The differentiation status of tumor cells, defined by histomorphological criteria, is a prognostic factor for survival of patients affected with intrahepatic cholangiocarcinoma (ICC). To strengthen the value of morphological differentiation criteria, we wished to correlate histopathological differentiation grade with expression of molecular biliary differentiation markers and of microRNAs previously shown to be dysregulated in ICC. We analysed a series of tumors that were histologically classified as well, moderately or poorly differentiated, and investigated the expression of cytokeratin 7, 19 and 903 (CK7, CK19, CK903), SRY-related HMG box transcription factors 4 and 9 (SOX4, SOX9), osteopontin (OPN), Hepatocyte Nuclear Factor-1 beta (HNF1β), Yes-associated protein (YAP), Epithelial cell adhesion molecule (EPCAM), Mucin 1 (MUC1) and N-cadherin (NCAD) by qRT-PCR and immunostaining, and of miR-31, miR-135b, miR-132, miR-200c, miR-221 and miR-222. Unexpectedly, except for subcellular location of SOX9 and OPN, no correlation was found between the expression levels of these molecular markers and histopathological differentiation grade. Therefore, our data point toward necessary caution when investigating the evolution and prognosis of ICC on the basis of cell differentiation criteria.
Citation: Demarez C, Hubert C, Sempoux C, Lemaigre FP (2016) Expression of Molecular Differentiation Markers Does Not Correlate with Histological Differentiation Grade in Intrahepatic Cholangiocarcinoma. PLoS ONE 11(6): e0157140. https://doi.org/10.1371/journal.pone.0157140
Editor: Heather Francis, Digestive Disease Research Center, Scott & White Healthcare, UNITED STATES
Received: February 9, 2016; Accepted: May 25, 2016; Published: June 9, 2016
Copyright: © 2016 Demarez et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The work was supported by the FRS-FNRS (Belgium; grant T.0072.14; http://www.fnrs.be), the Interuniversity Attraction Pole Programme (Belgian Science Policy BELSPO; grant PVII-47; http://www.belspo.be/). CD held a fellowship from Télévie (Belgium; grant 7.4597.14-F; http://www.fnrs.be/index.php/mecenat-prix/l-operation-televie). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Intrahepatic cholangiocarcinoma (ICC) is the second most common primary tumor of the liver. It represents less than 10% of cholangiocarcinoma cases, but in contrast to perihilar and distal cholangiocarcinoma, its incidence and mortality rates are rising [1–3]. When evaluating disease progression, several prognostic factors were proposed. These include tumor size, surgical strategy, serum markers, TNM staging, histological features, mutational profiles, gene expression signatures (mRNA, microRNA, lncRNA), or expression of individual genes at the mRNA, miRNA or protein level. Prognostic factors can also be evaluated combinatorially, for instance in the format of nomograms, to improve their prognostic value .
Because of its ease of use in the clinic, differentiation of ICC is an often considered prognostic factor. Histopathologically, the proportion of gland formation serves as a differentiation criterium: well-differentiated tumors, which are associated with better survival, exhibit greater than 95% glandular tissue, moderately differentiated tumors show 50% to 95% gland formation, and poorly differentiated tumors have less than 50% glandular tissue [5–8]. Interestingly, a number of markers that qualify as molecular differentiation markers because they were shown to promote differentiation of cholangiocytes or to be expressed specifically in the biliary lineage correspond to ICC prognostic factors. One would therefore expect that expression of molecular differentiation markers would overlap with the prognostic value of ICC. Consistently, Mazur and coworkers found that the transcription factor SRY-related HMG box transcription factor 9 (SOX9), which is required for bile duct morphogenesis and cholangiocyte polarity, is a prognostic factor for survival . Moreover, Terashi et al. showed that lack of osteopontin (OPN), which is expressed in the cholangiocytes and not in hepatocytes , is associated with poor patient outcome . However, SOX9 was not confirmed as a predictor of better survival in a different study , and not all molecular differentiation markers have a prognostic value: HNF1β, which is cholangiocyte-specific in adult liver and which stimulates biliary development , is not associated with survival rate of ICC [7, 13]. Again, Wang and coworkers showed that the transcription factor SOX4, which promotes cholangiocyte differentiation and bile duct formation , is unexpectedly a factor for poor prognosis of ICC .
These observations shed light on a discrepancy between the prognostic value and the expression of molecular differentiation markers and prompted us to consider that the expression of molecular markers of cholangiocyte differentiation would have a stronger predicting value if supported by histopathological differentiation criteria. Here we analysed a cohort of surgically resected ICCs and verified if histopathological differentiation grade is correlated with expression of molecular differentiation markers. For the latter, our analysis included transcription factors that are well characterized as differentiation regulators. We also included a set of microRNAs (miRNAs) which have been shown to be differentially expressed between the different tumor grades . Indeed, miRNAs are involved in cholangiocarcinoma where they modulate apoptosis, proliferation, migration and response to therapy [16–19] and several investigators characterized miRNA expression profiles in ICC to identify miRNA signatures for improved diagnosis and prognosis [20–24].
Surprisingly, we found that molecular differentiation markers are expressed at similar levels regardless of the differentiation status of the tumor. Therefore, our data highlight a disparity between molecular- and histology-based classification and suggest caution when classifying ICC tumor samples.
Materials and Methods
The work on human tissue samples was performed in compliance with the Belgian regulation, and with the 1975 Declaration of Helsinki. We studied 18 ICCs obtained from 16 patients (12 men; 4 women) surgically treated at the Cliniques Universitaires St Luc (Brussels, Belgium) by a surgeon who is not co-author of the present paper, prior to the start of the present analysis. An anonymous registration number was tagged on each sample and researchers who analysed the samples had no access to the name of the donor. The retrospective analysis of the surgical specimens was approved by the Commission on Biomedical Ethics of the Université Catholique de Louvain and Cliniques Universitaires Saint-Luc (Approval # 2013/08JUL/384). Informed consent is not required for retrospective analyses.
Immunohistochemistry and Immunofluorescence
Immunohistochemistry for cytokeratin (CK) 7, CK19 and H&E stainings were performed on 4 μm sections of formalin-fixed paraffin-embedded tumors and carried out with a standardized protocol on a Ventana Benchmark system (Ventana Medical Systems, Tucson, AZ, USA). Immunofluorescence was performed on 5 μm sections of formalin-fixed paraffin-embedded tumors. Briefly, tissue sections were deparaffinized and micro-wave heated for 10 min in 10 mM sodium citrate pH 6.0 for antigen unmasking. Sections were permeabilized for 5 min in phosphate-buffered saline (PBS)/0.3% Triton X-100 before blocking for 45 min in 0.3% milk/10% bovine serum albumin/0.3% Triton X-100 in PBS. Primary and secondary antibodies (listed in S1 Table) were diluted in blocking solution and incubated respectively at 4°C overnight and room temperature for 1 h. Pictures were taken with an Axiovert 200 fluorescent microscope using AxioVision system.
RNA extraction and RT-qPCR
Formalin-fixed paraffin-embedded tumors were sectioned at 10 μm and regions of interest were scraped from the glass slides with a surgical blade. Total RNA was isolated using the RecoverAll Total Nucleic Acid Isolation Kit for FFPE (#AM1975, Ambion). The procedure included a crosslink reversal step during RNA isolation and the use of random hexamers and short amplicon size for the RT-qPCR step . cDNA synthesis was performed with MMLV reverse transcriptase (#28025–13, Invitrogen) according to manufacturer’s protocol. Gene expression was quantified by qPCR using Kapa SYBR Fast 2X Universal Master Mix (#KK4601, Sopachem). The primers are listed in S2 Table. For measuring microRNA expression, specific stem-loop primers were used for reverse transcription and RT-qPCR was performed using a specific forward primer and a common universal reverse primer (S3 Table). mRNA and microRNA levels were normalized for β-ACTIN and for means of RNU6 and RNU1a respectively. mRNA and miRNA could be compared between tumor and non-tumor samples: mean Ct values for β-ACTIN were 20.21 and 21.70 in tumor and non-tumor, respectively; mean Ct values for means of U6 and U1A were 15.19 and 16.37 in tumor and non-tumor tissue.
Selection of tumor samples
We histologically classified 18 ICC samples from 16 patients (12 men and 4 women; age 39–75, mean: 60.2) as well (containing 95% glandular tissue), moderately (50% to 95% gland formation) or poorly differentiated (less than 50% glandular tissue) (Fig 1A–1C). None of the patients were receiving treatment, except for one of the two patients with a recurrence and who was treated with Gemcitabine. In one patient, the ICC contained both a well and a poorly differentiated area that were analysed separately. Two patients experienced a recurrence and each second tumor was also studied separately, leading to a total number of 19 samples of ICC characterized for their histological differentiation. We selected tumors areas in which CK7-positive cells represented more than 50% of the tumor surface (Fig 1D–1F), in order to reduce the contamination of epithelial cancer cells by non-epithelial cells during the RNA extraction procedure (see below). Tumor characteristics are described in Table 1.
(A-C) Tumors were classified as well-, moderately- and poorly-differentiated. (D-F) Immunohistochemical detection of CK7 in representative histologically-graded ICCs. Areas with at least 50% of CK7 staining were selected for further RNA extraction and immunofluorescence stainings. White scale bars = 200 μm. Black scale bars = 500 μm
Molecular differentiation markers in ICC: mRNA and microRNA expression
To look for a potential correlation between histopathological differentiation grade and molecular differentiation markers in ICC, we measured the expression of the latter in the tumor samples. We focused on transcription factors that were shown to be stimulators of biliary differentiation. Total RNA was extracted from formalin-fixed paraffin-embedded tissue: ICC tissue was delineated using CK7 and H&E staining in the adjacent sections, and non-tumor CK7-negative tissue from a distant sample of the same surgical specimen was used as control. We then measured by qRT-PCR the expression of biliary differentiation markers (SOX4, SOX9, and HNF1β) and of the hepatocyte marker HNF4α. All biliary markers were expressed at lower levels in the tumor tissue than in the adjacent non-tumor tissue (Fig 2A). HNF4α was low, confirming that the tumor samples were not contaminated with significant levels of hepatocytes.
(A) All biliary differentiation markers are expressed (qRT-PCR) at lower levels in ICC as compared to distant non-tumoral tissue. Low levels of HNF4α indicate lack of significant contamination with normal hepatocytes. (Mean ± SEM; *p<0.05, **p<0.01, ***p<0.001). (B-C) Lack of correlation between molecular differentiation marker expression and tumor grade when comparing (B) ICC samples or (C) the ratio of expression in tumor versus non-tumor tissue. Dots represent individual ICC samples and horizontal bars represent the mean of the samples.
Importantly, our analysis showed that none of the tested markers correlated with histological differentiation grade (Fig 2B). In addition, when the ratio of expression in tumor versus non-tumor tissue was analysed, again no correlation was found between molecular marker expression and histological grade of the ICCs (Fig 2C).
Since the transcription factors tested above did not correlate with ICC histological grade, we measured the expression of other markers (OPN, MUC1, CK903, YAP, EPCAM, NCAD and CK19) that are known to be biliary-specific and associated with ICC [26–32]. All those markers were expressed at lower levels in the tumor tissue than in the adjacent non-tumor tissue (Fig 3A). Moreover, as observed previously, we did not found any correlation between the biliary markers and the histological grade of the tumors (Fig 3B). Again, when considering the ratio of expression in tumor versus non-tumor tissue, no correlation was found (Fig 3C).
(A) All markers are expressed (qRT-PCR) at lower levels in ICC as compared to distant non-tumoral tissue. (Mean ± SEM; *p<0.05, **p<0.01, ***p<0.001). (B-C) Lack of correlation between molecular differentiation marker expression and tumor grade when comparing (B) ICC samples or (C) the ratio of expression in tumor versus non-tumor tissue. Dots represent individual ICC samples and horizontal bars represent mean of the samples.
We next investigated the expression of miRNAs in our ICC samples. Several studies have compared the expression of miRNAs in ICC and control tissue [20–24]. We here analysed miRNAs that were found to be dysregulated in earlier studies in which control tissue was compared with ICC, and whose dysregulation has been correlated with histological differentiation grade of ICCs (Table 2) .
Fig 4A shows that a subset of microRNAs (miR-200c, miR-31 and miR-135b) was expressed at higher levels in the tumor than in the non-tumor tissue. miR-222 was not detectable in our ICC samples. When considering differentiation of the tumors, no correlation was found between histological grade and expression of the selected miRNAs, except for miR-132, which was low in the poorly differentiated tumors (Fig 4B). However, when considering the ratio of expression in tumor versus non-tumor tissue, again no correlation was found with histological grade, even for miR-132 (Fig 4C).
MicroRNAs known to be dysregulated in ICC and differentially expressed among tumor grades were selected for qRT-PCR analysis. (A) Three microRNAs are up-regulated in ICC when compared to distant non-tumoral tissues. (Mean ± SEM; *p<0.05). (B-C) Lack of correlation between microRNA expression and tumor differentiation grade when (B) comparing ICC samples, or (C) when analysing the ratio of expression in tumor versus non-tumor tissue. Dots represent individual ICC samples.
Molecular differentiation markers in ICC—protein expression
Since mRNA expression does not necessarily correlate with protein expression, we analysed the expression of SOX9, OPN, CK19 and HNF1β by immunostaining on two representative samples of each histologically-defined differentiation grade (Fig 5). The results showed that the markers are expressed in the tumors regardless of their differentiation grade. SOX9, which is normally located in the nucleus, was restricted to the cytoplasm in well-differentiated ICC while it was expressed in both the cytoplasm and nucleus in moderately and poorly differentiated ICC (Fig 5A–5F). CK19 was expressed in all ICC samples (Fig 5G–5L). Similarly, OPN was expressed in all tumor types (Fig 5M–5R). However, in well-differentiated ICC, OPN was predominantly expressed at the apical pole of cells but a subset of cells also displayed cytoplasmic location (Fig 5M and 5N, white and yellow arrowheads). In contrast, OPN was observed mainly in the cytoplasm of cells of moderately differentiated tumors, while only a subset of cells expressed OPN at the apical pole (Fig 5O and 5P, white and yellow arrowheads). Poorly differentiated tumors have lost apico-basal polarity, preventing correct interpretation of polarity marker location. However, in those poorly differentiated tumors some cells expressed cytoplasmic OPN (Fig 5Q and 5R, yellow arrowheads). Therefore, overall expression of CK19, SOX9 and OPN protein does not correlate with histological differentiation grade, but subcellular location of SOX9 and OPN were related to ICC cell differentiation.
(A-F) SOX9 expression is detected in the cytoplasm of well-differentiated ICC cells (A-B) and both in the cytoplasm and nucleus of moderately and poorly differentiated tumor cells (C-F). (G-L) CK19 is expressed in all types of histologically-graded tumors. (M-R) OPN is detected in all types of histologically graded-ICC's. Its expression is mostly apical in well-differentiated tumor cells (M-N, white arrowheads), but cytoplasmic OPN is detected in moderately and poorly differentiated ICCs and in a subset of well-differentiated tumor cells (M-R, yellow arrowheads). (M’-R’) HNF1β expression inversely correlates with ICC differentiation grade. White scale bar = 100 μm, black scale bar = 500 μm.
Surprisingly, in contrast to the qRT-PCR data, HNF1β protein expression was high in poorly differentiated ICC (Fig 5Q–5R and 5Q'–5R') and, low and mislocated, i.e. non-nuclear, in well differentiated tumors (Fig 5M–5N and 5M'–5N'); moderately differentiated tumors showed intermediate expression levels (Fig 5O–5P and 5O'–5P').
In recent years, many efforts have been devoted to better classify ICCs and determine their prognosis. Since differentiation grade of the tumors, defined by histopathological criteria, is closely related to patient outcome [6, 7], in the present work we wished to strengthen histopathological differentiation criteria by correlating them with expression of molecular cholangiocyte differentiation markers. We assumed that well differentiated tumors would express high levels of cholangiocyte markers and that these levels would decrease in less differentiated tumors. However, our work unexpectedly revealed a lack of correlation between histopathological and molecular differentiation criteria in ICC.
Here, we analysed the expression of biliary markers in areas of ICCs where the tumoral cells (highlighted by the CK7 positive staining) represented more than 50% of the total area in order to limit contaminations by non-tumor cells. All our carcinoma express low levels of these markers irrespective of the tumor differentiation grade (Figs 2 to 4), thereby suggesting that the epithelial cells are undergoing a dedifferentiation process, at least at the molecular level. In line with this observation, Mazur and coworkers previously proposed that SOX9 expression decreases in the early step of ICC development . These results raise the question of the real differentiation status of the tumors, and may explain why some differentiation markers are associated with good or poor prognosis in distinct studies (e.g. SOX9; [7, 11]), and why some other differentiation markers such as HNF1β are not associated with prognosis [7, 13].
We further showed by RT-qPCR and immunofluorescence that SOX9 and OPN are expressed equally in all tumor types (Fig 5). However, their subcellular location depended on the differentiation grade of the ICC. SOX9 is located in the cytoplasm of well-differentiated tumors investigated here, whereas it was nuclear or cytoplasmic in moderately and poorly differentiated tumors. Cytoplasmic expression of SOX9 has already been observed in pancreatic adenocarcinoma affected with p53 mutation , in breast cancer cells  and in biliary tract cancers , suggesting that the location of SOX9 may depend on the mutational status of the cells. OPN is expressed in the cytoplasm of early biliary progenitors in the embryo, and becomes apical once the biliary cells become polarized . The dual cytoplasmic and apical location of OPN found in well-differentiated tumors may reflect distinct levels of molecular differentiation of the cells within the same tumor. Regarding HNF1β, Mazur and coworkers, could not correlate HNF1β expression with survival rates of ICC . This is at first sight at odds with our findings: since we found more HNF1β protein in poorly differentiated tumors, which are known to be associated with poor prognosis, our data would suggest that HNF1β is a marker of poor prognosis. However, combining our data with those of Mazur and coworkers would be in agreement with the overall conclusion of the present work, namely that molecular differentiation markers do not correlate with histopathological differentiation criteria. Finally, HNF1β protein expression is higher in poorly differentiated tumors, while HNF1β mRNA does not correlate differentiation grade. This points toward potential post-transcriptional regulation and further underscores the care needed when interpreting expression of molecular differentiation markers in ICC.
When analysing our set of tumors, we were unable to corroborate the findings by Plieskatt and co-workers, who showed that miR-200c, miR221, miR-31, miR-135b and miR-132 are differentially expressed between well differentiated, moderately differentiated and papillary carcinoma-type ICC [8, 33]. In line with this, several groups attempted to list dysregulated microRNAs in ICCs, but only little overlap can be found when comparing the proposed microRNA lists [8, 21–24, 33]. In addition, contradictory results were found. For instance, miR-200c was either down-regulated [21, 23] or upregulated [8, 33] (Fig 3). This points towards necessary care when interpreting microRNA expression in ICC. The broad range of control tissue used to investigate microRNA expression could contribute to the observed discrepancies. Indeed, control tissues selected by different authors were either normal hepatic bile duct tissue or total liver parenchyma, which are distinct tissues, and originated from liver associated with a variety of diseases such as cirrhosis, hepatitis, or infection with Opisthorchis viverrini [8, 33, 36]. In that context, Plieskatt and coworkers showed that in Opisthorchis viverrini-induced ICC, miRNAs in distal non-tumor tissue clustered closer to the tumor than to non-tumor tissue from control individuals . This is in line with our lack of correlation between miRNA or mRNA expression and ICC grade when considering the ratio of expression in tumor versus non-tumor tissue (Fig 3C). Indeed, our controls were from non-tumor tissue taken from the patient affected with ICC.
In this work we compared histologically-graded ICC rich in epithelial cells and found no correlation between expression of biliary molecular differentiation markers and the histological differentiation grade of the tumors. Therefore, our data point towards necessary caution when classifying ICC tumor samples and determining prognosis based on differentiation criteria.
S1 Table. List of primary and secondary antibodies used for immunostaining experiments.
S2 Table. List of primer sequences used for RT-qPCR.
The authors thank the members of the Lemaigre lab for discussions, and the Department of Pathology of the Cliniques Universitaires Saint Luc for technical support.
Conceived and designed the experiments: CD FPL CS. Performed the experiments: CD CS. Analyzed the data: CD FPL CS. Contributed reagents/materials/analysis tools: CD FPL CS CH. Wrote the paper: CD FPL CS CH.
- 1. Khan SA, Emadossadaty S, Ladep NG, Thomas HC, Elliott P, Taylor-Robinson SD, et al. Rising trends in cholangiocarcinoma: is the ICD classification system misleading us? J Hepatol. 2012;56(4):848–54. pmid:22173164.
- 2. Wadsworth CA, Lim A, Taylor-Robinson SD, Khan SA. The risk factors and diagnosis of cholangiocarcinoma. Hepatology international. 2013;7(2):377–93. Epub 2013/06/01. pmid:26201772.
- 3. Razumilava N, Gores GJ. Cholangiocarcinoma. Lancet (London, England). 2014;383(9935):2168–79. Epub 2014/03/04. pmid:24581682; PubMed Central PMCID: PMCPMC4069226.
- 4. Hyder O, Marques H, Pulitano C, Marsh JW, Alexandrescu S, Bauer TW, et al. A nomogram to predict long-term survival after resection for intrahepatic cholangiocarcinoma: an Eastern and Western experience. JAMA Surg. 2014;149(5):432–8. pmid:24599477.
- 5. Fan L, He F, Liu H, Zhu J, Liu Y, Yin Z, et al. CD133: a potential indicator for differentiation and prognosis of human cholangiocarcinoma. BMC Cancer. 2011;11:320. Epub 2011/07/30. pmid:21798073; PubMed Central PMCID: PMCPMC3161038.
- 6. Uenishi T, Kubo S, Yamamoto T, Shuto T, Ogawa M, Tanaka H, et al. Cytokeratin 19 expression in hepatocellular carcinoma predicts early postoperative recurrence. Cancer science. 2003;94(10):851–7. Epub 2003/10/15. pmid:14556657.
- 7. Mazur PK, Riener MO, Jochum W, Kristiansen G, Weber A, Schmid RM, et al. Expression and clinicopathological significance of notch signaling and cell-fate genes in biliary tract cancer. The American journal of gastroenterology. 2012;107(1):126–35. Epub 2011/09/21. pmid:21931375.
- 8. Plieskatt JL, Rinaldi G, Feng Y, Peng J, Yonglitthipagon P, Easley S, et al. Distinct miRNA signatures associate with subtypes of cholangiocarcinoma from infection with the tumourigenic liver fluke Opisthorchis viverrini. J Hepatol. 2014;61(4):850–8. pmid:25017828.
- 9. Antoniou A, Raynaud P, Cordi S, Zong Y, Tronche F, Stanger B, et al. Intrahepatic bile ducts develop according to a new mode of tubulogenesis regulated by the transcription factor SOX9. Gastroenterology. 2009;136(7):2325–33. Epub 2009/02/21. doi: S0016-5085(09)00300-X [pii]. pmid:19403103.
- 10. Terashi T, Aishima S, Taguchi K, Asayama Y, Sugimachi K, Matsuura S, et al. Decreased expression of osteopontin is related to tumor aggressiveness and clinical outcome of intrahepatic cholangiocarcinoma. Liver Int. 2004;24(1):38–45. Epub 2004/04/23. pmid:15101999.
- 11. Matsushima H, Kuroki T, Kitasato A, Adachi T, Tanaka T, Hirabaru M, et al. Sox9 expression in carcinogenesis and its clinical significance in intrahepatic cholangiocarcinoma. Dig Liver Dis. 2015. pmid:26341967.
- 12. Coffinier C, Gresh L, Fiette L, Tronche F, Schutz G, Babinet C, et al. Bile system morphogenesis defects and liver dysfunction upon targeted deletion of HNF1beta. Development. 2002;129(8):1829–38. Epub 2002/04/06. pmid:11934849.
- 13. Yu DD, Jing YY, Guo SW, Ye F, Lu W, Li Q, et al. Overexpression Of Hepatocyte Nuclear Factor-1beta Predicting Poor Prognosis Is Associated With Biliary Phenotype In Patients With Hepatocellular Carcinoma. Sci Rep. 2015;5:13319. pmid:26311117; PubMed Central PMCID: PMCPMC4550878.
- 14. Poncy A, Antoniou A, Cordi S, Pierreux CE, Jacquemin P, Lemaigre FP. Transcription factors SOX4 and SOX9 cooperatively control development of bile ducts. Dev Biol. 2015;402(2):136–48. doi: https://doi.org/10.1016/j.ydbio.2015.05.012 PMID: 26033091.
- 15. Wang W, Zhang J, Zhan X, Lin T, Yang M, Hu J, et al. SOX4 is associated with poor prognosis in cholangiocarcinoma. Biochem Biophys Res Commun. 2014;452(3):614–21. Epub 2014/09/03. pmid:25181339.
- 16. Wang XW, Heegaard NH, Orum H. MicroRNAs in liver disease. Gastroenterology. 2012;142(7):1431–43. pmid:22504185.
- 17. Szabo G, Bala S. MicroRNAs in liver disease. Nat Rev Gastroenterol Hepatol. 2013;10(9):542–52. pmid:23689081; PubMed Central PMCID: PMC4091636.
- 18. Papaconstantinou I, Karakatsanis A, Gazouli M, Polymeneas G, Voros D. The role of microRNAs in liver cancer. Eur J Gastroenterol Hepatol. 2012;24(3):223–8. pmid:22228372.
- 19. Munoz-Garrido P, Garcia-Fernandez de Barrena M, Hijona E, Carracedo M, Marin JJ, Bujanda L, et al. MicroRNAs in biliary diseases. World J Gastroenterol. 2012;18(43):6189–96. pmid:23180938; PubMed Central PMCID: PMCPMC3501766.
- 20. Meng F, Henson R, Lang M, Wehbe H, Maheshwari S, Mendell JT, et al. Involvement of human micro-RNA in growth and response to chemotherapy in human cholangiocarcinoma cell lines. Gastroenterology. 2006;130(7):2113–29. Epub 2006/06/10. doi: S0016-5085(06)00736-0 [pii] pmid:16762633.
- 21. Chen L, Yan HX, Yang W, Hu L, Yu LX, Liu Q, et al. The role of microRNA expression pattern in human intrahepatic cholangiocarcinoma. J Hepatol. 2009;50(2):358–69. pmid:19070389.
- 22. Kawahigashi Y, Mishima T, Mizuguchi Y, Arima Y, Yokomuro S, Kanda T, et al. MicroRNA profiling of human intrahepatic cholangiocarcinoma cell lines reveals biliary epithelial cell-specific microRNAs. J Nippon Med Sch. 2009;76(4):188–97. pmid:19755794.
- 23. Karakatsanis A, Papaconstantinou I, Gazouli M, Lyberopoulou A, Polymeneas G, Voros D. Expression of microRNAs, miR-21, miR-31, miR-122, miR-145, miR-146a, miR-200c, miR-221, miR-222, and miR-223 in patients with hepatocellular carcinoma or intrahepatic cholangiocarcinoma and its prognostic significance. Mol Carcinog. 2013;52(4):297–303. pmid:22213236.
- 24. Zhang MY, Li SH, Huang GL, Lin GH, Shuang ZY, Lao XM, et al. Identification of a novel microRNA signature associated with intrahepatic cholangiocarcinoma (ICC) patient prognosis. BMC Cancer. 2015;15:64. pmid:25880914; PubMed Central PMCID: PMCPMC4344737.
- 25. von Ahlfen S, Missel A, Bendrat K, Schlumpberger M. Determinants of RNA quality from FFPE samples. PLoS One. 2007;2(12):e1261. pmid:18060057; PubMed Central PMCID: PMCPMC2092395.
- 26. Mosnier JF, Kandel C, Cazals-Hatem D, Bou-Hanna C, Gournay J, Jarry A, et al. N-cadherin serves as diagnostic biomarker in intrahepatic and perihilar cholangiocarcinomas. Mod Pathol. 2009;22(2):182–90. pmid:18622386.
- 27. Aishima S, Asayama Y, Taguchi K, Sugimachi K, Shirabe K, Shimada M, et al. The utility of keratin 903 as a new prognostic marker in mass-forming-type intrahepatic cholangiocarcinoma. Mod Pathol. 2002;15(11):1181–90. pmid:12429797.
- 28. Sulpice L, Rayar M, Turlin B, Boucher E, Bellaud P, Desille M, et al. Epithelial cell adhesion molecule is a prognosis marker for intrahepatic cholangiocarcinoma. J Surg Res. 2014;192(1):117–23. pmid:24909871.
- 29. Braicu C. Molecular Markers in the Pathogenesis of Cholangiocarcinoma: Potential for Early Detection and Selection of Appropriate Treatment. Gastroenterology Research. 2009.
- 30. Sirica AE, Dumur CI, Campbell DJ, Almenara JA, Ogunwobi OO, Dewitt JL. Intrahepatic cholangiocarcinoma progression: prognostic factors and basic mechanisms. Clin Gastroenterol Hepatol. 2009;7(11 Suppl):S68–78. pmid:19896103; PubMed Central PMCID: PMCPMC3795391.
- 31. Marti P, Stein C, Blumer T, Abraham Y, Dill MT, Pikiolek M, et al. YAP promotes proliferation, chemoresistance, and angiogenesis in human cholangiocarcinoma through TEAD transcription factors. Hepatology. 2015;62(5):1497–510. pmid:26173433.
- 32. Pei T, Li Y, Wang J, Wang H, Liang Y, Shi H, et al. YAP is a critical oncogene in human cholangiocarcinoma. Oncotarget. 2015;6(19):17206–20. Epub 2015/05/28. pmid:26015398; PubMed Central PMCID: PMCPMC4627302.
- 33. Plieskatt J, Rinaldi G, Feng Y, Peng J, Easley S, Jia X, et al. A microRNA profile associated with Opisthorchis viverrini-induced cholangiocarcinoma in tissue and plasma. BMC Cancer. 2015;15:309. pmid:25903557; PubMed Central PMCID: PMCPMC4417245.
- 34. Huang L, Holtzinger A, Jagan I, BeGora M, Lohse I, Ngai N, et al. Ductal pancreatic cancer modeling and drug screening using human pluripotent stem cell- and patient-derived tumor organoids. Nat Med. 2015;21(11):1364–71. pmid:26501191.
- 35. Chakravarty G, Rider B, Mondal D. Cytoplasmic compartmentalization of SOX9 abrogates the growth arrest response of breast cancer cells that can be rescued by trichostatin A treatment. Cancer Biol Ther. 2011;11(1):71–83. pmid:21084857.
- 36. Razumilava N, Gores GJ. Classification, diagnosis, and management of cholangiocarcinoma. Clin Gastroenterol Hepatol. 2013;11(1):13–21 e1; quiz e3-4. pmid:22982100; PubMed Central PMCID: PMCPMC3596004.