Hepatitis virus B protein x (HBx) promotes hepatocellular carcinoma through hypoxia inducible factor (HIF)/lysyl oxidase like 2 (LOXL2) pathway

Abstract

Hepatocellular Carcinoma (HCC) being the third most lethal cancer in the world in 2008 is proposed to be highly associated with hepatitis B virus (HBV) infection. The distribution of high incident rate region of HCC and HBV were found to be very similar. It was also found that about 25% of chronic HBV patients would develop HCC. This strong association between HCC and HBV was believed to be related to HBx protein. HBx protein is encoded by HBV genome and is expressed in HBV infected hepatocytes. This multifunctional protein is able to interact with various cellular factors to modulate cellular transcription, protein degradation, calcium signaling, cell cycle, and apoptotic and proliferation pathways. Studies on HBx protein have proposed that overexpression of HBx protein can increase the proliferation of hepatocytes, while knocking down HBx protein reduces HCC tumorigenicity.

HBx can bind to and stimulate the transcription of the hypoxia inducible factor HIF1-α. HIF-1α is a transcriptional factor that is degraded in the presence of oxygen. Transcription target of HIF-1α include VEGF and LOXL2. LOXL2 was found overexpressed in various types of solid tumor including HCC. LOXL2 modifies the tumor microenvironment by cross-linking collagen to promote cancer cell local invasion and metastasis. In this study, we tried to explore whether HBV promotes HCC development and metastasis through the HBx/HIF-1α/LOXL2 pathway.

We utilized our HBx expressing cell line systems HepG2 (Tet-ON), HepG2 (Tet-OFF) and LO2 (Tet-ON) to mimic the HBx expressing liver and HCC cells. We have successfully demonstrated an up-regulation of HIF-1α protein and LOXL2 mRNA expression in our HBx expressing cell lines, and found that these up-regulation effects can be enhanced under hypoxic condition. These results confirmed the hypothesis that HBx can stabilize HIF-1α and up-regulate transcription of LOXL2.

However, the HBx/HIF-1α/LOXL2 pathway we confirmed in the cell line systems cannot be confirmed in human HCC clinical samples. Our in vivo study using 100 random human HCC samples provide no evidence for the correlations between HBx, GLUT1 (a HIF-1 marker), LOXL2 and cross-linked collagen. We believed that this unexpected result may be due to the complexity in human HCC cases where patients may have different clinical backgrounds. Furthermore, the expression of collagen, LOXL2 and GLUT1 can be affected by other factors which also can be found up-regulated in HCC tissues. We have given some potential solutions to this negative result, and hope to provide directions for further studies on the HBx/HIF-1α/LOXL2 pathway.