Hepatology - 2017 - Han - Circular RNA circMTO1 Acts As The Sponge of microRNA 9 To Suppress Hepatocellular Carcinoma

A MERICAN A SSOCIATION FOR
T HE STUDY OF LIVER D I S E ASES
HEPATOLOGY, VOL. 66, NO. 4, 2017
Circular RNA circMTO1 Acts

as the Sponge of MicroRNA-9
to Suppress Hepatocellular
Carcinoma Progression
Dan Han,1,2* Jiangxue Li,1* Huamin Wang,3 Xiaoping Su,2 Jin Hou,2 Yan Gu,2 Cheng Qian,2 Yun Lin,2 Xiang Liu,2
Mingyan Huang,2 Nan Li,2 Weiping Zhou,4 Yizhi Yu,2 and Xuetao Cao1–3
Noncoding RNAs play important roles in cancer biology, providing potential targets for cancer intervention. As a new
class of endogenous noncoding RNAs, circular RNAs (circRNAs) have been recently identified in cell development and
function, and certain types of pathological responses, generally acting as a microRNA (miRNA) sponge to regulate gene
expression. Identifying the deregulated circRNAs and their roles in cancer has attracted much attention. However, the
expression profile and function of circRNAs in human hepatocellular carcinoma (HCC) remain to be investigated. Here,
we analyzed the expression profile of human circRNAs in HCC tissues and identified circMTO1 (mitochondrial transla-
tion optimization 1 homologue; hsa_circRNA_0007874/hsa_circRNA_104135) as one circRNA significantly down-regulated
in HCC tissues. HCC patients with low circMTO1 expression had shortened survival. By using a biotin-labeled
circMTO1 probe to perform RNA in vivo precipitation in HCC cells, we identified miR-9 as the circMTO1-associated
miRNA. Furthermore, silencing of circMTO1 in HCC could down-regulate p21, the target of oncogenic miR-9, resulting
in the promotion of HCC cell proliferation and invasion. In addition, the tumor-promoting effect of circMTO1 silencing
was blocked by miR9 inhibitor. Intratumoral administration of cholesterol-conjugated circMTO1 small interfering RNA
promoted tumor growth in HCC-bearing mice in vivo. Conclusion: circMTO1 suppresses HCC progression by acting as
the sponge of oncogenic miR-9 to promote p21 expression, suggesting that circMTO1 is a potential target in HCC treat-
ment. The decrease of circMTO1 in HCC tissues may serve as a prognosis predictor for poor survival of patients. (HEPA-
TOLOGY 2017;66:1151-1164).
C
ircular RNAs (circRNAs) represent a new class characterized by covalently closed loop structures with
of endogenous noncoding RNAs that may reg- neither 50 to 30 polarity nor polyadenylated tail.(1,2) circR-
ulate gene expression in mammals. Unlike linear NAs are conserved and stable, and numerous circRNAs
RNAs terminated with 50 caps and 30 tails, circRNAs are seem to be specifically expressed in a given cell type or
Abbreviations: AFP, alfa-fetoprotein; ARID1B, AT-rich interaction domain 1B; circRNAs, circular RNAs; ELISA, enzyme-linked immunosorbent
assay; FAM13B, family with sequence similarity 13 member B; FARP2, FERM, ARH/RhoGEF and pleckstrin domain protein 2; FISH, fluorescence
in situ hybridization; HCC, hepatocellular carcinoma; ISH, in situ hybridization; kaps, K-adaptive partitioning statistical; miRNA, microRNA;
MMP2, matrix metalloproteinase 2; MTO1, mitochondrial translation optimization 1 homologue; PBS, phosphate-buffered saline; PCNA, proliferat-
ing cell nuclear antigen; RIP, RNA in vivo precipitation; siRNA, small interfering RNA.
Received July 27, 2016; accepted May 12, 2017.
Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1002/hep.29270/suppinfo.
Supported by grants from the National Key Basic Research Program of China (2014CB542101), the National Natural Science Foundation of China
(31390431, 81123006, and 81571564), and the CAMS Innovation Fund for Medical Sciences (2016-12M-1-003).
*The first two authors contributed equally to this work.
Copyright V C 2017 by the American Association for the Study of Liver Diseases.
View this article online at wileyonlinelibrary.com.

DOI 10.1002/hep.29270
Potential conflict of interest: Nothing to report.
1151
15273350, 2017, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.29270, Wiley Online Library on [03/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HAN, LI, ET AL. HEPATOLOGY, October 2017
developmental stage.(3,4) These cell-type– or late appearance of symptoms, and the prognosis of
developmental-stage–specific expression patterns indicate these HCC patients is still poor because of high fre-
that circRNAs may play important roles in many physio- quency of metastasis and recurrence. These challenges
logical and pathophysiological processes. Natural endoge- give rise to the urgent need to identify potential bio-
nous circRNAs are inherently resistant to exonucleolytic markers for prognosis predication and find new targets
RNA decay and contain selectively conserved microRNA to design a more powerful therapeutic approach. In the
(miRNA) target sites. So, circRNAs can serve as efficient last decade, numerous noncoding RNAs, including
miRNA sponges, interacting with miRNA to regulate miRNAs and long noncoding RNAs, have been found
gene expression. The dysregulation and functions of to be deregulated in HCC patients, having potential
miRNAs have been extensively investigated in almost applications in the clinic. However, the expression pro-
every biological process; however, the expression profiles file, function, and deregulation of circRNAs in HCC
and functions of the newly identified circRNAs in remain to be identified.
the specific biological activity still require further In our study, we analyzed the expression profile of
investigation. circRNAs in HCC tissues and identified that circRNA
circRNAs have been recently verified to be involved circMTO1 (mitochondrial translation optimization 1
in developmental processes, regulation of cell function, homologue; hsa_circRNA_0007874/hsa_circRNA_104135)
pathogenesis of heart diseases, and development of is significantly down-regulated in HCC tissues and closely
neurodegenerative diseases like Alzheimer’s disease.(5) related to the prognosis of HCC patients. We found that
In the study of carcinogenesis and progression, just like circMTO1 may function as the sponge of oncogenic miR-
the extensive involvement of miRNAs in cancer, the 9 to up-regulate p21 expression and consequently suppress
expression profiles and the roles of circRNAs in cancer HCC progression. Therefore, decrease of circMTO1 may
development and progression have attracted much serve as a biomarker for prognosis predication and as a
attention as an emerging filed. Recently, certain kinds potential therapeutic target for HCC patients.
of circRNAs have been found to be significantly
deregulated in gastric cancer, colorectal cancer, and
esophageal squamous cancer, and these deregulated
Materials and Methods
circRNAs are suggested to participate in cancer devel-
TISSUES AND CELL LINES
opment.(6-8) These observations indicate that circR-
NAs may be a new kind of potential biomarkers and The 289 samples of HCC and paired adjacent liver
therapeutic targets for cancer; however, elucidating the tissues were obtained from patients during operation.
deregulated circRNAs and identifying their functions All samples were collected with the consent of patients,
is still an ongoing process in cancer investigation. and the experiments were approved by ethics commit-
Hepatocellular carcinoma (HCC) is one of the lead- tee of Eastern Hepatobiliary Hospital (Shanghai,
ing causes of cancer-related death worldwide, espe- China). All tissue specimens were stored at 2808C
cially in China.(9) Only 30%-40% of HCC patients are until use. HCC cell lines HepG2, SMMC-7721,
amenable to curative resection, owing partially to the QGY-7701, and SK-Hep1 were cultured in
ARTICLE INFORMATION:
From the 1Institute of Immunology, Zhejiang University School of Medicine, Hangzhou, China; 2National Key Laboratory of Medical
Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China; 3Department of Immunology & Center
for Immunotherapy, Institute of Basic Medical Sciences, Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing,
China; and 4Third Department of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital, Shanghai, China.
ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO:

Xuetao Cao, M.D., Ph.D. Yizhi Yu, M.D., Ph.D.
Institute of Immunology, Zhejiang University School of Medicine National Key Laboratory of Medical Immunology &
866 Yuhangtang Road, Hangzhou 310058, China Institute of Immunology, Second Military Medical University
E-mail: caoxt@immunol.org 800 Xiangyin Road, Shanghai 200433, China
Tel: (186-571) 8820 8284; E-mail: yuyz@immunol.org
or Tel: (186-21) 5562 0605
1152
HEPATOLOGY, Vol. 66, No. 4, 2017 HAN, LI, ET AL.
Dulbecco’s modified Eagle’s medium with 10% fetal REAL-TIME qRT-PCR ANALYSIS
bovine serum (Gibco, NY, USA) as routine.(10)
Total RNA was extracted using TRIzol reagent (Invi-
trogen, Carlsbad, CA, USA). RNA concentration was
EXPRESSION PROFILE ANALYSIS measured by Nonodrop, and each paired sample was
OF circRNAs adjusted to the same concentration. Real-time qPCR
The circRNAs chip (Arraystar Human circRNAs was performed as described.(13,14) Glyceraldehyde 3-
chip; ArrayStar, Rockville, MD, USA), containing phosphate dehydrogenase was used as the internal con-
5396 probes specific for human circRNAs splicing trol. The primers are listed in Supporting Table S1.
sites, was used. After hybridization and washing with
samples, seven pairs of HCC samples (tumor tissues IN SITU HYBRIDIZATION
and matched nontumor tissues) were analyzed on the The expression level of circMTO1 in tissues was
circRNAs chips. Exogenous RNAs developed by the evaluated by in situ hybridization (ISH) using a specific
External RNA Controls Consortium were used as digoxin-labeled circMTO1 probe on tissue arrays,
controls. The Gene Expression Omnibus Accession which contained 116 HCC samples (Superchip Bio-
number is GSE97332. tech, Shanghai, China). A quantitative scanning app-
roach to evaluate staining and expression of circMTO1
circRNAs IN VIVO PRECIPITATION was Aperio ImageScope V11 from Leica Com-
(circRIP) pany,(15,16) and the positivity value 3100 represents
the expression of circMTO1. The sequence of the
Biotin-labeled circMTO1 probe (50 -AAA GGA circMTO1 probe for ISH was Digoxin-50 -CAT GAC
AGG ATT ACA TGA CAT CTG ACC CAA ATC TGA CCC AAA ACA ACC-30 -Digoxin.
AAC AAC CCC ACT GAC A-30 -biotin) was syn-
thesized by Sangon Biotech, and the circRIP assay was FLUORESCENCE IN SITU
performed as previously described by us with minor
modification.(11) CircMTO1-overexpressing HepG2
HYBRIDIZATION
cells were fixed by 1% formaldehyde for 10 minutes, Hybridization was performed overnight with
lysed, and sonicated. After centrifugation, 50 lL of the circMTO1 and hsa-miR-9-5p probes. Specimens were
supernatant was retained as input and the remaining analyzed on a Nikon inverted fluorescence microscope.
part was incubated with a circMTO1-specific probes- The circMTO1 probe for fluorescence in situ hybridi-
streptavidin dynabeads (M-280; Invitrogen) mixture zation (FISH) was 50 -GGA AGG ATT ACA TGA
overnight at 308C. On the next day, an M-280 dyna- CAT CTG ACC CAA AAC AAC CCC ACT
beads-probes-circRNAs mixture was washed and incu- GAC-30 , and the hsa-miR-9-5p probe for FISH was
bated with 200 lL of lysis buffer and proteinase K to 50 -TCT TTG GTT ATC TAG CTG TAT GA-30 .
reverse the formaldehyde cross-linking. Finally, the
mixture was added with TRIzol for RNA extraction HCC-BEARING NUDE MOUSE
and detection.(11) MODEL PREPARATION AND
EXPERIMENTS IN VIVO
SILENCING AND OVEREXPRESSION
The HCC-bearing male nude mice model, SMMC-
OF circMTO1
LTNM, were raised and passaged as described.(14,17) At
Small interfering RNA (siRNA) of circMTO1 was the third week after subcutaneous incubation of HCC
synthesized by GenePharma (Shanghai, China), tar- tumor mass, 10 optical density cholesterol-modified
geting to the junction region of the circMTO1 circMTO1 siRNA, control siRNA (Ribobio, Guang-
sequence. HepG2 and SMMC-7721 cells were trans- zhou, China), or phosphate-buffered saline (PBS) were
fected with circMTO1 siRNAs or overexpressing plas- intratumorally injected every 3 days for 2 weeks. Tumor
mids using INTERFERin or jetPRIME (Polyplus size was measured and serum alfa-fetoprotein (AFP)
transfection, France), respectively, as described.(12) was detected using enzyme-linked immunosorbent
The sequence of circMTO1 siRNA was 50 -UCU assay (ELISA; Autobio, Zhengzhou, China). All ani-
GAC CCA AAA CAA CCC CdTdT-30 . mal experiments were undertaken in accord with the
1153
National Institutes of Health Guide for the Care and results showed that circMTO1 expression was signifi-
Use of Laboratory Animals, with the approval of the cantly decreased in 87.4% (228 of 261) HCC tissues as
Scientific Investigation Board of Second Military Medi- compared to that in the matched nontumor tissues
cal University (Shanghai, China). (Fig. 2A,B). Although some circRNAs, including cir-
cARID1B (ARID1B, AT-rich interaction domain
IMMUNOHISTOCHEMISTRY 1B;), circFAM13B (family with sequence similarity 13
member B), and circFARP2 (FERM, ARH/RhoGEF
Immunohistochemistry was performed as des- and pleckstrin domain protein 2), were decreased in
cribed.(18) Tissues sections were incubated with anti- HCC tissues (Fig. 1B), the preliminary results showed
bodies against p21 (CST, Danvers, MA, USA), matrix that silencing of these circRNAs had no significant
metalloproteinase 2 (MMP2; CST), or proliferating cell effect on HCC cell growth (data not shown), and only
nuclear antigen (PCNA; CST) at 48C overnight. The silencing of circMTO1 had a significant effect to pro-
ChemMate DAKO EnVision Detection Kit (Dako, mote HCC cell proliferation. Therefore, the high per-
Santa Clara, CA, USA) was used for immunostaining. centage of circMTO1 decrease in HCC tissues
inspired us to focus on the biological significance of
STATISTICAL ANALYSIS circMTO1 in HCC progression.
The Student t test was performed to analyze whether
two experimental groups have significant difference DECREASED circMTO1
using P < 0.05 as the significant criteria. Survival analy- EXPRESSION IS SIGNIFICANTLY
sis was performed by Kaplan-Meier curves and log-rank CORRELATED WITH POOR
test for significance in GraphPad Prism 5. PROGNOSIS OF HCC PATIENTS
Because circMTO1 expression is significantly
Results decreased in HCC tissues, we next analyzed whether
circMTO1 decrease is correlated with the prognosis of
DEREGULATED circRNAs AND HCC patients. A cohort of 116 HCC patients with
DECREASE OF circMTO1 IN HCC survival data was included, and circMTO1 expression
TUMOR TISSUES was determined using ISH analysis on tissue array
slides. circMTO1 expression was also significantly
To investigate circRNA expression profile in HCC decreased in HCC tissues as compared to that in
tissues, we analyzed seven pairs of HCC tissue samples matched nontumor tissues (Fig. 2C). More important,
(seven HCC tissues and seven matched nontumor liver decreased circMTO1 expression in HCC tissues was
tissues) by using the CircRNA microArray (Supporting significantly correlated with poor prognosis of HCC
Fig. S1A). Through expression intensity sorting within patients, shown as Kaplan-Meier survival curve using
HCC tumor and nontumor groups, the 10 mostly median value of circMTO1 expression as the cut-off
increased and decreased circRNAs in HCC tumor tissues (Fig. 2D). Furthermore, we also used the K-adaptive
as compared to that in nontumor tissues are shown in partitioning statistical (kaps) algorithm in R software
Fig. 1A and Supporting Fig. S1B. Thus, the circRNA to calculate the most efficient cut-off value of
expression profiles of HCC and nontumor liver tissue circMTO1 expression (76.89),(19-21) which can most
were suggested in this study. Next, we examined the significantly distinguish the outcomes of these HCC
expression of these 20 mostly changed circRNAs in patients (Fig. 2E). Together, these data suggest that
HCC and matched nontumor tissue samples from 21 decreased circMTO1 expression in HCC is correlated
patients to confirm their expression, and expression of with poor prognosis of patients, and the circMTO1
the six mostly changed circRNAs are presented in Fig. expression value of 76.89 may be efficient in the prog-
1B. Among them, we found that circMTO1 expression nosis prediction of HCC patients.
was consistently and significantly decreased in HCC
tumor tissues as compared to that in matched controls circMTO1 BINDS miR-9
(Fig. 1B). Thus, we focused on the expression and role of
IN HCC CELLS
circMTO1 in HCC progression in this study.
The expression of circMTO1 in HCC tissues was Because circRNAs function mainly as miRNA
then confirmed in another 261 HCC patients, and the sponges to bind functional miRNAs and then regulate
1154
FIG. 1. Deregulated circRNAs in HCC tumor tissues. (A) The heat map showed the top 10 most increased and decreased circRNAs
in HCC tissues as compared to that in the matched nontumor tissues analyzed by circRNAs Arraystar Chip. (B) Relative expression
of the six indicated circRNAs from 21 HCC tumor tissues and adjacent nontumor tissues listed in (A) measured by qPCR. Abbrevia-
tions: IFNGR2, interferon gamma receptor 2; N, nontumor tissues; T, tumor tissues; UBAP2, ubiquitin-associated protein 2.

FIG. 2. Decrease of circMTO1 expression in HCC tissues and its correlation with prognosis of patients. (A) The differential expres-
sion of circMTO1 in HCC tissues and matched nontumor tissues of 261 patients as indicated. The median expression level of each
circMTO1 level was indicated by horizontal line in scatterplot figure. (B) The ratio of circMTO1 expression in tumor tissues to non-
tumor tissues was calculated from its expression level of 261 paired HCC patients’ samples. (C) The expression of circMTO1 in
HCC was analyzed by ISH on HCC tissue chip (116 cases). Intensity of circMTO1 expression was calculated from HCC tissue chip.
(D) Prognostic significance of circMTO1 expression for HCC patients was performed with ISH values by using the median value as
the cutoff, and the observation time was 80 months. (E) Prognostic significance of circMTO1 expression for HCC patients was per-
formed with ISH values by using the kaps algorithm in R software to calculate the most efficient cut-off value of circMTO1 expres-
sion (76.89). Abbreviations: N, nontumor tissues; T, tumor tissues.
1156
gene expression, we next examined the potential miR- expression of circMTO1 and the highest expression of
NAs associated with circMTO1. Through MiRanda miR-9, QGY-7701 cells showed moderate expression
Program prediction, 99 miRNAs were predicated and of the two RNAs, and HepG2 and SMMC-7721 cells
listed as possible targets of circMTO1 (Supporting Table showed higher expression of circMTO1 and lower
S2). In order to fish out the functional miRNAs which expression of miR-9 (Supporting Fig. S3A). Therefore,
may interact with circMTO1 in HCC cells, we used a HepG2 and SMMC-7721 cells were selected for silenc-
circMTO1-specific probe to perform RNA in vivo pre- ing of circMTO1, and SK-Hep1 and QGY-7701 cells
cipitation (RIP), which we established before to identify were selected for overexpression of circMTO1. We con-
the miRNAs actually binding with target mRNA.(11) structed siRNA, which covers the back-splicing region
We purified the circMTO1-associated RNAs by of circMTO1 for silencing. After transfection with
circMTO1-specific probe, and then screened the poten- circMTO1 siRNA into HepG2 and SMMC-7721
tial miRNAs which were predicted by the miRanda HCC cells, the efficiency of circMTO1 silencing was
Program. measured and confirmed by qPCR (Supporting Fig.
Among the 99 candidate miRNAs predicted by S3B). To convincingly show that our transfection had
miRanda, we mainly screened and analyzed 20 miR- no effects on other MTO1 splicing products, qPCR
NAs, whose expression and function have been impli- was performed to detect expression of another
cated in HCC, including miR-204, miR-760, miR- circMTO1 (has_circRNA_0132266, isoform2) and
4476, miR-6876, miR-152, miR-3686, miR-3159, mRNA level of MTO1 (linearMTO1). The expression
miR-199a, miR181a-2, miR-211, miR-218-2, miR- of circMTO1 isoform2 (has_circRNA_0132266) and
3155a, miR-6730, miR-15a, miR-199b, miR-223, MTO1 mRNA showed no significant change after
miR-3156, miR-4286, miR-4302, and miR-9 (Sup- transfection of circMTO1 siRNA, confirming the spe-
porting Fig. S2). By RIP circMTO1 pull-down experi- cificity of circMTO1 silencing (Supporting Fig. S3C).
ments, we purified the circMTO1-associated RNAs In addition, we constructed the circMTO1 overexpress-
and analyzed the 20 candidate miRNAs in the complex. ing plasmid (pCD-circMTO1) with circular frame and
We found a specific enrichment of circMTO1 and circMTO1 sequence, and found that circMTO1 could
miR-9 as compared to the controls (Fig. 3A,B), whereas be significantly overexpressed in QGY-7701 and SK-
the other miRNAs had no enrichment (Supporting Fig. Hep1 HCC cells (Supporting Fig. S3D).
S2), determining that miR-9 is the circMTO1- Silencing of circMTO1 was performed in HCC cells
associated miRNA in HCC cells. Although the contro- to determine its effect on HCC progression and to inves-
versial roles of miR-9 in promoting or inhibiting tumor tigate its interaction with miR-9 (Fig. 4A, Supporting
growth were reported,(22-24) miR-9 was frequently Fig. S3B). circMTO1 silencing significantly promoted
found to be increased in HCC and its increased expres- cell proliferation (Fig. 4B), enhanced cell invasion (Fig.
sion was closely correlated with poor prognosis of HCC 4C), and reduced apoptosis (Fig. 4D) of HepG2 and
patients.(25,26) Thus, the interaction between miR-9 and SMMC-7721 cells, respectively, whereas circMTO1
circMTO1 was then focused on and investigated. By overexpression promoted apoptosis of QGY-7701 and
FISH analysis in HCC tumor and matched nontumor SK-Hep1 HCC cells (Fig. 4D). Moreover, miR-9 over-
tissues, we found that circMTO1 was colocalized with expression photocopied the tumor-promoting effects of
miR-9 in the cytoplasm and this colocalization was circMTO1 silencing in HepG2 and SMMC-7721 (Fig.
decreased in tumor as compared to that in matched 4B-D). To verify the miRNA sponge function of circ-
nontumor tissues (Fig. 3C). Together, these results sug- MOT1, we examined miR-9 expression and found that
gest that circMTO1 can bind miR-9 in the cytoplasm. circMTO1 had no effect on the expression level of miR-9
(Supporting Fig. S4A,B). Thus, these results suggest that
circMTO1 INHIBITS circMTO1 suppresses HCC progression possibly by
PROLIFERATION AND INVASION reducing the oncogenic effect of miR-9 by sponge activity.
OF HCC CELLS IN VITRO
circMTO1 SUPPRESSES HCC
The roles of circMTO1 in HCC progression were GROWTH BY SPONGE ACTIVITY OF
then investigated. To choose the HCC cell lines used
miR-9 AND UPREGULATION OF p21
for silencing or overexpression of circMTO1, we
checked the expression of circMTO1 and miR-9 in In order to investigate whether circMTO1 exerts its
eight HCC cell lines. SK-Hep1 cells showed the lowest antitumor effect through sponge activity of miR-9, we
1157
examined the expression of tumor suppressor p21, the In order to functionally confirm that circMTO1
target of miR-9.(27-30) Both silencing of circMTO1 or suppresses HCC progression by sponge activity of
transfection of miR-9 mimic significantly reduced miR-9 and consequently up-regulating p21, miR-9
mRNA and protein expression levels of p21 in HCC inhibitor was used to examine whether the tumor-
cells (Fig. 4E,F). In addition, overexpression of promoting effect of circMTO1 silencing could be
circMTO1 could increase the mRNA and protein blocked by miR-9 knockdown. miR-9 inhibitor could
expression levels of p21 in QGY-7701 and SK-Hep1 significantly reduce the down-regulation of p21
HCC cells (Fig. 4E,F). These results suggest that mRNA and protein expression after silencing of
circMTO1 may exert its antitumor effect through pro- circMTO1 (Fig. 5A-C). More important, miR-9
tecting p21 from down-regulation by miR-9. inhibitor significantly blocked the circMTO1
FIG. 3. circMTO1 binds miR-9 in HCC cells. (A) circMTO1 in HCC cell lysis was pulled down and enriched with circMTO1
specific probe and then detected by qPCR. (B) miR-9 was pulled down and enriched with a circMTO1-specific probe and then
detected by qPCR. (C) miR-9 colocalized with circMTO1 in HCC adjacent nontumor and tumor tissues were detected by FISH.
Data are from three independent experiments (mean 6 SEM; A,B) or are representative of three independent experiments with simi-
lar results (C). Magnification, 3400. *P < 0.05; **P < 0.01. Abbreviation: DAPI, 40 ,6-diamidino-2-phenylindole.
1158
FIG. 4. circMTO1 suppresses HCC cell proliferation, invasion, and enhances apoptosis by up-regulating p21. (A) miR-9 mimic was
overexpressed in HepG2 and SMMC-7721 cells. (B-F) circMTO1 was silenced, overexpressed, or miR-9 mimic was overexpressed in
HepG2, SMMC-7721, QGY-7701, and SK-Hep1 HCC cells. Proliferation of HepG2 and SMMC-7721 cells were measured by 5-
ethynyl-20 -deoxyuridine incorporation (B). The invasion of HepG2 and SMMC-7721 cells was measured by transwell assays (C). The
apoptosis ratios of HepG2, SMMC-7721, QGY-7701, and SK-Hep1 cells were measured by fluorescence-activated cell sorting (D). The
mRNA level of p21 was measured by qPCR (E). The protein expression level of p21 in HepG2, SMMC-7721, QGY-7701, and SK-
Hep1 cells was analyzed by western blotting (F). Data are from three independent experiments (mean 6 SEM; A-E) or are representative
of three independent experiments with similar results (F). *P < 0.05; **P < 0.01; ***P < 0.001. Abbreviation: Ctrl, control.
1159
FIG. 5. miR-9 inhibitor blocks the tumor-promoting effects of circMTO1 silencing by up-regulating p21. (A) The efficiency of
circMTO1 silencing and miR-9 inhibition in SMMC-7721 cells was measured by qPCR. (B-E) After silencing of circMTO1, miR-9
inhibitor was added to SMMC-7721 cell culture. p21 expression in mRNA level was measured by qPCR (B). The p21 protein
expression level was analyzed by western blotting (C). Proliferation ratio of HCC cells was measured by 5-ethynyl-20 -deoxyuridine
incorporation (D). Apoptosis ratio of HCC cells was measured by fluorescence-activated cell sorting. (E). Data are from three inde-
pendent experiments (mean 6 SEM; A,B,D,E) or are representative of three independent experiments with similar results (C). *P <
0.05; **P < 0.01; ***P < 0.001. Abbreviation: Ctrl, control.
1160
silencing-mediated promotion of proliferation and beginning. Hence, the deregulated circRNAs in HCC
inhibition of apoptosis, of HCC cells (Fig. 5D,E). So, pathology still need further investigation, and we will
miR-9 inhibitor could block the HCC-promoting keep focusing on this issue in our future studies.
effect of circMTO1 silencing. These data convincingly The decreased circMTO1 expression in HCC was
demonstrate that circMTO1 suppresses HCC progres- determined to be significantly correlated with progno-
sion as a sponge of miR-9 to eliminate the miR-9 sis of HCC patients, and by using kaps algorithm, we
oncogenic effect through circMTO1/miR-9/p21 axis. obtained an efficient cut-off value of circMTO1
(76.89) in distinguishing the outcomes of HCC
INTRATUMORAL KNOCKDOWN patients. Although the exact applicable cut-off value
still needs to be confirmed in a large set of patients, the
OF circMTO1 ENHANCES HCC
circMTO1 expression value of 76.89 is somewhat
GROWTH IN VIVO practical in the prognosis prediction of HCC patients.
We next investigated the significance of circMTO1 Currently, a set of intracellular gene expression, includ-
silencing in HCC growth in vivo. HCC-bearing nude ing coding genes and noncoding RNAs, have been
mice (SMMC-LTNM) were initially prepared by determined to be correlated with prognosis of HCC
transplanting human HCC tissues to form subcutane- patients.(12,14) However, a conclusive or practical crite-
ous tumor in nude mice and maintained with conti- rion for the identification of HCC prognosis still
nuously subcutaneous passage.(11) After continuous requires further investigation. Quantitative examina-
intratumoral injection of cholesterol-conjugated tion of the expression and combined detection of these
circMTO1 siRNAs for 2 weeks, we found that tumor genes associated with prognosis may have considerable
growth was significantly accelerated and serum AFP potential in the accurate identification and prediction
was markedly increased in HCC-bearing SMMC- of prognosis in HCC patients. Moreover, because
LTNM mice (Fig. 6A,B). The expression of some of the noncoding RNAs, especially miRNAs, are
circMTO1 and p21 was markedly reduced in tumor stable in plasma, they can be the biomarkers for HCC
tissues of SMMC-LTNM mice 48 hours after intratu- diagnosis and treatment,(12,14,31-33) and some circular
moral injection of cholesterol-conjugated circMTO1 RNAs are enriched in exosomes.(34) Whether
siRNAs (Fig. 6C). Moreover, p21 and downstream circMTO1 can also be detected in plasma or even in
CDK2 expressions were inhibited by in vivo circulating exosomes and whether the circulating
circMTO1 is also associated with HCC development
circMTO1 knockdown, whereas cell invasion and pro-
are interesting and important questions for circRNA
liferation markers MMP2 and PCNA were up-
study, which require further investigation.
regulated (Fig. 6D and Supporting Fig. S4C). Thus,
The mechanisms for circRNA function in carcino-
these results demonstrate that circMTO1 can suppress
genesis and cancer progression have not been eluci-
HCC progression in vivo.
dated very clearly. circRNAs may regulate the
expression of oncogenic or tumor suppressive genes
Discussion through different targets, depending on the different
types of cancer or even at different stages.(35) The most
The roles of circRNAs in carcinogenesis and cancer reported function pattern for circRNAs is acting as a
progression have attracted much attention currently. miRNA sponge and forming the circRNA-miRNA-
Because their expression and function in HCC devel- mRNA axis.(3,36) For example, circRNA SRY was
opment are still largely elusive, we screened the differ- reported to be a tumor-related molecule in colorectal
entially expressed circRNAs between HCC and cancer and ovarian cancer by absorbing miR-138.(37)
matched nontumor liver tissues through chip micro- cir-ITCH functions as a sponge of miR-7, miR-17,
Array and focused on the role and underlying mecha- and miR-214 to suppress the wnt/b-catenin path-
nism of the decreased circMTO1 expression in HCC way,(8) and ciRS-7 participates in gastric cancer and
progression. Although the tumor-inhibitory effects of colorectal cancer by interacting with miR-7.(38-41)
circMTO1 in HCC were suggested in this study, we Together with miRNAs and their targetome, the
do not exclude the possibility that there may be other circRNA-miRNA-mRNA axis may function as an
important deregulated circRNAs, which also partici- extensive regulatory network in gene expression, and
pate in the development of HCC, because of the lim- their deregulation may cause disease progression
ited HCC tissue samples we screened at the including cancer development. This presumption may
1161
FIG. 6. Intratumoral silencing of circMTO1 promotes HCC growth in vivo. (A) Effect of intratumoral circMTO1 knockdown on
tumor growth in SMMC-LTNM mice (n 5 3). (B) Effects of intratumoral circMTO1 knockdown on AFP level of SMMC-LTNM
HCC-bearing nude mice. AFP was detected by ELISA before or 2 weeks after intratumoral injection of cholesterol-modified
circMTO1 siRNA, control siRNA, or PBS. (C) Expression of circMTO1 and p21 in SMMC-LTNM tumor tissue 48 hours after
intratumoral injection of cholesterol-modified circMTO1 siRNAs, control siRNAs, or PBS. (D) Immunostaining of p21, MMP2,
and PCNA expression in SMMC-LTNM tumor tissues after intratumoral circMTO1 knockdown at the fifth week. Data are from
three independent experiments (mean 6 SEM; A-C) or are representative of three independent experiments with similar results (D).
Magnification, 3200. *P < 0.05; **P < 0.01.
raise interesting future works to elucidate the regula- At present, miRNAs targeted by circRNAs are
tory network of noncoding RNAs and coding genes in mostly predicted and selected by computational algo-
cancer biology. rithms, which seems to be shortsighted. How to
1162
precisely figure out which miRNA truly binds to a par-

9) Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal
ticular circRNA is important for the investigation of A, et al. Global cancer statistics, 2012. CA Cancer J Clin 2015;
mechanisms of circRNA functions. We previously 65:87-108.
established miRIP as providing an effective way of 10) Yang L, Li N, Wang C, Yu Y, Yuan L, Zhang M, et al. Cyclin
L2, a novel RNA polymerase II-associated cyclin, is involved in
quickly finding miRNAs that can be interacted by the
pre-mRNA splicing and induces apoptosis of human hepatocellu-
interested RN.(11) Using the miRIP approach, the lar carcinoma cells. J Biol Chem 2004;279:11639-11648.
potential interaction between miR-92a and p21, and 11) Su X, Wang H, Ge W, Yang M, Hou J, Chen T, et al. An in
between miR-197 and signal transduction and activa- vivo method to identify microRNA targets not predicted by
computation algorithms: p21 targeting by miR-92a in cancer.
tor of transcription 3, were elucidated by our previous Cancer Res 2015;75:2875-2885.
study.(11,42) In this study, by performing RIP using a 12) Hou J, Wang P, Lin L, Liu X, Ma F, An H, et al. Micro-
circMTO1-specific probe, we found that miR-9 was RNA-146a feedback inhibits RIG-I-dependent type I IFN pro-
the circMTO1-binding miRNA and then functionally duction in macrophages by targeting TRAF6, IRAK1, and
IRAK2. J Immunol 2009;183:2150-2158.
proved that circMTO1 functioned as the sponge of the 13) Jeck WR, Sharpless NE. Detecting and characterizing circular
miR-9. Thus, the miRIP approach would be a practi- RNAs. Nat Biotechnol. 2014, 32:453-61.
cal and useful method to identify the true and definite 14) Hou J, Lin L, Zhou W, Wang Z, Ding G, Dong Q, et al.
Identification of miRNomes in human liver and hepatocellular
interaction between miRNAs and circRNAs, especially carcinoma reveals miR-199a/b-3p as therapeutic target for hepa-
considering that the circRNA-miRNA interaction tocellular carcinoma. Cancer Cell 2011;19:232-243.
network is relatively complicated. 15) Kalra J, Baker J. Using digital quantification of stained tissue
In summary, our findings reveal that circMTO1 microarrays as a medium-throughput, quantitative method for
measuring the kinetics of signal transduction. Methods Mol Biol
expression is significantly decreased in HCC and cor- 2017;1554:107-125.
related with poor prognosis of HCC patients. Func- 16) Chabot-Richards D, Martin D, Myers OB, Czuchlewski DR,
tionally and mechanistically, circMTO1 inhibits HCC Hunt KE. Quantitative image analysis in the assessment of dif-
fuse large B-cell lymphoma. Mod Pathol 2011;24: 1598-1605.
growth by the sponge activity on miR-9 and up-
17) Chen T, Yang M, Yu Z, Tang S, Wang C, Zhu X, et al. Small
regulation of p21 expression, indicating its tumor sup- GTPase RBJ mediates nuclear entrapment of MEK1/MEK2 in
pressive role in HCC development. Furthermore, the tumor progression. Cancer Cell 2014;25:682-696.
in vivo intervention of circMTO1 indicates its poten- 18) Wang H, Zhang W, Zuo Y, Ding M, Ke C, Yan R, et al. miR-
9 promotes cell proliferation and inhibits apoptosis by targeting
tial in HCC-targeted therapy. Our data suggest that LASS2 in bladder cancer. Tumour Biol 2015;36:9631-9640.
circMTO1 may have considerable potential as a prog- 19) Palumbo A, Avet-Loiseau H, Oliva S, Lokhorst HM,
nosis predictor and therapeutic target for HCC. Goldschmidt H, Rosinol L, et al. Revised international staging
system for multiple myeloma: a report from IMWG. J Clin
Oncol 2015;33:2863-2869.
REFERENCES 20) Kang HJ, Eo SH, Kim SC, Park KM, Lee YJ, Lee SK, et al.
Increased number of metastatic lymph nodes in adenocarcinoma
1) Chen LL, Yang L. Regulation of circRNA biogenesis. RNA of the ampulla of Vater as a prognostic factor: a proposal of new
Biol 2015;12:381-388. nodal classification. Surgery 2014;155:74-84.
2) Hentze MW, Preiss T. Circular RNAs: splicing’s enigma varia- 21) Wang C, Huang XF, Cai QQ, Cao XX, Duan MH, Cai H,
tions. EMBO J 2013;32:923-925. et al. Prognostic study for overall survival in patients with newly
3) Memczak S, Jens M, Elefsinioti A, Torti F, Krueger J, Rybak A, diagnosed POEMS syndrome. Leukemia 2017;1:100-106.
et al. Circular RNAs are a large class of animal RNAs with regu- 22) Ma L, Young J, Prabhala H, Pan E, Mestdagh P, Muth D,
latory potency. Nature 2013;495:1439-1449. et al. miR-9, a MYC/MYCN-activated microRNA, regulates E-
4) Jeck WR, Sorrentino JA, Wang K, Slevin MK, Burd CE, Liu J, cadherin and cancer metastasis. Nat Cell Boil 2010;12:247-256.
et al. Circular RNAs are abundant, conserved, and associated 23) Iorio MV, Ferracin M, Liu CG, Veronese A, Spizzo R,
with ALU repeats. RNA 2012;19:141-157. Sabbioni S, et al. MicroRNA gene expression deregulation in
5) Lukiw WJ. Circular RNA (circRNA) in Alzheimer’s disease
human breast Cancer. Cancer Res 2005; 65:7065-7070.
(AD). Front Genet 2013;4:307.
24) Chen P, Price C, Li Z, Li Y, Cao D, Wiley A, et al. miR-9 is
6) Li P, Chen S, Chen H, Mo X, Li T, Shao Y, et al. Using circu-
an essential oncogenic microRNA specifically overexpressed in
lar RNA as a novel type of biomarker in the screening of gastric
mixed lineage leukemia–rearranged leukemia. Proc Natl Acad Sci
cancer. Clin Chim Acta 2015;444:132-136.
USA 2013;110:11511-11516.
7) Huang G, Zhu H, Shi Y, Wu W, Cai H, Chen X, et al. cir-
25) Drakaki A, Hatziapostolou M, Polytarchou C, Vorvis C,
ITCH plays an inhibitory role in colorectal cancer by regulating
Poultsides GA, Souglakos J, et al. Functional microRNA high
the Wnt/beta-catenin pathway. PLoS One 2015; 10:e0131225.
8) Li F, Zhang L, Li W, Deng J, Zheng J, An M, et al. Circular throughput screening reveals miR-9 as a central regulator of liver
RNA ITCH has inhibitory effect on ESCC by suppressing the oncogenesis by affecting the PPARA-CDH1 pathway. BMC
Wnt/beta-catenin pathway. Oncotarget 2015;6:6001-6013. Cancer 2015;15:542.
1163
26) Zhen Y, Xinghui Z, Chao W, Yi Z, Jinwen C, Ruifang G. Sev- ternary complexes with p21 and CDK2. Nucleic Acids Res
eral microRNAs could predict survival in patients with hepatitis 2016;44:2846-2858.
B-related liver cancer. Sci Rep 2017;7:45195. 36) Thomas B, Hansen TI, Bettina H, Bramsen JB, Finsen B,
27) Rotkrua P, Akiyama Y, Hashimoto Y, Otsubo T, Yuasa Y. Damgaard CK, et al. Natural RNA circles function as efficient
MiR-9 downregulates CDX2 expression in gastric cancer cells. microRNA sponges. Nature 2013;495:384-388.
Int J Cancer 2011;129:2611-2620. 37) Long L, Huang G, Zhu H, Guo Y, Liu Y, Huo J. Down-regu-
28) Beuvink I, Boulay A, Fumagalli S, Zilbermann F, Ruetz S, lation of miR-138 promotes colorectal cancer metastasis via
O’Reilly T, et al. The mTOR inhibitor RAD001 sensitizes directly targeting TWIST2. J Transl Med 2013;11:916-926
tumor cells to DNA-damaged induced apoptosis through inhibi- 38) Ma C, Qi Y, Shao L, Liu M, Li X, Tang H. Downregulation
tion of p21 translation. Cell 2005;120:747-759. of miR-7 upregulates Cullin 5 (CUL5) to facilitate G1/S transi-
29) Zhang X, Ma W, Cui J, Yao H, Zhou H, Ge Y, et al. Regula- tion in human hepatocellular carcinoma cells. IUBMB Life 2013;
tion of p21 by TWIST2 contributes to its tumor-suppressor 65:1026-1034.
function in human acute myeloid leukemia. Oncogene 2015;34: 39) Zhang X, Hu S, Zhang X, Wang L, Zhang X, Yan B, et al.
3000-3010. MicroRNA-7 arrests cell cycle in G1 phase by directly targeting
30) Zhang Y, Liu Y, Duan J, Yan H, Zhang J, Zhang H, CCNE1 in human hepatocellular carcinoma cells. Biochem Bio-
et al. Hippocalcin-like 1 suppresses hepatocellular carcinoma phys Res Commun 2014;443:1078-1084.
progression by promoting p21(Waf/Cip1) stabilization by acti- 40) Zhang N, Li X, Wu CW, Dong Y, Cai M, Mok MT, et al.
vating the ERK1/2-MAPK pathway. HEPATOLOGY 2016;63: microRNA-7 is a novel inhibitor of YY1 contributing to colorec-
880-897. tal tumorigenesis. Oncogene 2013;32:5078-5088.
31) Guo W, Qiu Z, Wang Z, Wang Q, Tan N, Chen T, et al. 41) Zhao X, Dou W, He L, Liang S, Tie J, Liu C, et al. Micro-
MiR-199a-5p is negatively associated with malignancies and reg- RNA-7 functions as an anti-metastatic microRNA in gastric
ulates glycolysis and lactate production by targeting hexokinase 2 cancer by targeting insulin-like growth factor-1 receptor. Onco-
in liver cancer. HEPATOLOGY 2015;62:1132-1144. gene 2013;32:1363-1372.
32) Yuan SX, Wang J, Yang F, Tao QF, Zhang J, Wang LL, et al. 42) Wang H, Su X, Yang M, Chen T, Hou J, Li N, et al. Recipro-
Long noncoding RNA DANCR increases stemness features of hepa- cal control of miR-197 and IL-6/STAT3 pathway reveals miR-
tocellular carcinoma by derepression of CTNNB1. HEPATOLOGY 197 as potential therapeutic target for hepatocellular carcinoma.
2016;63:499-511. Oncoimmunology 2015;4:e1031440.
33) Xing Z, Lin A, Li C, Liang K, Wang S, Liu Y, et al. lncRNA
directs cooperative epigenetic regulation downstream of chemo-
Author names in bold designate shared co-first
kine signals. Cell 2014;159:1110-1125.
34) Li Y, Zheng Q, Bao C, Li S, Guo W, Zhao J, et al. Circular authorship.
RNA is enriched and stable in exosomes: a promising biomarker
for cancer diagnosis. Cell Res 2015;25:981-984.
35) Du WW, Yang W, Liu E, Yang Z, Dhaliwal P, Yang BB, et al.
Foxo3 circular RNA retards cell cycle progression via forming
Supporting Information
Additional Supporting Information may be found at
onlinelibrary.wiley.com/doi/10.1002/hep.29270/suppinfo.
1164

Hepatology - 2017 - Han - Circular RNA circMTO1 Acts As The Sponge of microRNA 9 To Suppress Hepatocellular Carcinoma

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hepatology - 2017 - Han - Circular RNA circMTO1 Acts As The Sponge of microRNA 9 To Suppress Hepatocellular Carcinoma

Uploaded by

Copyright:

Available Formats

A MERICAN A SSOCIATION FOR

T HE STUDY OF LIVER D I S E ASES

HEPATOLOGY, VOL. 66, NO. 4, 2017

Circular RNA circMTO1 Acts

View this article online at wileyonlinelibrary.com.

ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO:

precisely ﬁgure out which miRNA truly binds to a par-

You might also like