Biochimica et Biophysica Acta 1843 (2014) 1225–1236

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamcr

Peroxisome proliferator-activated receptor γ-mediated induction ofmicroRNA-145 opposes tumor phenotype in colorectal cancer

Anna Panza a, Carolina Votino b, Annamaria Gentile a, Maria Rosaria Valvano a, Tommaso Colangelo b,Massimo Pancione b, LuciaMicale c, GiuseppeMerla c, Angelo Andriulli a, Lina Sabatino b, Manlio Vinciguerra d,e,Clelia Prattichizzo f, Gianluigi Mazzoccoli e,1, Vittorio Colantuoni b,1, Ada Piepoli a,⁎,1

a Division of Gastroenterology and Research Laboratory, IRCCS Scientific Institute and Regional General Hospital “Casa Sollievo della Sofferenza”, S. Giovanni Rotondo, Foggia 71013, Italyb Department of Sciences and Technologies, University of Sannio, Via Port'Arsa, Benevento 82100, Italyc Medical Genetics Unit, IRCCS Scientific Institute and Regional General Hospital “Casa Sollievo della Sofferenza”, S. Giovanni Rotondo, Foggia 71013, Italyd University College London, Institute for Liver and Digestive Health, Division of Medicine, Royal Free Campus, London NW3 2PF, UKe Department of Medical Sciences, Division of Internal Medicine and Chronobiology Unit, IRCCS Scientific Institute and Regional General Hospital “Casa Sollievo della Sofferenza”,S. Giovanni Rotondo, Foggia 71013, Italyf Department of Medical Science and Surgery, University of Foggia, Italy

⁎ Corresponding author at: “Casa Sollievo della Soffere71013 San Giovanni Rotondo, FG,Italy. Tel.: +39 0882 41

E-mail address: a.piepoli@operapadrepio.it (A. Piepoli1 These senior authors contributed equally.

http://dx.doi.org/10.1016/j.bbamcr.2014.03.0030167-4889/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 December 2013Received in revised form 19 February 2014Accepted 3 March 2014Available online 11 March 2014

Keywords:miR-145PPARγSOX9Colorectal cancerInvasiveness

MicroRNAs (miRNAs) regulate diverse biological processes by inhibiting translation or inducing degradation oftargetmRNAs. miR-145 is a candidate tumor suppressor in colorectal carcinoma (CRC). Colorectal carcinogenesisinvolves deregulation of cellular processes controlled by a number of intertwined chief transcription factors, suchas PPARγ and SOX9. Since PPAR family members are able to modulate complex miRNAs networks, we hypothe-sized a role of miRNA-145 in the interaction between PPARγ and SOX9 in colorectal carcinogenesis.To address this issue,we evaluated gene expression in tissue specimens of CRC patients andwe took advantage ofin vitro models represented by CRC derived cell lines (CaCo2, SW480, HCT116, and HT-29), employing PPARγactivation and/or miRNA-145 ectopic overexpression to analyze how their interplay impact the expression ofSOX9 and the development of a malignant phenotype.Results: PPARγ regulates the expression of miR-145 by directly binding to a PPAR response element (PPRE) in itspromoter at−1207/−1194 bp from the transcription start site. The binding is essential for miR-145 upregula-tion by PPARγ upon rosiglitazone treatment. Ectopic expression of miR-145, in turn, regulates SOX9 expressionthrough the binding to specific seed motifs. The PPARγ-miR-145-SOX9 axis overarches cell cycle progression,invasiveness and differentiation of CRC derived cell lines.Together, these results suggest that miR-145 is a novel target of PPARγ, acts as a tumor suppressor in CRC celllines and is a key regulator of intestinal cell differentiation by directly targeting SOX9, a marker of undifferenti-ated progenitors in the colonic crypts.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Colorectal cancer (CRC) is the third leading cause of cancer-relateddeath worldwide, representing the third most common cancer in maleand the second in female subjects. Its incidence is steadily increasingin many transition and developing countries when compared to theUnited States and other western countries. This appears to be due tothe diffusion of unhealthy western lifestyles, such as smoking habit,physical inactivity and high calorie diets [1,2]. CRC arises from theintestinal epithelium, which is constantly self-renewing in the crypts

nza” Hospital, Viale Cappuccini,6281; fax: +39 0882 410271.).

of Lieberkühn, and requires a high rate of cell division to maintain epi-thelial homeostasis [3]. Activation of the Wnt pathway is the key factorthat controls crypt cells proliferation [4]. In particular, genes that aremembers of the Wnt signaling pathway, such as β-Catenin (CTNNB1)and SRY-box containing gene 9 (SOX9), are crucial for the maintenanceof undifferentiated progenitors in the crypts and for the preservation ofthe post-mitotic Paneth cells [3]. Another important factor that plays aregulatory role in the differentiation process of intestinal epithelial cellsis Peroxisome Proliferator-Activated Receptor gamma (PPARγ) [5].PPARγ is a member of the nuclear receptor superfamily of ligand-activated transcriptional factors and is expressed throughout the gastro-intestinal epithelium. It is becoming increasingly clear that mutations indevelopmentally regulated genes and transcription factors can causeCRC initiation and progression, and, on the other hand, represent possibletherapeutic targets. The implication of PPARγ in colorectal carcinogenesis

Table 1Clinical and pathological features of colorectal cancer patients.

n = 31 %

Age (years):Mean ± SD 66 ± 13

Gender:Male/Female 21/10

Tumor location:Proximala 15 48.4Distalb 16 51.6

Modified Dukes staging systemA 1 3.2B 11 35.5C 19 61.3D 0 /

MSI frequencyHigh 6 20.7Low 6 20.7Stable 14 58.6Missing 5 /

a Cecum, ascending colon, and proximal transverse.b Distal transverse, descending colon, sigmoid colon and rectum.

Fig. 1. miR-145 expression levels of in colon cancer tissue compared to matched normatissue and in colon cancer cell lines. (A) Expression levels of miR-145 in primary tumorsandmatched normal specimens from a cohort of 31 CRC patients.MiR-145 expression levelsof normal tissue are normalized to 1 and represented as a horizontal line. P-values b 0.05were considered statistically significant. (B) Expression levels of miR-145 inCaCo2, SW480, HCT116 and HT-29 cell lines compared to normal colon tissue. *p b 0.05**p b0.01, *** p b 0.001.

Table 2Clinicopathological parameters in CRC patients and relationship with gene expressionlevels. Bold data indicate statistical significance.

miR-145

n Median Q1 Q3 p Values

AGE (years) b60 8 0.183 0.091 0.218 0.15460–67 8 0.311 0.260 0.91768–75 9 0.427 0.240 0.734N75 6 0.475 0.075 0.846

Gender M 21 0.427 0.205 0.846 0.105F 10 0.234 0.049 0.265

Tumor Location Proximal 15 0.420 0.208 0.846 0.292Transverse 2 0.180 0.004 0.357Distal 14 0.245 0.049 0.734

Grading G1/G2 28 0.381 0.149 0.790 0.349G3/G4 3 0.208 0.203 0.240

Dukes A 1 0.004 0.004 0.004 0.230B 11 0.420 0.227 0.535C 19 0.263 0.164 0.846D 0 – – –

MSI statusa MSI-H 6 0.058 0.034 0.134 0.009MSI-L 6 0.317 0.164 0.846MSS 14 0.381 0.256 0.734

a Microsatellite instability (MSI) analysis was performed using the Bethesda panel ofmicrosatellite (BAT25, BAT26, D5S346, D17S250, and D2S123) evaluated by means of amultiplex-PCR and PAGE analysis. Tumors were then classified accordingly into microsatel-lite stable (MSS), microsatellite instability-low (MSI-L), and microsatellite instability-high(MSI-H).

1226 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

l

,

is still debated, considering the controversial results of experimentalstudies evaluating the effects of its activation in CRC cell lines, animalmodels and humans [6–9].

Mounting evidencepinpoints the role of severalmicroRNAs (miRNAs)in tumorigenesis. MiRNAs are a class of short (21–24 nucleotide [nt]long), noncoding RNAs that influence diverse biological functions byrepressing target genes during normal development, as well as duringphysiologic and pathologic responses. This regulation is mediated pre-dominantly by binding to the 3′-untranslated region (UTR) of mRNAs,resulting in translational inhibition or transcript destabilization [10].As a result, miRNAs elicit critical changes in gene expression programs,which have been reported to underlie diverse aspects of biology,including developmental timing, differentiation, proliferation, cell death,metabolism, embryogenesis, and oncogenesis [11,12].MiRNAs biogenesisis tightly controlled in a tissue- and time-specificmanner [13,14]; howev-er, the transcriptional regulation of miRNA genes has not been systemat-ically studied. So far, only few transcription factors, such as p53, nuclearfactor erythroid-2-related factor 2 (NRF2),myogenic transcription factorsand PPARα, have been shown to regulate specificmiRNAs inmammaliancells [15–17].

Deploying amiRNA regulatory network analysis on tissue expressionprofiling of CRC patients we recently identified miRNA-145 levels asstrongly decreased [14], in accordance with its role as a candidatetumor suppressor, often downregulated in lung, prostate, liver and gas-tric cancers [18–24].MiRNA-145 regulates chondrogenesis by repressingSOX9 expression [25], modulates embryonic stem cell differentiation bycontrolling stemness genes KLF4, OCT4, and SOX2 [26], and targets anumber of oncogenes, such as MYC, DFF45 and FLI[15,27,28]. Moreover,miR-145 reduces the expression of SOX9/ADAM17 eliminating the pop-ulation of tumor initiating cells in head and neck cancer tissues [29], anddown-regulates IRS1 protein levels hampering the expansion of coloncancer cells [30]. The expression of miR-145 depends on correct func-tioning of p53 signaling pathway [31]. Nevertheless, the mechanismsunderlying miR-145 regulation in the frame of colorectal cancer areelusive.

We have previously highlighted an important interaction betweenPPARγ and SOX9 in a gene regulatory network implicated in colorectal

Fig. 2. Interactions between variables performed by Pearson's product moment correlationon values normalized by logarithmic transformation. (A) Pearson's correlation test demon-strates significant correlation between miR-145 and SOX9 in complete set of CRC patients.(B) No significant correlation between miR-145 and PPARγ in complete set of CRC patientswas found using Pearson's correlation coefficient test. Statistical significancewas determinedby aMann–Whitney test. p=p value; R= Pearson's correlation coefficient; N= number ofCRC patients; and *p b 0.05.

1227A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

carcinogenesis, where PPARγ could influence SOX9 expression andsubcellular localization [9]. The molecular mechanisms underlying theirinteraction, however, are still unclear. Since miR-145 regulates theexpression of SOX9 during chondrogenic differentiation [25], and PPARfamily members are gatekeepers of the expression of specific miRNAssuch as the let-7c and the mir-17-92 cluster [16], we hypothesizedthat PPARγ could inhibit SOX9 in a miRNA-145 dependent manner incolorectal cancer cells.

To address this issue, we evaluated gene expression in tissuespecimens of CRC patients and we took advantage of an in vitro modelrepresented of PPARγ activation and/or miRNA-145-overexpression inCRC derived cell lines. The results show for the first time that miRNA-145 is a target of PPARγ and in turn regulates SOX9 expression,highlighting an important pathophysiological mechanism in the pro-cess of colorectal carcinogenesis.

2. Material and methods

2.1. Patients

We analyzed primary tumors andmatched normal tissue specimensfrom a cohort of 31 CRC patients undergoing curative surgery at ourInstitution (21 men and 10 women, mean age ± SD 66 ± 13 years).Clinical and pathological features of CRC patients are shown in Table 1.Tissue specimens, collected according to the guidelines of the localEthical Committee, were immediately frozen in liquid nitrogen andstored at −80 °C until nucleic acids and proteins extraction. Writteninformed consent was obtained from each patient in accordance withthe Institutional guidelines.

2.2. Cell cultures

CaCo2, SW480, HCT116, and HT-29 cell lines were kept at 37 °C and5% CO2,maintained in Dulbecco's Modified Eagle's Medium (DMEM) orin Minimum Essential Medium (MEM) Alpha Medium supplementedwith 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 ng/mlstreptomycin (Life Technologies Corporation). For rosiglitazone (Ros)treatment, all cell lines were treated with either vehicle (ethanol) or20 μM Ros for 24 h.

2.3. Quantitative real-timePCR (qPCR)

Total RNA including small RNA was extracted from human freshfrozen tissues and colon cancer cell lines using TRIzol Reagent (Life Tech-nologies Corporation) according to the manufacturer's instructions. FormicroRNA detection, 10 ng of RNA were reverse transcribed using theTaq-Man microRNA reverse transcription kit, including miR145 primers(Assay ID: 002278) and endogenous control RNU6B (Assay ID: 001093),in the reaction. Reverse transcription was followed by qPCR assay witha TaqMan PCR Universal master mix and the appropriate microRNA-specific TaqMan probe (Life Technologies Corporation). Expressionlevels of miR-145 were normalized to RNU6B and relative expressionlevels were calculated within each independent experiment using theformula 2−ΔΔCT.

2.4. Bioinformatic prediction of miR-145 targets

Putative miR-145 binding sites in PPARG genomic sequence werepredicted by the RNA22 program http://cbcsrv.watson.ibm.com/rna22.html.

2.5. Reagents

Mouse anti β-Actin, and rabbit polyclonal antibodies directedagainst SOX9 were purchased from Santa Cruz (D.B.A., Milan, Italy).

PPARγ rabbit monoclonal antibody was purchased from Cell Signaling.Rosiglitazonewas purchased fromCayman (AnnArbor,Michigan, USA).

2.6. Transfection assays

Synthetic miR-145 mimic (MSY0000437, QIAGEN) or miR-145-inhibitor (MIN0000437, QIAGEN)was transfected into the colon cancercell lines and examined at varying concentrations (1–25 nM) usingHiPerfect Transfection Reagent (QIAGEN), following themanufacturer'sinstructions. AllStar siRNA (1027280, QIAGEN) was used as negativecontrol. Transfection efficiency was estimated by observing cells usinglight microscopy 96 h after transfection and by evaluating miR-145 vsRNU6B expression levels.

To assess if SOX9 was regulated by miR-145, a miScript TargetProtector (MTP0003089, QIAGEN) designed to bind SOX9 mRNA(NM_000346) 3′UTR (GTTTTTGTTGAAAACAAACTGGAAACTTGTTTCTTTTTTT) was used. As negative control, miScript Target Protector thathas no homology to any known mammalian gene (MTP0000002,QIAGEN), was used. Target protectorswere used to a final concentrationof 100 μM.

Fig. 3. miR-145 expression evaluated by qPCR in colon cancer cell lines upon rosiglitazone treatment. Expression analysis by qPCR of miR-145 in HTC116, SW480, HT-29 and CaCo2 celllines upon rosiglitazone treatment at concentration of 20 μM for 24 h. *, p ≤ 0.05 value; **, p ≤ 0.01; ***, p ≤ 0.001. Dark gray bars correspond to vehicle treated cells, clear gray barscorrespond to rosiglitazone treated cells.

1228 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

2.7. Western blot analysis

CaCo2, SW480, HCT116, and HT-29 cell lines were harvested 48 and96 h after transfection,washed in PBS, and lysed in RIPA buffer (150mMNaCl, 50 MTris–HCl, 1% NP40, 0.1% SDS, 0.5% sodium deoxycholate, atpH 8). Protein lysates were processed for immunoblotting analyses aspreviously described [9].

2.8. Expression vectors and mutagenesis

The pCMV6-XL4 plasmid vector (OriGene, Rockville, MD) containingthe PPARγ coding region was used for Cignal™ PPAR (luc) reporter

Fig. 4.MiR-145 inhibitor transfection and rosiglitazone treatment assay. Expression analysisof SOX9 in CaCo2, SW480, HCT116 and HT-29 cell lines after transfection with miR-145 in-hibitor and rosiglitazone treatment at concentration of 20 μM for 24 h, by using qPCR. Trans-fection of CaCo2 and SW480 cells with miR-145 inhibitor increases SOX9 levels after Rostreatment compared to control (mock). Apartial rescuewasobserved inHT-29 cells,whereasin HCT-116 cells an opposite behaviorwas paradoxically displayed, with an increase in SOX9expression upon rosiglitazone treatment partially counteracted by miR-145 inhibitor.

assay (QIAGEN). The luciferase reporter containing the putative miR-145 promoter in pTransluc vector (P-miR-145) was supplied fromAddgene repository (Plasmid 21496).

A potential PPRE located at −1207 to−1194 bp from the transcrip-tion start site was identified in the miR-145 promoter by bioinformaticanalysis (http://www.classicrus.com/PPRE/). AmiR-145 deletion mutantwas then generated by deleting 13 bp spanning the PPRE (AGGCCAAGGCTCA) and cloned into pTransluc vector (P-miR-145-del). Briefly, thismutagenesis reactionwas conducted in a total volume of 50 μl containing100 ng DNA, 5 μl 10× buffer, 125 pmol (final) of each primer [For: 5′-CATTTTATAATTAAAGAGGGGTTAAGTGACTTAC-3′ and Rev: 5′-GTAAGTCACTTAACCCCTCTTTAATTATAAAATG-3′], 8 μMol (final) dNTPs and 1U Pfu Taq polymerase (Promega) and amplified using 95 °C × 2 min,and 18 cycles of 95 °C × 30 s, 55 °C × 1 min and 72 °C × 12 min as PCRconditions. One μl of Dpn1 (New England Biolabs) was added to digestparental DNA and 3 μl used to transform DH5α cells (Life TechnologiesCorporation). DNA sequencing of colony PCRswere performed to identifythe mutated clone. Plasmid preparation was obtained with Plasmid MidiKit (QIAGEN). The empty pTransluc control vector (P-control) was kindlyprovided by Dr. K.S. Kosik (MCDB, CSB).

2.9. Luciferase reporter assay

To evaluate the activation of PPARγ transcription factor, the Cignal™PPAR (luc) reporter assay kit (QIAGEN)was used. Transient transfectionwas performed in 96-well plates seeding CaCo2, SW480, HCT116, andHT-29 cells at a concentration of 2 × 104 cells per well. The cells wereco-transfected using Attractene Transfection Reagent (QIAGEN) with0.1 μg of pCMV6-XL4-PPARγ expression vector and 0.1 μg ofPPARreporter construct. The PPAR reporter was composed of a mixtureof an inducible transcription factor-responsive firefly luciferase reporterand a constitutively expressed Renilla construct (40:1). Twenty-fourhours post transfection, cells were treatedwith 20 μMof Ros inmediumcontaining 10% FBS for 24 h. Cells were collected and the ratio of fireflyto Renilla luciferase activity was analyzed for PPARγ activity using theDual-Glo Luciferase Reporter Assay System (Promega).

Fig. 5. Target protectors assay. CaCo2, SW480, HCT116 and HT-29 cell lines were cotransfected with miR-145 mimic plus a miScript Target Protector specific for the putative miR-145-binding site of SOX9 (Mimic + target protector), or solely miR-145 mimic or Negative Control miScript Target Protector. SOX9 mRNA levels were quantified by qPCR.

1229A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

For the miRNA-145 promoter assay the CaCo2, SW480, HCT116,and HT-29 cells were co-transfected with P-miR-145, P-miR-145-deland P-controlvectors. Briefly, cell lines were seeded in 96-well platesat a concentration of 2 × 105 cells per well 48 h before the transfectionand then co-transfected with each reporter vector (200 ng/well), 5 ngof the transfection control Renilla vector (pRL-SV40, Promega) and 0.5ml of Lipofectamine 2000. The transfected cells were treated with 20μM of Ros for 24 h. Firefly luciferase/Renilla luciferase ratios were mea-sured by using the Dual-Glo Luciferase Reporter Assay System.

2.10. Statistical analysis

miR-145 expression levels of CRC tissuewere comparedwith those ofadjacent normal mucosa, using the formula 2−ΔΔCt and values werereported as median, 25th percentile (first quartile, Q1) and 75th percen-tile (third quartile, Q3). For continuous variables normal distributionwasverified by the Shapiro–Wilk test and the one-sample Kolmogorov–Smirnov test. For normally distributed variables, hypotheses regardingdifferences among the groups were compared by means of the pairedt-test, Student's t-test or by means of the analysis of variance (ANOVA)as indicated. For non-normally distributed variables, hypotheses regard-ing differences among the groups were compared by means of a non-parametric Wilcoxon signed rank sum test, the Mann–Whitney ranksum test or the Kruskal–Wallis rank sum test as indicated. Statisticalanalyses were also performed on the subjects divided into quartilegroups and by using Pearson's chi-square test to evaluate interactionsbetween variables. p-Values b 0.05 were considered statistically signifi-cant. For protein expression levels and in vitro experiments, the resultsare expressed as means ± SE. All analyses were performed using SPSSStatistical Package v17 (SPSS Inc, Chicago, IL, USA).

3. Results

3.1. Analysis of miR-145 levels in colorectal tumor tissues and cell lines

We measured miR-145 expression levels by qPCR in 31 pairedtumor-normalCRC tissue samples obtained from the same surgically re-moved colonic specimens (9). We observed miR-145down-regulation(median = 0.27, Q1–Q3 = 0.16–0.73, p = 0.0009) in tumor tissuesmatched to normal tissues. Median values, 25th and 75th percentile,and extremes of RNU6B-Ct value/target miR-Ct value are shown inFig. 1A. We also correlated miR-145 expression levels with the clinico-pathological features of the patients: a statistically significant relation-ship was observed between the low miR-145 expression and the MSIstatus, in particular with MSI-H (p = 0.005) (Table 2). We then exam-ined miR-145 expression in four human CRC derived cell lines, CaCo2,HCT116, HT-29 and SW480: we found a barely detectable level in allcell lines, when normalized to normal colon tissue (Fig. 1B). To under-stand the involvement of miR-145 in the interaction of PPARγ withSOX9, we assessed the correlation between miR-145 expression andthe genes encoding these transcription factors (9). A significant inversecorrelation was found between the expression of miR-145 and SOX9mRNA (Pearson correlation coefficient r = −0.412, p = 0.021)(Fig. 2A). No statistically significant correlation was found betweenmiR-145 and PPARγ mRNA (Fig. 2B).

3.2. PPARγ-inducedmiR-145 targets SOX9

Rosiglitazone-mediated activation of PPARγ induces SOX9 down-regulation in CRC cell lines (9). Since miR-145 is negatively correlatedwith SOX9 expression (Fig. 2A), we hypothesized that the ectopicexpression of miR-145 could inhibit SOX9 expression upon PPARγ

1230 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

ligand activation. To this aim, we first tested the effects of rosiglitazoneon miR-145 expression in the four CRC cell lines. Treatment withrosiglitazone (20 μM) for 24 h resulted in a statistically significantupregulation ofmiR-145 in the cell lines examined (Fig. 3). In particular,a higher increase was observed in SW480 (3.2-fold, p= 0.01), in CaCo2(2.6-fold, p = 0.00002), in HT-29 (2.35-fold, p = 0.00003); a lowerncrease was detected in HCT116 (1.5-fold, p = 0.03). We confirmedPPARγ activation by rosiglitazone using a specific Cignal Reporterassay for PPARγ (Supplementary Fig. 1). Subsequently, to evaluate apotential role of miR-145 in SOX9 regulation through PPARγ activation,a loss-of-miR-145 function was applied by transfecting cells with the

Fig. 6. miR-145 promoter transcription is regulated by PPARγ. Experiments were performedputative PPAR response element (PPRE) binding site in the promoter of the miR-145 at the lPPRE, which was deleted in P-miR-145-del. (B) Luciferase assays using P-miR-145p-Luc repnormalized to the internal transfection control Renilla and then divided by the average valuemiR-145 promoter (PPRE) in CaCo2, SW480, HCT116, and HT-29 cell lines. Relative enrichmexperiments, each performed in triplicate (p b 0.05 vs. control).

miR-145 specific inhibitor upon rosiglitazone treatment for 24 h. Asshown in Fig. 4, rosiglitazone treatment significantly lowered SOX9mRNA expression in CaCo2 and HT-29 but not in SW480 and HCT-116cell lines as we have previously reported [9] (Fig. 4): transfectionwith the miR-145 inhibitor induced an increase of SOX9 expressionabove control levels in CaCo2 (1.4-fold, p = 0.0007) and SW480 (2.2-fold, p b0.00001), whereas a partial rescue was observed in HT-29 cells(Fig. 4). HCT-116 cells paradoxically displayed an opposite behavior,with an increase in SOX9 expression upon rosiglitazone treatment[9], partially counteracted by miR-145 inhibitor. Altogether, our re-sults (Figs. 3 and 4) suggest that miR-145 expression is induced by

in transfected CaCo2, SW480, HCT116 and HT-29 cells. (A) Schematic description of theocation of −1194 to −1207 bp. The promoter construct P-miR-145 contains a putativeorter. For derivation of the relative luciferase levels, the raw luciferase activity was firstof the basal construct P-control. (C) ChIP analysis shows high PPARγ enrichment at theent is normalized to control IgG. The data in (B) and (C) were from three independent

Fig. 7. Bioinformatic prediction ofmiR-145 targets,miR-145 transfection assays andWestern blot analysis. (A) PutativemiR-145binding sites in PPARγ genomic sequence predicted by theRNA22 program. (B) Expression analysis by qPCR of SOX9 in CaCo2, SW480, HCT116 and HT-29 cell lines after transfection with miR-145 mimic at 96 h. (C) Western Blot analysisperformed to evaluate protein levels of PPARγ at 96 h after transfection with miR-145 mimic.

1231A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

rosiglitazone-mediated PPARγ activation and inhibits SOX9 expressionin CaCo2 and HT-29, but not in SW480 and HCT-116, colon cancer celllines.

To test the hypothesis that miR-145 could directly bind SOX9mRNA 3′UTR and inhibit its expression, we employed miScripttarget protectors, which are single-stranded, modified RNAs thatspecifically interfere with the interaction of a miRNA with a singleendogenous target, leaving unaffected all other targets. We usedmiScript target protectors designed for the putative miR-145-binding site in the SOX9 mRNA 3′UTR. Whereas miR-145 mimiccould efficiently downregulate SOX9 mRNA in CaCo2, SW480,HCT116, and HT-29 cells, co-transfection of the miR-145 mimicwith its target protector was able to reverse the effects of themiR-145 mimic resulting in increased SOX9 mRNA levels that wasnot significantly different from the untransfected or negativecontrol cells (Fig. 5).

3.3. Transcriptional regulation of miR-145 by PPARγ

The prerequisite for the transcriptional regulation of a given gene byPPARγ is the presence of a PPAR response element (PPRE) in the regu-latory regions [33]. By inspecting miR-145 promoter region, we identi-fied a putative PPRE (AGGCCAAGGCTCA), located at −1207–1194 bp(Fig. 6A) from the transcription start site (TSS). To ascertain the func-tionality of this PPRE as a response element, a 1.5 kb long DNA fragmentcorresponding to the promoter regionwas cloned in a luciferase report-er plasmid, defined P-miR-145. This sequencemotif was thendeleted bysite-directed mutagenesis, generating the plasmid P-miR-145-del(Fig. 6A). CaCo2, SW480, HCT116, and HT-29 cells were transfectedwith P-miR-145 and P-miR-145-del constructs and exposed to 20 μMrosiglitazone. The transcriptional activity of the plasmid carrying thewild type promoter increased upon rosiglitazone treatment, while novariations in luciferase activity were observed with themutant plasmid

Fig. 8. Cell cycle by flow cytometry. Cell-cycle profile was examined by flow cytometry and percentages of cells in G0/G1 (D), S (E), and G2/M (F) phase in CaCo2, HT-29, HTC116 andSW480 cell lines transfected with miR-145 mimic or inhibitors are indicated in the graphs.

1232 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

indicating that the predicted PPRE is responsible for PPARγ trans-activation of miR-145 expression in vitro (Fig. 6B). To validate thebinding of PPARγ to miR-145 promoter, we performed chromatin

immunoprecipitation (ChIP; Supplementary Materials and Methods[32]) analysis in the CRC cell lines. Consistent with PPRE-driven tran-scriptional activity, rosiglitazone treatment induced PPARγ recruitment

Fig. 9.MTT viability assay at 24, 48, 72 and 96 h in CaCo2, SW480, HTC116 and HT-29 cells transfected with miR145 mimic.

1233A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

to the PPRE site upstream of the miR-145 TSS; an isotypic IgG antibodywas used as a negative control (Fig. 6C). These data demonstrate thatmiR-145 is a novel PPARγ transcriptional target.

3.4. Prediction and evaluation of miR-145 putative binding sites intoPPARγ gene

Since interlinking positive and negative feedback loops have beendescribed [15,34] for various miRNAs, we evaluated whether miR-145was able to regulate PPARγ expression. The RNA22 computationalmethod was used to predict putative miR-145 binding sites on thePPARγ gene. An alignment of human PPARγ gene at the predictedmiR-145 binding sites is shown in Fig. 7A. Two putative miR-145 bind-ing sites located upstream and downstream of the PPARγ gene, werepredicted. To assess if miR-145 targets the putative binding sites inthe PPARγ gene, we performed transfection experiments using specificsynthetic miRNAs. Transfection was optimized and the efficiency deter-mined after 96 h bymeans of qPCR. Ectopic expression of miR-145 at 96h did not induce any significant change in PPARγ mRNA in any of thecell lines with the exception of the HT-29 cell line, where we observeda significant down-regulation (FC = −1.4-fold, p = 0.0003) (Fig. 7B).In this cell line, the same change was also detected at PPARγ proteinlevel upon ectopic expression of miR-145 (Fig. 7C). Taken togetherthese data suggest that only in HT-29CRC cell line PPARγ level is bal-anced by a feedback regulation involving miR-145.

3.5. Effects of miR-145 on cell cycle, cell proliferation and invasiveness inCRC cell lines

To investigatewhethermiR-145was able to affect cell cycle progres-sion via SOX9 down-regulation, CaCo2, SW480, HCT116 and HT-29 celllines were transfected with a miR-145 mimic or inhibitor for 96 h. InCaCo2 and SW480 cells transfected with miR-145 mimic as well asinhibitor we observed an increase of the cells in the G0/G1 phase anda reduction of those in the S phase when compared to control cells(Fig. 8). In HT-29 cells transfectedwithmiR-145mimic as well as inhib-itor an increase in the number of cells in the G2/M phase was observedwhen compared with the control cells. Interestingly, miR-145 mimictransfected HCT116 cells exhibited increase in the number of cells insub-GO/G1 phase, an index of pre-apoptosis, and diminution of thosein G0/G1 and S phases when compared with the control cells (Fig. 8).We then evaluated whether miR-145 could affect cell proliferation bycarrying out the MTT assay at four time points (24, 48, 72, and 96 h)upon miR-145 mimic transfection. The ectopic expression of miR-145did not elicit variations in cell growth (Fig. 9). Finally, we investigatedwhethermiR-145 could affect cell motility andmigration by performingthe wound-healing and trans-well assay, respectively. A 30% reduction

in the ability of the cells to induce wound closure was obtained at 72h in HT-29, SW480 and CaCo2 cells transfected with the miR-145mimic, while a 50% reduction was observed in HCT116 cells in compar-ison with control cells (Fig. 10A). In the trans-well assay, HT-29 cellstransfected with miR-145-mimic showed only marginal reduction ininvasiveness at 72 h, while all the other cell lines examined displayeda significant reduction (about 50%) of invading cells (Fig. 10B). Thesedata suggest that miR-145 is implicated in the control of cell migrationand motility, features that are indicative of a differentiated phenotype.To test this hypothesis, we assessed the expression of a series of epithe-lial and mesenchymal markers upon miR-145over-expression. mRNAlevels of epithelial E-cadherin, CK20 and ZO1were increased, particular-ly in HCT116 cells (Fig. 10C); in contrast, the mesenchymal markersSnail, ZEB1, N-cadherin were reduced in all cell lines, suggesting thatmiR-145 sustains a differentiated phenotype and counteracts the inva-sive ability of the cells. Vimentin was significantly reduced in SW480and in HCT-116, and is not expressed in HT-29 and CaCo2. Hematoxylinand eosin staining showed that miR-145 caused a shift towards moreepithelial morphological characteristics with the tendency to formcell-aggregates (data not shown).

4. Discussion

Transcriptional regulators play a key role in the process of colorectalcarcinogenesis, and PPARγ is a leading actor in this scenario. Here, weshow that this nuclear receptor transcriptionally regulates in a directfashion miR-145, which in turn is a fundamental hinge of cell apoptosisand differentiation, and is deregulated in CRC. MiR-145 is controlled byPPARγ and may function as a tumor suppressor directly targeting anddown-regulating SOX9 in colon cancer, influencing the process of intes-tinal differentiation at early stage. We found a severe down-regulationof miR-145 expression in CRC patients, in particular in those withMSI-H status. Defects in the DNA mismatch repair system are presentin about 15% of CRC patients, leading to MSI and generation of a largenumber of substitutions, as well as insertions, deletions, andmutations,which target microsatellite sequences, resulting in truncation or otheralterations of the protein product. The expression of mRNAs with suchframeshift mutations is reduced by the presence of premature stopcodons and results in degradation of some of themutantmRNA throughthe nonsense-mediated decay pathway. In CRC patients, the MSI statusmay influence the response to adjuvant chemotherapy, and in particularirinotecan-containing regimens assure better outcome in patientssuffering from MSI-H tumors when compared to 5-fluorouracil-containingtreatments [35]. Furthermore, the significant inverse corre-lation between the decreased expression of miR-145 and the increasedexpression of SOX9 suggested that miR-145 and SOX9 are coordinatelyregulated in colorectal carcinogenesis. Accordingly, in four CRC derived

1234 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

1235A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

cell lines (CaCo2, HT-29, HCT-116 and SW480) miR-145 expressionwas barely detectable when compared to normal colon tissue, was up-regulated by PPARγ upon rosiglitazone treatment, and induced SOX9downregulation by direct interaction.

The results of our experiments convincingly demonstrate thatmiR-145 was also able to regulate PPARγ expression by a feedbackmechanism specifically in the HT-29 cell line, where a cell specifictranscriptional regulatory mechanism could be envisaged.

MiR-145 upregulation did not change the proliferation ability ofthe CRC cell lines, in agreement with data reported in the gastric cancercell lines BGC-823, SGC-7901, MKN-45 and MKN-28[22], but not inthe LS174T CRC cell line, where miR-145 reduced proliferation [28].Variations of themiR-145 levels influenced cell cycle and induced mor-phological changes. MiR-145 ectopic expression differently influencedcell distribution along the cell cycle in the four CRC cell lines analyzed,highlighting a cell type-specific pattern of response despite a commoninduced downregulation of SOX9, which plays a fundamental role inCRC onset and progression [36]. Furthermore, we provided evidencethat miR-145 has a prominent role on cell motility, invasion and epitheli-al/mesenchymal transition, in agreement with data previously reported[22]. In particular, Gao and coauthors showed that N-cadherin is a directtarget of miR-145, which was found capable to suppress gastric cancercell invasiveness downregulating N-cadherin protein expression throughtranslation inhibition [22]. Accordingly, in our study over-expression ofmiR-145 in HT-29 and HCT116 cells induced significant phenotypicchanges towards epithelial features (E-cadherin, CK20 and ZO1 upregu-lation)while reducingor abolishingmesenchymal stemcell like features(Snail, ZEB1, N-cadherin and Vimentin). These latter effects are mostlikely due to a corresponding decrease of SOX9, a key player of theWNT/β-catenin pathway involved in the maintenance of crypt stem cellself-renewal[36]. Consistently, miR145 maintains a differentiated stateby silencingOCT4, SOX2, andKLF4,master regulators of pluripotency [26].

In conclusion, in this study we demonstrated that in colon cancercell lines PPARγ activation induces the expression of miR-145, whichin turn modulates the expression of SOX9, influencing cell-cycle pro-gression, invasiveness and differentiation status. This newly identifiedmolecularmechanism of colorectal malignancymay lead to therapeuticadvancements. In several cancer models the transcriptional activationof PPARγ through specific agonists is capable to inhibit cell prolifera-tion, stimulate terminal differentiation, promote apoptosis and hinderinflammation. On the other hand, PPARγ agonists seem to exert alsopro-tumorigenic effects, and many clinical trials with PPARγ agonistsyielded contrasting results [37]. Modulation of PPARγ signaling path-way in associationwith othermolecular targetsmight enhance the use-fulness of the effects induced by chemical or biological agents actingthrough this nuclear receptor. Targeting of PPARγ-miR-145-SOX9 axisbrings on physiological effects that could play a chemopreventive or che-motherapeutic role without the harmful pro-carcinogeneticactions deriv-ing from the exclusive PPARγ activation.

SupplementaryMaterials&Methods and Data are available online, atdoi: http://dx.doi.org/10.1016/j.bbamcr.2014.03.003.

Acknowledgments

The study was supported by the “5 × 1000” voluntary contribution,by a MFAG-AIRC (13419) grant (MV) and by grants from the Ministryof Health, Italy (RC1203GA55, RC1203GA56, RC1203ME46 and

Fig. 10. Effect ofmiR-145 onHT-29, HCT116, SW480 and CaCo2 cellmigration, invasion and pheafter transfection ofHT-29,HCT116, SW480 and CaCo2 cellswith themiR-145mimic or scramblfrom transfection counted in at least five microscopic fields. The microphotographs are represethe epithelial andmesenchymalmarkers E- andN-cadherin, CK20, ZO1, Snail, ZEB1 andVimentilevels detected in scrambled transfected cellswere used as calibrator and illustrated as a dotted rcellular phenotypes influencedby their variations. Arrows indicate a positive effect, clades indicaas mean ± SD from at least 3 independent experiments performed in triplicates. *, p ≤ 0.05; *

RC1302ME31) and IRCCS Scientific Institute and Regional General Hospi-tal “Casa Sollievo della Sofferenza”, Opera di Padre Pio da Pietrelcina, SanGiovanni Rotondo (FG), Italy.

References

[1] J. Ferlay, H.R. Shin, F. Bray, D. Forman, C. Mathers, D.M. Parkin, Estimates of worldwideburden of cancer in 2008: GLOBOCAN 2008, Int. J. Cancer 127 (2010) 2893–2917.

[2] A. Jemal, M.M. Center, C. DeSantis, E.M. Ward, Global patterns of cancer incidence andmortality rates and trends, Cancer Epidemiol. Biomarkers Prev. 19 (2010) 1893–1907.

[3] S.A. Lin, N. Barker, Gastrointestinal stem cells in self-renewal and cancer, J.Gastroenterol. 46 (2011) 1039–1055.

[4] C. Crosnier, D. Stamataki, J. Lewis, Organizing cell renewal in the intestine: stemcells, signals and combinatorial control, Nat. Rev. Genet. 7 (2006) 349–359.

[5] J. Rageul, S. Mottier, A. Jarry, Y. Shah, S. Théoleyre, D. Masson, F.J. Gonzalez, C.L.Laboisse, M.G. Denis, KLF4-dependent, PPARgamma-induced expression of GPA33in colon cancer cell lines, Int. J. Cancer 125 (2009) 2802–2809.

[6] I.A. Voutsadakis, Peroxisome proliferator-activated receptor gamma (PPARγ) andcolorectal carcinogenesis, J. Cancer Res. Clin. Oncol. 133 (2007) 917–928.

[7] M. Pancione, L. Sabatino, A. Fucci, V. Carafa, A. Nebbioso, N. Forte, A. Febbraro, D.Parente, C. Ambrosino, N. Normanno, L. Altucci, V. Colantuoni, Epigenetic silencingof peroxisome proliferator-activated receptor γ is a biomarker for colorectal cancerprogression and adverse patients' outcome, PLoS One 5 (2010) e14229.

[8] S. Ogino, K. Shima, Y. Baba, K. Nosho, N. Irahara, S. Kure, L. Chen, S. Toyoda, G.J.Kirkner, Y.L. Wang, E.L. Giovannucci, C.S. Fuchs, Colorectal cancer expression ofperoxisomeproliferator-activated receptor gamma(PPARG, PPARgamma) is associatedwith good prognosis, Gastroenterology 136 (2009) 1242–1250.

[9] A. Panza, V. Pazienza, M. Ripoli, G. Benegiamo, A. Gentile, M.R. Valvano, B. Augello, G.Merla, C. Prattichizzo, F. Tavano, E. Ranieri, P. di Sebastiano, M. Vinciguerra, A.Andriulli, G. Mazzoccoli, A. Piepoli, Interplay between SOX9, β-catenin and PPARγactivation in colorectal cancer, Biochim. Biophys. Acta 1833 (2013) 1853–1865.

[10] J.R. Pekow, U. Dougherty, R. Mustafi, H. Zhu, M. Kocherginsky, D.T. Rubin, S.B.Hanauer, J. Hart, E.B. Chang, A. Fichera, L.J. Joseph, M. Bissonnette, miR-143 andmiR-145 are downregulated in ulcerative colitis: Putative regulators of inflamma-tion and protooncogenes, Inflamm. Bowel Dis. 18 (2012) 94–100.

[11] B.N. Davis, A. Hata, microRNA in Cancer: the involvement of aberrant microRNAbiogenesis regulatory pathways, Genes Cancer 1 (2010) 1100–1114.

[12] C. Liu, K. Kelnar, B. Liu, X. Chen, T. Calhoun-Davis, H. Li, L. Patrawala, H. Yan, C. Jeter,S. Honorio, J.F. Wiggins, A.G. Bader, R. Fagin, D. Brown, D.G. Tang, The microRNAmiR-34a inhibits prostate cancer stem cells and metastasis by directly repressingCD44, Nat. Med. 17 (2011) 211–215.

[13] J. Krol, I. Loedige, W. Filipowicz, The widespread regulation of microRNA biogenesis,function and decay, Nat. Rev. Genet. 11 (2010) 597–610.

[14] A. Piepoli, F. Tavano, M. Copetti, T. Mazza, O. Palumbo, A. Panza, F.F. di Mola, V.Pazienza, G. Mazzoccoli, G. Biscaglia, A. Gentile, N. Mastrodonato, M. Carella, F.Pellegrini, P. di Sebastiano, A. Andriulli, Mirna expression profiles identify driversin colorectal and pancreatic cancers, PLoS One 7 (2012) e33663.

[15] M. Sachdeva, S. Zhu, F. Wu, H. Wu, V. Walia, S. Kumar, R. Elble, K. Watabe, Y.Y. Mo,p53 represses c-Myc through induction of the tumor suppressor miR-145, Proc.Natl. Acad. Sci. U. S. A. 106 (2009) 3207–3212.

[16] Y.M. Shah, K. Morimura, Q. Yang, T. Tanabe, M. Takagi, F.J. Gonzalez, Peroxisomeproliferator-activated receptor alpha regulates amicroRNA-mediated signaling cascaderesponsible for hepatocellular proliferation, Mol. Cell. Biol. 27 (2007) 4238–4247.

[17] A. Singh, C. Happel, S.K. Manna, G. Acquaah-Mensah, J. Carrerero, S. Kumar, P.Nasipuri, K.W. Krausz, N. Wakabayashi, R. Dewi, L.G. Boros, F.J. Gonzalez, E.Gabrielson, K.K. Wong, G. Girnun, S. Biswal, Transcription factor NRF2 regulatesmiR-1 and miR-206 to drive tumorigenesis, J. Clin. Invest. 123 (2013) 2921–2934.

[18] A. Pagliuca, C. Valvo, E. Fabrizi, S. di Martino, M. Biffoni, D. Runci, S. Forte, R. DeMaria, L. Ricci-Vitiani, Analysis of the combined action of miR-143 and miR-145on oncogenic pathways in colorectal cancer cells reveals a coordinate program ofgene repression, Oncogene 32 (2013) 4806–4813.

[19] W.C. Cho, A.S. Chow, J.S. Au, Restoration of tumour suppressor hsa-miR-145 inhibitscancer cell growth in lung adenocarcinoma patients with epidermal growth factorreceptor mutation, Eur. J. Cancer 45 (2009) 2197–2206.

[20] M. Ozen, C.J. Creighton, M. Ozdemir, M. Ittmann, Widespread deregulation ofmicroRNA expression in human prostate cancer, Oncogene 27 (2008) 1788–1793.

[21] A. Karakatsanis, I. Papaconstantinou, M. Gazouli, A. Lyberopoulou, G. Polymeneas, D.Voros, Expression of microRNAs, miR-21, miR-31, miR-122, miR-145, miR-146a,miR-200c, miR-221, miR-222, and miR-223 in patients with hepatocellularcarcinoma or intrahepatic cholangiocarcinoma and its prognostic significance, Mol.Carcinog. 52 (2013) 297–303.

[22] P. Gao, A.Y. Xing, G.Y. Zhou, T.G. Zhang, J.P. Zhang, C. Gao, H. Li, D.B. Shi, Themolecular mechanism of microRNA-145 to suppress invasion-metastasis cascadein gastric cancer, Oncogene 32 (2013) 491–501.

notypic characteristics. (A) Relativewound closure (residualwound area) assessed at 72 hed control; (B) Trans-well assay: thehistograms report thenumber of invading cells at 72 hntative of the trans-well assay after crystal violet staining of the cells; (C) qPCR analysis ofn (VIM) at 72 h inHT-29, HCT116, SW480 andCaCo2 cells transfectedwas reported.mRNAed line; (D) Schemedepicting the axis hardwired by PPARγ, miRNA-145 and SOX9and thete a negative effect. EMT= epithelial-to-mesenchymal transition. All data are represented*, p ≤ 0.01.

1236 A. Panza et al. / Biochimica et Biophysica Acta 1843 (2014) 1225–1236

[23] O. Slaby,M. Svoboda, P. Fabian, T. Smerdova, D. Knoflickova,M. Bednarikova, R. Nenutil,R. Vyzula, Altered expression of miR-21, miR-31, miR-143 and miR-145 is related toclinicopathologic features of colorectal cancer, Oncology 72 (2007) 397–402.

[24] T. Schepeler, J.T. Reinert,M.S. Ostenfeld, L.L. Christensen, A.N. Silahtaroglu, L. Dyrskjøt, C.Wiuf, F.J. Sørensen, M. Kruhøffer, S. Laurberg, S. Kauppinen, T.F. Ørntoft, C.L. Andersen,Diagnostic and prognostic microRNAs in stage II colon cancer, Cancer Res. 68 (2008)6416.

[25] B. Yang, H. Guo, Y. Zhang, L. Chen, D. Ying, S. Dong, MicroRNA-145 regulateschondrogenic differentiation of mesenchymal stem cells by targeting SOX9, PLoSOne 6 (2011) e21679.

[26] N. Xu, T. Papagiannakopoulos, G. Pan, J.A. Thomson, K.S. Kosik, MicroRNA-145 regu-lates OCT4, SOX2, and KLF4 and represses pluripotency in human embryonic stemcells, Cell 137 (2009) 647–658.

[27] J. Zhang, H. Guo, G. Qian, S. Ge, H. Ji, X. Hu, W. Chen, MiR-145, a new regulator of theDNA fragmentation factor-45 (DFF45)-mediated apoptotic network, Mol. Cancer 9(2010) 211.

[28] J. Zhang, H. Guo, H. Zhang, H. Wang, G. Qian, X. Fan, A.R. Hoffman, J.F. Hum, S. Ge,Putative tumor suppressor miR-145 inhibits colon cancer cell growth by targetingoncogene Friend leukemia virus integration 1 gene, Cancer 117 (2011) 86–95.

[29] C.C. Yu, L.L. Tsai, M.L. Wang, C.H. Yu, W.L. Lo, Y.C. Chang, G.Y. Chiou, M.Y. Chou, S.H.Chioum, miR145 targets the SOX9/ADAM17 axis to inhibit tumor initiating cells andIL-6-mediated paracrine effects in head and neck cancer, Cancer Res. 73 (2013)3425–3440.

[30] B. Shi, L. Sepp-Lorenzino, M. Prisco, P. Linsley, T. deAngelis, R. Baserga, MicroRNA145 targets the insulin receptor substrate-1 and inhibits the growth of colon cancercells, J. Biol. Chem. 282 (2007) 32582–32590.

[31] L. Boominathan, The guardians of the genome (p53, TA-p73, and TA-p63) are regu-lators of tumor suppressor miRNAs network, Cancer Metastasis Rev. 29 (2010)613–639.

[32] G. Bolasco, R. Calogero, M. Carrara, M.A. Banchaabouchi, D. Bilbao, G.Mazzoccoli, M. Vinciguerra, Cardioprotective mIGF-1/SIRT1 signaling induceshypertension, leukocytosis and fear response in mice, Aging (Albany NY) 4(2012) 402–416.

[33] E. John, A. Wienecke-Baldacchino, M. Liivrand, M. Heinäniemi, C. Carlberg, L.Sinkkonen, Dataset integration identifies transcriptional regulation of microRNAgenes by PPARγ in differentiating mouse 3 T3-L1 adipocytes, Nucleic Acids Res. 40(2012) 4446–4460.

[34] K. Doberstein, N. Steinmeyer, A.K. Hartmetz, W. Eberhardt, M. Mittelbronn, P.N.Harter, E. Juengel, R. Blaheta, J. Pfeilschifter, P. Gutwein, MicroRNA-145 targets themetalloprotease ADAM17 and is suppressed in renal cell carcinoma patients, Neo-plasia 15 (2013) 218–230.

[35] M.M. Bertagnolli, D. Niedzwiecki, C.C. Compton, H.P. Hahn, M. Hall, B. Damas, S.D.Jewell, R.J. Mayer, R.M. Goldberg, L.B. Saltzm, R.S. Warren, M. Redston, Microsatelliteinstability predicts improved response to adjuvant therapy with irinotecan,fluorouracil, and leucovorin in stage III colon cancer: Cancer and Leukemia GroupB Protocol 89803, J. Clin. Oncol. 27 (2009) 1814–1821.

[36] A.D. Gracz, S.T. Magness, Sry-box (Sox) transcription factors in gastrointestinalphysiology and disease, Am. J. Physiol. Gastrointest. Liver Physiol. 300 (2011)G503–G515.

[37] J.M. Peters, Y.M. Shah, F.J. Gonzalez, The role of peroxisome proliferator-activatedreceptors in carcinogenesis and chemoprevention, Nat. Rev. Cancer 12(2012) 181–195.