Carcinogenesis vol.35 no.10 pp.2175–2182, 2014
doi:10.1093/carcin/bgu110
Advance Access publication May 15, 2014
MicroRNA-21 promotes tumour malignancy via increased nuclear translocation
of β-catenin and predicts poor outcome in APC-mutated but not in APC-wild-type
colorectal cancer
Po-Lin Lin1, De-Wei Wu2, Chi-Chou Huang3,4,
Tsung-Ying He1, Ming-Chih Chou1,3,4, Gwo-Tarng Sheu1
and Huei Lee2,*
1
Institute of Medicine, College of Medicine, Chung Shan Medical University,
Taichung 402, Taiwan, 2Graduate Institute of Cancer Biology and Drug
Discovery, College of Medical Science and Technology, Taipei Medical
University, Taipei 115, Taiwan, 3Department of Medicine, Chung Shan Medical
University, Taichung 402, Taiwan and 4Department of Surgery, Chung Shan
Medical University Hospital, Taichung 402, Taiwan
*To whom correspondence should be addressed. Graduate Institute of Cancer
Biology and Drug Discovery, College of Medical Science and Technology,
Taipei Medical University, Room 5, 12F, No. 3, Park Street, Nangang District,
Taipei 115, Taiwan. Tel: +886 2 2736 1661; Fax: +886 2 2255 8562;
Email: [email protected]
MiR-21 has been associated with poor prognosis in colon adenocarcinomas. However, in our preliminary data, the prognostic
value of miR-21 levels was observed only in adenomatous polyposis
coli (APC)-mutated tumours, not in APC-wild-type tumours. We
explored whether β-catenin nuclear translocation was synergistically promoted by miR-21 in APC-mutated cells but not in APCwild-type cells. We enrolled 165 colorectal tumour to determine
APC mutation, miR-21 levels and nuclear β-catenin expression
by direct sequencing, real-time PCR and immunohistochemistry. Overall survival and relapse-free survival were analysed by
Kaplan–Meier and Cox regression models. The mechanistic action
of β-catenin nuclear translocation modulated by miR-21 and its
effect on cell invasion were evaluated in a cell model. Positive
nuclear β-catenin expression was more commonly occurred in
APC-mutated tumours than in APC-wild-type tumours. High
miR-21 levels were relatively more common in tumours with
positive nuclear β-catenin expression than in those with negative
nuclear β-catenin expression. APC-mutated tumours with high
miR-21 levels had shorter overall survival and relapse-free survival periods compared with others. However, the prognostic value
of miR-21 levels was not observed in APC-wild-type tumours.
Phosphorylation of β-catenin at Ser552 via the miR-21-mediated
PTEN/AKT axis plays a critical role in β-catenin nuclear translocation in APC-mutated cells but not in APC-wild-type cells.
Moreover, nuclear β-catenin expression increased by miR-21 is
responsible for the capability of invasiveness. In summary, nuclear
translocation of β-catenin increased by miR-21 promotes tumour
malignancy and a poor outcome in APC-mutated patients but not
in APC-wild-type colorectal cancer.
Introduction
Colorectal cancer has the highest incidence and is the third leading
cause of cancer death in Taiwan; however, the prognosis of patients
with colorectal cancer is relatively more favourable when compared
with other cancers (1). Therefore, establishing reliable prognostic and
therapeutic markers is useful to improve the outcome in colorectal
cancer patients. The majority of colorectal tumours have a mutation
Abbreviations: AKT, protein kinase B (PKB); APC, adenomatous polyposis
coli; GSK-3β, glycogen synthase kinase-3β; HR, hazard ratio; KRAS, V-Kiras2 Kirsten rat sarcoma viral oncogene homolog; miR-21, microRNA-21;
OS, overall survival; PTEN, phosphatase and tensin homolog; RFS, relapsefree survival; TCF, T-cell factor; TGF-βR2, transforming growth factor beta,
receptor 2.
in a key regulatory factor of the Wnt/β-catenin pathway, resulting in
activation of the pathway. The mutations were most often in adenomatous polyposis coli (APC) and rarely occurred in β-catenin (2–6). Up
to 80% of tumours have a nuclear accumulation of β-catenin to cooperate with T-cell factor (TCF) and in turn up-regulate the expression
of downstream genes of the Wnt/β-catenin pathway, such as c-Myc
and cyclin D1 (7–10). Activation of downstream genes of the Wnt/βcatenin pathway by APC mutation was shown to be linked with the
initiation of colorectal tumourigenesis (11,12). However, the prognostic significance of APC mutation in colorectal cancer has not yet
been reported. This observation suggests that the APC mutation could
not be fully responsible for β-catenin nuclear translocation through
the activation of the Wnt/β-catenin pathway and consequently for the
promotion of tumour malignancy and the poor outcome in colorectal
cancer.
MicroRNA-21 (MiR-21) overexpression is commonly observed in
human cancers, including colorectal cancer (13,14). MiR-21 expression levels have been associated with poor prognosis and poor chemotherapeutic response in colon adenocarcinoma patients (15). Loss of
the miR-21 allele reduces tumourigenesis via up-regulation of Spry2,
phosphatase and tensin homolog (PTEN) and programmed cell death
4 (PDCD4) and inhibits phosphorylation of extracellular signal-regulated kinases (ERK), protein kinase B (AKT) and c-Jun N-terminal
kinases (JNK) (16). Phosphorylation of β-catenin at Ser552 by AKT
plays an important role in increased nuclear translocation and transcriptional activity of the β-catenin/TCF axis (17). However, our
preliminary data showed that the prognostic significance of miR-21
expression levels was observed only in colorectal cancer patients
whose tumours harboured an APC mutation. Therefore, we expected
that the β-catenin/TCF axis, targeted by miR-21 via the PTEN/AKT
pathway, might promote tumour progression and metastasis in APCmutated colorectal cancer via increased nuclear translocation of
β-catenin and, in turn, results in a poor outcome.
Materials and methods
Study subjects
This study included 165 resected tumours from colorectal cancer patients
recruited from the Colorectal Division, Department of Surgery, Chung Shan
Medical University Hospital, Taichung, Taiwan, between 2000 and 2007. All
subjects were unrelated ethnic Chinese residents of central Taiwan and provided written informed consent approved by the Institutional Review Board.
Among these patients, 30 had tumour metastases in the liver (12 cases), omentum (5 cases), lung (3 three cases), hypopharynx, bone, left para-aortic lymph
node, pelvis, rectum and mesentery (all one case); the metastatic data of five
patients were not available. In total, three patients had tumours that metastasized to more than one organ. No patient received chemotherapy or radiotherapy before surgical resection. The overall survival (OS) of patients was
calculated from the date of surgery. Board-certified pathologists conducted
examinations for the pathological stage of each case. Information pertaining to
personal characteristics was collected from hospital reports.
Real-time PCR analysis of miR-21 and cyclin D1 mRNA expression levels
DNase I-treated total RNA was subjected to miRNA real-time PCR analysis with the TaqMan miRNA Reverse Transcription Kit, miRNA assays and
an Real-Time Thermocycler 7500 (Applied Biosystems, Foster City, CA).
RNU6B was used as the small RNA reference housekeeping gene. The primers
used for real-time PCR analysis of cyclin D1 mRNA expression are shown in
Supplementary Table 1, available at Carcinogenesis Online. The miR-21 levels in tumours that were higher than the median value were defined as ‘high’,
whereas levels lower than the median value were defined as ‘low’.
Immunohistochemical staining
The β-catenin expression in paraffin sections of colorectal tumours from
primary colorectal cancer patients was evaluated by immunohistochemistry
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
2175
P.-L.Lin et al.
using β-catenin antibodies purchased from Santa Cruz Biotechnology (Santa
Cruz, CA) and immunohistochemical procedures described previously (18).
Samples did not treat with the primary antibody served as negative controls.
The immunohistochemical staining scores were defined as described previously (19,20), and the signal intensities were evaluated independently by three
observers. Immunostaining scores were defined as the cell staining intensity
(0 = nil; 1 = weak; 2 = moderate and 3 = strong) multiplied by the percentage
of labelled cells (0–100%), leading to scores from 0 to 300. A score ≥150 was
defined as ‘high’ immunostaining, whereas a score <150 was defined as ‘low’.
APC sequencing
APC mutations in colorectal cancer patients were detected in the mutation
cluster region. The mutations were determined by direct sequencing of PCR
products amplified from the DNA of tumour cells isolated by the colorectal
tumour tissues. These primers are shown in Supplementary Table 1, available
at Carcinogenesis Online. The PCR products were sequenced using an Applied
Biosystems 3100 Avant Genetic Analyser (Applied Biosystems), and the same
primers were used for both the PCR and the DNA sequencing. All mutations
were confirmed by direct sequencing of both DNA strands.
Cell lines
LoVo, HCT15 and HT29 cells were kindly provided by Dr S.-G.Shiah (National
Institute of Cancer Research, National Health Research Institutes, Miaoli, Taiwan).
HCT116 cells were kindly provided by Dr C.C.Chang (Institute of Biomedical
Sciences, National Chung Hsing University, Taichung, Taiwan). SW480 cells were
obtained from the Bioresource Collection and Research Center, Food Industry
Research and Development Institute (Hsinchu, Taiwan). HCT15, HCT116 and
HT29 colon cancer cell lines were maintained in RPMI-1640 (HyClone, Logan,
UT), SW480 colon cancer cell lines were maintained in L-15 (HyClone), and
LoVo colorectal cancer cells were maintained in F12-K (HyClone). The medium
contained 10% fetal bovine serum supplemented with penicillin (100 U/ml) and
streptomycin (100 mg/ml). These cells were cultured according to the suppliers’ instructions and were stored used at passages 5–20. Once resuscitated, cell
lines were routinely authenticated (once every 6 months; cells were last tested in
December 2012) through cell morphology monitoring, growth curve analysis, species verification by isoenzymology and karyotyping, identity verification using
short tandem repeat profiling analysis and contamination checks.
Plasmid construction, transfection and stable clone selection
β-catenin and APC shRNA were purchased from the National RNAi
Core Facility, Academia Sinica, Taiwan, and the sequences were shown
in Supplementary Table 1, available at Carcinogenesis Online. The seeding sequences of two β-catenin shRNA were located at 3′UTR of β-catenin
(Supplementary Figure 1A, available at Carcinogenesis Online). However,
the plasmid-derived exogenous β-catenin did not contain 3′UTR. Two stable
clones of β-catenin-knockdown HCT15 and HCT116 cells were selected by
puromycin. The expression vector of β-catenin and APC were purchased from
Addgene (Cambridge, MA). Non-specific shRNA control of the scramble
sequence was used as the control in the knockdown experiment, and a vector control was used as the control of β-catenin overexpression. β-cateninS552A was generated using the QuickChange site-directed mutagenesis
system (Stratagene, Santa Clara, CA); the primers are shown in Supplementary
Table 1, available at Carcinogenesis Online. The transfection and stable clone
selection procedures were as described previously (19).
MiR-21 precursor and inhibitor transfection
Six-well plates were used to confluence the cells. MiR-21 precursor (premiR-21, 20–40 nmol/l per well; Ambion, Foster City, CA), miR-21 inhibitor
(40–80 nmol/l per well; Ambion) and negative control cells (Ambion) were
transfected using Lipofectamine 2000 transfection reagent (Invitrogen, Foster
City, CA) according to the manufacturers’ protocols. Transfection efficiency
was evaluated by real-time PCR.
Luciferase reporter assay
Cells were transfected with TOPFLASH and FOPFLASH (Millipore, Billerica,
MA). Luciferase assays were performed using the Luciferase Reporter Assay
System (Promega, Madison, WI) 24 h after transfection. Normalized luciferase
activity was reported as the ratio of luciferase activity/β-galactosidase activity.
Boyden chamber invasion assay
A Boyden chamber with a pore size of 8 µm was used for the in vitro cell invasion assay. Cells (5 × 104) in 0.5% serum containing the culture medium were
plated in the upper chamber, and 10% fetal bovine serum was added to the culture medium in the lower chamber as a chemoattractant. The upper side of the
filter was covered with 0.2% Matrigel (Collaborative Research, San Francisco,
CA) diluted in RPMI-1640. After 24 h, cells on the upper side of the filter were
removed, and cells that adhered to the underside of the membrane were fixed in
2176
95% ethanol and stained with 10% Giemsa dye. The number of invasive cells
was counted. Ten contiguous fields of each sample were examined to obtain a
representative number of cells that invaded across the membrane. The procedures and methods were as previously described (21).
Statistical analysis
All statistical analyses were performed using the SPSS statistical software
program as described previously (version 11.0; SPSS, Chicago, IL) (19,22).
Statistical tests were examined by two-sided analysis of the variance, and
P values <0.050 were considered statistically significant.
Results
Nuclear β-catenin expression correlated with APC mutations and
miR-21 expression in tumours from colorectal cancer patients
The expression levels of miR-21 and APC mutations in 165 colorectal
tumours were examined by real-time PCR and direct sequencing to
evaluate whether both molecules could be associated with patients’
clinical parameters. As shown in Supplementary Tables 2 and 3,
available at Carcinogenesis Online, miR-21 expression levels and
APC mutation were not associated with clinical parameters, including age, gender, tumour site, cigarette smoking and tumour stage.
β-catenin was evaluated by immunohistochemistry and the representative immunostaining results are shown in Supplementary Figure 2,
available at Carcinogenesis Online. There was a higher prevalence of
positive nuclear β-catenin expression in APC-mutated tumours than
in APC-wild-type tumours (P = 0.003; Table I). Interestingly, high
miR-21 levels were relatively more common in tumours with positive nuclear β-catenin, but this did not reach statistical significance
(P = 0.182; Table I). When tumours were divided into four subgroups
by two parameters of APC mutation and miR-21, the highest prevalence of positive nuclear β-catenin expression was observed in APCmutated tumours with high miR-21 levels compared with the other
three subgroups (85.7% versus 45.6, 25.0 and 28.0%, P < 0.001; Table
I). These results suggest that miR-21 might promote nuclear translocation of β-catenin in patients with APC-mutated tumours.
MiR-21 expression levels are associated with poor OS and relapse-free
survival in APC-mutated patients but not in APC-wild-type patients
As previously mentioned, miR-21 expression levels have been associated with poor prognosis in colon adenocarcinoma patients (13,15).
Herein, the usefulness of miR-21 expression levels was evaluated for
predicting the outcome in colorectal cancer. Kaplan–Meier analysis
showed that miR-21 and APC mutations were not associated with OS
and relapse-free survival (RFS) in patients studied (data not shown).
When tumours were divided into four subgroups by both parameters
(P = 0.045 for OS; P = 0.019 for RFS; Figure 1A and B), APC-mutated
tumours with high miR-21 levels had shorter OS and RFS periods
than APC-wild-type tumours with high miR-21 levels (P = 0.035 for
OS; P = 0.024 for RFS; Figure 1A and B) or APC-mutated tumours
with high miR-21 levels (P = 0.031 for OS; P = 0.044 for RFS;
Figure 1A and B). The prognostic significance was not observed
between APC-mutated patients with high miR-21 levels versus
APC-wild-type patients with low miR-21 levels (P = 0.392 for OS;
P = 0.412 for RFS; Figure 1A and B) and APC-mutated patients with
low miR-21 levels versus APC-wild-type patients with high miR-21
levels (P = 0.547 for OS; P = 0.701 for RFS; Figure 1A and B).
In addition, poorer OS and RFS in APC-mutated patients with high
miR-21 levels was still observed compared with the combination of
the other three subgroups (P = 0.019 for OS, Figure 1C; P = 0.014
for RFS, Figure 1D). Cox regression analysis indicated that APCmutated patients with high miR-21 levels had a hazard ratio (HR) of
1.98 and 1.59 for OS and RFS, respectively, compared with APCmutated patients with low miR-21 tumours (95% confidence interval
[CI]: 1.14–3.24, P = 0.045 for OS; 95% CI: 1.04–3.56, P = 0.047
for RFS; Table II). However, the prognostic value of miR-21 levels for OS and RFS was not observed in APC-wild-type patients
(Table II). Interestingly, APC-mutated patients with high miR-21
levels had poorer OS and RFS than did APC-wild-type patients with
MiR-21 enhances β-catenin nuclear translocation
Table I. Relationships between β-catenin protein nucleus expression and
clinical-pathological parameters in colorectal cancer patients
Nucleus β-catenin
Parameters
Case no.
165
Age
≤65
76
89
>65
Gender
Female
76
Male
89
Tumour site
Proximal colon
40
Distal colon
54
Rectum
71
Smoking status
Non-smoking
109
Smoking
56
Stage
I + II
69
III
66
IV
30
APC mutation
No
105
Yes
60
miR-21
Low
84
High
81
APCmut/miR-21 status
No/low
57
No/high
48
Yes/low
25
Yes/high
35
Negative (%)
Positive (%)
90 (54.5)
75 (45.5)
36 (47.4)
54 (60.7)
40 (52.6)
35 (39.3)
0.087
40 (52.6)
50 (56.2)
36 (47.4)
39 (43.8)
0.648
21 (52.5)
30 (55.6)
39 (54.9)
19 (47.5)
24 (44.4)
32 (45.1)
0.954
58 (53.2)
32 (57.1)
51 (46.8)
24 (42.9)
0.631
39 (56.5)
35 (53.0)
16 (53.3)
30 (43.5)
31 (47.0)
14 (46.7)
0.910
67 (63.8)
23 (38.3)
38 (36.2)
37 (61.7)
0.003
49 (59.8)
41 (49.4)
33 (40.2)
42 (50.6)
0.182
31 (54.4)
36 (75.0)
18 (72.0)
5 (14.3)
26 (45.6)
12 (25.0)
7 (28.0)
30 (85.7)
<0.001
P
high miR-21 levels (HR: 2.73, 95% CI: 1.37–5.46, P = 0.004 for OS;
HR: 2.88, 95% CI: 1.46–5.57, P = 0.02 for RFS; Table II). These
results suggest that miR-21 levels may be useful to predict the outcome in APC-mutated patients but not in APC-wild-type patients.
Cell invasion capability is dependent on miR-21 expression in
colon cancer cells with APC or β-catenin mutation
Five types of colon cancer cells (HCT15, SW480, HT29, LoVo and
HCT116) were used to test whether miR-21 expression could modulate cell invasion capability; HCT15, SW480, HT29 and LoVo possessed APC mutations, and HCT116 harboured β-catenin codon 45
mutations. As shown in Figure 2A, the highest miR-21 expression
level was found in HCT116 cells, followed by LoVo, HT29, SW480
and HCT15 cells. MiR-21 expression was knocked down by miR21 inhibitor in HCT116 and LoVo cells and ectopically expressed
by miR-21 precursor in HCT15 and SW480 cells (Figure 2B). As
expected, miR-21 expression levels decreased in miR-21 knockdown
HCT116 and LoVo cells and increased in miR-21 overexpressing
HCT15 and SW480 cells. Boyden chamber and MTT assays showed
that the invasion and cell growth capability decreased markedly in
miR-21 knockdown HCT116 and LoVo cells and increased in miR-21
overexpressing HCT15 and SW480 cells in a dose-dependent manner (Figure 2C). Representative invasive cells on Matrigel membranes
are shown in Figure 2C (right panel). However, cell proliferation was
not affected by miR-21 precursor and miR-21 inhibitor in these cells
(Supplementary Figure 3, available at Carcinogenesis Online). These
results clearly indicate that miR-21 promotes invasiveness but does
not affect cell proliferation in colon cancer cells.
MiR-21 promotes β-catenin nuclear translocation via increased
phosphorylation of β-catenin at Ser552
Phosphorylation of β-catenin at Ser552 by AKT promotes β-catenin
transcriptional activity and, in turn, enhances tumour invasion (17).
APC-mutated HCT15 with wild-type β-catenin and APC-wild-type
HCT116 with mutated β-catenin were used to determine whether
miR-21 could promote nuclear translocation of β-catenin. HCT15
and HCT116 cells were treated with two doses of miR-21 precursor
and miR-21 inhibitor, respectively. The nuclear/cytoplasmic fraction
showed that nuclear β-catenin expression increased by miR-21 overexpression in HCT15 cells and decreased by miR-21 knockdown in
HCT116 cells; however, the opposite phenomenon were observed in
cytoplasmic β-catenin expression in both cells with miR-21 overexpression and knockdown (Figure 3A). α-Tubulin and Sp1 were used
as loading protein controls of cytoplasmic and nuclear fractions of
colon cancer cells. As expected, cyclin D1 and phosphorylated (p)AKT expression increased and PTEN expression decreased in miR-21
overexpressing HCT15 cells (Figure 3A). Conversely, the decrease of
cyclin D1 and p-AKT and increase of PTEN expression was observed
in miR-21 knockdown HCT116 cells (Figure 3A). Luciferase reporter
activity assay showed that TCF-reporter (TOPFLASH) activity and
cyclin D1 mRNA expression levels were dose dependently elevated by miR-21 precursor treatment in HCT15 and HCT116 cells
(Figure 3B); however, TCF-reporter activity was not detected in
the TCF-mutated reporter (FOPFLASH) system (Figure 3B). These
results suggest that increased phosphorylation of β-catenin at Ser552
by the miR-21-mediated PTEN/AKT axis promotes β-catenin nuclear
translocation and its transcriptional activity.
Phosphorylation of β-catenin at Ser552 is responsible for β-catenin
nuclear translocation and its transcriptional activity in APCmutated but not in APC-wild-type colon cancer cells
To confirm the efficacy of β-catenin silenced by shRNA, two
β-catenin shRNA were used to knock down β-catenin in HCT15
and HCT116 cells. TCF-reporter activity assay showed that a similar reduction of TCF-reporter activity was observed in both cells
transfected with two β-catenin shRNA (Supplementary Figure 1B,
available at Carcinogenesis Online). To exclude the clonal effects,
two stable clones of β-catenin knockdown of HCT15 and HCT116
were established and a similar reduction of TCF-reporter activity was
seen in two stable clones of both cells (Supplementary Figure 1C,
available at Carcinogenesis Online). We next explored whether
phosphorylation of β-catenin at Ser552 could be responsible for its
nuclear translocation and transcriptional activity in APC-mutated
but not APC-wild-type colon cancer cells. To test the hypothesis,
one of β-catenin-knockdown (shβ-catenin) HCT15 stable clone with
mutated APC was used to introduce the expression vector of wild-type
β-catenin or β-catenin with the S552A mutation (β-catenin-S552A)
and/or treatment of the miR-21 precursor. Western blotting showed
that nuclear β-catenin expression increased markedly in HCT15
cells with wild-type β-catenin transfection plus miR-21 precursor
treatment, but nuclear β-catenin expression was slightly elevated
by wild-type β-catenin alone (Figure 4A, upper panel). However,
nuclear β-catenin expression was not detected in HCT15 cells transfected with β-catenin-S552A transfection and/or combined with the
miR-21 precursor (Figure 4A, upper panel). Consistently, cytoplasmic β-catenin disappeared in HCT15 cells with wild-type β-catenin
transfection plus the miR-21 precursor (Figure 4A, upper panel).
Cytoplasmic β-catenin expression was observed in HCT15 cells with
β-catenin-S552A transfection and/or combined with the miR-21 precursor (Figure 4A, upper panel). A marked increase in TCF-reporter
activity and invasion capability were observed in HCT15 cells with
wild-type β-catenin plus the miR-21 precursor compared with
other treatments (Figure 4A, lower panel, and Figure 4B). Nuclear
β-catenin expression level, TCF-reporter activity and invasion capability were not elevated by the miR-21 precursor or miR-21 inhibitor in the shβ-catenin HCT116 stable clone with wild-type β-catenin
transfection, when compared with the shβ-catenin HCT116 stable
clone without wild-type β-catenin transfection and/or the miR-21 precursor or inhibitor co-treatment (Figure 4B). In addition, shβ-catenin
HCT116 stable clones with wild-type β-catenin transfection were
co-transfected with APC shRNA (shAPC) and/or miR-21 precursor
2177
P.-L.Lin et al.
Fig. 1. Kaplan–Meier survival curves of APC mutation, miR-21 and both for OS and RFS in colorectal cancer patients. Survival curves of APC mutation
(Mutated/Wild-type), miR-21 (High/Low) and both for (A) OS and (B) RFS of all studied patients and the four subgroups by miR-21 expression levels and APC
mutation status. The P value of APC-mutated/high miR-21 versus APC-wild-type/high miR-21 levels (P = 0.035 for OS; P = 0.024 for RFS), APC-mutated/
high miR-21 versus APC-mutated /low miR-21 levels (P = 0.031 for OS; P = 0.044 for RFS), APC-mutated/high miR-21 versus APC-wild-type/low miR-21
(P = 0.392 for OS; P = 0.412 for RFS) and APC-mutated/low miR-21 versus APC-wild-type/high miR-21 (P = 0.547 for OS; P = 0.701 for RFS). (C) OS and (D)
RFS between APC mutation and miR-21 high expression patients compared with others (P = 0.019 for OS; P = 0.014 for RFS).
Table II. Cox regression analysis for the combined effects of APC mutation and miR-21 on OS and RFS in colorectal cancer patients
APC mutation
/miR-21
OS
RFS
Case no.
5-year
survival (%)
Median
survival
(month)
HR
No/high
Yes/high
Yes/low
No/low
Yes/low
No/low
No/high
Yes/high
41
35
25
48
25
48
41
35
66
42
64
39
64
39
66
42
79
30
77
48
77
48
79
30
1
2.73
1.30
1.44
1
0.99
0.71
1.98
HR: adjusted by age, genders, smoking status and stage.
2178
95% CI
P
1.37–5.46
0.49–3.40
0.72–2.91
0.004
0.605
0.306
0.44–5.38
0.46–6.29
1.14–3.24
0.994
0.469
0.045
Case no.
5-year
survival
(%)
Median
survival
(month)
HR
45
33
23
54
23
54
45
33
65
36
57
39
57
39
65
36
77
28
40
47
40
47
77
28
1
2.88
1.89
1.54
1
0.84
0.55
1.59
95% CI
P
1.46–5.57
0.78–4.62
0.79–2.08
0.002
0.162
0.191
0.39–1.85
0.24–1.26
1.04–3.56
0.669
0.159
0.047
MiR-21 enhances β-catenin nuclear translocation
Fig. 2. MiR-21 promotes invasive capability of colon cancer cells. (A) Relative miR-21 expression levels of four colon cancer cell lines determined by real-time
PCR. (B) HCT15, SW480, HCT116 and LoVo cells were treated with various doses of miR-21 precursor and inhibitor for 48 h, respectively. Then the miR-21
expression level was determined by real-time PCR. (C) The Boyden chamber invasion assay was performed to evaluate the changes of invasiveness in HCT15
and SW480 cells with miR-21 precursor treatment and in HCT116 and LoVo cells with miR-21 inhibitor treatment when compared with their parental cells. The
invasive cells on Matrigel membranes are shown as representative pictures in the left panel and as a quantitative graph in the right panel.
to examine whether miR-21 could enhance β-catenin nuclear translocation, TCF-reporter activity and invasion capability. As shown
in Figure 4C, the levels of phosphorylated β-catenin at Ser552 in
shβ-catenin HCT116 stable clone were markedly increased by wildtype β-catenin plus shAPC transfection compared with shβ-catenin
HCT116 stable clone with wild-type β-catenin transfection. The level
of phosphorylated β-catenin at Ser552 was more elevated by the miR21 precursor in the shβ-catenin HCT116 stable clone with wild-type
β-catenin plus shAPC transfection. Interestingly, the TCF-reporter
activity and invasion capability was dependent on the levels of phosphorylated β-catenin at Ser552 in the shβ-catenin HCT116 stable
clone with wild-type β-catenin plus shAPC transfection and/or miR21 precursor. On the other hand, no influence of the miR-21 precursor
or miR-21 inhibitor on β-catenin nuclear translocation was observed
in the shAPC HCT15 stable clone with wild-type APC transfection even though the phosphorylated β-catenin expression level was
elevated by miR-21 precursor (Supplementary Figure 4, available at
Carcinogenesis Online). Collectively, these results indicate that miR21 promotes β-catenin nuclear translocation, TCF-reporter activity
and invasion capability in APC-mutated (or silencing) but not in APCwild-type colon cancer cells.
Discussion
MiR-21 has been associated with poor prognosis via the promotion
of colorectal tumour progression and metastasis. However, the role of
miR-21 in colorectal tumour malignancy remains unclear. In the present study, an increase in β-catenin nuclear translocation by miR-21
was responsible for miR-21-mediated invasiveness in APC-mutated
cells but not in APC-wild-type colon cancer cells. Among patients,
nuclear β-catenin expression positively correlated with APC mutation and miR-21 levels in colorectal tumours. The prognostic value
of miR-21 levels on OS and RFS was observed only in APC-mutated
patients but not in APC-wild-type patients. In addition, consistent with
a previous study (23), the correlation of miR-21 expression with APC
mutation and nuclear β-catenin expression was observed in advanced
patients of this study population (Supplementary Table 4, available
at Carcinogenesis Online). Therefore, we suggest that miR-21 levels
may serve to predict the outcome in APC-mutated patients but not in
APC-wild-type patients.
APC functions in a normal colon to negatively regulate Wnt signalling through targeting of β-catenin for proteasome degradation
(24). Wild-type APC plays a central role in a destruction complex that
includes Axin, glycogen synthase kinase-3β (GSK-3β) and casein
kinase 1 (CK1) (25,26). This complex directs a series of phosphorylation events on β-catenin that targets it for ubiquitination and
subsequent proteolysis (25). The canonical mechanism of β-catenin
regulation involves a destruction complex where β-catenin is phosphorylated by GSK-3β at the Thr41, Ser45 and Ser37 sites (27,28).
Our data showed that the expression levels of phosphorylation of
β-catenin at Ser37 and Thr41/Ser45 did not change by miR-21 overexpression in HCT15 and HCT116 cells (Supplementary Figure 5,
available at Carcinogenesis Online), suggesting that GSK-3β might
play a minor role in miR-21-mediated β-catenin nuclear translocation.
A recent report using zebrafish and human cells showed that
homozygous loss of APC causes failed intestinal cell differentiation
due to the elevation of C-terminal-binding protein-1 (CtBP1) (6).
Increased nuclear β-catenin and intestinal proliferation require the
additional activation of V-Ki-ras2 Kirsten rat sarcoma viral oncogene
homolog (KRAS) (6). Therefore, expression of CtBP1 elevated by
APC loss contributes to adenoma initiation, whereas KRAS activation and β-catenin nuclear localization promote adenoma progression
to carcinoma (6). Therefore, APC loss alone results in cytoplasmic
β-catenin accumulation but does not enhance β-catenin nuclear translocation. Phosphorylation of β-catenin at Ser552 by AKT plays an
important role in increased nuclear translocation and transcriptional
activity of the β-catenin/TCF axis (17). Consistently, the expression
levels of phosphorylated β-catenin at Ser552 in HCT15 and HCT116
2179
P.-L.Lin et al.
Fig. 3. MiR-21 promotes β-catenin/TCF transcriptional activity in colon cancer cells. A precursor and inhibitor of miR-21 were respectively treated with (A)
HCT15 and HCT116 for 48 h, and then western blotting was used to determine β-catenin expression in whole-cell extracts and cytoplasmic and nuclear protein
fractions. β-actin, α-tubulin and Sp1 were used as protein-loading controls of whole-cell extracts and cytoplasmic and nuclear fractions, respectively. (B)
TOPFLASH and FOPFLASH were, respectively, co-transfected with miR-21 precursor and inhibitor into HCT15 and HCT116 for 48 h. Reporter activity was
determined using luciferase reporter assay. β-gal served as an internal control. The changes of cyclin D1 mRNA expression levels by the miR-21 precursor and
inhibitor were evaluated by real-time PCR.
2180
MiR-21 enhances β-catenin nuclear translocation
Fig. 4. Phosphorylation of β-catenin at Ser552 is responsible for β-catenin nuclear translocation in colon cancer cells with mutant APC or with mutant β-catenin.
(A) Wild-type β-catenin, β-catenin S552A and/or miR-21 precursor were transfected into the HCT15 shβ-catenin stable clone, and western blotting was used
to evaluate β-catenin expression in whole-cell extracts and cytoplasmic and nuclear protein. TOPFLASH was respectively transfected into those groups for
24 h, and then the reporter activity was determined using the luciferase reporter assay. β-gal served as an internal control. The Boyden chamber invasion assay
was performed to evaluate changes of invasiveness. The invasive cells on Matrigel membranes are shown as representative pictures in the upper panel and a
quantitative graph in the lower panel. (B) Wild-type β-catenin, miR-21 precursor and miR-21 inhibitor were transfected into the shβ-catenin HCT116 stable
clone, and western blotting was used to evaluate β-catenin expression in whole-cell extracts and the cytoplasmic and nuclear fraction of cells (upper panel).
TOPFLASH was co-transfected into the shβ-catenin HCT116 stable clone with different treatments for 24 h, and then reporter activity was determined by a
luciferase reporter activity assay. β-gal served as an internal control (middle panel). The Boyden chamber invasion assay was performed to evaluate the changes
of invasiveness. Representative invasive cells on Matrigel membranes and a quantitative graph are shown in the lower panel. (C) Wild-type β-catenin, APC
shRNA and miR-21 precursor were co-transfected into the shβ-catenin HCT116 stable clone, and western blotting was used to evaluate β-catenin expression in
whole-cell extracts and the cytoplasmic and nuclear fraction of cells (upper panel). TCF-reporter activity (TOPFLASH transfection for 24 h) was evaluated by
the luciferase reporter activity assay in the shβ-catenin HCT116 stable clone with different treatments (middle panel). β-gal served as an internal control. The
Boyden chamber invasion assay was performed to evaluate the invasiveness of the shβ-catenin HCT116 stable clone with different treatments. Representative
invasive cells on Matrigel membranes are shown in the lower panel, and a quantitative graph is shown in the upper panel (bottom panel).
cells increased by miR-21 overexpression in a dose-dependent manner (Figure 3A). This may partially explain why the elevation of
nuclear β-catenin expression was observed in miR-21 overexpressing
HCT15 and HCT116 cells even though both cells harboured KRAS
mutations (29).
PTEN is targeted by miR-21 and, in turn, activates the PI3K/AKT
signalling pathway (30). β-Catenin activation synergizes with PTEN
loss to promote tumour progression and metastasis in melanomas,
bladder, endometrial and prostate cancers (31–34). Yu et al. showed
that miR-21 induced colon cancer cell stemness via activation of the
Wnt/β-catenin signalling pathway by down-regulating transforming
growth factor beta, receptor 2 (TGF-βR2) (35). Recently, an animal
model study showed that TGF-βR2 knockout alone was not sufficient
for intestinal tumour formation, and lack of PTEN alone had a weak
effect on intestinal tumour induction (36). However, the combination
of TGF-βR2 inactivation with PTEN loss led to malignant tumours
in both the small intestine and the colon in 86% of the mice and to
metastasis in 8% of the tumour-bearing mice (36). Therefore, inactivation of the TGF-β signalling pathway combined with loss of PTEN
cooperatively promotes intestinal tumour progression (36). In the present study, we provided evidence that miR-21 promoted the invasive
capability of APC-mutated colon cancer cells via increased β-catenin
nuclear localization and transcriptional activity of the β-catenin/TCF
signalling pathway. We also observed that the TCF-reporter activity
enhanced by miR-21 overexpression in HCT15 cells was restored
by NF-κB inhibitor BAY and PI3K/AKT inhibitor LY94002 but not
by ERK inhibitor AZD6244. Similarly, the TCF-reporter activity
promoted by miR-21 overexpression in HCT116 cells decreased by
BAY and LY294002 but not by AZD6244 (Supplementary Figure 6,
available at Carcinogenesis Online). These results are consistent with
previous studies showing that miR-21 expression is regulated by the
NF-κB signalling pathway (37). Phosphorylation of β-catenin at
Ser552 is regulated through AKT activation due to PTEN targeted by
miR-21 (38,39). Therefore, miR-21 activation of TCF-reporter activity might occur partially through the NF-κB/PI3K/AKT cascade.
In the present study, APC mutations in colorectal cancer patients
were detected in the mutation cluster region. The majority of APC
mutations in colorectal cancer occur in the mutation cluster region (40).
The APC mutation rate in this study population (60 of 165, 36.4%) was
lower than in other reports (~80%) (40–42). Immunohistochemical
data indicated that nuclear β-catenin expression in colorectal tumours
positively correlated with APC mutations and miR-21 expression
2181
P.-L.Lin et al.
levels (Table I). Moreover, the highest prevalence of positive nuclear
β-catenin expression was observed in colorectal tumours that had
high miR-21 levels combined with APC mutation when compared
with the other three subgroups (Table I). These results are consistent
with the cell model findings and showed that miR-21 overexpression
may promote β-catenin nuclear localization in APC-mutated cells but
not in APC-wild-type cells. Therefore, miR-21 levels may be an independent prognostic indicator in APC-mutated colorectal cancer but
not in APC-wild-type colorectal cancer. However, this observation is
not consistent with a previous report indicating that miR-21 levels
may predict a poor outcome in all colon adenocarcinoma patients;
this could be attributed to ~80% of Caucasian colon cancer patients
having APC mutations (41,42).
In summary, we demonstrated that miR-21 may promote β-catenin
nuclear translocation and β-catenin/TCF transcriptional activity and,
in turn, enhance invasiveness in APC-mutated cells but not in APCwild-type colon cancer cells. Correspondingly, APC-mutated colorectal cancer patients whose tumours have high miR-21 levels showed
poorer OS and RFS compared with those with low miR-21 expression
levels, but the prognostic value was not observed in APC-wild-type
colorectal cancer patients.
Supplementary material
Supplementary Tables 1–6 and Figures 1–6 can be found at http://
carcin.oxfordjournals.org/
Funding
National Science Council (NSC100-2314-B-038-043-MY3 and
NSC101-2632-B-040-001-MY3), Taiwan; Taipei Medical University
(TMU101-AE1-B10), Taipei, Taiwan.
Conflict of Interest Statement: None declared.
References
1.Ministry of Health and Welfare. (2012) Cause of Death Statistics in Taiwan.
Ministry of Health and Welfare, Executive Yuan, Taiwan.
2.Ko,J.M. et al. (1999) Genomic instability and alterations in Apc, Mcc and
Dcc in Hong Kong patients with colorectal carcinoma. Int. J. Cancer, 84,
404–409.
3.Christie,M. et al. (2013) Different APC genotypes in proximal and distal sporadic colorectal cancers suggest distinct WNT/β-catenin signalling
thresholds for tumourigenesis. Oncogene, 32, 4675–4682.
4.Conlin,A. et al. (2005) The prognostic significance of K-ras, p53, and APC
mutations in colorectal carcinoma. Gut, 54, 1283–1286.
5.Jass,J.R. et al. (2003) APC mutation and tumour budding in colorectal cancer. J. Clin. Pathol., 56, 69–73.
6.Phelps,R.A. et al. (2009) A two-step model for colon adenoma initiation
and progression caused by APC loss. Cell, 137, 623–634.
7.Fodde,R. et al. (2010) Nuclear beta-catenin expression and Wnt signalling:
in defence of the dogma. J. Pathol., 221, 239–241.
8.He,T.C. et al. (1998) Identification of c-MYC as a target of the APC pathway. Science, 281, 1509–1512.
9.Tetsu,O. et al. (1999) Beta-catenin regulates expression of cyclin D1 in
colon carcinoma cells. Nature, 398, 422–426.
10. Günther,K. et al. (1998) Predictive value of nuclear beta-catenin expression for the occurrence of distant metastases in rectal cancer. Dis. Colon
Rectum, 41, 1256–1261.
11. Burgess,A.W. et al. (2011) Wnt signaling and colon tumorigenesis–a view
from the periphery. Exp. Cell Res., 317, 2748–2758.
12. Hassan,A.B. et al. (2005) Colorectal cancer prognosis: is it all mutation,
mutation, mutation? Gut, 54, 1209–1211.
13. Kanaan,Z. et al. (2012) Plasma miR-21: a potential diagnostic marker of
colorectal cancer. Ann. Surg., 256, 544–551.
14. Pan,X. et al. (2010) MicroRNA-21: a novel therapeutic target in human
cancer. Cancer Biol. Ther., 10, 1224–1232.
15. Schetter,A.J. et al. (2008) MicroRNA expression profiles associated with
prognosis and therapeutic outcome in colon adenocarcinoma. JAMA, 299,
425–436.
2182
16. Ma,X. et al. (2011) Loss of the miR-21 allele elevates the expression of its
target genes and reduces tumorigenesis. Proc. Natl Acad. Sci. USA, 108,
10144–10149.
17. Fang,D. et al. (2007) Phosphorylation of beta-catenin by AKT promotes
beta-catenin transcriptional activity. J. Biol. Chem., 282, 11221–11229.
18. Cheng,Y.W. et al. (2007) Human papillomavirus 16/18 E6 oncoprotein is
expressed in lung cancer and related with p53 inactivation. Cancer Res.,
67, 10686–10693.
19. Sung,W.W. et al. (2011) A polymorphic -844T/C in FasL promoter predicts
survival and relapse in non-small cell lung cancer. Clin. Cancer Res., 17,
5991–5999.
20. Wu,D.W. et al. (2012) Loss of TIMP-3 promotes tumor invasion via
elevated IL-6 production and predicts poor survival and relapse in HPVinfected non-small cell lung cancer. Am. J. Pathol., 181, 1796–1806.
21. Wu,D.W. et al. (2010) Paxillin predicts survival and relapse in nonsmall cell lung cancer by microRNA-218 targeting. Cancer Res., 70,
10392–10401.
22. Wu,D.W. et al. (2011) Reduced p21(WAF1/CIP1) via alteration of p53DDX3 pathway is associated with poor relapse-free survival in early-stage
human papillomavirus-associated lung cancer. Clin. Cancer Res., 17,
1895–1905.
23.(1988) The regulation of cell proliferation and differentiation: from basic
science to clinical implications. Papers from a meeting. April 1987, Ulm,
West Germany. Scand J Gastroenterol Suppl., 151, 1–129.
24. Rubinfeld,B. et al. (1996) Binding of GSK3beta to the APC-beta-catenin
complex and regulation of complex assembly. Science, 272, 1023–1026.
25. Li,V.S. et al. (2012) Wnt signaling through inhibition of β-catenin degradation in an intact Axin1 complex. Cell, 149, 1245–1256.
26. Behrens,J. et al. (1998) Functional interaction of an axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science, 280, 596–599.
27. Hagen,T. et al. (2002) Expression and characterization of GSK-3 mutants
and their effect on beta-catenin phosphorylation in intact cells. J. Biol.
Chem., 277, 23330–23335.
28. Hagen,T. et al. (2002) Characterisation of the phosphorylation of betacatenin at the GSK-3 priming site Ser45. Biochem. Biophys. Res. Commun.,
294, 324–328.
29. Edkins,S. et al. (2006) Recurrent KRAS codon 146 mutations in human
colorectal cancer. Cancer Biol. Ther., 5, 928–932.
30. Meng,F. et al. (2007) MicroRNA-21 regulates expression of the PTEN
tumor suppressor gene in human hepatocellular cancer. Gastroenterology,
133, 647–658.
31. Damsky,W.E. et al. (2011) β-catenin signaling controls metastasis in Brafactivated Pten-deficient melanomas. Cancer Cell, 20, 741–754.
32. Ahmad,I. et al. (2011) β-Catenin activation synergizes with PTEN loss to
cause bladder cancer formation. Oncogene, 30, 178–189.
33.van der Zee,M. et al. (2013) Alterations in Wnt-β-catenin and Pten signalling play distinct roles in endometrial cancer initiation and progression. J.
Pathol., 230, 48–58.
34. Persad,S. et al. (2001) Tumor suppressor PTEN inhibits nuclear accumulation of beta-catenin and T cell/lymphoid enhancer factor 1-mediated transcriptional activation. J. Cell Biol., 153, 1161–1174.
35. Yu,Y. et al. (2012) MicroRNA-21 induces stemness by downregulating
transforming growth factor beta receptor 2 (TGFβR2) in colon cancer cells.
Carcinogenesis, 33, 68–76.
36. Yu,M. et al. (2014) Inactivation of TGF-beta signaling and loss of PTEN
cooperate to induce colon cancer in vivo. Oncogene, 33, 1538–1547.
37. Shin,V.Y. et al. (2011) NF-κB targets miR-16 and miR-21 in gastric cancer:
involvement of prostaglandin E receptors. Carcinogenesis, 32, 240–245.
38. Han,M. et al. (2012) Antagonism of miR-21 reverses epithelial-mesenchymal transition and cancer stem cell phenotype through AKT/ERK1/2 inactivation by targeting PTEN. PLoS One, 7, e39520.
39. Xiong,B. et al. (2013) MiR-21 regulates biological behavior through the
PTEN/PI-3 K/Akt signaling pathway in human colorectal cancer cells. Int.
J. Oncol., 42, 219–228.
40. Miyoshi,Y. et al. (1992) Somatic mutations of the APC gene in colorectal
tumors: mutation cluster region in the APC gene. Hum. Mol. Genet., 1,
229–233.
41. Bisgaard,M.L. et al. (2004) Familial adenomatous polyposis patients without an identified APC germline mutation have a severe phenotype. Gut, 53,
266–270.
42. Rowan,A.J. et al. (2000) APC mutations in sporadic colorectal tumors:
A mutational “hotspot” and interdependence of the “two hits”. Proc. Natl
Acad. Sci. USA, 97, 3352–3357.
Received November 27, 2013; revised April 3, 2014;
accepted April 18, 2014