HCC is a hypervascular tumor, and expression of the hypoxia inducible factor (HIF) and some of its target genes have been reported to be associated with a poor prognostic phenotype. To explore the importance of the hypoxic response in HCC, we are using a comparative genomic approach to characterize the hypoxic gene expression profile in freshly isolated mouse hepatocytes. We found more than 1800 significantly regulated genes (P less than 0.001) in primary cultures of mouse hepatocytes maintained under hypoxic conditions over 24 hr. In the first 12 hr, the most important response to hypoxic conditions was up-regulation of genes involved in angiogenesis, blood vessel formation and blood coagulation, while the transition to an anaerobic metabolism was more prominent after 24 hr. Genes that showed at least 2-fold expression differences between hypoxic and normoxic conditions were selected to define the hypoxia gene expression signature. 504 orthologous genes derived from the hypoxic signature were used to perform hierarchical cluster analysis of 139 human HCC. Two subsets of genes were identified. The first subset of 104 genes was implicated in cell cycle and apoptosis regulation (Gadd45a, Cdk4, Map4k4, Dusp1, Csk2), blood coagulation (Plg, F2, F9, F13b, Hc, Agt, Serpinc1, Serpinf1) and immune response (C1r, C8a, C8b, C8g, C9, Rarres2) among other functions, and was able to predict HCC with a good prognosis. The second subset of 62 genes also associated with cell cycle progression (Cdk6, Ches1, Ccng1, Ngfb), apoptosis regulation (Bcl6b, Pik3r1) and angiogenesis (Vegf, Id3), predicted HCC with poor prognosis. HIF-1, a key transcription factor induced by hypoxia, is a heterodimer consisting of the hypoxic regulated factor HIF-1a and the constitutively expressed aryl hydrocarbon receptor nuclear translocator (known as HIF-1a ). To study which of the two identified subsets of genes is regulated specifically by HIF-1, we generated liver-specific HIF-1a knockout mice and plan to use freshly isolated HIF-1a-/- hepatocytes to identify more precisely the HIF-1a dependent and independent sets of genes regulated by normoxic and hypoxic conditions. To address the importance of HIF-1a in liver carcinogenesis, we are using two HIF-1a conditional knockout mouse models in which deletion of the exons 13 to 15 of HIF-1a, essential for the responsiveness to hypoxia, is achieved by either Alb-cre or Mx1-cre transgenes. The use of these mice in combination with the well characterized model of DEN-induced hepatocarcinogenesis,will allow us to address the contribution of HIF-1a to the different stages of tumor development. We have previously shown that activation of AP-1 transcription factors is a characteristic feature of HCC subtype expressing hepatoblast traits. We have now started to evaluate the relevance of AP-1 (Jun/Fos) gene expression signature for human HCC. We performed performed gene expression profiling of liver samples(normal and tumors) derived from WT (Junfl/fl) and Jun-KO mice. In total, 364 genes were found to be differentially expressed (P less than 0.01) in tumors derived from Jun-KO vs. WT animals, with 60% being down-regulated. The up-regulated genes were strongly associated with apoptosis. This subset of genes was referred to as """"""""Jun signature in HCC"""""""". Applying a comparative functional genomics approach, we found orthologous genes of the Jun signature in human, and then integrated the gene profiles from mouse tumor samples and human HCC (n=139). The Jun signature successfully discriminated human HCC into clusters displaying either WT or Jun-KO signatures. Notably, integration of the Jun signature with the human data revealed a clear and significant association of Jun-KO and Jun-WT signatures, respectively, with either Hepatocyte (HC) or Hepatoblast (HB) signatures which have been described previously. The Jun signature was significantly associated with patient survival. The gene expression signature in tumors derived from Jun-KO livers was close to the group of human HCC with a better survival. These tumors exhibited a greater apoptotic index. In the future, we will validate the Jun signature by analysis of enrichment of specific binding sites in the promoter regions of this subset of genes. A similar approach is currently being used in JNK conditional knockout mice.
