Cancer Stem Cells (CSCs) and Mechanisms of Their Regulation: Implications for Cancer Therapy

Abstract

The identification of small subpopulations of cancer stem cells (CSCs) from blood mononuclear cells in human acute myeloid leukemia (AML) in 1997 was the landmark observation for recognizing the potential role of CSCs in tumor aggressiveness. Two critical properties contribute to the functional role of CSCs in the establishment and recurrence of cancerous tumors: their self-renewal capacity and their potential to differentiate into unlimited heterogeneous populations of cancer cells. These findings suggest that CSCs may represent novel therapeutic targets for the treatment and/or prevention of tumor progression as they appear to be involved in cell migration, invasion, metastasis, and treatment resistance, all of which lead to poor clinical outcomes. The identification of CSC-specific markers, the isolation and characterization of CSCs from malignant tissues, and targeting strategies for the destruction of CSCs provides a novel opportunity for cancer research. Described in this overview is the potential implication of several common CSC markers in the identification of CSC subpopulation restricted to common malignant diseases e.g., leukemia, breast, prostate, pancreatic and lung cancers. The role of microRNAs (miRNAs) in the regulation of CSC function is also discussed, as are several methods commonly used in CSC research. The potential role of the anti-diabetic drug metformin that has been shown to have effects on CSCs, and known function as an anti-tumor agent, provides an example of this new class of chemotherapeutics.

1. Introduction

While cancer stem cells (CSCs) were recognized several decades ago, it is only in the the last 15 years that they have been identified and characterized in hematological malignancies, such as leukemia (Bonnet and Dick, 1997) and other tumors. This led to an increased interest in the potential role of CSCs in tumor aggressiveness, treatment resistance, and tumor recurrence (relapse) and metastasis (Rasheed and Matsui, 2012;Sarkar et al., 2009;Yu et al., 2012b). Like normal pluripotent stem cells, CSCs are long-lived, and display quiescent potentials in a dormant state, and are responsible for angiogenic induction, apoptotic resistance, self-renewal and differentiation. These characteristics are orchestrated by a rare sub-population of tumor cells within the total tumor cells present in a tumor mass, namely CSC cells that express stem cell marker genes, including Oct4, Sox2, Nanog, c-kit, ABCG2, and ALDH (Charafe-Jauffret et al., 2009;Croker et al., 2009;Prud'homme, 2012;Yu et al., 2012a). These characteristics suggest that CSCs themselves contribute to tumor development and progression. While the pathogenic effects of CSCs remains to be elucidated, it is widely believed that intrinsic and extrinsic alterations in the stem cell tumor microenvironment, together with mutations and epigenetic regulations, are mainly responsible for the development of CSCs that are involved in tumor initiation and progression (Figure 1) (Bao et al., 2012a).

Figure 1.

The implication of stem cells in the development and progression of tumor. NSC: normal stem cells; CSC: cancer stem cells; Hh: hedgehog; Bmil-1: polycomb complex protein; EZH2: enhancer of zeste homolog 2; miRNAs: microRNAs.

It is known that CSCs constitute only a small percentage (0.05-1%) of tumor cells within a tumor mass containing heterogeneous population of tumor cells within the tumor microenvironment (Li et al., 2012;Yu et al., 2012a;Yu et al., 2012b). These CSCs have the capacity for self-renewal, giving rise to uncontrolled amplification of differentiated cell populations with altered molecular and cellular phenotypes. These eventually lead to the heterogeneous primary and metastatic tumor cells within a tumor mass that may be resistant to therapeutics and contribute to tumor recurrence (Li et al., 2012;Yu et al., 2012a). It is interesting to note that CSCs are important in the prognosis of many malignant diseases which has been demonstrated by the finding that they are present in the majority of malignant tumor tissues, and appear to be resistant to chemo-radiation therapy as compared to their differentiated progenies (Creighton et al., 2010;Lee et al., 2008). This may explain why tumor regression alone does not correlate with overall survival rate in cancer patients (Creighton et al., 2010). Rather, it appears that tumor recurrence/relapse occurs because of the presence and sustenance of CSCs within the tumor microenvironment, even after conventional cancer therapy, which suggests that CSCs plays a critical role in treatment resistance, tumor metastasis and recurrence. The identification of CSC-specific markers, the isolation and characterization of CSCs from malignant tissues, and the development of strategies for targeted eradication of CSCs represent an important opportunity in cancer research.

Described in the present report are the implications of several common CSC markers and their relevance to common malignant diseases such as leukemia, breast, prostate, pancreatic and lung cancers. The importance of microRNAs (miRNAs) in the regulation of CSC characteristic is also considered, as are the methods commonly employed in CSC research. As an example of the possible success of this approach, the anti-tumor activity of metformin, an anti-diabetic agent and potential CSC regulator, is discussed.

2. CSCs and tumor aggressiveness

Cancer stem cells were first identified and characterized in the bone marrow of AML patients in 1997. Subsequent clinical and laboratory studies provided additional evidence supporting the role of CSCs in drug resistance and cancer metastasis, thereby contributing to the poor outcomes experienced by patients with pancreatic, prostate, liver, breast, and brain tumors (Bauerschmitz et al., 2008;Lee et al., 2008;Matsui et al., 2008). For example, CD133⁺ pancreatic CSC cells co-express the CXC chemokine receptor CXCR4 at the invading margins of human ductal pancreatic tumors (Lee et al., 2008;Narducci et al., 2006;Klein et al., 2001). Both CD133⁺CXCR4⁻ and CD133⁺CXCR4⁺ CSC cells isolated from human pancreatic tumors are able to generate and reconstitute primary tumors in mouse xenograft models. However, only the CD133⁺CXCR4⁺ cells display a significant metastatic capacity in this animal model. As the inhibition of CXCR4 in these pancreatic CSCs prevents metastasis in this xenograft mouse model (Hermann et al., 2007), it appears that CD133⁺CXCR4⁺ CSCs play a critical role in tumor metastasis although it is not the case for CD133⁺CXCR4⁻ cells. Moreover, the CSC-like glioma cells also contribute to resistance to radio-therapy via preferential activation of DNA damage response pathways, and by increasing DNA repair capacity (Bao et al., 2006). Similarly, a subpopulation of glioma cells with the phenotype of CD133⁺, a CSC marker for various tumors, is enhanced after glioma radiation (Bao et al., 2006). The CD133⁺ glioma cell population surviving after treatment with ionizing radiation displays a significant increase in the proportions of CSCs relative to the large numbers of CD133⁻ tumor cells, suggesting that they are responsible for the radiation-resistant phenotype in gliomas (Bao et al., 2006).

It has been reported that mouse mammary tumor CSCs are responsible for cisplatin resistance (Shafee et al., 2008) and that CSCs in human colorectal cancer tissue are responsible for resistance to chemotherapeutic agents (Dylla et al., 2008). Moreover, human breast cancer cells containing aldehyde dehydrogenase (ALDH)⁺ CSC cells have an increased metastatic capacity with distinct CSC molecular phenotypes, e.g., ALDH, Notch-2, and CXCR1 (Charafe-Jauffret et al., 2009). This suggests that CSCs are also involved in the regulation of breast cancer metastasis.

The involvement of CSCs in drug resistance has been demonstrated in pancreatic, colon, breast, and brain tumors. Moreover, CSC-containing tumors display greater tumorigenic and metastatic potential in vitro and in vivo than non-CSC cancer cells. It has been documented that human pancreatic cancer tissues contain a small subpopulation of CD133⁺ CSC phenotypic cells that are exclusively tumorigenic and highly resistant to standard chemotherapy (Hermann et al., 2007). Elimination of these CSC populations suppresses the metastatic phenotype of pancreatic tumors without modulating their tumorigenic potential (Hermann et al., 2007). Similarly, CSCs in pancreatic tumor tissues are associated with drug resistance and metastatic potential with pancreatic tumor cells having a CD44 positive CSC phenotype, which correlates with tumor histological grade and poor clinical outcomes (Hong et al., 2009). These findings suggest that the CSCs, promote tumor aggressiveness. Given these findings CSCs appear to be an excellent target for treating malignancies.

3. Identification of CSC markers in common malignant diseases

The identification and characterization of CSCs in malignant diseases provides insights as to ways in which to selectively inhibit or eradicate CSCs as a treatment for tumor aggressive phenotypes. Several common stem cell markers, including CSC-specific markers such as CD34, CD44, CD123, CD133, Oct4, Sox2, Nanog, c-kit, ABCG2, and ALDH have been identified in the CSC populations isolated from a wide variety of malignant diseases.

3.1. Leukemia

Leukemia is one of the most commonly diagnosed malignant diseases in children and adults (Siegel et al., 2012). This includes acute myeloid leukemia (AML), chronic myeloid leukemia (CML), acute lymphoblastic leukemia (ALL), and multiple myeloma (MM). Within each of these groups there is significant patient-to-patient heterogeneity in leukemic blast cell morphology. For instance, AML is classified into seven French-American-English subtypes, according to the maturation stage of the acute leukemias, the preferential expression of multi-lineage cell markers, and morphology among individual patients (Bennett et al., 1985;Warner et al., 2004). Targeting these different leukemic blasts, especially leukemia CSCs, while avoiding normal hematopoietic stem cells, provides an opportunity for the treatment or eradication of these conditions.

A small subpopulation of leukemic cells namely CSCs displaying the CD34⁺CD38⁻ cell surface phenotype were first identified in bone marrow samples of AML patients (Bhatia et al., 1997;Bonnet and Dick, 1997) using flow cytometry-based sorting. The injection of 5,000 leukemic CSCs developed human leukemia in immune compromised mice (Bhatia et al., 1997;Bonnet and Dick, 1997). These small subpopulations of leukemia CSC cells also display a greater capacity for stem cell self-renewal as compared to normal adult bone marrow cells.

This seminal discovery led to the identification and characterization of CSCs in other cancers. Thus, several stem cell markers like CD34, CD38, HLA-DR, and CD71, which are shared with normal hematopoietic stem cells, have been identified in leukemic cells (Dick, 2005;Warner et al., 2004). Some cell surface markers, such as CD90 (Thy-1), are differentially expressed between normal hematopoietic stem cells and leukemia CSCs, with CD90 being under-expressed in leukemic CSCs (Johnsen et al., 2009;Warner et al., 2004). This suggests that CD90 may be a useful marker for distinguishing leukemic CSCs from normal hematopoietic stem cell sub-populations (Blair et al., 1997;Warner et al., 2004). Loss of expression of stem cell factor c-kit, also known as CD117, is a consistent feature of leukemic CSCs (CD34⁺c-kit⁻) isolated from AML, but not of normal bone marrow hematopoietic stem cells (CD34⁺c-kit⁺) (Blair and Sutherland, 2000). It has also been reported that CD123 (IL-3Rα) is a unique marker of leukemic CSCs (Jordan et al., 2000) and that it is essential in enhancing cell survival pathways in leukemia (Testa et al., 2004). The expression of CD123 is significantly increased in the CD34⁺CD38⁻ CSC population isolated from leukemia patients (Budel et al., 1989;Testa et al., 2004). However, expression of CD123 is undetectable in CD34⁺CD38⁻ CSC cells isolated from the bone marrow of normal subjects (Jordan et al., 2000). Moreover, activation of NF-κB increases the number of quiescent leukemia CSC populations, without affecting normal hematopoietic stem cell subpopulations (Warner et al., 2004). These findings suggest that there are cell markers unique for leukemia CSCs that are not found in normal hematopoietic stem cells.

Monoclonal antibodies against CD44, CD47 and CD123 are effective in eliminating leukemic CSCs in AML. This finding is consistent with the inhibition of the IL-3-mediated signaling pathway between leukemic stem cells and supporting cells in NOD/SCID mice (Jin et al., 2006;Jin et al., 2009;Majeti et al., 2009). Targeting these leukemic CSC related markers or proteins could provide an effective strategy for the eradication of CSC populations in leukemia. The identification and characterization of leukemic CSC-specific markers that are not shared with normal hematopoietic stem cells makes it possible to identify and develop novel therapeutic agents for the management of human leukemias.

3.2. Breast cancer

Breast cancer is the second most deadly malignancy for females, with one in eight women expected to develop the condition in their lifetime (Siegel et al., 2012). Breast cancer affects 121 per 100,000 people, with a greater incidence among African Americans than other ethnic groups. Thanks to early detection and more effective treatment options the survival rate for breast cancer increased from 84% to 90% between 1987 and 2007 (Siegel et al., 2012). Nonetheless, breast cancer remains a major cause of cancer-related death because of treatment resistance and metastatic disease, especially for those diagnosed at advanced stages of the disease.

Breast CSCs are a small subpopulation (0.1-1%) of breast cancer cells in primary tumors. A rare subset of breast CSC has a high capacity for self-renewal and is able to initiate tumorigenesis when transplanted into NOD/SCID mice (Al-Hajj et al., 2003;Klonisch et al., 2008). Several common CSC markers, including CD44, CD133, ALDH, c-kit, ESA and ABCG2, have been identified in primary breast cancer specimens (Prud'homme, 2012).

A critical role for CD44⁺ in the development of this condition is indicated by the finding that injection of less than 100 breast CSCs with the phenotype of CD44⁺/CD24⁻ can result in 85% tumor formation in xenograft models, while injection of more than 10,000 non-CSC breast adenocarcinoma cells fails to do so (Al-Hajj et al., 2003;Patrawala et al., 2005). Moreover, CD133⁺ breast CSCs have characteristics similar to CD44⁺/CD24⁻ CSCs while CD133⁻ breast adenocarcinoma cells do not generate tumors in mouse tumor xenograft models (Klonisch et al., 2008) As there is no overlap in cell surface proteins between CD133⁺ and CD44⁺/CD24⁻ CSCs (Al-Hajj et al., 2003;Patrawala et al., 2005), there is no universal CSC marker for each type of breast cancer. However, there is some overlap of CSCs among patients between ALDH⁺ and CD44⁺/CD24⁻ cell subpopulations (Yu et al., 2012a). Data indicate that breast cancer CSCs with a CD44⁺CD24⁻ALDH⁺ phenotype have greater tumorigenic potential than CSCs with the CD44⁺CD24⁻ or ALDH⁺ phenotype (Ginestier et al., 2007). As ALDH is not expressed with CD44 and CD133 in ovarian tumors, it appears that CSCs with the phenotype of CD44⁺, CD24⁻, CD133⁺, ALDH⁺ have the most pronounced tumorigenic potential in breast cancer, making these CSC subpopulations are attractive targets for the treatment of breast cancer.

Expression of mucin 1 (MUC-1), a mediator of the growth of undifferentiated human embryonic stem cells (Hikita et al., 2008) that is overexpressed in human estrogen positive and negative breast carcinomas, is associated with breast tumor cells and its side population cells, one type of so called CSC cells identified by Hoechst 33342 dye (Kufe, 2012;Engelmann et al., 2008). This suggests that it is a potential CSC marker for breast cancer.

3.3. Prostate cancer

Prostate cancer is the most commonly diagnosed malignant disease in males, being the second leading cause of cancer-related death for men in the United States (Siegel et al., 2012). As most prostate cancer patients are treatable, the survival rate has increased significantly over the years, although fatalities still occur, especially with aggressive phenotypes that are resistant to chemotherapy. Considerable effort has been expended to identify and characterize prostate CSC populations so as to target them in treating this condition. Subpopulations of CSCs (<0.1%) have been identified in primary prostate cancers that preferentially express the cell surface markers, CD44, CD133, stem cell antigen 1 (Sca-1), collagen receptor α2β1^hi, CK5/14, CK8/18, CD49f, and ABCG2 (Klonisch et al., 2008). However, different CSC subpopulations in prostate cancer are dependent on the enrichment of CSCs in different cellular compartments (Kasper, 2009). For example, the expression of CD44 in prostate CSCs is present in most prostate cancer basal cells, whereas CD133, α2β1^hi, CK5/14, CK8/18 and ABCG2 are expressed in less than 1% of basal cells (Klonisch et al., 2008). Nevertheless, there is some overlap of CSCs between the α₂β₁^hi, CD44⁺, and ABCG2⁺ cell populations (Patrawala et al., 2007). Subpopulations of CSCs in prostate cancer have high proliferative capacity, increased clonogenic potential, and a greater capacity for tumorigenesis and metastasis in xenograft models in vitro and in vivo. Injection of 1,000 prostate cancer cells with the CD44⁺CD24⁻ CSC phenotype consistently develops tumors in NOD/SCID mice, whereas the injection of non-CSC cancer cells does not (Hurt et al., 2008). It has been reported that Sca-1, a cell surface protein, is a potential CSC marker that is unique to prostate cancer. It is found primarily in the proximal regions of the organ (Zhou et al., 2007) in the same site as the CSC niche, a very small or limited area of tumor tissues where CSCs resides closely in a similar, permissive environment to retain the high capacity of self-renewal and multi-lineage differentiation potential as well as protect them from diverse genotoxic insults (Tsujimura et al., 2002). The Sca-1⁺ prostate epithelial cells that display an increased activation of Akt signaling by transfection of constitutively active AKT can initiate prostate cancer in xenograft models (Xin et al., 2005). In addition, PTEN, a known tumor suppressor that targets primarily the PI3K/Akt/mTOR signaling, and c-kit are proposed to be potential markers of CSCs in prostate cancer (Klonisch et al., 2008). The expression of these proteins is down-regulated in prostate cancer cells, especially prostate CSCs.

3.4. Pancreatic cancer

Pancreatic ductal adenocarcinoma is currently the fourth leading cause of cancer-related deaths in the United States (Siegel et al., 2012). It is also one of the most lethal malignancies, with a 5-year survival rate of less than 5% (Edwards et al., 2005). It is estimated that each year in the United States, 44,000 patients are diagnosed with pancreatic cancer (greater than 90% as pancreatic ductal adenocarcinoma), with 37,000 dying from this malignancy. Due to the absence of specific symptoms, the lack of early sensitive detection tests, and its rapid and insidious growth, pancreatic cancer is typically diagnosed at an advanced and incurable stage. Thus, the overall survival of patients is approximately 5–6 months, even with conventional therapy for locally advanced and metastatic disease.

Pancreatic CSC populations account for less than 1% of all pancreatic cancer cells. They have the capacity for self-renewal and uncontrolled potential of differentiated progeny. Pancreatic CSC populations express the cell surface markers CD44⁺, CD24⁺ and epithelial-specific antigen (ESA)⁺ (Lee et al., 2008;Klonisch et al., 2008). When transplanted into NOD/SCID mice, a CD44⁺CD24⁺ESA⁺ CSC subpopulation isolated from human primary pancreatic cancers readily formed tumors, while cancer cells lacking these cell surface markers were poorly tumorigenic (Lee et al., 2008;Klonisch et al., 2008). The CSC marker-positive cells display a 100-fold increased capacity for the development of tumors and exhibit tumor morphology similar to primary pancreatic cancer. Moreover, these CSCs maintain their cell surface marker phenotype after repeated passages as xenografts in immunocompromised mice (Li et al., 2007;Lee et al., 2008). Unlike normal pancreatic epithelial cells and non-CSC-like cancer cells, these pancreatic cancer CSC phenotypic cells also display a strong transcriptional up-regulation of sonic hedgehog (SHH) and the polycomb group (PCG) gene family member Bmi-1. All of these are known mediators for maintaining CSC characteristics (Li et al., 2007;Lee et al., 2008). Pancreatic cancers contain 1–3% of CD133⁺ cancer cells, some of which show high expression of CXCR4, a pro-invasive marker. These CD133⁺CXCR4⁺ cells, but not CD133⁺CXCR4⁻ cells, have significant metastatic capacity. The selective inhibition by AMD3100 of CXCR4 signaling in CXCR4⁺ CSC cells blocks tumor tissue invasion (Hermann et al., 2007), suggesting a potential role of CXCR4 in pancreatic tumor metastasis. Accordingly, it remains possible that there is more than one type of CSC sub-population in pancreatic cancer tissues, which would be consistent with the known heterogeneity of most human tumors (Klonisch et al., 2008).

Putative human pancreatic CSC-like cells derived from MiaPaCa-2 sphere-forming cells obtained from the mouse xenograft tumors showed increased expression of CD44, EpCAM, and enhancer of zeste homolog 2 (EZH2), which is an epigenetic mediator involved in the regulation of CSC characteristics (Bao et al., 2012b). Expression of EZH2 increases in various tumors, including pancreatic cancer (Bao et al., 2012b;Toll et al., 2010). Injection of as few as 5,000 of these CSC-like sphere cells into SCID mice produces tumors within 2–3 weeks, whereas it is necessary to inject 10⁷ non-CSC parental cancer cells to achieve comparable tumor incidence. As these pancreatic CSC-like cells display an increased capacity for migration and invasion, aggressiveness, and self-renewal capacity (Bao et al., 2012b), it is possible that EZH2 can be used as a CSC marker and might be a therapeutic target in the treatment of pancreatic cancer.

3.5. Lung cancer

In the United States, 226,000 people are diagnosed with lung cancer each year, with 160,000 individuals dying annually from this malignancy (Siegel et al., 2012). Although there has been significant progress over the past decade in the diagnosis and the treatment of this condition, prognosis remains poor due to treatment resistance, rapid tumor growth, and metastatic capacity. A small subpopulation of lung CSC cells appears to be responsible for the aggressive phenotypes of lung cancer. This group expresses typical stem cell markers, such as CD133, CD44, ALDH, Oct4, and Nanog (Eramo et al., 2010;Wu et al., 2012). The remarkable heterogeneity among lung cancers in terms of cell origin, biology, etiology, and molecular/genetic pathogenesis influences treatment strategies and prognosis. This makes it important to be able to distinguish non-small cell lung cancer (NSCLC), which accounts for 40% of lung tumors, and small cell lung cancer (SCLC), which accounts for 20% of lung tumors, on the basis of CSC characteristics. Although there are controversies regarding lung cancer CSC markers, CD133, CD44, ALDH, Oct4, Nanog, and ABCG2 have been associated with NSCLC, whereas only CD133 and ALDH have been found in association with SCLC (Kitamura et al., 2009;Nurwidya et al., 2012;Wu et al., 2012). High levels of the tumor metastatic marker urokinase plasminogen activator receptor (u-PAR), podocalyxin-like protein 1 (PODXL-1), a marker of embryonic and hematopoietic stem cells, and B cell-specific Mo-MULV integration site 1 (Bmil-1), a member of polycomb group protein family, are considered potential markers for CSCs in SCLC ((Kitamura et al., 2009). Lung CSCs with the CD44⁺CD133⁺ phenotype display an increased capacity for self-renewal and show unlimited differentiated progeny of heterogeneous populations of NSCLC cancer cells in comparison with cancer cells lacking these markers (Eramo et al., 2010). Expression of ABCG2, a marker for drug resistance, epithelial-specific antigen (ESA), stem cell factor c-kit, and CXCR4 is increased in lung CSC subpopulations of NSCLC (Sung et al., 2008;Ho et al., 2007;Wu et al., 2012). The atypical protein kinase C (PKC) iota, an oncogene for NSCLC, is required for bronchioalveolar stem cell growth in vitro and in vivo and for lung tumor formation in the PCK iota-deficient K-ras mouse model (Regala et al., 2009). Expression of Rac-1, a small GTP binding protein of the Rho family, is increased in CSC populations associated with NSCLC (Akunuru et al., 2011). Because Rac-1 is a signal transducer for several oncogenic pathways involved in cell survival, proliferation, migration and invasion in tumors, it appears to be essential for the K-ras mediated tumor growth of lung tumors (Kissil et al., 2007). In addition, RAC-1 is a downstream target of PKC iota, with PCK iota/Rac-1 signaling required for the activation of the K-ras signaling pathway in NSCLC tumors (Fields and Regala, 2007). Therefore, Rac-1 and/or PKC iota are potential CSC markers for NSCLC. Inasmuch as CD133, CD44, and ALDH are common markers for both normal and cancer stem cells, there are no selective markers for lung CSC populations. This makes it necessary to identify and use multiple markers for characterizing the lung CSC subpopulations (Wu et al., 2012).

4. The role of miRNAs in the regulation of CSC characteristics

MicroRNAs (miRNAs), non-protein-coding RNAs 18–24 nucleotides in length, are post-transcriptional regulators of mRNAs that binds to specific sites in the 3’ untranslated region (3’-UTR) of their target mRNAs. This attachment results in either mRNA degradation or inhibition of protein synthesis (Liu and Tang, 2011), making miRNAs useful tools in characterizing the role of select proteins in cell function. Moreover, miRNAs have an important role in tumorigenesis, with alterations in their expression in cancers associated with clinical outcome, therapy resistance, and tumor recurrence/relapse (Fabbri et al., 2007;Liu and Tang, 2011). It is also known that miRNAs regulate the CSC characteristics by affecting signaling pathways and CSC signature genes. Particular miRNAs, which are differentially expressed in CSCs or CSC-like cells in various tumors, are potential CSC markers, as discussed below.

4.1. Let-7/miR-200

The members of the Let-7 and miR-200 family participate in the development and progression of tumors by targeting multiple cell signaling pathways involved in cell survival. Moreover, it has been demonstrated that the expression levels of let-7 and miR-200 are related to clinical outcomes in cancer patients (Olson et al., 2009;Peter, 2009;Wendlandt et al., 2012). As the expression of both of these miRNA family members is either lost or significantly reduced in leukemia, breast, prostate, pancreatic, and lung cancers, it is possible that they may be tumor suppressors. Indeed, let-7 family members are negative regulators of the epithelial-to-mesenchymal transition (EMT), a developmental event associated with treatment resistance, metastasis, and tumor recurrence. These effects are similar to those observed with CSCs that are, in part, mediated through regulation of PTEN and CSC gene signature marker Lin28B in pancreatic and prostate cancer cells (Kong et al., 2010;Chang et al., 2011;McCarty, 2012;Li et al., 2009;Peter, 2009). The miR-200 family members also inhibit the EMT phenotype by directly targeting the EMT regulators, ZEB1 and ZEB2 (Peter, 2009;Kent et al., 2009;Li et al., 2009), and, like CSC, miR-200 decreases the expression of Bmil-1, Suz12, and Notch-1, known regulators of CSC and EMT phenotypes and function, in various cancer cells (Bao et al., 2011;Iliopoulos et al., 2010;Leal and Lleonart, 2012). The let-7 family members also inhibit the expression of EZH2, a major epigenetic component of the polycomb repressive complex 2 (PRC2) that inhibits the expression of developmental genes in embryonic and adult stem cells (Kong et al., 2012b). Thus, EZH2 may be a regulator in maintaining the characteristics of CSCs (Chang and Hung, 2012). Down-regulation of let-7 and miR-200 family members occurs in CSCs of various tumors, including breast cancer (Golestaneh et al., 2012;Yu et al., 2007;Shimono et al., 2009), suggesting that the let-7 and miR-200 families play critical roles in the regulation of CSCs via regulation of multiple signaling pathways implicated in survival and the acquisition of EMT. For this reason, these miRNAs could serve as CSC markers and therapeutic targets.

4.2. miR-21

It appears that miR-21 acts as an oncogenic molecule by targeting multiple survival signaling pathways (Pang et al., 2010). Increased expression of miR-21 in pancreatic, prostate, lung, and breast cancer tumors, as well as in leukemia, is associated with poor clinical outcomes (Dillhoff et al., 2008;Moriyama et al., 2009) and the suppression of PTEN expression (Ali et al., 2010;Bao et al., 2012b). In addition, miR-21displays anti-apoptotic activity and enhances the proliferative, invasive and angiogenic potentials in a wide variety of tumor cells (Moriyama et al., 2009;Zhang et al., 2007). Expression of miR-21 is increased significantly in CSC populations as compared to non-CSC cancer cells while forced overexpression of miR-21 by its mimic cells enhances survival of bone marrow mesenchymal stem cells (Golestaneh et al., 2012;Han et al., 2012). Conversely, functional loss of miR-21 increased apoptosis of mesenchymal stem cells (Nie et al., 2011).

4.3. miR-34a

It has been reported that miR-34a is under-expressed in a variety of tumors, including those associated with pancreatic, prostate, breast and lung cancers, and with leukemia (Kent et al., 2009;Kong et al., 2012a). Decreased levels of miR-34a are also thought to be associated with poor clinical outcomes (Kent et al., 2009;Kong et al., 2012a). Mounting evidence suggests that miR-34a may act as a tumor suppressor by inhibiting cell survival, proliferation, invasion, and metastasis, which are mediated, in part, through activation of p53 and inactivation of Cyclin D1, E2F1/2, and CDK6 in tumor cells (Aranha et al., 2011;Guo et al., 2011;Lodygin et al., 2008;Sun et al., 2008;Wang et al., 2011). It has been shown that miR-34a inhibits the expression of CSC signature genes such as CD44, CD133, and Notch-1, which is consistent with its ability to attenuate the self-renewal capacity of many tumor cells (Kong et al., 2012a;Liu and Tang, 2011;Nalls et al., 2011). The expression of miR-34a has been found to be significantly decreased in CD133+ glioma CSC-like cells (Sun et al., 2012). This suggests the loss of miR-34a acts as a tumor suppressor in the regulation of the CSC function, and points to the possibility of employing miR-34a as a CSC marker and therapeutic target. A strategy to up-regulate miR-34a in tumors would be a novel approach for the treatment of cancer.

5. Common experimental methods used in CSC research

Summarized on Table 1 are several common methods or techniques that have been used to better understand the biology of CSCs.

Table 1.

Most Common Experimental Methods/Assays in CSC Research

Methods/Assays	Biological Significance in CSC Research
1. Clonogenic assay (Ali et al., 2010)	Evaluation of proliferative potential of CSCs
2. Sphere formation assay (Bao et al., 2012b)	Evaluation of CSC self-renewal capacity
3. Immunostaining assay for FACS*, image microscopy, or flow cytometry (Kong et al., 2009)	Isolation of pure CSC subpopulation and/or evaluation of the expression of CSC markers
4. Real-time RT-PCR assay (Bao et al., 2012b)	Evaluation of the expression of CSC marker mRNAs and CSC-related miRNAs
5.Mouse xenograft tumor assay of CSC or CSC-like sphere cells (Bao et al., 2012b)	Evaluation of CSC or CSC-like sphere cells-initializing tumor formation

^*FACS (fluorescence activated cell sorting).

6. The anti-diabetic drug metformin as a potential anti-tumor agent targeting CSC subpopulations

Metformin, a biguanide, oral hypoglycemic agent, is the most commonly used drug for the treatment of type 2 diabetes mellitus (DM). Metformin reduces blood glucose levels by the down-regulation of hepatic gluconeogenesis and up-regulation of glucose uptake in peripheral tissues, such as skeletal muscle and fat (Shaw et al., 2005). It also enhances the insulin sensitivity of tissues, thereby reducing insulin levels overall.

Epidemiological and clinical studies indicate a reduced incidence of breast and pancreatic cancers in DM patients taking metformin, and that its consumption improves the clinical outcome of cancer patients (Bowker et al., 2006;Heikkinen et al., 2007;Landman et al., 2010;Libby et al., 2009;Monami et al., 2009). While these findings suggest that metformin has anti-tumor effects, the mechanism of this action remains undefined. One possibility is that the metformin-induced increase in insulin sensitivity inhibit cancer cell growth by activation of AMP kinase (AMPK) which, in turn, inhibits the PI3K/Akt/mTOR signaling pathway via phosphorylation of mTOR (Dowling et al., 2007;Zakikhani et al., 2006) and a rapid inhibition of cellular protein synthesis and growth (Cazzaniga et al., 2009;Goodwin et al., 2009;Martin-Castillo et al., 2010). Moreover, metformin can directly inhibit tumor cell growth and proliferation by regulating the cyclin D1-mediated cell cycle, p53 expression, and phosphorylation in breast and pancreatic cancers (Ben, I et al., 2010;Feng et al., 2007). Metformin can also decrease the production of inflammatory cytokines, including TNF-α, IL-6, and VEGF by inactivation of NF-κB and HIF-1α (Ersoy et al., 2008;Lund et al., 2008;Huang et al., 2009). It is also possible that the anti-tumor activity of metformin in vitro and in vivo may be associated with inhibition of the insulin/IGF-1 pathway through AMPK activation (Kisfalvi et al., 2009;Rozengurt et al., 2010) by inactivation of breast CD44⁺/CD24⁻ CSC cells and the EMT phenotype (Hirsch et al., 2009;Vazquez-Martin et al., 2010) or by inhibiting cell growth, clonogenic potential, migration/invasion, and CSC self-renewal capacity in gemcitabine-resistant pancreatic cancer cells (Bao et al., 2012c). It has also been found that metformin inhibits the expression of the CSC surface markers CD44 and EpCAM, expression of CSC genes such as EZH2, Notch-1, Nanog, and Oct4, and the miRNA expression of let-7 and miR-200 family in the CSC-like sphere cells of gemcitabine-resistant cells (Bao et al., 2012c). These findings indicate that the anti-tumor effects of metformin may involve the targeting of CSC subpopulations, providing additional proof in support of the importance of CSC cells in cancer.

7. Perspective and conclusion

A considerable body of evidence supports the hypothesis that a very small population of CSCs is associated with an aggressive tumor phenotype characterized by increased cell survival, migration, invasion, metastatic capacity, treatment resistance, and tumor recurrence, all of which ultimately contribute to poor prognosis. Although there have been efforts to characterize CSCs, their pathogenesis and molecular interactions in the tumor microenvironment are not well defined. The identification of CSC-specific markers, and the isolation and characterization of CSCs in malignant tissues will provide insights that will be of value in designing strategies for the development of chemotherapeutics that is expected to reduce tumor aggressiveness by targeting CSCs. Drug-induced modification of CSC-associated markers will modulate the phenotype and consequent function of these cells. For example, miRNAs such as, let-7, miR-200, miR-21 and miR-34a are possible targets as they play key roles in CSC regulation via multiple signaling pathways that regulate cell growth and survival. Because these miRNAs can be differentially expressed in the CSCs or CSC-like cells of various tumors, they may also be useful as CSC markers. The validity of this approach is suggested by the finding that metformin, an anti-diabetic drug, displays anti-tumor effects that may be due to the targeted elimination of CSCs. Additional clinical and preclinical work is required to demonstrate conclusively the therapeutic benefit of metformin, and CSC-targeting drugs in general, for the management of particular cancers.