Zebularine

DNMT1-mediated methylation of BEX1 regulates
stemness and tumorigenicity in liver cancer
Graphical abstract
Non-CSC HCC
Wnt pathway activation
liver CSC self-renewal
BEX1 expression level
DNMT1-mediated methylation
Highlights
 BEX1 is an oncofetal protein and a novel stem cell marker for
hepatoblastoma and hepatocellular carcinoma.
 BEX1 levels are regulated by expression of DNA
methyltransferase 1.
 BEX1 inhibition can sensitize liver cancer cells to sorafenib
treatment.
 BEX1 enhances stemness through beta-catenin signaling.
 BEX1 upregulates beta-catenin by blocking RUNX3, a tran￾scription factor that downregulates beta-catenin levels.
Authors
Qian Wang, Ning Liang, Tao Yang, .,
Zhangqian Chen, Xianli He, Chaoxu Liu
Correspondence
[email protected] (G. Wang),
[email protected] (Z. Chen),
[email protected] (X. He),
[email protected] (C. Liu).
Lay summary
Cancer stem cells (CSCs) contribute to a
high rate of cancer recurrence, as well
as resistance to conventional therapies.
However, the regulatory mechanisms
underlying their self-renewal remains
elusive. Herein, we have reported that
BEX1 plays a key role in regulating CSC
properties in different types of liver
cancer. Targeting BEX1-mediated Wnt/
b-catenin signaling may help to address
the high rate of recurrence, and het￾erogeneity of liver cancer.

https://doi.org/10.1016/j.jhep.2021.06.025

© 2021 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. J. Hepatol. 2021, -, 1–12
Research Article
Hepatic and Biliary Cancer
DNMT1-mediated methylation of BEX1 regulates stemness and
tumorigenicity in liver cancer
Qian Wang1,†
, Ning Liang2,†
, Tao Yang3,†
, Yuedan Li4,†
, Jing Li5
, Qian Huang6
, Chen Wu7
Ligang Sun2
, Xile Zhou8
, Xiaobin Cheng8
, Long Zhao8
, Gang Wang9,
*, Zhangqian Chen10,
Xianli He11,
*, Chaoxu Liu8,
Department of Anorectal Surgery, The First Affiliated Hospital, Zhengzhou University, Zhengzhou 450052, China; 2
Department of General
Surgery, The 75th Group Army Hospital, Dali 671000, China; 3
Department of Pain Treatment, Tangdu Hospital, Air Force Military Medical
University, Xi’an 710032, China; 4
Department of Medicinal Chemistry and Pharmaceutical Analysis, School of Pharmacy, Air Force Military
Medical University, Xi’an, Shaanxi 710032, China; 5
College and Hospital of Stomatology, Xi’an Jiao Tong University, Xi’an 710000, China; 6
Department of Obstetrics and Gynecology, The 75th Group Army Hospital, Dali 671000, China; 7
Department of General Surgery, Huashan
Hospital, Fudan University, Shanghai 201907, China; 8
Department of Colorectal Surgery, The First Affiliated Hospital of Zhejiang University,
Hangzhou 310006, China; 9
Department of General Surgery, The 74th Group Army Hospital, Guangzhou 510318, China; 10State Key Laboratory
of Cancer Biology, National Clinical Research Center for Digestive Diseases and Xijing Hospital of Digestive Diseases, Air Force Military Medical
University, Xi’an 710032, Shaanxi, China; 11Department of General Surgery, Tangdu Hospital, Air Force Military Medical University, Xi’an
710032, Shaanxi, China
Background & Aims: Hepatoblastoma (HB) and hepatocellular
carcinoma (HCC) both exhibit notable cancer stem cell (CSC)
features. Moreover, the development of both diseases is closely
associated with the presence of CSCs. We investigated the role of
brain-expressed X-linked protein 1 (BEX1) in regulating the CSC
properties of HB and a subtype of HCC with high CSC features
(CSC-HCC).
Methods: Stemness scores were analyzed in 5 murine HCC
models. A subpopulation of BEX1-positive cells and BEX1-
negative cells were sorted from HCC cell lines, and subjected to
transcriptome analysis. The expression and function of BEX1 was
examined via western blotting, sphere formation assays, and
xenograft tumor models.
Results: We identified BEX1 as a novel CSC marker that was
required for the self-renewal of liver CSCs. Furthermore, zebu￾larine, a potent DNMT1 inhibitor, can induce the reactivation of
BEX1 by removing epigenetic inhibition. Notably, BEX1 was
highly expressed in patients with HB and CSC-HCC, but not in
patients with non-CSC HCC. Moreover, DNMT1-mediated
methylation of the BEX1 promoter resulted in differential BEX1
expression patterns in patients with HB, CSC-HCC, and non-CSC￾HCC. Mechanistically, BEX1 interacted with RUNX3 to block its
inhibition of b-catenin transcription, which led to the activation
of Wnt/b-catenin signaling, and stemness maintenance in both
HB and CSC-HCC. In contrast, downregulated BEX1 expression
released RUNX3 and inhibited the activation of Wnt/b-catenin
signaling in non-CSC-HCC.
Conclusion: BEX1, under the regulation of DNMT1, is necessary
for the self-renewal and maintenance of liver CSCs through
activation of Wnt/b-catenin signaling, rendering BEX1 a poten￾tially valuable therapeutic target in both HB and CSC-HCC.
Lay summary: Cancer stem cells (CSCs) contribute to a high rate
of cancer recurrence, as well as resistance to conventional ther￾apies. However, the regulatory mechanisms underlying their
self-renewal remains elusive. Herein, we have reported that
BEX1 plays a key role in regulating CSC properties in different
types of liver cancer. Targeting BEX1-mediated Wnt/b-catenin
signaling may help to address the high rate of recurrence, and
heterogeneity of liver cancer.
© 2021 European Association for the Study of the Liver. Published by
Elsevier B.V. All rights reserved.
Introduction
The theory of cancer stem cells (CSCs) suggests that cells are
hierarchically organized within tumors, and that a subpopula￾tion of stem-like cells, which have self-renewal and differentia￾tion characteristics, are responsible for cancer progression,
recurrence, and metastasis.1 Several biomarkers have been
identified for their applicability in the differentiation of liver
CSCs, including CD133, Sal-like protein 4 (SALL4), and epithelial
cell adhesion molecule (EpCAM). In liver CSCs, one or more
highly conserved stemness signaling pathways, such as Wnt/b￾catenin, Notch, and Hedgehog, are excessively activated. Addi￾tionally, transcription factors such as NANOG, OCT4, SOX2, and c￾MYC also play critical roles in the self-renewal of CSCs.2 c-MYC is
a transcription factor that is aberrantly expressed in liver cancer,
including hepatoblastoma (HB) and hepatocellular carcinoma
(HCC).3–5 HB is the most common type of liver cancer reported in
infants under 3 years of age. In adults, HCC is the most common
Keywords: Brain-expressed X-linked protein 1; SALL4; Wnt/b-catenin signaling;
cancer stem cell-like properties.
Received 9 December 2020; received in revised form 4 May 2021; accepted 14 June 2021;
available online xxx
* Corresponding authors. Addresses: Chaoxu Liu, Department of General Surgery, The
First Affiliated Hospital of Zhejiang University, Hangzhou, China; Tel: +86-21-685079;
Fax: +86-21-6405875; (C. Liu), or Department of General Surgery, Tangdu Hospital,
Air Force Military Medical University, Xi’an, 710032, China, (X. He), or State key
Laboratory of Cancer Biology, National Clinical Research Center for Digestive Diseases
and Xijing Hospital of Digestive Diseases, Air Force Military Medical University, Xi’an,
710032, Shaanxi, China, (Z. Chen), or Department of General Surgery, the 74th group
army hospital, Guangzhou, 510318, China. (G. Wang).
E-mail addresses: [email protected] (G. Wang), [email protected]
(Z. Chen), [email protected] (X. He), [email protected] (C. Liu). † These authors contributed equally to this work.

https://doi.org/10.1016/j.jhep.2021.06.025

Journal of Hepatology 2021 vol. – j 1–12
Research Article
Hepatic and Biliary Cancer
form of liver cancer documented. Although there are marked
differences between HB and HCC in terms of etiology, histolog￾ical phenotype, and genetic background, the development of
both diseases is closely related to the presence of CSCs.
Brain-expressed X-linked protein 1 (BEX1), an intracellular
signal transducer or regulator, was first characterized by its
decreased expression in F9 teratoma cells following treatment
with retinoic acid.6 Vilar et al. have reported that BEX1 is
involved in the differentiation of neural stem cells.7 Addition￾ally, BEX1 expression is closely related to the repair of skeletal
muscle after injury, suggesting that it may exert an effect on
adult stem cells.8 Nevertheless, whether BEX1 regulates the
biological behavior of liver CSCs and hepatocarcinogenesis re￾mains unclear.
Available evidence suggests that DNA methylation is a key
epigenetic reprogramming mechanism that plays a vital role in
stem cell biology.9 For example, DNA methyltransferase in￾hibitors can significantly improve the overall efficiency of the
reprogramming process. Moreover, activation of pluripotency￾associated genes by DNA methyltransferase inhibitors in neuro￾sphere cells has been reported.10 DNA methylation is performed
by the following 3 main types of DNA methyltransferases
(DNMTs): DNMT1, DNMT3A, and DNMT3B. DNMT3A and
DNMT3B play a key role in establishing de novo DNA methyl￾ation, whereas DNMT1 maintains DNA methylation through cell
division.11 Considering that BEX1 is subjected to epigenetic
silencing in a variety of tumors,12 we sought to examine whether
BEX1 was also under epigenetic regulation in HCC. In this study,
we identified BEX1 as an oncofetal protein, and a novel stem cell
marker for HB and HCC. Importantly, our results suggest that
BEX1 plays a key role in regulating CSC properties in HB, CSC￾HCC, and non-CSC HCC. Therefore, BEX1 may serve as a
biomarker for diagnosis and as a potential drug target in HB and
CSC-HCC.
Materials and methods
Antibodies and reagents
PT3-EF1a-c-Myc, pT3-myr-AKT-HA, pT3-N90-beta-catenin, and
pT3-EF1a-c-Met were kindly gifted by Xin Chen,5 and were ob￾tained from Addgene (Cambridge, MA, USA). pCMV/SB10 was a
gift from Perry Hackett and was obtained from Addgene.13 In￾formation on plasmids, antibodies, and primer sequences are
provided in the CTAT methods section.
Mice and HCC induction
Standard Sleeping Beauty–mediated hydrodynamic injections
were performed as per methods previously described.14 For
establishment of the patient-derived xenograft model, primary
HCC samples were harvested for xenograft establishment as per
previously reported protocols.15 In accordance with NIH guide￾lines, all animal experiments were approved by the Institutional
Animal Care and Use Committee at Zhejiang University.
Statistical analysis
SPSS 17.0 software (IBM, Armonk, NY) was used for statistical
analyses, and p <0.05 was considered statistically significant. An
unpaired Student’s t test was performed to examine differences
between groups.
Results
BEX1 is a human oncofetal protein and stem-cell marker in
liver cancer
In this study, we aimed to screen for potential targets that may
be associated with CSC properties in HCC, using murine HCC
models with high CSC properties. To this end, we first estab￾lished various murine HCC models, and subsequently compared
the stemness scores of each model. We then induced the
development of liver tumors via hydrodynamic injection with
various oncogenic plasmids. Additionally, diethyl nitrosamine￾induced and obesity-associated murine HCC models were also
used in this study (Fig. S1A). Immunohistochemistry (IHC)
revealed that CSC markers exhibited remarkable positivity in c￾MYC-induced HCC (Fig. S1B). Additionally, hepatoma cells
collected from fresh samples harvested from the 5 murine HCC
models were cultured in vitro to determine their sphere￾forming ability. Spheroid formation assay results showed that
c-MYC-induced HCC possessed remarkable CSC properties
(Fig. S1C). Moreover, the critical pathways that regulate CSC
properties are significantly activated in c-MYC-induced HCC
(Fig. S1D-G). Therefore, we used the c-MYC mouse model to
explore potential targets that might be associated with CSC
properties in HCC (Fig. 1A). Transcriptomic analysis of c-MYC￾driven HCC showed that expression levels of 2,587 genes were
upregulated, and those of 2,733 genes were downregulated
(Fig. S1D). Among the top 3 upregulated genes associated with
stemness, we focused on BEX1 expression (Gene ID: 55859),
whose depletion significantly inhibited oncosphere formation
of Huh7 cells (Fig. 1B). IHC staining and immunoblotting results
showed that BEX1 and SALL4 protein levels were significantly
higher in c-MYC-driven HCC tissues than those in peritumor
tissues (Fig. 1C-D). We also detected the expression levels of
BEX1 in other murine HCC models. As shown in Fig. S2, the
expression of BEX1 in c-MYC-induced murine HCC was signif￾icantly higher than that in other murine HCCs. To further
explore the physiological and pathological role of BEX1 in the
liver, we examined BEX1 protein levels in human fetal liver,
adult liver, and human HCC tissues. As shown in Fig. 1E and
Fig. S3, BEX1 protein levels were relatively high in fetal liver
tissues, but were weakly positive in the adult counterparts,
suggesting that BEX1 was an oncofetal protein. Since HB is an
embryonal tumor, we examined the expression level of BEX1 in
86 patients with HB (cohort 1). BEX1 expression was signifi-
cantly upregulated in HB (Fig. 1F). HB possesses notable CSC
properties (Fig. S4). Using IHC, we analyzed the relationship
between BEX1 expression and CSC markers of HCC in a cohort
of 186 patients (cohort 2). The results showed that the rate of
BEX1-positive cases was 85.3% in SALL4-positive cases, 90.5% in
CD133-positive cases, 88.4% in EpCAM-positive cases, and 68.4%
in SOX4-positive cases. This indicated that BEX1 expression was
markedly associated with CSC markers of HCC (Fig. 1G). More￾over, a positive correlation between BEX1 and CSC markers was
also confirmed in The Cancer Genome Atlas HCC cohort (n =
371) (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi) (Fig. 1H). To
determine the expression of BEX1 in Huh7 CSC subsets, CD133+
and CD133- CSCs were sorted from cells. Indeed, BEX1 protein
expression was significantly higher in CD133+ CSCs than that
observed in CD133- cells (Fig. S5A). Moreover, upregulated
protein levels of BEX1 and CSC-related markers were found in
spheroids when compared with the adherent cells, suggesting
2 Journal of Hepatology 2021 vol. - j 1–12
Research Article Hepatic and Biliary Cancer
that BEX1 might play a key role in the maintenance of CSC
properties (Fig. S5B). To understand the behavior of tumor cells
expressing BEX1, we isolated BEX1-positive tumor cells
(BEX1Pos) using a promoter-reporter strategy.16 A lentiviral
vector containing the human BEX1 promoter was used to drive
the expression of the green fluorescent protein (GFP) reporter
gene (LvPBex1-GFP) (Fig. 1I). Immunofluorescence assay results
showed that SALL4 (red), a known CSC biomarker of liver can￾cer, was highly expressed in BEX1Pos cells (green). This indi￾cated that BEX1Pos cells might represent liver CSCs (Fig. 1J).
Moreover, we observed that SALL4 was also highly expressed in
oncosphere cells derived from BEX1Pos cells isolated from
patient-derived primary HCC cells (Fig. 1K). Taken together,
these results suggest that BEX1 can be used as a novel CSC
marker in HCC.
BEX1 promotes tumorigenesis and stemness in liver cancer
We generated BEX1 knockout mice (referred to as Bex1-/-) using
the CRISPR/Cas9 method (Fig. 2A and Fig. S6A-G). To investigate
whether BEX1 expression was involved in hepatocarcinogenesis,
Bex1-/- mice and wild-type (WT) littermates were subjected to
hydrodynamic transfection with oncogenic plasmids, to establish
Fig. 1. BEX1 is a human oncofetal protein and stem cell marker in liver cancer. (A) RNA sequencing analysis of CSC-related markers in c-MYC-induced tumor
and peri-tumor tissues (n = 3 per group). (B) The top 3 CSC-related genes were subjected to knockout in Huh7 HCC cells, and their sphere formation ability was
tested via in vitro assays. (C, D) Representative images of western blotting and IHC analyses showing upregulation of BEX1 and SALL4 in c-MYC-driven tumor
lesions compared with peri-tumor tissues (n = 3 per group). Upper scale bars: 50 lm; lower scale bars: 20 lm. (E) Representative H&E images (upper panel) and
IHC staining results of BEX1 (lower panel) in human fetal livers and adult livers (n = 20 per group). Scale bars: 50 lm. (F) Representative H&E images (upper
panel) and IHC staining of BEX1 (lower panel) in 86 human HB and peri-tumor tissues. Scale bars: 50 lm. (G) The relationship between BEX1 expression and CSC
markers in 186 patients with HCC was evaluated by IHC analysis. Scale bars: 50 lm. (H) The relationship between BEX1 and CSC markers was further evaluated in
TCGA HCC cohort (n = 371). (I) Schematic of lentiviral vector LvPBex1-GFP, in which GFP expression was controlled by the human BEX1 promoter. (J) Immuno-
fluorescence analysis of SALL4 expression in FACS-purified BEX1+ cells from Huh7 and Hep3B cell lines. Scale bars: 20 lm. (K) Immunofluorescence analysis of
SALL4 expression in oncospheres derived from BEX1+ tumor cells isolated from patient-derived primary HCC cells. Scale bars: 20 lm. *p <0.05, ***p <0.001 by
using 2-tailed Student’s t test. CSC, cancer stem cell; HB, hepatoblastoma; HCC, hepatocellular carcinoma; HE: hematoxylin-eosin staining; IHC, immunohisto￾chemistry; P, peri-tumor tissues; T, c-MYC-induced tumor tissues; TCGA, The Cancer Genome Atlas.
Journal of Hepatology 2021 vol. - j 1–12 3
an HCC model (Fig. 2B). As shown in Fig. 2C, tumor onset was
reduced in Bex1-/- mice compared to WT mice. Furthermore, the
survival time was prolonged in Bex1-/- mice compared to that in
WT mice (Fig. 2D). Additionally, the liver weight, tumor volume,
serum aspartate aminotransferase, and alanine aminotransferase
levels were significantly reduced in Bex1-/- mice compared with
WT mice at 7 weeks (Fig. 2E). We further explored the impact of
BEX1 deletion on critical pathways that regulate CSC properties,
such as Hedgehog, Notch, and Wnt/b-catenin signaling. The re￾sults showed that Gli1, Hes1, NICD, and b-catenin protein
expression was significantly downregulated in Bex1-/- liver tissue
compared to that in WT liver tissue, suggesting that BEX1
expression was necessary for the activation of the pathways that
regulate CSC properties (Fig. 2F). To investigate whether the
delayed hepatocarcinogenesis induced by the loss of BEX1
expression was also observed in other murine HCC models (e.g.
AKT-driven, AKT/c-MET-driven, and CAT/c-MET-driven HCC),
hydrodynamic injections were performed in WT or Bex1-/- mice.
We observed that BEX1 deletion significantly delayed the
development of HCC in all murine HCC models (Fig. S6H-P), thus
implying a key function of BEX1 in hepatocarcinogenesis in
multiple settings. Additionally, liver-specific transfection of
BEX1-overexpression vectors using hydrodynamic injection was
performed to clarify the role of BEX1 in HCC development. A c￾A 1
Fig. 2. BEX1 promotes tumorigenesis and stemness in liver cancer. (A) Transgenic mice were identified by PCR screening. Homozygous: 1 band at 682 bp (#1,
4); heterozygous: 2 bands at 682 bp and 878 bp (#6, 7, 8); wild-type: 1 band at 878 bp (#2, 3, 5, 9). (B) Study design (n = 6 per group). (C) Tumor formation in WT
and Bex1-/- mice at different time points following oncogenic plasmid injection. Scale bars: 50 lm. (D) Survival curve. (E) Liver weight, tumor volume, and serum
AST and ALT levels in Bex1-/- and WT groups were measured at 7 weeks. (F) Immunoblotting analysis of the indicated proteins in WT and Bex1-/- liver tissues. (G)
Spheroid formation assays were performed in Huh7 cells transfected with Lv-BEX1, and in Hep3B cells transfected with Lv-shBEX1. Scale bars: 20 lm. (H) Flow
cytometric analysis of the EpCAM+ cell population in treatment groups, as indicated. (I) The invasion of hepatoma cells was examined by performing an invasion
chamber assay. Scale bars: 20 lm. (J) At 7 weeks, the mRNA expression of CSC-associated markers in tumors isolated from Bex1-/- and WT mice were examined by
reverse transcription PCR. (K) Limiting dilution xenograft formation of Huh7 cells infected with Lv-BEX1 or Lv-Control (n = 6 per group). *p <0.05, **p <0.01, ***p
<0.001 by using 2-tailed Student’s t test. ALT, alanine aminotransferase; AST, aspartate aminotransferase; CSC, cancer stem cell; Ctrl, control; FITC, fluorescein
isothiocyanate; Lv, lentiviral; M, marker; OE, overexpression; WT, wild-type.
4 Journal of Hepatology 2021 vol. - j 1–12
Research Article Hepatic and Biliary Cancer
MYC-induced HCC model was used in this study. A liver-specific
adeno-associated viral vector (AAV8-TBGp-BEX1) or a control
vector (AAV8-TBGp-control) were administered to c-MYC mice.
Notably, mice given AAV8-TBGp-BEX1 showed aggravated liver
tumor burden, indicating that BEX1 was a key regulator in the
development of HCC (Fig. S7).
Previous studies have shown that BEX1 is involved in the
functional regulation of neural stem cells.17,18 However, whether
BEX1 regulates CSC properties in HCC remains unclear. Spheroid
formation assays revealed the existence of an increased number
of spheroids in cells overexpressing BEX1, and a reduced number
of spheroids in BEX1 knockdown cells (Fig. 2G). Flow cytometric
analysis also revealed an increased proportion of EpCAM+ cells in
cells overexpressing BEX1, and a reduced proportion of EpCAM+
cells in BEX1 knockdown cells (Fig. 2H). BEX1 overexpression
enhanced hepatoma cell invasion, whereas BEX1 knockdown
inhibited the invasion of hepatoma cells (Fig. 2I). Importantly,
the expression of CSC-related markers was significantly
decreased in Bex1-/- mice compared with that observed in WT
mice at 7 weeks (Fig. 2J). An in vivo limiting dilution assay
revealed that hepatoma cells expressing BEX1 exhibited
enhanced tumor initiation capacity in BALB/c nude mice
compared with the capacity exhibited by control cells (Fig. 2K).
Additionally, overexpression of BEX1 in normal human liver HL-
7702 cells was achieved. The introduction of BEX1 caused he￾patocyte transformation in vitro and tumor initiation in vivo,
while the control cells did not present with such occurrences
(Fig. S8A-C). Moreover, intense alpha-fetoprotein (AFP) staining
was demonstrated by the tumor (Fig. S8D). To validate the
function of BEX1 expression in regulating the self-renewal of
CSCs in HCC, we inhibited the expression of BEX1 in BEX+ tumor
cells and investigated their ability to self renew. Spheroid for￾mation assays revealed that the self-renewal of BEX1+ cells was
inhibited by transfection with sh-BEX1 (Fig. S8E-G). Further￾more, restoration of the expression of BEX1 in BEX1- tumor cells
was achieved, and the results showed an enhanced self-renewal
capacity in these cells following infection with Lv-BEX1
(Fig. S8H-J). We then compared the tumorigenicity of BEX1-
cells with and without BEX1 overexpression in vivo. When BEX1
expression was restored, the tumorigenic ability of BEX1- cells
was found to be similar to that exhibited by the BEX1+ cell
control (Fig. S8K-L). Taken together, these results suggest that
BEX1 promotes tumorigenesis and stemness in liver cancer.
BEX1 expression upregulation attributed to DNMT1
inhibition may facilitate functional stemness enhancement
Accumulating evidence suggests that DNA methylation plays a
vital role in maintaining the stem cells’ ability to self-renew.19
Since BEX1 is subjected to epigenetic silencing in a variety of
tumors,20 we sought to examine whether BEX1 levels were
regulated by the expression of DNMT1 in HCC. According to the
criteria and algorithm described by Dahiya,21 the BEX1 promoter
in number
of spheres
Fig. 3. BEX1 expression upregulation attributed to DNMT1 inhibition may facilitate functional stemness enhancement. (A) MethHC database was used to
predict the CpG island in BEX1, which extends from -57 to -335 bp from the TSS. Each red tick mark represents 1 CpG site. Arrows indicate the TSS. (B) Alteration
of CpG island methylation in both Huh7 and Hep3B cell lines. Bisulfite sequencing evaluation of the 26 CpG sites in the BEX1 promoter was performed using 10
clones of each cell line. Black dot: methylated DNA; white dot: unmethylated DNA. (C) Huh7 and Hep3B cells were treated with zebularine for 3 days, and the
methylation of the BEX1 promoter was analyzed by conducting methylation-specific PCR. (D) Huh7 and Hep3B cells were subjected to treatment with zebularine
or with a vehicle control as indicated, and the BEX1 expression level was measured by reverse transcription PCR. (E) Immunoblotting of the indicated proteins in
Huh7 cells. (F, G) Flow cytometric analysis of the EpCAM+ Huh7 cell population in treatment groups, as indicated. (H, I) Sphere formation ability of Huh7 cells in
treatment groups, as indicated. Scale bars: 20 lm. *p <0.05, **p <0.01 by using 2-tailed Student’s t test. FITC, fluorescein isothiocyanate; M, methylated status; TSS,
transcription start site; U, unmethylated status.
Journal of Hepatology 2021 vol. - j 1–12 5
A Hepatoblastoma (BEX1-positive rate = 88.4%)
Correlation between BEX1 and DNMT1 expression
in hepatoblastoma tissues
BEX1
expression
Lv-DNMT1
Fig. 4. The differential expression pattern of BEX1 in hepatoblastoma, CSC-HCC, and non-CSC-HCC patients is through an epigenetic reprogramming
mechanism. Representative IHC images of BEX1 and DNMT1 expression in tumor tissues of (A) 86 patients with hepatoblastoma, (B) 88 patients with CSC-HCC
and (C) 98 with non-CSC HCC. Scale bars: 50 lm. (D) Immunoblotting of the indicated proteins in WRL68 hepatoblastoma cells (WRL68), CD133+ Huh7 stem cells
6 Journal of Hepatology 2021 vol. - j 1–12
Research Article Hepatic and Biliary Cancer
considered for conducting bisulfite sequencing analysis. The re￾sults showed that the BEX1 promoter was hypermethylated in
HCC cells (Fig. 3B). Additionally, demethylation treatment with
zebularine (ZEB), a potent DNMT1 inhibitor, remarkably restored
BEX1 expression in Huh7 cells in a dose-dependent manner,
reaching a plateau at a concentration of 100 lM (Fig. S9).
Therefore, this concentration (ZEB, 100 lM) was used in the
subsequent experiments. Methylation-specific PCR, reverse
transcription PCR, and immunoblotting experiments demon￾strated that ZEB could induce the reactivation of BEX1 by
removing epigenetic inhibition (Fig. 3C-E). We then investigated
whether BEX1 upregulation, induced by DNMT1 inhibition,
could enhance functional stemness. Flow cytometric analysis
revealed an increased proportion of EpCAM+ Huh7 cells
following ZEB treatment, and a reduced proportion of EpCAM+
Huh7 cells with BEX1 knockdown (Fig. 3F-G). Furthermore, Huh7
cells’ capacity for self-renewal was enhanced following ZEB
treatment, which was attenuated in Huh7 cells with BEX1
knockdown (Fig. 3H-I). These results suggest that DNMT1 in￾hibitors can induce the reactivation of BEX1 by removing
epigenetic inhibition. Enhanced BEX1 expression further pro￾moted self-renewal in HCC.
The differential expression pattern of BEX1 in HB, CSC-HCC,
and non-CSC HCC patients through an epigenetic
reprogramming mechanism
CD133 is recognized as a stem cell biomarker in liver cancer.22
Using a CD133 staining score, 186 patients with HCC from
cohort 2 were defined as negative, weakly positive, moderately
positive, and strongly positive. The non-CSC HCC group (cohort 2,
n = 98) demonstrated negative and weakly positive expression,
while the CSC-HCC group (cohort 2, n = 88) demonstrated
moderate and strongly positive expression. Additionally, 86 pa￾tients with HB from cohort 1 were included in the analysis. Using
IHC and immunoblotting analyses, we found that patients with
HB and CSC-HCC had a higher BEX1-positive rate than patients
with non-CSC HCC (Fig. 4A-C and Fig. S10A-C). Moreover, there
existed an inverse correlation between DNMT1 and BEX1 in HB,
CSC-HCC, and non-CSC HCC (Fig. 4A-C). These results suggested
that the expression pattern of BEX1 in HB, CSC-HCC, and non￾CSC HCC were different, which might be due to the different
expression levels of DNMT1. Next, we investigated the expres￾sion pattern of BEX1 in different types of liver cancer cells,
including WRL68 HB cells (WRL68), CD133+ Huh7 stem cells
(CD133+
), and HCC cell lines (Huh7 and Hep3B). In line with the
results obtained for the clinical samples, immunoblotting anal￾ysis revealed that BEX1 expression in WRL68 and CD133+ stem
cells was significantly higher than that in Huh7 and Hep3B cells.
However, WRL68 and CD133+ stem cells expressed low levels of
DNMT1 (Fig. 4D). Bisulfite sequencing analysis revealed that the
BEX1 promoter was hypomethylated in WRL68 and CD133+ stem
cells (Fig. 4E). To further explore the role of DNMT1 in the
regulation of the expression of BEX1, overexpression of DNMT1
in WRL68 and CD133+ stem cells was achieved. DNMT1 over￾expression promoted the hypermethylation of BEX1 (Fig. 4F), and
downregulated BEX1 mRNA and protein expression in both
WRL68 and CD133+ stem cells (Fig. 4G). Our study suggested that
increased DNMT1 expression might lead to the hyper￾methylation of BEX1 and reduced BEX1 expression in non-CSC
HCC. In contrast, decreased DNMT1 expression may lead to the
hypomethylation of BEX1, and increased BEX1 expression in HB
and CSC-HCC. Additionally, high BEX1 expression was associated
with a poor prognosis in patients with HB and HCC (Fig. S10D-G).
In addition to DNA methylation, we explored whether other
epigenetic modifications were involved in the regulation of BEX1
expression. Our preliminary results suggested that m6
A RNA
methylation and histone acetylation were not likely to play a role
in regulating BEX1 expression (Fig. S11 and 12). Moreover,
common risk factors (such as HBV, HCV, dietary aflatoxin B1
exposure, and alcohol intake) and somatic mutations (such as
TP53, PTEN, and BRAF) may not be closely associated with BEX1
reactivation in human HCC (Fig. S13). However, as depicted in
Fig. S14, BEX1 expression is regulated by the transcription factor
c-MYC. Since both BEX1 and AFP are oncofetal proteins, an as￾sociation between them may exist. Our results showed that there
was a positive correlation between BEX1 expression and AFP
serum levels in patients with HCC (Fig. S15).
BEX1 is associated with sorafenib benefit in liver cancer
CSC properties are closely associated with chemoresistance.23
Considering that BEX1 can enhance CSC properties in liver can￾cer, the relationship between BEX1 expression and chemo￾resistance was investigated subsequently. As shown in Fig. S16A,
BEX1 overexpression resulted in the development of hepatoma
cell resistance to sorafenib (Sor)-induced growth inhibition. In
contrast, BEX1 knockdown rendered hepatoma cells sensitive to
Sor-induced growth inhibition (Fig. S16B). We next performed
mouse experiments to validate the above results. To this end,
hydrodynamic injections were performed in WT and Bex1-/-
mice. Three weeks post-injection, at the beginning phase of tu￾mor growth,5 these mice were subjected to treatment with Sor
or vehicle for an additional 3 weeks (Fig. 5A). At 6 weeks, all WT
mice became moribund due to the high tumor burden. Notice￾ably, the BEX1-/-+Sor group displayed the most considerable in￾hibition of HCC development (Fig. 5B). Moreover, compared to
the other 3 groups, the tumor mass, tumor volume, and serum
aspartate aminotransferase and alanine aminotransferase levels
were significantly decreased in the BEX1-/-+Sor group (Fig. 5C).
Results obtained from the mouse xenograft model consistently
showed that BEX1 inhibition could trigger the sensitization of
liver cancer cells to Sor treatment (Fig. 5D). Next, we evaluated
the relationship between BEX1 expression and Sor response in 5
patients with HCC. Based on analysis using patient-derived xe￾nografts, we observed that the xenografts derived from tumors
with high levels of BEX1 and CSC marker expression were
resistant to Sor treatment. In contrast, compared with the vehicle
treatment, Sor treatment almost resulted in a blockade of growth
of xenografts derived from tumors with low expression of the
indicated proteins (Fig. 5E-F). Moreover, early intrahepatic
metastasis was observed in patients with HCC and high BEX1
(CD133+
), and HCC cell lines (Huh7 and Hep3B). (E) Bisulfite sequencing-based evaluation of the 26 CpG sites in the BEX1 promoter was performed using 10 clones
of the WRL68 and CD133+ cell lines. Black dot: methylated; white dot: unmethylated. (F) WRL68 cells or CD133+ cells transfected with Lv-DNMT1 were subjected
to methylation-specific PCR analysis. (G) The mRNA and protein expression of BEX1 in WRL68 and CD133+ cells over expressing DNMT1. *p <0.05, **p <0.01 by
using 2-tailed Student’s t test. CSC, cancer stem cell; EV, empty vector; HCC, hepatocellular carcinoma; IHC, immunohistochemistry; Lv, lentiviral; M, methylated
status; U, unmethylated status.
Journal of Hepatology 2021 vol. - j 1–12 7
expression (Fig. 5F, bottom panel). To further assess the clinical
relevance of BEX1 expression and sensitivity to Sor treatment in
patients with HCC, we determined BEX1 expression in HCC
samples obtained from 85 postoperative HCC patients who had
received adjuvant Sor treatment (cohort 3). Kaplan-Meier anal￾ysis showed prolonged overall survival in patients with HCC and
low BEX1 levels who received adjuvant Sor treatment (Fig. S17A).
Furthermore, we also investigated the expression of BEX1 in
primary tumors obtained from another cohort of patients who
had received Sor treatment after HCC relapse (cohort 4, n = 78).
These results showed that low BEX1 expression in primary HCC
was associated with prolonged survival in patients who received
Sor treatment for recurrent tumors (Fig. S17B).
BEX1-mediated stemness enhancement is b-catenin￾dependent
To identify the downstream pathway that is associated with
BEX1-mediated self-renewal activity, we performed global gene
expression profiling in purified BEX1+ and BEX1- cells. KEGG
pathway analysis showed that Wnt/b-catenin signaling was
significantly enriched in BEX1+ cells (Fig. 6A). Considering that
the Wnt/b-catenin pathway plays a key role in maintaining
stemness in liver cancer,24 we investigated whether this pathway
was involved in BEX1-mediated self-renewal activity. Since the
classical Wnt/b-catenin pathway is activated by either the Wnt
ligand or by glycogen synthase kinase 3b (GSK3b), we tested
whether BEX1 affected the expression of Wnt, and determined
the phosphorylation levels of GSK3b. Our results showed that
neither Wnt expression nor GSK3b phosphorylation was affected
by the overexpression of BEX1 in Huh7 cells or by BEX1 knock￾down in Hep3B cells. This indicated that BEX1-mediated acti￾vation of the Wnt/b-catenin pathway was independent of the
Wnt ligand and GSK3b phosphorylation (Fig. S18). TCF/LEF
luciferase reporter assays showed that overexpression of BEX1
increased the transcriptional activity of b-catenin, whereas BEX1
knockdown attenuated the transcriptional activity of b-catenin
CD133 CT
LHL 2.5 y RHL 1 y LHL 1.5 y LHL 3.5 y RHL 4.5 y
BEX1 EpCAM SALL4 -catenin
Fig. 5. BEX1 is associated with sorafenib benefit in liver cancer. (A) Study design. (B) Representative images of mouse liver tissue from the indicated groups at
6 weeks (n = 6 per group). Scale bars: 1 cm. (C) Tumor mass, tumor volume, and serum ALT and AST levels of the indicated groups. (D) Left panel: images of
xenografts derived from nude mice subcutaneously injected with Huh7-shCtrl and –shBEX1 cells administered with sorafenib or with a vehicle control (n = 6 per
group). Middle panel: quantification of tumor volume. Right panel: quantification of tumor mass. (E) PDXs were treated with sorafenib (60 mg/kg body weight) or
vehicle for 24 days (n = 3 per group), and xenograft growth was monitored. (F) Protein levels of BEX1 and CSC-related proteins in PDXs treated with Sor or Veh
were assessed by IHC. Representative CT images of 5 patients with HCC are shown in the bottom row. Red arrows indicate the sites of tumor recurrence. Scale
bars: 50 lm *p <0.05, **p <0.01, ***p <0.001 by using 2-tailed Student’s t test. ALT, alanine aminotransferase; AST, aspartate aminotransferase; CSCs, cancer stem
cells; IHC, immunohistochemistry; LHL, left hepatic lobectomy; PDXs, patient-derived xenografts; RHL, right hepatic lobectomy; shCtrl, shControl; Sor, sorafenib;
Veh, vehicle.
8 Journal of Hepatology 2021 vol. - j 1–12
Research Article Hepatic and Biliary Cancer
(Fig. 6B). This indicates that BEX1 may regulate stemness by
activating the Wnt/b-catenin pathway. Moreover, both PCR and
immunoblotting analyses suggested that BEX1 overexpression
increased the expression levels of b-catenin and its downstream
target genes c-Myc and cyclin D1, while downregulation of BEX1
demonstrated the opposite effect (Fig. 6C-D). Spheroid assays
confirmed that b-catenin knockdown reduced the enhancement
of BEX1-mediated spheroid formation in both Huh7 and Hep3B
cells (Fig. 6E-F). In conclusion, these results suggest that activa￾tion of the Wnt/b-catenin pathway is necessary for BEX1-
mediated stemness enhancement.
BEX1 blocks RUNX3-mediated inhibition exerted on b-catenin
transcription
Next, we explored the mechanism by which BEX1 induced
transcription of b-catenin. After a comprehensive review of the
literature, it was concluded that RUNX3 is an important tran￾scription factor that downregulates b-catenin levels in a variety
of human carcinomas. Moreover, using the JASPAR core database,
it was observed that RUNX3 bound directly to the b-catenin
promoter. We further verified that RUNX3 overexpression
significantly inhibited the expression of b-catenin (Fig. 7A) and
TCF/LEF luciferase activity (Fig. 7B). Moreover, RUNX3 knock￾down in Huh7 cells promoted spheroid formation (Fig. 7C).
Additionally, immunoprecipitation assays revealed that BEX1
and RUNX3 formed a protein complex in Huh7 cells (Fig. 7D).
Based on this interaction, we hypothesized that BEX1 might
prevent RUNX3 from inhibiting b-catenin transcription. Chro￾matin immunoprecipitation assays confirmed that BEX1 over￾expression inhibited the binding of RUNX3 to the b-catenin
promoter (Fig. 7E). To verify these results, we constructed a
luciferase reporter vector in which luciferase expression was
controlled by the b-catenin promoter. Luciferase assays showed
that RUNX3 overexpression weakened BEX1-induced luciferase
activity. However, the counteractive interaction between RUNX3
and BEX1 was eliminated in the luciferase reporter harboring a
mutated RUNX3 element (Fig. 7F). These results suggest that
BEX1 interacts with RUNX3 to block its inhibition of b-catenin
transcription. Furthermore, we determined the expression of
RUNX3 in patients from cohorts 1 and 2 using IHC. Spearman’s
correlation analysis indicated that the level of RUNX3 in the
nucleus was positively correlated with DNMT1 expression, but
A BEX1pos/BEX1Neg
123123 Wnt/β-catenin signaling
Cell cycle
0 Fold difference in number of spheres
Fig. 6. BEX1-mediated stemness enhancement is b-catenin-dependent. (A) Left panel: Heat map of significantly differentially expressed genes between BEX1+
and BEX1- Huh7 cells. Red and green colors indicate high and low expression levels, respectively. Right panel: KEGG pathway analysis of differentially expressed
genes in the BEX1+ and BEX1- Huh7 cells. (B) BEX1 overexpression promoted transcriptional activity of b-catenin, whereas BEX1 knockdown reduced tran￾scriptional activity of b-catenin. (C) The mRNA and (D) protein expression of b-catenin, Cyclin D1, and c-Myc in Huh7 cells transfected with Lv-BEX1 or vector
control and Hep3B cells transfected with sh-BEX1 or vector control. (E, F) b-catenin knockdown in Huh7 and Hep3B cells reduced the spheroid-forming ability of
cells induced by BEX1 overexpression. Scale bars: 20 lm *p <0.05, **p <0.01 by using 2-tailed Student’s t test. si-Ctrl, si-Control.
Journal of Hepatology 2021 vol. - j 1–12 9
was negatively correlated with BEX1 expression (Fig. S19). Taken
together, these results indicate the occurrence of DNMT1-BEX1-
RUNX3 signaling in liver cancer.
Discussion
In this study, we found that BEX1 was highly expressed in pa￾tients with HB and CSC-HCC, but only expressed at low levels in
patients with non-CSC HCC. Furthermore, the differential
expression pattern of BEX1 in these 3 types of liver cancer was
attributable to an epigenetic reprogramming mechanism,
namely DNA methylation mediated by DNMT1. Mechanistically,
BEX1 established an interaction with RUNX3 to block its inhi￾bition of b-catenin transcription, which led to the activation of
Wnt/b-catenin signaling, and promoted the self-renewal and
maintenance of CSCs in HB and CSC-HCC. In contrast, a reduced
expression of BEX1 in non-CSC HCC activated RUNX3 and
inhibited the activation of Wnt/b-catenin signaling. Our results
suggest that BEX1 plays different roles in regulating CSC prop￾erties in different types of liver cancer (Fig. 7G).
A previous study has reported that BEX1 expression is
involved in the differentiation of neural stem cells and is
necessary for recovery of neurons from injury.7 Moreover, BEX1
plays an important role in the repair of skeletal muscle after
injury, suggesting that it may regulate the activity of adult stem
cells.8 In this study, we have revealed that BEX1 is an oncofetal
protein, and can be considered as a novel stem cell marker for HB
and HCC. Importantly, elevated BEX1 expression was associated
with a poor prognosis in patients with HCC and HB. Previous
studies have indicated that tumor cells with CSC properties are
highly prone to distant metastasis and are characterized by
decreased intercellular adhesion and enhanced motility and in￾vasion. Our results showed that metastatic tumors exhibited
upregulated BEX1 expression, indicating that metastasis might
be driven by tumor cells with high CSC properties and elevated
BEX1 expression.
Accumulating evidence has demonstrated that epigenetic
reprogramming, such as DNA methylation, plays an important role
in stem cell biology.25 For instance, Mikkelsen et al. reported that
treatment with DNMT inhibitors markedly improved the overall
efficiency of the reprogramming process.26 The inhibition of DNA
methylation may cause the transcriptional activation of chromatin
structure, which further affects gene expression and influences cell
Wnt pathway activation
liver CSC self-renewal
BEX1 expression level
DNMT1-mediated methylation
Fig. 7. BEX1 blocks RUNX3-mediated inhibition exerted on b-catenin transcription. (A) Reverse transcription PCR and western blot analysis of b-catenin
expression in Huh7 cells transfected with Lv-RUNX3 and Vector control. (B) The luciferase reporter assay was conducted in Huh7 cells to determine the inhibitory
role of RUNX3 on the transcriptional activity of b-catenin. (C) RUNX3 knockdown causes an increased oncosphere-forming ability in Huh7 cells. Scale bars: 20 lm.
(D) IP assay was performed with anti-BEX1 or anti-RUNX3 antibodies using Huh7 cell lysates. The presence of BEX1 or RUNX3 in IPs was evaluated by
immunoblotting analysis. (E) ChIP analysis of the b-catenin promoter in Vector and Lv-BEX1 Huh7 cells using the indicated antibodies (left panel). The quantified
results are shown in the right panel. (F) Luciferase assays were performed to confirm the direct regulation of the b-catenin promoter by RUNX3 and BEX1. (G)
Schematic diagram showing that BEX1 promotes tumorigenesis by enhancing CSC features through Wnt/b-catenin signaling in liver cancer. *p <0.05, **p <0.01 by
using 2-tailed Student’s t test. ChIP, chromatin immunoprecipitation; IP, immunoprecipitation; CSC, cancer stem cell.
10 Journal of Hepatology 2021 vol. - j 1–12
Research Article Hepatic and Biliary Cancer
fate decision.9 Our results revealed that demethylation remarkably
restored BEX1 expression in HCC cells, indicating that BEX1 levels
were regulated by the expression of DNMT1. In addition to DNA
methylation, we explored whether other epigenetic modifications
were involved in the regulation of BEX1 expression. Our pre￾liminary results showed that m6
A RNA methylation and histone
acetylation were not likely to play a key role in the regulation of
BEX1 expression. However, the specific mechanism remains un￾clear and it warrants further study in the future.
Aberrant Wnt/b-catenin activation is involved in maintaining
stemness in various tumors, including liver cancer.27 In this
study, we performed global gene expression profiling of purified
BEX1+ and BEX1- cells, and found that Wnt/b-catenin signaling
activity was frequently altered in purified BEX1+ cells. Further￾more, b-catenin knockdown weakened the BEX1-mediated Wnt/
b-catenin activation and impaired stemness enhancement in
liver cancer cells, indicating that the Wnt/b-catenin pathway was
necessary for BEX1-mediated stemness enhancement.
Based on results obtained via chromatin immunoprecipita￾tion assays, it was confirmed that RUNX3 bound directly to the b￾catenin promoter. Previous studies have indicated that RUNX3
inhibits transcription of b-catenin by binding to its promoter.28
Considering that BEX1 could establish an interaction with
RUNX3, it is conceivable that this interaction may block RUNX3
inhibition of b-catenin transcription. Indeed, our findings suggest
that BEX1 interacts with RUNX3 to block its inhibition of b-cat￾enin transcription, which leads to the activation of Wnt/b-cat￾enin signaling and regulates self-renewal and maintenance of
CSCs. Taken together, BEX1 can be considered an excellent target
for the formulation of treatment strategies for liver cancers,
especially HB and CSC-HCC. Further analyses of BEX1-mediated
mechanisms may provide insights into a potential therapeutic
strategy for liver cancers.
Abbreviations
AFP, alpha-fetoprotein; BEX1, brain-expressed X-linked protein
1; CSCs, cancer stem cells; DNMT, DNA methyltransferase; HB,
hepatoblastoma; HCC, hepatocellular carcinoma; IHC, immuno￾histochemistry; SALL4, spalt like transcription factor 4; Sor, sor￾afenib; ZEB, zebularine.
Financial support
This work was supported by grants from the National Natural
Science Foundation of China (No. 1670587 and 81903075) and
Shanghai Excellent Youth training program (2018YQ62).
Conflict of interest
The authors declare that they have no conflicts to disclose.
Please refer to the accompanying ICMJE disclosure forms for
further details.
Authors’ contributions
CXL, XLH, ZQC and GW designed the experiments, supervised the
study and revised the paper; QW, NL and TY performed animal
studies; QW wrote the first draft of the paper; YDL, CW and JL
performed cellular experiments; QH, XLZ, XBC and LZ performed
animal studies; LGS and QH aided in clinical sample collection;
all authors have reviewed the final version of the manuscript and
approved it for publication.
Data availability statement
The authors declare that all data supporting the findings of this
study are available in the article, as well as in the Supplementary
Information file.
Acknowledgments
The authors thank Professor Xin Chen and Professor Perry
Hackett for providing the plasmids. We thank Professor Qiang Ai
for providing bioinformatics analysis.
Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jhep.2021.06.025.
References
Author names in bold designate shared co-first authorship
[1] Llovet JM, Zucman-Rossi J, Pikarsky E, Sangro B, Schwartz M, Sherman M,
et al. Hepatocellular carcinoma. Nat Rev Dis Prim 2016;2:16018.
[2] Gabay M, Li Y, Felsher DW. MYC activation is a hallmark of cancer initi￾ation and maintenance. Cold Spring Harb Perspect Med 2014;4:a014241.
[3] Shachaf CM, Kopelman AM, Arvanitis C, Karlsson A, Beer S, Mandl S, et al.
MYC inactivation uncovers pluripotent differentiation and tumour
dormancy in hepatocellular cancer. Nature 2004;431:1112–1117.
[4] Chow EK, Fan LL, Chen X, Bishop JM. Oncogene-specific formation of
chemoresistant murine hepatic cancer stem cells. Hepatology
2012;56:1331–1341.
[5] Liu P, Ge M, Hu J, Li X, Che L, Sun K, et al. A functional mammalian target
of rapamycin complex 1 signaling is indispensable for c-Myc-driven
hepatocarcinogenesis. Hepatology 2017;66:167–181.
[6] Gu Y, Wei W, Cheng Y, Wan B, Ding X, Wang H, et al. A pivotal role of BEX1
in liver progenitor cell expansion in mice. Stem Cell Res Ther 2018;9:164.
[7] Khazaei MR, Halfter H, Karimzadeh F, Koo JH, Margolis FL, Young P, et al.
Bex1 is involved in the regeneration of axons after injury. J Neurochem
2010;115:910–920.
[8] Koo JH, Smiley MA, Lovering RM, Margolis FL. Bex1 knock out mice show
altered skeletal muscle regeneration. Biochem Biophys Res Commun
2007;363:405–410.
[9] Raggi C, Factor VM, Seo D, Holczbauer A, Gillen MC, Marquardt JU, et al.
Epigenetic reprogramming modulates malignant properties of human
liver cancer. Hepatology 2014;59:2251–2262.
[10] Ruau D, Ensenat-Waser R, Dinger TC, Vallabhapurapu DS, Rolletschek A,
Hacker C, et al. Pluripotency associated genes are reactivated by
chromatin-modifying agents in neurosphere cells. Stem Cells
2008;26:920–926.
[11] Jones PA. Functions of DNA methylation: islands, start sites, gene bodies
and beyond. Nat Rev Genet 2012;13:484–492.
[12] Lee CH, Wong TS, Chan JY, Lu SC, Lin P, Cheng AJ, et al. Epigenetic regu￾lation of the X-linked tumour suppressors BEX1 and LDOC1 in oral
squamous cell carcinoma. J Pathol 2013;230:298–309.
[13] Ivics Z, Hackett PB, Plasterk RH, Izsvák Z. Molecular reconstruction of
Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in
human cells. Cell 1997;91:501–510.
[14] Wang G, Wang Q, Liang N, Xue H, Yang T, Chen X, et al. Oncogenic driver
genes and tumor microenvironment determine the type of liver cancer.
Cell Death Dis 2020;11:313.
[15] Wu Y, Wang J, Zheng X, Chen Y, Huang M, Huang Q, et al. Establishment
and preclinical therapy of patient-derived hepatocellular carcinoma
xenograft model. Immunol Lett 2020;223:33–43.
[16] Liu C, Liu L, Chen X, Cheng J, Zhang H, Shen J, et al. Sox9 regulates self￾renewal and tumorigenicity by promoting symmetrical cell division of
cancer stem cells in hepatocellular carcinoma. Hepatology 2016;64:117–
129.
[17] Vilar M, Murillo-Carretero M, Mira H, Magnusson K, Besset V, Ibáñez CF.
Bex1, a novel interactor of the p75 neurotrophin receptor, links neuro￾trophin signaling to the cell cycle. EMBO J 2006;25:1219–1230.
[18] Carter BD. A Bex-cycle built for two. EMBO Rep 2006;7:382–384.
[19] Dan J, Rousseau P, Hardikar S, Veland N, Wong J, Autexier C, et al. Zscan4
inhibits maintenance DNA methylation to facilitate telomere elongation
in mouse embryonic stem cells. Cell Rep 2017;20:1936–1949.
Journal of Hepatology 2021 vol. – j 1–12 11
[20] Liu F, Huang W, Hong J, Cai C, Zhang W, Zhang J, et al. Long noncoding
RNA LINC00630 promotes radio-resistance by regulating BEX1 gene
methylation in colorectal cancer cells. IUBMB Life 2020;72:1404–1414.
[21] Li LC, Dahiya R. MethPrimer: designing primers for methylation PCRs.
Bioinformatics 2002;18:1427–1431.
[22] Kohga K, Tatsumi T, Takehara T, Tsunematsu H, Shimizu S, Yamamoto M,
et al. Expression of CD133 confers malignant potential by regulating
metalloproteinases in human hepatocellular carcinoma. J Hepatol
2010;52:872–879.
[23] Xiang DM, Sun W, Zhou T, Zhang C, Cheng Z, Li SC, et al. Oncofetal HLF
transactivates c-Jun to promote hepatocellular carcinoma development
and sorafenib resistance. Gut 2019;68:1858–1871.
[24] Guo Y, Wu Z, Shen S, Guo R, Wang J, Wang W, et al. Nanomedicines
reveal how PBOV1 promotes hepatocellular carcinoma for effective gene
therapy. Nat Commun 2018;9:3430.
[25] Gutzat R, Rembart K, Nussbaumer T, Hofmann F, Pisupati R,
Bradamante G, et al. Arabidopsis shoot stem cells display dynamic tran￾scription and DNA methylation patterns. EMBO J 2020;39:e103667.
[26] Mikkelsen TS, Hanna J, Zhang X, Ku M, Wernig M, Schorderet P, et al.
Dissecting direct reprogramming through integrative genomic analysis.
Nature 2008;454:49–55.
[27] Huang M, Zhang D, Wu JY, Xing K, Yeo E, Li C, et al. Wnt-mediated
endothelial transformation into mesenchymal stem cell-like cells induces
chemoresistance in glioblastoma. Sci Transl Med 2020;12:eaay7522.
[28] Guo X, Dai X, Liu J, Cheng A, Qin C, Wang Z. Circular RNA circREPS2 acts as
a sponge of miR-558 to suppress gastric cancer progression by
regulating RUNX3/beta-catenin signaling. Mol Ther Nucl Acid
2020;21:577–591.
12 Journal of Hepatology 2021 vol. – j 1–12
Research Article Hepatic and Biliary Cancer