Hepatocyte nuclear factors (HNF) play a critical role in development of the liver. Their roles during liver tumorigenesis and progression of hepatocellular carcinomas (HCC) are, however, poorly understood. To address the role of HNFs in tumor progression, we generated a new experimental model in which a highly differentiated slow-growing transplantable mouse HCC (sgHCC) rapidly gives rise in vivo to a highly invasive fast-growing dedifferentiated variant (fgHCC). This in vivo model has allowed us to investigate the fundamental mechanisms underlying HCC progression. A complete loss of cell polarity, a decrease in cell-cell and cell-extracellular matrix (ECM) adhesion, elevation of telomerase activity, and extinction of liver-specific gene expression accompanies tumor progression. Moreover, cells isolated from fgHCCs acquired the ability to proliferate rapidly in culture. These alterations were coupled with a reduced expression of several liver transcription factors including HNF4, a factor essential for hepatocyte differentiation. Forced re-expression of HNF4␣1 in cultured fgHCC cells reversed the progressive phenotype and induced fgHCC cells to reestablish an epithelium and reform cell-ECM contacts. Moreover, fgHCC cells that expressed HNF4␣1 also re-established expression of the profile of liver transcription factors and hepatic genes that are associated with a differentiated hepatocyte phenotype. Importantly, re-expression of HNF4␣1 in fgHCC reduced the proliferation rate in vitro and diminished tumor formation in congenic recipient mice. In conclusion, loss of HNF4 expression is an important determinant of HCC progression. Forced expression of this factor can promote reversion of tumors toward a less invasive highly differentiated slow-growing phenotype. (HEPATOLOGY 2004;39:1038 -1047 H epatocellular carcinoma (HCC) is one of the world's most common cancers, with chronic hepatitis B and C infection and prolonged exposure to hepatocarcinogens being the major risk factors. 1 The development of a malignant phenotype is considered a multi-step process that results from the accumulation of genetic alterations. The consequence of these alterations is the evolution of a tumor towards a more aggressive phenotype. HCC progression from a well differentiated to a less differentiated form is accompanied by striking changes in the morphological and genetic properties of the cells and is one of the most critical steps of liver tumorigenesis. 2,3 Hepatic tumor progression is defined by a decrease in differentiation, an extinction of tissue-specific gene expression, acceleration of cell proliferation, loss of epithelial morphology, increased invasiveness, and ultimately metastasis. While a variety of inter-and intracellular signaling pathways essential for the control of liver function and proliferation have been described, the mo-

Dedifferentiation and epithelial-mesenchymal transition are important steps in epithelial tumor progression. A central role in the control of functional and morphological properties of different cell types is attributed to tissue-specific transcription factors which form regulatory cascades that define specification and differentiation of epithelial cells during embryonic development. The main principles of the action of such regulatory systems are reviewed on an example of a network of hepatocyte nuclear factors (HNFs) which play a key role in establishment and maintenance of hepatocytes--the major functional type of liver cells. HNFs, described as proteins binding to promoters of most hepatospecific genes, not only control expression of functional liver genes, but are also involved in regulation of proliferation, morphogenesis, and detoxification processes. One of the central components of the hepatospecific regulatory network is nuclear receptor HNF4alpha. Derangement of the expression of this gene is associated with progression of rodent and human hepatocellular carcinomas (HCCs) and contributes to increase of proliferation, loss of epithelial morphology, and dedifferentiation. Dysfunction of HNF4alpha during HCC progression can be either caused by structural changes of this gene or occurs due to modification of up-stream regulatory signaling pathways. Investigations preformed on a model system of the mouse one-step HCC progression have shown that the restoration of HNF4alpha function in dedifferentiated cells causes partial reversion of malignant phenotype both in vitro and in vivo. Derangement of HNFs function was also described in other tumors of epithelial origin. We suppose that tissue-specific factors that underlie the key steps in differentiation programs of certain tissues and are able to receive or modulate signals from the cell environment might be considered as promising candidates for the role of tumor suppressors in the tissue types where they normally play the most significant role.

BackgroundRNA-seq is a useful tool for analysis of gene expression. However, its robustness is greatly affected by a number of artifacts. One of them is the presence of duplicated reads.ResultsTo infer the influence of different methods of removal of duplicated reads on estimation of gene expression in cancer genomics, we analyzed paired samples of hepatocellular carcinoma (HCC) and non-tumor liver tissue. Four protocols of data analysis were applied to each sample: processing without deduplication, deduplication using a method implemented in SAMtools, and deduplication based on one or two molecular indices (MI). We also analyzed the influence of sequencing layout (single read or paired end) and read length. We found that deduplication without MI greatly affects estimated expression values; this effect is the most pronounced for highly expressed genes.ConclusionThe use of unique molecular identifiers greatly improves accuracy of RNA-seq analysis, especially for highly expressed genes. We developed a set of scripts that enable handling of MI and their incorporation into RNA-seq analysis pipelines. Deduplication without MI affects results of differential gene expression analysis, producing a high proportion of false negative results. The absence of duplicate read removal is biased towards false positives. In those cases where using MI is not possible, we recommend using paired-end sequencing layout.