Carcinogenesis vol.33 no.12 pp.2391–2397, 2012
doi:10.1093/carcin/bgs288
Advance Access publication September 16, 2012
Downregulation of miR-144 is associated with colorectal cancer progression via
activation of mTOR signaling pathway
Takeshi Iwaya1,2,*,†, Takehiko Yokobori3,†, Naohiro
Nishida1, Ryunosuke Kogo1, Tomoya Sudo1, Fumiaki
Tanaka1, Kohei Shibata1, Genta Sawada1, Yusuke
Takahashi1, Masahisa Ishibashi1, Go Wakabayashi2,
Masaki Mori4 and Koshi Mimori1
1
Department of Surgery, Kyushu University Beppu Hospital, 4546
Tsurumihara, Beppu 874-0838, Japan, 2Department of Surgery, Iwate
Medical University, Morioka 020-8505, Japan, 3Department of General
Surgical Science, Graduate School of Medicine, Gunma University, 3-3922 Showa-machi, Maebashi, Gunma 371-8511, Japan and 4Department of
Gastroenterological Surgery, Graduate School of Medicine, Osaka University,
2-2 Yamadaoka, Suita 565-0871, Japan
*To whom correspondence should be addressed. Tel: +81 977 1650;
Fax: +81 977 1651;
Email: [email protected]
The mammalian target of rapamycin (mTOR) is a downstream
integrator of essential pathways. mTOR signaling is frequently
dysregulated in a variety of human cancers, and in silico analysis
has revealed two miR-144 binding sites in the mTOR 3′ untranslated region. We investigated the clinicopathologic magnitude of
the mTOR pathway regulating microRNA, miR-144 in colorectal
cancer (CRC) cases. The regulation of mTOR by miR-144 was
examined with inhibitor miR-144-transfected cells. We also investigated changes in sensitivity to the mTOR inhibitor, rapamycin,
in inhibitor miR-144-transfected cells. Quantitative RT-PCR was
used to evaluate the clinicopathologic significance of miR-144
expression in 137 CRC. Furthermore, we assessed the correlation between CRC prognosis and the expression of 16 genes in the
Akt/mTOR pathway. In vitro assays showed that mTOR is a direct
target of miR-144, and downregulation of miR-144 facilitated
proliferation of CRC cell line, HT29. In addition, the viability of
HT29 cells with downregulated miR-144 expression was significantly reduced with rapamycin treatment. Low expression levels
of miR-144 were associated with enhanced malignant potential
such as venous invasion (P = 0.0013), liver metastasis (P = 0.08),
liver recurrence (P = 0.0058) and poor prognosis (P = 0.0041).
Multivariate analysis indicated that low miR-144 expression
was an independent prognostic factor for survival. Among many
genes consisting of the mTOR pathway, only high expression of
Rictor was associated with poor prognosis of CRC. miR-144 is a
meaningful prognostic marker. Downregulation of miR-144 leads
to poor prognosis of CRC patients via activation of the mTOR
signaling pathway.
Introduction
The mammalian target of rapamycin (mTOR) kinase acts downstream
of phosphoinositide 3-kinase/Akt to regulate cellular growth, metabolism and the cytoskeleton, and its signaling pathway is frequently
dysregulated in a variety of human cancers (1,2). Rapamycin and rapamycin derivatives have long been employed for immunosuppression
and, more recently, as anticancer treatments. The drug is now clinically used for the treatment of advanced renal cell cancer, and Phase
III trials are under way for breast cancer, gastric cancer, hepatocellular
Abbreviations: CI, confidence interval; CRC, colorectal cancers; LMD,
laser microdissection; miRNAs, micro RNAs; mTOR, mammalian target of
rapamycin; RR, relative risk; UPL, Universal Probe Library Probe; 3′UTR, 3′
untranslated region.
†
These authors contributed equally to this work.
carcinoma, pancreatic neuroendocrine tumors and lymphoma, following favorable results of Phase II studies (3–8).
Since 60% of colorectal cancers (CRC) exhibit high levels of activated Akt (9), the association between mTOR and CRC has also been
investigated. It has been reported that mTOR was highly activated in
glandular elements of CRC and in colorectal adenomas with highgrade intraepithelial neoplasia, with a correlation between immunohistochemical staining intensity and depth of infiltration (10). mTOR
exists in two functionally distinct complexes: mTORC1 (containing
mTOR, Raptor, etc.) and mTORC2 (containing mTOR, Rictor, etc.).
Gulhati et al. demonstrated increased expression of mTOR, Raptor
and Rictor mRNA in more advanced stages of CRC. In addition,
mTOR, Raptor and Rictor protein levels were significantly elevated
in stage IV CRCs (11). These clinical observations have led to the
association of higher grade CRC malignancies with increased expression of mTOR and its complexes. Activation of Akt/mTOR signaling
is also frequently observed in CRCs, and mTOR inhibitors have an
anti-proliferative effect in several CRC cell lines (11,12). Therefore,
mTOR kinase inhibitors, such as rapamycin, have been investigated as
part of the therapeutic regimen for CRC patients. In order to predict
the efficacy of rapamycin for CRC patients, it is important to know
the value of mTOR expression status as a prognostic or susceptibility
marker in clinical settings of CRC.
Micro RNAs (miRNAs) are 19- to 25mer non-coding RNAs that
incompletely bind the 3′ untranslated region (UTR) of multiple target mRNAs, enhancing their degradation and inhibiting translation.
miRNAs possess normal biological functions, such as regulation of
proliferation, differentiation and apoptosis. Moreover, dysregulation
of miRNAs plays a critical role in carcinogenesis and cancer progression (13). Many miRNAs are present at lower levels in cancer tissue
than in normal tissue, a state that contributes to cancer progression
(14). In silico analysis of miRNA-target mRNA prediction algorithm
(TargetScan 6.0; http://www.targetscan.org/) revealed two miR-144
binding sites in the mTOR 3′UTR region with perfect Watson–Crick
matches at miRNA positions 1–7 and 2–8 (Supplementary Figure 1A,
available at Carcinogenesis Online). These sites raise the possibility
that miR-144 is involved in the mTOR signaling pathway. In this study,
we evaluated miR-144 expression status, its clinical significance in
CRC and the association between miR-144 and the mTOR pathway.
Materials and methods
Experimental studies
Cell lines and cell culture. The human colorectal cancer cell lines, HT29 and
CaR-1, were obtained from the American Type Culture Collection and the
Japanese Collection of Research Bioresources. These cell lines were maintained in RPMI 1640 containing 10% fetal bovine serum with 100 U/ml penicillin and 100 U/ml streptomycin sulfates and cultured in a humidified 5% CO2
incubator at 37°C.
Evaluation of mTOR mRNA expression in colorectal cancer cells. For RNA
analysis, each cell line was seeded at 1 × 105 cells per well in a volume of
2 ml in 6-well flat-bottomed microtiter plates. Total RNA from these cell lines
was isolated using the mirVana miRNA Isolation Kit (Ambion) following 48 h
incubation. Quantitative real-time reverse transcriptase PCR (qRT-PCR) was
performed with the Universal Probe Library Probe (UPL; Roche Diagnostics)
to measure mTOR mRNA expression. Primer sequences corresponding to UPL
and RT-PCR protocols were described previously (15).
Transfection of miR-144 inhibitor (Anti-miR-144). Either Anti-miR-144 or
Anti-miR negative control (Ambion miRNA Inhibitors, Applied Biosystems)
was transfected into HT29 cells at 30 nmol/l (final concentration) using
Lipofectamine RNAiMAX (Invitrogen Life Technologies) according to the
manufacturer’s instructions.
Cell proliferation and viability assays. Logarithmically growing HT29
cells were transfected with Anti-miR-144 or Anti-miR negative control and
seeded at 4.0 × 103 cells per well in 96-well flat-bottomed microtiter plates in
© The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
2391
T.Iwaya et al.
a final volume of 100 μl of culture medium per well. Cells were incubated in
a humidified atmosphere (37°C and 5% CO2) for 24, 48 and 72 h after initiation of transfection. 3-(4,5-Dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium
bromide assays were used to measure cell proliferation during each time
period, as described previously (15). To evaluate changes in sensitivity to
an mTOR inhibitor, the cells were seeded in 96-well plates and exposed to
rapamycin (100 nM to 1.0 μM) for 72 h, and then cell number were measured using the CellTiter-Glo Luminescent assay (Promega) according to the
manufacturer’s instructions, and luminescence was recorded with a luminometer (BioTek FLx800, BioTek Instruments, Inc). The number of cells
used per experiment is determined empirically. These assays were carried
out with six replicates.
Construction of reporter plasmids and the luciferase reporter assay. To
construct a luciferase reporter plasmid, a full length fragment of the mTOR
3′UTR or 580 bp region of Rictor 3′UTR including predictive miR-144 binding site (Supplementary Figure 1C, available at Carcinogenesis Online) was
subcloned into pmirGlo Dual-luciferase miRNA Target Expression Vector
(Promega) located 5′ to the firefly luciferase. The nucleotide sequences of
the constructed plasmids were confirmed by DNA sequencing. For luciferase
reporter assays, HT29 or CaR-1 cells were seeded in a 96-well plate and then
cotransfected with the pmirGlo-mTOR 3′UTR (or Rictor 3′ UTR) construct
and miR-144 (Pre-miR-144) or Pre-miR negative control (Ambion). Assays
were performed 48 h after transfection using the Dual-Luciferase Reporter
Assay System (Promega). The firefly luciferase signals were normalized to
the Renilla luciferase signals. Transfections were repeated three times in
independent experiments.
Protein expression analysis. Western blots were used to confirm mTOR
expression in Anti-miR-144-transfected cells. The following primary antibodies and dilutions were used: mTOR rabbit monoclonal antibody (Cell Signaling
Technology, Inc) at 1:1000 and Beta-tubulin rabbit monoclonal antibody (Cell
Signaling Technology, Inc) at 1:1000. Detailed protocols are described previously (15).
Clinical cases
Patients and sample collection. A total of 280 CRC samples [137 were used in
bulk (set 1), 86 used as pure tissues separated by laser microdissection (LMD)
(set 2), and 145 were also used as bulk (set 3) in which 88 samples overlapped
with (set 1)] obtained during surgery were used after obtaining informed consent. These sets of CRC cases and assays performed in each set were schematized in the flow chart (Supplementary Figure 2, available at Carcinogenesis
Online). All patients underwent resection of the primary tumor at Kyushu
University Hospital at Beppu and affiliated hospitals between 1992 and 2007.
Written informed consent was obtained from all patients. All patients had a
clear histologic diagnosis of colorectal cancer and were closely followed every
3 months. The follow-up periods ranged from 0.1 to 12.3 years, with a mean of
3.8 years in set 1; from 0.1 months to 3.2 years, and a mean of 2.1 years in set
2; and from 0.1 to 15.4 years, and a mean of 3.6 years in set 3. Resected cancer
tissues were immediately cut and stored in RNAlater (Ambion) or embedded
in Tissue-Tek OCT (Optimum Cutting Temperature) medium (Sakura), frozen in liquid nitrogen and kept at –80°C until RNA extraction. Frozen tissue
specimens were homogenized in guanidium thiocyanate, and total RNA was
obtained by ultracentrifugation through a cesium chloride cushion. cDNA was
synthesized from 8.0 µg of total RNA with M-MLV Reverse Transcriptase
(Invitrogen, Carlsbad, CA). Clinicopathologic factors and clinical stage were
classified using both Duke’s and TNM system of classification. All sample
data, including age, gender, histology, tumor depth, lymph node metastasis,
lymphatic invasion, vascular invasion, liver metastasis, Duke’s classification
and postoperative liver recurrence, were obtained from the clinical and pathologic records (Table I and Supplementary Table 1, available at Carcinogenesis
Online).
Evaluation of miR-144 expression in clinical samples. miR-144 expression was assayed using qRT-PCR in 137 CRC patients (set1). cDNA was
synthesized from 10 ng of total RNA using TaqMan MicroRNA hsa-miR144-specific primers (Applied Biosystems) and a TaqMan Micro-RNA
Reverse Transcription Kit (Applied Biosystems) (15). Relative quantification
of miRNA expression was calculated using the 2-ΔΔCt method. The raw data
were presented as the relative quantity of target miRNA, normalized with
respect to RNU6B and compared with a reference sample.
Survival analyses using microarray data. Using our previous expression
array data from LMD samples obtained from 86 CRC patients (set 2), we
examined the association between the expression status of the mTOR signaling pathway genes and CRC prognosis. Gene expression arrays have been
deposited in the National Center for Biotechnology Information (NCBI)
Gene Expression Omnibus (GEO) database with accession code GSE21815.
Detail information of LMD and cDNA microarray was described in the
Supplementary Information, available at Carcinogenesis Online.
2392
Table I. miR-144 expression and clinicopathologic factors
Factors
Age (mean ± SD)
Gender
Male
Female
Histological gradea
Well
Others
Depth of tumor invasionb
m, sm, mp, ss
se, si
Lymph node metastasis
Absent
Present
Lymphatic invasion
Absent
Present
Venous invasion
Absent
Present
Liver metastasis
Absent
Present
Dukes classification
A, B
C, D
Liver recurrence
(Dukes A, B, C)
Absent
Present
miR-144/RNU6B
High expression
Low expression
P value
n = 68
n = 69
66.5 ± 10.4
65.9 ± 12.3
0.79
40
28
43
26
0.68
26
42
30
39
0.53
48
20
54
15
0.3
41
27
38
31
0.53
41
27
42
27
0.94
64
4
51
18
0.0013**
62
6
56
13
0.08
40
28
33
36
0.2
58
2
43
10
0.0058**
a
Well: well-differentiated type, Others: moderately differentiated, poorly
differentiated and mucinous type.
b
Tumor invasion of mucosa (m), submucosa (sm), muscularis propria (mp),
subserosa (ss), penetration of serosa (se) and invasion of adjacent structures (si).
*P < 0.05, **P < 0.01.
Evaluation of Rictor expression in clinical samples. qRT-PCR was performed with the UPL (Roche Diagnostics) to measure Rictor mRNA expression in 145 CRC samples (set 3). Primer sequences corresponding to UPL and
RT-PCR protocols were described previously (15).
Statistical analysis
Data from the qRT-PCR analysis and in vitro-transfected cell assays were
analyzed with JMP 5. Overall survival rates were calculated actuarially
according to the Kaplan–Meier method and were measured from the day
of surgery. Differences between groups were estimated using the chi-square
test, Student’s t-test, repeated-measures analysis of variance test and the
log-rank test. Variables with a value of P < 0.05 in univariate analysis were
used in a subsequent multivariate analysis based on the Cox proportional
hazards model for survival and logistic regression model for postoperative
liver recurrence. All differences were statistically significant at the level of
P < 0.05.
Results
Repression of miR-144 facilitates proliferation of colorectal cancer
cells via mTOR upregulation
In silico analysis of miRNA-target mRNA prediction algorithm
(TargetScan 6.0) revealed several miRNA families that targeted mTOR
and broadly conserved among vertebrates. Among these families, only
miR-144 singly had two binding sites in the mTOR 3′UTR region
(Supplementary Figures 1A and 3, available at Carcinogenesis Online).
An mTOR 3′UTR–luciferase construct was generated to determine if
the mTOR gene is a direct target of miR-144. Cotransfectants expressing both miR-144 and the mTOR 3′UTR showed a significant reduction in luciferase activity (P < 0.05; Student’s t-test) compared with
control HT29 and CaR-1 cells (Figure 1A). Furthermore, to confirm
miR-144 expression in colorectal cancer
Fig. 1. (A) Luciferase analysis in HT29 and CaR-1. mTOR 3′UTR luciferase vector + miR-144 transfectants showed lower luciferase activities than did control
cells. Relative luciferase level: (Sample Luc/Sample Renilla)/(Control Luc/Control Renilla). Luc, raw firefly luciferase activity; Renilla, Renilla activity (internal
transfection control); Pre-miR n.c., Pre-miR negative control. (B) miR-144 (left) and mTOR mRNA (right) expression after treatment with a negative control or
Anti-miR-144 in HT29 cells (qRT-PCR). (C) Western blot analysis of mTOR in HT29 cells transfected with Anti-miR-144 or a negative control (n.c.). Proteins
were normalized to the level of beta-tubulin. (D) Left, proliferation rates after treatment with Anti-miR-144 in HT29. n.c., negative control; Right, enhancement
of rapamycin sensitivity in HT29 cells repressed miR-144 expression. Anti-miR-144 or Anti-miR negative control were transfected with or without rapamycin
(0.1, 0.5 and 1.0 μM) treatment and cell growth was measured after a 72 h incubation. Absorbance at 0 h was assigned a value of 1. Figure 1A, 1B and 1D: The
error bar represents the standard deviation from six replicates. (E) Kaplan–Meier overall survival curves of colorectal cancer patients based on the level of miR144 expression.
the direct interaction between miR-144 and its binding sites on mTOR
3′UTR, we also constructed a luciferase reporter plasmid in which the
region included two miR-144 binding sites were deleted. The activity of the reporter construct without miR-144 binding sites was unaffected by simultaneous transfection with Pre-miR-144 (Supplementary
Figure 4, available at Carcinogenesis Online). Reduction of miR-144
expression by Anti-miR-144 transfection induced an elevation in mTOR
mRNA expression and protein in HT29 cells (Figure 1B and 1C).
Simultaneously, downregulation of miR-144 also facilitated cellular
capacity for proliferation (Figure 1D, left). Unlike HT29, downregulation of miR-144 with Anti-miR-144 transfection did not facilitate cellular
capacity for proliferation in other CRC cells, such as neither RKO nor
SW480 (Supplementary Figure 5A, available at Carcinogenesis Online).
Endogenous miR-144 expression was observed much higher in HT29
cells than RKO/SW480 cells (Supplementary Figure 5B, available at
Carcinogenesis Online). Therefore, we speculated that the capability
of the cellular proliferation by anti-miR-144 might be only affected in
HT29 strongly expressed miR-144, however, RKO and SW480 with low
endogenous miR-144 did not change cellular capacity for proliferation
by downregulation of miR-144.
miR-144 knockdown increased rapamycin sensitivity in HT29 cells
We evaluated changes in sensitivity to an mTOR inhibitor, rapamycin, in HT29 cells transfected with Anti-miR-144 and compared with
parent or control cells. Luminescent cell viability assays were performed to evaluate the growth inhibitory effect of rapamycin in HT29.
Downregulation of miR-144 expression with rapamycin significantly
reduced cell viability (Figure 1D, right); indicating that CRC cells
with simultaneous mTOR overexpression and low miR-144 expression
might be more sensitive to rapamycin.
Clinicopathologic significance of miR-144 mRNA expression in
colorectal cancers
In this study, patients with expression levels of miR-144 that were below
the median value of 0.2688 (normalized to RNU6B) were assigned to
the low expression group (n = 69), whereas those with expression
values above the median were assigned to the high expression group
(n = 68). Clinicopathologic factors were analyzed in relation to miR144 levels (Table I). The miR-144 low expression group showed more
frequent venous invasion (P = 0.0013), postoperative liver recurrence
(P = 0.0058) and a tendency of frequent liver metastases at the time
of operation (P < 0.1) than the high expression group (chi-square
test). However, no significant differences were observed regarding
age, gender, histology, tumor depth, lymphatic invasion, lymph node
metastasis or Duke’s classification (Table I). In the overall survival
curve, patients in the miR-144 low expression group (5-year survival
rate, 53.2%) had a significantly poorer prognosis than those in the
miR-144 high expression group (81.5%, P = 0.0041; log-rank test;
Figure 1E). Univariate analysis of overall survival revealed that
the relative level of miR-144 expression, tumor depth, lymph node
metastasis, lymphatic/venous invasion and liver metastasis were
prognostic predictors. Variables with a P < 0.05 were selected for
multivariate analysis. Multivariate analysis showed that the level of
miR-144 expression was an independent prognostic predictor [relative
risk (RR): 2.60, 95% confidence interval (CI): 1.17–6.01, P = 0.019;
Cox hazard proportional model; Table II]. Univariate analysis of
postoperative liver recurrence in Dukes ABC patients revealed that
the relative level of miR-144 expression, lymph node metastasis and
venous invasion were predictors. Multivariate analysis showed that
only the level of miR-144 expression was an independent variables
that could predict liver recurrence (RR: 5.59, 95% CI: 1.27–39.35,
P = 0.0391; logistic regression model; Table III). Furthermore, in
2393
T.Iwaya et al.
Table II. Univariate and multivariate analysis for overall survival (Cox proportional hazard model)
Clinicopathologic variable
Gender (male/female)
Histology gradea (well/others)
Depthb (m, sm, mp, ss/se, si)
Lymph node metastasis (negative/positive)
Lymphatic invasion (negative/positive)
Venous invasion (negative/positive)
Liver metastasis (negative/positive)
miR-144 expression (high/low)
Univariate analysis
Multivariate analysis
RR
95% CI
P value
RR
95% CI
P value
0.93
0.83
2.71
2.73
2.21
1.70
2.45
2.28
0.65–1.29
0.58–1.16
1.93–3.86
1.86–4.32
1.57–3.23
1.15–2.42
1.71–3.43
1.16–4.73
0.67
0.28
<0.0001***
<0.0001***
<0.0001***
0.009**
<0.0001***
0.016*
—
—
1.96
2.00
1.51
1.19
1.58
2.60
—
—
1.29–3.01
1.31–3.23
1.03–2.3
0.77–1.8
1.03–2.42
1.17–6.01
—
—
0.0018**
0.0011**
0.0345*
0.43
0.037*
0.019*
RR, relative risk; CI, confidence interval. aWell: well-differentiated type, Others: moderately differentiated, poorly differentiated and mucinous type.
b
Tumor invasion of mucosa (m), submucosa (sm), muscularis propria (mp), subserosa (ss), penetration of serosa (se) and invasion of adjacent structures (si)
*P < 0.05, **P < 0.01, ***P < 0.001.
Dukes BC patients, only the miR-144 expression status was predictors
for liver recurrence in univariate analysis (RR: 6.67, 95% CI: 1.594–
45.738, P = 0.020; Table IV).
Association between expression of the mTOR signaling pathway
genes and CRC prognosis
We demonstrated that downregulation of miR-144 expression status
exerted an influence on the prognosis of CRC patients, and that miR144 directly targeted mTOR and repressed its expression. However,
it has not been demonstrated that mTOR expression can function as a
prognostic marker in CRC patients. Thus, we examined the association between the expression status of 16 genes in the mTOR signaling
pathway and CRC prognosis using previous expression array data
from LMD samples obtained from another set of 86 CRC patients
(set 2). Upstream molecules of mTOR (PDK1, Akt, TSC1, TSC2 and
Rheb), common molecules in the two mTOR complexes (mTOR,
MLST8 and Deptor), specific molecules in mTORC1 (Raptor) and
mTORC2 (Rictor, PRR5 and MAPKAP1), and downstream targets of
mTORC (P70S6K, 4EBP1, RhoA and Rac1) were evaluated. Patients
with values less than the median expression level of each molecule
were assigned to a low expression group (n = 43), and those with
expression values above the median were assigned to a high expression group (n = 43). Of these molecules, only Rictor expression was
associated with CRC prognosis. Patients in the high Rictor expression group had a significantly poorer prognosis than those in the
low Rictor expression group (P = 0.0032; log-rank test; Figure 2A).
None of the other molecules under examination were associated with
CRC prognosis (Figure 2A). Expression status was not associated
with CRC prognosis at any other cut off levels in almost all of the
mTOR pathway genes with the exception of Rictor (Supplementary
Table III. Univariate and multivariate analysis for liver recurrence in Dukes ABC (logistic regression model)
Clinicopathologic variable
Gender (male/female)
Histology gradea (well/others)
Depthb (m, sm, mp, ss/se, si)
Lymph node metastasis (negative/positive)
Lymphatic invasion (negative/positive)
Venous invasion (negative/positive)
miR-144 expression (high/low)
Univariate analysis
Multivariate analysis
RR
95% CI
P value
RR
95% CI
P value
0.49
0.41
1.54
4.12
3.02
5.30
6.74
0.10–1.75
0.09–1.48
0.32–5.76
1.21–16.36
0.90–10.90
1.38–19.44
1.67–45.36
0.3029
0.2062
0.548
0.0289*
0.0763
0.0117*
0.0171*
—
—
—
3.82
—
2.89
5.59
—
—
—
1.03–16.14
—
0.67–11.75
1.27–39.35
—
—
—
0.0506
—
0.1394
0.0391*
RR, relative risk; CI, confidence interval.
a
Well: well-differentiated type, Others: moderately differentiated, poorly differentiated and mucinous type. bTumor invasion of mucosa (m), submucosa (sm),
muscularis propria (mp), subserosa (ss), penetration of serosa (se) and invasion of adjacent structures (si).
*P < 0.05.
Table IV. Univariate analysis for liver recurrence in Dukes BC (logistic regression model)
Clinicopathologic variable
Gender (male/female)
Histology gradea (well/others)
Depthb (m, sm, mp, ss/se, si)
Lymph node metastasis (negative/positive)
Lymphatic invasion (negative/positive)
Venous invasion (negative/positive)
miR-144 expression (high/low)
Univariate analysis
RR
95% CI
P value
0.62
0.62
0.83
1.82
1.64
3.79
6.67
0.128–2.332
0.128–2.332
0.171–3.172
0.512–7.383
0.474–6.073
0.960–14.454
1.594–45.738
0.5068
0.5068
0.8008
0.3667
0.4372
0.0501
0.0200*
RR, relative risk; CI, confidence interval.
a
Well: well-differentiated type, Others: moderately differentiated, poorly differentiated and mucinous type.
b
Tumor invasion of mucosa (m), submucosa (sm), muscularis propria (mp), subserosa (ss), penetration of serosa (se) and invasion of adjacent structures (si).
*P < 0.05.
2394
miR-144 expression in colorectal cancer
Fig. 2. (A) Kaplan–Meier overall survival curves of 86 colorectal cancer patients (set 2) based on the expression level of mTOR signaling pathway genes,
using expression array data from LMD samples. (B) Kaplan–Meier overall survival curves of 145 colorectal cancer patients (set 3) based on the level of Rictor
expression, using qRT-PCR data from bulk samples. (C) Luciferase analysis in HT29. Rictor 3′UTR luciferase vector + miR-144 transfectants showed lower
luciferase activities than did control cells. Relative luciferase level: (Sample Luc/Sample Renilla)/(Control Luc/Control Renilla). Luc, raw firefly luciferase
activity; Renilla, Renilla activity (internal transfection control); Pre-miR n.c., Pre-miR negative control.
Figure 6, available at Carcinogenesis Online). In the validation study
of 145 CRC (set 3), patients in the high Rictor expression group
(n = 73; 5-year survival rate, 64.8%) had a poorer prognosis than
those in the low Rictor expression group (n = 72; 82.7%, P = 0.0438;
log-rank test; Figure 2B). Univariate analysis of overall survival
revealed that the level of Rictor expression, histology grade, tumor
depth, lymph node metastasis, lymphatic/venous invasion and liver
metastasis were prognostic predictors. Multivariate analysis showed
that only the liver metastasis was an independent prognostic predictor (Supplementary Table 2, available at Carcinogenesis Online).
As well as 3′UTR region of mTOR, cotransfectants expressing both
miR-144 and the Rictor 3′UTR showed a significant reduction in
luciferase activity (P < 0.05; Student’s t-test) compared with control
HT29 (Figure 2C).
Discussion
In this study, we focused on the clinical significance of miR-144
expression and the association between miR-144 and mTOR pathway
in CRC. The gene encoding miR-144 is located on chromosome 11,
and encompasses a non-coding transcriptional unit that also encodes
miR-451. miR-144 has a passenger strand (miR-144*) as well as many
other miRNAs, whereas miR-451 span the loop region and has no passenger strand. Kalimutho et al. revealed that significant upregulation
of miR-144* in both feces and tissue samples from CRC patients and
proposed that miR-144* represented a novel fecal-based diagnostic
marker for CRC screening (19). However, the expression status of
miR-144 in solid cancers, including CRC, has not been well characterized. miR-144 and miR-451 form a miRNA cluster with robust
expression in erythroid cells, such that miR-144/451 deficiency results
in erythroid hyperplasia, ineffective erythropoiesis and mild anemia.
Within the cluster, however, miR-144 does not play a pivotal role in
terminal erythropoiesis and may possess an as yet unknown function
(20). Since the maturation mechanism of miR-144 differs from that
of miR-451 in erythroid cells (21), it is possible that the two miRNAs may also have different functions in other tissues. Indeed, unlike
miR-144, the expression status of miR-451 is not associated with
CRC prognosis in our qRT-PCR analysis of 95 CRC tissue samples
(Supplementary Figure 7, available at Carcinogenesis Online) and
mTOR pathway genes had no predictive miR-451 binding sites in their
3′UTR regions by in silico analysis. For these reasons, we particularly carried out detailed analysis of miR-144, and demonstrated that
downregulation of miR-144 is associated with poor prognosis in CRC
patients. Several reports have demonstrated an association between
dysregulation of miR-144 and malignancies. Wan et al. carried out
a comprehensive meta-analysis of miRNA expression microarray
datasets from 28 published tumor studies that included 1843 cancer
samples and 1097 non-cancer samples including 16 cancer types and
corresponding non-tumor controls, and 879 miRNAs were investigated in more than three studies (22). This analysis revealed 52 dysregulated miRNAs consisting of 29 downregulated miRNAs and 23
upregulated miRNAs in cancers. miR-144 was widely downregulated
in 30 studies of various cancer types. Although CRCs were also evaluated in three studies included in the meta-analysis, downregulation
2395
T.Iwaya et al.
of miR-144 in CRC compared with non-cancer samples was not
demonstrated. Data from the this study showing lower expression of
miR-144 in advanced CRC suggest that downregulation of miR-144 is
involved in the progression of CRC.
Although miR-144 is apparently repressed in a variety of cancers, the
clinical significance of miR-144 expression status in malignancies is not
fully understood, nor had the target mRNA been identified. Sureban
et al. have proposed that DCAMKL-1, a microtubule-associated kinase
expressed in postmitotic neurons, negatively regulates miR-144 and
that Notch-1 is one of the downstream target of miR-144 in a human
pancreatic cancer cell line (23). In this study, in silico analysis found
two miR-144 binding sites in the 3′UTR of mTOR with perfect matches
at miRNA positions 2–8 and 1–7 (mTOR 3′UTR positions 130–136
and 780–786, respectively; Supplementary Figure 1A, available at
Carcinogenesis Online). This study presents the first evidence for a
direct interaction between the mTOR 3′UTR and miR-144 in CRC cell
lines using a luciferase reporter assay. Because many genes are involved
in the AKT/mTOR signaling pathway, it is possible that miR-144
regulates not only mTOR but also other targets in the pathway, which
transforms CRC cells into a more malignant phenotype. The association
between mTOR and other molecules in the mTOR signaling pathway,
and prognosis of CRC patients was evaluated using our previous
microarray expression data, in which tumor samples were purified by
LMD in 85 CRC patients with an average follow-up period of 2.1 years.
The majority of the evaluated genes in the mTOR signaling pathway,
including mTOR, were not associated with CRC prognosis (Figure 2A).
Only Rictor expression was associated with CRC patient prognosis.
Patients in the high Rictor expression group significantly had a poorer
prognosis than those in the low expression group (Figure 2 A and B).
Rictor is a major component of mTORC2. Disruption of mTORC2 by
targeting Rictor, via small interfering RNA, inhibited human coronary
artery endothelial cell migration (24). The same could not be said for
mTORC1 when Raptor was targeted. Masri et al. have also reported
that glioma cell lines and tissues exhibited Rictor overexpression,
which resulted in elevated mTORC2 activity and promoted anchorageindependent growth, cellular motility and in vivo growth (25). Therefore,
high Rictor expression may influence poorer prognosis by contributing
to metastasis of CRC. Different from the result of miR-144, multivariate
analysis showed that the level of Rictor expression could not be an
independent prognostic predictor (Supplementary Table 2, available at
Carcinogenesis Online). Interestingly, Rictor has a predicted miR-144
binding site in its 3′UTR region according to the in silico miRNA-target
mRNA prediction analysis as well as mTOR (Supplementary Figure 1B,
available at Carcinogenesis Online), and we also demonstrated a
direct interaction between the Rictor 3′UTR and miR-144 (Figure 2C).
Furthermore, it was recently shown that mTORC1 and mTORC2
regulate the epithelial–mesenchymal transition in CRC by regulating
actin cytoskeleton rearrangements involved in CRC motility via RhoA
and Rac1 signaling (11). RhoA and Rac1 also contain predicted miR-144
binding sites in their 3′UTR regions (Supplementary Figure 1B, available
at Carcinogenesis Online). These results suggest that downregulation of
miR-144 is associated with poor CRC prognosis by dysregulating both
mTOR and other molecules in the mTOR pathway such as Rictor, RhoA,
and Rac1 and miR-144 is a meaningful prognostic marker for CRC.
According to the result of the association between miR-144 expression status and clinicopathologic factors, downregulation of miR-144
may be involved in the hematogenous liver metastasis and recurrence
(Table II). Multivariate analysis of liver recurrence in Dukes ABC
patients revealed that only the level of miR-144 expression was an
independent variable that could predict liver recurrence (Table III).
Liver recurrence (n = 12) was observed in Dukes BC patients and not
in Dukes A. In Dukes BC patients, only the miR-144 expression status
was predictors for liver recurrence in univariate analysis (Table IV).
The benefit of chemotherapy for CRC patients at stage Dukes C is
accepted (26), whereas its role in patients at stage Dukes B remains
controversial (27,28). Thus, the expression status of miR-144 may be
one of the useful indicators for the adjuvant therapy (e.g. chemotherapy) for CRC patients because of determining individual risk of liver
recurrence.
2396
Finally, we assessed the change of rapamycin sensitivity in HT29
cells transfected with Anti-miR-144. The transfectants were more sensitive to rapamycin compared with control cells, although the proliferation of CRC cells was facilitated by downregulation of miR-144.
This indirect evidence of an association between miR-144 and mTOR
supported the direct interaction observed with the luciferase assay.
Since additive effect by combination use of rapamycin and miR-144
for CRC treatment will be expected, we also evaluated the changes
in CRC cells by overexpression of miR-144. Unexpectedly, overexpression of miR-144 by Pre-miR-144 transfection could not suppress
mTOR expression and the proliferation rate of CRC cells. Moreover,
changes in sensitivity to rapamycin were not also observed in PremiR-144-transfected CRC cells (Supplementary Figure 8, available at
Carcinogenesis Online). These results indicated that exogenous miR144 could not repress CRC cell growth via mTOR pathway with or
without rapamycin. In contrast, low miR-144 expression cancer cells
may be more sensitive to rapamycin treatment. Although the efficacy
of rapamycin has been demonstrated for several types of metastatic
cancers, a sensitive marker for rapamycin has not been established.
Further studies about the correlation between miR-144 expression status and rapamycin sensitivity are warranted.
This study demonstrated that downregulation of miR-144 led to
poor prognosis of CRC patients via activation of the mTOR signaling
pathway. Among mTOR pathway genes, only the level of miR-144
expression was an independent prognostic predictor. Our results indicate that miR-144 can be a useful prognostic marker and an indicator
for adjuvant therapy in CRC patients.
Supplementary material
Supplementary Tables 1 and 2, Figures 1–8 and supplementary information can be found at http://carcin.oxfordjournals.org/
Funding
This work was supported by JSPS KAKENHI Grant Number
23591937 and 38th Daiwa Securities Health Foundation Grant.
Conflict of Interest Statement: None declared.
References
1.Vivanco,I. et al. (2002) The phosphatidylinositol 3-kinase AKT pathway in
human cancer. Nat. Rev. Cancer, 2, 489–501.
2.Guertin,D.A. et al. (2007) Defining the role of mTOR in cancer. Cancer
Cell, 12, 9–22.
3.Motzer,R.J. et al. (2008) Efficacy of everolimus in advanced renal cell
carcinoma: a double-blind, randomised, placebo-controlled phase III trial.
Lancet, 372, 449–456.
4.Ellard,S.L. et al. (2009) Randomized phase II study comparing two schedules of everolimus in patients with recurrent/metastatic breast cancer:
NCIC Clinical Trials Group IND.163. J. Clin. Oncol., 27, 4536–4541.
5.Doi,T. et al. (2010) Multicenter phase II study of everolimus in patients
with previously treated metastatic gastric cancer. J. Clin. Oncol., 28,
1904–1910.
6.Yao,J.C. et al. (2011) Everolimus for advanced pancreatic neuroendocrine
tumors. N. Engl. J. Med., 364, 514–523.
7.Zhu,A.X. et al. (2011) Phase ½ study of everolimus in advanced hepatocellular carcinoma. Cancer, 117, 5094–5102.
8.Witzig,T.E. et al. (2011) A phase II trial of the oral mTOR inhibitor everolimus in relapsed aggressive lymphoma. Leukemia, 25, 341–347.
9.Roy,H.K. et al. (2002) AKT proto-oncogene overexpression is an early
event during sporadic colon carcinogenesis. Carcinogenesis, 23, 201–205.
10. Zhang,Y.J. et al. (2009) mTOR signaling pathway is a target for the treatment of colorectal cancer. Ann. Surg. Oncol., 16, 2617–2628.
11. Gulhati,P. et al. (2011) mTORC1 and mTORC2 regulate EMT, motility,
and metastasis of colorectal cancer via RhoA and Rac1 signaling pathways.
Cancer Res., 71, 3246–3256.
12. Gulhati,P. et al. (2009) Targeted inhibition of mammalian target of rapamycin signaling inhibits tumorigenesis of colorectal cancer. Clin. Cancer Res.,
15, 7207–7216.
miR-144 expression in colorectal cancer
13. Calin,G.A. et al. (2006) MicroRNA signatures in human cancers. Nat. Rev.
Cancer, 6, 857–866.
14. Volinia,S. et al. (2006) A microRNA expression signature of human solid
tumors defines cancer gene targets. Proc. Natl. Acad. Sci. U.S.A., 103,
2257–2261.
15.Nishida,N. et al. (2011) MicroRNA-125a-5p is an independent prognostic
factor in gastric cancer and inhibits the proliferation of human gastric cancer
cells in combination with trastuzumab. Clin. Cancer Res., 17, 2725–2733.
16. Nishida,K. et al. (2005) Global analysis of altered gene expressions during
the process of esophageal squamous cell carcinogenesis in the rat: a study
combined with a laser microdissection and a cDNA microarray. Cancer
Res., 65, 401–409.
17. Mori,M. et al. (2002) Analysis of the gene-expression profile regarding the
progression of human gastric carcinoma. Surgery, 131(1 suppl.), S39–S47.
18.Quackenbush,J. (2002) Microarray data normalization and transformation.
Nat. Genet., 32(suppl.), 496–501.
19. Kalimutho,M. et al. (2011) Differential expression of miR-144* as a novel
fecal-based diagnostic marker for colorectal cancer. J. Gastroenterol., 46,
1391–1402.
20. Rasmussen,K.D. et al. (2010) The miR-144/451 locus is required for erythroid homeostasis. J. Exp. Med., 207, 1351–1358.
21. Cheloufi,S. et al. (2010) A dicer-independent miRNA biogenesis pathway
that requires Ago catalysis. Nature, 465, 584–589.
22. Wang,W. et al. (2011) Human tumor microRNA signatures derived from
large-scale oligonucleotide microarray datasets. Int. J. Cancer, 129,
1624–1634.
23. Sureban,S.M. et al. (2011) DCAMKL-1 regulates epithelial-mesenchymal
transition in human pancreatic cells through a miR-200a-dependent mechanism. Cancer Res., 71, 2328–2338.
24. Moss,S.C. et al. (2010) Rapamycin regulates endothelial cell migration
through regulation of the cyclin-dependent kinase inhibitor p27Kip1. J.
Biol. Chem., 285, 11991–11997.
25. Masri,J. et al. (2007) mTORC2 activity is elevated in gliomas and promotes growth and cell motility via overexpression of rictor. Cancer Res.,
67, 11712–11720.
26. Allegra,C. et al. (2007) End points in advanced colon cancer clinical trials:
a review and proposal. J. Clin. Oncol., 25, 3572–3575.
27.Chang,G.J. et al. (2007) Lymph node evaluation and survival after curative
resection of colon cancer: systematic review. J. Natl. Cancer Inst., 99, 433–441.
28. Mamounas,E. et al. (1999) Comparative efficacy of adjuvant chemotherapy
in patients with Dukes’ B versus Dukes’ C colon cancer: results from four
National Surgical Adjuvant Breast and Bowel Project adjuvant studies
(C-01, C-02, C-03, and C-04). J. Clin. Oncol., 17, 1349–1355.
Received April 2, 2012; revised September 6, 2012; accepted September 7,
2012
2397