3,6-Dihydroxyflavone Suppresses Breast Carcinogenesis by Epigenetically Regulating miR-34a and miR-21

Breast cancer is one of the most serious health threats to women worldwide. In the United States, 230,480 new cases of invasive breast cancer were diagnosed in 2011. The lifetime risk of being diagnosed with breast cancer is 12.5% for women in the United States, and approximately 39,520 women are expected to die from this disease annually (1). Despite advances in screening, diagnosis, and therapy, breast cancer continues to pose an enormous global healthcare problem.

miRNAs are a class of 20- to 25-nucleotide-long noncoding RNAs that negatively regulate gene expression by binding to the 3′ untranslated region (3′UTR) of target messenger RNAs, causing translational repression or degradation. miRNAs have been well established to be aberrantly expressed in carcinogenesis and progression, where they function as tumor suppressors or act as oncogenes. Our study has revealed the global upregulation of miR-21 and downregulation of miR-34a in breast carcinogenesis (2). As a proapoptotic transcriptional target of p53, miR-34a has been shown to behave as a tumor suppressor by repressing genes involved in various oncogenic signaling pathways. On the contrary, miR-21 is a notable oncogene, which is consistently overexpressed in a wide range of cancers and plays a key role in resisting programmed cell death in cancer cells (3). We have previously selected a promising anticancer agent, the flavonol 3,6-dihydroxyflavone (3,6-DHF), in pharmacodynamic experiments, and demonstrated that it is a potent natural chemopreventive agent against breast carcinogenesis. IN addition, the previous study also revealed that 3,6-DHF treatment significantly upregulates miR-34a and downregulates miR-21 in breast cancer cells, yet the upstream and downstream events of the anticancer mechanism remain unclear.

Increasing evidence supports that the epigenetic variability at specific transcription regulation sites appears to be susceptible to modulation by diet and nutrition; some dietary components may suppress tumorigenesis and development by affecting the process of DNA methylation and histone modifications, which modulates the expression of certain key genes and the activation of crucial signaling pathways (4). In the last two decades, the study of epigenetic modification emerged as one of the major areas of cancer treatment. The key processes responsible for epigenetic regulation are DNA methylation and modifications in chromatin and posttranscriptional gene regulation by noncoding RNA. Thus, we hypothesized that 3,6-DHF may regulate the expressions of miR-34a and miR-21 by modulating DNA methylation or histone modification.

The PI3K/Akt/mTOR signaling pathway is critical to normal cellular physical progress and also commonly overactivated in cancer progression, especially in breast carcinogenesis and metastasis (5, 6). In fact, this signaling pathway acts as a key integration point between the extrinsic and intrinsic cellular environments (7, 8); it regulates a broad spectrum of cellular processes and consequently plays an important role in breast tumorigenesis and development (9, 10). Currently, targeting the PI3K/AKT/mTOR signaling pathway has become an effective strategy for chemoprevention and cancer treatment (11, 12). Studies indicated that miR-34a and miR-21 play an important role in regulating the PI3K/Akt/mTOR signaling pathway. The tumor suppressor gene PTEN acts as the principal negative regulator of the PI3K pathway. Notably, PTEN is a direct target of miR-21, which negatively regulates its expression (13). Conversely, Notch-1, a direct target of miR-34a, is also known to interact with the PI3K/Akt signaling pathway (14). The aberrant activation of Notch signaling is an early event in breast cancer (15, 16).

In the present study, we investigated the underlying epigenetic mechanisms by which 3,6-DHF regulates the expression of miR-34a and miR-21 in breast carcinogenesis and in breast cancer cells. Furthermore, we determined the effects of 3,6-DHF on the PI3K/Akt/mTOR signaling pathway in vivo and in vitro, and explored the roles of miR-34a and miR-21.

3,6-DHF was purchased from Alfa Aesar; DMEM/F12 medium and FBS were purchased from HyClone; Trizol reagent, horse serum, gentamicin, insulin, Lipofectamine 2000, and Opti-Mem were purchased from Invitrogen; epidermal growth factor (EGF) was purchased from PeproTech Inc.; all antibodies were purchased from Cell Signaling Technology. 1-methyl-1-nitrosourea (MNU), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), benzo[a]pyrene (B[a]P), cholera enterotoxin, hydrocortisol, protein A/G-agarose, dimethylsulfoxide (DMSO), PBS, and other chemicals were purchased from Sigma-Aldrich. The miRNA first-strand cDNA synthesis Kit and miRNA Real-Time PCR Assay Kit were purchased from aidlab. Anti–miR-34a oligonucleotides were purchased from Exiqon. pcDNA3-miR21 was purchased from Addgene.

Female Sprague-Dawley (SD) rats (aged 42–48 days, 145–165 g) and BALB/c nude mice (aged 42–48 days, 15–20 g) were obtained from the Medical Experimental Animal Center of the Third Military Medical University [SCXK-(army)-2007-015]. These animals were bred and maintained in accordance with our institutional guidelines for the use of laboratory animals. The animal rooms were maintained at 25°C with 50% relative humidity and a 12-hour light/12-hour dark cycle. All of the animal procedures were approved by the Animal Ethics Committee of the Third Military Medical University. Experimental models of carcinogenesis in rats and xenografted MDA-MB-231 cells in athymic mice were processed as reported previously (2)

Human breast epithelial MCF-10A cells were maintained in complete medium (DMEM/F12 medium supplemented with mitogenic additives, including 100 ng/mL cholera enterotoxin, 10 μg/mL insulin, 0.5 μg/mL hydrocortisol, 20 ng/mL EGF, and 5% horse serum). Breast cancer cell MDA-MB-231 and MCF-7 were grown in DMEM/F12 medium, MDA-MB-453 were grown in RPMI 1640 medium, supplemented with 10% FBS. All the cultures were maintained in a humidified atmosphere of 5% CO₂/95% air at 37°C. All cell lines were purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China) in October, 2013. All cell lines have been tested and authenticated by DNA (short tandem repeat genotyping) profiling in May, 2014.

The cellular breast carcinogenesis model was processed as reported previously (17). Briefly, MCF10A cells were treated with NNK and B[a]P (each at 100 pmol/L) and with different concentrations of 3,6-DHF (0, 5, 10, and 20 μmol/L); the cells were subcultured every 3 days. The cancer-associated properties of the treated cells were evaluated using the reduced dependence on growth factors (RDGF) assay, anchorage-independent cell growth (AIG) assay, and scratch/wound-healing assay.

The activities of mTOR and p-Akt (S473) were measured with a colorimetric K-LISA mTOR activity assay kit (Calbiochem) and a PathScan Phospho-Akt (S473) ELISA kit (Cell Signaling Technology), respectively. Briefly, 10⁷ treated cells were prepared and measured. The assays were performed as per the manufacturer's manual.

The cell lysates were prepared using RIPA buffer; equal amounts of cellular proteins were resolved by electrophoresis in 10% or 12% SDS-polyacrylamide gels for Western immunoblotting with specific antibodies. The antigen–antibody complexes on the filters were detected by chemiluminescence.

Briefly, slides were rinsed with PBS and cells were fixed with 2% paraformaldehyde and permeabilized in methanol. After washing with PBS, slides were blocked with 2% donkey serum. Primary and secondary antibodies were incubated in 5% donkey serum. Then the slides were immediate analyzed by a laser confocal scanning microscopy.

Xenografted breast tumors of MDA-MB-231 cells in athymic mice, breast tissues, and the tumors of MNU-treated rats were all obtained from the previous study (2). The tissue sections (4-μm-thick) were placed onto treated slides, heat-fixed, deparaffinized, rehydrated, boiled in citrate buffer for antigen retrieval, and treated with 3% hydrogen peroxide to block the endogenous peroxidase activities. After washing with PBS, the slides were blocked with 2% donkey serum for 0.5 hours and then incubated with an antibody against p-Akt(T308) (dilution 1:200) at 4°C overnight. The primary antibody was omitted for the negative controls. The secondary biotinylated antibody was then applied, and the signal was developed using a modified avidin–biotin complex immunoperoxidase staining procedure. Counterstaining was performed with Trypan blue or Harris hematoxylin. Immunostaining density was quantified using Image J analysis.

Anti–miR-34a oligonucleotides (5′-ACAACCAGCTAAGACACTGCC-3′) were obtained from Exiqon. The pcDNA6.2-GW/miR-21 (pcDNA3-miR21) (Plasmid 21114) plasmid was provided by Addgene. The cells were transfected using Lipofectamine 2000 in Opti-Mem according to the manufacturer's protocol. The medium was replaced 8 hours later, and the cells were collected for the subsequent experiments 48 hours after transfection. The final concentrations of oligonucleotides were 100 nmol/L.

The total RNA was extracted using Biozol reagent. The miRNA first-strand cDNA synthesis Kit and miRNA Real-Time PCR Assay Kit (aidlab) were used to quantify the miRNA transcripts in our study following the manufacturer's instructions. Each reaction sample was run in triplicate. The expression of U6 small nucleolar RNA was used as an internal control. The relative expression level for each miRNA was calculated with the comparative CT method (2^−ΔΔCt).

The genomic DNA was subjected to sodium bisulfite modification using the EZ DNA Methylation-Gold Kit (ZYMO Research) according to the manufacturer's instructions. The modified DNA was eluted at a final volume of 10 μL, and 1.5 μL were used for methylation-specific PCR (MSP). The PCR primers used to detect the CpG-methylation of the miR-34a promoter were previously established. MSP primer sequences for CarT and MDA-MB-231 cells: M: Forward 5′-GGTTTTGGGTAGGCGCGTTTC-3′, Reverse 5′-TCCTCATCCCCTTCACCGCCG-3′; U: Forward 5′-IIGGTTTTGGGTAGGTGTGTTTT-3′, Reverse 5′-AATCCTCATCCCCTTCACCACCA-3′. MSP primer sequences for MCF-7 and MDA-MB-453 cells: M: Forward 5′-ATGAGGATTAGGATTTCGGAG, Reverse 5′-AACGCATAAAAACGACGACAA; U: Forward 5′-GGGGATGAGGATTAGGATTTT, Reverse 5′-CAAACAAAACACATAAAAACAACA. The reactions were carried out using the iQ 5 Multicolor Real-Time PCR Detection System (BioRad) at 95°C for 10 minutes; followed by 35 cycles of 30 seconds at 95°C, 30 seconds at 60°C, 30 seconds at 72°C (30 cycles); and a final 7-minute extension at 72°C. Each amplification product (5 μL) was resolved in a 2.5% agarose gel by electrophoresis (Invitrogen). The electrophoresis conditions used were 80 V for 60 minutes. The gel was directly visualized under UV illumination after electrophoresis.

For the DNA methyltransferase (DNMT) activity assay, nuclear extracts of 10⁶ treated cells were prepared using a nuclear extraction reagent (Pierce) following the manufacturer's instructions. The DNMT activity was determined in the nuclear extracts using the EpiQuik DNA Methyltransferase Activity Assay Kit (Epigentek Inc.) following the manufacturer's protocol. For DNMT1 inhibitor testing, 1 μL of DNMT1 enzyme (Epigentek Inc.) was incubated with different concentrations of 3,6-DHF at 37°C for 60 minutes, and the DNMT activity was determined using the EpiQuik DNA Methyltransferase 1 Activity/Inhibitor Screening Assay Core Kit (Epigentek Inc.).

The blind docking was performed with Autodock4 by setting grid sizes. The receptor site was prepared with Sybyl (Tripos) using the NMR structure 2NPU model 1 from the Protein Data Bank (www.pdb.org). The grid size for the docking site was expanded to include the entire DNMT1 molecule, and 3,6-DHF was docked.

Chromatin immunoprecipitation (ChIP) was carried out according to the instructions of the EZ-ChIP Chromatin immunoprecipitation Kit (Millipore). After ChIP, the DNA precipitated by the target antibody was detected with qRT-PCR. PCR was conducted in a final volume of 25 μL containing 12.5 μL of 2× SYBR Mix, Taq DNA Polymerase (BioEasy), 1 μL each of forward primer and reverse primers (10 μmol/L), and 6 μL of DNA template under the following conditions: the template was first denatured at 94°C for 10 minutes, then subjected to 50 cycles of amplification (94°C for 20 seconds, 60°C for 1 minute), 95°C for 2 minutes, 72°C for 1 minute, 95°C for 30 seconds, and 55°C for 10 seconds (repeat 80 times), 30°C for 1 minute. After real-time PCR, relative data quantification was performed using the 2^−ΔΔCt method, and the result was calculated in the form of % Input, which was given by the following formula: %Input = 2^{(Ctinput−CtChIP)} × input dilution factor × 100. Real-time PCR primers were synthesized by SBS Genethech Co., Ltd. The purified DNAs were amplified with the following primer pairs: 5′-TGC TGTTTGGTCTCAGTA-3′and 5′-GGCAAGTTAACGAAAAGAA-3′ for miR-21.

The results are presented as the mean ± SD from at least three independent experiments. The tumor incidences were compared using the χ² test. The other data were analyzed by one-way ANOVA followed by the Tukey test for multiple comparisons. Differences were considered significant at P < 0.05.

Using cancer-associated properties as the target endpoints, we assessed the effects of 3,6-DHF on the chronic breast carcinogenesis induced in vitro by the carcinogens NNK and B[a]P. As shown in Fig. 1A and B, human breast epithelial MCF10A cells treated with carcinogens (CarT) for 30 days showed aberrantly increased cell survival adapted to RDGF and AIG, indicating cellular carcinogenic transformation. Compared with CarT, cells cotreated with carcinogens and 3,6-DHF (C-DHF) for 30 days exhibited a significantly lower acquisition of RDGF and AIG. Similarly, the wound-healing assay showed that CarT cells required apparently increased mobility and proliferation to heal the wound, which can be suppressed by 3,6-DHF cotreatment (Fig. 1C). These data suggest the chemopreventive effects of 3,6-DHF on carcinogen-induced breast carcinogenic transformation.

The qRT-PCR data revealed the significant downregulation of miR-34a and upregulation of miR-21 in cellular breast carcinogenesis, which could be mitigated by 3,6-DHF cotreatment (Fig. 2A and B). The data also showed that 3,6-DHF treatment effectively modulates the expressions of miR-34a and miR-21 in breast cancer cells MDA-MB-231, MCF-7, and MDA-MB-453 (Fig. 2C and D).

The inactivation of miR-34a in multiple types of malignancies, including breast cancer, has been attributed to the hypermethylation of the promoter. We detected the effect of 3,6-DHF on the methylation status of the miR-34a promoter in breast cancer cells and in carcinogenesis. The MSP data showed that the hypermethylation of miR-34a promoter was significantly inhibited in 30-day C-DHF and 3,6-DHF–treated breast cancer cells compared with CarT and the control, respectively (Fig. 3A). Because DNMTs catalyze the process of DNA methylation, we assessed whether 3,6-DHF changed the DNMT activity in breast cancer cells. Our study found that 3,6-DHF acted in a dose-dependent manner to decrease the DNMT activity in breast cancer cells (Fig. 3B). We then performed the DNMT1 inhibitor testing, and the data (Fig. 3C) showed that 3,6-DHF is an effective DNMT1 inhibitor that can significantly inhibit the DNMT1 activity in a dose-dependent manner. Furthermore, we blind-docked 3,6-DHF to the DNMT1 target using Autodock4 by setting grid sizes that included the entire DNMT1 molecule. As shown in Fig. 3D, 3,6-DHF docked to the putative cytosine pocket with the lowest binding energies of −7.1 kcal/mol. This pocket is the active region that facilitates binding with hemimethylated DNA, a process that is required for subsequent methylation.

In addition, we performed ChIP-qPCR analysis for histone modifications associated with the active transcription state (H3K9-14ac and H3K27ac) and the inactive transcription state (H3K27me3) on the miR-21 promoter. The results showed that 3,6-DHF treatment significantly lowers the H3K9-14ac on the miR-21 promoter, while it does not apparently affect other histone modifications (Fig. 3E). These findings indicated that 3,6-DHF upregulates miR-34a expression via demethylation, whereas downregulates miR-21 expression by modulating histone modification.

Given the important role of the PI3K/Akt/mTOR signaling pathway in tumorigenesis, we evaluated the effect of 3,6-DHF on the signaling in breast carcinogenesis. Western blot detections (Fig. 4A) showed that the levels of PI3K, p-Akt, and p-mTOR in CarT cells significantly increased in a time-dependent manner, indicating the dramatic promotion of the PI3K/Akt/mTOR signaling pathway in carcinogenesis. Cotreatment with 3,6-DHF could effectively suppress the cancer-promoting signaling pathway. We then measured the activity of Akt and mTOR kinase with an ELISA-based kinase activity assay. In accordance with the Western blot detections, the data (Fig. 4B and C) confirmed that the activity of Akt and mTOR in CarT cells significantly increased in a time-dependent manner, which could be counteracted by 3,6-DHF cotreatment. Furthermore, we detected the p-Akt level of breast tissues and tumors in MNU-treated rats via immunohistochemistry. The results (Fig. 5A) showed that the p-Akt level significantly increased in breast carcinogenesis in vivo, which could be effectively inhibited by 3,6-DHF administration [20 mg/kg, intragastric (i.g.) administration].

We also evaluated the effect of 3,6-DHF on the PI3K/Akt/mTOR signaling pathway in breast cancer cells in vitro and in vivo. As shown in Fig. 6A, 3,6-DHF treatment could significantly downregulate the levels of PI3K (p85, p110), p-Akt (Thr308, Ser473), and p-mTOR (S2448, S2481) and consequently reduce the activity of Akt (Fig. 6B) and mTOR (Fig. 6C) in a dose-dependent manner. In accordance with these findings, the immunohistochemistry for p-Akt(T308) in xenografted breast tumors of MDA-MB-231 cells showed that 3,6-DHF administration (20 mg/kg, i.g.) could significantly suppress the activation of Akt in vivo (Fig. 5B). Furthermore, to support our findings, we further investigated the effects of 3,6-DHF on the downstream targets of the PI3K/Akt/mTOR signaling pathway. As expected, the detections showed that 3,6-DHF treatment could significantly suppress the activation of the downstream targets, including ribosomal p70-S6 kinase (S6K70), ribosomal protein S6 (rpS6), eukaryotic translation initiation factor (eIF)4B, and eukaryotic translation initiation factor 4E-binding protein (4EBP) 1 (Figs. 5C and 6D).

We detected the effects of 3,6-DHF on the levels of PTEN, Notch-1, and Hes-1 because they are targets of miR-21 and miR-34a. Notably, PTEN is the key inhibitor of the PI3K/Akt signaling pathway; on the contrary, Notch-1 is an activator of this signaling. The results indicated that 3,6-DHF treatment significantly upregulates PTEN and downregulates Notch-1 and Hes-1 in carcinogenesis (Fig. 7A) and in breast cancer cells (Fig. 7B).

To explore the roles of miR-34a and miR-21, we blocked the effects of 3,6-DHF on the two miRNAs by transfecting the cells with a specific inhibitor or plasmid. As shown in Fig. 2C and D, a locked nucleic acid oligonucleotide complementary to the miR-34a sequence (TC_anti-34a) blocked miR-34a function, while the pcDNA6.2-GW/miR-21 plasmids led to the substantial production of miR-21 in MDA-MB-231 cells. The results indicated that the inhibition of miR-34a or overexpression of miR-21 significantly reduced the effects of 3,6-DHF on Notch-1 and PTEN, and consequently weakened the suppression of 3,6-DHF on the activation of Akt and mTOR (Fig. 7C and D). Taken together, these findings indicated that 3,6-DHF suppresses the PI3K/Akt/mTOR signaling pathway partially by modulating miR-34a and miR-21.

Epidemiologic studies and systematic analyses indicated that diets rich in fruits and vegetables are associated with a reduced risk of breast cancer (18, 19). Dietary flavonoids, a group of polyphenolic compounds, have been identified as potential cancer-preventive components of fruits and vegetables (20, 21). Flavonoids can be categorized into six major subclasses based on their range and structural complexity: flavonols, flavones, flavan-3-ols, flavanones, anthocyanins, and isoflavones. Flavonols mainly exist in onions, broccoli, tea, and various common fruits; flavones are common in aromatic herbs, celery, and chamomile tea; flavan-3-ols are present in cocoa, red wine, grapes, apples, green tea, and other fruits; flavanones can be found in oranges and other citrus fruits; anthocyanidines are abundant in colored berries, black currants; and isoflavones are found in soy food (22, 23). In fact, our previous study revealed that flavonols and flavones, but not the total flavonoids, flavan-3-ols, flavanones, or anthocyanins, reduce breast cancer risk (24). Further studies should be developed to select effective anticancer compounds from flavonoids and uncover the mechanisms of their relevance to cancer prevention. The findings will provide useful insight and evidence for dietitians and other healthcare professionals when discussing diet and cancer prevention.

Our preliminary screening of anticancer compounds from flavonoids identified 3,6-DHF as a promising anticancer agent and an effective natural chemopreventive agent against breast carcinogenesis. Based on these findings, the present study further revealed that 3,6-DHF epigenetically regulates miR-34a and miR-21, and consequently represses the PI3K/Akt/mTOR signaling pathway. Our study suggested a new strategy for suppressing the critical cancer-promoting signaling pathway in the treatment of cancer via modulating miRNAs. Moreover, our findings will likely provide useful insight and evidence that can be used by registered dietitians and other healthcare professionals when discussing diet and cancer prevention with patients. However, the anticancer effects of 3,6-DHF were only demonstrated in animal and cell culture studies. Because human clinical trials examining the chemopreventive potential of 3,6-DHF have not been conducted, more studies are warranted to confirm the results.

DNA methylation is responsible for regulating gene expression and interacting with the nucleosomes that control DNA packaging. This process can affect entire domains of DNA. The gene silencing of tumor suppressors is frequently associated with the hypermethylation of the CpG islands in promoter regions, but the initial trigger of hypermethylation is unknown. CpG islands typically span the promoter region and first exon of approximately 60% of all genes. The majority of these CpG islands are unmethylated in the normal cell, which correlates with active gene transcription, in comparison with the bulk of the genome, which is methylated. However, these methylation states are often reversed in cancer cells: the global levels of DNA methylation become hypomethylated in conjunction with the DNA hypermethylation of the CpG islands associated with gene promoters, especially tumor suppressors. MiR-34a was demonstrated to be a tumor suppressor, a direct transcriptional target of p53, and a component of the p53 transcriptional network (25, 26). Notably, the epigenetic silencing of miR-34 by aberrant CpG methylation frequently occurs in various types of human cancer, including breast cancer (27, 28). Our data indicated that 3,6-DHF demethylates miR-34 and increases its expression in breast cancer cells, a mechanism that may be associated with the direct inhibition of DNMT1 activity.

Next, the histone modifications in the chromatin structure also play an important role in gene regulation and carcinogenesis. Histone modifications are specifically characterized by the genomic regulatory regions, such as inactive promoters that are enriched in trimethylated H3 at lysine 27 (H3K27me3) or trimethylated H3 at lysine 9 (H3K9me3) and active promoter regions that are enriched in acetylated H3 at lysine 27 (H3K27ac; ref. 29). Our study showed that 3,6-DHF significantly affects the histone modification in breast cancer cells, lowering the H3K9-14ac on the miR-21 promoter, which may contribute to the downregulation of miR-21. A deep investigation of the underlying mechanism is interesting.

No potential conflicts of interest were disclosed.

Conception and design: H. Chang, M. Mi

Development of methodology: X. Peng, Y. Gu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Peng, H. Chang, Y. Gu, J. Chen, Q. Xie

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Chang

Writing, review, and/or revision of the manuscript: X. Peng, H. Chang, M. Mi

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Yi, J. Zhu, Q. Zhang

Study supervision: M. Mi

The authors thank Elsevier WebShop for the English language editing of the article.

This work was supported by the grant from the National Natural Science Foundation of China (81372974) and Chongqing Fundamental and Advanced Research Project (cstc2013jcyjA10083; to H. Chang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.