Prevention of breast cancer by dietary polyphenols—role of cancer stem cells

Abstract

Breast cancer is a common malignancy with poor prognosis. Cancer cells are heterogeneous and cancer stem cells (CSCs) are primarily responsible for tumor relapse, treatment-resistance and metastasis, so for breast cancer stem cells (BCSCs). Diets are known to be associated with carcinogenesis. Food-derived polyphenols are able to attenuate the formation and virulence of BCSCs, implying that these compounds and their analogs might be promising agents for preventing breast cancer. In the present review, we summarized the origin and surface markers of BCSCs and possible mechanisms responsible for the inhibitory effects of polyphenols on BCSCs. The suppressive effects of common dietary polyphenols against BCSCs, such as curcumin, epigallocatechin gallate (EGCG) and related polyphenolic compounds were further discussed.

Introduction

Breast cancer is one of the most common malignancies and a primary reason for cancer-related death among women worldwide, and the proportion of breast cancer in newly confirmed cancer cases for women is 29% in the United States (Burnett et al. 2017). Similar to other types of cancers, metastasis, recurrence and drug resistance are leading barriers for breast cancer treatment. Conventional therapy of breast cancer includes chemotherapy (e.g. paclitaxel, docetaxel, adriamycin, cyclophosphamide), radiotherapy and surgery, which show poor prognosis and suffer from long-term side effects. Thus numerous efforts have been made to address these problems.

CSCs are just a minor proportion of cells in a tumor, and accumulating studies have confirmed that CSCs mainly account for tumor metastasis, regrowth and drug resistance (Dean 2009; Yin et al. 2011; Flemming 2015; Iqbal et al. 2016; Liang 2016; Su et al. 2016). Consistently, BCSCs are main reasons of the drug resistance and relapse of breast cancers. Therefore, targeting CSCs combined with differentiated cancer cells has become a promising therapeutic measure to prevent cancers. However, available chemotherapeutic drugs focus on killing differentiated cancer cells, which account for the majority of cells in tumors, rather than the fairly quiescent CSCs, and drugs targeting CSCs remain to be developed (Boman and Wicha 2008). A wide range of evidences show that fruits and vegetables that are rich in various bioactive phytochemicals such as polyphenols decrease the cancer incidence and relapse risk (Demark-Wahnefried et al. 2008; Lanou and Svenson 2010). Interestingly, close to half of currently approved anticancer drugs are either natural phytochemicals (isolated from plants, seeds and so on) or their derivatives (Nobili et al. 2009; Newman and Cragg 2012). Moreover, dietary polyphenols possess prominent effects on inhibiting tumorigenesis and potentiating drug sensitivity of CSCs (Vanden Berghe 2012).

Herein, we focus on discussing the origins and surface markers of BCSCs and mechanisms by which the common dietary polyphenols (curcumin, EGCG, resveratol, quercetin, genistein, and 6-shogaol), and their analogs or nano-formula target BCSCs.

Origins of CSCs/BCSCs

CSCs share the similar special surface tags, self-renewal abilities along with regulatory signaling with normal stem cells. Thus CSCs could spring from the malignant conversion of dormant normal stem cells that have accumulated oncogenic mutations over time, and this increasing mutational burden plays a vital role in tumorigenesis and development of tumor and disrupting proper cell differentiation (Dean, Fojo, and Bates 2005; Yang, Xu, et al. 2017). Besides genetic mutations, other factors such as epithelial mesenchymal transition (EMT) program and epigenetic changes (abnormal methylation, histone modification) may also involve in the transformation of stem cells into CSCs (Garg 2017; Wainwright and Scaffidi 2017).

Another hypothesis for the origin of CSCs is “misplacement of somatic stem cell theory”. It assumes that a handful of embryonal stem cell-like cells exist in the blood or other tissues. If they move at an inappropriate time and/or displaced (suffer from damaging environmental factors), they would turn into CSCs (Justyna Gil, Pesz, and Siadek 2008). Similar theory proposes that stem cell misplaces in the stroma owing to basement membrane lesion, and these misplaced stem cells are the root sources of the invasive tumors (Wang, Li, et al. 2013).

Besides, traditional cancer therapy also induces no-CSCs into CSC-like cells in various cancers such as breast cancer, and the drug-induced CSC-like properties (therapeutic resistance) will lose once chemotherapy is removed, implying that this conversion is reversible and transient (Goldman et al. 2015; Doherty et al. 2016).

BCSCs markers

In order to study the biology of BCSCs, accurately identifying and isolating BCSCs are necessary. Thus far, a series of surface markers have been used for flow cytometry separation of BCSCs. The most widely used surface markers are cluster of differentiation (CD) 44⁺, CD24⁻, aldehyde dehydrogenase 1 (ALDH1)⁺ and CD133⁺.

CD44 is a multifunctional transmembrane glycoprotein, which binds to hyaluronic acid and components of extracellular matrix (ECM), and is implicated in adhesion between cells and interactions among cells and matrices (Naab, Ricks-Santi, and Khan 2016). CD44 is related to metastasis, cancer invasion and multidrug resistance by interacting with P-glycoprotein (Miletti-Gonzalez et al. 2005; Pokharel et al. 2016. CD24 is a glycosylphosphatidylinositol-anchored membrane protein that is commonly overexpressed in human cancer cells (Ma et al. 2015), which is related to malignancy, chemoresistance, and metastasis of breast cancer cells (Kwon et al. 2015; Ma et al. 2015; Deng et al. 2017). CD44⁺CD24^−/low is a set of markers firstly used for identifying BCSCs, which has been widely used and is reliable to isolate cancer cells with stem cell-like characteristics, highly connected with the malignance of breast cancer (Ma et al. 2014; Wang, Wang, et al. 2017). The stemness of cells expressing CD44⁺CD24^−/low was proven by tumorigenicity tests (tumorigenesis in xenotransplanted mice), colony/mammosphere formation, migration, and invasion assays (Elizabeth Louie et al. 2010; Li, Ma, et al. 2017). However, some breast cancer cell lines do not express CD44⁺/CD24⁻ (Li, Ma, et al. 2017), implying that additional BCSCs markers should be established and verified.

ALDH1 is another commonly used marker of identifying BCSCs population. ALDH1 is able to convert retinol into retinoic acid, acting as an important detoxifying protease, and suppression of ALDH1 expression decreases the stemness of breast cancer cell lines and BCSCs (Kim et al. 2013). Moreover increased ALDH1 expression is commonly found in malignant breast stem/progenitor cells, which has a positive correlation with poor prognosis (Ginestier et al. 2007). Thus, ALDH1+ maybe a more helpful indicator for forecasting breast cancer metastasis in comparison with CD44⁺/CD24⁻ (Zhong et al. 2014). Interestingly, 1500 ALDH + cells without the CD44⁺CD24^−/low phenotype were capable of initiating tumors in immunodeficient mouse, while CD44⁺CD24^−/low ALDH – cells did not, suggesting that ALDH1 is an more effective marker to characterize BCSCs compared with CD44⁺CD24^−/low (Ginestier et al. 2007; Liu et al. 2014). In addition, a tiny amount of ALDH1-positive cells also express CD44⁺CD24^−/low marker, which were capable of producing tumors in vivo more frequently (Ginestier et al. 2007; Liu et al. 2014). After examining their ability of self-renewal and tumorigenesis, high CD44/CD24 ratio combined with ALDH1+ were proposed to be alternative markers which were able to identify BCSCs more accurately (Li, Ma, et al. 2017).

CD133, a transmembrane glycoprotein, is also an important surface marker of BCSCs. It was conceived as a specific marker of hematopoietic stem cells at first (Yin et al. 1997; Tume et al. 2016). However, growing evidences show that CD133 is also expressed or hyper-expressed in many cancers including breast cancer (Liu et al. 2013; Jang et al. 2017). Moreover, CD133⁺ cells in breast carcinoma are often characterized by higher drug tolerance, invasive and metastatic capacity, higher tumor-initiating and self-renewal ability (Liu et al. 2013; Tume et al. 2016; Sansone et al. 2016), suggesting that CD133 is closely related to stemness of BCSCs. In addition, CD133 is up-regulated and facilitates multidrug resistance by enhancing phosphatidylinositol-3-kinase (PI3K)-Akt signaling (Xi et al. 2016). Also, CD133 is involved in tumor metastasis via promoting epithelial-mesenchymal transition (Ding et al. 2014). In terms of the clinical manifestation, people suffering CD133⁺ mammary neoplasm usually have higher mortalities, poorer overall survival and disease-free survival (Li, Yin, et al. 2017), and nuclear translocation of CD133 may be responsible for poor prognosis of patients with triple-negative breast cancer (Cantile et al. 2013). Thus, CD133⁺ might be a useful surface marker for identifying BCSCs as well as a helpful indicator for predicting prognosis.

Other markers such as Protein C receptor (PROCR)⁺, MUC1⁺/CD24⁺, CD44⁺/Vimentin⁺, CD44⁺/Osteonectin⁺, CD24⁺/CK18⁺, CD24⁺/GATA3⁺, PROCR⁺/ESA⁺, and proteosome^low have also been proposed (Engelmann, Shen, and Finn 2008; Vlashi et al. 2009; Park et al. 2010; Pece et al. 2010; Wang, Cai, et al. 2015). Nevertheless, studies on BCSCs markers are still ongoing owing to the fact that BCSCs markers often vary depending on diverse subtypes of breast cancer cells, histologic stages and heterogeneity within tumors (Tsang et al. 2012). In addition, clinical relevance of these identified BCSCs biomarkers remains controversial and requires further studies (Lu et al. 2011).

Signaling pathways regulating BCSCs self-renewal

CSCs share the characteristics with other stem cells such as self-renewal and differentiation. By symmetric or asymmetric cell division, CSCs can repopulate themselves and generate clonal daughter cells (Li et al. 2011). Several key pathways are implicated in regulating BCSCs self-renewal including Wnt/β-catenin, Notch, Hedgehog and Transforming growth factor-β (TGF-β) pathways.

Wnt/β-catenin signaling

Wnt/β-catenin signaling, a highly conserved pathway, plays a vital role in modulating cell propagation and differentiation (Kahn 2014). However, this pathway is commonly deregulated and aberrantly activated during carcinogenesis (Anastas and Moon 2013), which promotes clonal expansion. In breast cancer, this aberrant signaling facilitates selfrenewal and associated properties (metastasis, multi-drug resistance, invasiveness) of BCSCs. Blocking Wnt/β-catenin pathway by knocking down miR-142, a potent effector for activating this signaling, decreased tumor-initiating ability of BCSCs and mammosphere formation by BCSCs (Isobe et al. 2014). Similarly, stemness, self-renewal and proportion of BCSCs were significantly reduced when treated with pyrvinium pamoate, a suppressor of Wnt pathway (Xu et al. 2016). In addition, inhibiting Wnt/β-catenin by Let-7 could make BCSCs more chemosensitive to tamoxifen accompanied by reduced self-renewal and tumorigenic ability (Sun et al. 2018). Thus targeting Wnt/β-catenin signaling would be an effective strategy to suppress BCSCs.

Notch signaling

Notch signaling is associated with cell differentiation, propagation and is indispensible in the development of human organs (Miele 2006). In tumors, the signaling is frequently hyperactive and closely related to self-renewal of CSCs including BCSCs (Bu et al. 2013; Pal et al. 2017). BCSCs bearing ALDH⁺ express a higher level of Notch-1, a receptor of Notch signaling, which also promotes EMT. However, the promoting effect is inhibited when Notch-1 is suppressed by Psoralidin (Pal et al. 2017). Similarly, both receptors (Notch-1,-2) and ligands of Notch signal are highly expressed in renal CSCs, meanwhile, enforced expression of Notch-1 makes renal CSCs more chemoresistant and upregulatse self-renewal ability, suggesting Notch signaling facilitates self-renewal and stemness of CSCs (Xiao et al. 2017). In addition, disrupting Notch signaling by γ-secretase suppressor in mammary tumor could markedly decrease the number of BCSCs (Mamaeva et al. 2016). In short, Notch signaling is also an effective target for inhibiting BCSCs.

Hedgehog signaling

Hedgehog signaling is critical for morphogenesis during embryonic development (Gritli-Linde et al. 2002; Monkkon and Lewis 2017). However, deregulation of this pathway is associated with carcinogenesis, and the signaling is abnormal and hyperactive in nearly 1/4 human cancers, including mammary cancer (Monkkon and Lewis 2017). Hedgehog signaling is also closely related to self-renewal ability, oncogenicity and stemness of CSCs/BCSCs. Consistently, members in hedgehog signaling were highly expressed in BCSCs (cells bearing CD24⁺/CD24⁻/Lin⁻), and the highly activated hedgehog signaling promoted self-renewal of BCSCs, which was illustrated by increased size and number of mammosphere in vitro and tumor in vivo (Liu et al. 2006). In addition, activation of hedgehog signaling by long non-coding RNAs promoted BCSCs generation, increased the expression of stemness markers (SOX2 and OCT4) and tumorigenicity of BCSCs (Zhou et al. 2016). On the other hand, when hedgehog signaling is suppressed by nitidine chloride (Sun et al. 2016), genistein (Fan et al. 2013), salinomycin (Lu et al. 2015), properties and numbers of BCSCs are also inhibited. Therefore, regulating the abnormal hedgehog signaling is a promising way to fight against BCSCs.

Transforming growth factor-β (TGF-β) signaling

Transforming growth factor-β (TGF-β) signaling is indispensable for proper development of tissues and organs. In breast cancer, this signaling is associated with metastasis (Zhao et al. 2018), migration and invasion (Zhao et al. 2018). TGF-β signaling is frequently activated in BCSCs and contributes to maintaining quiescent state of BCSCs (Tang et al. 2015). In addition, TGF-β signaling was responsible for the augment of EMT and self-renewal of BCSCs induced by paclitaxel, which could be abolished by suppressing TGF-β signaling (Park et al. 2015). Studies focusing on the effects of this signaling on BCSCs are limited, though many studies about promoting effects of the signaling on CSCs self-renewal have been reported in other cancers (Fu, Li, and Hao 2017).

Epigenetic mechanisms associated with BCSCs self-renewal and differentiation

Though analogous to normal stem cells, these pathways mentioned above modulating self-renewal and differentiation are commonly aberrant in CSCs, due to epigenetic and/or genetic alterations. In normal stem cells, promoters of genes regulating cell differentiation usually contain the bivalent marks: H3K4me3 for gene activation and H3K27me3 for gene suppression (Bernstein et al. 2006). When normal stem cells differentiate, the H3K27me3 marker of the lineage specific genes is removed while the H3K4me3 marker of the target genes is retained, initiating a specific differentiation program (Boyer et al. 2006; Sauvageau and Sauvageau 2008). While in CSCs, the genes regulating differentiation are normally suppressed by epigenetic suppression, including aberrant DNA methylation and H3K27 methylation catalyzed by enhancer of zeste homolog 2 (EZH2), the methyltransferase of the PRC2 complex, thereby blocking differentiation and permanently positioning cells to self-renewal (Easwaran et al. 2012). Indeed, EZH2 is commonly overexpressed in many types of cancers (glioblastoma cancer, acute myeloid leukemia, small cell lung cancer and others), which causes blockage of differentiation and is relevant to unfavorable prognosis (Suva et al. 2009; Tanaka et al. 2012; Sato et al. 2013). Meanwhile anti-cancer genes are usually inhibited by abnormal DNA methylation in both leukemia (Rahmani et al. 2018) and solid tumors (Lv et al. 2015; Liang and Weisenberger 2017).

Additionally, microRNAs (miRNAs) are endogenous and non-coding small RNAs with approximately 21–25 nucleotides in lengths. MiRNAs often function as a modulator of gene expression (Bartel 2004; Ha and Kim 2014). Consistently, miRNAs are also dysregulated in cancer cells via epigenetic mechanisms, rendering CSCs to undergo self-renewal instead of differentiation (Tuna, Machado, and Calin 2016). Restoring the expression of miR-34a, an important regulator of stem cell-like characteristics in human pancreatic CSCs, reduces CSCs characteristics including proliferation, self-renewal and EMT (Nalls et al. 2011). Based on their effects on CSCs, there are two categories of miRNAs: pluripotent miRNAs and pro-differentiation miRNAs. Pluripotent miRNAs such as miR-290 and miR-9 favor self-renewal and proliferation of stem cells but attenuate differentiation (Wang, Du, et al. 2013). Pro-differentiation miRNAs (let-7 and miR-470) facilitate differentiation (Wang, Du, et al. 2013). MiRNAs may target Wnt/β-catenin signaling, Notch signaling and other factors to regulate CSCs self-renewal and differentiation. For examples, miRNAs (miR-142 and miR-663) are able to suppress adenomatous poliposis coli (APC), which activate canonical Wnt signaling (Wang and Xu 2010; Isobe et al. 2014). In BCSCs, let-7 miRNAs are decreased, but upregulated during differentiation; it is up-regulation hinders BCSCs proliferation, mammosphere formation, neoplasm growth as well as metastasis in NOD/SCID mice, showing that let-7 suppresses BCSCs self-renewal (Yu et al. 2007). In human BCSCs, the expression of miR-146 is increased (Shimono et al. 2009). The Wnt receptors Frizzled-6 expression levels and low lipoprotein receptor related protein 6 are augmented through downregulation of Zinc Ring finger 3 by miR-146, thereby activating Wnt signaling (Deng et al. 2015). MiRNAs also regulate Notch signaling pathway. MiR-146a, which is upregulated in human BCSCs, activates Notch signaling pathway by targeting Numb, an inhibitor of this pathway (Kuang et al. 2009; Forloni et al. 2014). MiRNA-34a which is downregulated in human breast cancer, is capable of suppressing Notch signaling pathway, thereby inhibiting self-renewal ability, metastatic potential as well as drug resistance of BCSCs (Kang et al. 2015). In addition, Hippo signaling pathway is modulated via miRNAs in breast cancer cells as well. MiRNA-125a indirectly modulates tafazzin, which is a typical effector of Hippo signaling, by targeting leukemia inhibitory factor receptor (LIFR). The miR-125a-LIFR axis could effectively regulate the homeostasis of nonmalignant and malignant breast epithelial stem cells via altering the Hippo signaling (Nandy et al. 2015). Transcription factors required for maintaining pluripotency including Oct4, Sox2, and Nanog could be targeted by miRNAs such as miR-296, thus tilting the balance between self-renewal and differentiation (Tay et al. 2008). Also, miR-590-5p down-regulates Sox2 expression, suppressing BCSCs property (mammosphere formation, tumorigenicity) (Zhou et al. 2017).

Food-derived polyphenols targeting BCSCs

Accumulating studies confirm that dietary polyphenols are promising agents to fight against BCSCs. The inhibitory effects of food-derived polyphenols on BCSCs include decreasing percentage of ALDH + cells, CD44⁺/CD24⁻ cells or cells bearing other BCSCs markers, and reducing mammosphere formation and tumorigenesis in vivo (Dontu et al. 2003). These reported roles of food-derived polyphenols targeting BCSCs by various mechanisms including blocking pathways enhancing BCSCs self-renewal (Wnt/β-catenin and Hedgehog signaling), altering BCSCs niches (Signal transducer and activator of transcription3(STAT3)-mediated inflammatory signaling, changing growth factors such as hypoxia-inducible factor (HIF) and vascular endothelial growth factor (VEGF), suppressing ATP-binding cassette (ABC) drug transporters (ABCG2, ABCC1 and P-glycoprotein), regulating anti-cancer miRNAs (miR-16, -200c, miR148a), downregulating heat shock protein 27 and heat shock 70kDa protein 5, decreasing estrogen receptor-α36 (ER-α36),and activating intracellular signaling pathways including RAS/ERK and PI3K/AKT pathways, and suppressing adipogenesis and lipid metabolism (Figure 1). The inhibitory effects of dietary polyphenols on BCSCs and potential molecular targets are summarized below (Table 1).

Figure 1.

Mechanisms contributing to anti-BCSCSs effects of dietary polyphenols.

Note: represents inhibition; represents promotion.

Table 1.

Anti-BCSCs effects of food-derived polyphenols.

Compound	Polyphenol concentration	Effects on CSCs	Molecular targets	References
	15 μM	inhibit migration, invasion and EMT	a↑E-cadherin/β-catenin complex; ↓nuclear β-catenin; b↑epithelial markers (E-cadherin, cytokeratin-18 and -19); ↓β-catenin transcriptional targets (cyclin D1, c-Myc, and Slug); ↓EMT-promoting factors (vimentin, MMP-2 and MMP-9)	(Mukherjee et al., 2014)
	10–40 μM 100 mg/kg/d	inhibit mammosphere formation; induce autophagy; inhibit tumor growth in vivo; inhibit ALDH + BCSCs tumorigenicity in vivo	↓β-catenin and cyclin D1 ↑LC3-II, Beclin1 and Atg7	(Fu et al. 2014)
	50 μM 25 mg/kg/d	reduce invasive ability inhibit tumor growth	↑miR-16, -141, -143, and -200c ↑Argonaute2 (Ago2) ↓Zeb1 ↑E-cadherin	(Hagiwara et al. 2012)
	50-100 μM 22.4 kg/body weight	reduce mammosphere formation; reduce breast cancer cells viability; induce BCSCs apoptosis; inhibit tumor growth in vivo suppress lipid syntheses	↓fatty acid synthase ↑pro-apoptotic genes (DAPK2, BNIP3)	(Pandey et al. 2011)
	50-100 μM	decrease BCSCs proliferation; induce BCSCs apoptosis; down-regulate lipogenesis; inhibit mammosphere formation;	↓SREBP1 ↑pro-apoptotic genes (DAPK2, BNIP3) ↓SREBP1 downstream genes (ACLY, ACC1 and FAS)	(Pandey et al. 2013)
	25,50 μM	attenuate migrational and invasive ability decrease proportion of BCSCs inhibit mammosphere formation	↓Cyclin D1and c-Myc ↓MMP-2 and -9 ↓Sox-2 and CD44 ↓p-Akt and p-Stat3	(Suh, Kim, and Surh 2018)
Resveratrol analog HS-1793	0.6–2.5 μM 1.5 mg/kg	reduce invasive ability enhance radiosensitivity and apoptosis induced by radiotherapy	↓Oct4, KLF4 and Sox2 protein ↓VEGF ↓HIF-1α	(Choi et al. 2016)
	12.5–50 μM	inhibit size and number of mammo-spheres; decrease population of ALDH + cells; suppress EMT and migration;	↓Hsp27 ↓N-cadherin and twist ↑E-cadherin	(Li et al. 2011)
	12.5-50 μM	repress BCSCs inhibit tumor growth inhibit VM capability	↓Hsp27	(Lee et al. 2014)
	20 μM	inhibition migration and population of ALDH + cells	↓Hsp27	(Lee et al. 2012)
	0.7 μM	decrease population of CD44+/CD24− cells improve chemosensitivity	↓nuclear Y-box binding protein 1 ↓P-glycoprotein	(Li, Zhao, Wang, Yuan, et al. 2018)
	50 μM	decrease population of CD44+/CD24− cells inhibit mammosphere formation inhibit tumor growth	↓PI3K/Akt/mTOR	(Li, Zhou, Wang, Liu, et al. 2018)
	10-40 μM	inhibit mammosphere formation; inhibit CD44+/CD24− cells	↓ER-α36 ↓EGFR	(Pan et al. 2016)
	20–160 μg/mL 16.5 mg/kg	induce apoptosis of the ALDH + SUM-149 cells; inhibit mammosphere formation; inhibit tumor growth in xenograft mouse model	↓lymphangiogenesis-promoting genes ↓VEGF-D	(Mineva et al. 2013)
EGCG analogs	10,20 μM	inhibit mammosphere formation; inhibition CD44+ high/CD24− low cells		(Chen et al. 2012)
	0-30 μM 0–50 mg/kg	inhibit mammosphere formation; decrease ALDH + levels in xenograft tumors; decrease CD44+CD24− population;	↓Smo ↓Gli1	(Fan et al. 2013)
	2 μM, and 40 nM	induce differentiation of BCSCs; decrease CD44+/CD24−/ESA+ cells;	↑differentiated cell markers (E-cadherin, α-smooth muscle actin and Claudin-1 genes) ↓stem cell markers (Fibronectin and Snail); ↑ phospho-Akt308/473 ↑ phospho-ERK1/2 ↑Amphiregulin ↑phospho-β-catenin	(Liu et al. 2016)
	2 μM, and 40 nM	inhibit adipogenesis inhibit mammosphere formation;	↑mammary tumor suppressor (PTEN and E-cadherin) ↓Pparγ ↓fatty acid synthase ↑estrogen receptor b	(Montales et al. 2013)
	2 μM, and 40 nM 1–40 μM	inhibit mammosphere formation inhibit primary and secondary mammo-sphere formation; decrease CD44+/CD24− cells proportion;	↑PTEN ↓Cleaved Notch1 ↓Hes1 and Cyclin D1	(Montales et al. 2012) (Ray, Vasudevan, and Sengupta 2015)
	5-25 μM	inhibit mammosphere formation; decrease CD44 expression of BCSCs; increase chemosensitivity of BCSCs; induce cell necrosis phenomena;	↓hedgehog ↑p-β-catenin ↓CD44 ↓c-Myc ↓cyclin D1	(Wu, Hong, et al. 2015)
	–	inhibit metastasis; inhibit tumor growth; inhibit mammosphere formation; suppress cell migration and mobility	↓PI3K/AKT signaling ↓MAPK/ERK signaling ↓STAT3 signaling	(Vuong et al. 2016)
	10 μM 20 mg/kg Body Weight	decrease mesenchymal features; inhibit mammosphere formation; inhibit tumor growth;	↓TGFβ-SMADs Signal ↑miR-148a ↓DNMT1 and DNMT3a	(Jiang et al. 2016)
	25, 50 μM 50 mg/kg/d	inhibit self-renewal and multi-differentiation; reduce BCSCs proportion; inhibit tumor growth;	↓heat shock 70kDa protein 5 ↓β-catenin/ABCG2 signaling	(Wang Wang, et al. 2014)
	25, 50 μM 50 mg/kg/d	block self-renewal BCSCs; reduce BCSCs population; induce BCSCs G0/G1 phase arrest; inhibit tumor growth;	↓DNMT1 methyltransferase ↑WIF1 ↓β-catenin signaling	(Wang Wang, et al. 2015)
Curcumin + piperine	5,10 μM	inhibit mammosphere formation; reduce the percent of cells bearing ALDH1+	↓Wnt signaling	(Kakarala et al. 2010)
Curcumin + Mitomycin C	5–40 μM 100 mg/kg	nhibit mammosphere formation; reduce cells expressing CD44+CD24−/low; inhibit tumor growth in vivo	↓ABCG2 and ABCC1	(Zhou et al. 2015)
Curcumin + EGCG	10 μM, 10 μM	inhibit mammosphere formation; inhibit invasion; reduce CD44−expressing subpopulation	↓p-STAT3 ↓nuclear STAT3-NFkB	(Chung Seyung 2015)

^a↓Represents down-regulation;

^b↑Represents down-regulation.

Abbreviations: ACC1, acetyl CoA carboxylase 1; ALDH, aldehyde dehydrogenase; ACLY, ATP-citrate lyase; BCSCs, breast cancer stem cells; EMT, epithelial mesenchymal transition; EGFR, epidermal growth factor receptor; ER-α36, estrogen receptor-α36; FAS, fatty acid synthase; GSK3β, glycogen synthase kinase 3β; Hsp, Heat shock protein; MMP, metal matrix proteinase; PPARγ, Peroxisome proliferator-activated receptor γ; SREBP1, sterol regulatory element binding proteins1; VEGF, vascular endothelial growth factor.

Curcumin

Curcumin is a dietary polyphenol from turmeric with anticarcinogenic effects (Wang et al. 2016; Wang, Hang, et al. 2017. Curcumin is also an effective natural phytochemical inhibiting the proliferation of BCSCs together with suppressing the bulk breast cancer cells (Li and Zhang 2014). Curcumin can reduce CSCs in various cancers by inhibiting signaling pathways promoting self-renewal, particularly the Wnt/β-catenin signaling (Li and Zhang 2014; Wu, Guo, et al. 2015; Yang, Wang, et al. 2017; Zhu, Yang, et al. 2017). Curcumin inhibited BCSCs migration by inhibiting β-catenin nuclear translocation, attenuating the expression of β-catenin transcriptional targets including pro-EMT factors, along with the increased formation of E-cadherin/β-catenin complex and suppressed expression of epithelial markers (Mukherjee et al. 2014). In another study, treatment with curcumin and piperine separately or jointly repress mammosphere formation, reduce the percentage of ALDH1⁺ cells, and also block secondary/tertiary mammospheres formation by BCSCs. These effects are mediated by suppression of dysregulated Wnt signaling (Kakarala et al. 2010). Additionally, in BT-549 breast cancer cells, curcumin-load POCA4C6 micelles inhibited BCSCs by interruption of androgen receptor signaling and Wnt/β-catenin signaling (Chen, Li, et al. 2017).

Apart from regulating self-renewal pathway of BCSCs, curcumin also attenuates BCSCs by other mechanisms. Generally, BCSCs are resistant to chemotherapy, which is partially caused by ATP-binding cassette (ABC) drug transporters. Curcumin was able to inhibit ABCG2 and ABCC1, thereby decreasing anti-cancer efficiency of chemotherapeutic agents, and combined use of curcumin and Mitomycin C effectively eliminated CD44⁺CD24^−/low cells as well as inhibited BCSCs-derived tumor growth (Zhou et al. 2015). Likewise, MCF-7 cells co-treated by curcumin (10 μM) and EGCG (2 μM) were susceptible to apoptosis and sensitive to doxorubicin treatment (Wang, Chen, et al. 2014).

Resveratrol

Resveratrol is one of the most studied polyphenols with anti-cancer effects including suppressing CSCs (Sahin et al. 2016; Chen, Lien, et al. 2017; McCubrey et al. 2017). Resveratrol could preferentially attenuate BCSCs, which was illustrated by suppression of ALDH + cells proportion in SUM159 and MCF-7 cell lines and xenograft breast tumors. Resveratrol also inhibited mammospheres formation and tumorigenicity of BCSCs, and facilitate autophagy at the same time, primarily by means of disrupting Wnt/β-catenin pathway (Fu et al. 2014). In another study, MR-3, a natural methoxylated compound of this polyphenol, was found to suppress invasive ability of breast cancer cells (MCF-7), owing to the blockage of EMT progress by attenuating PI3K/AKT pathway and reducing levels of nuclear β-catenin (Tsai et al. 2013). Additionally, cancer cells commonly show up-regulated expression of adipogenic genes, which may promote energy supply to rapidly growing tumor cells through enhanced β-oxidation and also lipids needed for membrane synthesis (Kuhajda 2000). Hence, suppression of adipogenic gene expression inhibits tumorigenesis of BCSCs. Resveratrol induces apoptosis and inhibits mammosphere formation and xenograft tumor growth of BCSCs (CD24⁻/CD44⁺/ESA⁺), which is associated with suppression of lipogenesis (Pandey et al. 2011). Similarly, key lipogenic genes were upregulated in ductal carcinoma and its cancer stemlike cells, which was disrupted by resveratrol in a xenograft mice model (Pandey et al. 2013). Emerging evidences show that various miRNAs are potential molecule targets to inhibit BCSCs, such as suppression of brain metastasis of BCSCs by miR-7 (Okuda et al. 2013), increasing tumorigenicity of BCSCs by miR-221/222 (Li et al. 2016), inhibiting BCSCs stemness and drug resistance by miR-375 or miR-519d (Okuda et al. 2013; Li et al. 2016; Fu, Li, and Hao 2017; Shindo et al. 2017; Xie et al. 2017). Resveratrol could also promote the expression of anti-cancer miRNAs such as miR-16 and -200c, thus inhibiting BCSCs stemness. Moreover, resveratrol increased the level of Argonaute2, potentiating tumor-suppressive function of miRNAs (Hagiwara et al. 2012). Besides resveratrol is capable of repressing BCSCs through restoring BCSCs niches, including suppressing the secretion of HIF and VEGF, and proliferation of cancer-associated fibroblasts. HIF regulates oxygen homeostasis, which is implicated in drug resistance and the consolidation of cancer microenvironment (Warfel and El-Deiry 2014). HIF promotes VEGF that is indispensable for sustaining tumor growth (Auguste et al. 2005; Liu et al. 2012). HS-1793, a resveratrol analog, reduced the expression of HIF-1a and VEGF under hypoxic conditions in mouse breast cancer FM3A cells, and suppressed hypoxia-induced CSCs properties, including decreased invasive ability along with suppressed levels of stemness markers (Oct4, Klf4 and Sox2 proteins). Furthermore HS-1793 also promoted apoptosis initiated by ionizing radiation in hypoxic FM3A cells (Choi et al. 2016). Cancer-associated fibroblasts, the key component of tumor microenvironment or CSCs niche, contribute to malignancy of cancer cells and self renewal activity of CSCs (Adisetiyo et al. 2014). Resveratrol was able to abolish metastatic and invasive ability, inhibit mammosphere formation of breast cancer cells and decrease BCSCs proportion by suppressing cancer-associated fibroblasts (Suh, Kim, and Surh 2018).

Genistein

Genistein, a dietary isoflavone phytoestrogen mostly found in soybeans, shows promising antitumor effects (Banerjee et al. 2008). Daily consumption of genistein reduces the incidences of breast cancer among women in eastern countries (Iwasaki et al. 2008). Genistein reduces CSCs in mammary cancer (Liu et al. 2016), ovarian cancer (Ning et al. 2014) and gastric cancer (Yu et al. 2014). Genistein inhibited the number and size of mammospheres and decreased the percentage of cells bearing CD44⁺CD24⁻ in treated MCF-7 cells, and also lowered the proportion of ALDH + cells in xenograft tumors of genistein treated nude mice, which might be due to the suppression of Hedgehog-Gli1 pathway in vivo (Fan et al. 2013). Like resveratrol, genistein could also decrease mammary adipogenesis and upregulate phosphatase and tensin homolog (PTEN) and E-cadherin, both of which are mammary tumor suppressors, in female mice; genistein-treated adipocytes suppressed the formation of MCF-7 mammospheres (Montales et al. 2013). BCSCs are more multi-drug resistant than differentiated cancer cells, a feasible measure to suppress BCSCs is facilitate them differentiation. Genistein (2 μM or 40 nM) could induce the differentiation of BCSCs, confirmed by adverse morphological alteration of mammospheres and upregulated expression of differentiated cell markers in MCF-7 cells, which might attribute to the suppression of PI3K/Akt and MEK/ERK signaling by genistein (Liu et al. 2016). Additionally, genistein inhibited mammospheres formation by MCF-7 and MDA-MB-231 cells partially through suppressing PTEN/PI3K/Akt pathway (Montales et al. 2012).

Quercetin

Quercetin, a dietary flavonoid found in many plant foods, displays excellent antitumor activity (Kashyap et al. 2016). Recently, small heat shock proteins 27 (HSP27), a member of a heat shock proteins family, was found to be beneficial to maintain CSCs, and CSCs stemness can be inhibited by downregulation of HSP27 (Lu et al. 2016; Yasuda et al. 2017). Moreover, both Hsp27 and phosphorylated Hsp27 were upregulated in ALDH + BCSCs compared with ALDH-non-BCSCs. Quercetin could act as an inhibitor of Hsp27, thereby decreasing self-renewal of BCSCs and reducing the population of ALDH + cells (Li et al. 2011). Similarly, 20 μM of quercetin displayed a synergistic effect with a low dose of geldanamycin, a Hsp90 inhibitor, in suppression of migration and population of ALDH + BCSCs, further illustrating the synergistic effects of Hsp90 and Hsp27 suppression in inhibition of BCSCs tumorigenesis (Lee et al. 2012). In addition, vascularization of tumors was suppressed by quercetin through targeting epidermal growth factor (EGF)/Hsp27 signaling (Lee et al. 2014). Apart from targeting Hsp27, quercetin was able to effectively reduce number and stemness of BCSCs through blocking PI3K/Akt/mTOR, a pathway regulating self-renewal of CSCs (Li, Zhou, Wang, Liu, et al. 2018). Furthermore, P-glycoprotein pumping drugs out of cells, is also responsible for multidrug resistance of cancer cells, and nuclear translocation of Y-box binding protein 1 contributes to upregulation of P-glycoprotein and stemness of CSCs (Oda et al. 2003; Oda et al. 2007; To et al. 2010). Quercetin could block Y-box binding protein 1 nuclear translocation and down-regulate P-glycoprotein, thus improving chemosensitivity of breast cancer cells and inhibiting proportion of BCSCs (Li, Zhao, Wang, Yuan, et al. 2018).

EGCG

EGCG, the primary bioactive polyphenol in green tea, exhibits potent effects on fighting against cancers, and inhibits CSCs in prostate cancer, lung cancer and others (Fujiki et al. 2017). EGCG also attenuates BSCSs by various mechanisms. STAT3 mediates cytokine-induced signaling, which forms a feedforward circuit with IL6/Nuclear Factor-κB (NF-κB) axis, promoting the tumorigenesis of BCSCs (Iliopoulos, Hirsch, and Struhl 2009; Kise, Kinugasa-Katayama, and Takakura 2016). Hence, targeting STAT3 can inhibit stemness of BCSCs. Consistently, combined treatment of 10 μM EGCG and 10 μM curcumin decreased the formation of mammospheres and CD44-expressing subpopulation, suppressed cell invasiveness of MDAMB-231 and MCF7-HER2 cells. Moreover, Co-treatment with EGCG and curcumin blocked STAT3 phosphorylation, preventing STAT3-mediated inflammatory signaling and suppressing CD44 expression of BCSCs (Chung Seyung 2015). Targeting ER-α36 is also a potential way to eradicate BCSCs. ER-α36, a variant of ERα, is highly expressed in ER⁻ breast cancer cells (Zhang et al. 2011), which is critical to keep stemness of both ER⁺ and ER⁻ breast cancer cells (Kang et al. 2011; Deng et al. 2014). Accordingly, 20 μM EGCG could effectively decrease the expression of ER-β36, thereby inhibiting tumor sphere growth and reducing CD44⁺/CD24⁻ cells population of ER⁻ breast cancer cells (Pan et al. 2016). Additionally, EGCG can inhibit BCSCs by blocking lymphangiogenic factor and lymphangiogenesis. VEGF-D is a critical lymphangiogenic factor, which promotes lymphangiogenesis and lymph node metastasis of cancers (Stacker et al. 2014). Lymphatic vessels facilitate tumor metastasis (Li and Li 2015), and lymphatic metastasis is a major form of tumor metastasis. EGCG downregulated expression of VEGF-D which inhibited the lymphangiogenic potential, robustly inhibiting mammosphere formation and neoplasm growth in vivo (Mineva et al. 2013). Octa-acetate derivative of EGCG also displayed inhibiting effects on BCSCs self-renewal, confirmed by reduced mammospheres formation (Chen et al. 2012).

Besides, EGCG could dose-dependently repress Wnt signaling by targeting HMG-box protein 1, a suppressor of Wnt signaling, thus attenuating tumorigenicity and invasiveness of mammary cancer cells (Kim et al. 2006), which suggested that EGCG was capable of inhibiting self-renewal of BCSCs. In addition, EGCG is capable of suppressing CSCs self-renewal in other cancers (Zhu, Jiang, et al. 2017; Fujiki et al. 2018).

6-Shogaol

6-Shogaol, derived from dried ginger, has potent anticancer activity (Warin et al. 2014; Anwar et al. 2016). 6-Shogaol blocks Hedgehog and/or Wnt/β-catenin in BCSCs, which reduces their primary and secondary mammospheres formation and decreasing CD44⁺CD24^−/low cells in mammospheres. Moreover, downregulation of Notch signaling is also involved in anti-BCSCs effects of 6-shogaol (Ray, Vasudevan, and Sengupta 2015). These polyphenolic compounds disrupt hedgehog and Akt/GSK3β signaling pathway, thereby attenuating stemness of BCSCs (Wu, Hong, et al. 2015).

Other dietary polyphenols

Gallic acids, commonly found in fruits (pomegranate, mango and blueberry), have various biological activity (Kosuru et al. 2018). Fermented blueberry juice that is rich in Gallic acids could effectively attenuate self-renewal ability and migration potential of BCSCs by preventing IL6/STAT3 axis, a typical inflammatory signaling (Vuong et al. 2016). Glabridin, a polyphenol extracted from Glycyrrhiza glabra, is a promising anti-cancer agent. Recently, Jiang et al reported that glabridin (10 μM) could potently suppress BCSCs-like properties such as reducing mammospheres formation, ameliorating mesenchymal phenotype and inhibiting tumorigenesis in vivo. These restraining effect ascribed to up-regulation of miR-148a by glabridin, which further inhibited TGFμ-SMADs Signaling, a pathway associated with BCSCs-like characteristics (Jiang et al. 2016). Isoliquiritigenin, a natural polyphenol in licorice, exerts anti-tumor effects in lung cancer (Jung et al. 2014), bladder cancer (Moreno-Londono, Bello-Alvarez, and Pedraza-Chaverri 2017) and breast cancer (Peng et al. 2017). Isoliquiritigenin could make BCSCs more chemosensitive and inhibit BCSCs propagation and self-renewal by directly targeting heat shock 70 kDa protein 5, suppressing β-catenin/ABCG2 axis (Wang, Wang, et al. 2014). Besides isoliquiritigenin could also attenuate BCSCs by epigenetic regulation. Isoliquiritigenin effectively decreased BCSCs proportion, led to BCSCs arrest and blocked tumor growth in vivo by blocking DNA methyl-transferase1, which consequently led to up-regulation of Wnt inhibitory factor by means of promoter demethylation (Wang, Wang, et al. 2015).

Nano-technology to enhance polyphenol bioavailability against BCSCs

Although polyphenols mentioned above display excellent anti-cancer effects, the effectiveness of dietary polyphenols is limited due to their instability and lower bioavailability (Naksuriya et al. 2014). Thus, alternative delivery methods are needed, of which nano-drug delivery system is a good choice. These nano-durgs, also called nano-formulas, are often developed by coating hydrophobic or low bioavailable drugs with nanocarriers, and the size of which are often 10–100 nm, thus enhancing the solubility of hydrophobic chemotherapeutic agents, promoting drug permeation into neoplasm, lowering adverse effects and facilitating targeted delivery of therapeutic agents (Naksuriya et al. 2014; Fatima et al. 2016). Up to now, the reported nano-drugs made of dietary ployphenols targeting BCSCs are mainly curcumin nano-drug, and they are summarized in Table 2. Curcumin nano-medicine (C-SSM) is composed of curcumin coated with a biodegradable and no-toxic wall material (polyethylene glycol-grafted phospholipids), and the surface of the nano-medicine was further conjugated with vasoactive intestinal peptide (VIP), which binds to a receptor commonly found on the surface membranes of cancer cells, thus achieving targeted delivery. Indeed, the nano-medicine displayed lower IC₅₀ value than free drug. This curcumin nano-drug (5 μM) effectively suppressed mammosphere formation, demonstrating that targeting BCSCs by the new nano-formula is promising (Gulcur et al. 2013). In another study, both biocompatible and stable nanoparticle delivery systems for curcumin were developed by embedding curcumin with amphiphilic phosphorylated calixarene POCA4C6. This complex could gradually release curcumin depending on pH with minimal toxicity in vivo. Compared with free curcumin, it was more potent to reduce proliferation, invasive and migrating capacity of BT-549 cancer cells and tumor formation in xenografted mouse, and also reduced the population of CD44⁺/CD133⁺ cells (Chen, Li, et al. 2017). Given that CSCs commonly exist in the internal space of tumor, where is difficult for hydrophobic drug to reach owing to the lack of capillaries, specific ligands have been used in nano-drug systems to facilitate their internalization into tumors (Torchilin 2014). Meanwhile, because most of BCSCs highly express CD44, which is a hyaluronic acid, nano anti-BCSCs drugs which could effectively bind to CD44 has been developed. A cholesteryl-hyaluronic acid (CHA) nanogel-drug conjugates embedding curcumin displayed promising effects in suppressing the malignancy of floxuridine-resistant MDA-MB-231 which highly expresses CD44. This novel conjugates were capable of penetrating into mammary tumor effectively and showed more potent cytotoxicity in comparison with free drugs (Wei et al. 2013). Similarly, a more intelligent and effective curcumin delivery system against BCSCs was designed by Yang and colleagues. This co-releasing paclitaxel and curcumin nano-drug could automatically change its surface charge, remove its polyethylene glycol coat to decrease its size during circulation and reaching tumor sites. The new drug could more effectively inhibit mammospheres formation of MCF 7 breast cancer cell line, and reduce CD44⁺/CD24⁻ and ALDH1⁺ cancer cells in mammospheres compared with the traditional combination therapy of paclitaxel (Yang, Sun, et al. 2017).

Table 2.

Anti-BCSCs effects of food-derived nano-polyphenols.

Nano-phytochemicals	Phytochemicals concentration	Effects on CSCs	Molecular targets	References
Curcumin-loaded POCA4C6 micelles (CPM)	3 μM, 5 mg/kg	inhibit proliferation, invasion, migration, tumor spheroid formation; induce apoptosis;	↓nuclear β-catenin; ↓androgen receptor	(Chen, Li, et al. 2017)
co-releasing paclitaxel and curcumin micellar system(PPBV micelles)	0–2.5 μg/mL 10 mg/kg	inhibit mammosphere formation; reduce the percent of cells bearing ALDH1+ or CD44+/CD24−; inhibit tumor growth		(Yang, Sun, et al. 2017)
CHA nanogel-drug conjugates embedding curcumin	–	inhibit mammosphere formation;		(Wei et al. 2013)
Curcumin nano-medicine (C-SSM)	5μM	inhibit mammosphere formation; reduce CD44+CD24−/low cells		(Gulcur et al. 2013)

^a↓Represents down-regulation.

Abbreviations: HIF-1α, Hypoxia-inducible factor-1α;VEGF, vascular endothelial growth factor.

Additionally, nano-drugs made of other dietary polyphenols are also reported, although these studies mainly focus on preventing cancer cells rather than CSCs/BCSCs. A nanodrug made by embedding EGCG with polylactic acid-polyethylene glycol was 10 times more effective in inhibiting cancer cells than free EGCG (Siddiqui et al. 2013). A more advanced EGCG nano-fomula was designed later by introducing a special polymer which could specifically recognize prostate membrane antigen, achieving targeted delivery of EGCG. Indeed, the new nano-EGCG displayed better suppressing effects on cancer cells than free drug (Siddiqui et al. 2013). Nano-quercetin was acquired by encapsulating quercetin with lecithin, which was able to improve bioavailability of drug by facilitating drug pass through cell membrane, so lower doses of the naon-drug could effectively induce MCF7 cell apoptosis compared to using quercetin alone (Minaei et al. 2016).

In short, polyphenols not only can be obtained from diets, but also administered through nano drug delivery system to specifically target cancer cells and BCSCs.

Conclusions and perspectives

Food-derived polyphenolic compounds have gained extensive attentions owing to their significant anti-cancer and anti-CSCs effects. Polyphenolic compounds are effective in suppressing the stemness of BCSCs, blocking mammosphere formation and neoplasm growth, as well as suppressing proliferation, migration, invasion and EMT. Polyphenolic compounds are able to inhibit cancer cells and BCSCs through multiple mechanisms, including suppression of Wnt/β-catenin pathway, preventing inflammation along with altering the expression of miRNAs to reduce the stemness of BCSCs. Therefore, polyphenolic compounds are promising in reducing and preventing cancer development, especially breast cancer development. However, several key questions remain to be addressed. Though a number of possible reasons have been identified to account for the antineoplastic functions of polyphenolic compounds, these beneficial effects are likely derived from a key mechanism, which is primarily responsible for the various anti-cancinogenesis and anti-BCSCs effects of polyphenolic compounds. Up to date, this key mechanism explaining the various effects of polyphenolic compounds in preventing/reducing carcinogenesis remain to be established. Mitochondria play a central role on cell fate commitment by orchestrating a series of metabolic events (Siu and Alway, 2005; Alaynick, 2008); mitochondrial malfunction is common and CSCs preferentially rely on glycolytic metabolism, which render cancer cells and CSCs escaping mitochondria-mediated apoptosis (Loureiro et al. 2013; Ju et al. 2014; Shen et al. 2015). Polyphenols, redoxactive compounds, are able to elicit an energy restricted state, which ameliorate mitochondrial dysfunction and induce mitochondrial biogenesis (de Oliveira et al. 2016). Furthermore, polyphenols are able to reprogram metabolic ways of CSCs to oxidative metabolism by modulating mitochondria-related enzyme (hexokinase), which is helpful in suppressing the stemness of CSCs and facilitating mitochondria-mediated apoptosis (Wang, Fan, et al. 2015). Green tea, which is rich in EGCG, suppressed BCSCs proliferation by regulating mitochondrial metabolism, rendering cancer cell more quiescent (Bonuccelli, Sotgia, and Lisanti 2018). Thus improvement in mitochondriogenesis and mitochondrial function due to dietary polyphenols may have a central role in inhibiting cancer cells/CSCs, which needs to be further confirmed in the future. In addition, the most available studies were conducted in cultured cells and in rodents, where high doses of polyphenolic compounds were commonly used. Whether these findings in vitro and in rodents can be translated into humans remain to be tested in vivo. Also, most polyphenolic compounds have low bioavailability and also not quite stable. Thus, how to effectively administer these compounds to humans are another area needing to be further studied. Finally, to accurately quantify the anti-BCSC effects of phytochemicals, BCSCs need to be isolated. Up to now, inconsistency in the definition of BCSCs remains and the biological identity of BCSCs need to be further explored. BCSCs may represent a stage where differentiated cancer cells gain stemness, or they may represent a unique pool of stem-like cells which are the sources of all differentiated cancer cells. The expression of BCSCs markers often vary depending on diverse subtypes of breast cancer, histologic stage and heterogeneity within a tumor (Tsang et al. 2012), which adds further complexity.