Tumor-specific marker genes for intrahepatic cholangiocarcinoma: utility and mechanistic insight

Cholangiocarcinoma (CCA) is the second most common primary hepatic malignancy. It is classified into intrahepatic and extrahepatic forms, the latter including perihilar (involvement of right and/or left hepatic ducts and their union to form the common hepatic duct) and distal bile duct CCA. Growth patterns are differentiated into periductal-infiltrating, mass-forming and papillary or intraductal growth. The different forms of CCA are distinct and differ clinically, etiologically, pathophysiologically and in management [1]. Global incidence rates of intrahepatic cholangiocarcinoma (ICC) have significantly increased while annual incidence rates of the more common extrahepatic form has remained relatively stable throughout the last four decades [2]. The prognosis of CCA is devastating with survival of <24 months following diagnosis. The only potentially curative treatment options currently are surgical. Unfortunately, the majority of patients are diagnosed at an advanced, unresectable stage due to the initial silent clinical character of this malignancy.

Despite advances in radiologic and laboratory diagnostic tests, the diagnosis of ICC remains highly challenging. It is an adenocarcinoma and as such often mimics metastasis to the liver. Many clinicians and pathologists would, therefore, prefer to have an objective, definitive approach for the diagnosis of this neoplasm. Immunohistochemical markers such as CK7 and CK19 are overexpressed in 100% of cholangiocarcinomas [3-5]. However, they are not specific for this cancer and can be expressed by hepatocellular carcinoma (HCC) or metastatic adenocarcinoma [6]. Therefore, there is a continued need for new diagnostic markers with better specificity.

Nishino et al. report in this issue of the journal the identification of three genes differently expressed in ICC compared to normal hepatic tissue, HCC and chronic liver disease: claudin-4 (CLDN-4), insulin-like growth factor-binding protein 5 (IGFBP-5) and biglycan (BGN). The utility of using all three markers for the diagnosis of ICC is impressive with a receiver operating curve displaying an area under the curve of close to one (near perfection). However, the interpretation of their results is limited due the number and choice of tissues. Initial SAGE-data are obtained from only one ICC sample and the overall number of ICC samples was a mere 16. The most important control, metastatic adenocarcinoma to the liver, is missing from these studies. Thus, whether these gene products can be used to positively diagnose ICC and exclude metastasis is not addressed by this study, a critical limitation.

In this study, non-malignant hepatic tissue from patients with hepatic colorectal cancer metastases served as normal liver controls and non-malignant hepatic tissue samples from HCC patients served as chronic liver disease control. The generalization of chronic liver disease is problematic as the molecular signature likely differs depending upon the etiology of the chronic liver disease. ICC is a highly desmoplastic, paucicellular tumor, and microdissection would be preferable in studies which evaluate gene expression in these tumors. The use of immunohistochemistry to confirm upregulation of the gene products in the cancer cells assuages this latter concern. Nevertheless, this battery of markers can likely be employed to differentiate HCC, and/or mixed tumors of the liver. From a scientific perspective, the results are not surprising. CLDN-4 belongs to the claudin family which are components of tight junctions [7]. CLDN-4 has been reported to be upregulated in a variety of different epithelial malignancies including cholangiocarcinoma and has been linked to invasiveness in certain cancers [7-11]. Lodi et al. showed that CLDN-4 is not expressed in normal hepatocytes and HCC, weakly in cholangiocytes and highly expressed in cholangiocarcinoma [8]. Hence, the data by Nishino et al. presented in this issue of the journal confirm the results of previous studies. However, Lodi et al. also point out that CLDN-4 expression does not differentiate between biliary tract cancers and metastatic adenocarcinomas. BGN is matrix proteoglycan important for organization of collagenous tissue and modulation of cell adhesion. Its overexpression is characteristic for mesenchymal cells [12] and likely reflects the highly desmoplastic character of cholangiocarcinomas.

IGFBP-5 is one of the six members of the IGFBP-family. IGFBP are important members of the insulin growth factor (IGF)-axis which can bind IGF-I and II, and thereby inhibit the interaction of these growth factor with their receptors. Proteolytic cleavage results in release of IGF-I/II followed by receptor binding and activation. IGFBP-5 can bind to extracellular matrix proteins which results in a decrease of its IGF binding affinity. Similarly, IGFBP-5 as been reported to be upregulated in a variety of different malignancies including breast-, prostate-, ovarian cancer and other malignancies. Nishino et al’s observation of IGFBP-5 in ICC is interesting from a biologic perspective. Recently, it was reported that ICC express the insulin growth factor-I receptor and elevated biliary IGF-I concentration is suggested as a novel marker for identification of cholangiocarcinoma in patients with obstructive cholestasis [13, 14]. IGFBP-5 has been reported to function by different IGF-independent and –dependent mechanisms resulting for example in enhancement of local growth factors stimulation [15]. Nishino et al.’s results, therefore, provide further mechanistic insight into the biology of IGF-I and its regulators in ICC.

In summary, Nishino et al. have helped identify markers which distinguish ICC from HCC and may provide mechanistic information regarding the biology of ICC. A comparison of the findings with metastatic adenocarcinomas to the liver is needed before this panel of markers can be used clinically. Biologically, they form the basis and justification for further functional studies of IGF-I and IGFBP-5 in ICC. More work is encouraged on the biology of ICC; it is well within the domain of hepatology, and as a profession we need to address this cancer [16, 17].