Altered miRNA expression in high-fat diet-induced prostate cancer progression
Altered miRNA expression in high-fat diet-induced prostate cancer progression
Taketoshi Nara 2
Shintaro Narit 2
Huang Mingguo 1 2
Toshiaki Yoshioka 0
Atsushi Koizumi 1 2
Kazuyuki Numakura 2
Hiroshi Tsuruta 2
Atsushi Maeno 2
Mitsuru Sait o 2
Takamitsu Inoue 1 2
Norihiko Tsuchiya 3
Shigeru Satoh 2
Tomonori Habuchi 1 2
0 Department of Urology, Yamagata University School of Medicine , Yamagata, 990-9585 Japan
1 Department of Molecular Pathology and Tumor Pathology, Akita University School of Medicine , 1-1-1 Hondo, Akita, 010-8543 Japan
2 Department of Urology
3 AMED-CREST, Japan Agency for Medical Research and Development , AMED
Recent evidence suggests that a high-fat diet (HFD) plays an important role in prostate carcinogenesis; however, underlying mechanisms largely remain unknown. Here, we investigated microRNA (miRNA) expression changes in murine prostate cancer (PCa) xenografts using two different diets: HFD and control diet. We then assessed the roles and targets of altered miRNAs in HFD-induced PCa progression. We identified 38 up- and 21 downregulated miRNAs in xenografts under HFD conditions using the miRCURY LNA™ microRNA array. The differences in 10 candidate miRNAs were validated using quantitative RT-PCR. We focused on miR-130a because the expression levels were significantly lower in the three PCa cell lines in comparison with benign prostate PINT1B cells. PCa cells cultured in a medium containing HFD mouse serum were associated with significantly higher cell proliferation rates and lower miR-130a expression levels. Further, miR-130a modulated MET expression in PCa cells, and MET was overexpressed i nin vitro and in vivo HFD-induced PCa progression models. Moreover, ectopic miR-130a downregulated AR in LNCaP cells and DICER1 in PC-3 and DU145 cells, respectively. In human tissues, as elucidated using laser capture microdissection, the mean miR-130a expression level in cancer epithelium was significantly lower than that in normal epithelium. Furthermore, cytoplasmic MET in PCa tissues was overexpressed in patients with higher body mass index. In conclusion, a substantial number of miRNAs was altered in HFD-induced PCa growth. Specifically, miR-130a was attenuated in HFD-induced PCa progression with MET overexpression. miRNAs thus have implications in the mechanism, prevention and treatment of HFD-induced PCa progression.
Prostate cancer (PCa) is the second most commonly diagnosed environmental factors, is associated with PCa carcinogenesis,
cancer in men and the sixth most lethal cancer at a global as reported by several epidemiological and mechanistic stu-d
level 1(). Interestingly, a wide variation exists internationally ies (4,5). However, the underlying mechanisms largely remain
for both PCa incidence and mortality2)(, and immigrants from unknown.
Asia have an intermediate risk of PCa incidence between ind-ig Several mechanisms including insulin/insulin-like growth f-ac
enous Asians and those born in the USA3(). These results sug- tor-1 axis, sex hormones, cytokine signaling and oxidative stress
gest that lifestyle and environmental factors play an important have been postulated as having an impact on diet and/or obesity
role in PCa progression and aggressiveness. Dietary fat, which in carcinogenesis 6(,7). We previously reported the relationship
is presumed to be one of the strongest influential lifestyle and between PCa carcinogenesis and a specific diet including a high-fat
receptacles were placed on top of the cages to control food intake, and fresh
food was given without opening the cages three times a week. Body weight
and tumor volume were weekly measured throughout the duration of the
experiment. Tumor volume was calculated using the following formula:
length (cm) × weight (cm) × height (cm) × 0.5236. At 14 weeks, the animals
were killed, and xenograft tumors were excised and stored at –80°C until
further use. The institutional review board of Akita University School of
Medicine approved all animal experiments.
RNA isolation for microarray analysis
induced PCa progressionin vitro, in vivo and in human samples.
Materials and methods
Quantitative RT-PCR (qRT-PCR)
RNA fractions were extracted from LNCaP xenografts and cell lines using
the mirVana isolation kit for qRT-PCR. TaqMan miRNA assays were used to
quantify the expression levels of hsa-miR-215, hsa-miR-494, hsa-miR-1225a,
hsa-miR-665, hsa-miR-3170, hsa-miR-27a, hsa-miR-130a, hsa-miR-24,
hsamiR-222 and hsa-miR-152 (Applied Biosystems), according to the
manuFor microarray analyses, total RNA was extracted from each of the three
diet (HFD), carbohydrate diet and specific fatty acids usi ningvitro tumors including those with enhanced growth in the HFD group and with
and mouse models ( 8–11). HFD enhanced tumor growth by modu- tmyirpiVcaanl,aslmoiwRNgAroiwsotlhatinionthkeitC,Dagccroourdp.inTogttaol RtNheAsmwaneruefaicsotularteerd’s uinsisntgr-utche
lating the expression of a substantial number of mRNAs related tion (Ambion). Total RNA concentration was evaluated using ND-1000
to cancer, apoptosis and growth factor signaling, while regulating (Nanodrop Technologies, Wilmington), and RNA integrity was analyzed
the proliferation and invasion of PCa8,(10). MicroRNAs (miRNAs) using Agilent Bioanalyzer 2100 (Agilent Technologies, Cheshire, UK).
are small non-coding RNAs of 19–25 nucleotides that can inhibit
mRNA translation and/or negatively regulate mRNA stabil1it2y).( Microarray analysis
Dysregulation of miRNA expression is common in various can- We delegated microarray analyses to Filgen (Aich, Japan). miRNA expr-es
cers, and accumulating data have revealed multiple functions of sion was generated via the miRCURY LNATM microRNA Array, fifth
genmiRNAs as oncogenes or tumor suppressors (13,14). In PCa, sev- eration platform (Exiqon, Denmark). One microgram of total RNA was
eral miRNAs have been proposed to be aberrantly expressed and labeled with dyes spectrally equivalent to the CTMy3fluorophores using
have been suggested to functionally contribute to tumorigenesis the miRCURYTM Array Power Labelling kit (Exiqon, Denmark). Microarrays
(15,16). In addition, recent studies have suggested that miRNAs with labeled samples were hybridized at 56°C for 16 h and washed using
contribute to many cellular processes caused by dietary factors the miRCURYTM Array Wash Buffer kit (Exiqon, Denmark). Slides were
scanned using the Genepix 4000B scanner (Axon Instruments). The flu-o
(17,18). However, to the best of our knowledge, there are no reports rescence signal of each spot was quantified using the Array Pro Analyzer
regarding the relationship between diet-induced miRNA expr-es ver. 4.5 (Media Cybernetics, Bethesda). Quantified intensity was analyzed
sion and PCa progression 1(8). using a software package (Microarray Data Analysis Tool ver. 3.2, Filgen).
Here we hypothesized that HFD modulates several cancer- Significant changes in miRNA expression were defined as a 1.5
foldrelated miRNAs in PCa progression. Based on this hypothesis, we change and 20% false discovery rate in comparison with each group19().
investigated the differential miRNA profiles by microarray ana-ly Minor sequences of miRNAs and Exiqon-specific miRNAs named ‘miR
sis in a murine model with an LNCaP PCa xenograft using two d-if plus’ were omitted from further analyses. A heat map for differentially
ferent diets: HFD and control diet (CD). Furthermore, we focused expressed miRNAs was generated by using software Cluster 3.0 and Java
on the roles of candidate miRNAs and/or their targets on HFD- Treeview (Michael Eisen Laboratory, Stanford University, CA).
Human PCa LNCaP, PC-3 and DU145 cells were obtained from the American
Type Cell Culture Collection (Manassas, VA) or RIKEN BioResource Center facturer’s instructions. Briefly, single-stranded cDNA was generated from
(Tsukuba, Japan). Cells were authenticated with short-tandem repeat RNA samples by reverse transcription using the TaqMan microRNA Reverse
analysis by JCRB Cell Bank (Osaka, Japan) Applied Biosystems (Foster City, Transcription kit (Applied Biosystems). qRT-PCR was performed with the
CA), Bio-Synthesis (Lewisville, TX), respectively. LNCaP and DU145 cells TaqMan Universal PCR Master Mix and primers of the TaqMan microRNA
were obtained in 2011 and authenticated in 2015. PC-3 cells were obtained assay using the Takara Thermal Cycler Dice® Real Time System, acco-rd
in 2013 and authenticated in 2015. The normal prostate epithelial cell line ing to the manufacturer’s instructions. The reaction was performed under
PNT-1B was provided by Professor N Maitland (York, UK). Cells were c-ul the following conditions: pre-incubation at 95°C for 10 min, followed by 40
tured in RPMI 1640 medium (Invitrogen, Carlsbad, CA) or Dulbecco's mod-i cycles each of 95°C for 15 s and 60°C for 1 min. Relative miRNA expression
fied Eagle's medium (Invitrogen) containing 10% fetal bovine serum in a data were analyzed using the 2Δ-ΔCT method (20) with RNU6 as an
endog5% CO2 humidified incubator at 37°C. enous control2(1). Each experiment was performed in triplicate.
Cell proliferation assay
We previously reported that the tumor growth of LNCaP xenografts was- sig Sera from mice fed with HFD and CD without any xenografts for 14 weeks
nificantly higher in HFD-fed cells than in CD-fed cel1l0s)(. We used the same were collected, clarified by filtration (SLGV004SL) (Millipore, Billerica,
model to assess the expression of miRNAs in xenografts. Briefly, 6-week- MA) and used for an in vitro cell proliferation assay; a non-radioactive
old athymic BALB/c-nu/nu mice were obtained from Japan SLC (Shizuoka, MTT-based cell proliferation assay kit (Roche) was used, according to the
Japan), and three mice were housed per cage in a pathogen-free enviro-n manufacturer’s instructions. The proliferation assays were performed in
ment. Each mouse was fed an autoclaved CE-2 diet (Japan SLC) and subcu- triplicate. A total of 6.25 × 120PC-3 or 1.25 × 103 DU145 cells were seeded
taneously inoculated with LNCaP cells in the hind limb. For each injection, into each well of a 96-well plate and incubated for 96 h with fresh media
1 × 106 LNCaP cells were re-suspended in 0.25 ml of ice-cold BD Matrigel containing 2.5% mouse serum. Absorbance was measured using an ELISA
(BD Bioscience, Bedford, MA), and then, 0.25 ml of RPMI medium was added. reader (BIO RAD, Hercules, CA).
Four weeks after the injection, 24 mice with a palpable tumor were r-an Transfection with synthetic miRNA and an miR
domly divided into two different groups: HFD and CD. The two types of sp-e
cific diets were obtained by Test Diet (Purina Mills Test Diets, Richmond, IN). inhibitor
As shown in Supplementary Table 1, available atCarcinogenesis Online, HFD LNCaP, PC-3 and DU145 cells were seeded in 6-well plates and incubated
and CD reaped 59.9 and 41.6% calories from fats, respectively. The feeding overnight. In total, 30 pmol of Pre-miR miRNA precursor molecules
(Ambion, Huston, TX) mimicking hsa-miR-130a or control non-spe- Results
cific miRNA were transfected into cell lines using Lipofectamine 2000
(Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions. HFD enhanced tumor growth in PCa LNCaP
The miRCURY LNA™ microRNA inhibitor for hsa-miR-130a and the xenografts
negative control microRNA inhibitor (scrambled) oligonucleotides were
purchased from Exiqon, Denmark. Inhibitor transfections (150 pmol)
into PC-3 and DU145 cells were performed using Lipofectamine 2000
When we assessed the impact of the two diets on tumor growth
of LNCaP xenografts, the tumor growth rate of xenografts in
mice fed with HFD was significantly higher than that in those
fed with CD (P = 0.009, Supplementary Figure 1A, available at
Western blot analysis Carcinogenesis Online); this result was consistent with that of
Total proteins were isolated using complete Lysis-M buffer (Roche our previous study (10). At 14 weeks, the mean tumor volume
Diagnostic). Protein concentration was measured using the ND-1000 of LNCaP xenografts was significantly higher in the HFD group
method (Thermo Fisher Scientific). Equal amounts of protein lysates (HFD group, 2.57 ± 0.22 cm 3 versus CD group, 1.58 ± 0.20 cm3,
were separated by SDS-PAGE and transferred using the iBot® Blotting P = 0.003, Supplementary Figure 1B, available atCarcinogenesis
System (Invitrogen). The membranes were blocked for 1 h at room Online). The mean body weight of mice fed with HFD at 14
temperature with a buffer containing 2% bovine serum albumin in
Tris-buffered saline with 0.1% Tween-20. The membranes were inc u- weeks tended to be higher than that of those fed with CD (HFD
bated overnight in the diluted antibodies and blocked with seco-nd group, 20.89 ± 0.53 g versus CD group, 19.13 ± 0.59 g,P = 0.060,
ary IgG antibody for 1 h. Specific proteins were detected using the ECL Supplementary Figure 1C, available atCarcinogenesis Online).
prime western blotting detection reagent (Amersham Biosciences, Next, to eliminate the effect of body weight on tumor growth, we
Buckinghamshire, UK). The monoclonal/polyclonal antibodies MET, estimated the mean ratio of tumor volume to body weight. The
SOCS3, DICER1, MAPK androgen receptor (AR) andβ-actin were pur- tumor growth adjusted for body weight was significantly higher
chased from Cell Signaling Technology. in the HFD group (HFD group, 0.12 ± 0.0097 cm3/g versus CD
group, 0.083 ± 0.011 cm 3/g, P = 0.020, Supplementary Figure 1D,
Laser capture microdissection for human prostate
available atCarcinogenesis Online). These results suggested that
tissue samples HFD enhanced the tumor growth of PCa LNCaP xenografts,
regardless of the impact of changes in body weight.
Fresh human prostate tissues were obtained from eight patients who
underwent radical prostatectomy without any preoperative hormonal
and chemotherapeutic treatment at Akita University Hospital. The-tis Candidate miRNAs involved in tumor growth of
sues were snap-frozen in liquid nitrogen and stored at −80°C until further human PCa xenografts under HFD conditions
use. All slides were reviewed by a pathologist (T.Y.), who was blinded to
the patient’s clinical background for the identification of PCa foci as well After image quantification and normalization of micro-ar
as adjacent normal or benign hyperplasia epithelial tissue. Frozen tissue ray data, we searched for miRNAs with consistent differential
was embedded in an optimal cutting temperature compound. Serial fr-o expression levels between the HFD and CD groupsn( = 3, each).
zen sections (8 µm) were cut using a clean blade at −20°C and mounted Based on our miRCURY LNATM microarray data, we found 38
upon glass slides. The sections were then stained with HistoGene Staining and 21 downregulated miRNAs, with a false-positive discovery
Solution using the HistoGene LCM Frozen Section Staining kit (Arcturus rate of <0.2 and fold change of >1.5 (positive or negative) in xe-n
Engineering, Mountain View, CA), according to the manufacturer’s inst-ruc ografts under HFD conditionsF(igure 1; Supplementary Tables 2
tions. Glands of normal and cancerous prostatic epithelia, specified by a
pathologist, were dissected with the LMD6500/7000 Leica Microsystems® and 3, available atCarcinogenesis Online).
(Leica AS LMD, Leica, Wetzlar, Germany), were collected in 6µ0l RNA lysis We selected five markedly up- and downregulated miRNAs
solution of the RNAqueou®-Micro kit (Ambion) according to the man-u each to confirm the results obtained from the miRNA array an-al
facturer’s instructions. The institutional review board of Akita University ysis (Table 1). All up- or downregulated miRNAs in the array were
School of Medicine approved all experiments, and human prostate sam- up- or downregulated as per qRT-PCR tooT(able 1), although
ples were utilized with written informed consent. there were differences in the ratio of the miRNA level between
the miRNA array results and qRT-PCR results. miR-130a,
miRImmunohistochemistry 24, miR-27a and miR-665 showed robust expression changes
PCa samples were obtained from 60 patients who underwent radical between the two groups as per both the miRNA array results and
prostatectomy without any preoperative hormonal and chemotherape-u qRT-PCR results, and these were selected for further analyses as
tic treatment at Akita University Hospital. The patient’s body mass index previous studies have reported their importance in the develo-p
(BMI) was obtained from medical records. The MET (C-28) polyclonal ment and progression of several types of cancer2(2,23).
antibody (Santa Cruz) was used as the primary antibody at a dilution of
1:100. Immunohistochemical staining was performed as described previ- Expression of candidate miRNAs in PCa cells and in
(oSu.Nsl. yan(8d). MEv.aHl.)ubaltiinodneadntdo stchoeripnagtiwenetr’es bpearcfkogrrmoeudndwaitnhdtchleiniincvoepsattihgoalt-oogris a benign immortalized prostate cell line
cal information. To assess MET immunoreactivity, we used the following
scoring systems. Cytoplasmic intensity of cancer epithelium was scored
and stratified into four groups: negative (0), weak1),( moderate (2) and
strong 3(). Percentage of cytoplasmic MET-positive cells in cancer epit-he
lium was stratified into five groups: 0% (0), <25%1(), <50% (2), <75% (3) and
≤100% (4). The total immunoreactivity score was calculated by multipl-ica
tion of the two values, score = staining intensity × fraction of positive cells,
as described previously 8().
We examined the expression levels of miR-130a, miR-24,
miR27a and miR-665 in four prostate cell lines including and-ro
gen-sensitive PCa LNCaP, androgen-independent PCa PC-3 and
DU145 and in a benign immortalized prostate cell line PINT1B
by qRT-PCR (Figure 2A–D). The expression of miR-130a, miR-24
and miR-27a downregulated in HFD xenografts was significantly
lower in LNCaP cells than in PNT1B cellPs =( 0.008, 0.008 and
0.008, respectively, Figure 2A–C), whereas the expression of
miR-665 upregulated in HFD xenografts was significantly higher
in LNCaP cells than in PNT1B cellsP (= 0.008, Figure 2D).
miR130a levels were significantly lower in PC-3 and DU145 cells in
comparison with PINT1B cells P( = 0.008 and 0.008, respectively,
Figure 2Aand B). On the other hand, there were no differences in
Statistical analyses were performed using SPSS ver. 22. All values are-pre
sented as mean ± SE. Unpaired Student’st-test and Mann–Whitney test
were used to compare differences between the groups. Differences were
considered significant ifP values were <0.05.
the expression of miR-27a and miR-24 between PNT1B cells and
in PC-3 and DU145 cellsF(igure 2Band C).The expression of
miR665 was significantly higher in PC-3 cells than in PNT1B cells
(P = 0.032), whereas there was no difference in the expression of
miR-665 between PNT1B and DU145 cells P( = 0.917, Figure 2D).
We further focused on miR-130a as the relative expression of
miR-130a was significantly lower in all the three cancer cell lines
in comparison with the benign cell line, indicating its potential
role as a suppressive miRNA in PCa progression.
HFD-fed mouse sera promoted cell proliferation and inhibited miR-130a expression in PCa cells
To confirm the functional role of HFD on PCa progression
and miR-130a expression in HFD-induced PCa progression,
an in vitro cell proliferation assay was performed with two
PCa cell lines, PC-3 and DU145, using fresh media containing
2.5% mouse serum from mice fed with either HFD or CD. PC-3
and DU145 cells cultured in media containing 2.5% HFD-fed
mouse serum showed significantly higher cell proliferation
levels than those cultured in media containing 2.5% CD-fed
mouse serum ( P = 0.002 and 0.041, respectively, Figure 3A
and B). Regarding miRNA expression in an in vitro model, the
expression levels of miR-130a in PC-3 and DU145 cells cu-l
tured in media containing 2.5% HFD-fed mouse serum were
significantly lower than those cultured in media containing
2.5% CD-fed mouse serum ( P = 0.029 and 0.009, respectively,
Figure 3C and D). Therefore, sera from mice fed with HFD
enhanced PCa cell proliferation and downregulated miR-130a
expression in PCa cells.
Prediction of target genes and proteins of miR-130a in PCa cell lines
Post-transcriptional silencing of target genes by miRNAs can
occur either by cleavage of homologous target mRNAs or by
inhibition of target protein synthesis24(). Computational
predictions indicate that one miRNA may target hundreds of
genes and that >50% human protein-coding genes could be
regulated by miRNAs (25). To identify the effectors of miR-130a
in PCa, we performed protein expression analysis in the three
FAi ghueraet 1m.aHpeawtamsagpenfoerraatletderbeadsemdiRoNnAdeixffperreesnstioianlliny HreFgDulaantdedCDmixReNnAosgrwaiftths. a PCa cell lines transfected with either has-miR-130a precursor
false-positive discovery rate of <0.2 and fold change of >1.5 (positive or negative) molecule or control. Along with the transfection experiment,
in LNCaP xenografts of mice fed with HFD compared with mice fed with CD.This we selected several proteins for candidates of downstream
was done using Cluster 3.0 and Java Treeview. CD, control diet; HFD, high-fat diet. effectors of miR-130a based on the TargetScan algorithm
and/or previous studies that have reported miR-130a t-ar
gets (22,26,27). In particular, we focused on the miR-130/MET
Table 1. Quantitative RT-PCR results to validatethe expression of top interaction because two previous studies had revealed that
5 up- and down-regulated miRNAs in microRNA array analysis miR-130a directly targeted MET in cancer cell2s6(,28). We
Microarray results Quantitative RT-PCR faolulntdhrteheaPtCtahceelplroltineeins terxapnressfesciotnedowfiMthEThswaa-msidReNcAre-1a3s0ead-pirne
HFD/CD ratio HFD/CD ratio cursor molecule in comparison with control miRNA precursor
molecule (Figure 4A). Furthermore, AR expression in LNCaP
Oncogenic (upregulated) miRNAs cells transfected with miR-130a was lower than that in LNCaP
hsa-miR-215 6.1 1.4 cells transfected with control, and DICER1 was downregulated
hsa-miR-494 2.7 1.1 in PC-3 and DU145 cells F(igure 4A). These results suggested
hsa-miR-1255a 2.5 1.1 that MET is a promising target of miR-130a in PCa cells, and
hsa-miR-665 2.0 4.0 cell-dependent targets may exist in PCa cells transfected with
hsa-miR-3170 1.9 1.4 miR-130a.
Suppressive (downregulated) miRNAs To explore whether MET overexpression was observed du-r
hsa-miR-27a 0.37 0.58 ing HFD-induced PCa progression, we assessed the expression
hhssaa--mmiiRR--12340a 00..5422 00..6552 of MET genes and proteins inin vivo and in vitro HFD-induced
hsa-miR-222 0.53 0.67 PCa progression models described previously F(igure 4B–D).
hsa-miR-152 0.54 0.58 The relativeMET mRNA expression was significantly higher in
LNCaP xenografts of mice fed with HFD than in those of mice
fed with CD (P = 0.002, Figure 4B). The MET protein expression with the negative controPl (= 0.004 and P = 0.025, respectively;
level was higher in the three PCa cell lines cultured with 2.5% Figure 4F and G).
HFD-fed mouse serum compared with those cultured with 2.5%
CD-fed mouse serum ( Figure 4C). In addition, the mRNA expres- Expression levels of miR-130a and MET in
sion of MET was significantly higher in PC-3 and DU145 cells human PCa
with media containing HFD-fed mouse serum compared with Using surgically resected human prostate tissues from patients
that in those with CD-fed mouse serum P( = 0.002 and 0.002, with PCa, miR-130a and MET expression levels were assessed. To
respectively,Figure 4D and E). evaluate the expression of miR-130a in human PCa tissues,
To confirm the MET modulation by miR-130a in PCa cells, 130a levels were measured in cancerousn( = 8) and normal (n = 8)
we performed a MET mRNA expression analysis in PC-3 and epithelia separated by laser capture microdissection by qRT-PCR.
DU145 cells treated with either the miR-130a inhibitor or its All specimens yielded RNA of sufficient quantity (>8.25 ngµl/) and
negative control. As expected, the relatiMveET mRNA expres - most showed a relatively high purity with an O26D0/280 value of ~2.0.
sion was significantly higher in the PC-3 and DU145 cells tra-ns In human tissues, the mean miR-130a expression level of cancer
fected with the miR-130a inhibitor than in those transfected epithelia was significantly lower than that of normal epithelia
(P = 0.023, Figure 5A). As obesity is strongly associated with d-ie overexpression of its potential target, MET. In addition,
miRtary fat 2(9), we assessed the relationship between MET expre-s 130a expression was decreased in human PCa epithelium, and
sion in radical prostatectomy specimens and BMIF(igure 5B–D). cytoplasmic MET in PCa epithelium was overexpressed in PCa
Cytoplasmic MET expression varied in PCa epithelia, and the patients with a higher BMI. The potential mechanism revealed
immunoreactivity scores ranged from 0 to 12. Representative in this study is summarized in Figure 6.
images are shown in Figure 5B. The median BMI value was 26, In this study, miR-130a was found to be one of the promising
and the mean immunoreactivity score of MET was significantly targets in HFD-induced PCa progression. Bolelt al. (22) recently
higher in patients with BMI ≥ 26n( = 29) than in those with BMI assessed the expression profile of miRNAs in PCa cells and
< 26 (n = 31, P = 0.043, Figure 5C). Based on the total immuno- human PCa. In their study, they found downregulation of
miRreactivity score, we divided the patients into three groups: low 130a in LNCaP cells compared with benign prostate cell lines
(scored as 0–3,n = 19), moderate (scored as 4–6,n = 23), and high and a high differentiation grade of human PCa. In chronic ly-m
(scored as 8–12,n = 18). The mean BMI was significantly higher in phocytic leukemia research, Kovalevaet al. (27) revealed that
patients with high MET expression compared with that in those miR-130a was downregulated in primary cancer cells and that
with low MET expression P( = 0.028, Figure 5D). In addition, the its expression was associated with cell survival. These results
immunoreactivity score of MET was not associated with major were consistent with our results showing that low miR-130a
clinical variables, including preoperative prostate specific antigen expression plays a role in cancer progression. Regarding the
(PSA) levels, clinical stage, pathological Gleason score, preop-era impact of diet on miR-130a expression in disease models, Kim
tive total serum cholesterol level, preoperative serum triglycerideet al. assessed the expression of miR-130a and miR-130b in
level and prostate specific antigen recurrence after prosta-tec white adipose tissues of mice with HFD-induced inflammation
tomy (Supplementary Table 4, available atCarcinogenesis Online (30). They found that miR-130a expression in white adipose
and supplementary Figure 2, available atCarcinogenesis Online). tissues of HFD-fed mice was significantly higher than that in
CD-fed mice, which is discordant with the present result that
Discussion dHiFsDcraetpatenncuyamteasy mreiRs-u1l3t0afreoxmprtehsesioexnisintemniccee oxfentiosgsruaef-tssp.eTchifiisc
Here, we demonstrated that HFD enhanced the tumor growth expression profiles or cancer-related modulations of
miRof PCa LNCaP xenografts and resulted in the modulation of a 130a. In this study, we did not elucidate the exact mechanism
substantial number of miRNAs. Among these miRNAs, miR-130a on HFD-induced miR-130a attenuation but did identify se-v
was markedly attenuated in HFD-induced PCa progression, with eral potential mechanisms, including roles for specific lipids,
modulating miR-130a expression or diet. Furthermore, miR-130a
has been reported to inhibit autophagy by reducing autopha-go
some formation and an effect mediated by downregulation of
the genesATG2B and DICER1 (27). In our study, ectopic miR-130a
was observed to downregulate DICER1 expression in PC-3 and
DU145 cells. As DICER1 is also a potential target of miR-130a, its
functional role in autophagic response in PCa should be further
We could not delineate the functional roles of all diffe-ren
tially expressed miRNAs identified through our miRNA array
analyses. miR-24 and miR-27a were other related miRNA cand-i
dates for HFD-induced PCa progression. Both of these have been
previously reported to be downregulated in PC2a2(,36), whereas
the expression of the miR-27a24-2 cluster has been reported to
be upregulated in several types of cancers, including breast c-an
cer, gastric adenocarcinoma and PCa3(7,38). The fact that
GATA2, which has been shown to be inhibited by miR-24/27a 3(9), is
bound to many AR-binding regions is of interest4(0). Regarding
candidate oncogenic miRNAs, the expression of miR-665 was
Figure 6. Proposed schema of the molecular mechanisms of HFD-induced
prostate cancer growth with the modulation of miRNAs. Downregulation of miR- upregulated in HFD xenografts in the present study. A study on
130a, miR-21 and miR-27 and upregulation of miR-665 were observed during esophageal cancer revealed that miR-665 was upregulated in
the HFD-induced progression of PCa. HFD potentially attenuated miR-130a with esophageal squamous cell cancer tissues compared with
nonMET, AR and DICER1 overexpression in HFD-induced PCa cell proliferation. tumor counterparts 4(1). However, still, there is insufficient
information about the functional role of miR-665 in biological
inflammation, insulin signaling and adipokines for diet- and/ processes. Further studies are warranted to clarify the role of
or obesity-modulated cancer progression31(). Further study is candidate miRNAs identified by us as having different expre-s
required to explore how HFDs attenuate miR-130a expression sions in HFD and CD xenografts.
during PCa progression. Our study has several limitations. First, the diets used c-on
We further found that the expression of MET was dow-n tained different fat sources and ingredients other than fat. The
regulated in all three PCa cell lines transfected with ectopicindependent ingredients may have some effects on prostate
miR-130a. In addition, the expression of MET in xenografts tumor progression. Second, we used single cell lines for the
under HFD conditions was higher than that in those under CD miRNA expression profile analysis using a xenograft model,
conditions, and thein vitro results revealed HFD-induced MET although we performed microarray analysis for each of three
upregulation in PCa cells. Acunzeot al. (26) reported that human independent tumors to avoid sample bias. Last, the sample size
miR-130a has a matched region in the 3′-untranslated region to of prostate specimens was relatively small. Furthermore, the
human MET and revealed that enhanced expression of miR-130a relationship between MET expression and other crude indic-a
reduced MET protein and mRNA levels in A549 lung cancer cells. tors of fat intake in patients with PCa who are treated using
Moreover, Sundaram et al. (32) found that MET correlated with radical prostatectomy should be examined in future studies.
obesity-induced basal-like breast cancer tumor onset in null-ipa In conclusion, a substantial number of miRNAs were
modurous mice. Based on the results of our study, MET is a potential lated in HFD-induced PCa tumor growth. Attenuation of
miRtarget of diet- and/or obesity-modulated miR-130a in cancer 130a under HFD conditions may enhance PCa progression,
progression. However, we did not confirm the direct binding of partly by targeting MET. Our results enable successful cl-arifi
miR-130a to MET 3′UTR using luciferase assay as described in the cation of the molecular mechanisms underlying HFD-induced
previous studies (26,28). Furtherin vitro study is needed to prove PCa progression. Targeting miRNAs may have a potential role for
the direct targeting of MET by miR-130a in PCa cell lines. A p-re novel therapeutic and prevention strategies in PCa.
vious PCa-based study showed that MET expression was higher
in tumor samples from patients with castration-resistant PCa
(CRPC) compared with tumor samples from those who had not Supplementary material
yet undergone androgen deprivation therapy33(). Cabozantinib Supplementary Tables 1–4 and Figures 1–3 can be found at
(XL-184), an orally bioavailable tyrosine kinase inhibitor with http://carcin.oxfordjournals.org/
potent activity against MET and VEGF receptor 2, reportedly has
clinical activity in men with CRPC34(). Collectively, miR-130a
and MET may be attractive therapeutic targets for PCa patients, Funding
and we provide evidence on the preclinical therapeutic potential This study was supported in part by research grafts from
Grantof miR-130a using in vitro and in vivo models. in-Aid for Scientific Research (Kakenhi, No. 25462466, 25293332,
Ligand-independent activation of AR is one of the critical 26670695 and 16H02679) and Japan Agency for Medical Research
steps of castration-resistant PCa35(). Interestingly, our study and Development (AMED-CREST).
indicated that miR-130a overexpression in LNCaP cells reduced
AR expression. Several AR coregulators, for example GTF2H1, Acknowledgements
ROCK1 and STX6, have been reported to be modulated by
miR130a (22). It is necessary to validate the impact of miR-130a We greatly thank Yoko Mitobe, Yukiko Sugiyama and Yuka
on AR activation by assessing the expression of downstream Izumida for their technical assistance.
targets of AR in LNCaP and other AR-positive PCa cell lines by Conflict of Interest Statement: None declared.
References 22. Boll, K. et al. (2013) MiR-130a, miR-203 and miR-205 jointly repress key
oncogenic pathways and are downregulated in prostate carcinoma.
1. Ferlay, J. et al. (2010) Estimates of worldwide burden of cancer in 2008: Oncogene, 32, 277–285.
GLOBOCAN 2008. Int. J. Cancer, 127, 2893–2917. 23. Nygren, M.K. et al. (2014) Identifying microRNAs regulating B7-H3 in
2. Center, M.M. et al. (2012) International variation in prostate cancer breast cancer: the clinical impact of microRNA-29c. Br. J. Cancer, 110,
incidence and mortality rates. Eur. Urol., 61, 1079–1092. 2072–2080.
3. Shimizu, H. et al. (1991) Cancers of the prostate and breast among Ja-pa 24. Gregory, R.I. et al. (2005) MicroRNA biogenesis and cancer. Cancer Res.,
nese and white immigrants in Los Angeles County. Br. J. Cancer, 63, 65, 3509–3512.
963–966. 25. Li, S.C. et al. (2007) Intronic microRNA: discovery and biological imp-li
4. Giovannucci, E. et al. (1993) A prospective study of dietary fat and risk cations. DNA Cell Biol., 26, 195–207.
of prostate cancer. J. Natl. Cancer Inst., 85, 1571–1579. 26. Acunzo, M. et al. (2012) miR-130a targets MET and induces TRAIL-se-n
5. Kobayashi, N. et al. (2008) Effect of low-fat diet on development of pr-os sitivity in NSCLC by downregulating miR-221 and 222. Oncogene, 31,
tate cancer and Akt phosphorylation in the Hi-Myc transgenic mouse 634–642.
model. Cancer Res., 68, 3066–3073. 27. Kovaleva, V. et al. (2012) miRNA-130a targets ATG2B and DICER1 to
6. Venkateswaran, V. et al. (2010) Diet and prostate cancer: mechanisms inhibit autophagy and trigger killing of chronic lymphocytic leukemia
of action and implications for chemoprevention. Nat. Rev. Urol., 7, 442– cells. Cancer Res., 72, 1763–1772.
453. 28. Zhou, Y.M. et al. (2014) MiR-130a overcomes gefitinib resistance by t-ar
7. Allott, E.H. et al. (2013) Obesity and prostate cancer: weighing the- evi geting met in non-small cell lung cancer cell lines. Asian Pac. J. Cancer
dence. Eur. Urol., 63, 800–809. Prev., 15, 1391–1396.
8. Narita, S. et al. (2008) Candidate genes involved in enhanced growth of 29. Bray, G.A. et al. (1998) Dietary fat intake does affect obesity! Am. J. Clin.
human prostate cancer under high fat feeding identified by microarray Nutr., 68, 1157–1173.
analysis. Prostate, 68, 321–335. 30. Kim, C. et al. (2013) TNFalpha-induced miR-130 resulted in adipocyte dy-s
9. Huang, M. et al. (2011) Overexpression of Fn14 promotes androgen- function during obesity-related inflammation. FEBS Lett., 587, 3853–3858.
independent prostate cancer progression through MMP-9 and cor-re 31. Louie, S.M. et al. (2013) Mechanisms linking obesity and cancer. B-io
lates with poor treatment outcome. Carcinogenesis, 32, 1589–1596. chim. Biophys. Acta, 1831, 1499–1508.
10. Huang, M. et al. (2012) A high-fat diet enhances proliferation of pr-os 32. Sundaram, S. et al. (2013) Role of HGF in obesity-associated tumorige-n
tate cancer cells and activates MCP-1/CCR2 signaling. Prostate, 72, esis: C3(1)-TAg mice as a model for human basal-like breast cancer.
1779–1788. Breast Cancer Res. Treat., 142, 489–503.
11. Huang, M. et al. (2014) Diet-induced macrophage inhibitory cytokine 1 33. Pfeiffer, M.J. et al. (2011) Steroidogenic enzymes and stem cell markers
promotes prostate cancer progression. Endocr. Relat. Cancer, 21, 39–50. are upregulated during androgen deprivation in prostate cancer. Mol.
12. Ambros, V. (2004) The functions of animal microRNAs. Nature, 431, Med., 17, 657–664.
350–355. 34. Smith, D.C. et al. (2013) Cabozantinib in patients with advanced pr-os
13. Ventura, A. et al. (2009) MicroRNAs and cancer: short RNAs go a long tate cancer: results of a phase II randomized discontinuation trial. J.
way. Cell, 136, 586–591. Clin. Oncol., 31, 412–419.
14. Croce, C.M. (2009) Causes and consequences of microRNA dysregul-a 35. Scher, H.I. et al. (2005) Biology of progressive, castration-resistant p-ros
tion in cancer. Nat. Rev. Genet., 10, 704–714. tate cancer: directed therapies targeting the androgen-receptor si-gnal
15. Bonci, D. et al. (2008) The miR-15a-miR-16-1 cluster controls prostate c-an ing axis. J. Clin. Oncol., 23, 8253–8261.
cer by targeting multiple oncogenic activities. Nat. Med., 14, 1271–1277. 36. Szczyrba, J. et al. (2011) Downregulation of Sec23A protein by
miRNA16. Liu, C. et al. (2011) The microRNA miR-34a inhibits prostate cancer 375 in prostate carcinoma. Mol. Cancer Res., 9, 791–800.
stem cells and metastasis by directly repressing CD44. Nat. Med., 17, 37. Fletcher, C.E. et al. (2012) Androgen-regulated processing of the
211–215. oncomir miR-27a, which targets prohibitin in prostate cancer. Hum.
17. Zhang, J. et al. (2009) Maternal high fat diet during pregnancy and-lac Mol. Genet., 21, 3112–3127.
tation alters hepatic expression of insulin like growth factor-2 and key 38. Liu, T. et al. (2009) MicroRNA-27a functions as an oncogene in gastric
microRNAs in the adult offspring. BMC Genomics, 10, 478. adenocarcinoma by targeting prohibitin. Cancer Lett., 273, 233–242.
18. Saini, S. et al. (2010) Diet, microRNAs and prostate cancer. Pharm. Res., 39. Wang, F. et al. (2014) A regulatory circuit comprising GATA1/2 switch
27, 1014–1026. and microRNA-27a/24 promotes erythropoiesis. Nucleic Acids Res., 42,
19. Das, S. et al. (2012) Nuclear miRNA regulates the mitochondrial genome 442–457.
in the heart. Circ. Res., 110, 1596–1603. 40. Böhm, M. et al. (2009) A role for GATA-2 in transition to an aggressive
20. Livak, K.J. et al. (2001) Analysis of relative gene expression data using phenotype in prostate cancer through modulation of key
androgenreal-time quantitative PCR and the 2(-Delta Delta C(T)) method. M-eth regulated genes. Oncogene, 28, 3847–3856.
ods, 25, 402–408. 41. Kong, K.L. et al. (2012) MicroRNA-375 inhibits tumour growth and
21. Watahiki, A. et al. (2011) MicroRNAs associated with metastatic pr-os metastasis in oesophageal squamous cell carcinoma through repre-ss
tate cancer. PLoS One, 6, e24950. ing insulin-like growth factor 1 receptor. Gut, 61, 33–42.