Inhibitory Effects of Tangeretin, a Citrus Peel-Derived Flavonoid, on Breast Cancer Stem Cell Formation through Suppression of Stat3 Signaling

Breast cancer stem cells (BCSCs) are responsible for tumor chemoresistance and recurrence. Targeting CSCs using natural compounds is a novel approach for cancer therapy. A CSC-inhibiting compound was purified from citrus extracts using silica gel, gel filtration and high-pressure liquid chromatography. The purified compound was identified as tangeretin by using nuclear magnetic resonance (NMR). Tangeretin inhibited cell proliferation, CSC formation and tumor growth, and modestly induced apoptosis in CSCs. The frequency of a subpopulation with a CSC phenotype (CD44+/CD24−) was reduced by tangeretin. Tangeretin reduced the total level and phosphorylated nuclear level of signal transducer and activator of transcription 3 (Stat3). Our results in this study show that tangeretin inhibits the Stat3 signaling pathway and induces CSC death, indicating that tangeretin may be a potential natural compound that targets breast cancer cells and CSCs.


Introduction
Breast cancer is the most frequently diagnosed cancer and the leading cause of cancer-related death among females [1]. Triple-negative breast cancer (TNBC) accounts for approximately 10-15% of all diagnosed breast cancers [2], and is defined as ER-, PR-and HER2-negative breast cancer. Women with TNBC have a high risk of recurrence within three years of diagnosis, and the mortality rate is increased for five years after diagnosis [3]. Cancer stem cells (CSCs) are relatively resistant to chemotherapy, radiotherapy and hormone therapy [4]. CSCs have functional roles in self-renewal and differentiation [5]. CSCs are responsible for the processes of cancer initiation, metastasis and cancer relapse [6]. Therefore, breast CSCs may contribute to drug resistance and relapse [7]. The CD44 + /CD24 − being the most common cell-surface phenotype of breast CSCs can facilitate invasion, migration and proliferation [8].
Signal transducer and activator of transcription 3 (Stat3) plays a role in the inflammatory response and is a member of the seven-member Stat protein family (Stat1, 2, 3, 4, 5a, 5b, and 6) that is activated by growth factors [9]. Stat3 undergoes alternative splicing into Stat3α (92 kDa) and the isoform Stat3β  formation in the presence of the HPLC-purified sample. Cancer cells were treated with the HPLCpurified sample. Images show representative mammospheres, and were imaged by microscopy (scale bar: 100 μm).

Tangeretin Suppresses the Proliferation of MDA-MB-231 and MCF-7 Cells and the Formation of Mammospheres
Breast cancer cells were incubated with various concentrations of tangeretin for 24 h. The antiproliferative function of tangeretin was assayed. Tangeretin inhibited cell proliferation, as shown

Tangeretin Suppresses the Proliferation of MDA-MB-231 and MCF-7 Cells and the Formation of Mammospheres
Breast cancer cells were incubated with various concentrations of tangeretin for 24 h. The antiproliferative function of tangeretin was assayed. Tangeretin inhibited cell proliferation, as shown in Figure 3A. To test whether tangeretin can suppress mammosphere formation, we treated cancer cells with tangeretin. Tangeretin inhibited mammosphere formation, as shown in Figure 3B. Tangeretin suppressed the migration and colony formation of cancer cell lines ( Figure 3C,D). Our data indicated that tangeretin inhibited cancer hallmarks (proliferation, migration and colony formation) and mammosphere formation. in Figure 3A. To test whether tangeretin can suppress mammosphere formation, we treated cancer cells with tangeretin. Tangeretin inhibited mammosphere formation, as shown in Figure 3B. Tangeretin suppressed the migration and colony formation of cancer cell lines ( Figure 3C,D). Our data indicated that tangeretin inhibited cancer hallmarks (proliferation, migration and colony formation) and mammosphere formation. Representative data were collected. The data from triplicate experiments are represented as the mean ± SD; *p < 0.05, **p < 0.01.

Tangeretin Modestly Induces Apoptosis in Mammospheres and Inhibits Mammosphere Proliferation
Late apoptosis in mammospheres was modestly induced by 80 μM of tangeretin ( Figure 4A). Tangeretin reduced the transcript level of stem cell marker genes (Oct3/4, Sox2, and Nanog gene) ( Figure 4B). To test whether tangeretin suppresses mammosphere growth, we treated mammospheres with tangeretin and counted the number of cancer cells derived from mammospheres. Tangeretin treatment inhibited mammosphere growth ( Figure 4C). Our data showed that tangeretin, which disregulates the Stat3/Sox2 signaling pathway, was essential for inhibiting the proliferation of BCSCs. The data from triplicate experiments are represented as the mean ± SD; * p < 0.05, ** p < 0.01.

Tangeretin Modestly Induces Apoptosis in Mammospheres and Inhibits Mammosphere Proliferation
Late apoptosis in mammospheres was modestly induced by 80 µM of tangeretin ( Figure 4A). Tangeretin reduced the transcript level of stem cell marker genes (Oct3/4, Sox2, and Nanog gene) ( Figure 4B). To test whether tangeretin suppresses mammosphere growth, we treated mammospheres with tangeretin and counted the number of cancer cells derived from mammospheres. Tangeretin treatment inhibited mammosphere growth ( Figure 4C). Our data showed that tangeretin, which disregulates the Stat3/Sox2 signaling pathway, was essential for inhibiting the proliferation of BCSCs. Molecules 2020, 25, x FOR PEER REVIEW 5 of 14 We induced mammosphere formation and treated mammospheres with tangeretin. After treatment, apoptotic cells were examined using Annexin V/PI staining. (B) Transcription levels of CSC markers, including the Nanog, Sox2 and Oct4 genes, determined in tangeretin-and DMSO-treated mammospheres using CSC marker-specific primers and real-time PCR. β-actin acts as an internal control. The data shown represents the mean ± SD of three independent experiments. *p < 0.05 vs. DMSO-treated control. (C) Tangeretin inhibited mammosphere growth. Mammospheres with or without tangeretin were dissociated into single cells, and the single cells were plated in 6 cm dishes in equal numbers. The cells were examined one, two and three days later.

Tangeretin Decreases Tumor Growth In Vivo
As tangeretin has antiproliferative effects, we examined whether tangeretin suppresses tumor formation in a nude mouse model. There was no significant body weight difference between control and tangeretin-treated mice ( Figure 5A). At each time point, the tumor volume ( Figure 5B) and weight ( Figure 5C) of the tangeretin-treated nude mice were smaller than those of the untreated nude mice.  We induced mammosphere formation and treated mammospheres with tangeretin. After treatment, apoptotic cells were examined using Annexin V/PI staining. (B) Transcription levels of CSC markers, including the Nanog, Sox2 and Oct4 genes, determined in tangeretin-and DMSO-treated mammospheres using CSC marker-specific primers and real-time PCR. β-actin acts as an internal control. The data shown represents the mean ± SD of three independent experiments. * p < 0.05 vs. DMSO-treated control. (C) Tangeretin inhibited mammosphere growth. Mammospheres with or without tangeretin were dissociated into single cells, and the single cells were plated in 6 cm dishes in equal numbers. The cells were examined one, two and three days later.

Tangeretin Decreases Tumor Growth In Vivo
As tangeretin has antiproliferative effects, we examined whether tangeretin suppresses tumor formation in a nude mouse model. There was no significant body weight difference between control and tangeretin-treated mice ( Figure 5A). At each time point, the tumor volume ( Figure 5B) and weight ( Figure 5C) of the tangeretin-treated nude mice were smaller than those of the untreated nude mice.

Tangeretin Decreases Tumor Growth In Vivo
As tangeretin has antiproliferative effects, we examined whether tangeretin suppresses tumor formation in a nude mouse model. There was no significant body weight difference between control and tangeretin-treated mice ( Figure 5A). At each time point, the tumor volume ( Figure 5B) and weight ( Figure 5C) of the tangeretin-treated nude mice were smaller than those of the untreated nude mice.
Molecules 2020, 25, x FOR PEER REVIEW 6 of 14 and tangeretin-treated mice were assayed after sacrifice at day 35. The data are presented as the mean ± SD of three independent experiments. **p < 0.05 versus the DMSO-treated control group.

Tangeretin Inhibits the Stat3 Signaling Pathway and Reduces the Sox2 Level in Mammospheres
To investigate the cellular mechanism by which tangeretin inhibits mammosphere formation, we assessed the expression levels of Stat3 and pStat3 in mammospheres. Our results showed that tangeretin decreased the total protein levels of Stat3 and pStat3 in BCSCs ( Figure 7A). The protein level of phospho-Stat3 was significantly reduced in the cytosol and nucleus of mammosphere cells. The Stat3 protein level was also decreased, as shown in Figure 7B. Additionally, we investigated Stat3 probe DNA binding to tangeretin-treated nuclear extracts by EMSA. We examined Stat3 probe DNA binding to mammosphere nuclear proteins using a Stat probe. Tangeretin reduced Stat3 DNA binding ( Figure 7C, # 3). The specificity of Stat3 binding was determined using a self-competitor (100×) ( Figure 7C, # 4) or a mutated Stat oligo (100×) ( Figure 7C, # 5). The band indicated by arrows is Stat3 and the specific DNA complex. To examine the effect on Stat3 on mammosphere formation, we performed mammosphere formation using siRNA of Stat3. Our data showed that Stat3 reduction decreased mammosphere formation ( Figure 7D). To analyze the cellular function of tangeretin, after tangeretin treatment, we checked the Sox2 level because Stat3 dimer activated Sox2 gene [18]. Our

Tangeretin Inhibits the Stat3 Signaling Pathway and Reduces the Sox2 Level in Mammospheres
To investigate the cellular mechanism by which tangeretin inhibits mammosphere formation, we assessed the expression levels of Stat3 and pStat3 in mammospheres. Our results showed that tangeretin decreased the total protein levels of Stat3 and pStat3 in BCSCs ( Figure 7A). The protein level of phospho-Stat3 was significantly reduced in the cytosol and nucleus of mammosphere cells. The Stat3 protein level was also decreased, as shown in Figure 7B. Additionally, we investigated Stat3 probe DNA binding to tangeretin-treated nuclear extracts by EMSA. We examined Stat3 probe DNA binding to mammosphere nuclear proteins using a Stat probe. Tangeretin reduced Stat3 DNA binding ( Figure 7C, # 3). The specificity of Stat3 binding was determined using a self-competitor (100×) ( Figure 7C, # 4) or a mutated Stat oligo (100×) ( Figure 7C, # 5). The band indicated by arrows is Stat3 and the specific DNA complex. To examine the effect on Stat3 on mammosphere formation, we performed mammosphere formation using siRNA of Stat3. Our data showed that Stat3 reduction decreased mammosphere formation ( Figure 7D). To analyze the cellular function of tangeretin, after tangeretin treatment, we checked the Sox2 level because Stat3 dimer activated Sox2 gene [18]. Our data showed that tangeretin reduced the level of transcripts and protein of Sox2 ( Figure 7E). Sox2 plays a role in CSC progression [19]. Our data showed that the Stat3/Sox2 signal is important for mammosphere formation. Our data showed that tangeretin, which disregulates the Stat3/Sox2 signaling pathway, was essential for inhibiting the proliferation of BCSCs (Figure 8).
Molecules 2020, 25, x FOR PEER REVIEW 7 of 14 data showed that tangeretin reduced the level of transcripts and protein of Sox2 ( Figure 7E). Sox2 plays a role in CSC progression [19]. Our data showed that the Stat3/Sox2 signal is important for mammosphere formation. Our data showed that tangeretin, which disregulates the Stat3/Sox2 signaling pathway, was essential for inhibiting the proliferation of BCSCs (Figure 8). downregulated MDA-MB-231 cells were cultured for seven days using mammosphere media. Images were obtained by microscrope at 100× magnification. (E) The transcriptional level of the Sox2 gene in MDA-MB-231 was determined in tangeretin-treated mammospheres. A Sox2-specific primer was used for real-time RT-qPCR. Western blot analysis of mammosphere under tangeretin. β-actin served as an internal control. The data are presented as the mean ± SD of three independent experiments. **p < 0.01 versus the DMSO-treated control group.

Discussion
It has been postulated that high consumption of fruits can prevent more than 20% of all cancer cases [20]. This preventive effect is predominantly mediated by phytochemicals interacting with specific target proteins that play important roles in cancer [21][22][23]. Citrus, one of the most important food sources of phytochemicals with health benefits, has many biological properties and controls key pathways involved in pathologies such as cancer [24][25][26]. The combination CD44 + /CD24 − has emerged as the most important marker for BCSC isolation, and the CD24 population of MDA-MB-231 is low [27].
First, we purified a CSC inhibitor from citrus. Assay-based fractionation and several chromatographic methods isolated one compound from a citrus powder, tangeretin. Tangeretin is the major flavonoid of citrus. It also has antioxidant, anti-inflammatory and anticancer properties [17]. Tangeretin modestly induces apoptosis in bladder cancer cells through mitochondrial dysfunction [28]. Tangeretin and nobiletin induces G1 cell cycle arrest but not apoptosis in breast and colon cancers [29]. Nobiletin inhibits CD36-dependent tumor angiogenesis, migration, invasion and sphere formation through the CD36/Stat3/Nf-Κb signaling axis [30,31]. Quercetin suppresses breast CSCs through its inhibition of the PI3K/Akt/mTOR signaling pathway [32]. Despite numerous studies, there are no studies on tangeretin-induced antiproliferative and anti-CSC effects. Our results

Discussion
It has been postulated that high consumption of fruits can prevent more than 20% of all cancer cases [20]. This preventive effect is predominantly mediated by phytochemicals interacting with specific target proteins that play important roles in cancer [21][22][23]. Citrus, one of the most important food sources of phytochemicals with health benefits, has many biological properties and controls key pathways involved in pathologies such as cancer [24][25][26]. The combination CD44 + /CD24 − has emerged as the most important marker for BCSC isolation, and the CD24 population of MDA-MB-231 is low [27].
First, we purified a CSC inhibitor from citrus. Assay-based fractionation and several chromatographic methods isolated one compound from a citrus powder, tangeretin. Tangeretin is the major flavonoid of citrus. It also has antioxidant, anti-inflammatory and anticancer properties [17]. Tangeretin modestly induces apoptosis in bladder cancer cells through mitochondrial dysfunction [28]. Tangeretin and nobiletin induces G1 cell cycle arrest but not apoptosis in breast and colon cancers [29]. Nobiletin inhibits CD36-dependent tumor angiogenesis, migration, invasion and sphere formation through the CD36/Stat3/Nf-Kb signaling axis [30,31]. Quercetin suppresses breast CSCs through its inhibition of the PI3K/Akt/mTOR signaling pathway [32]. Despite numerous studies, there are no studies on tangeretin-induced antiproliferative and anti-CSC effects. Our results showed that tangeretin suppresses the proliferation of BCSCs. Tangeretin inhibited mammosphere formation in breast cancer cells ( Figure 3) and modestly induced apoptosis in CSCs (Figure 4).
Our data showed the reduction of CD44 + /CD24 − subpopulation, mammosphere formation, colony formation and tumor formation. It is known that tangeretin did not inflict damaging effects sufficient to result in a reduced capacity to survive and proliferate. However, inhibition of the growth of breast cancers without inducing cancer cell death may be advantageous in treating breast tumors. Breast cancer cells resumed growth similar to untreated control within a day of tangeretin removal [29]. The Stat3 protein is essential for the maintenance of BCSCs [4]. The acetylated derivative of tangeretin (5-AcTMF) had anticancer effects on human glioblastoma multiforme cells through blockade of Stat3 signaling [33]. Extracellular IL-8 protein is a factor for BCSCs formation [13]. We investigated Stat3 signaling under tangeretin treatment. Tangeretin suppressed the total protein levels of Stat3 and pStat3. The nuclear protein levels of Stat3 and pStat3 were also decreased by tangeretin. We assessed Sox2 transcript levels in BCSCs under tangeretin treatment and confirmed that the Sox2 mRNA level was decreased in the treated samples. In addition, the protein level of Sox2 was decreased in treated cells compared with control cells (Figure 7). Finally, tangeretin had an inhibitory effect on tumor growth in a breast cancer xenograft model. The tangeretin-treated group had a smaller tumor size and lower tumor weight than the control group.
In our study, tangeretin inhibited BCSC formation and targeted BCSCs by inhibiting the Stat3/Sox2 signaling pathway. Tangeretin is a possible therapeutic agent for breast cancer and BCSCs.

Reagents
Silica gel 60 and TLC plates were purchased from MERK (Darmstadt, Germany) and Sephadex LH-20 was obtained from Pharmacia (Uppsala, Sweden). High-pressure liquid chromatography was performed with a Shimadzu application system (Shimadzu, Kyoto, Japan). Cell viability was measured using the EZ-Cytox Cell Viability Assay Kit (DoGenBio, Seoul, Korea). Tangeretin was obtained from ChemFaces Co. (Hubei, China).

Plant Materials
Citrus peel was collected from Jeju Island, South Korea. The citrus peel was freeze-dried, and the dried citrus was ground. The samples (No. 2017_030) were deposited in the Department of Biomaterials, Jeju National University, JeJu-Si, South Korea.

Extraction and Isolation of an Inhibitor
Citrus powder was extracted with 100% methanol. The bioassay-based isolation protocol is summarized in Figure 1A. The methanol extracts were mixed with water, and the methanol was evaporated. The water-suspended samples were extracted with equal volumes of ethyl acetate. The ethyl acetate-concentrated part was loaded onto a silica gel column (3 × 35 cm) and fractionated with solvent (chloroform-methanol, 20:1) ( Figure S1). The three parts were divided and assayed by evaluating mammosphere formation. The #2 part potentially inhibited mammosphere formation. The #2 part was loaded onto a Sephadex LH-20 open column (2.5 × 30 cm) and eluted in three fractions ( Figure S2). The three parts were obtained and assayed by evaluating mammosphere formation. Part # 3 showed inhibition of mammosphere formation. Part #3 was isolated using preparatory TLC (glass plate; 20 × 20 cm) and developed in a TLC glass chamber. Individual bands were separated from the silica gel plate. Each fraction was assayed by evaluating mammosphere formation ( Figure S3). The #1 fraction was loaded onto a Shimadzu HPLC instrument (Shimadzu, Tokyo, Japan). HPLC used an ODS 10 × 250 mm C18 column (flow rate; 3 mL/min). The mobile phase was water and acetonitrile. For elution, the acetonitrile proportion was initially set at 20%, increased to 80% at 20 min and finally increased to 100% at 40 min ( Figure S4).

Structural Analysis of the Purified Sample
The chemical structures of the isolated compounds were determined by mass and nuclear magnetic resonance (NMR) measurements. The molecular weight was established as 372 by ESI-mass spectrometry, which showed a quasimolecular ion peak at m/z 373.3 [M + H] + in the positive mode ( Figure S9). The 1 H NMR spectrum in CDCl 3 exhibited signals due to four aromatic methine protons at δ 7.86 (2H) and 7.01 (2H), which are attributable to a 1,4-disubstituted benzene, one aromatic singlet methine at δ 6.59, and five methoxy groups at δ 4.09, 4.01, 3.94, 3.93 and 3.87. In the 13  All proton-bearing carbons were assigned by the HMQC spectrum, and the 1 H-1 H COSY spectrum established a partial structure, 1,4-disubstituted benzene ( Figures S6, S7 and S9). The chemical structure was determined to be from the HMBC spectrum, which exhibited long-range correlations from the methine proton at δ 6.59 to the carbons at δ 177.3, 161.2, 123.8 and 114.8, and from the methine protons at δ 7.86 to the carbon at δ 161.2, implying that this compound was a polymethoxylated flavone ( Figures S8 and S9). Finally, long-range correlations from the five methoxy proton peaks to the oxygenated sp 2 quaternary carbons established the structure of this compound as that of tangeretin ( Figure 2).

Colony Formation Assay
MDA-MB-231 and MCF-7 cells (2 × 10 3 cells/well) were seeded in a 6-well plate, treated with different concentrations of tangeretin in DMEM and maintained for 7 days at 37 • C in a 5% CO 2 incubator. The grown colonies were washed with 1× PBS three times, fixed for 10 min using 3.7% formaldehyde, treated for 20 min and stained for 20 min with 0.05% crystal violet.

Wound-Healing Assay
MDA-MB-231 cells were seeded in a 6-well plate at 2 × 10 6 cells/well. A scratch was made by using a microtip after the cells had grown into a monolayer. After the cells were washed two times with 1× PBS, the cancer cells were cultured with tangeretin in fresh DMEM for 24 h. Photomicrographs of the wounded areas were acquired using a light microscope [36].

Flow Cytometry Analysis
After incubating with tangeretin for 24 h, cancer cells were dissociated using 1× trypsin/EDTA. We used a previously described method [36]. In total, 1 × 10 6 cells were cultured with anti-CD44-FITC and anti-CD24-PE antibodies (BD, San Jose, CA, USA) on ice for 30 min. The cancer cells were centrifuged and washed two times with 1× FACS buffer and analyzed on an Accuri C6 (BD, San Jose, CA, USA).

Gene Expression Analysis
Total RNA was extracted from cancer cells and purified, and real-time RT-qPCR was performed using a real-time One-step RT-qPCR kit (Enzynomics, Daejeon, Korea). We used a previously described method [37]. The specific primers are listed in Table S1.

Western Blot Analysis
Protein samples were extracted from mammospheres and cancer cells. After electrophoresis on a 12% SDS-PAGE gel, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Burlington, MA, USA). The membranes were blocked in Odyssey blocking buffer at room temperature for 1 h and then incubated overnight with primary antibodies. The antibodies were anti-phospho-Stat3, anti-lamin B (Cell Signaling, Danvers, MA, USA), anti-Stat3, anti-Sox2 and anti-β-actin (Santa Cruz Biotechnology, Dallas, TA, USA). After membranes were washed three times using Tris-buffered saline/Tween 20, all membranes were incubated with IRDye 680RD-and IRDye 800W-labeled secondary antibodies for 1 h at room temperature, and the signal images were determined with an Odyssey CLx (LI-COR, Lincoln, NE, USA).

Electrophoretic Mobility Shift Assays (EMSAs)
Nuclear proteins were prepared as described previously [38]. EMSAs were performed with an Odyssey Infrared EMSA kit (LI-COR, Lincoln, NE, USA) according to the manufacturer's instructions. IRDye 700-labeled forward and reverse strands of the Stat3 oligonucleotide were annealed. The IRDye 700-labeled Stat3 oligonucleotide was incubated with nuclear extracts in a final volume of 20 µL at room temperature for 20 min. The samples were electrophoresed on a 6% polyacrylamide nondenaturing gel, and EMSA data were visualized on an ODYSSEY CLx system (LI-COR, Lincoln, NE, USA).

Xenograft Transplantation
Twelve female nude mice were injected with two million MDA-MB-231 cells with or without an additional tangeretin (2.5 mg/kg) injection. Tumor volume was estimated for 35 days using the formula (width 2 × length)/2. The mouse experiments were performed as described previously [39]. Animal care and animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (JNU-IACUC; Approval Number 2017-003) of Jeju National University. Female nude mice (4 weeks old) were purchased from OrientBio (Daejeon, South Korea) and cultured in mouse facilities for 1 week.

SiRNA of Stat3
To examine the inhibitory function of Stat3 on mammosphere formation, breast cancer cells were transfected with siRNAs targeting human Stat3 (Bioneer, Daejeon, South Korea). The Stat3 siRNAs (NM_1145658) and a scrambled siRNA were purchased from Bioneer (Daejeon Cor., South Korea). For siRNA transfection, cancer cells were cultured and transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The protein levels of Stat3 were determined via immunoblot analysis.

Statistical Analysis
All data were analyzed with GraphPad Prism 7.0 software (GraphPad Prism, Inc., San Diego, CA, USA). All data from three independent experiments are reported as the mean ± standard deviation (SD). Data were analyzed using one-way ANOVA. A p-value of less than 0.05 was considered significant.

Conclusions
A CSC-inhibiting compound from citrus extracts was purified using silica gel, gel filtration, TLC, and HPLC. The compound was identified as tangeretin. Tangeretin inhibited cell proliferation, CSC formation and tumor growth, and modestly induced apoptosis in CSCs. The size of the CSC subpopulation (CD44 + /CD24 − ) was reduced by tangeretin. Tangeretin reduced the total level and phosphorylated nuclear level of Stat3 Tangeretin decreased the transcript levels of Sox2 and Sox2 protein in mammospheres. Our results in this study show that tangeretin inhibits the Stat3/Sox2 signaling pathway and induces CSC death, indicating that tangeretin may be a potential natural compound targeting breast cancer.
Supplementary Materials: The following are available online. Specific primer sequences for real-time RT-PCR are described in Table S1. Isolation and structure analysis of CSC inhibitor are described at Figures S1-S9.