Aldehyde dehydrogenase 1, a functional marker for identifying cancer stem cells in human nasopharyngeal carcinoma

Cancer Letters 330 (2013) 181–189

Contents lists available at SciVerse ScienceDirect

Cancer Letters journal homepage: www.elsevier.com/locate/canlet

Aldehyde dehydrogenase 1, a functional marker for identifying cancer stem cells in human nasopharyngeal carcinoma q Aibing Wu a,b,1, Weiren Luo a,c,1, Qianbing Zhang a, Zhixiong Yang b, Gong Zhang a, Siyi Li c, Kaitai Yao a,⇑ a

Cancer Research Institute, Southern Medical University, Guangzhou, PR China Cancer Center of Affiliated Hospital, Guangdong Medical College, Zhanjiang, PR China c Department of Pathology, Secondary Clinical College, Guangdong Medical College, Dongguan, PR China b

a r t i c l e

i n f o

Article history: Received 12 September 2012 Received in revised form 26 November 2012 Accepted 27 November 2012

Keywords: ALDH1 Nasopharyngeal carcinoma Cancer stem cell Epithelial–mesenchymal transition Prognosis

a b s t r a c t Aldehyde dehydrogenase 1 (ALDH1) activity has now been employed successfully as a cancer stem cells (CSCs) marker in various tumors. We wanted to investigate whether ALDH1 can be used as a putative CSCs marker and a powerful prognostic factor in nasopharyngeal carcinoma (NPC). Here, we isolated ALDH1positive cells from human NPC cell lines (5-8F and CNE2) and found that ALDH1-positive cancer cells grew faster and had higher clone formation efficiency (0.435 ± 0.15; 0.431 ± 0.025 vs. 0.131 ± 0.007; 0.121 ± 0.126), differentiation capability and had higher migration (233.00 ± 5.29; 228.60 ± 9.34 vs. 123.20 ± 7.70 vs. 97.20 ± 6.61) than those of ALDH1-negative cancer cells in vitro. In addition, ALDH1- positive cancer cells formed significantly more tumor spheres. Our in vivo experiments showed that only 5  103 ALDH1-positive NPC cells were required to induce tumors. Notably, ALDH1-positive cells are enriched in the side-population (SP) cells, and stem cells-like genes OCT4, BMI1, KLF4 and SOX2 are preferentially expressed in ALDH1-positive cells. Immunohistochemical results showed that the expression of ALDH1 correlated significantly with T classification (P = 0.011), N classification (P = 0.005), M classification (P = 0.002) and clinical stage (P = 0.001). Interestingly, ALDH1 expression in the tumor correlated significantly with epithelial–mesenchymal transition (EMT) markers including vimentin expression and loss of E-cadherin (P = 0.003 and 0.008, respectively). Furthermore, multivariate analyses showed that ALDH1 expression was an independent prognostic indicator (P = 0.032). Taken together, for the first time, we demonstrate that ALDH1 could be a novel stem cells marker and a valuable predictor of poor survival NPC. Ó 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Nasopharyngeal carcinoma (NPC) is one of the most common cancers in southern China; the incidence is up to 25 per 100,000. But this cancer is rare in Europe and the USA with an incidence of 0.5–2 per 100,000. Intermediate incidence is seen in Southeast Asia, the Mediterranean Basin, and the Arctic [1]. Mortality from this disease remains high because of the development of therapeutic resistance and distant metastasis. This result may be due to the existence of cancer stem cells (CSCs). The CSCs hypothesis suggests that a rather small subset of cells in tumors possess stem cell prop-

q Contributions: Participated in the conception and design of the study and the critical revision of the manuscript for important intellectual content: Y.K., W.A., L.W. Performed the data collection and analysis: W.A., L.W., Z.Q. Interpreted the data and produced the draft of the manuscript: W.A., L.W., Y.Z., L.S., Z.G. Obtained funding for the study: Y.K. All authors have read and approved the manuscript. ⇑ Corresponding author. Address: Cancer Research Institute, Southern Medical University, 1838 Guangzhou Road North, Guangzhou 510515, PR China. Tel./fax: +86 20 61648225. E-mail address: [email protected] (K. Yao). 1 These authors contributed equally to this work.

0304-3835/$ - see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.canlet.2012.11.046

erties. Recently, CSCs have been hypothesized to be responsible for tumorigenesis, therapeutic resistance, recurrence and distant metastasis [2,3]. The identification of these cells seems to be important for a better understanding of tumor progression and for the development of more effective therapies. Diverse cell surface markers have been used for the identification of CSCs in human tumors. Bonnet and Dick [4] first isolated CSCs using specific surface markers CD34+ and CD38 in patients with acute myeloid leukaemia. After implanting those cells into non-obese/severe-combined immunodeficient (NOD/SCID) mice, they declared that a minor subpopulation of CD34+/CD38 cells are capable of initiating tumor development, but cells with a typical leukemic phenotype CD34+/CD38+ are not capable of initiating tumor development in NOD/SCID mice. The existence of CSCs were extended to solid tumors, Al-Hajj et al. [5] found that only the population of CD44+/CD24/low lineage–cells could initiate the process of carcinogenesis in immunodeficient mice. Those cells possessed capacities of self-revival, differentiation, unlimited proliferation, and tumorigenesis, whereas CD44+CD24+ cells did not have the capacities. Subsequently, CSCs have been isolated and identified from a variety of malignancies, including brain cancer [6], prostate

182

A. Wu et al. / Cancer Letters 330 (2013) 181–189

cancer [7], lung cancer [8], melanoma [9], ovarian cancer [10], colorectal cancer [11,12], gastric cancer [13] and head and neck squamous cancer [14]. We have initially demonstrated that mouse nasopharyngx and human NPC tissues contained stem cells by label-retaining cell (LRC) approach [15]. Our current study focuses on the discovery of potential stem cell marker of NPC. Aldehyde dehydrogenase 1 (ALDH1) belongs to the aldehyde dehydrogenase superfamily which is responsible for the oxidation of aldehydes to their corresponding carboxylic acids [16,17]. Only recently, scholars have found high ALDH1 activity can be used to identify and isolate CSCs. ALDH1 was first used as marker of cancer stem cells in hematopoietic cells [18]. In solid tumors, Ginestier et al. discovered that ALDH1 could be enriched in cancer stem cell in breast cancer [19]. ALDH1-positive cells possess self-renewal, small cell tumorigenicity, and differentiation potential, which accounted for only 5% of the total number of breast cancer cells. More importantly, they found that the expression of ALDH1 was associated with poor clinical outcome of patients. Subsequently, several studies have shown that ALDH1 can also serve as stem cell markers of head and neck squamous carcinoma [20], lung cancer [21,22], prostate cancer [23], pancreatic cancer [24] and colorectal cancer [25]. However, there are no observations about the existence and function of ALDH1 activity in NPC thus far. Based on this background, the purpose of this study was to investigate the biological function of ALDH1 in vivo and vitro, and the relationship between ALDH1 and clinical outcomes in patients with NPC.

containing ALDH substrate (BAAA, 1 l mol/l per 1  106 cells) and incubated for 40 min at 37 °C. As a negative control, for each sample of cells an aliquot was treated with 50 mmol/l diethylaminobenzaldehyde (DEAB), a special ALDH inhibitor. The sorting gates were established using as negative controls the cells stained with PI only, for viability, the ALDEFLUOR-stained cells treated with DEAB and the staining with secondary antibody alone. 2.4. Cell proliferation Cell proliferation was analyzed using an MTT assay (Sigma, St. Louis, USA). Briefly, 1  103 cells were seeded into a 96-well plate with quadruplicate repeat for each condition. After 24 h of incubation, MTT reagent was added to each well and incubated for 4 h. The formazan crystals formed by viable cells were then solubilized in DMSO and measured at 490 nm for the absorbance (A) values. Each experiment was performed in triplicate. 2.5. Colony formation assay and tumor sphere culture About 200 cells were added to each well of a 6-well culture plate, and each cell group contained 2 wells. After 2 weeks of incubation, cells were washed twice with PBS and stained with Giemsa solution. The number of colonies containing P50 cells was counted under a microscope. The colony formation efficiency was calculated as: efficiency = (number of colonies/number of cells inoculated)  100%. Each experiment was performed in triplicates. Tumor sphere culture was performed as previously described [28]. Single-cells (1  104) were plated onto a 6-well ultra-low attachment plate (Corning, Corning, NY) in serum-free DMEM-F12, supplemented with 20 ng/mL basic fibroblast growth factor, 20 ng/mL epidermal growth factor, and B27 supplement (1:50 dilution; Invitrogen). After 7 days of culture, the number of tumorspheres formed (diameter >40 lm) was counted under an inverted microscope. Each experiment was performed in triplicates. 2.6. Transwell invasion assays

2. Materials and methods 2.1. Patients and tissue samples We collected 122 paraffin-embedded NPC samples from the Department of Pathology, the People’s Hospital of Gaozhou City, Affiliated Hospital of Guangdong Medical College, China. Subjects comprised 92 males and 30 females with ages ranging from 15 to 73 years. The clinical follow-up time of patients ranged from 8 to 92 months [26]. For the use of these clinical materials for research purposes, prior written informed consent from all the patients and approval from the Ethics Committees of the People’s Hospital of Gaozhou City were obtained. Histological classification and clinicopathologic staging of the samples was performed according to the rules of WHO histological classification. 2.2. Immunohistochemistry (IHC) Paraffin sections (3 lm) from the samples were deparaffinized in 100% xylene and re-hydrated in a descending ethanol series and water according to standard protocols. Heat induced antigen retrieval was performed in 10 mM citrate buffer for 2 min at 100 °C. Endogenous peroxidase activity and non-specific antigens were blocked with a peroxidase blocking reagent containing 3% hydrogen peroxide and serum, followed by incubation with rabbit anti-human ALDH1 antibody (Abcam, dilution: 1:250) overnight at 4 °C. After washing, the sections were incubated with biotin-labeled goat anti-rabbit antibody for 10 min at room temperature, and were subsequently incubated with streptavidin-conjugated horseradish peroxidase (HRP) (Maixin Inc, China). The peroxidase reaction was developed using 3, 3-diaminobenzidine chromogen solution in DAB buffer substrate. Sections were identified with DAB and counterstained with hematoxylin, mounted in neutral gum, and analyzed using a bright field microscope. The immunohistochemically stained tissue sections were reviewed and scored separately by two pathologists blinded to the clinical parameters. Expression of ALDH1 in the nucleus and in the cytoplasm was independently evaluated. For cytoplasmic staining, the score was evaluated according to the sum of cytoplasm staining intensity and the percentage of positive staining areas of cells. The staining intensity was scored as previously described [26,27]. Using this method of assessment, we evaluated the expression of ALDH1 in astrocytomas by staining index (scored as 0, 1, 2, 3, 4, 6, or 9). The staining index score = 9 was used to define tumors with high expression, staining index = 6 was used to define tumors with moderate expression and staining index 64 as tumors with low expression of ALDH1. 2.3. Aldefluor assay and separation of the ALDH1-positive and ALDH1-negative cells performed by FACS The identification of ALDH1 activity from 5-8F and CNE2 cells using the Aldefluor assay (Stem Cell Technologies, Durham, NC, USA) was followed by fluorescenceactivated cell sorting analysis. Cells were suspended in ALDEFLUOR assay buffer

Cells growing in the log phase were treated with trypsin and re-suspended as a single-cell solution. A total of 1  105 cells were seeded on a fibronectin-coated polycarbonate membrane inserted in a transwell apparatus (Corning Inc., Corning, NY). In the lower chamber, 600 ll of RPMI 1640 with 10% FBS was added as a chemoattractant. After the cells were incubated for 12 h, the insert was washed with PBS, and cells on the top surface of the insert were removed by a cotton swab. Cells adhering to the lower surface were fixed with methanol, stained with Giemsa, and counted under a microscope in five predetermined fields (100). All assays were independently repeated at least three times. 2.7. Animals’ model SCID mice (4–6 weeks) were purchased from the Model Animal Research Center of Nanjing University. All experiments were approved by the animal care committee of southern medical university. Five mice per group of scid mice were injected with ALDH1-positive, ALDH1-negative cells and unsorted parental cell at doses of 500, 5000, 104, and 105, respectively. The mice were observed for 6 weeks to allow tumor growth. Tumors were photographed. Tumor tissue was collected, fixed in 10% formaldehyde, and embedded in paraffinfor immunohistochemistry staining to assess tumor pathology. 2.8. Quantitative RT-PCR (qRT-PCR) To investigate the mRNA levels of stem cells and EMT-related genes, total RNA of ALDH1-positive and ALDH1-negative cells was reversely transcribed using PrimeScriptÒ RT reagent Kit (TaKaRa). Quantitative real-time PCR (qPCR) was performed using SYBRÒ Premix Ex Taq™ II (TaKaRa) on a StrataGene Mx3005p System. The primers (50 ? 30 ) were: OCT4 forward primer: GTGGAGAGCAACTCCGATG and reverse primer: TGCTCCAGCTTCTCCTTCTC; SOX2 forward primer: CGAGTGGA AACTTTTGTCGGA and reverse primer: TGTGCAGCGCTCGCAG; KLF4 forward primer: CCGCTCCATTACCAAGAGCT and reverse primer: ATCGTCTTCCCCTCTTTGGC; BMI1 forward primer: AAATGCTGGAGAACTGGAAAG and reverse primer: CTGTGGATGAGGAGACTGC. The data were normalized by using ARF5 as a reference gene (forward primer, ATCTGTTTCACAGTCTGGGACG; reverse primer, CCTGCTTGTTGGC AAATACC). 2.9. Immunofluorescence assay Cells were grown on glass coverslips for 24 h at 37 °C, washed twice with PBS, and fixed in 4% paraformaldehyde (Sigma) for 30 min at room temperature. The cells were washed two additional times with PBS and incubated with 1% bovine serum albumin for 30 min at room temperature. After washing with PBS, cells were incubated 2 h at 4 °C with ALDH1 (Abcam, dilution 1:200), ABCG2 (Santa, dilution 1:50), CK5/6 (Santa, 1:200), OCT4 (Santa, dilution 1:50), Bmi-1 (Santa, dilution 1:100), KLF (Santa, dilution 1:50), SOX2 (Santa, dilution 1:300), E-cadherin (BD,

A. Wu et al. / Cancer Letters 330 (2013) 181–189 dilution 1:500), vimentin (BD, dilution 1:400). After washing with PBS, cells were incubated 60 min at 4 °C in the dark with goat anti-rabbit IgG and goat anti-mouse IgG (Multisciences, China). After extensive washing with PBS, cell nuclei were counterstained with DAPI (Boisynthesis biotechnology, Beijing, China). The results were observed and representative images were photographed using a FV300 confocal microscope (Olympus, Tokyo, Japan). 2.10. Side-population (SP) analysis Cells growing in the log phase were treated with 0.25% trypsin, washed twice with PBS, resuspended in ice-cold RPMI 1640 culture with 2% FBS, at a concentration of 1  106 cells/mL. Hoechst 33342 dye was then added at a final concentration of 5 Ag/mL and incubated at 37 °C in a 5% CO2 for 90 min in the dark with intermittent shaking. Following this incubation, cells were washed with ice-cold PBS, stained with propidium iodide (1 lg/ml), and maintained at 4 °C for flow cytometry analyses and for sorting of side population (SP) fraction using a FACSAria Flow cytometer (Beckton Dickson). The Hoechst dye was excited with an UV laser at 351–364 nm, and its fluorescence was measured with a 515-nm side population filter (Hoechst blue) and a 608 EFLP optical filter (Hoechst red). A 540 DSP filter was used to separate the emission wave lengths. 2.11. Statistical analysis All data were analyzed for statistical significance using SPSS 13.0 software (SPSS Inc., Chicago, IL). ALDH1-positive and ALDH1-negative cells with two-tailed student’s t test for statistical significance. One-way ANOVA was used to determine the differences between the groups for tumor spheres culture. The Mann–Whitney U test was applied examine the relationship between ALDH1 expression levels and clinicopathologic characteristics. Survival analysis was performed using the Kaplan–Meier method. The Multivariate Cox proportional hazards method was used to analyze the relationship between the variables and patient’s survival time. A P value of less than 0.05 was considered statistically significant.

3. Results 3.1. 5-8F and CNE2 cell lines contain a small fraction of ALDH1positive cells Using the Aldefluor assay and the fluorescence-activated cell sorting (FACS) analysis, we isolated ALDH1-positive and ALDH1negative cells from 5-8F and CNE2 cell lines. Results showed that there was only a small fraction of ALDH1-positive cells (1.96 ± 0.69%, n = 8; 1.97 ± 0.16%, n = 8; respectively) in parental NPC cells (Fig. 1A). 3.2. ALDH1-positive cells feature stem cell-Like characteristics in vitro To study the biological function of ALDH1-positive and ALDH1negative cells in vitro, unsorted parental cells, ALDH1-positive and ALDH1-negative cells were cultured for 7 days utilizing an MTT assay, and found ALDH1-positive cells grew more rapidly than unsorted parental cells and ALDH1-negative cells lines (P < 0.05; Fig. 1B). The results indicated that ALDH1-positive cells possessed stronger proliferative capacity. The clonogenic potential of ALDH1-positive and ALDH1-negative cells and their individual proliferative capacity were tested by plating the cells. After 14 days of culture, most clones had reached >50 cells. We counted the clone number and found that ALDH1-positive cells formed more colonies than those of ALDH1negative cells. Statistical analysis showed significant differences in the number of colonies between the ALDH1-positive cells and ALDH1-negative cells (87.33 ± 2.51; 86.33 ± 5.03 vs. 26.33 ± 1.52; 24.33 ± 2.51). Furthermore, ALDH1- positive cells formed colonies that were significantly larger than ALDH1-negative cells (P < 0.05; Fig. 1C). Tumor sphere is a method of identification and isolation of cancer stem cell. We cultured cells under serum-free medium with bFGF, EGF and B27 and found that ALDH1-positive cells formed significantly more tumor spheres than unsorted parental cells and ALDH1-negative cells (P < 0.05; Fig. 1D). We next analyzed the expression of cytokeratin 5/6(CK-5/6) protein in ALDH1-positive

183

cells and ALDH1-negative cells by immunofluorescence (IF). We found that the expression of CK5/6 in ALDH1-positive cells was significantly higher than in ALDH1-negative cells (Supplementary Fig. 1). These finding indicated that ALDH1-positive cancer cells possessed the capacities of proliferation, self-renewal, and differentiation. Cell migration is a key step during tumor development and metastasis. We tested the ability of ALDH1-positive and ALDH1negative cells to migrate through the 8 lm pores on the polycarbonate membrane. Compared with ALDH1-negative cells, ALDH1positive cells showed higher invasion activity (P < 0.05; Fig. 1E). We detected expression of ALDH1 in tumor spheres by immunocytochemistry (Fig. 1F). We found that ALDH1 is high expression in tumor spheres. 3.3. Tumor formation and self-renewal of ALDH1-positive cancer cells in vivo To evaluate the tumorigenic capability of ALDH1-positive, ALDH1-negative cells and parental cells, five mice per group of scid mice were injected with ALDH1-positive, ALDH1-negative cells and unsorted parental cell at doses of 500, 5000, 104, and 105, respectively. The results showed that 5000 ALDH1-positive cells in xenotrans-planted mice all resulted in the generation of visible tumors 6 weeks after injection, but 105 ALDH1-negative cells did not induce tumor formation (Fig. 2A, Table1). We also found that the dose of 1  105 ALDH1-positive cancer cells yielded larger tumors than parental cells (5-8F and CNE2) (1.06 ± 0.14, 1.09 ± 0.13 cm3 vs. 0.256 ± 0.039, 0.264 ± 0.054 cm3) (P < 0.05). (Fig. 2B) This indicated that ALDH1-positive cells possessed a higher tumorigenic ability than ALDH1-negative cells. In addition, we analyzed the expression of ALDH1 in primary xenograft originating from ALDH1-positive cells and parental unsorted cells by immunohistochemistry. The results showed that the expression of ALDH1 protein in primary xenograft originating from ALDH1-positive cells was higher than parental unsorted cells (Fig. 2C). 3.4. Stem cell-related genes are preferentially expressed in ALDH1positive cells To further determine whether ALDH1-positive cells might represent more primitively in NPC, the expression levels of some stem cell-associated genes in ALDH1-positive and ALDH1-negative cells were evaluated. Real-time PCR analysis indicated that the expression levels of OCT4, BMI1, KLF4 and SOX2 mRNA in ALDH1-positive cells were significantly higher than ALDH1-negative cells (Fig. 2D). Further immunofluorescence analysis revealed that the expression levels of OCT4, BMI1, KLF4 and SOX2 were up-regulated in ALDH1positive cells than ALDH1-negative cells (Fig. 2E). 3.5. ALDH1-positive cells are enriched in the side-population (SP) cells It has been reported that SP cells could enrich stem cells of nasopharyngeal carcinoma. The Bcrp1/ABCG2 gene is an important determinant of the SP phenotype [29]. To investigate the relationship between ALDH1-positive cells and SP cells, real-time PCR analysis revealed that the expression levels of ALDH1 in SP cells were higher than those in non-SP cells. Similarly, the expression levels of ABCG2 in ALDH1-positive cells were higher than that in ALDH1-negative cells (Fig. 3A). Immunofluorescence analysis showed that the cytoplasm of ALDH1-positive cells had strong immunoreactivity for ABCG2 when compared to that in ALDH1negative cells (Fig. 3B). Compared with non-SP cells cells, SP cells also had strong immunoreactivity for ALDH1 (Fig. 3B. We also found that SP cells represented about (24.90% ± 0.36 vs. 24.17% ± 0.32) of ALDH1-positive cells (Fig. 3C). Theses results

184

A. Wu et al. / Cancer Letters 330 (2013) 181–189

Fig. 1. ALDH1-positive nasopharyngeal carcinoma (NPC) cells have tumor stem cell properties in vitro. (A) Analyzing and sorting ALDH1-positive and ALDH1-negative cells from 5-8F and CNE2 cell lines via FACS. Cell were incubated with ALDEFLUOR substrate (BAAA) and the specific inhibitor of ALDH, DEAB, was used to establish the baseline fluorescence of these cells (P1) and to define the ALDEFLUOR positive region (P2). (B) Cell growth curve of unsorted parental cells, (C) analysis of cell colony numbers in clonogenicity assays of ALDH1-positive and ALDH1-negative. ALDH1-positive cell populations formatted larger and more colonies compared with ALDH1-negative cells. The assay was performed three times. (D) Analysis of the formation abilities of sphere bodies from various groups under serum-free medium with bFGF, EGF and B27. Original magnification: 100 and 200. (E) The migrating capability of ALDH1-positive and ALDH1-negative cells was examined by Transwell chamber assay. (F) Immunofluorescence staining reveals the expression of ALDH1 in tumor sphere. Asterisk indicates value compared with that of control with P < 0.05.

indicated the overlap between the ALDH1-positive cells and SP cells. 3.6. ALDH1 expression and the correlation with EMT indicators in NPC As shown in Fig. 4A, immunohistochemical results showed that the localization of ALDH1 was observed mainly in the cytoplasm of primary cancer cells, among which, 6.6% (8/122) was negative expressed, 86.1% (105/122) had low to moderate expressed, and 7.3% (8/122) showed strong signal (Supplementary Table1). Furthermore, we found that ALDH1 was highly expressed in the invasive front of tumors, spindle-shaped cells infiltrating into the surrounding stroma and invading into blood vasculature. It was noted that in most cases, tumor cells exhibited strongly ALDH1 expression, anomalous positivity of vimentin and low expression of E-cadherin (Fig. 4C). High ALDH1 expression was closely associated with high vimentin expression (Spearman correlation coefficient, 0.240; P = 0.008) and low E-cadherin expression (Spearman correlation coefficient, 0.264; P = 0.003). Moreover, immunofluorescence and Quantitative RT-PCR analysis indicated that the expression of vimentin was overexpressed and E-cadherin expression was aberrantly inhibited in ALDH1-positive cells compared with ALDH1-negative cells (Supplementary Fig. 2). 3.7. Correlation of ALDH1 expression with clinicopathologic characteristics in NPC The relationship between clinicopathologic characteristics and ALDH1 expression levels in individuals with NPC are summarized in Table 2. The expression levels of ALDH1 were not found to be

significantly associated with patient’s age, sex and Histologic subtype. On the contrary, the expression of ALDH1 was positively correlated with T classification (P = 0.011), lymph node metastasis (N classification) (N0–N1 vs. N2–N3) (P = 0.005), M classification (P = 0.002) and clinical stage (I–II vs. III–IV) (P = 0.001) in NPC patients. 3.8. Survival analysis To investigate the prognostic value of ALDH1 expression for NPC, we assessed the association between the expression levels and patient survival using Kaplan–Meier analysis with the log-rank test. In 122 NPC cases with prognosis information, we observed that the level of ALDH1 protein expression was negatively correlated with the overall survival of NPC patients (Fig. 4B). Patients with higher levels of ALDH1 expression had poorer survival rates than those with lower levels of ALDH1 expression (P < 0.001). In addition, T/N/M classifications and clinical stages were also significantly correlated with patients’ survival (P < 0.05, all). To determine independent prognostic factors for survival in NPC, multivariate analysis was subsequently performed. Our results indicated that ALDH1 expression was an independent prognostic factor for NPC (P = 0.032). 4. Discussion The CSCs concepts elucidate not only the issue of tumor initiation and development, tumor’s ability to metastasize and reoccur, but also the ineffectiveness of conventional cancer therapy. Therefore, treatment directed at eliminating cancer stem cells or induc-

A. Wu et al. / Cancer Letters 330 (2013) 181–189

185

Fig. 2. The ALDH1-positive cell population xenografted in NOD/SCID mice has cancer stem cell properties. (A). ALDH1-positive and unsorted parental cells (5-8F and CNE2) were implanted into SCID mice. After 6 weeks, the dose of 105 ALDH1-positive cancer cells yielded larger tumors (red arrow), whereas the same dose of unsorted parental cell generated a small tumor mass (blue arrow). (B) The size of tumors were evaluated in ALDH1+cells and unsorted parental cells. (C) Immunohistochemistry (IHC) staining for ALDH1 of a representative primary xenograft originating from ALDH1-positive cells and parental NPC cells. (D) Relative mRNA expression levels of OCT-4, BMI-1, KLF4 and SOX2 in ALDH1-positive and ALDH1negative cells populations were determined by real time PCR. (E) FACS-sorted ALDH1-positive cells and ALDH1-negative cells were analyzed with the indicated antibodies by immunofluorescence (Green, antibody staining; blue, DAPI). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1 Tumorigenesis of ALDH1-positive, ALDH1-negative cells and parental unsorted cells in SCID mice after 6 weeks by subcutaneous injection. Cell number 5-8F cells 5-8F ALDH1-positive cells 5-8F ALDH1-negative cells CNE2 cells CNE2 ALDH1-positive cells CNE2 ALDH1-negative cells

500 0/6 0/6 0/6 0/6 0/6 0/6

5000 0/6 6/6 0/6 0/6 6/6 0/6

1  104 0/6 6/6 0/6 0/6 6/6 0/6

1  105 6/6 6/6 0/6 6/6 6/6 0/6

ing their proper differentiation may be an effective way to cure cancer. However, up to now, there are few articles that are involved in the investigation of nasopharyngeal normal and cancer stem cells [15,30]. There is an urgent need for new molecular markers to isolate and identify stem cells in NPC.

In this study, we for the first time successfully isolated ALDH1positive cells from nasopharyngeal carcinoma cells and then evaluated the biological function of ALDH1-positive cells by in vitro and in vivo approaches. Our in vitro experiments showed that ALDH1positive cancer cells grew faster and had higher clone formation efficiency, differentiation capability and had higher migration than those of ALDH1-negative cancer cells. We also found that ALDH1positive cancer cells formed significantly more tumor spheres than unsorted parental cells and ALDH1-negative cancer cells. Taken together, these findings implicate that ALDH1-positive cells display higher properties of cancer stem cell-like cells, including high capacities for proliferation, self-renewal, differentiation and migration in vitro. It is generally accepted that high tumorigenic capability is the gold standard for evaluating tumorigenicity of CSCs. Ginestier et al. [19] have demonstrated that 500 ALDH1-positive cells are

186

A. Wu et al. / Cancer Letters 330 (2013) 181–189

Fig. 3. The correlation between ALDH1-positive cell and SP cells. (A) Analysis of ALDH1 and ABCG2 genes expression by Real-time PCR. (B) FACS-sorted ALDH1-positive cells and ALDH1-negative cells were analyzed with ALDH1 and ABCG2 antibodies by Immunofluorescence. Green, antibody staining; blue, DAPI. (C) Analyzing and sorting ALDH1positive and ALDH1-negative cells from side population (SP) cells and non side population (non-SP) cells via FACS. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

capable of tumor initiation in NOD-SCID mouse. Similarly, our in vivo experiments showed that only 5  103 ALDH1-positive NPC cells were required to induce tumors, whereas 1  105 ALDH1-negative cells did not, indicating that the tumor formation ability of ALDH1-positive cells was significantly greater than that in the latter. Therefore, we suggest that this population possesses properties classically attributed to cancer stem cell-like cells. Common molecules have been demonstrated to exist between CSCs and embryonic stem cells (ESCs) because both of them have the capacity of self-renewing [31,32]. To further validate the capacities of cancer stem cell-like cells in ALDH1-positive cells of NPC cells lines, we investigated the expression of stem cell-associated genes by real-time PCR and immunofluorescence. Our results showed that the expressions of OCT4, SOX2, BMI1 and KIF4 in ALDH1-positive cells were significantly increased than ALDH1negative cells, suggesting ALDH1-positive cells have more stem cells-like properties. A subset of stem cells, termed the ‘‘side population’’ (SP), have been isolated from stem cells of various normal tissues [29] as well as in tumor cells [33,34], and enrich the stem cell population. Therefore, SP analysis can be a useful tool to identify CSCs. In few years, Wang et al. reported that SP cells in human NPC cell line CNE-2 had stem cell characteristics in vitro and also had a strong ability to form tumors in vivo [30]. This data indicates that SP cells could be enriched in stem cells of NPC. Given this evidence, we further analyzed the relationship between ALDH-positive cells and SP cells. More interestingly, our results showed that the expression levels of ALDH1 in SP cells were higher than those in non SP cells. On the other hand, SP cells represented more population of ALDH1-

positive cells compared with non SP cells in NPC cell lines. Our data demonstrate the positive relationship between the ALDH1-positive cells and SP cells. ABCG2, the ATP-binding cassette transporter protein, has been proved to be associated with the SP phenotype [29,33,35,36]. In the present study, it is of note that the expression levels of ABCG2 were highly exhibited in ALDH1-positive cells. Consequently, these data strongly suggest that ALDH1-positive cells should be targeted during the treatment of NPC patients because of inherently high resistance to chemotherapy. To our knowledge, the correlation between ALDH1-positive cells and SP cells in tumor cells is not fully understood. To confirm ALDH1 expression and its prognostic significance in patient specimens, immunohistochemical analysis of 122 NPC samples was performed. Consistent with other group studies, ALDH1 protein was found to be expressed in the cytoplasm of tumor cells [22,23]. Interestingly, we found that ALDH1 was highly expressed in neoplastic spindle cells, which often present at the invasive tumor fronts, or infiltrate into the surrounding tissues. Recently, we have showed that these spindle-shaped cells possessed invasive and metastasis properties with epithelial–mesenchymal transition (EMT) [37]. It is now widely accepted that EMT plays an important role during tumor invasion and metastasis, and cancer cells acquire a more aggressive phenotype via EMT [38,39]. More importantly, we here showed that the expression of ALDH1 correlated significantly with EMT-associated proteins including E-cadherin and vimentin expression. In vitro, immunofluorescence and quantitative RT-PCR analysis showed that E-cadherin expression was greatly inhibited, while vimentin expression was strongly up-regulated in ALDH1-positive cells compared with ALDH1-nega-

A. Wu et al. / Cancer Letters 330 (2013) 181–189

187

Fig. 4. Expression of ALDH1 protein and the association with EMT indicators in 122 NPCs. (A) Different expression levels of ALDH1 were examined in NPC samples. (B) Kaplan–Meier survival analysis of overall survival according to ALDH1 expression. Moderate/high ALDH1 expression group had significantly worse survival than negative/low expression group. The log-rank test was used to calculate P values. (C) ALDH1 was highly expressed in the invasive fornt, spindle cells and tumor cells of blood vessel. One representative case (No. 30115) showed that ALDH1 expression correlated negatively with E-cadherin expression, positively with vimentin expression. (All images, 400.)

tive cells. As expected, our results showed that ALDH1 expression correlated closely with the aggressive behaviors of NPC, such as tumor growth, lymph node metastasis and clinical stage. For example, ALDH1 protein was highly expressed in patients with T3–T4 stage compared with those in T1–T2 stage, suggesting these populations could be endowed with highly motile and migratory potentials to infiltrate into the surrounding tissues. Therefore, we speculate that might be endowed with migratory and invasive capabilities due to EMT. In fact, some studies have demonstrated a correlation between CSCs and EMT [40,41]. Interestingly, we found that high ALDH1 expression was associated significantly

with distant metastasis. It is well known that NPC patients generally tend to have distant metastasis and have a worse prognosis compared to others without distant metastasis [42]. Combined with in vitro and vivo assays described above, we hypothesize that EMT plays an essential role in endowing migratory CSCs with the more migratory and invasive potentials to spread from the primary site to lymph nodes or to other distant organs. Finally, we showed in the present study that ALDH1 expression was associated inversely with patients’ overall survival. Patients with higher expression of ALDH1 protein had a shorter overall survival time. Most importantly, multivariate analyses showed that

188

A. Wu et al. / Cancer Letters 330 (2013) 181–189

Table 2 Correlation between the clinicopathologic characteristics and expression of ALDH1 protein in NPC. Characteristics

Cases

ALDH1 expression Negative/low grade (%)

Moderate/high grade (%)

P

Gender Male Female

92 30

55 (59.8) 19 (63.3)

37 (40.2) 11 (36.7)

0.730

Age (years) <50 P50

74 48

48 (64.9) 26 (54.2)

26 (35.1) 22 (45.8)

0.237

Histologic subtype DNKC 19 UDC 103

9 (47.4) 63 (62.1)

10 (52.6) 39 (37.9)

0.228

T classification T1–T2 T3–T4

58 64

41 (70.7) 33 (51.6)

17 (29.3) 31 (48.4)

0.011

N classification N0–N1 N2–N3

70 52

50 (71.4) 24 (46.2)

20 (28.6) 28 (53.8)

0.005

M classification M0 M1

107 15

67 (62.6) 3 (20.0)

40 (37.4) 12 (80.0)

0.002

Clinical stage I–II III–IV

33 89

28 (84.8) 46 (51.7)

5 (15.2) 43 (38.3)

0.001

DNKC, differentiated nonkeratinizing carcinoma; UDC, undifferentiated carcinoma; T, tumor size; N, lymph node; M, distant metastasis; ALDH1, Aldehyde dehydrogenase 1.

increased expression of ALDH1 was a significant predictor of poor prognosis for NPC patients. Consistent with our findings, ALDH1 expression has been considered as a valuable prognostic biomarker in breast cancer [19] and lung cancer [22]. Therefore, our findings imply that detection of the ALDH1 expression might be a useful biomarker for identifying a poor prognosis in NPC. ALDH1 may also provide a therapeutic target for developing specific agents for eradiating NPC cancer stem cell-like cells and thus could potentially yield efficient therapeutic approaches. In conclusion, we show first that these populations of ALDH1positive cells performed by Aldefluor assay are endowed with extensive proliferation, capable of self-renewal and of generating tumors. Therefore, ALDH1-positive cells possess the properties of cancer stem cell-like cells. In addition, we also demonstrate ALDH1 expression as an independent prognostic factor. These findings provide a novel tool for the study of NPC cancer stem cell-like cells and a potential therapeutic target for treatment of NPC patients. Acknowledgments This work was supported by funding from National Natural Science Foundation of China (NSFC)-Guangdong Joint Fund [u0732006] and National Natural Science Foundation of China [81201672 and 81202125]. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.canlet.2012. 11.046. References [1] L. Zhang, Y.X. Zhu, Y. Wang, C.P. Huang, Y. Wu, Q.H. Ji, Salvage surgery for neck residue or recurrence of nasopharyngeal carcinoma: a 10-year experience, Ann. Surg. Oncol. 18 (2011) 233–238 (PubMed: 20737217).

[2] H. Clevers, The cancer stem cell: premises, promises and challenges, Nat. Med. 17 (2011) 313–319 (PubMed: 21386835). [3] S.M. Cabarcas, L.A. Mathews, W.L. Farrar, The cancer stem cell niche – there goes the neighborhood?, Int J. Cancer 129 (2011) 2315–2327 (PubMed: 21792897). [4] D. Bonnet, J.E. Dick, Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell, Nat. Med. 3 (1997) 730– 737 (PubMed: 9212098). [5] M. Al-Hajj, M.S. Wicha, A. Benito-Hernandez, S.J. Morrison, M.F. Clarke, Prospective identification of tumorigenic breast cancer cells, Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 3983–3988 (PubMed: 12629218). [6] S.K. Singh, C. Hawkins, I.D. Clarke, J.A. Squire, J. Bayani, T. Hide, R.M. Henkelman, M.D. Cusimano, P.B. Dirks, Identification of human brain tumour initiating cells, Nature 432 (2004) 396–401 (PubMed: 15549107). [7] A.T. Collins, P.A. Berry, C. Hyde, M.J. Stower, N.J. Maitland, Prospective identification of tumorigenic prostate cancer stem cells, Cancer Res. 65 (2005) 10946–10951 (PubMed: 16322242). [8] C.F. Kim, E.L. Jackson, A.E. Woolfenden, S. Lawrence, I. Babar, S. Vogel, D. Crowley, R.T. Bronson, T. Jacks, Identification of bronchioalveolar stem cells in normal lung and lung cancer, Cell 121 (2005) 823–835 (PubMed: 15960971). [9] D. Fang, T.K. Nguyen, K. Leishear, R. Finko, A.N. Kulp, S. Hotz, P.A. Van Belle, X. Xu, D.E. Elder, M. Herlyn, A tumorigenic subpopulation with stem cell properties in melanomas, Cancer Res. 65 (2005) 9328–9337 (PubMed: 16230395). [10] A.A. Wani, N. Sharma, Y.S. Shouche, S.A. Bapat, Nuclear–mitochondrial genomic profiling reveals a pattern of evolution in epithelial ovarian tumor stem cells, Oncogene 25 (2006) 6336–6344 (PubMed: 16732329). [11] C.A. O’Brien, A. Pollett, S. Gallinger, J.E. Dick, A human colon cancer cell capable of initiating tumour growth in immunodeficient mice, Nature 445 (2007) 106– 110 (PubMed: 17122772). [12] M. Wang, J. Xiao, J. Jiang, R. Qin, CD133 and ALDH may be the molecular markers of cholangiocarcinoma stem cells, Int. J. Cancer 128 (2011) 1996– 1997 (PubMed: 20568110). [13] Y.Y. Zhan, J.P. He, H.Z. Chen, W.J. Wang, J.C. Cai, Orphan receptor TR3 is essential for the maintenance of stem-like properties in gastric cancer cells, Cancer Lett., 2012. http://dx.doi.org/pii:S0304-3835(12)00580-0 (PubMed: 23043761). [14] M.E. Prince, R. Sivanandan, A. Kaczorowski, G.T. Wolf, M.J. Kaplan, P. Dalerba, I.L. Weissman, M.F. Clarke, L.E. Ailles, Identification of a subpopulation of cells with cancer stem cell properties in head and neck squamous cell carcinoma, Proc. Natl. Acad. Sci. U.S.A. 104 (2007) 973–978 (PubMed: 17210912). [15] H.B. Zhang, C.P. Ren, X.Y. Yang, L. Wang, H. Li, M. Zhao, H. Yang, K.T. Yao, Identification of label-retaining cells in nasopharyngeal epithelia and nasopharyngeal carcinoma tissues, Histochem. Cell Biol. 127 (2007) 347–354 (PubMed: 17139514). [16] N.E. Sladek, Human aldehyde dehydrogenases: potential pathological, pharmacological, and toxicological impact, J. Biochem. Mol. Toxicol. 7 (2003) 7–23 (PubMed: 12616643). [17] S.A. Marchitti, C. Brocker, D. Stagos, V. Vasiliou, Non-P450 aldehyde oxidizing enzymes: the aldehyde dehydrogenase superfamily, Expert Opin. Drug. Metab. Toxicol. 4 (2008) 697–720 (PubMed: 18611112). [18] D.J. Pearce, D. Taussig, C. Simpson, K. Allen, A.Z. Rohatiner, T.A. Lister, D. Bonnet, Characterization of cells with a high aldehyde dehydrogenases activity from cord blood and acute myeloid leukemia samples, Stem Cells 23 (2005) 752–760 (PubMed: 15917471). [19] C. Ginestier, M.H. Hur, E. Charafe-Jauffret, F. Monville, J. Dutcher, M. Brown, J. Jacquemier, P. Viens, C.G. Kleer, S. Liu, A. Schott, D. Hayes, D. Birnbaum, M.S. Wicha, G. Dontu, ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome, Cell Stem Cell. 1 (2007) 555–567 (PubMed: 18371393). [20] Y.C. Chen, Y.W. Chen, H.S. Hsu, L.M. Tseng, P.I. Huang, K.H. Lu, D.T. Chen, L.K. Tai, M.C. Yung, S.C. Chang, H.H. Ku, S.H. Chiou, W.L. Lo, Aldehyde dehydrogenase 1 is a putative marker for cancer stem cells in head and neck squamous cancer, Biochem. Biophys. Res. Commun. 385 (2009) 307–313 (PubMed: 19450560). [21] D. Ucar, C.R. Cogle, J.R. Zucali, B. Ostmark, E.W. Scott, R. Zori, B.A. Gray, J.S. Moreb, Aldehyde dehydrogenase activity as functional marker for lung cancer, Chem. Biol. Interact. 178 (2009) 48–55 (PubMed: 18952074). [22] F. Jiang, Q. Qiu, A. Khanna, N.W. Todd, J. Deepak, L. Xing, H. Wang, Z. Liu, Y. Su, S.A. Stass, R.L. Katz, Aldehyde dehydrogenase 1 is a tumor stem cell-associated marker in lung cancer, Mol. Cancer Res. 7 (2009) 330–338 (PubMed: 19276181). [23] T. Li, Y. Su, Y. Mei, Q. Leng, B. Leng, Z. Liu, S.A. Stass, F. Jiang, ALDH1A1 is a marker for malignant Prostate stem cells and predictor of prostate cancer patients’ outcome, Lab Invest. 90 (2010) 234–244 (PubMed: 20010854). [24] C. Li, D.G. Heidt, P. Dalerba, C.F. Burant, L. Zhang, V. Adsay, M. Wicha, M.F. Clarke, D.M. Simeone, Identification of pancreatic cancer stem cells, Cancer Res. 67 (2007) 1030–1037 (PubMed: 17283135). [25] E.H. Huang, M.J. Hynes, T. Zhang, C. Ginestier, G. Dontu, H. Appelman, J.Z. Fields, M.S. Wicha, B.M. Boman, Aldehydedehydrogenase1 Is a marker for normal and malignant human colonic stem cells (SC) and tracks SC Cover population during colon tumorigenesis, Cancer Res. 69 (2009) 3382–3389 (PubMed: 19336570). [26] W.R. Luo, A.B. Wu, W.Y. Fang, S.Y. Li, K.T. Yao, Nuclear expression of Ncadherin correlates with poor prognosis of nasopharyngeal carcinoma, Histopathology 61 (2012) 237–246 (PubMed: 22385354).

A. Wu et al. / Cancer Letters 330 (2013) 181–189 [27] W. Luo, W. Fang, S. Li, K. Yao, Aberrant expression of nuclear vimentin and related epithelial–mesenchymal transition markers in nasopharyngeal carcinoma, Int. J. Cancer 131 (2012) 1863–1873 (PubMed: 22307379). [28] G. Dontu, W.M. Abdallah, J.M. Foley, K.W. Jackson, M.F. Clarke, M.J. Kawamura, M.S. Wicha, In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells, Genes Dev. 17 (2003) 1253–1270 (PubMed: 12756227). [29] S. Zhou, J.D. Schuetz, K.D. Bunting, A.M. Colapietro, J. Sampath, J.J. Morris, I. Lagutina, G.C. Grosveld, M. Osawa, H. Nakauchi, B.P. Sorrentino, The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype, Nat. Med. 7 (2001) 1028–1034 (PubMed: 11533706). [30] J. Wang, L.P. Guo, L.Z. Chen, Y.X. Zeng, S.H. Lu, Identification of cancer stem cell-like side population cells in human nasopharyngeal carcinoma cell line, Cancer Res. 67 (2007) 3716–3724 (PubMed: 17440084). [31] J.E. Dick, Stem cell concepts renew cancer research, Blood 112 (2008) 4793– 4807 (PubMed: 19064739). [32] T. Reya, S.J. Morrison, M.F. Clarke, I.L. Weissman, Stem cells, cancer, and cancer stem cells, Nature 414 (2001) 105–111 (PubMed: 11689955). [33] C. Hirschmann-Jax, A.E. Foster, G.G. Wulf, J.G. Nuchtern, T.W. Jax, U. Gobel, M.A. Goodell, M.K. Brenner, A distinct ‘‘side population’’ of cells with high drug efflux capacity in human tumor cells, Proc. Natl. Acad. Sci. U.S.A. 101 (2004) 14228–14233 (PubMed: 15381773). [34] A. Hadnagy, L. Gaboury, R. Beaulieu, D. Balicki, SP analysis may be used to identify cancer stem cell populations, Exp. Cell Res. 312 (2006) 3701–3710 (PubMed: 17046749).

189

[35] M. Kim, H. Turnquist, J. Jackson, M. Sgagias, Y. Yan, M. Gong, M. Dean, J.G. Sharp, K. Cowan, The multidrug resistance transporter ABCG2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is overexpressed in hematopoietic stem cells, Clin. Cancer Res. 8 (2002) 22–28 (PubMed: 11801536). [36] Y. Shi, X. Fu, Y. Hua, Y. Han, Y. Lu, J. Wang, The side population in human lung cancer cell line NCI-H460 is enriched in stem-like cancer cells, PLoS One 7 (2012) e33358 (PubMed: 22428030). [37] W.R. Luo, X.Y. Chen, S.Y. Li, A.B. Wu, K.T. Yao, Neoplastic spindle cells in nasopharyngeal carcinoma show features of epithelial–mesenchymal transition, Histopathology 61 (2012) 113–122 (PubMed: 22486228). [38] J.P. Thiery, Epithelial–mesenchymal transitions in tumour progression, Nat. Rev. Cancer 2 (2002) 442–454 (PubMed: 12189386). [39] C. Scheel, R.A. Weinberg, Phenotypic plasticity and epithelial–mesenchymal transitions in cancer and normal stem cells?, Int J. Cancer 129 (2011) 2310– 2314 (PubMed: 21792896). [40] S.A. Mani, W. Guo, M.J. Liao, The epithelial–mesenchymal transition generates cells with properties of stem cells, Cell 133 (2008) 704–715 (PubMed: 18485877). [41] D. Sarkar, B. Shields, M.L. Davies, J. Müller, J.A. Wakeman, BRACHYURY confers cancer stem cell characteristics on colorectal cancer cells, Int. J. Cancer 130 (2012) 328–337 (PubMed: 21365650). [42] P.M. Teo, W.H. Kwan, W.Y. Lee, S.F. Leung, P.J. Johnson, Prognosticators determining survival subsequent to distant metastasis from nasopharyngeal carcinoma, Cancer 77 (1996) 2423–2431 (PubMed: 8640688).