Cyclohexylmethyl Flavonoids Suppress Propagation of Breast Cancer Stem Cells via Downregulation of NANOG

Breast cancer stem cells (CSCs) are highly tumorigenic and possess the capacity to self-renew. Recent studies indicated that pluripotent gene NANOG involves in regulating self-renewal of breast CSCs, and expression of NANOG is correlated with aggressiveness of poorly differentiated breast cancer. We initially confirmed that breast cancer MCF-7 cells expressed NANOG, and overexpression of NANOG enhanced the tumorigenicity of MCF-7 cells and promoted the self-renewal expansion of CD24−/lowCD44+ CSC subpopulation. In contrast, knockdown of NANOG significantly affected the growth of breast CSCs. Utilizing flow cytometry, we identified five cyclohexylmethyl flavonoids that can inhibit propagation of NANOG-positive cells in both breast cancer MCF-7 and MDA-MB231 cells. Among these flavonoids, ugonins J and K were found to be able to induce apoptosis in non-CSC populations and to reduce self-renewal growth of CD24−/lowCD44+ CSC population. Treatment with ugonin J significantly reduced the tumorigenicity of MCF-7 cells and efficiently suppressed formation of mammospheres. This suppression was possibly due to p53 activation and NANOG reduction as either addition of p53 inhibitor or overexpression of NANOG can counteract the suppressive effect of ugonin J. We therefore conclude that cyclohexylmethyl flavonoids can possibly be utilized to suppress the propagation of breast CSCs via reduction of NANOG.


Introduction
Breast cancer is a leading cause of cancer death among women, as cancer recurrence and metastasis occur frequently in breast cancer patients [1,2]. Accumulating evidence indicates that CD24 −/low CD44 + breast cancer cells, also referred to as "tumorigenic breast cancer cells" [3,4], "breast cancer stem cells (CSCs)" [5], and "stem-like breast cancer cells" [6], possess stem cell characteristics, display resistance to conventional therapies, and have high tumor-initiating and metastatic ability [3,4,[7][8][9]. Therefore, the presence of breast CSCs has been suggested to be the underlying cause of breast cancer recurrence and metastasis [2,8,9]. In order to improve breast cancer therapeutics, efforts are now being directed towards identifying strategies that target breast CSCs [2,9].
Accumulating evidence supports that self-renewal regulators of normal stem cells may govern clinical behavior of human cancer [10,11]. For example, embryonic stem cell (ESC) signature is associated with poor clinical outcome in patient of breast cancer patients [12]. Among the regulatory genes involved in pluripotent maintenance of ESCs, NANOG was found to express a NANOGP8 retrogene 2 Evidence-Based Complementary and Alternative Medicine locus in a wide variety of somatic and cancer cells [13][14][15]. Recent work has shown that NANOG was functionally involved in human tumor development and in regulating cancer stemness [15,16]. Knockdown of NANOG significantly reduced the tumorigenic potentials of various cancer cells including breast cancer [17]. NANOG has also been identified in breast cancer cells and was found to mediate multidrug resistance via activation of STAT3 signaling [18] suggesting that NANOG is a potential target for breast cancer therapeutics.
Herbal medicine has been proposed for utilizing a complementary approach for control of breast cancer recurrence and metastasis [19,20]. However, whether the activity of breast CSCs can be suppressed by treatment of herbal medicine has never been addressed. In Chinese traditional medicine, the roots of the fern Helminthostachys zeylanica (L.) Hook. (Ophioglossaceae), known as "Ding-Di-U-Gon", is used as antipyretic and antiphlogistic agent to treat inflammatory diseases, various hepatic disorders, and possibly malignancy in pancreas [21][22][23]. The rhizome of this medicinal fern is also named as "tunjuk langit" in India which has been used as a folk medicine to treat pulmonary disease and even to cure impotency by the tribal people [24]. In Malaysia, the rhizome is used as an antidiarrheal agent and chewed with areca for whooping cough relief [25]. However, efforts to evaluate the efficacy of such treatment on CSCs and to identify responsible principles of its effect on cancer were scarce.
In the present study, a group of natural cyclohexylmethyl flavonoids isolated from the rhizomes of H. zeylanica had been examined. Utilizing flow cytometry, we identified five members of natural cyclohexylmethyl flavonoids that can inhibit expansion of NANOG + cells. Among these cyclohexylmethyl flavonoids, ugonins J and K, which were the main components of the ethyl acetate-soluble extract of the rhizomes of H. zeylanica, were able to suppress propagation of CD24 −/low CD44 + breast cancer stem cells both in vitro and in vivo.

Chemicals. Doxorubicin (Dox) was obtained from
Sigma. Ugonins (J-S) were isolated and purified from the rhizomes of Helminthostachys zeylanica [21]. All of the ugonins used in the experiments were repurified by reversed-phase HPLC to ensure the purity >99%.

Flow Cytometric
Analysis. Cells were trypsinized and washed three times with PBS before resuspension in Hanks' Balanced Salt Solution (HBSS; Invitrogen) containing 2% FBS and 10 mM HEPES (Invitrogen). The cell density was adjusted to 10 6 /100 L in staining buffer before being stained with antibodies FITC-conjugated anti-CD24 (BD Biosciences) and APC-conjugated anti-CD44 (BD Biosciences) for 30 minutes. In some experiments, MCF-7 cells were stained with anti-NANOG antibodies (Cell Signaling) followed by staining with FITC-conjugated goat anti-rabbit IgG (BD Biosciences). Stained cells were analyzed utilizing FACSCalibur flow cytometry (BD Biosciences) after the addition of propidium iodide (2 g/mL) to exclude dead cells.

Immunofluorescent
Staining. MCF-7 cells (5 × 10 4 cell/ well) were seeded in the 24-well plate and cultured overnight. After cells were treated with different compounds for different time course, cells were fixed by 4% PFA (Sigma) for 30 minutes at room temperature and permeabilized at room temperature in 0.1% Triton X-100 for 30 minutes. After blocking with 2% Roche blocking reagent, the cells were incubated with primary antibody overnight at 4 ∘ C and with secondary antibody for 2 hours at room temperature. The primary antibodies were used at the following dilutions: rabbit anti-NANOG 1 : 100 (Cosmo Bio USA, Inc) and rabbit antiphospho-p53 S15 1 : 400 (Cell Signaling). Cells were counter-stained with Hoechst dye (Sigma) to visualize the cell nuclei. Images of the immunostaining were obtained using a fluorescence microscopy (Leica Microsystems Inc).

Establishment of NANOG-Knockdown Cells.
The lentiviral shNANOG construct (TRCN0000004884) was obtained from the National RNAi Core Facility (Institute of Molecular Biology/Genomic Research Center, Academia Sinica), and the lentivirus was generated as described in the previous section. MCF-7 cells were infected with shNANOG lentivirus, and then NANOG-knockdown cells were selected in -MEM containing 1 g/mL puromycin.

Analysis of the Promoter and p53-Binding Site of NANOG and NANOGP8.
To analyze the elements upstream of NANOG and NANOGP8, the 5-kb upstream sequences of the translation start sites of NANOG and NANOGP8 were retrieved from the human RefSeq files (NC 000012 and NC 000015, resp.). The p53MH program (PMID: 12077306) was employed to detect possible P53-binding site within the 5-kb sequence. The top 100 possible p53-binding sites were extracted. For the identification of the most likely binding site, the threshold of the percentage of maximum possible score was set as 80%. The prediction of the promoter region was carried out with CoreBoost HM (PMID: 18997002). The score of 0.7 was set as a cutoff value for the plausible promoter region.

Establishment of Orthotropic Tumor Xenografts in SCID
Mice. All animal experiments were approved by the Academia Sinica Institutional Animal Care and Utilization Committee. Four-week-old female SCID mice purchased from BioLASCO were used to carry out MCF-7 xenograft experiments. For tumorigenicity assay, eighteen mice were divided into three groups (6 mice/group) and were injected in the mammary fat pad with Control, NANOGoverexpressing, or NANOG-knockdown MCF-7 cells (1×10 6 cells/60 L). To determine if ugonin J can suppress tumor growth, eighteen mice were divided into three groups (6 mice/group) and were injected in the mammary fat pad with MCF-7 cells (2 × 10 5 cells/60 L). When the tumor volume reached 50 mm 3 (set as Day 0), the tumor-bearing mice were then administered a weekly dose of doxorubicin (12 mg/kg, dissolved in 100 L of DMSO) or ugonin J (50 mg/kg, dissolved in 100 L of DMSO) interperitoneally for a total of 4 doses. Body weight of mice and tumor size were measured weekly.

Histology and Immunohistochemistry.
Tumor tissues were fixed overnight at room temperature with 3.5% formaldehyde solution containing 68.6% EtOH and 4.8% acetic acid (FAA fixative) prior to being processed and embedded in paraffin. 4 m thick sections were cut and mounted on Superfrost plus slides (Thermo Scientific). For immunohistochemical staining, sections were subjected to antigen retrieval in Citric-acid based buffer (Vector Laboratories) at 95 ∘ C for 20 minutes. The sections were then permeabilized with 0.1% (v/v) Triton X-100 in PBS for 30 min and incubated in 2% blocking buffer (Roche) before being incubated sequentially with primary, HRP-conjugated secondary antibodies. Super Sensitive Polymer HRP IHC Detection System (Biogenex Laboratories) was used to visualize the positive cells. Sections were counterstained with hematoxylin and mounted with Entellan Neu (Merck). The primary antibodies were used at the following dilutions: rabbit anti-Nanog 1 : 150 (Cosmo Bio), mouse anti-MUC 1 1 : 100 (Abcam), and mouse anti-HCAM/CD44 1 : 100 (SantaCruz).

Invasion
Assay. 1 × 10 4 of MCF-7 cells suspended in serum-free medium with or without ugonins J or K was seeded into the top chamber of the matrigel-coated insert (Millicell, 24-well plate, 8 m, Millipore) in 100 L serum-free medium. In the lower chamber, the well was filled with serum-containing medium which was used as a chemoattractant. After 24-hour incubation, cells that did not invade through the pores were removed by a cotton swab. Cells on the lower surface of the membrane were fixed with methanol and stained with Giemsa solution (Merck). The number of invasive cells/each well was counted under a light microscope. Data are representative of three independent experiments. * * * < 0.001, * * < 0.01 versus compared control.

Statistical Analysis.
Experiments were repeated at least three times with consistent results. Statistical differences between groups were determined by unpaired Student's t test. The statistical significance was set at * < 0.05, * * < 0.01, * * * < 0.001. FACS data were analyzed by FlowJo software (Ashland, OR, USA). The statistical analysis for fluorescent staining used MetaMorph imaging analytical software (Molecular Devices).

A Critical Role of NANOG in Modulating Proliferation and
Tumorigenicity of Breast Cancer Cells. We initially investigated whether expression of NANOG plays an important role in breast cancer growth. To address this question, we generated NANOG-overexpressing and NANOG-knockdown MCF-7 cell lines. As shown in Figure 1(a), RNA interferencemediated NANOG knockdown reduced breast cancer. And overexpression of NANOG slightly increased the overall growth rate. To further determine if NANOG is the key component modulating self-renewal capability and tumorigenicity of the tumorigenic breast cancer cells, we carried out the mammosphere-forming assay and orthotropic tumor xenografts experiments in female SCID mice. As shown in Figures 1(b) and 1(c), NANOG-overexpressing cells formed 20% more of mammospheres and generated twofold larger tumor xengrafts than control MCF-7 cells. In contrast, knockdown of NANOG not only significantly reduced the ability to form mammospheres, but also dramatically reduced the tumorigenicity of MCF-7 cells. Immunohistochemical analysis of tumor xengrafts (Figure 1(d)  Since NANOG knockdown suppressed mammosphere formation and reduced levels of SOX2 and MUC1 in tumor xengrafts, we next tried to determine if propagation of CD24 −/low CD44 + breast CSC subpopulation in MCF-7 cells is also regulated by NANOG [3,4]. As shown in Figure 2, we found that overexpression of NANOG increased the proportion of CD24 −/low CD44 + CSC subpopulation in MCF-7 cells from 6.8% to 25.8%. In contrast, NANOG knockdown reduced the proportion of CD24 −/low CD44 + CSC subpopulation in MCF-7 cells from 6.8% to 1.88%. These data indicated that NANOG played an important role in modulating selfrenewal and tumorigenicity of breast CSC subpopulations.

Identification of Bioactive Cyclohexylmethyl Flavonoids
Targeting NANOG + Breast Cancer Cells. We have explored that NANOG possibly played a critical role in modulating self-renewal expansion and tumorigenicity of breast CSCs. We therefore then assume identification of bioactive natural components from herb medicine that can suppress that NANOG would be beneficial for developing a complementary approach for control of breast CSC-driven recurrence and metastasis. Since the dietary flavonoids were reported to possess the ability to suppress the prostate CSCs via inhibiting NANOG [28], a group of natural cyclohexylmethyl flavonoids isolated from the rhizomes of H. zeylanica that had been examined. Initially, an MTT colorimetric assay was used to determine cytotoxicity of cyclohexylmethyl flavonoids to two breast cancer cell lines (MCF-7 and MDA-MB231) ( Table 1). Among these flavonoids, ugonins J and K were found to display cytotoxicity (IC 50 < 25 M) to breast cancer cells. In contrast, these two ugonins were less cytotoxic to normal foreskin fibroblasts (HFF). Utilizing flow cytometry, we identified five members of natural cyclohexylmethyl flavonoids that inhibited expansion of NANOG + population in both MCF-7 and MDA-MB 231 cells (Figures 3(a) and  3(b)). Among these natural cyclohexylmethyl flavonoids, based on using immune-fluorescent staining, we validated that either treatment of ugonins J or K, both compounds were the main component of the ethyl acetate-soluble extract of the rhizomes of H. zeylanica, significantly reduced the expression level of NANOG and MUC1 in MCF-7 cells (Figure 3(c)).

Downregulation of NANOG Mediates the Suppressive Effect of Ugonin J on Propagation of Breast Cancer Stem
Cells. The ability of formation of mammospheres is known as one of self-renewal characteristics of breast CSCs; we then determined if treatment of ugonin J can suppress mammosphere-forming ability. In comparing with NANOG overexpression increased mammosphere formation, pretreatment with ugonin J completely inhibited formation of 6 Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine 7 mammospheres in control MCF-7 cells (Figure 4(a)). In contrast, NANOG overexpression partially counteracted the suppressive effect of ugonin J on mammosphere formation. We then determined if ugonin J can reduce the malignant features of MCF-7 cells including invasion ability, and IL-6 secretion led to STAT3 phosphorylation in mammospheres (Figure 4(b)) [29]. Treatment with ugonin J or K for 24 hours significantly suppressed invasion ability of MCF-7 cells. Moreover, treatment of 28-day mammospheres with ugonin J for 24 hours significantly reduced IL-6 secretion and STAT3 phosphorylation in both control mammospheres and NANOG-overexpressing mammospheres (Figures 4(c) and 4(d)). These results suggest that ugonin J-mediated downregulation of NANOG may be a key event affecting the propagation of breast CSCs.

P53-Dependent Pathway Mediates Downregulation of NANOG by Ugonin J Treatment. NANOG is a pluripotent regulator of embryonic stem cells, and previous studies have
shown that p53 binds to the promoter of NANOG and suppresses NANOG expression after DNA damage [30,31]. However, it has recently been shown that there are 11 NANOG pseudogenes [32]. NANOGP8 has been recognized as a retrogene and was recently found to be expressed in various cancer tissues and several cancer cell lines including breast cancer MCF-7 cells. We therefore determined if NANOGP8 can also be regulated by p53. The p53MH program was employed to detect possible P53-binding site within the 5-kb sequence in the NANOG and NANOGP8 promoter regions. As shown in Figure 5, both NANOG and NANOGP8 promoter regions contained several potential binding site for p53. To further determine if p53 pathway can be activated by ugonin J treatment. Time-course experiments were performed and showed that treatment of ugonin J (Figure 6(a)) did in fact increase phosphorylation of p53 at ser15 and ser392 and also activated the apoptotic pathway, as evidenced by cleaved forms of Poly (ADP-ribose) polymerase (PARP) and caspase 9 in western blot analysis. In order to determine whether the downregulation of NANOG in MCF-7 cells with ugonin J treatment was directly mediated by p53, we generated p53-overexpressing MCF-7 cells. A 60% reduction of NANOG + cells was found in p53-overexpressing MCF-7 cells, combined treatment with ugonin J further reduced 90% of NANOG + cells. In contrast, treatment of pifithrin-(p53 inhibitor) rescued the reduction of NANOG induced by ugonin J (Figures 6(b) and 6(c)). The results suggested that activation of p53 pathway mediated the effect of ugonin J on suppression of the NANOG expression. confirmed that treatment with ugonin J suppressed NANOG expression. In contrast, some Dox-treated cancer cells still expressed NANOG (Figure 7(c)) which can explain how tumors can still be slightly propagated (Figure 7(a)). These results suggest that ugonin J can suppress the propagation of breast CSCs in vivo via reduction of NANOG.

Discussion
NANOG is a transcriptional factor that plays key roles in the self-renewal and maintenance of pluripotency in embryonic stem cells [31]. There are 11 NANOG pseudogenes [32]. NANOGP8 has been recognized as a retrogene and was recently found to be expressed in various cancer tissues and several cancer cell lines including the MCF-7 cells used in the current study. We have previously shown that activation of p53 by disrupting porphyrin homeostasis in embryonic stem cells resulted in suppression of NANOG expression [33]. In the current work, we observed a similar phenotype, where treatment of MCF-7 cells with cyclohexylmethyl flavonoids induced activation of p53, which in turn led to the reduction of NANOG expression. This suggests that NANOG expression is regulated by a similar mechanism in both breast CSCs and embryonic stem cells. Recent work further indicates that NANOG could be upregulated by beta-catenin through interaction with Oct3/4 [34]. We have evaluated the possibility by immunohistochemical analysis and Top/Fop flash assay (data not shown) and found that ugonin J treatment decreased the level of beta-catenin in tumor xengraft. However, the activity of beta-catenin was extremely low in both MCF-7 and MDA-MB231 cells. We therefore proposed that it is possible that Ugonin J treatment causes concomitant downregulation of beta-catenin and NANOG in breast cancer, but, in absence of wnt/beta-catenin, ugonin J is capable to downregulate NANOG expression through p53 activation. P53, a wellknown tumor suppressor protein, involves regulating cell cycle, senescence, and apoptosis responses against the cell suffering from stress such as hypoxia or DNA damage. In most cancers, p53 is either lost or mutated to allow cancer cells to expand and progress [35]. Recent reports raised the possibility to suppress tumor growth by restoring wild-type p53 to cancer cells [36]. Our current work further highlights the importance of restoring the function of p53 in CSCs. Recent work has further demonstrated that NANOG transcribed from the NANOGP8 locus is important in tumorigenesis [16]. RNA interference-mediated NANOG knockdown inhibited tumor development in xenograft animals and decreased long-term clonal and clonogenic growth of cancer cells [16,17]. These results are consistent with our findings that overexpression of NANOG enhances the overall growth rate of MCF-7 cells and downregulation of NANOG by ugonin J treatment suppresses propagation of breast CSCs. However, the mechanisms, involved in regulating transcription of NANOG from the NANOGP8 locus during breast carcinogenesis, remain to be determined.
We have tried to determine the structure-activity relationship (SAR) of several cyclohexylmethyl flavonoids with high potency to suppress NANOG that may possess the specific structural features. We proposed that 6, 6-dimethyl-2-methylene-cyclohexylmethyl groups on the C-6 position are important for the potency of ugonins J and K to suppress propagation of breast CSCs. In contrast, the bulky isoprenyl group attached to position 2 of the B ring (found in ugonins M, N and O) may reduce the potency. In addition, the free rotation of the bulky isoprenyl moiety (ugonins J and K) may contribute more stereohindrance and lipophilic properties compared with the cyclized moiety with C-4 by ether-linkage (ugonins L and S), as the double bond in the cyclohexane (ugonin P) disrupts the chair form of the cyclohexane moiety and reduces its lipophilicity. And one hydroxyl group attached to the cyclohexyl ring (ugonins Q and R) might increase the hydrophilicity, which would also reduce the potency. It has been reported that derivatives of ambrein and agelasine that possessed cyclohexylmethyl groups are capable to suppress the expansion of multiple cancer cell lines [37,38]. This may explain why ugonins J and K exhibited relatively high potency to suppress propagation of breast CSCs. CD24 −/low CD44 + breast CSCs have been suggested to be the underlying cause of breast cancer recurrence and are a critical target for breast cancer therapies. H. zeylanica have been used in Chinese traditional medicine for treating inflammatory diseases and various hepatic disorders. In the present study, we have identified two cyclohexylmethyl flavonoids, ugonins J and K, which were the main components of the rhizomes of H. zeylanica and were able to suppress propagation of breast CSCs in mammosphere cultures and in tumor xengrafts. The current work also found that the suppressive effect of ugonin J on propagation of breast CSCs was mediated by activation of p53 which in turn led to reduction of NANOG. Overexpression of NANOG counteracted the suppressive effect of ugonin J. The current findings suggest that the rhizomes of H. zeylanica can possibly be used as complementary medicine for reducing CSCmediated breast cancer recurrence.