas, the score was 0 for normal control and NDMA group, and +1 and + 2 in OV group from 21 days to 1 month and at 3–6 months, respectively. The extent of fibrosis was prominently associated with the proliferation of bile duct epithelial cells.
Alteration in collagen genes expression and hydroxyproline levels, and myofibroblast related to CCA development
Expression of collagen I and III mRNA was assessed by quantitative RT-PCR (Fig. 2a). At 4 months, significant increase in mRNA expression of collagen I was observed in the order of OV + NDMA > OV > NDMA groups, and OV + NDMA synergistically increased its level at 5 and 6 months. In addition, expression of collagen I in the OV group, and collagen III in the OV + NDMA and OV groups peaked at 1 month.
Liver hydroxyproline level in OV + NDMA, OV and NDMA groups was significantly higher than control, and increased with time. Notably, from 3 months post-treatment, OV + NDMA synergistically increased plasma hydroxyproline level (Fig. 2b). Hydroxyproline content in the liver was positively associated with that in the plasma (r = 0.773, p = 0.0001). The levels of hydroxyproline in the liver and in the plasma were positively associated with collagen I gene expression (r = 0.700; p = 0.001 and r = 0.829; p = 0.0001, respectively)
Expression of α-SMA, a marker of myofibroblast, was assessed by indirect immunohistochemistry (Fig. 2c). Positive immunoreactivity of α-SMA was observed mainly in the endovascular cells in NDMA-treated group, and in the cytoplasm of myofibroblast at the basement of epithelial bile duct (21 days and 1 month), invasive front of inflammed tissue (3–6 months), and fibrotic area in OV-treated group. In OV + NDMA group, α-SMA expression was observed in invasive front of inflammed tissue at Day 21. The highest immunoreactivity of α-SMA expression was observed in the stroma cells of malignant tissue at 6 months postinfection.
To evaluate the expression level of α-SMA, we confirmed by Western blot (Fig. 2d), which exhibited higher levels of α-SMA in OV + NDMA group compared to normal control and NDMA group. The number of α-SMA-positive cells was significantly larger in OV + NDMA group than those in the other groups. By 3 months postinfection, the combination of OV + NDMA synergistically increased the number of α-SMA-positive cells (Fig. 2e). The number of α-SMA-positive cells was significantly associated with collagen I expression (r = 0.647, p = 0.035), grading score of fibrotic tissue (r = 0.740, p = 0.046) and the percentage of CK-19 expression, a marker of bile duct proliferation (r = 0.084, p = 0.001).
Association of MMP-9 expression with iNOS expression
Expression of iNOS and MMP-9 activity in hamsters with OV infection and/or NDMA treatment are shown in Figure 3. Expression of MMP-9 was assessed by indirect immunohistochemistry (Fig. 3a). The immunoreactivity of MMP-9 was also observed in the cytoplasm of myofibroblast at the basement of epithelial bile duct, invasive front of inflammed tissue, stroma cells, and tumor cells in OV + NDMA group. Immunoreactivity of MMP-9 was observed in the order of OV + NDMA > OV > NDMA groups. MMP-9 expression increased in parallel with α-SMA expression and CCA development.
iNOS expression was assessed by Western blot (Fig. 3b). NDMA or OV infection alone induced only slight iNOS expression, and OV + NDMA apparently increased iNOS expression (Fig. 3b). MMP-9 activity in the liver and plasma was assessed by zymography (Fig. 3c). In the liver of OV + NDMA group, MMP-9 activity at 3–6 months was higher than that in OV infection, whereas its intensity unchanged in NDMA group. In the plasma, a marked increase in MMP-9 activity was observed in OV + NDMA group from 21 days and prominently at 4 months. In NDMA group and OV group, active band of MMP-9 was observed only at 21 days, and at 21 days to 1 month, respectively. The relative MMP-9 activity in the liver was significantly correlated with iNOS expression (r = 0.630; p = 0.001).
Colocalization of iNOS with MMP-9 was observed in the cytoplasm of inflammatory cells (large cells like macrophage), and tumor cells in inflammation tissue (1 month), early CCA (3 months), invasive front area (4 months) and invasive CCA (6 months) in OV + NDMA group (Fig. 3d).
Association of MMP-9 expression with Rac1 expression
MMP-9 and Rac1 mRNA expression was assessed by quantitative RT-PCR (Fig. 4a). From 3 months, the expression of MMP-9 and Rac1 increased markedly in a time-dependent manner in OV + NDMA group. In addition, in OV group, the peak of these transcript levels was observed at 1 month. MMP-9 mRNA level was positively correlated with Rac1 expression (r = 0.532, p = 0.023).
Rac1 protein expression was assessed by Western blot (Fig. 4b). The increase in Rac1 expression was observed in OV + NDMA group from 3 months, whereas slight expression was observed in OV group, NDMA group and control. The relative MMP-9 activity in the liver was significantly correlated with Rac1 protein expression (r = 0.369, p = 0.045).
The expression of MMP-9 was observed in the cytoplasm of inflammatory cells, fibroblast, stroma cells and tumor cells in OV + NDMA group [hyperplasia (21 days to 1 month), early CCA (3 months) and malignant CCA (4–6 months)] (Fig. 4c). The expression of Rac1 was observed in the nucleus of bile duct epithelial cells, inflammatory cells, and tumor cells, except in some tumor cells of malignant tissues (Fig. 4c).
Rac1 expression and oxidative and nitrative DNA damage
Colocalization of Rac1 with α-SMA, iNOS and oxidative and nitrative DNA lesions in OV + NDMA group is shown in Figure 5. Rac1 was observed in the nucleus of epithelial cells, inflammatory cells, hyperplastic cells, and tumor cells of early and malignant CCA. The expression of α-SMA was observed in the cytoplasm of fibroblast and stroma cells. Intensity of Rac1 and α-SMA expression in malignant CCA was higher than that in early CCA and hyperplasia (Fig. 5a). Expression of iNOS in the cytoplasm and Rac1 mainly in the nucleus were colocalized in bile duct epithelial cells and inflammatory cells in hyperplasia and early and malignant CCA (Fig. 5b). 8-Nitroguanine was colocalized with Rac1 (Fig. 5c) and 8-oxodG (Fig. 5d) in the nucleus of bile duct epithelial cells, inflammatory cells and tumor cells.
Profiles of MMP-2, uPA, TIMP-1, TIMP-2, TNF-α and TGF-β mRNA expression
Profiles of MMP-2, uPA, TIMP-1, TIMP-2, TNF-α and TGF-β mRNA expression were assessed by quantitative RT-PCR (Fig. 6a). At 4–6 months, OV + NDMA synergistically induced mRNA expression of MMP-2, uPA, TIMP-1 and TIMP-2, which were associated with tumor progression. Synergistic expression of TNF-α and TGF-β by OV + NDMA was observed at 5–6 months and 6 months, respectively. In OV or NDMA-treated groups, the expression levels of MMP-2, TIMP-1, TIMP-2, TNF-α and TGF-β reached the peaks at 21 days or 1 month and at 5–6 months postinfection.
The activity of MMP-2 and uPA in homogenated liver was assessed by zymography (Fig. 6b). In OV + NDMA group, the active form of MMP-2 was increased in a time-dependent manner after 3 months, whereas only the latent form was observed in OV- or NDMA-treated groups. Plasminogen activator zymography showed a band of 55 kDa and the activity of uPA in liver tissue was found only in OV + NDMA group and increased with time after 3 months.
We examined a time profile of fibrosis and the expression of related molecules during CCA development in an animal model. The combination of OV infection and NDMA administration synergistically induced fibrosis and CCA development in hamsters in a time-dependent manner. The strong and diffuse expression of CK-19 confirms the ontogeny and tumorigenesis of bile duct epithelial cells. These findings raised an idea that fibrosis may be involved in CCA formation and progression. An increased level of hydroxyproline in the liver and plasma, and the expression of collagen I and α-SMA were associated with fibrosis. OV-induced fibrogenesis through chronic inflammation9–11 and NDMA-caused liver injury and fibrosis via immunological mechanisms18 may synergistically promote fibrogenesis and carcinogenesis. Our findings could be supported by the previous observation that primary sclerosing cholangitis, a chronic cholestatic liver disease characterized by inflammation and fibrosis of the bile ducts,19 and congenital hepatic fibrosis20 are risk factors of CCA where OV is not endemic. Likewise, in endemic area of OV infection, fibrosis in the portal areas is one characteristic, which is most common in OV-associated CCA, and may contribute to enhanced susceptibility to CCA associated with severe liver fluke infection.21 Our carcinogenesis model agrees with an animal model of thioacetamide-induced hepatic fibrosis leading to not only hepatocellular carcinoma, but also intestinal-type CCA in rats.22
Myofibroblasts are the key cells involved in repair of epithelial injuries and producing MMPs. Chronic OV infection with administration of NDMA may stimulate a pathological repair process, leading to persistent fibrotic ECM that is not completely degraded by MMPs.23 OV and NDMA synergistically induced the expression of MMP-9 and α-SMA, and the distribution of these molecules was associated with liver fibrosis and CCA development. MMP-9 may stimulate myofibroblast-mediated fibrosis leading to tumor formation, and then carcinoma-associated myofibroblasts may promote to tumor progression24 by secretion of proteases such as MMP-9.7
MMPs not only facilitate tumor progression, but also are involved in tumor formation at early stage.25 Our data showed that OV + NDMA synergistically increased MMP-9 activity in the liver at 3 months. MMP-9 activity was also prominently increased in the plasma in OV + NDMA group, suggesting that MMP-9 can be used as a diagnostic marker at the initial stage of CCA.26 An increase in MMP-9 activity may be explained by increased expression of its mediator TNF-α and activator uPA in OV + NDMA group. In addition, MMP-2 expression was prominently increased during 3–6 months in OV + NDMA group, suggesting that liver injury and fibrosis are involved in CCA progression. Moreover, TIMP-1 and TIMP-2 genes also increased prominently during 4–6 months in this group, suggesting that TIMPs inhibit MMPs activity and mediate slow degradation of ECM leading to the accumulation of fibrosis and tumor progression.27 In contrast, TGF-β expression decreased at 3–4 months and then increased at 6 months, suggesting that TGF-β is involved in immunosuppression, contributing to CCA progression.28
NO is known to activate MMPs,29 potentiating matrix degradation. Our results showed that MMP-9 and iNOS expression was commonly found in inflammatory cells and stroma cells at inflammatory areas at 21 days to 1 month, and CCA occurred at 3–6 months, suggesting that iNOS expression and NO production mediated MMP-9 activation, leading to CCA development via ECM degradation. The correlation of the expression of iNOS and MMP-9 and fibrosis in OV + NDMA-induced CCA may be supported by the model of chronic carbon tetrachloride administration in iNOS knockout mice.30 Recently, it has been reported that a specific inhibitor of iNOS can prevent thioacetamide-induced hepatic fibrosis in rats,31 and effectively reduce CCA incidence in hamsters with choledochojejunostomy and injection with a carcinogen N-nitrosobis(2-oxopropyl)amine.32 Moreover, an antisense inhibitor of MMP-9 attenuates angiogenesis, human prostate cancer cell invasion and tumorigenicity.33 These reports suggest that NO-mediated MMP-9 expression contributes to carcinogenesis.
mRNA and protein levels of MMP-9 were associated with the expression of Rac1 at the transcriptional and translation levels, implying that MMP-9 mediates Rac1 expression. Stromal cells may promote carcinoma cell invasion across basement membranes involving stimulation of MMP-9 and Rac1.34 Expression of MMPs and Rac1 is involved in ROS production leading to DNA damage.8 Notably, our result showed that Rac1 was colocalized with iNOS and 8-nitroguanine, suggesting that activation of Rac1 leads to CCA formation and progression via nitrative DNA damage. Therefore, reciprocal activation of MMP-9 and iNOS may play the key role in OV-induced carcinogenesis. NO reacts with the superoxide anion (O) to form highly reactive peroxynitrite (ONOO−), which mediates oxidative and nitrative DNA damage in OV-associated CCA,10, 11 and tissue injury is mediated by MMPs activaton.9 This idea may be explained by our results that MMP-9 activity was correlated with Rac1 leading to DNA damage. Colocalization of 8-nitroguanine with 8-oxodG was found in nucleus of bile duct epithelial cells, inflammatory cell and tumor cells, implying that genomic instability via DNA damage enhances tumor initiation, promotion and progression.35 In addition, chronic bile duct injury associated with fibrotic matrix microenvironment and alteration of tumor relevant genes may be involved in CCA.36 Moreover, NDMA is a carcinogen causing DNA methylation.37 Therefore, the combination of OV and NDMA enhances fibrosis-mediated carcinogenesis and/or fibrosis-inducing molecules cause carcinogenesis in parallel with fibrosis via MMP-9- and Rac1-dependent DNA damage.
On the basis of our results and supporting evidence, we proposed the mechanism of cholangiocarcinogenesis induced by OV infection and NDMA administration in Figure 7. OV and NDMA mediate the migration of myofibroblasts producing MMP-9. MMP-9 mediates Rac1 expression, leading to generation of reactive oxygen and nitrogen species, resulting in DNA damage. In addition, NO mediates the activation of MMP-9, which induces additional myofibroblast activation and fibrosis. Therefore, CCA may develop through continuous MMP-9- and Rac1-dependent DNA damage associated with fibrogenic process. We suggest that Rac1 and MMP-9 may serve as new therapeutic targets and potential markers of CCA at early stages.
The authors thank Helmut Bartsch, Professor of Toxicology (emeritus), University of Heidelberg for critical reading of the manuscript.
- 9Time profiles of the expression of metalloproteinases, tissue inhibitors of metalloproteases, cytokines and collagens in hamsters infected with Opisthorchis viverrini with special reference to peribiliary fibrosis and liver injury. Int J Parasitol 2009; 39: 825–35., , , , , .