AGR2 Promotes Metastasis In Hepatocellular Carcinoma And Characterization Of HDGF Interactome

This dissertation consists of two chapters:AGR2 promotes metastasis in hepato-cellular carcinoma and characterization of HDGF interactome.Chapter One:AGR2 promotes metastasis in hepatocellular carcinoma.Hepatocellular carcinoma (HCC) is one of the most common and aggressive ma-lignant tumor worldwide. High rate of recurrence and metastasis of HCC remains one of the major obstacles for HCC treatment, because the molecular basis of HCC me-tastasis is still poorly understood, with few effectual diagnostic biomarkers and the-rapeutic targets thus far. Therefore to find the HCC metastasis-related genes and cla-rify the metastasis mechanism would be crucial for the treatment of HCC.At the beginning, comparative proteome analysis was performed in order to cha-racterize the molecular profiling of metastatic HCC. The expressed proteomes were compared among four human hepatocellular carcinoma syngeneic-alike cell lines with increased metastasis potentials (MHCC97L, MHCC97H, HCCLM3和HCCLM6), by two-dimensional gel electrophoresis (2DE) followed with liquid chromatography ion trap mass spectrometry analysis.46 leading differentially expressed protein spots were identified, and particularly the expression of AGR2 (anterior gradient homolog 2) was apparently altered in HCC cell lines with metastatic potential.AGR2 was overexpressed in many types of cancers, and it was supposed to pro-mote tumor growth and metastasis. However, the function and mechanism of AGR2 in tumors is still unclear, especially in the hepatocellular carcinoma. Thus, in this chap-ter, we tried to figure out whether the overexpression of AGR2 would also exhibit in a metastatic HCC, and by what molecular mechanism AGR2 contributes to these ma-lignant features, such as cell invasion and metastasis.1. AGR2 was up-regulated in metastatic cell lines and HCC samples. The expression of AGR2 was up-regulated with the increased HCC metastatic potential in HCC cell lines identified by 2DE and Western blot. We also evaluated the protein ex-pression of AGR2 in a panel of non-metastasis HCC cell lines (SNU398, Hep3B, HepG2, QGY7703, QSG7701 and L02), most of the non-metastasis HCC cell lines expressed low levels of AGR2 in contrast with that of MHCC97L. We then analyzed AGR2 mRNA expression profile in 48 HCC samples including 24 pri-mary HCCs with metastasis and 24 HCCs without metastasis using real-time PCR. The expression level of AGR2 in HCC with metastasis was apparently higher than that in HCC without metastasis (P<0.001). The expression level of AGR2 was also positive related to tumor capsule (P=0.002) and TNM staging (P=0.003).2. High expression of AGR2 increased invasive ability of HCC cells in vitro. Endo-genous AGR2 of MHCC97H was disrupted by transiently transfected with siRNA, and AGR2 disruption caused approximately 40% decrease in the invasiveness of MHCC97H. MHCC97L and HepG2 cells were stably infected with both AGR2-containing retroviruses and blank retrovirus, AGR2 overexpression did enhance a significant increase in the invasiveness for MHCC97L (30%) and HepG2 (50%), respectively. These results suggested clearly that AGR2 could be a pro-metastatic gene and is partially responsible for cellular invasion in HCC cells.3. Overexpression of AGR2 promoted tumor growth and metastasis of HCC in vivo. Three groups of HepG2 cells were subcutaneously implanted into three different regions on mice flanks. HepG2 cells with pBABE-AGR2 transfected showed promotion in both size and weight in tumor growth, statistically differed from those of vector control. In the orthotropic implantation model, implanted tumor survived in all nude mice. According to autopsy and microscopic examination, 62.5%(5/8) mice in the AGR2 high expressed group have intrahepatic metastases much higher than the control groups (12.5%). Lung metastasis was detected in 1 (12.5%) nude mice of the AGR2 high expressed group, while no tumor nodules were detected in the lung of the control mice.4.18 proteins were identified to form complex with AGR2 by SBP/FLAG-TAP coupled with LC-MS/MS. We used SBP/FLAG-TAP-LC-MS/MS for identifica-tion of proteome-wide binding partners of AGR2. After two rounds of purifica-tion sequentially over anti-FLAG M2 agarose beads and then Streptavidin beads, the proteins on Streptavidin beads were digested by trypsin, then separated by nano-LC and identified eventually by LTQ MS. To ensure an analytical reprodu-cibility, three independent purifications were repeated. By the end, there were 18 potential-interactors of AGR2 were found. To confirm some important interac-tions between AGR2 and potential binding partners, TAK1, TAB1, STK4, and PDCD6 were then selected for co-immunoprecipitation experiment. The protein data were entered into Ingenuity Pathways Analysis (IPA) tools and analyzed by Core Analysis. AGR2 interacting proteins focused on the MAPK and apoptosis pathways.5. The pro-metastatic effect of AGR2 on HCC is likely via phosphorylation of ERK1/2. AGR2 does not affect apoptosis. Endogenous AGR2 disruption with siRNA in MHCC97H or stable exogenous expression of AGR2 in MHCC97H and HepG2 did not affect apoptosis. AGR2 could promote the phosphorylation levels of ERK1/2. In HEK 293T, AGR2 alone did not have a significant effect on the changes in the phosphorylation levels of ERK1/2(p44/p42), SAPK/JNK(p54/46) and p38. When TAK1 and TAB1 were co-transfected with AGR2 in HEK 293T, these coexisted proteins were able to increase the phospho-rylation levels of JNK and ERK1/2. However, the phosphorylation levels of JNK and ERK1/2 were higher than those without AGR2, indicating that the AGR2 promote the phosphorylation levels of JNK and ERK1/2 likely. Stable expression of AGR2 in HepG2 cells could increase the phosphorylation levels of ERK1/2 but not of JNK. These results implied that the function of AGR2 might due to coordinate and cooperate with TAK1 and TAB1 and then promote phosphoryla-tion of ERK1/2.Cancer metastasis is a multi-step, multiple genes involved process, and thus one gene or protein cannot work alone. In this chapter, we found AGR2 worked as a pro-metastatic gene in hepatocellular carcinoma. We also identified 18 proteins might work together with AGR2, and the pro-metastatic effect of AGR2 on HCC is likely via phosphorylation of ERK1/2. However, we still did not know whether AGR2 was an upstream regulator or a downstream executor, and its detail function and regulating pathway need further investigation.Chapter Two:Characterization of HDGF interactome.Hepatoma-derived growth factor (HDGF), is a nuclear protein with both mitogenic and angiogenic activity. It is involved in many normal tissues development, and overexpression of HDGF is found in a number of human cancers. Lots of studies have been done about the effects of HDGF on cancer occurrence and development, howev-er, the mechanism of its mitogenic and angiogenic activity still remains unknown. Knowledge about HDGF-interactors is also very limited, therefore the analysis of HDGF interactome will be important for its function and mechanism integration process.1. The feasibility study of SBP and FLAG tag to be used in the HDGF complex pu-rification. First, immunofluorescence experiment showed that SBP/FLAG tag did not change HDGF nuclear location; then, we found that SBP/FLAG tag did not interfere with the DNA binding ability of HDGF by EMSA; at the same time, the SBP/FLAG tag did not block protein binding, because several known HDGF in-teracting proteins such as C12orf11 and C15orf44 were also identified by SBP/FLAG-TAP coupled with LC-MS/MS in this work. Then the purification ef-ficiency was determined by Western blot. Without cross-linking the recovery rate of FLAG and SBP tag were 52% and 43%, and under cross-linking condition, the recovery rate were 23% and 40%. The purification efficiency of FLAG tag was significantly affected by cross-linking.2. Proteome-wide identification of HDGF interacting proteins by SBP/FLAG-TAP coupled with LC-MS/MS. The cells were lysed and then HDGF complex was pu-rified by the anti-FLAG M2 agarose beads and then Streptavidin beads. The complexes on Streptavidin beads were boiled with 1×SDS-loading buffer and se-parated by SDS-PAGE. Silver stained bands were excised and digested in gel, then identified by LC-MS/MS. Four independent experiments were performed. Totally,132 proteins which appeared in at least two experiments with at least 3 unique peptides were identified.3. The protein-protein interactions of HDGF depend on its HATH domain. The in-teractome of the HATH domain (1～98aa of HDGF) and NO-HATH (99～240aa of HDGF) were characterized by SBP/FLAG-TAP coupled with LC-MS/MS. The HATH domain could bind 95 proteins, among which 88 proteins were detected at least once in the HDGF experiments. The NO-HATH only had 5 proteins identi-fied which maybe the contaminated proteins during sample preparation.4. The identity and specificity of the identified proteins were checked by Co-IP and fluorescent co-localization analysis. Histone H2B, PARP-1, PRKDC, DDX5, SFRS1, HDAC1 and SSRP1 could co-immunoprecipitate with HDGF and HATH but not with NO-HATH and vector control. We also chose four proteins (DDX5, SFRS1, HDAC1 and SSRP1) to validate their co-localization with HDGF and its HATH domain.5. Characterization of RNA in the HDGF complex. HDGF complex was purified with Streptavidin beads and was then extracted by Trizol reagent. The purified RNA was reverse transcripted and cloned into T vector. Thirty clones were se-quenced,28S rRNA and 18S rRNA were the main components, PRELID1 and SFSWAP mRNA were also identified.6. The protein-RNA interaction of HDGF depends on the HATH domain. We ex-tracted the RNA from HDGF, HATH and NO-HATH complex by SBP-RIP. RNA extracted from HDGF and HATH complex had almost the same bands, but RNA from the control and NO-HATH complex did not show visible band. And the 28S rRNA,18S rRNA, PRELID1 and SFSWA could only be amplified from the HDGF and HATH cDNA but not from the control and NO-HATH by RT-PCR.7. Biological process analysis of HDGF interactome showed that HDGF was a mul-tifunctional protein and might participate in many cellular activities, such as ri-bosome biogenesis, RNA processing and DNA damage repair. We also found that HDGF might regulate transcription with two different mechanisms:modulating chromatin structure and acting together with transcriptional coactivators or core- pressors.In this study, through a tandem affinity purification strategy (TAP) followed with LC-MS/MS,132 proteins were shown to form complex with HDGF. RNA especially rRNA was also found in the HDGF complex through SBP-tag based RNA co-immunoprecipitation (RIP) assay. We then confirmed that these interactions were mainly mediated by the HATH domain. The interactome shows HDGF is a multifunc-tional protein and might participate in many celluar events. The identification of HDGF interactome will be useful to further our understanding of HDGF-mediated cellular functions.