Chemopreventive role of olive oil in colon carcinogenesisby targeting noncoding RNAs and methylation machinery

Neha Nanda 1,2,3, Safrun Mahmood2, Alka Bhatia2, Akhtar Mahmood4 and Devinder Kumar Dhawan3

1Department of Pathology, Sol Goldman Pancreatic Cancer Research Center, The Johns Hopkins University School of Medicine, Baltimore,

Maryland, USA2Department of Experimental Medicine and Biotechnology, Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India3Department of Biophysics, Panjab University, Chandigarh, India4Department of Biochemistry, Panjab University, Chandigarh, India

Epigenetic therapy induced by dietary components has become a strong interest in the field of cancer prevention. Olive oil, a

potent dietary chemopreventive agent, control colon cancer, however, its role in epigenetic therapy remains unclear. Thus, we

aimed to investigate the effect of olive oil in a preclinical model of colon cancer by targeting genetic and epigenetic

mechanisms. DMH was used to induce colon cancer in rats; while olive oil was given to separate group of rats along with DMH

treatment. Tumor burden and incidence in DMH and DMH + olive oil-treated rats was observed by macroscopic examination

and histoarchitectural studies. Potent anti-inflammatory, anti-angiogenic and pro-apoptotic activity of olive oil was explored by

gene expression and immunohistochemical studies. The effect of olive oil on epigenetic alterations was examined by detecting

promoter methylation with MS-HRM and dysregulation of miRNA by TaqMan MicroRNA Assay. We observed that olive oil

administration lowered tumor incidence and inhibited the development of tumors in DMH-treated rats. Olive oil markedly

decreased the expression of inflammatory and angiogenic markers and restored the expression of pro-apoptotic markers in

DMH-treated rats. Furthermore, the inverse relationship between gene expression and DNA methylation, deviant miRNA pattern

and miRNA silencing mediated by aberrant DNA methylation was also seen in DMH-treated rats, which was potentially

reversible upon olive oil treatment. Our study concludes that olive oil may play a role in the epigenetic therapy by altering NF-

κB and apoptotic pathways via targeting noncoding RNAs and methylation machinery that affecting epigenome to prevent

colon carcinogenesis.

IntroductionColorectal cancer is the third most common cancer and oneof the leading causes of death worldwide, being responsiblefor about 10% of total cancer-related mortality.1 A hallmarkof CRC lies in its pathogenetic heterogeneity, attained throughgenetic instability, epigenetic alterations and dysregulation ofnoncoding RNA that contribute to its malignant status leadingto the activation of distinct pathways.2

Genetic studies have proven the role of key components ofinflammation such as tumor necrosis factor (TNF-α) andnuclear factor kappa B (NF-κB) in the onset of carcinogenesis.NF-κB regulates the expression of its target genes that areinvolved in angiogenesis (VEGF) and metastasis (MMP-9) forthe progression of CRC.3 Numerous studies have indicated thatNF-κB may inhibit apoptosis either via inhibition of p53 activa-tion or by regulating the anti-apoptotic proteins which in turnblock the caspase cascade.4 Therefore, targeting inflammatorypathway and apoptotic machinery may play an increasinglyimportant role in devising future strategies for cancer therapy.5,6

It is increasingly evident that heritable losses of gene func-tions may be mediated by genetic as well as by epigeneticalterations in colon cancer. The most widely studied epige-netic abnormality in tumorigenesis is DNA methylation.7 Therelationship between DNA methylation and gene expression ishowever complex. Higher levels of gene expressions are oftenassociated with promoter hypomethylation,8 while lower levelsof gene expression are related to promoter hypermethylation.9

The interplay of DNA hypermethylation and hypomethylationdemonstrates two independent functions of significant associ-ations with both molecular and clinicopathological parametersin colorectal cancer.10 Recent studies have shown that epige-netic mechanisms also play an important role in the

Key words: chemoprevention, colon cancer, methylation, miRNA,

olive oil

Conflict of interest: All the authors have no conflict of interest.

DOI: 10.1002/ijc.31837History: Received 26 Apr 2018; Accepted 15 Aug 2018;

Online 28 Aug 2018

Correspondence to: Prof. Devinder Kumar Dhawan, Department of

Biophysics, Panjab University, Chandigarh-160014, India,

E-mail: dhawan@pu.ac.in; Tel.: 91-172-2534121, 2534119 or

Dr. Safrun Mahmood, Geneticist, Department of Experimental

Medicine and Biotechnology, Research Block-B, Postgraduate

Institute of Medical Education and Research (PGIMER), Chandigarh

160012, India, E-mail: mahmoodpgi@gmail.com; Tel.: 91-172-

2747403

International Journal of Cancer

IJC

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

dysregulation of miRNA in the cancer pathogenesis.11,12 miR-NAs act as both oncogenes or tumor-suppressors, dependingupon the inhibition of target genes. Mir-143 and mir-145 arekind of these miRNAs which are downregulated in lung can-cer, breast cancer and colon cancer.13,14 Since mir-143 andmir-145 are involved in the development of colon carcinogen-esis, so they seem to be good targets to study for preventiveapproach against colorectal cancer.

Furthermore, epigenetic alterations are more dynamic aswell as potentially reversible than genetic modifications andcan be influenced by dietary interventions. The growing evi-dence indicated that various dietary agents such as turmeric,green tea and soybean have been shown to modulate DNAmethylation alterations and control incipient malignancy.15

Some reports suggested that the polyphenolic content of thesedietary agents including olive oil, can block the initiation pro-cess or reverse the promotion stage of multistep carcinogene-sis.16,17 Olive oil is a major component of the Mediterraneandiet with numerous beneficial effects due to various antioxi-dants which include Vitamin E (alpha-tocopherol) and phe-nolic compounds such as tyrosol, hydroxytyrosol andoleuropein.18,19 Luteolin, a flavonoid found in olive oilexhibits anti-tumorigenic properties by promoting apoptosisand cell cycle arrest in various cancers such as lung cancer,oesophageal cancer and colon cancer.20 Several in vitro studiessuggest that, besides antioxidant ability, olive oil phenols exertanti-inflammatory and anti-neoplastic activities.21 Recently,Franca et al. revealed the protective effect of olive oil in coloncancer model by decreasing cell proliferation.22 However, veryfew studies were undertaken to explore the effect of olive oilin modulating epigenetic pathways. It has been demonstratedin a recent study that olive oil modulates the expression oftumor suppressor gene (CB1) by altering methylation patternin colon cancer cells, which further provides an evidence forthe chemopreventive activity of olive oil in colon cancer.23

However, Rodríguez-Miguel et al. revealed the tumor-enhancing effects of high-fat diets by using corn oil and oliveoil on breast cancer model by modulating epigenetic patterns.While the effect of olive oil was weaker on breast cancer pro-gression as compared to corn oil by altering histone modifica-tion patterns and reduced hypomethylation levels. However,further studies are required to explore this complex mecha-nism on breast cancer progression by high-fat diets.24 Never-theless, to our knowledge, no study has reported the role ofolive oil in epigenetic therapy by targeting both

hypermethylation and hypomethylation during multistepcolorectal carcinogenesis, especially in vivo conditions.

Therefore, in the present study, we executed a comprehen-sive and systematic analysis to investigate the effect of oliveoil on colon carcinogenesis by targeting genetic and epigeneticmechanisms. The study analyzed the promoter hypermethyla-tion/hypomethylation of various genes involved in the patho-genesis of colon cancer which further were reversible uponolive oil treatment. Furthermore, the present study wasdesigned to look for the interface between DNA methylationand miRNA expression associated with transcriptional silenc-ing and loss of expression of genes encoding for altered cellu-lar pathways in an animal model of CRC.

Materials and MethodsExtra virgin olive oilExtra virgin olive oil was purchased from company Figaro,Deoleo S.A, Spain. The nutrition value for extra virgin oliveoil per 100 g was: Energy 900Kcal, Protein 0 g, Total carbohy-drate 0 g, Total fat 100 g (Saturated fat 13 g, Monosaturatedfat 79 g, Polyunsaturated fat 8 g, Trans fat 0 g), Cholesterol0 mg and Sodium 0 mg.

Animals and experimental designAll animal experiments were performed in accordance withprotocols approved by the Institutional Animal EthicsCommittee of PGIMER, Chandigarh vide ReferenceNo. 50/IAEC/247. Healthy male Sprague Dawley rats withbody weights ranging from 180 to 200 g were divided intothree groups with each group having 12 rats. Group I ratsserved as normal controls and received weekly subcutaneous(s.c.) injections of normal saline. Rats in Group II wereinjected subcutaneously once in a week with freshly prepared1, 2-dimethylhydrazine (DMH) in normal saline at a dosageof 30 mg/kg body weight, for two different time durations of10 and 20 weeks.25,26 Olive oil treatment was given to GroupIII rats, thrice a week at a dose of 1 g/kg body weight throughoral gavaging in addition to DMH treatment as was given togroup II rats for 10 and 20 weeks (Fig. 1). Olive oil treatmentwas started along with DMH treatment.

Body weightThe body weights of rats belonging to normal control and allthe treated groups were recorded every week and maintainedproperly throughout the study.

What’s new?Epigenetic therapy induced by dietary components has strong potential in cancer prevention. Olive oil presents some

chemopreventive activity in colon cancer, but its potential role in epigenetic therapy remains unclear. This study shows that

olive oil prevents colon cancer by reducing the expression of inflammatory and angiogenic genes and enhancing apoptotic

genes and miRNAs expression in a preclinical model. Moreover, the interface between DNA methylation and miRNA expression

associates with transcriptional silencing of genes encoding for altered cellular pathways in colon cancer. Olive oil acts as a

demethylating agent for tumor suppressor genes and hypermethylating agents for tumor promoter genes.

Nanda et al. 1181

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

Processing of tissuesAt the end of the treatment periods (10 and 20 weeks), theanimals were sacrificed under anesthetization with milddiethyl ether. A complete autopsy was performed and all theviscera were examined with the naked eye for any gross path-ological changes. Colons were immediately removed from therats, flushed with ice-cold physiological saline and opened lon-gitudinally. A desired portion of the colon was fixed in 10%formal saline for histopathological analysis. For DNA, RNAand miRNA isolation, the normal colonic tissue from controland tumorous tissue from treated groups were stored at -80 �C.

Tumor or polyp analysisTumor or polyp incidence was calculated as the percentage ofrats with tumors or polyps per group.27 Tumor multiplicitywas calculated as an average number of tumors or polypsfrom each rat in a group.28 Tumor growth was monitored bymeasuring the tumor or polyp length (L) and width (W) byusing calipers and tumor area was calculated by using the for-mula: A = (L × W). Ellipsoid tumor volume was calculated aslength × width 2× π/6 as described by Pettan-Brewer et al.29

Histoarchitectural studiesTissues from normal, as well as tumorous regions of thecolons of animals, were isolated and fixed in 10% paraformal-dehyde in PBS, embedded in paraffin and 4 μm thick sectionswere cut. Further, paraffin sections were stained with hema-toxylin and eosin for routine histological analysis.

ImmunohistochemistryTissues in paraffin block were cut into 4 μm thick sectionsand were processed for immunostaining. The sections wereincubated with primary antibodies, diluted (1:100) in 2.5%normal horse serum (NHS; Vector Labs), left overnight at4 �C, then stained with biotin-conjugated secondary antibodyfollowed by incubation with peroxidase-streptavidin complexusing Vectastain® Universal Quick Kit (Vector Labs). Immu-nostaining was performed by using 3, 30-diaminobenzidine(DAB; Sigma) according to the manufacturer’s instructions.The sections were counterstained with hematoxylin for 1 min,mounted with DPX, and a coverslip was permanently addedfor light microscopy. The after primary antibodies were usedin the present study; NF-κB (Cat No. PC137; Calbiochem(EMD Millipore)), VEGF (Cat No. sc-7269; Santa Cruz

Biotechnology, USA), MMP-9 (Cat No. sc-6840; Santa CruzBiotechnology, USA), Caspase-3 (Cat No. orb 10237; Biorbyt,UK) and Caspase-9 (Cat No. orb1024; Biorbyt, UK). Imageswere analyzed and quantified by the Image-Pro Plus 6.0 soft-ware (Media Cybernetics, Bethesda, MD, USA).

Reverse transcription and quantitative real-time PCR (qPCR)Total RNA was extracted using Trizol reagent according to themanufacturer’s instructions (Invitrogen). After quantification,purified RNA was used to synthesize the first strand cDNA usingthe Applied Biosystems cDNA synthesis kit. Quantitative PCRwas performed by LightCycler® 480 (Roche) using LightCycler®

480 SYBR Green I master (Roche) detection method after themanufacturer’s protocol. Relative quantification was performedusing the Ct method and by using b-actin as a reference.30

Primer sequences were as follows: β-actin mRNA forward 50-TTGCTGATCCACATCTGCTG-30, reverse 50-GACAGGATG-CAGAAGGAGAT-30; Caspase-3 mRNA forward 50-GACAG-GATGCAGAAGGAGAT-30, reverse 50-CATGGGATCTGTTTCTTTGC-30; Caspase-9 mRNA forward 50-CTGAGCCAGATGCTGTCCCATA-30, reverse 50-CCAAGGTCTCGATGTACCAG-GAA-30; NF-κB NFkB mRNA forward 50-GAAGCACAGATAC-CACTAAG-30, reverse 50-AGAGTTCAGCCTCATAGAAG-30;VEGF mRNA forward 50-TGTGCGGGCTGCTGCAATGAT-30,reverse 50-TGTGCTGGCTTTGGTGAGGTTTGA-30; MMP-9mRNA forward 50-TGCTCCTGGCTCTAGGCTAC-30, reverse50- TTGGAGGTTTTCAGGTCTCG-30.

miRNA quantificationmiRNA from colonic tissues was isolated by using mirVanamiRNA Isolation Kit (Ambion). Quantification of the matureform of miRNAs was performed with TaqMan MicroRNAAssay Kit, in accordance with the manufacturer’s instructions(Applied Biosystems). The U6 small nuclear RNA was used asan internal control. The results were interpreted as relativefold change of miRNA expression and were determined using2-ΔΔCp. Primer sequences were as follows:

rno-miR-143-3p 50-UGAGAUGAAGCACUGUAGCUCA-30

rno-miR-145-5p 50-GUCCAGUUUUCCCAGGAAUCCCU-30.

Bisulfite modification and methylation-sensitive high-resolution melt analysisDNA extraction and bisulfite modification. Genomic DNAwas extracted from frozen tissues using a ZymoBead™ Geno-mic DNA Kit (Zymo Research, Orange, CA, USA) accordingto the manufacturer’s instructions. Bisulfite modification ofgenomic DNA was carried out by using the EpiTect® FastDNA Bisulfite Kit (Qiagen), to detect methylation sites. A100% methylated DNA was prepared by methylating the ratgenomic DNA in vitro with the CpG methylase enzyme SssI(New England Biolabs) and an unmethylated DNA (0%) wasisolated from the peripheral blood mononuclear cells of ahealthy rat, were subjected to the bisulfite modification

Figure 1. Timeline of the experiment. [Color figure can be viewed atwileyonlinelibrary.com]

1182 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

procedure to create a series of standards containing methyla-tion levels of 100, 90, 75, 50, 25, 10 and 0%, which were usedas reference controls.

Methylation sensitive high-resolution melt analysis (MS-HRM). The PCR amplification and high resolution meltinganalyses were performed on the 7,500 Fast Step One instru-ment with SDS Software v1.4 (Applied Biosystems) using Melt-Doctor™HRM Master Mix (Applied Biosystems) detectionmethod after the manufacturer’s protocol. Data was analyzedby using the Applied Biosystems MS-HRM software (v.3.0.1.).The normalized HRM profiles allow estimation of methylationlevels of unknown samples run along with the 0–100% methyl-ated standards.31 Primers were designed according to the prin-ciples outlined by Wojdacz and Dobrovic.32 Primers were asfollows: Caspase-3 forward 50-TGGTTTATTTAGTTATTTT-GYGTTATG-30, reverse 50-CRTAACCACCTTCCRATTAA-30;Caspase-9 forward 50-TAGTTATGGAGGAGGTTGATYG-30,reverse 50-TAGTTATGGAGGAGGTTGATYG-30; NFkB for-ward 50-TAGTTATGGAGGAGGTTGATYG-30, reverse 50-ATACCRCCAACAACATCTTC-30; VEGF forward 50-ATTG-TYGTTTAATTGAGATTTTGGT-30, reverse 50-CACACCRCATTAAAAACACA-30; MMP-9 forward 50-TAYGGATAGAA-GAGAAGGGATAGT-30, reverse 50-CTTACCTAACAACCR-CAAAAC-30; miR-143 forward 50-YGTTGTAGTAATTTGTGGTGAT-30, reverse 50- AACRACCATTAAATAAACCAAC-30;miR-145 forward 50-GTYGGATGTAGGGAAAATT-30, reverse50-AAATTCCTAAAAAAACTAAACCRT-30.

Statistical methodsAll statistical analysis was performed using SPSS and Graph-Pad Prism 5.0 software (La Jolla, CA, USA). To determinewhether the data is normally distributed or not, we assessednormality of the data by using Shapiro–Wilk Test due to smallsample size. The significance value of the test (p-value) wasgreater than 0.05 (the chosen alpha level), thus the nullhypothesis was accepted and the data was considered to benormally distributed. Therefore, statistical analysis betweenthe groups was done by using ANOVA, followed by NewmanKeul’s test and Duncan’s test. The results were represented asMean � SD. The correlation between promoter methylationand expression levels, and the association between miRNAand gene expression levels were calculated by using Pearsoncorrelation coefficient in a parametric fashion. Values withp < 0.05 were considered statistically significant.

ResultsBody weight of animalsThe variation in the body weights of the animals subjected todifferent treatments are shown in Table 1. A steady increase inthe body weights was observed from 0 week to 10 weeks ofDMH treatment as compared to normal controls, whereas theincrease was significant (p < 0.05; p < 0.01) after 6 weeks oftreatment. Also, a significant increase in body weights was

observed with olive oil treatment in DMH treated rats as com-pared to DMH-alone treatment after 2 weeks (Table 1a). How-ever, a significant decrease (p < 0.01; p < 0.001) in body weightswas observed after 10 weeks till 20 weeks of DMH treatmentas compared to normal controls. Whereas DMH + olive oiltreatment caused a significant gain (p < 0.001) in body weightsof rats as compared to DMH-alone treatment at 20 weeks(Table 1b).

Olive oil treatment inhibits colon cancer growthIn the present study, we observed a pronounced increase incolonic thickness, vascular congestion and some pedunculatedor sessile polypoid lesions in rats subjected to DMH as well ascombined DMH + olive oil treatments at 10 weeks. However,multiple tumors were grossly observed in rats after 20 weeksof DMH treatment. Interestingly, no evidence of any tumorformation was seen in rats treated with DMH + olive oil at20 weeks, whereas some polypoid lesions were observed inthese rats (Fig. 2a). Further, it was found that treatment withDMH alone resulted in increased tumor incidence, tumormultiplicity and tumor size (measured by tumor area and

Table 1. Effect of DMH and DMH + olive oil treatments for 10 and 20

weeks on body weights of rats

Weeks Normal Control DMH DMH + Olive oil

10 weeks on body weights of rats

0 196.5 196.67 192

2 203.42 � 11.67 233.50 � 11.79 268.5 � 14.3

4 223.12 � 18.79 247.75 � 14.26 299.5 � 18.6xb

6 227.67 � 5.66 263.08 � 5.54 317.66 � 21.6yb

8 240.17 � 2.12 287.58 � 7.42xa 319.5 � 29.9

10 251.68 � 3.51 295.17 � 0.94xa 352.66 � 7.54yb

20 weeks on body weights of rats

0 196.00 194.17 199

2 193.00 � 0.94 204.58 � 4.83 256.87 � 23.15zb

4 208.33 � 4.71 217.58 � 5.77 292.12 � 13.25zb

6 200.63 � 2.26 233.08 � 4.12xa 313.25 � 6.01zb

8 221.82 � 1.20 244.67 � 3.30 312.5 � 30.75zb

10 237.92 � 2.24 234.75 � 7.90 343.25 � 7.77zb

12 261.62 � 18.69 250.58 � 3.89 338.75 � 18.03zb

14 286.77 � 16.36 255.17 � 0.24xa 345.5 � 1.76zb

16 306.40 � 2.97 260.42 � 0.35za 362.5 � 7.77zb

18 319.07 � 3.21 271.75 � 8.60za 354.25 � 3.18zb

20 342.00 � 5.89 266.08 � 0.82za 352.62 � 1.23zb

Weight in grams.Data are presented as Mean � SD.n = 6 for normal control group and n = 6 for each treatment group.xp < 0.05,yp < 0.01,zp < 0.001,avs control,bvs DMH.

Nanda et al. 1183

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

tumor volume). However, olive oil administration along withDMH, resulted in marked decrease in tumor or polyp inci-dence, multiplicity and size (Fig. 2b–e). At 10 weeks, the inci-dence rate of polyps was 100% in DMH-treated while, 66.66%in DMH + olive oil treated rats. Tumor formation wasobserved in 20 weeks DMH-treated rats with 100% incidence,however, there was no evidence of tumors in normal controland olive oil-treated rats. Whereas DMH + olive oil treatedrats showed 50% of polyp incidence at 20 weeks, therebyshowing improvement in 20 weeks as compared to 10 weeksof treatments with olive oil (Fig. 2b). With regard to tumor orpolyp multiplicity, olive oil (p < 0.001) significantly inhibitedtumor formation in the colon as compared to DMH-alonetreated rats at 20 weeks (Fig. 2c). Tumor or polyp area wassignificantly reduced in animals administered DMH + oliveoil as compared to those given DMH-alone for 20 weeks(p < 0.001) (Fig. 2d). Similarly, Tumor or polyp volume wasalso significantly decreased with olive oil treatment in DMH-treated rats at 20 weeks (p < 0.001) (Fig. 2e).

Further, while conducting the histoarchitectural examina-tion, 10 weeks DMH-treated colons confirmed villous ade-noma by an increase in inflammation, high-grade hyperplasiaand villi-form transformation of the mucosa as compared tocontrols which conserved their normal mucosal architecture.Moreover, 20 weeks DMH-treated rats showed poorly differ-entiated carcinoma, characterized by scanty cytoplasm, a high

mitotic rate and an ovoid to angulated shape cells that containmucin vacuoles in their cytoplasm as compared to their con-trol counterparts. However, all DMH + olive oil treated ratsshowed lymphoid infiltration, mild atypia and minimalinflammation in the mucosa of the colon. Importantly, nosign of malignancy was observed in these animals at 10 and20 weeks. Therefore, olive oil treatment was markedly effectiveto prevent colon cancer growth at both the stages (Fig. 2f ).

Anti-inflammatory and pro-apoptotic roles of olive oil incolon carcinogenesisOlive oil treatment reduces mRNA expression of NF-κB, VEGFand MMP-9. mRNA expression of NF-κB was significantlyupregulated in DMH-treated rats in comparison with normalcontrols (~7.5-fold, p < 0.001; ~10.9-fold, p < 0.001) at 10 and20 weeks. However, olive oil treatment significantly attenuatedthe expression levels of NF-κB mRNA in DMH + Olive at10 weeks (~4.5-fold, p < 0.01) and 20 weeks (~8.5-fold,p < 0.001) in comparison to DMH treatments at both timeduration respectively, as shown in Figure 3a. In addition, VEGFmRNA expression was also significantly increased in DMH-treated rats compared to normal controls (~2.2-fold, p < 0.05;~7.6-fold, p < 0.001) at 10 and 20 weeks, respectively. While,mRNA expressions of VEGF was significantly downregulated inolive oil treated rats in comparison to DMH-treated rats at10 weeks (~twofold, p < 0.05) and 20 weeks (~sevenfold,

Figure 2. Olive oil treatment prevents colon cancer growth. (a) Photomicrographs of normal control, DMH-treated, and DMH + olive-treatedcolon; (b) Tumor or polyp incidence was revealed by percentage of rats with tumors or polyps per group; (c) Tumor or polyp multiplicity foreach treatment group is expressed as average number of tumors or polyps from each rat in a group; (d) Tumor area was calculated as,A = (L × W); (e) Tumor volume as length × width2× π/6 in all treatments. Data are presented as Mean � SD; (f ) Hematoxylin and eosin stainof tissue from control and treated rats, Magnification ×200; (n = 6). [Color figure can be viewed at wileyonlinelibrary.com]

1184 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

p < 0.01) (Fig. 3b). Moreover, MMP-9 mRNA expression wasalso significantly increased in DMH-treated rats compared tonormal controls (~2.2-fold, p < 0.01; ~ninefold, p < 0.001) at10 and 20 weeks, respectively. Whereas mRNA expression ofMMP-9 was significantly downregulated in DMH + olive oiltreated rats in comparison to DMH-treated rats at 10 weeks(~2.4-fold, p < 0.05) and 20 weeks (~7.6-fold, p < 0.05) (Fig. 3c).

Olive oil treatment increases caspase-3 and caspase-9 mRNAexpression. Expression of caspase-3 mRNA was significantlydecreased in DMH-treated rats as compared to normal controls(~0.6-fold, p < 0.001) at 10 weeks and (~0.7-fold, p < 0.001) at20 weeks. In contrast, olive oil treated rats had significantlyhigher caspase-3 mRNA expression levels at 10 weeks(~1.4-fold, p < 0.001) and 20 weeks (~threefold, p < 0.001) ascompared to DMH-treated rats in both time durations, respec-tively (Fig. 3d). Similarly, mRNA expression of caspase-9 wassignificantly decreased in DMH-treated rats (~0.65-fold,p < 0.001) at 10 weeks and (~0.95-fold, p < 0.001) at 20 weeks.However, olive oil treatment cause significantly higher expres-sion level of caspase-9 mRNA at 10 weeks (~0.6-fold, p < 0.01)and 20 weeks (~3.6-fold, p < 0.001) as compared to DMH-treated rats (Fig. 3e).

Olive oil modifies localization and expression of pro- andanti-apoptotic proteins. Further, we studied the effect ofolive oil on the expression of various pro- and anti-apoptoticproteins in colonic tissues. Immunohistochemical analysis of

colonic sections in DMH-treated rats revealed an enhancedexpression of anti-apoptotic protein NF-κB p65, mainly in thenuclei of epithelial cells, while more cytoplasmic staining wasobserved in control and olive oil treated groups at both10 and 20 weeks. The expression of VEGF and MMP-9 wereincreased in the cytoplasm and membrane of epithelial cells inDMH-treated rats. However, reduced immunostaining ofVEGF and MMP-9 was observed in DMH + olive oil treatedrats which were essentially similar to that in controls. How-ever, a markedly reduced cytoplasmic staining of apoptoticproteins, that is caspase-3 and caspase-9 was observed inDMH-treated rats relative to controls, while olive oil treat-ment proved beneficial as it resulted in a moderate to highincrease in expression of these apoptotic proteins inDMH + olive oil treated rats (Fig. 4a). The quantification ofthese proteins in all the groups was shown in Figure 4b. Theresults of protein expression were consistent with gene expres-sion results.

Olive oil modulates aberrant methylation pattern in coloncarcinogenesisTo evaluate the effect of olive oil on DNA methylation alter-ations, we used Methylation Sensitive-High Resolution Melt-ing analysis to simultaneously detect hypomethylation andhypermethylation.

Olive oil induces hypermethylation of NF-κB, VEGF and MMP-9.On the basis of methylation percentage, we observed

Figure 3. Effect of olive oil on gene expression was assessed by qPCR in colon cancer. (a) NF-κB mRNA expression pattern in normal controland treated animals; (b) VEGF mRNA; (c) MMP-9 mRNA; (d) Caspase-3 mRNA and (e) Caspase-9 mRNA. Data are presented as Mean � SD;(n = 6). [Color figure can be viewed at wileyonlinelibrary.com]

Nanda et al. 1185

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

1186 Chemopreventive role of olive oil in colon carcinogenesis

Can

cerTherapy

andPrevention

hypomethylation in the CpG island within the promoterregion of NF-κB in DMH-treated rats (20%, p < 0.05) com-pared to controls, which however was increased considerablyupon olive oil treatment at 10 (p < 0.05) and 20 weeks(p < 0.01). Similarly, the promoter region of MMP-9 showedhypomethylation in DMH-treated rats (~10%, p < 0.001) incomparison to controls. However, olive oil caused a significantincrease in methylation of the MMP-9 promoter at 10 weeks(61%, p < 0.001) and 20 weeks (81%, p < 0.001). We alsoobserved promoter hypomethylation of VEGF with DMHtreatment and reverse pattern with olive oil, albeit the changeswere nonsignificant (Fig. 5).

Olive oil induces demethylation of caspase-3 and caspase-9.We observed hypermethylation in the promoter region ofcaspase-3 in DMH-treated rats at 10 weeks (74%, p < 0.05)and 20 weeks (81%, p < 0.001) as compared to controls.While, the olive oil treatment in combination with DMH at20 weeks, caused demethylation in caspase-3 gene (40.87%,p < 0.001). We also found a highly significant promoter meth-ylation of caspase-9 with DMH treatment (~75%, p < 0.001)in comparison to normal controls, which however wasdecreased upon olive oil treatment (~57%) (Fig. 5).

miR-143 and miR-145 expression in colon cancerTo characterize the expression of miR-143 and miR-145 incolonic tissues, we used TaqMan qRT–PCR. As shown inFigure 6, the expression of miR-143 and miR-145 was signifi-cantly lower in DMH-treated rats (p < 0.001) as compared tocontrols. However, olive oil treatment significantly increasedthe expression of miR-143 and miR-145 (p < 0.001) inDMH + Olive oil-treated rats as compared to DMH-treatedrats at both 10 and 20 weeks.

Methylation of miR-143 and miR-145 in colon cancerSince DMH treatment showed nearly complete loss of miR-143and miR-145 expression; therefore, we analyzed whether thesemiRNAs may be silenced in colon cancer by an epigenetic mech-anism. Notably, we observed promoter regions of miR-143 andmiR-145 were heavily hypermethylated (89.5%, p < 0.01; 80.7%,p < 0.001) respectively, compared to controls. However, olive oiltreatment reduced the methylation levels of these miRNAs,though the effect was significant only on mir-145 (Fig. 7a, b).

Correlation analysisCorrelation between miRNA levels and gene expression ofCaspase-3, Caspase-9, NF-κB, VEGF and MMP-9. There wasa significantly strong positive correlation was observed

between mir-143 levels and mRNA expression of caspase-3(r = 0.9601) and caspase-9 (r = 0.9978). Similarly, a strongpositive correlation was observed between mir-145 levels andgene expression of caspase-3 (r = 0.9038) and caspase-9(r = 0.8601). Although, the positive correlation was signifi-cantly strong for caspase-3 only, but these results clearly dem-onstrated that miRNA and caspases levels tend to increase ordecrease together with or without treatments (Table 2).

However, a moderate negative correlation was observedbetween mir-143 and gene expression of NF-κB (r = −0.5903),VEGF (r = −0.5126) and MMP-9 (r = −0.445). While, a moder-ate to strong negative correlation between mir-145 levels andmRNA expression of NF-κB (r = −0.8583), VEGF (r = −0.7385)and MMP-9 (r = −0.6794). Thus, these results suggested thatincrease in miRNA levels tends to decrease of NF-κB, VEGFand MMP-9 expression or vice versa (Table 2).

Correlation between promoter methylation and expression of NF-κB, VEGF, MMP-9, Caspase-3, Caspase-9, mir-143 and mir-145.Further, we computed the correlation between DNA methyla-tion levels and gene expression of NF-κB, VEGF, MMP-9,caspase-3 and caspase-9 in all the groups by Pearson correla-tion coefficient. A significant strong negative correlation wasobserved for NF-κB (r = −0.924) and moderate to strong cor-relation was observed for MMP-9 (r = −0.685) and VEGF(r = −0.810). In addition, there was a strong negative correla-tion between methylation levels and gene expression ofcaspase-3 (r = −0.756), however, a weak negative correlationwas observed in caspase-9 (r = −0.186). Thus, the results con-firmed the inverse relationship between gene expression andmethylation levels of above genes (Table 3).

Moreover, while assessing the correlation between pro-moter methylation and miRNA levels, we observed a signifi-cant strong negative correlation for mir-145 (r = −0.911) anda weak correlation for mir-143 (−0.294) (Table 3), which indi-cate that increase of methylation levels tend to decrease ofmiRNA levels for mir-143 and mir-145 (Table 3).

DiscussionThe relationship between dietary factors and chemopreventionis incredibly important, that can target the multifactorial pro-cess of colon carcinogenesis. Colon cancer can be summed upto represent a complex interplay between genetic and epige-netic abnormalities. Recent studies indicated that epigeneticalterations are modifiable by dietary factors due to theirreversible nature. Thus, epigenetic therapy has been identifiedas a promising strategy for reducing cancer incidence andmortality. In the present study, we have simultaneously

Figure 4. (a) Immunohistochemistry for colonic tissues from normal control and treated rats. Representative immunostained sectionsdemonstrating NF-κB, VEGF, MMP-9, Caspase-3 and Caspase-9 localization and expression. In each case, the signal for protein is representedby brown color due to DAB and blue signal due to hematoxylin counterstain. Magnification ×200; (n = 6). (b) Quantitativeimmunohistochemical expression of (A) NF-κB; (B) VEGF; (C) MMP-9; (D) Caspase-3 and (E) Caspase-9. Data are presented as Mean � SD;(n = 6). [Color figure can be viewed at wileyonlinelibrary.com]

Can

cerTherapy

andPrevention

detected two opposite epigenetic phenomena (hypomethyla-tion and hypermethylation), and have shown that methylationis inversely related to gene expression. We demonstrated forthe first time that olive oil has the demethylating activity tothe hypermethylated genes (caspase-3 and caspase-9), and itmay also act as demethylation inhibitor if its targets would behypomethylated genes (NF-κB, VEGF and MMP-9). More-over, we observed the interrelationship between DNA methyl-ation and miRNA expression pattern which may be associatedwith mir-143/145 silencing and also analyzed the demethylat-ing activity of olive oil to these miRNAs.

In the present study, DMH caused a decrease in bodyweight after 10 weeks till 20 weeks of treatment in DMH-

treated rats as compared to normal controls, while olive oiltreatment caused a significant increase in body weight inDMH + olive oil treated rats at both 10 and 20 weeks. Thesefindings are similar to our earlier findings where we haveobserved a decrease in body weight in DMH-treated rats ascompared to normal controls,33 in addition, the study gainssupport from Franca et al., where they reported an increase inbody weight with olive oil treatment.22 Macroscopically, it wasobserved that DMH administration caused an increased rateof tumor incidence, tumor multiplicity and tumor size (mea-sured by tumor area and tumor volume). However, a signifi-cant decrease in tumor growth was observed with olive oiltreatment. Our results are supported by reports published in

Figure 5. Percentage methylation of the genes in normal control and treated groups. Representative samples show (A) Aligned melt curve,(B) Difference plot normalized to the 0% methylated standard DNA for methylated standard curves of 100 to 0% (DNA standards with 100 to0% methylation) represented in different colors and differential fluorescence, and Percentage change in methylation of (C) NF-κB; (D) MMP-9;(E) VEGF; (F) Caspase-3 and (G) Caspase-9 promoter region using Methylation sensitive-high resolution melting (MS-HRM) analysis. Data arepresented as Mean � SD; (n = 6). [Color figure can be viewed at wileyonlinelibrary.com]

1188 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

the literature which demonstrated that DMH treatmentenhanced tumor incidence, tumor multiplicity and tumorgrowth.34,35 However, Sánchez-Fidalgo et al. reported thatextra virgin olive oil fed female C57BL/6 mice showed lesstumor incidence and multiplicity, and confirmed that EVOOdiet has protective/preventive effect in the Ulcerative Colitis-associated colorectal cancer.36 The results were confirmed byhistoarchitectural studies that showed a clear demonstrationof inhibition of tumor growth by administration of olive oil.Based on the results of H&E staining of the colonic tissues, itis well documented that administration of DMH in vivo pro-motes colon carcinogenesis.33,35 However, Sánchez-Fidalgo

et al. demonstrated that the number of adenocarcinomas andhigh-grade dysplasia were reduced by 22% and 56%, respec-tively with extra virgin olive oil (EVOO) diet in dextransodium sulphate-treated female C57BL/6 mice. It was demon-strated that a diet rich in EVOO effectively prevented thedevelopment of dysplasia and/or cancer in the experimentalmodel.36

A connection between inflammation and tumor develop-ment by activation of nuclear factor kappa B (NF-κB), wassuggested by Wang et al.3 Earlier, studies have reported thatconstitutive NF-κB activation and its upregulation increasesthe risk of colorectal cancer by triggering inflammatory

Figure 6. Effects of olive oil on miRNA expression in normal control and treated rats. (A) mir-143 and (B) mir-145 expression levels by TaqManqRT-PCR. Data are presented as Mean � SD; (n = 6).

Figure 7. Percentage change in promoter methylation of (A) mir-143 and (B) mir-145 using Methylation sensitive-high resolution melting (MS-HRM) analysis. Data are presented as Mean � SD; (n = 6). [Color figure can be viewed at wileyonlinelibrary.com]

Nanda et al. 1189

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

pathway. Furthermore, NF-κB may contribute to the progres-sion of CRC by regulating the expression of diverse targetgenes that are involved in angiogenesis, metastasis andinvasion.37–39 On similar lines, we also observed increasedexpression of NF-κB and its target genes (VEGF and MMP-9)in DMH-treated rats at 10 and 20 weeks, while we witnessedthe decreased expressions of NF-kB, VEGF and MMP-9 witholive oil treatment. We also validated these findings by immu-nohistochemical analysis and found the similar observationswith all the treatments as that observed at the gene level. It iswell documented that the promoter regions of MMP-9 andVEGF genes possess binding sites for NF-κB, which may formthe basis for the regulation of these molecules via NF-κB-mediated pathway.39–41 Therefore, these findings implicateNF-κB inhibition as an important therapeutic target in CRC.Various studies have reported that olive oil decreases theexpression of NF-kB in metabolic syndrome and chronicinflammatory diseases. Recently, Boss et al. have providedshreds of evidence that olive oil and its polyphenols have ananti-inflammatory role in various diseases via inhibiting NF-κB activation.42 Therefore, our data further validate the abovefindings in context with colon carcinogenesis wherein olive oiladministration prevents inflammatory cascade via inhibitingNF-κB activation in a chronic inflammatory disease like colo-rectal cancer.

Gene regulation in cancer can be disrupted either throughgenetic alterations, or epigenetic alterations.43 Methylation isthe broadly studied epigenetic alteration in cancer.44 Methyla-tion is catalyzed by DNA methyltransferase (DNMT)enzymes,45 which play a central role in maintaining the estab-lished methylation patterns in cell division.46 Recently, Nagar-aju et al. suggested that NF-κB overexpression also resulted inoverexpression of DNMT-1 in pancreatic cell lines.47 SinceDNMT-1 transcription is regulated by Sp-1/NF-κB,48 thusNF-κB might be a better target for inhibition of DNMT-1 andmethylation. Hence, inhibition of DNMT-1 via NF-κB meth-ylation was associated with overexpression of silenced tumorsuppressor genes. Our study is understandably the first of itskind that demonstrated the hypomethylation of NF-κB and itstarget genes in colorectal cancer rat model. Several studieshave indicated that controlling transcription via epigeneticmodulation on the promoter of VEGF and MMP-9 genes pro-vides a novel approach for the treatment of cancer.49,50 Shu-keir et al.51 demonstrated that DNA hypomethylation controlsthe activation of multiple tumor-promoting genes such asMMP-2 and VEGF and suggested that hypermethylation ther-apy might be a possible approach to target the demethylationmachinery. On similar lines, we also observed hypomethyla-tion of VEGF and MMP-9 in DMH-treated rats, while wefound that olive oil acted as a demethylation inhibitor andsuppressed the tumor growth via inhibition of tumor-promoting genes (NF-κB, VEGF and MMP-9).

In regard to the olive oil acts as a demethylating inhibitorfor NF-kB, VEGF and MMP-9, we desired to mention thatsince VEGF and MMP-9 are targets of NF-kB which are dem-onstrated by various earlier studies, so their methylation pat-tern might be dependent on NF-kB levels. In this regard,correlation studies also provide a strong evidence, where weobserved a significantly strong negative correlation betweenmethylation pattern and gene expression level of NF-kB,which may lead to the moderate inverse correlation for VEGFand MMP-9. Since NF-kB is a master regulator and regulatesvarious inflammatory or anti-apoptotic pathways, thus itsaberrant methylation pattern modulates various factors.DNMT-1 could be one factor that plays a major role in NF-kB pathway or NF-kB regulates its downstream molecules byDNMT-1 as suggested by the literature, thus it is worthwhile

Table 3. Correlation between expression and promoter methylation

levels of Caspase-3, caspase-9, NF-kB, VEGF, MMP-9, mir-143 and

mir-145

Pearson Correlation

r Sig. (2-tailed)

Caspase-3 −0.756 0.139

Caspase-9 −0.186 0.764

NF-kB −0.924* 0.025

VEGF −0.810 0.097

MMP-9 −0.685 0.202

mir-143 −0.294 0.631

mir-145 −0.911* 0.031

Correlation is significant at*p < 0.05.

Table 2. Correlation between mir-143/145 levels and gene expression of Caspase-3, caspase-9, NF-kB, VEGF and MMP-9

Caspase-3 Caspase-9 NF-kB VEGF MMP-9

mir-143 Pearson Correlation r = 0.9601** r = 0.9978*** r = −0.5903 r = −0.5126 r = −0.4445

Sig. (2-tailed) p = 0.0095 p = 0.0001 p = 0.2947 p = 0.3771 p = 0.4532

mir-145 Pearson Correlation r = 0.9038* r = 0.8601 r = −0.8583 r = −0.7385 r = −0.6794

Sig. (2-tailed) p = 0.0353 p = 0.0615 p = 0.0627 p = 0.1541 p = 0.2071

Correlation is significant at*p < 0.05,**p < 0.01,***p < 0.001.

1190 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

to explore the DNMT-1/NF-kB pathway in future studies.However, we could not find the exact mechanism behind thealtered methylation level of NF-kB by olive oil, but as we havementioned earlier, this is a first study which mentioned theDNA methylation level of NF-kB in colon cancer and itshypermethylation by olive oil. Thus, our study forms a step-ping stone for further elucidation of the mechanism behindthe altered methylation level of NF-kB by olive oil.

Cancer cells are believed to be resistant to treatments thatinduce cell death due to different mechanisms, and one of themechanisms is pro-apoptotic gene silencing mediated by pro-moter DNA hypermethylation. The present study demon-strated that gene expression of caspase-3 and -9 wasdownregulated in DMH treated rats, while the expressionlevels of these genes were augmented with olive oil treatmentat 10 and 20 weeks. Similar to gene expression results, ourimmunostaining data corroborated the expression and locali-zation of caspase-3 and caspase-9 proteins in the colonic sec-tions of controls, DMH and DMH + olive oil treated rats. Theloss of expression of caspases in colorectal cancer was welldocumented by various in vitro and in vivo studies. However,a study by Odabasoglu et al. demonstrated the pro-apoptoticrole of olive oil by activation of caspase-2, −3, −8 and − 9 inTi-implanted rabbits.52 Moreover, they provided an evidence,which showed that besides caspase-dependent mechanism,olive oil-induced apoptosis was also caused by its antioxidantproperty. Recently, Leong et al. revealed that a natural product(Aglaia exima) triggered apoptosis by activation of caspasecascade and inhibiting NF-kB translocation in colon cancercell lines.53 The plausible reason behind the above findingsmight be the NF-κB promoted cell survival by blocking thecaspase cascade via anti-apoptotic genes4 or anti-apoptoticmembers of the Bcl family that inhibit cytochrome c releasewhich further blocks the intrinsic apoptotic pathway.5 Besidesthe above mechanisms, caspase expression may be impairedby epigenetic alterations such as promoter hypermethylation,thus hypermethylation appears to be a primary mechanism ofdisabling the caspase cascades in tumors.54 Recently, Hervouetet al. 55 in an in vitro study also showed that Decitabine5-azacytidine, a demethylating agent, reversed the process ofhypermethylation on caspase-9 promoter resulted in the initi-ation of apoptosis. The present data also indicated thedemethylating activity of olive oil, which resulted in a re-expression of these genes, in rat colon, which further verifiedby correlation studies, indicating the inverse relationshipbetween methylation and gene expression of caspases. Thus, itis a significant finding with regard to chemoprevention activ-ity of olive oil, which has not been reported, earlier.

We wish to mention that there are ample of studies whichfocused on the chemopreventive activity of olive oil, mainlyepidemiological studies. Since olive oil is an important dietaryfactor in Mediterranean diet, and there is some evidence whichindicates that the incidence rate of colorectal cancer is muchlower in Mediterranean countries as compared to Northern

countries such as USA and UK.56 Recently, some case–controlstudies revealed that Mediterranean diet reduced the risk ofcolorectal cancer and lowered the mortality rate.57–59 Moreover,a clinical trial reported in JAMA Internal Medicine revealedthat Mediterranean diet supplemented with extra virgin oliveoil showed a 62% lowered risk of breast cancer.60 However,very few studies have been undertaken in exploring the molecu-lar mechanism behind the anti-cancer activity of olive oil. Pro-moter methylation is one of those mechanisms which are notsuccessfully elucidated yet in regard to olive oil. As we dis-cussed that olive oil modulates hypermethylation of tumor sup-pressor genes (caspase-3 and caspase-9) and acts as ademethylating agent, it must be through the NF-kB-mediatedinhibition of DNMT-1 by preventing NF-kB activation and fur-ther restoring the expression of tumor suppressor genes. Unfor-tunately, we did not check the methylation level andtranscriptional expression of DNMT-1, as our study was mainlyfocused on the analyses of various inflammatory and apoptoticmarkers, though we did rely on literature in regard to the roleof DNMT-1/NF-kB in methylation.

Beside methylation, miRNA dysregulation is strongly asso-ciated with colon carcinogenesis.61 Recently, Chen et al.revealed the differentially expressed miRNAs which are con-sistently deregulated in colorectal cancer. miR-143 and miR-145 were one of those miRNAs which showed significantlydecreased expression in CRC tissues as compared to normaltissues.62 Similarly, we also observed the significant down-regulation of miR-143 and miR-145 in colons of DMH treatedrats, while the expression of these miRNAs was restored uponolive oil treatment. These results are also supported by a num-ber of other studies reported in literature, which determinedthat levels of miR-143 and -145 are downregulated in coloncancer cell lines and CRC patients,63,64 though the transientoverexpression of these miRNAs was associated withincreased caspase −3, −8 and − 9 activities and decreasedextracellular-regulated protein kinase 5, nuclear factor-kB andBcl-2 protein expression.65,66 Similarly, we also observed thepositive correlation between mir-143/145 and caspase-3/9,which strongly indicated that upregulation or downregulationof these tumor suppressor miRNAs is directly related toincrease or decrease of caspases. Moreover, the negative corre-lation was observed between mir-143/145 and NF-κB or itstarget genes (VEGF and MMP-9), showed the inverse rela-tionship between them. The mechanism involving the down-regulation of these miRNAs may be the transactivation ofmir-143 and mir-145 from NF-κB and p53 transcriptional fac-tors respectively as suggested by Zhang et al. and Fenget al.66,67 It is likely that olive oil also modulates the NF-κB/p53 pathway which may result in mir-143 and -145 re-expression leading to its chemopreventive activity. Recently, ithas been demonstrated by some in vitro studies that restoringthe expression of miR-143 and mir-145 induced the tumorcell sensitization to various chemotherapeutic drugs and theup-regulation of these tumor suppressor miRNAs inhibited

Nanda et al. 1191

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

the colon cancer cell proliferation and cell cycle progres-sion.68,69 These reports suggested that restoring the levels ofmiR-143 and miR-15 could potentially be used as therapeuticsignatures for colon cancer therapy.

In our study, we also understand the dysregulation ofmicroRNA regulatory network in cancer and observed thehypermethylation of microRNAs that undergo silencing incolorectal cancer. Various reports indicate that transcriptionalsilencing of microRNAs may be related to the aberrant DNAmethylation of a CpG island of 50 regulatory regions located inthe host gene 50UTR.70 Recent studies demonstrated that pro-moters of tumor suppressor miRNAs were aberrantly hyper-methylated in different cancers, which lead to the silencing ofmiRNA expression and further its restoration by decreasing

methylation level provided a therapeutic potential.71,72 Donzelliet al. observed the hypermethylation on miR-145 promoterwhich further led to its downregulation in brain metastasis.73

Moreover, miRNA silencing mediated by aberrant promoterDNA methylation can potentially be reversed by hypomethylat-ing agents, and hence may present a new therapeutic target incancer.74 Since, we also observed a significant inverse correla-tion between promoter methylation and mir-145 expression,which confirmed that aberrant hypermethylation of this tumorsuppressor miRNA may trigger its downregulation, which fur-ther reversed by olive oil treatment. However, very few studiesare reported in the literature on the effect of chemopreventiveagents involving miRNA expression modulated by aberrantDNA methylation, thus our study is first of its kind to unravelthe regulatory role of methylation on miRNA expression. Butfurther studies are needed to substantiate this line of the sug-gested mechanism.

ConclusionsThe current study revealed the potential effect of olive oil inpreventing the colon carcinogenesis by unraveling the mecha-nisms involved in an aberrant epigenetic pattern in experi-mental animals. The mode of action of olive oil may involveinhibition of NF-κB mediated pathway and restoration of apo-ptotic pathway via regulation of miRNAs and methylationpattern as proposed in Figure 8. Thus, the current findingsopen up a new vista for epigenetic approach by targeting non-coding RNAs and other epigenetic mechanisms with the die-tary intake to control incipient malignancy.

AcknowledgmentsThe authors thank Mr. Sunil Kumar for helping to carry out the animalwork and Ms. Rekha Verli for her technical assistance. The authors thankIndian Council of Medical Research (ICMR), New Delhi, India, for finan-cial support to the research work.

References

1. American Cancer Society. Colorectal cancer facts &figures 2014–2016, Atlanta: American CancerSociety, 2014.

2. Link A, Balaguer F, Shen Y, et al. Curcumin mod-ulates DNA methylation in colorectal cancer cells.PLoS One 2013;8:57709.

3. Wang S, Liu Z, Wang L, et al. NF-kappa B signal-ing pathway, inflammation and colorectal cancer.Cell Mol Immunol 2009;6:327–4.

4. Deveraux QL, Reed JC. IAP family proteins-suppressors of apoptosis. Gene Dev 1999;13:239–52.

5. Karin M, Cao Y, Greten F, et al. NF-κB in cancer:from innocent bystander to major culprit. NatureRev Cancer 2002;2:301–10.

6. Ghavami S, Hashemi M, Ande SR, et al. Apopto-sis and cancer: mutations within caspase genes. JMed Genet 2009;46:497–510.

7. Baylin SB, Esteller M, Rountree MR, et al. Aber-rant patterns of DNA methylation, chromatin for-mation and gene expression in cancer. Hum MolGenet 2001;10:687–92.

8. Kass SU, Landsberger N, Wolffe AP. DNA meth-ylation directs a time-dependent repression oftranscription initiation. Curr Biol 1997;7:157–65.

9. Clark SJ, Melki J. DNA methylation and genesilencing in cancer: which is the guilty party?Oncogene 2002;21:5380–7.

10. Frigola J, Solé X, Paz MF, et al. Differential DNAhypermethylation and hypomethylation signaturesin colorectal cancer. Hum Mol Genet 2005;14:319–26.

11. Lopez-Serra P, Esteller M. DNA methylation-associated silencing of tumor suppressor micro-RNAs in cancer. Oncogene 2012;31:1609–22.

12. Suzuki H, Maruyama R, Yamamoto E, et al. DNAmethylation and microRNA dysregulation in can-cer. Mol Oncol 2012;6:567–78.

13. Gao W, Yu Y, Cao H, et al. Deregulated expressionof miR-21, miR-143 and miR-181a in non smallcell lung cancer is related to clinicopathologiccharacteristics or patient prognosis. Biomed Phar-macother 2010;64:399–8.

14. Bockmeyer CL, Christgen M, Müller M, et al.MicroRNA profiles of healthy basal and luminalmammary epithelial cells are distinct and reflectedin different breast cancer subtypes. Breast CancerRes Treat 2011;130:735–45.

15. Li Y, Tollefsbol TO. Impact on DNA methylationin cancer prevention and therapy by bioactive die-tary components. Curr Med Chem 2010;17:2141–51.

16. Moga MA, Dimienescu OG, Arvatescu CA, et al.The role of natural polyphenols in the preventionand treatment of cervical cancer-an overview.Molecules 2016;17:21.

17. Reddy BS. Introduction: novel approaches to theprevention of colon cancer by nutritional manipu-lation and chemoprevention. Cancer EpidemiolBiomarkers Prev 2000;9:3.

18. Tripoli E, Giammanco M, Tabacchi G, et al. Thephenolic compounds of olive oil: structure, biolog-ical activity and beneficial effects on humanhealth. Nutr Res Rev 2005;18:98–112.

Figure 8. Hypothetical model represents genetic and epigeneticapproach of olive oil to prevent colon carcinogenesis. [Color figurecan be viewed at wileyonlinelibrary.com]

1192 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

19. Waterman E, Lockwood B. Active componentsand clinical applications of olive oil. Altern MedRev 2007;12:331–42.

20. Rahman MA, Amin AR, Shin DM. Chemopreven-tive potential of natural compounds in head andneck cancer. Nutr Cancer 2010;62:973–87.

21. Casaburi I, Puoci F, Chimento A, et al. Potentialof olive oil phenols as chemopreventive and thera-peutic agents against cancer: a review of in vitrostudies. Mol Nutr Food Res 2013;57:71–3.

22. França Fda S, Bodack C, Costa A, et al. The effectof dietary oils on the development of aberrantcrypt foci, bifido bacteria and fecal pH. In Vivo2014;28:197–203.

23. Di Francesco A, Falconi A, Di Germanio C, et al.Extra virgin olive oil up-regulates CB1 tumor sup-pressor gene in human colon cancer cells and inrat colon via epigenetic mechanisms. J Nutr Bio-chem 2015;26:250–8.

24. Rodríguez-Miguel C, Moral R, Escrich R, et al.The role of dietary extra virgin olive oil and cornoil on the alteration of epigenetic patterns in therat DMBA-induced breast cancer model. PLoSOne 2015;10:e0138980.

25. Soler AP, Miller RD, Laughlin KV, et al. Increasedtight junctional permeability is associated with thedevelopment of colon cancer. Carcinogenesis 1999;20:1425–31.

26. Chadha VD, Vaiphei K, Dhawan DK. Zinc medi-ated normalization of histoarchitecture and anti-oxidant status offers protection against initiationof experimental carcinogenesis. Mol Cell Biochem2007;304:101–8.

27. Banerjee S, Bueso-Ramos C, Aggarwal BB. Sup-pression of 7,12-dimethylbenz(a)anthracene-induced mammary carcinogenesis in rats by res-veratrol: role of nuclear factor-kappaB, cyclooxy-genase 2, and matrix metalloprotease 9. CancerRes 2002;62:4945–54.

28. Imaoka T, Nishimura M, Teramoto A, et al. Coop-erative induction of rat mammary cancer by radia-tion and 1-methyl-1-nitrosourea via the oncogenicpathways involving c-Myc activation and H-rasmutation. Int J Cancer 2005;115:187–93.

29. Pettan-Brewer C, Morton J, Cullen S, et al. Tumorgrowth is suppressed in mice expressing a trun-cated XRCC1 protein. Am J Cancer Res 2012;2:168–77.

30. Livak KJ, Schmittgen TD. Analysis of relative geneexpression data using real-time quantitative PCRand the 2 (−DeltaDelta C (T)) method. Methods2001;25:402–8.

31. Migheli F, Stoccoro A, Coppede F, et al. Compari-son study of MS-HRM and pyrosequencing tech-niques for quantification of APC and CDKN2Agene methylation. PLoS One 2013;8:e52501.

32. Wojdacz TK, Dobrovic A. Methylation-sensitivehigh resolution melting (MS-HRM): a newapproach for sensitive and high-throughputassessment of methylation. Nucleic Acids Res2007;35:e41.

33. Nanda N, Dhawan DK, Bhatia A, et al. Doxycy-cline promotes carcinogenesis & metastasis viachronic inflammatory pathway: an in vivoapproach. PLoS One 2016;11:e0151539.

34. Li W, Li CB. Effect of oral Lactococcuslactis con-taining endostatin on 1,2- dimethylhydrazine-induced colon tumor in rats. World J Gastroen-terol 2005;11:7242–NaN.

35. Yu LN, Zhang QL, Li X, et al. Tiam1 transgenicmice display increased tumor invasive and

metastatic potential of colorectal cancer after1,2-dimethylhydrazine treatment. PLoS One 2013;8:73077.

36. Sánchez-Fidalgo S, Villegas I, Cárdeno A, et al.Extra-virgin olive oil enriched diet modulatesDSS-colitis-associated colon carcinogenesis inmice. Clin Nutr 2010;29:663–73.

37. Horst D, Budczies J, Brabletz T, et al. Invasionassociated up-regulation of nuclear factor kappa Btarget genes in colorectal cancer. Cancer 2009;115:4946–58.

38. Vaiopoulos AG, Athanasoula KC,Papavassiliou AG. NF-κB in colorectal cancer. JMol Med (Berl) 2013;91:1029–37.

39. Gambhir S, Vyas D, Hollis M, et al. Nuclear factorkappa B role in inflammation associated gastroin-testinal malignancies. World J Gastroenterol 2015;21:3174–83.

40. Kunnumakkara AB, Diagaradjane P, Guha S,et al. Curcumin sensitizes human colorectal can-cer xenografts in nude mice to gamma-radiationby targeting nuclear factor-kappa B-regulatedgene products. Clin Cancer Res 2008;14:2128–36.

41. Kunnumakkara AB, Guha S, Krishnan S, et al.Curcumin potentiates antitumor activity of gemci-tabine in an orthotopic model of pancreatic can-cer through suppression of proliferation,angiogenesis, and inhibition of nuclear factor-kap-paB-regulated gene products. Cancer Res 2007;67:3853–61.

42. Boss A, Bishop KS, Marlow G, et al. Evidence tosupport the anti-cancer effect of olive leaf extractand future directions. Nutrients 2016;8:513.

43. Jones PA, Laird PW. Cancer epigenetics comes ofage. Nat Genet 1999;21:163–7.

44. Baylin SB, Herman JG. DNA hypermethylation intumorigenesis: epigenetics joins genetics. TrendsGenet 2000;16:168–74.

45. Taberlay PC, Jones PA. DNA methylation andcancer. Prog Drug Res 2011;67:1–23.

46. Robertson KD. DNA methylation, methyl-transferases, and cancer. Oncogene 2001;20:3139–55.

47. Nagaraju GP, Zhu S, Wen J, et al. Novel syntheticcurcumin analogues EF31 and UBS109 are potentDNA hypomethylating agents in pancreatic can-cer. Cancer Lett 2013;341:195–203.

48. Liu S, Liu Z, Xie Z, et al. Bortezomib inducesDNA hypomethylation and silenced gene tran-scription by interfering with Sp1/NFkappa B-dependent DNA methyltransferase activity inacute myeloid leukemia. Blood 2008;111:2364–73.

49. Kim JY, Hwang JH, Zhou W, et al. The expressionof VEGF receptor genes is concurrently influencedby epigenetic gene silencing of the genes andVEGF activation. Epigenetics 2009;4:313–21.

50. Farias LC, Gomes CC, Rodrigues MC, et al. Epige-netic regulation of matrix metalloproteinaseexpression in ameloblastoma. BMC Clin Pathol2012;12:11.

51. Shukeir N, Pakneshan P, Chen G, et al. Alterationof the methylation status of tumor-promotinggenes decreases prostate cancer cell invasivenessand tumorigenesis in vitro and in vivo. CancerRes 2006;66:9202–10.

52. Odabasoglu F, Yildirim OS, Aygun H, et al. Dif-fractaic acid, a novel proapoptotic agent, induceswith olive oil both apoptosis and antioxidativesystems in Ti-implanted rabbits. Eur J Pharmacol2012;674:171–8.

53. Leong KH, Looi CY, Loong XM, et al. Cycloart-24-ene-26-ol-3-one, a new Cycloartane isolatedfrom leaves of Aglaia exima triggers tumournecrosis factor-receptor 1-mediated Caspase-dependent apoptosis in colon cancer cell line.PLoS One 2016;11:e0152652.

54. Fulda S, Kufer MU, Meyer E, et al. Sensitizationfor death receptor- or drug-induced apoptosisby re-expression of caspase-8 through demethyl-ation or gene transfer. Oncogene 2001;20:5865–77.

55. Hervouet E, Cheray M, Vallette FM, et al. DNAmethylation and apoptosis resistance in cancercells. Cell 2013;2:545–73.

56. Trichopoulou A, Lagiou P, Kuper H, et al. Cancerand Mediterranean dietary traditions. Cancer Epi-demiol Biomarkers Prev 2000;9:869–73.

57. Rosato V, Guercio V, Bosetti C, et al. Mediterra-nean diet and colorectal cancer risk: a pooledanalysis of three Italian case-control studies. Br JCancer 2016;115:862–5.

58. Jacobs S, Harmon BE, Ollberding NJ, et al.Among 4 diet quality indexes, only the alternateMediterranean diet score is associated with bettercolorectal cancer survival and only in AfricanAmerican women in the multiethnic cohort. JNutr 2016;146:1746–55.

59. Jones P, Cade JE, Evans CEL, et al. The Mediter-ranean diet and risk of colorectal cancer in theUKWomen’s cohort study. Int J Epidemiol 2017;46:1786–96.

60. Toledo E, Salas-Salvadó J, Donat-Vargas C, et al.Mediterranean diet and invasive breast cancer riskamong women at high cardiovascular risk in thePREDIMED trial: a randomized clinical trial.JAMA Intern Med 2015;175:1752–60.

61. Kasinski AL, Slack FJ. Epigenetics and genetics.MicroRNAs en route to the clinic: progress in val-idating and targeting microRNAs for cancer ther-apy. Nat Rev Cancer 2011;11:849–64.

62. Chen X, Shi K, Wang Y, et al. Clinical value ofintegrated-signature miRNAs in colorectal can-cer: miRNA expression profiling analysis andexperimental validation. Oncotarget 2015;6:37544–56.

63. Meng WJ, Yang L, Ma Q, et al. MicroRNAexpression profile reveals miR-17-92 and miR-143-145 cluster in synchronous colorectal cancer.Medicine (Baltimore) 2015;94:e1297.

64. Borralho PM, Kren BT, Castro RE, et al.MicroRNA-143 reduces viability and increasessensitivity to 5-fluorouracil in HCT116 humancolorectal cancer cells. FEBS J 2009;276:6689–700.

65. Zhang J, Wang L, Li B, et al. miR-145 downregu-lates the expression of cyclin-dependent kinase6 in human cervical carcinoma cells. Exp TherMed 2014;8:591–4.

66. Zhang X, Liu S, Hu T, et al. Up-regulatedmicroRNA-143 transcribed by nuclear factorkappa B enhances hepatocarcinoma metastasis byrepressing fibronectin expression. Hepatology2009;50:490–9.

67. Feng Z, Zhang C, Wu R, et al. Tumor suppressorp53 meets microRNAs. J Mol Cell Biol 2011;3:44–50.

68. Gomes SE, Simões AE, Pereira DM, et al. miR-143 or miR-145 over-expression increasescetuximab-mediated antibody-dependent cellularcytotoxicity in human colon cancer cells. Oncotar-get 2016;7:9368–87.

Nanda et al. 1193

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention

69. Su J, Liang H, Yao W, et al. MiR-143 and MiR-145 regulate IGF1R to suppress cell proliferationin colorectal cancer. PLoS One 2014;9:e114420.

70. Grady WM, Parkin RK, Mitchell PS, et al. Epige-netic silencing of the intronic microRNA hsa-miR-342 and its host gene EVL in colorectal can-cer. Oncogene 2008;27:3880–8.

71. Vrba L, Munoz-Rodriguez JL, Stampfer MR, et al.miRNA gene promoters are frequent targets ofaberrant DNA methylation in human breast can-cer. PLoS One 2013;8:54398.

72. Dou L, Zheng D, Li J, et al. Methylation-mediatedrepression of microRNA-143 enhances MLL-AF4oncogene expression. Oncogene 2012;31:507–17.

73. Donzelli S, Mori F, Bellissimo T, et al. Epigeneticsilencing of miR-145-5p contributes to brainmetastasis. Oncotarget 2015;6:35183–201.

74. Wang LQ, Liang R, Chim CS. Methylation oftumor suppressor microRNAs: lessons from lym-phoid malignancies. Expert Rev Mol Diagn 2012;12:755–65.

1194 Chemopreventive role of olive oil in colon carcinogenesis

Int. J. Cancer: 144, 1180–1194 (2019) © 2018 UICC

Can

cerTherapy

andPrevention