Lung Cancer Stem Cells: Current Progress and Future Perspectives

Cancer stem cells (CSCs) have attracted a lot of interest in the field of cancer research in recent years. Lung CSCs share many characteristics with the normal pluripotent stem cells, such as self-renewal and multi-potent abilities. Identification of normal adult lung stem cells and their response to injury has led us to the lung cancer stem cells, partially based on the knowledge that lung CSCs as well as the normal lung stem cells share similar markers and/or location in the airway tree. Several studies have identified CD133, CD44, ALDH (aldehyde dehydrogenase) and ABCG2 (ATP-binding cassette sub-family G member 2) as lung cancer stem cell markers, all of which are validated CSC markers in multiple other cancer types as well. Embryogenesis signaling pathways, such as the hedgehog, wnt and notch pathways has also been implicated as determinants of lung CSC phenotype. With the central role of CSCs in tumor recurrence, metastasis and drug-resistance, targeting CSC markers and/or signaling pathways to eradicate lung cancer and enhance patient outcome is an attractive approach. This review summarizes our current understanding of CSC markers and signaling pathways in lung cancer.


Introduction
Lung cancer is one of the most common malignancy in the US and worldwide and a leading cause of cancer-related deaths [1]. Lung cancer consists of heterogeneous groups in terms of pathological features and is classified into the following two major types, Small Cell Lung Carcinoma (SCLC) and Non-Small Cell Lung Carcinoma (NSCLC). NSCLC includes squamous cell carcinoma and adenocarcinoma.
It is believed that the existence of Cancer Stem Cells (CSCs) within the tumors is responsible for the majority of biological characteristics that are associated with aggressive behavior of tumors/ cancer cells. CSC theory proposes that a tumor cell subpopulation has self-renewal capacity, cancer-initiating ability and multi-potent differentiation ability. In recent years, substantial experimental evidence has been reached in support of the role of CSCs. There are many concepts in cancer that can be explained by the stem cell theory like tumor self-renewal, tumor heterogeneity, and tumor relapse after treatment and resistance to conventional chemotherapies. CSCs were first described in human hematopoietic cancer. They have been studied in other solid tumors as well as in lung cancer [2][3][4][5]. Clinical specimens from SCLC and adenocarcinoma patients have been found to include a small subpopulation of cells (<1.5%) that are capable of forming colonies when grown on agar. Upon their intracranial injection into athymic nude mice, they produced cancers with characteristics identical to those of the original specimens [5]. Researchers have used different methods in investigate cancer stem cells, such as flow cytometry, because CSCs express tissue-specific cell surface markers, for example, CD133, ALDH, ABCG2, and CD44. All of these will be discussed in this article. The ability of ABCG2, expressed in stem cell populations to efflux the fluorescent Hoechst 33342 dye, forms the basis for side population assay. Distinct "low" Hoechst 33342 dye staining pattern cells are the so called Side Population Cells (SPC). They have been described in a variety of tumor types as being enriched in stem-like properties [6]. SPC are also found to have the ability to self-renew in addition to having a high level of human telomerase reverse transcriptase (hTERT). Telomerase is an important marker of cellular immortalization in cancers [6].
CSCs are resistant to chemotherapeutic drugs and can be selected from the parental population of human lung tumor cells by treatment with drugs. Cells which survived drug treatment have all the properties of CSCs. They have been reported to possess high clonogenic efficiency, enrichment of SPC and expression of CD133 marker, all of which are associated with CSCs [7]. carcinoma frequently arises in the proximal part of the airways and expresses Keratin 5 [13]. In the middle airway, Clara cells, which serve to protect and detoxify the bronchiolar epithelium, have been suggested by some to be progenitor cells [23] due to their role in maintenance of both secretory and ciliated cell types after oxidant-mediated damage. Clara cells also possess multi-potent capacity for differentiation. However, it is believed that most Clara cells do not possess the ability to self-renew [20]. Pulmonary Neuroendocrine Cells (PNEC) is specialized airway epithelial cells that produce specific neuropeptides and are grouped in clusters, termed Neuroepithelial Bodies (NEBs). Repair from airway injury is associated with PNEC hyperplasia [16]. However, PNEC do not have multi-potent differentiation ability [24]. SCLC, which frequently arises in the midlevel bronchioles, has been shown to express stem cell marker CD44 [13]. Moreover, similarly to PNECs, SCLC exhibits primitive neuroendocrine features, such as the expression of calcitonin gene-related peptide [24]. In the distal airways, including terminal bronchioles and alveoli, cells in the bronchoalveolar duct junction have been found to give rise to diverse populations of cells after injury [19]. Lung adenocarcinoma and bronchoalveolar carcinoma have been associated with stem cells from the bronchoalveolar duct junction region [19].
Stemness phenotype model theory suggests that all cancer cells have stem cell properties, and that the stemness is modulated by microenvironment [25]. This type of analysis suggests that stem cells should be constantly generated in order to not disappear after continuous passages. One of the main proposals of this theory is that all cancer cell phenotypes must originate from one mother cell. This clonal origin likely to be the case for the A549 cell line [26]. The stemness phenotype theory proposes that all CSCs and non-CSC should undergo symmetrical division and then comes the role of the microenvironment to change the phenotype of the cell [26]. It has been demonstrated that ectopic expression of miR-17-92, which is present in embryonic stem cells, promotes proliferation and influences differentiation of lung epithelial progenitor cells [20]. Matrixmetalloproteinase-10 (MMP10) (stromelysin 2) has been shown to have a role in the maintenance and tumorigenicity of mouse lung CSCs. Moreover, it has been shown that this tumorigenicity is independent of MMP10 in the tumor microenvironment. It is rather more dependent on MMP10 produced by cancer cells [27].
Epithelial-to-Mesenchymal Transition (EMT) is a complex program by which epithelial cells lose their characteristics, such as cell-to-cell adhesion, polarity and lack of motility. The cells also gain mesenchymal features, including motility and resistance to apoptosis. EMT role in acquisition of stem cell characteristics has been studied in mammary epithelial cells [9] and other solid tumors [28,29]. However, in the lung, EMT and CSCs are still not adequately evaluated. EMT signaling pathways, such as Wnt, Notch and Hedgehog, are also involved in normal and cancerous stem cell renewal and maintenance [30]. Our own work on TGF-β-induced EMT in NSCLC cells has demonstrated a role of sonic hedgehog (Shh) signaling [31]. Another study found that after four weeks of exposure of immortalized human bronchial epithelial cells to tobacco carcinogens, EMT was demonstrated by reduced expression of E-cadherin and increased expression of ZEB1 (Zinc finger E-box-binding homeobox 1) beside the change in cells morphology, while it took the cells 12 weeks to gain stem cell-like traits like CD44 high/CD24 low marker and spheroids formation [32].

Role of Microenvironment
Stem cell populations are established in niches, which represent the microenvironment that interacts with stem cells to regulate stem cell self-renewal and differentiation. It remains unclear whether CSCs form their own microenvironment or take advantage of the pre-existing tissue environment [33]. The solid microenvironment is comprised of different mesenchymal cell types that are recruited by cancer cells to enhance their survival, growth, invasion and dissemination. These include endothelial cells of the blood and lymphatic circulation, inflammatory cells, fibroblasts and others [34]. Fibroblasts are known to produce different chemokines that modulate tumor expansion, invasion, angiogenesis, and activation of extracellular matrixassociated cytokines and growth factors. Some of them are commonly associated with self-renewal [33,35].
It has been suggested that Mesenchymal Stem Cells (MSC) in the bone marrow travel to the tumor site [36] when attracted by chemoattractive molecules, such as SDF-1 (stromal cell-derived factor-1), SCF (stem cell factor)/c-kit, VEGF/VEGFR (vascular endothelial growth factor/vascular endothelial growth factor receptor) and others [37]. In the tumor sites, they play different roles, including contribution to blood vessel formation [38], immunomodulation [38] and promotion of metastatic behavior as established in breast cancer cells when mixed together with the MSCs [39]. However, MSC may suppress the tumor through downregulation of Wnt signaling pathway in breast cancer [40]. In addition to these cells, the solid microenvironment also contains factors secreted into the bloodstream, which also influences the tumor cell proliferation and differentiation. These include TGF-β, which is involved in maintenance of stem cell phenotypes and the EMT [41], Platelet-Derived Growth Factor (PDGF), fibroblast growth factor, insulin like growth factor-1, and epidermal growth factor [33].

Lung CSC Markers
Currently, there are relatively few CSC markers that have been validated. However, extensive studies have led to the identification

Airway Cells capable of stem-like behavior Cell response to injury Normal stem cell marker Associated cancer Cancer marker
Proximal airway (trachea and Bronchi) Basal cells [10] Self-renewal and multipotent abilities [11] Keratin 5 Keratin14 Transcription factor-63 Trp-63 [12] Squamous cell carcinoma Keratin 5 [13] Midlevel airway (Bronchioles) Clara cells [14] Multi-potent but not selfrenewal [14] Secretoglobin, scgb1a1 (CCSP or CC10) [12] Variant Clara cells [15] Multi-potent and selfrenewal [17] Lack cytochrome P4502F2 [15] Small cell lung cancer Calcitonin gene-related peptide Pulmonary Neuroendocrine Cells (PNEC) [16] Self-renewal [17] Calcitonin gene-related peptide [20] Distal airway (alveoli) Bronchoalveolar duct junction cell Self-renewal [18] CCSP [19] Bronchiolar adenocarcinoma of various CSCs that differ from other cells in the tumor. Most CSCs express multiple markers at the same time and using one marker to define CSC is not possible. For example in cell lines A549 and H446, CD133 positive and negative populations contain the same amount of CSCs [42]. Studying CSC markers may present new insight that will improve current lung cancer therapy and better patient prognosis. In this review we will highlight the current knowledge for some of the markers.

CD133
CD133 is the most frequently demonstrated marker for CSCs. It is a cell surface glycoprotein that consists of five transmembrane domains and two large glycosylated extracellular loops. Chen et al. [43] studied CD133 positive cells and CD133 negative cells from tissue samples of 10 NSCLC patients and cell lines A549, H1299, CCL-1, CCL-5 and C299. Results showed that CD133 positive cells displayed higher ability of self-renewal, tumor initiation and drug resistance. They also expressed higher level of Oct-4 (octamer-binding transcription factor 4) [43], which is a transcription factor expressed in embryonic stem cells [44]. Another study demonstrated that exposing A549 cells to cytotoxic concentration of cisplatin resulted in an eight-fold increase of CD133 cells fraction [45]. This study also confirmed the findings in vivo using mice with six different lung cancer xenografts that were treated with cisplatin, which supports the idea that chemotherapy is not effective in eliminating CD133 positive cancer initiating cells. Investigation of the CD133 in the human lung cancer cell lines A549, H157, Hee6, Calu-1, H292 and H446 showed that CD133 cell expressing stem cell characteristics were found only in H446 SCLC, but CD133 positive cells were not found in other NSCLC cell lines [46]. In another study, increased levels of CD133-expressing cells were found in NSCLC specimens, where they seemed to contribute to vasculogenesis [47]. However, in resected early stage NSCLC, CD133 had no prognostic value [48].

ALDH
Aldehyde Dehydrogenase (ALDH) is another marker connected with stem cell-like properties. Its enzymes are responsible for acetaldehyde oxidation and control the differentiation of normal stem cells [49]. ALDH-1 has been suggested as the specific marker for human lung adenocarcinoma [50]. Another study showed that ALDH1A1 positive lung cancer cells displayed resistance to gefitinib, an EGFR (epidermal growth factor receptor) tyrosine kinase inhibitor, compared to ALDH1A1 negative lung cancer cells [51]. Unlike CD133, increased cancer stem cell ALDH has been found to be helpful in tumor staging and provided prognostic information [52]. Another study demonstrated that there is also a correlation between ALDH1A1 and poor patient prognosis [53]. Analysis of various lung cancer cell lines and tumors reported that the majority of NSCLC comprise of a subpopulation of cells with increased ALDH activity, which is associated with the expression of ALDH1A1. Furthermore, lung cancer cells that expressed ALDH were shown to be highly tumorigenic and clonogenic in addition to being capable of self-renewal in comparison to lung cancer cells that do not express ALDH. The same study also revealed that there was increased expression of the Notch pathway transcript in ALDH + lung cancer cells. Suppressing the Notch pathway resulted in a considerable decrease in ALDH + cells, tumor cell proliferation and cell clonogenicity [53]. These insights suggest that ALDH functions in selecting for a subpopulation of self-renewing NSCLC stem-like cells with a greater possibility of being tumorigenic. Further exploration of the Notch pathway and the ALDH component in lung cancer may yield more insight into lung cancer stem cell maintenance.

ABCG2
ABCG2 is an ATP binding cassette transporter, which can pump chemotherapeutic drugs out of the cell, resulting in decreased concentrations of drugs inside the cell [54]. ABCG2 uses energy from ATP to drive the transport of substances across the cell membrane, including drug metabolites and other compounds. It is highly expressed in a subpopulation of stem cells and further research suggests that there is a link between ABCG2 and unfavorable prognosis in various tumors [55]. Moreover, recent studies have shown that ABCG2 is expressed in many cancer stem cells, including lung cancer. Evidence indicates that ABCG2 functions in protecting stem cells and increase the survival of hematopoietic stem cells in hypoxia [56]. In addition, further studies found that ABCG2 plays a cytoprotective role in cardiac SP cell populations placed in oxidative stress and they also contribute to the proliferative capacity of human hematopoietic progenitors in vitro and in vivo [57]. Overall, ABCG2 observes a role in preserving stem cells by contributing to their survival ability and proliferation.
Stem cell isolation from six human lung cancer cell lines (H460, H23, HTB-58, A549, H441 and H2170), using flow cytometry and Hoechst 33342 dye efflux assay, showed a side population with elevated expression of ABCG2 and increased resistance to drugs [6]. It has been recently demonstrated that Low Molecular Weight Heparin (LMWH) reduced lung side population cell colony formation ability and ABCG2 expression. Furthermore, combination of LMWH and cisplatin could overcome drug resistance and induce apoptosis [58].

CD44 and CD24
CD44, which is a frequently studied cell surface marker, is a hyaluronic acid receptor that is expressed by almost every tumor cell [59]. It promotes cell migration and is generally associated with proteins that are critical in regulating cell adhesion, proliferation, growth, migration and motility [60]. Moreover, CD44 is associated with signaling cascades that stimulate tumor initiation. The CD44 gene also experiences alternative splicing to encode different proteins in varying cancer subtypes, which makes it an ideal surface marker for isolating CSCs from carcinomas [61]. CD24 is a heat stable antigen that is another cell surface marker expressed in numerous tumors. It plays a role in cell-to-cell and cell-to-matrix interactions and is also involved in cell adhesion and metastasis, which suggests that it may be an essential marker in tumor prognosis and diagnosis [62]. It also enhances cell proliferation and stimulates tumor cells to adhere to fibronectin, collagen and laminin. High levels of CD24 are usually associated with tumor progression and metastasis. CD44 and CD24 have been utilized together or with other markers to isolate CSCs from solid tumors, though their usage is restricted to a particular few cancer types due to their lack of universal expression [63]. In vitro and in vivo studies of CD44 positive lung cancer cells showed the ability of these cells to initiate tumors [64]. Moreover, CD44 positive cells were also associated with EMT markers, such as SNAI1 (snail homolog 1), CDH2 (cadherin 2) and VIM (vimentin), which lend more evidence to the role EMT may play in maintaining stemness.

Role of Signaling Pathways in Lung CSCs
Hh pathway, wnt pathway and Notch pathway regulate proliferation and differentiation during embryogenesis. It was hypothesized that the same pathways that rule normal stem cell self-renewal could also rule cancer stem cell self-renewal. In the following section we discuss these pathways in the context of lung CSCs.

Role of Hh pathway
The Hedghog (Hh) signaling pathway is activated when the Shh ligand binds to transmembrane patched (PTCH1) proteins. In the absence of the ligand, PTCH1 represses Smo (smoothened), which prevents the activation of Hedgehog. However, after binding of the sonic hedgehog ligand to PTCH1, the interaction of PTCH1 and Smo is altered and Smo is no longer inhibited. This in turn activates several members of the GLI family of transcription factors. SCLCs maintain their malignant phenotype in vitro and in vivo through liganddependent Hedgehog pathway activation [65]. The hedgehog-signaling pathway has been also shown to be activated in NSCLC and correlates with histological type and prognostic parameters [66].
The Hh pathway regulates many basic processes, such as stem cell maintenance, cell differentiation and cell proliferation. Activation of the Hh pathway has been shown to lead to tumorigenesis in many carcinomas, including lung cancer. In addition, paracrine Hh signaling has also been shown to regulate proliferation of CSCs and enhance proliferation and tumor invasion [67,68]. Research suggests that tumor growth and propagation are dependent on a small group of CSCs with similar characteristics to normal tissue stem cells and are thus regulated by the same signaling molecules [69]. Hh signaling is associated with the regulation of self-renewal of CSCs in several cancers, including lung cancer. Furthermore, it has been shown to induce tumor metastasis by contributing to EMT, which is the transforming of polarized epithelial cells into mesenchymal cells, which are more prone to invasive growth and metastasis. Through the up regulation of transcription factor SNAIL and the down regulation of E-cadherin, Hh helps induce EMT in numerous cell lines [70]. This is in agreement with our earlier results with the cell line A549 where induction of EMT by TGF-β is marked by activation of Shh signaling [31]. Targeting the Hh signaling pathway may provide new therapeutic options for many different cancer types. Although the mechanisms behind the Hh signaling pathway is not completely understood, research has shown that aberrant Hh signaling does cause tumor growth and proliferation, increased tumor aggression and frequency of metastasis.

Role of Wnt pathway
Wnt signaling mediates many cellular processes, including proliferation, differentiation, apoptosis and motility [71]. The best understood Wnt signaling pathway is the canonical pathway, where Wnt ligands bind to a cell surface receptor complex causing the phosphorylation of disheveled family proteins (Dvl). The Dvl then activates GSK-3 (Glycogen Synthase Kinase 3) and CK1 (Casein Kinase 1), which mediates the degradation of β-catenin molecules, resulting in the accumulation of β-catenin in the cytoplasm. Some β-catenin is able to enter the nucleus and interact with TCF/LEF (Transcription Factor/ Lymphoid Enhancer-Binding Factor 1) family transcription factors to promote specific gene expression [72]. It has been shown that Wnt1 and Wnt2 are overexpressed in NSCLC cell lines and primary tumors [73].
Teng et al. [74] revealed regulation of a main regulator of selfrenewal and differentiation in embryonic stem cells, OCT-4, after treating lung cancer cell line A549 with cisplatin for two days. This was accompanied by an increased expression of β-catenin. When LiCl, a GSK-3b inhibitor, was used as an activator of the Wnt signaling pathway, an enhanced expression of β-catenin was evident. β-catenin accumulated in the cytoplasm and also translocated to the nucleus. Looking at the CSC characteristics, a dramatic increase in the proliferation, migration and clone formation abilities was observed.
On the other hand, when the pathway was blocked by knocking down β-catenin, a decrease was observed in the proliferation, clone formation, migration and drug resistance abilities.

Role of Notch pathway
Ligand proteins binding to the extracellular domain of Notch receptor induce proteolytic cleavage, including a final cleavage operated by a γ-secretase complex. This releases the active forms of Notch receptors, the intracellular Notch (Notch IC ), which translocate to the nucleus and form a complex with a DNA-binding protein resulting in the expression of various target genes. The four Notch family members may have very different and opposing activities in vivo [75][76][77], which explains why Notch pathway activation can have either oncogenic or tumor suppressor effects. A study found controversy between in vitro and in vivo inhibition of Notch using γ-secretase inhibitor MRK-003. In vitro, the Notch pathway inhibition decreased cancer cell line NCI-H1299 and NCI-H1435 clonogenic potential and this effect was reversed by expression of a constitutively active form of Notch3. In vivo, there was no evident effect on tumorigenicity from induced expression of a dominant-negative Notch pathway inhibitor. This supports the idea that different Notch isoforms have different roles and suppression of each of them differs from global suppression of Notch signaling [78].
Notch proteins caused a profound growth arrest in SCLC, according to one study [79]. On the other hand, several studies demonstrated the role of Notch-3 in lung cancer cell proliferation. Moreover, inhibition of the Notch3 pathway using a dominantnegative receptor dramatically decreased growth in soft agar [76,77]. Hypoxia increases expression of Notch direct downstream genes and the Notch IC domain interacts with Hypoxia Inducible Factor 1α (HIF-1alpha). Consequently, hypoxia, through the Notch signaling pathway, keeps the cell in the undifferentiated state [80].
Eliasz et al. [81] hypothesized a possible cross-talk between Notch-1 and IGF-1 signaling in lung adenocarcinoma. They demonstrated that that Notch-1-mediated pro-survival function in hypoxic lung adenocarcinoma through activation of Akt-1, which in turn inhibits Phosphatase and Tensin (PTEN) homolog expression and induces Insulin-like Growth Factor 1 Receptor (IGF-1R). As a result, hypoxia supports cancer stem cell survival and causes resistance to anticancer therapy.

Targeting CSC Biomarkers and Signaling Pathways for Therapy
The combination of anti-psychotic drug trifluoperazine with gefitinib has been demonstrated to overcome drug resistance in lung CSCs by targeting CD133/CD44 markers [82].
One study, using FISH in ALDH positive and negative A549, H460 and H1299 cell lines, showed that lung CSCs possess longer telomeres than their differentiated counterparts [83]. Using telomerase inhibitor MST312, it was demonstrated that the response to treatment differs according to the length of telomere, and cells with longer telomeres (H1299>H460, and CSCs>non-CSCs) were more sensitive to MST312mediated cell growth inhibition. A decrease in ALDH positive cells was shown after treatment.
Therapeutic targeting of the Notch pathway in ALDH positive cells resulted in a significant decrease in ALDH positive lung cancer cells, which implicates the Notch pathway in lung cancer stem cell maintenance [53]. According to another study, the inhibition of Wnt1  [84]. MRK-003, a gamma-secretase inhibitor, was proven to be able to inhibit Notch3 signaling, resulting in inhibition of lung cancer cell lines proliferation and increased apoptosis in vitro and in vivo using mouse xenograft models [85]. GDC-0449 is shown to be effective in inhibiting lung cancer cell lines HCC (adenocarcinoma) and H1339 (small-cell-lung carcinoma) by targeting hedgehog pathway in the side population cells in these cancers. Moreover, it enhanced the effect of cisplatin [86].
CSCs produce SCF that, by binding to c-kit, could stimulate CSC proliferation. Imatinib blocks the ability of SCF to activate c-kit receptors, thus resulting in a profound inhibition of CSC growth. In contrast to CSCs, imatinib had no detectable effect on bulk NSCLC cells due to the lack of the c-kit expression by the vast majority of these cells. As a result, cisplatin-imatinib treatment of NSCLC cells growing in vitro led to potent inhibition of both non-CSCs and CSCs [87]. All these studies point to an effective management of lung cancer subtypes through targeted approach against CSCs.

Conclusions
CSCs, which are subpopulation of cancer cells, have distinct markers and signaling pathways, play a major role in lung cancer initiation, resistance to chemotherapy and relapse after treatment. Several studies have been conducted to characterize as well as target this population of cells in order to better control the cancer recurrence after chemotherapy. However, several questions remain unanswered. These include better understanding of the exact role of the signaling pathways in normal and malignant stem cells, the interactions between cancer stem cells and the microenvironment and whether targeting the microenvironment will better control the cancer progression. Also, the translation of laboratory research to clinical applications and therapies is the next big challenge. The preliminary information is encouraging and further research on CSCs will broaden our knowledge of lung cancer cell biology and will ultimately improve therapeutic options for the effective clinical management of lung cancer.