| According to the latest studies,circular RNAs?circRNAs?widely exist in human normal tissues,cancerous tissues,serum and plasma.Furthermore,they have been reported to play important roles in the development of cancers,such as bladder,gastric and colon cancer.Hovever,the role of circRNAs in hepatocellular carcinoma?HCC?was unclear.In the present study,we explored the function and diagnostic value of circRNAs in HCC.Our study was consisted of two parts.In the first part of our study,we detected the expression of circRNAs in five pairs HCC and peri-tumor tissues by high-throughput sequencing and further explored the function of cSMARCA5?the circRNA derived from the exons 15 and 16 of the SMARCA5 gene,CircBase ID: hsacirc0001445?in HCC.Firstly,with RNA-sequencing,we found 4727 highly abundant?average RPM?0.1 in HCC or peri-tumoor tissues?circRNAs.Most of them were from exons and less than 1000 nucleotides.Importantly,there were 513 novel circRNAs in human liver tissues and 236 differentially expressed circRNAs between HCC and paired peri-tumor tissues.Among them,108 was upregulated and 128 was downregulated in HCC tissues,compared with peri-tumor tissues.We successfully validated five upregulated?cDNAJC6,cGPC3,cME1,cPVT1 and cSATB2?and five downregulated?cCYP2C8,cKCNN2,cLIFR,cPIK3R1 and cSMARCA5?circRNAs by qRT-PCR?quantitative real-time polymerase chain reaction?,agarose gel electrophoresis and Sanger sequencing.Secondly,using RNase R digestion,actinomycin D treatment,FISH?fluorescence in situ hybridization?and qRT-PCR,we proved that cSMARCA5 was stable,circular,highly abundant,predominantly located in cytoplasma and downregulated in HCC.Thirdly,by constructing cSMARCA5-overexpressing vectors,qRT-PCR,Northern blot,RNA interference and RNA immunoprecipitation,et al,we found that the biogenesis of cSMARCA5 was promoted by the pairing of I14RC?reverse complementary sequences in intron 14?and I16RC?reverse complementary sequences in intron 16?,while DHX9(?DExH-Box Helicase 9,an abundant nuclear RNA helicase?could inhibit the production of cSMARCA5 by binding to I14 and I16 RC and inhibiting their pairing.In addition,the mRNA and protein level of DHX9 was upregulated and was was negatively associated with the expression of cSMARCA5 in human HCC tissues.Furthermore,in a cohort with 163 HCC patients,we found that the lower expression of cSMARCA5 in HCC was associated with aggressive clinicopathological characteristics?Edmondson's grade,tumor size,microvascular invasion,TNM stage and BCLC stage?and predicted poorer overall survival and recurrence free survival for HCC patients after hepatectomy.Moreover,we successfully overpxpressed cSMARCA5 by using cSMARCA5-overexpressing vector and knocked down the knocked down the expression of cSMARCA5 in HCC cell lines by using small interfering RNAs?siRNAs?targeting the back-splice sequence.The overexpression of cSMARCA5 retarded the growth and metastasis of HCC cells in vitro and in vivo.The silence of of cSMARCA5 promoted the growth and metastasis of HCC cells in vitro.Finally,by using miRanda miRNA target prediction tool,RIP with an antibody against AGO2,circRIP?circRNAs in vivo precipitation?,dual-luciferase reporter assay,biotincoupled miRNA capture and RNA FISH,we proved that cSMARCA5 could act as a sponge for miR-17-3p and miR-181b-5p.Next,we found that cSMARCA5 could retard the growth and metastasis of HCC cells mainly by protecting TIMP3?a tumor suppressor and common target of miR-17-3p and miR-181b-5p?from degradation by miR-17-3p and miR-181b-5p.In conclusion,cSMARCA5,a circular RNA,inhibits the progression of HCC and serves as a pontential therapeutic target for HCC.In the second part of our study,we used microarray to screen circRNAs in five plasma samples from patients with HBV-related HCC and five plasma samples from patients with chronic hepatitis B.We found 371 differentially expressed circRNAs.Among them,326 were upregulated,and 45 were downregulated in plasma samples from patients with HBVrelated HCC,compared with the plasma samples from patients with chronic hepatitis B.After overlapping the 326 upregulated circRN with the circRNAs expressed in HCC tissues,we obtained 10 candidate circRNAs.By qRT-PCR,agarose gel electrophoresis,Sanger sequencing,RNase R digestion and incubation at room temperature,we successfully identified four of them?hsacirc0000976,hsacirc0003506,hsacirc000775 and hsacirc0139897?and proved that they were circular and stable in plasma and liver tissues.We used qRT-PCR and ?-actin was used as the inter control when detecting the expression of these four circRNAs in liver tissues.As there was no accepted internal control when detecting the expression of RNAs in plasma,we used absolute quantification when detecting the expression of these four circRNAs in plasma.As a result,the expression of hsacirc0000976 and hsacirc0007750 in HCC plasma was higher than that in healthy,CHB or liver cirrhosis plasma,their expression in pre-hepatectomy HCC plasma was higher that in post-hepatectomy HCC plasma,their expression in HCC tissues was higher then that in peri-tumor tissues and their expression in pre-hepatectomy HCC plasma was positively correlated with that in HCC tissues.Although the expression of hsacirc0139897 showed no significant difference between HCC tissues and peri-tumor tissues,its expression in HCC plasma was higher than that in healthy,CHB or liver cirrhosis plasma,its expression in prehepatectomy HCC plasma was higher that in post-hepatectomy HCC plasma,and its expression in pre-hepatectomy HCC plasma was positively correlated with that in HCC tissues.We supposed that it was due to that HCC tissues tend to secrete more hsacirc0139897 into plasma than normal liver tissues do.The expression of hsacirc0003506 in HCC,healthy,CHB and liver cirrhosis plasma showed no significant difference.Therefore,we chose hsacirc0000976,hsacirc0007750 and hsacirc0139897 for further study.Using ROC curve?Euclidean distance?analysis from Cutoff Finder?http://molpath.charite.de/cutoff/?,we determined that the optimal cut off values of hsacirc0000976,hsacirc0007750 and hsacirc0139897 for the diagnosis of HCC were 1067,4324 and 1108 copies/ml plasma,respectively.Using logistic regression,we combined these three circRNAs as a panel?termed as CircPanel?.In a training set with 313 participants?158 with HBV-related HCC,50 with HBVrelated liver cirrhosis,52 with CHB and 53 healthy controls?and a validation set with 306 participants?152 with HBV-related HCC,50 with HBV-related liver cirrhosis,54 with CHB and 50 healthy controls?,we used the sensitivity,specificity,and area under the receiver operating characteristic curve?AUC?to evaluate diagnostic performance of CircPanel.CircPanel showed a higher diagnostic accuracy than AFP?cut off: 20 ng/ml?to distinguish individuals with HCC from controls?including healthy,CHB and liver cirrhosis?in both training set?AUC 0.863 [95% CI 0.819–0.907] vs 0.790 [0.738–0.842],P=0.036?and validation set?AUC 0.843 [0.796–0.890] vs 0.747 [0.691–0.804],P=0.011?,and the combination of CircPanel and AFP further improved diagnostic accuracy in both training set?AUC 0.878 [0.836–0.920] vs 0.790 [0.738–0.842],P=0.010?and validation set?AUC 0.863 [0.819–0.908] vs 0.747 [0.691–0.804],P=0.002?.CircPanel also showed a higher diagnostic accuracy than AFP to distinguish individuals with Small-HCC?solitary,diameter ?3cm?from controls in both training set?AUC 0.862 [0.796–0.928] vs 0.680 [0.589–0.770],P=0.001?and validation set?AUC 0.838 [0.776–0.900] vs 0.699 [0.613–0.785],P=0.011?,and the combination of CircPanel and AFP further improved diagnostic accuracy in both training set?AUC 0.873 [0.817–0.929] vs 0.680 [0.589–0.770],P<0.001?and validation set?AUC 0.874 [0.823–0.925] vs 0.699 [0.613–0.785],P=0.001?.Importantly,CircPanel also showed a high diagnostic accuracy to distinguish individuals with AFP-negative HCC from controls in both both training set?AUC 0.838 [0.761–0.914]?and validation set?AUC 0.857 [0.793–0.922]?,and to distinguish individuals with AFP-negative Small-HCC from controls in both training set?AUC 0.881 [0.797–0.965]?and validation set?AUC 0.897 [0.827–0.968]?.We also divided the participants from the control group into healthy,CHB and liver cirrhosis groups and evaluated the diagnostic performance of CircPanel for HCC vs healthy,HCC vs CHB,HCC vs liver cirrhosis,Small-HCC vs healthy,Small-HCC vs CHB,SmallHCC vs liver cirrhosis,AFP-negative HCC vs healthy,AFP-negative HCC vs CHB,AFPnegative HCC vs liver cirrhosis,AFP-negative Small-HCC vs healthy,AFP-negative SmallHCC vs CHB and AFP-negative Small-HCC vs liver cirrhosis.To sum up,CircPanel,a combination of three circRNAs,has a high diagnostic accuracy in HCC and can be used in the diagnosis of Small-HCC and AFP-negative HCC. |
PDF Full Text Request