We are using a global genomic approaches to profile clinical specimens that are associated with different stages of liver diseases. For example, we have identified a unique signature that may be useful in diagnosing patients with early onset of liver cancer. We have also developed and validated a unique molecular signature based on the mRNA gene expression of metastatic primary hepatocellular carcinoma (HCC) specimens to predict prognosis and metastasis of HCC patients. Importantly, we found that this molecular signature could identify those patients who were most at risk for recurrence even in patients with early stage disease. We have also developed a unique molecular prognostic signature based on mRNA gene expression of the liver microenvironment of HCC patients. We found that a predominant humoral cytokine profile occurs in the metastatic liver microenvironment and that a shift toward anti-inflammatory/immune-suppressive responses may promote HCC metastases. Interestingly, the tumor signature is principally different from that of liver microenvironment. In addition, we have used molecular profiling to identify five genes that may serve as biomarkers for early onset of HCC, especially for those that are negative for alpha-fetoprotein. We have also explored the role of small non-coding RNAs, termed microRNAs, in HCC metastasis and survival. We found that certain microRNA expression changes are associated with metastasis and could significantly predict patient survival and relapse even in early stage disease. We have also found that the expression of certain microRNAs (e.g. microRNA-26) are gender-related. In fact, patients with low microRNA-26 expression had poor survival and were better responders to interferon therapy than those with normal expression. We have since developed a qRT-PCR-based matrix template and scoring algorithm (MIR26-DX) to assign patients into either low or high microRNA-26 groups. Patients with low microRNA-26 levels selected by the template were those that responded favorably to interferon-alpha therapy. We have now initiated a multi-center randomized control clinical trial in China based on these findings (NCT01681446). We have also recently used integrative approaches to identify HCC driver genes. For example, we have combined high-resolution, array-based comparative genomic hybridization and transcriptome analysis of HCC samples and found a 10-gene signature associated with chromosome 8p loss and poor outcome. The signature was validated in 2 independent HCC cohorts and breast cancer cohorts. Functional in vitro and in vivo studies demonstrated that three gene products among the 10-gene signature have tumor suppressive properties. We have also integrated metabolite and mRNA profiles to define key signaling events that can alter the fitness of EpCAM+ AFP+ HCC cancer stem cells. Our analysis revealed tumor-specific and stem-cell-like-specific metabolites linked to patient survival along with correlating significant genes in the stem cell-like tumor subgroup. We found a cluster of fatty acids and 273 surrogate genes that could predict HCC survival in independent HCC patients. Gene network analysis and functional data revealed that stearoyl CoA desaturase (SCD), a key enzyme involved in fatty acid biosynthesis, and its related metabolites were highly elevated in stem cell-like HCC and are associated with HCC survival and may functionally contribute to HCC stemness and aggressiveness. To aid in the integration of multiple omics data, we have proposed a novel conceptual framework to discover patterns of miRNA-gene networks which could be used for patient stratification in HCC. This integrative subgraph mining approach, called iSubgraph, allows for transformation of microarray data into graph representation that encodes miRNA and gene expression levels and the interactions between them as well. The iSubgraph algorithm was able to detect cooperative regulation of miRNAs and genes even if it occurred only in some patients. Next, the miRNA-mRNA modules were used to predict HCC subgroups via patient clustering by mixture models. The class predictions are highly stable and the HCC subgroups identified by the algorithm have different survival characteristics. Pathway analyses of the miRNA-mRNA co-modules identified by the algorithm demonstrate key roles of Myc, E2F1, let-7, TGFB1, TNF and EGFR in HCC subgroups. Thus, our method can integrate various omics data derived from different platforms and with different dynamic scales to better define molecular tumor subtypes. Serum alpha-fetoprotein (AFP) is normally highly expressed in the liver during fetal development but is reactivated in 60% of HCC tumors and associated with poor patient outcome. We hypothesized that AFP+ and AFP- tumors differ biologically. Using microarray-based global microRNA profiling, we found that miR-29 family members were the most significantly down-regulated miRNAs in AFP+ tumors. We also found a significant inverse correlation between miR-29 and DNMT3A gene expression suggesting that they might be functionally antagonistic. Moreover, global DNA methylation profiling revealed that AFP+ and AFP- HCC tumors have distinct global DNA methylation patterns and that increased DNA methylation is associated with AFP+ HCC. Experimentally, we found that AFP expression induces cell proliferation, migration and invasion along with inhibition of miR-29a and induction of DNMT3A. AFP also inhibited transcription of the miR-29a/b-1 locus and this effect is mediated through c-MYC binding to the transcript of miR-29a/b-1. Further, AFP expression promotes tumor growth of AFP- HCC cells in nude mice. Our findings indicate that tumor biology differs considerably between AFP+ HCC and AFP- HCC and that AFP is a functional antagonist of miR-29, which may contribute to global epigenetic alterations and poor prognosis in HCC. While HCC is the most frequent form of primary liver cancer worldwide, with rising incidence in the western countries, the incidence of cholangiocarcinoma (CCA), a bile-duct-related cancer and second most frequent PLC, is prevalent, especially in the north-east area of Thailand. Although genomics-based studies have been performed to profile human tumors among several populations, such studies have not yet been explored among patients in Thailand. We have thus initiated the Thailand Initiative for Genomics and Expression Research in Liver Cancer (TIGER-LC) to provide a comprehensive global analysis of genomic alterations related to the primary liver cancer types in Thai liver cancer patients. The findings of this study will not only allow us to more clearly understand the biological signaling related to primary liver cancer types and subtypes, but will also allow us to identify relevant health disparity genomic biomarkers that are related to etiological risk factors or ethnic groups by comparing our results with that of other populations. Furthermore, our study may identify novel, clinically useful biomarkers or targets to improve the health and outcome of patients suffering from this deadly disease. Thus, TIGER-LC is expected to have a major global health impact on improving the outcome of patients with liver cancer. Our findings have been extremely fruitful and offer useful tools for personalized patient management and also challenge the current paradigm of tumor evolution. Clearly, expression profiling has expanded our knowledge of the global changes that occur in liver cancer, and has provided numerous insights into the molecular mechanisms of this disease. In addition, these studies will undoubtedly contribute to the establishment of novel markers with potential diagnostic and prognostic value, as well as potential therapeutic targets for direct clinical intervention.
Showing the most recent 10 out of 48 publications
