miR-200a Regulates Nrf2 Activation by Targeting Keap1 mRNA in Breast Cancer Cells*

NF-E2-related factor 2 (Nrf2) is an important transcription factor that activates the expression of cellular detoxifying enzymes. Nrf2 expression is largely regulated through the association of Nrf2 with Kelch-like ECH-associated protein 1 (Keap1), which results in cytoplasmic Nrf2 degradation. Conversely, little is known concerning the regulation of Keap1 expression. Until now, a regulatory role for microRNAs (miRs) in controlling Keap1 gene expression had not been characterized. By using miR array-based screening, we observed miR-200a silencing in breast cancer cells and demonstrated that upon re-expression, miR-200a targets the Keap1 3′-untranslated region (3′-UTR), leading to Keap1 mRNA degradation. Loss of this regulatory mechanism may contribute to the dysregulation of Nrf2 activity in breast cancer. Previously, we have identified epigenetic repression of miR-200a in breast cancer cells. Here, we find that treatment with epigenetic therapy, the histone deacetylase inhibitor suberoylanilide hydroxamic acid, restored miR-200a expression and reduced Keap1 levels. This reduction in Keap1 levels corresponded with Nrf2 nuclear translocation and activation of Nrf2-dependent NAD(P)H-quinone oxidoreductase 1 (NQO1) gene transcription. Moreover, we found that Nrf2 activation inhibited the anchorage-independent growth of breast cancer cells. Finally, our in vitro observations were confirmed in a model of carcinogen-induced mammary hyperplasia in vivo. In conclusion, our study demonstrates that miR-200a regulates the Keap1/Nrf2 pathway in mammary epithelium, and we find that epigenetic therapy can restore miR-200a regulation of Keap1 expression, therefore reactivating the Nrf2-dependent antioxidant pathway in breast cancer.

Due to its role in protecting against genotoxic stress, it has been hypothesized that Nrf2 could be a target for chemopreventative strategies. Nrf2 is found down-regulated in breast cancer cell lines and breast cancer patient samples when compared to healthy mammary epithelial cells (9). However, Nrf2 is not down-regulated in all cancers. For example, in lung and pancreatic cancers, mutation of Keap1 disrupts Keap1/Nrf2 interaction, leading to constitutive activation of Nrf2 (10,11). In these cancers, it is thought that hijacking of detoxifying pathways, through constitutive hyperactivation of Nrf2, promotes resistance to chemotherapy. Nevertheless, in support of a protective role for Nrf2, several studies have found that Nrf2-deficient mouse models show increased carcinogen sensitivity and cancer incidence (12)(13)(14)(15). Furthermore, other studies suggest additional roles for Nrf2 in opposing tumorigenesis. Nrf2 knockdown in lung cancer cells has been shown to result in loss of E-cadherin and increased cell motility (16). Similarly, the activation of the Nrf2 pathway in renal tubular epithelial cells (undergoing immunosuppressive induction) has been shown to prevent epithelial to mesenchymal transition by maintaining E-cadherin expression (17). Finally, it was found that knockdown of Nrf2 can facilitate hepatocellular carcinoma and hepatocellular adenocarcinoma cell migration by activating transforming growth factor ␤ (TGF-␤)/Smad signaling pathways (16). Thus, in addition to protecting against genotoxicity, Nrf2 may oppose tumorigenesis by suppressing tumor cell invasion. Regardless, the exact role of Nrf2 in tumorigenesis is likely to be cancer-specific and stage-specific.
In addition to somatic mutation, other mechanisms have been found that impact Keap1 expression in cancers. It was shown that CpG methylation of the Keap1 promoter can affect Keap1 expression in gliomas (19). Additionally, aberrant mRNA splicing has been found to impact Keap1 expression in prostate cancer (20). Due to a limited understanding of Keap1 regulation in breast cancers, we sought to determine what if any role miRs play in the regulation of the Keap1/Nrf2 pathway in breast cancer cells. Of particular importance to our studies, there has been no previous characterization of Keap1 regulation by miRs.
miRs are ϳ22-nucleotide single-stranded non-coding RNAs that suppress gene expression post-transcriptionally by binding to miR response elements within 3Ј-UTRs of target mRNAs. miR regulation results in inhibition of target gene expression through mRNA degradation and/or translational inhibition (21)(22)(23). Aberrant expression of miRs is closely associated with various human diseases, including breast cancer (23)(24)(25). Among miRs dysregulated in breast cancers, the miR-200 family is often found down-regulated (26,27) and subject to aberrant epigenetic silencing (28 -30). The miR-200 family has been found to play a critical role in the maintenance of the epithelial phenotype by targeting ZEB1, SIP1, and SIRT1, thus preventing the silencing of E-cadherin (30,31). Within the miR-200 family, there are two clusters: miR-200b/200a/429 located on chromosome 1p36, and miR-200c/141 located on chromosome 12p13 (32). SAHA (vorinostat) is an HDAC inhibitor that targets Class I/II HDACs that is being evaluated in multiple clinical trials for treatment of advanced solid tumors and hematological malignancies and is currently approved for the treatment of cutaneous T-cell lymphoma. HDAC inhibitors, such as SAHA, function by increasing histone (and non-histone) acetylation, which can lead to re-expression of epigenetically silenced genes (33,34). It has recently been demonstrated that treatment of breast cancer cell lines with histone deacetylase inhibitors can result in the re-expression of epigenetically silenced miRs (28 -30).
Here we show that miR-200a regulates Keap1 expression by binding to the Keap1 mRNA 3Ј-UTR. It was previously demonstrated that epigenetic mechanisms played a key role in miR-200a loss in breast cancer. We now demonstrate that epigenetic silencing of miR-200a contributes to dysregulation of Keap1 and interferes with the Keap1/Nrf2 pathway. Treatment of breast cancer cell lines with epigenetic therapy (SAHA) resulted in miR-200a restoration and Keap1 down-regulation, which ultimately led to reactivation of the Nrf2-dependent antioxidant pathways.
Quantitative Real-time PCR-Total RNA from cell lines was extracted for analysis as previously described (30,35). Briefly, total RNA was extracted with TRIzol reagent (Invitrogen). cDNA was synthesized from 1 g of total RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen). Total RNA from mouse mammary gland tissue was isolated with the RNeasy Lipid Tissue Midi Kit (Qiagen, Valencia, CA) following manufacturer's instructions. Quantitative real-time PCR (qRT-PCR) was carried out using a Light Cycler 480II (Roche Applied Science) with primers for Keap1 (forward, 5Ј-TACGATGTG-GAAACAGAGACGTGGA-3Ј; reverse, 5Ј-TCAACAGGTA-CAGTTCTGGTCAATCT-3Ј) and NQO1 (forward, 5Ј-AGGC-TGGTTTGAGCGAGTTC-3Ј; reverse, 5Ј-ATTGAATTCGG-GCGTCTGCTG-3Ј). mRNA levels were normalized to levels of GAPDH mRNA as before (30). In addition, small RNA was converted to cDNA from 1 g of total RNA using the First-Strand Synthesis Kit (SABiosciences, Frederick, MD). Follow-up miR analysis was performed by qRT-PCR using miRspecific (miR-200a) primer sets (SABiosciences) with normalization to U6 snRNA levels as an internal control. miR expression was screened for an 88-miR panel using the Human miFinder Array (SABioscience).
Luciferase assays were performed as described previously (30,38). Cells were transfected with luciferase reporter plasmids along with Renilla luciferase phGR-TK (Promega, Madison, WI) as an internal control using Lipofectamine 2000. The luciferase activity was measured 48 h after transfection using the Dual-Luciferase reporter assay system (Promega).
Analysis of Keap1 mRNA Stability-Keap1 mRNA stability assays were performed as previously described (35). Briefly, cells were transfected with miR-200a expression vector or empty control vector. 48 h after transfection, cells were incubated with Act D (5 g/ml) and harvested at subsequent time intervals (0, 1, 2, and 4 h). Total RNA was extracted, and cDNA was synthesized as described above. qRT-PCR analysis was used to monitor Keap1 mRNA decay. Housekeeping gene GAPDH, which shows little to no decay over 4 h, was used as an internal control. Results from Act D assays were processed

miR-200a Targets Keap1 in Human Breast Cancer Cells
using Prism 4.0 software (GraphPad, La Jolla, CA) to calculate Keap1 mRNA half-life.
Western Blotting-Western blotting was performed using anti-Keap1 (E-20, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) goat polyclonal antibody and anti-Nrf2 rabbit polyclonal antibody (H-300, Santa Cruz Biotechnology, Inc.). Nuclear extraction was performed using the ProteoJET TM cytoplasmic and nuclear protein extraction kit (Fermentas, Glen Burnie, MD). Anti-proliferating cell nuclear antigen rabbit polyclonal antibody (Santa Cruz Biotechnology, Inc.) was used as a nuclear marker and a loading control. Actin antibody (Santa Cruz Biotechnology, Inc.) was used to normalize protein expression.
Chromatin Immunoprecipitation (ChIP)-The ChIP assay was carried out as described previously (30,38). Briefly, cells were cross-linked with 1% formaldehyde followed by sonication. Soluble chromatin was collected and incubated with antibodies against Nrf2 (Santa Cruz Biotechnology, Inc.) overnight at 4°C for immunoprecipitation. Rabbit IgG was used as negative control. The recovered DNA was then analyzed by qRT-PCR using primers flanking the ARE of the human NQO1 promoter (39). Results were normalized to input controls.
Anchorage-independent Cell Growth Assay-The soft agar assay was performed in 6-well plates (in duplicate) as described previously (35). Each well consisted of a bottom base layer (0.6% agarose diluted in DMEM) (Bio-Rad) and top layer (0.3% agarose diluted in DMEM) containing 5 ϫ 10 4 cells. We added a few drops of DMEM to the solidified top layer. The top layer was replenished on a weekly basis. After 3 weeks, cells were stained with 0.05% crystal violet overnight at 37°C. Colonies were visualized and counted with light microscopy. Colonies larger than 50 m in diameter were counted from four random 4ϫ objective fields.
Immunohistochemistry (IHC)-Adult female C57BL6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and were housed under controlled temperature, humidity, and lighting conditions. The animals were maintained in facilities approved by the Institutional Animal Care and Use Committee. Four treatment groups were included in the study: 1) an E 2 -treated group, which was treated with 20 mg/kg E 2 for 28 days by mammary fat pad injection; 2) an E 2 and SAHA-cotreated group, which was treated with 20 mg/kg E 2 for 28 days by mammary fat pad injection and with 50 mg/kg SAHA by oral treatment for days 14 -28; 3) a control group, which was injected with ethanol in the fat pad for 28 days; and 4) a control group treated only with SAHA (50 mg/kg) on days 14 -28. Formalin-fixed and paraffin-embedded mammary gland tissue sections were prepared for immunohistochemistry staining as described previously (30). Animals from each treatment group were examined. Polyclonal rabbit anti-Ki67 (H-300, Santa Cruz Biotechnology, Inc.), polyclonal goat anti-Keap1 antibody (E-20, Santa Cruz Biotechnology, Inc.), and polyclonal rabbit anti-Nrf2 antibody (H-300, Santa Cruz Biotechnology, Inc.) were applied and followed by a biotin-conjugated donkey antigoat or goat anti-rabbit secondary antibody (Santa Cruz Biotechnology, Inc.). An avidin-biotin peroxidase substrate kit (Vector Laboratories, Burlingame, CA) was used to develop brown precipitate. Hematoxylin was utilized for staining of nuclei. For quantitative analysis, we calculated the percentage of positively stained mammary epithelial cells.
Statistical Analysis-Statistical analysis was performed using Student's t test, and p values of Ͻ0.01 were considered significant. Data were represented as mean Ϯ S.D. GraphPad Prism 4.0 software was used for all data analysis.

Identification of miR-200a as a Potential Regulator of Keap1
in Breast Cancer Cells-We began our studies interested in probing possible regulation of the Keap1/Nrf2 pathway by miR networks in mammary epithelium. Previously, Nrf2 has been found down-regulated in multiple breast cancer cell lines and patient samples (9). To determine if miRs might play some role in the dysregulation of the Keap1/Nrf2 pathway, we first examined the expression profile of 88 highly characterized miRs in metastatic breast cancer cell line MDA-MB-231 compared with that of the MCF-10A non-tumorigenic mammary epithelial cell line. Fig. 1A shows that the results of our expression profiling revealed up-or down-regulation of 18 miRs by 4-fold or greater in MDA-MB-231 breast cancer cells. Among these 18 miRs were two members examined from the miR-200 family (miR-141 and miR-200c; chromosome 12), which were found to be the most significantly down-regulated miRs (miR-141 Ͼ800-fold, miR-200c Ͼ400-fold) examined with this array. Importantly, when exploring the TargetScan5.1 miR target prediction algorithm, miR-141 was predicted to target a conserved site within the Keap1 3Ј-UTR. In fact, only one conserved targeting site was identified within the Keap1 3Ј-UTR, a site matching the seed sequences for miR-200a and miR-141 (which share identical seed sequences) (Fig. 1C). Based on the predicted targeting information and our array data, we further examined dysregulation of the miR-200 family by examining miR-200a (chromosome 1) expression by qRT-PCR. We found that miR-200a levels were also significantly decreased in breast cancer cell lines MDA-MB-231 and Hs578T (invasive, triplenegative breast cancer cell line) compared with non-tumorigenic MCF-10A cells (Fig. 1B, p Ͻ 0.01).
Keap1 mRNA 3Ј-UTR Contains a Validated miR-200a-targeted Site-In order to test whether miR-200a could regulate the predicted target site within the Keap1 mRNA 3Ј-UTR, we performed luciferase reporter assays by cloning the 3Ј-UTR of Keap1 into luciferase reporter constructs. We also constructed mutant luciferase reporters by mutating 3 bases of the predicted miR-200a target seeding site ((T(C to G)A(G to C)T(G to C)TT) with PCR-based mutagenesis to disrupt miR-200a targeting. We examined reporter activity in HEK293T cells, which do not express detectable levels of miR-200a (data not shown). HEK293T cells were co-transfected with wild type or mutant pSGG Keap1 3Ј-UTR luciferase vectors (along with phGR-TK Renilla luciferase vectors for normalization) in addition to miR-200a expression vector or empty vector control. Co-transfection of miR-200a expression vector was found to decrease wild type Keap1 3Ј-UTR reporter activity by 85% (p Ͻ 0.01) compared with transfection of empty vector controls (Fig. 1C). However, co-transfection of miR-200a expression vector did not significantly alter mutant Keap1 3Ј-UTR reporter activity (p Ͼ 0.05) (Fig. 1C). This result provides validation that miR-  Fig. 2A). Next, we confirmed that miR-200a targeting of Keap1 mRNA also resulted in decreased Keap1 protein levels by Western blotting (Fig. 2B).
Collectively, these results demonstrate that miR-200a targets the 3Ј-UTR of Keap1 mRNA, resulting in destabilization of Keap1 mRNA and a corresponding decrease in Keap1 protein levels. This establishes miR-200a as a potentially important regulator of Keap1 expression and, hence, the Keap1/Nrf2 pathway.
miR-200a Targeting of Keap1 Promotes Nrf2 Nuclear Translocation-In order to confirm if miR-200a regulation of Keap1 impacts Nrf2, we transfected breast cancer cell lines (MDA-MB-231 and Hs578T) with miR-200a expression vector and examined Nrf2 expression in whole-cell lysates and nuclear extracts with Western blotting. Upon overexpression of miR-200a, we found an increase in Nrf2 expression in both whole cell lysates (total Nrf2) and nuclear extract (nuclear Nrf2) compared with control transfections (Fig. 2, B and C). Next, to examine if miR-200a up-regulation of Nrf2 levels restored the function of Nrf2 as the key transcription factor regulating the expression of detoxifying enzymes (e.g. NQO1), we examined an NQO1-ARE promoter luciferase reporter. Co-transfection

miR-200a Targets Keap1 in Human Breast Cancer Cells
of miR-200a expression vector was found to increase NQO1-ARE promoter activity, as measured by luciferase assay (Fig. 2D, p Ͻ 0.01). Finally, we performed qRT-PCR to examine NQO1 (an Nrf2 target gene) mRNA levels after miR-200a transfection. NQO1 mRNA expression in miR-200a-overexpressing breast cancer cell lines was significantly up-regulated (MDA-MB-231 Ͼ2-fold; Hs578T Ͼ7-fold) compared with control transfected cells (Fig. 2E, p Ͻ 0.01) (30). In addition, we found that SAHA treatment decreased Keap1 mRNA levels in both cell lines (Fig. 3B, MDA-MB-231 Ͼ2-fold decrease and Hs578T Ͼ3-fold decrease, p Ͻ 0.01). Finally, we examined Keap1 expression by Western blotting and found that SAHA treatment also resulted in downregulation of Keap1 protein levels in both cell lines (Fig. 3C).
We next examined the specificity with which SAHA treatment had resulted in Keap1 mRNA down-regulation. We treated the Hs578T cell line with synthetic miR-200a inhibitor (antagomiR-200a) and re-examined the impact of SAHA treatment on Keap1 mRNA levels. Transfection with antagomir-200a (50 nM) abolished SAHA treatment-induced Keap1 mRNA downregulation (Fig. 3D). This indicates that SAHA treatment of breast cancer cell lines results in the down-regulation of Keap1 specifically through actions of miR-200a and provides evidence supporting endogenous miR-200a targeting of Keap1.
Next, we again wanted to confirm that miR-200a targeting of Keap1 would result in Nrf2 activation. However, this time we wanted to determine if SAHA treatment would activate Nrf2 through stimulating endogenous miR-200a re-expression. Because SAHA treatment resulted in increased Nrf2 protein levels (Fig. 3C), we then addressed whether increased levels of

miR-200a Targets Keap1 in Human Breast Cancer Cells
Nrf2 correlated with increased localization of Nrf2 to the promoters of antioxidant response genes. Using ChIP, we found that SAHA treatment dramatically increased Nrf2 localization to the NQO1 promoter in Hs578T cells (Fig. 3E). We confirmed the specificity of this response by co-treating with antagomiR-200a, which considerably inhibited SAHA treatment-induced binding of Nrf2 to the NQO1 promoter. The increased occupation of the NQO1 promoter after SAHA treatment was found to correlate with elevated NQO1 mRNA expression (Fig. 3F, Ͼ30fold, p Ͻ 0.01) as determined by qRT-PCR. Again, co-treatment with antagomiR-200a was shown to significantly inhibit the effects of SAHA treatment. It is important to point out that SAHA treatment can affect the expression of many genes. Although the impact of SAHA treatment on Keap1 appears to act through miR-200a (Fig. 3D), treatment with antagomiR-200a does not fully reverse SAHA treatment-induced Nrf2 activity or increased NQO1 gene transcription (Fig. 3, E and F). Therefore, we cannot rule out a Keap1-independent (thus miR-200a-independent) mechanism through which SAHA treatment may activate Nrf2 separate from miR-200a targeting of Keap1 mRNA. In summary, treatment of breast cancer cell lines with epigenetic therapy (SAHA) was found to restore functional Nrf2 expression by reversing miR-200a epigenetic silencing, leading to subsequent Keap1 mRNA down-regulation.
Both Nrf2 Restoration and SAHA Treatment Reduce Anchorageindependent Growth of Breast Cancer Cell Lines-To investigate the functional effects of restoring Nrf2 expression on breast cancer cell lines, we utilized soft agar assays to examine the impact on anchorage-independent cell growth. We tested both SAHA treatment and overexpression of Nrf2 using an Nrf2-myc expression vector (37) (Fig. 4A) and evaluated the impact on colony formation of breast cancer cells in soft agar. After 3 weeks, colony formation in SAHA-treated MDA-MB-231 and Hs578T cells was decreased by 90% (Fig. 4, B and C, p Ͻ 0.01) compared with DMSO-treated controls. Interestingly, MDA-MB-231 and Hs578T cells overexpressing Nrf2 also showed a Ͼ90% reduction in colony formation (Fig. 4, B and C, p Ͻ 0.01) compared with empty control vectortransfected cells. This indicates that up-regulation of Nrf2 can interfere with the anchorage-independent growth of breast cancer cell lines. Because SAHA treatment can impact the expression of many genes, we can only conclude that the impact of SAHA on Keap1/Nrf2 signaling may in part contribute to its ability to decrease colony formation in soft agar.
SAHA Treatment Up-regulates miR-200a and Leads to Increased Nrf2 Expression in Vivo-Finally, using a mouse model of mammary carcinogenesis, we explored miR-200a regulation of Keap1 in vivo and examined the chemopreventative potential for SAHA to regulate Nrf2 activation. We performed

miR-200a Targets Keap1 in Human Breast Cancer Cells
estrogen (E 2 ) injections into fat pads of mouse mammary glands (20 mg/kg E 2 for 28 days) in an attempt to induce the dysregulation of Keap1/Nrf2 expression within a carcinogen-induced model of mammary epithelial transformation. It is well known that estrogen is a carcinogen in breast cancer development and that estrogen treatment promotes mammary epithelial hyperplasia through estrogen receptor signaling and estrogen metabolite-induced genotoxicity (41)(42)(43). Compared with controltreated mice, E 2 -treated mouse mammary glands show extensive mammary epithelial proliferation and disorganization and disappearance of luminal structure, indicating the formation of hyperplastic lesions (see Fig. 4D) (41)(42)(43).

miR-200a Targets Keap1 in Human Breast Cancer Cells
We subsequently examined Keap1 and Nrf2 expression in mammary epithelium of E 2 -treated mice and in mice co-treated with E 2 and SAHA. Using IHC staining, we found significantly higher expression of Keap1 and significantly lower Nrf2 expression in E 2 -treated mammary epithelium compared with control mice (Fig. 4D). With this result, we established that our carcinogen-induced model of mammary epithelial transformation indeed showed dysregulation of the Keap1/Nrf2 pathway. Next, we found that co-treatment with SAHA significantly reversed E 2 -induced dysregulation of Keap1/Nrf2 and resulted in reduced Keap1 and increased Nrf2 expression compared with mice treated with E 2 only (Fig. 4D). Again, in control mice, SAHA treatment by itself did not result in significant changes to Keap1 or Nrf2 expression (Fig. 4D).
Next, we examined Keap1 mRNA and miR-200a expression in mammary glands of E 2 -treated and SAHA-co-treated mice by qRT-PCR. At the mRNA level, Keap1 was found to be significantly up-regulated in mammary glands of E 2 -treated mice (Ͼ2-fold, p Ͻ 0.01) when compared with control mice. Cotreatment with SAHA was found to lower Keap1 mRNA expression in mammary glands when compared with mice treated with E 2 only (Fig. 4E, Ͼ10-fold, p Ͻ 0.01). We next investigated miR-200a levels in the mammary glands of E 2 -treated and SAHA-co-treated mice. Interestingly, in E 2 -treated mice, miR-200a was significantly down-regulated in mammary glands (Ͼ10-fold, p Ͻ 0.01) when compared with control mice. In addition, miR-200a was found to be significantly up-regulated in the mammary glands of SAHA and E 2 -co-treated mice when compared with mice treated with E 2 only (Fig. 4F, Ͼ8-fold, p Ͻ 0.01).
Collectively, these data show that SAHA treatment can stabilize miR-200a expression and prevent Keap1 overexpression in mammary glands in vivo, ultimately preventing Nrf2 downregulation. This confirms, in a mouse model of carcinogenesis, a miR-200a/Keap1/Nrf2 pathway that may play an important cancer-preventative function in mammary epithelium that may be lost during estrogen-dependent initiation and progression of breast cancers.

DISCUSSION
The Nrf2 antioxidant pathway can serve as an important cellular defense against oxidative stress, genotoxicity, and potentially tumor formation (44,45). Keap1 serves as the key regulator and repressor of this pathway (7). The mechanisms that regulate Keap1 expression, however, remain poorly understood. In our study, for the first time, miR regulation of Keap1 expression was examined. The TargetScan5.1 prediction algorithm identified one conserved miR targeting site within the Keap1 mRNA 3Ј-UTR, a site complementary to the miR-200 family members miR-200a and miR-141 (which share identical seed sequences). We have evaluated this prediction and found miR-200a to be involved in the regulation of Keap1 expression at the posttranscriptional level, having validated miR-200a targeting of the 3Ј-UTR of Keap1 mRNA with luciferase reporter assays.
miR-200 family members have been heavily investigated in breast cancer because they play a pivotal role in the maintenance of epithelial phenotype through targeting of ZEB1, SIP1, and SIRT1 mRNAs (30,31). The miR-200 family, including miR-200a (36,46), is often found to be down-regulated in breast cancer cells ( Fig. 2A) (26,27). Because down-regulation of Nrf2 has been found in breast cancers and is often preceded by the dysregulation of Keap1 (9,45,47), we investigated the possible relationship between these two phenomena and found that loss of miR-200a correlated with Keap1 up-regulation and Nrf2 down-regulation in breast cancer cell lines.
Aberrant epigenetic silencing of gene expression has been found to be a prominent mechanism for miR dysregulation in cancers, where epigenetic therapies are often found to be capable of restoring expression of silenced miRs (48 -52). DNA hypermethylation has been found to be associated with miR-200 family silencing in breast cancer (28). We find that treatment with epigenetic therapy (the HDAC inhibitor SAHA) can lead to re-expression of miR-200a in breast cancer cells. We also find that re-expression of miR-200a after SAHA treatment correlates with Keap1 down-regulation and activation of the Nrf2 pathway.
Having established that epigenetic therapy can manipulate the Keap1/Nrf2 pathway, we next examined what functional consequence this would have on tumorigenic properties of breast cancer cell lines, such as anchorage-independent growth. We found that Nrf2 restoration (either directly or by SAHA treatment) decreased soft agar colony formation in MDA-MB-231 and Hs578T cell lines. The mechanism by which Nrf2 restoration inhibits cell growth in soft agar is currently unknown, but previously, Nrf2 was shown to inhibit tumor cell growth by inducing cell cycle arrest in late G 1 phase (18) and was also shown to interfere with the TGF-␤ pathway (16). It is possible that these and additional mechanisms may provide a means for Nrf2 inhibition of breast cancer cell growth in soft agar. Regardless, our findings offer further support for a role for Nrf2 in the inhibition of tumor cell growth (47).
Finally, we confirmed our in vitro findings in vivo using a mouse model of mammary gland carcinogenesis. We found that treatment of mouse mammary glands with high levels of the carcinogen estrogen resulted in a hyperproliferative phenotype. We discovered that treatment with high levels of estrogen induced miR-200a down-regulation, Keap1 overexpression, and Nrf2 down-regulation in mammary epithelium. Conversely, we found that co-treatment with SAHA provided a chemopreventative effect, reducing estrogen-induced mammary epithelial hyperproliferation. We found that the protective effects of SAHA treatment might be mediated through stabilization of miR-200a expression and down-regulation of Keap1 and thus act through the restoration of Nrf2 expression. This provides compelling evidence for the existence of a miR-200a/ Keap1/Nrf2 pathway in normal mammary epithelium that may be dysregulated in breast cancer.
In conclusion, our study not only demonstrates a novel mechanism by which Keap1 expression is regulated but also provides insights as to what might contribute to the loss of Nrf2 activity in breast cancer by providing evidence supporting a role for miR-200a silencing in the overexpression of Keap1. Finally, we show that epigenetic therapy (HDAC inhibitors) can inhibit breast cancer cell growth, in part, through activation of the

miR-200a Targets Keap1 in Human Breast Cancer Cells
Nrf2-dependent antioxidant pathway, mediated by miR-200a re-expression and the targeting of the Keap1 3Ј-UTR.