Abstract
Recent studies suggest that ALDH1 is a putative marker for HNSCC-derived cancer stem cells. However, the regulation mechanisms that maintain the stemness and metastatic capability of HNSCC- cells remain unclear. Initially, HNSCC- cells from HNSCC patient showed cancer stemness properties, and high expression of Bmi1 and Snail. Functionally, tumorigenic properties of HNSCC- cells could be downregulated by knockdown of Bmi-1. Overexpression of Bmi-1 altered in expression property cells to that of cells. Furthermore, knockdown of Bmi-1 enhanced the radiosensitivity of radiation-treated HNSCC- cells. Moreover, overexpression of Bmi-1 in HNSCC- cells increased tumor volume and number of pulmonary metastatic lesions by xenotransplant assay. Importantly, knock-down of Bmi1 in HNSCC- cells significantly decreased distant metastases in the lungs. Clinically, coexpression of Bmi-1/Snail/ALDH1 predicted the worst prognosis in HNSCC patients. Collectively, our data suggested that Bmi-1 plays a key role in regulating Snail expression and cancer stemness properties of HNSCC- cells.
1. Introduction
Head and neck squamous cell carcinoma (HNSCC), including oral squamous cell carcinoma (OSCC), is the sixth most prevalent type of malignancy worldwide and accounts for approximately 8% to 10% of all cancers in Southeast Asia [1, 2]. HNSCC-related mortality is mainly caused by cervical lymph node metastasis, and occasionally by distant organ metastasis [3].
The epithelial-mesenchymal transition (EMT) is a process in which epithelial cells lose their polarity and adopt a mesenchymal phenotype [4]. This process is thought to be a critical step in the induction of tumor metastasis and malignancy [5]. Mani et al. demonstrated that induction of EMT results in cells that have stem cell properties and generates cells with properties similar to breast cancer stem cells [6]. Snail, a member of the zinc-finger transcription factor family, is one of the master regulators that promotes EMT and mediates invasiveness as well as metastasis in many different types of malignant tumors [7, 8]. The aldehyde dehydrogenase (ALDH) family of enzymes is comprised of cytosolic isoenzymes that oxidize intracellular aldehydes and contribute to the oxidation of retinol to retinoic acid in early stem cell differentiation [9]. Recently, ALDH has been reported to be a unique marker of head and neck cancer stem cells (CSC) [10, 11]. ALDH1 was also found to co-localize with other CSCs-related markers, including MMP-9, CD44, and CK14, at the invasive front of the tumor [12]. We previously reported the isolation of ALDH1-positive cells from patients with HNSCC [13]. These HNSCC-ALDH1+ cells displayed the radioresistance and represented a reservoir of cells that have the proliferative potential to generate tumors [13]. ALDH1+-lineage cells underwent EMT and endogenously co-expressed Snail [13]. These findings suggested that Snail expression may regulate the tumorigenesis, radiochemoresistance, and cancer stem cell properties of malignant HNSCC tumors [13]. However, the molecular mechanisms involved in mediating metastasis and tumor malignancy of HNSCC-CSC through the regulation of Snail remain unknown.
Bmi-1 is a member of the Polycomb (PcG) family of transcriptional repressors that mediate gene silencing by regulating chromatin structure [14]. Bmi-1 is essential for maintaining the ability of neural, hematopoietic, and intestinal stem cells to self-renew [15–17]. Bmi-1 was identified as a proto-oncogene that cooperates with MYC to promote the generation of lymphoma [18]. Bmi-1 also inhibited MYC-induced apoptosis by repressing the Cdkn2a locus [19]. Additionally, Bmi-1 has been verified as a predictor of prognosis in bladder cancer [20], prostate cancer [21], brain cancer [22, 23], breast cancer [24], pancreatic cancer [25], and lung cancer [26]. Bmi-1 has been demonstrated to play a role in the tumorigenesis of HNSCC [27, 28]. Bmi-1 has also been reported to be involved in tumor metastasis [29, 30]. Recently, an elegant study by Song et al. showed that Bmi-1 can directly promote EMT and malignancy in nasopharyngeal carcinoma by regulating Snail [31]. The goal of this study was to clarify the relationship between Bmi-1, Snail, and ALDH1 in HNSCC or HNSCC-associated CSC and the involved molecular mechanisms.
2. Materials and Methods
2.1. Isolation and Cultivation of HNSCC-Derived ALDH1+ and ALDH1- Cells from HNSCC Patients
This study followed the tenets of the Declaration of Helsinki. All samples were obtained after patients provided informed consent. The study was approved by the Institutional Ethics Committee/Institutional Review Board of Taipei Veterans General Hospital. The information of HNSCC patients has been previously described in Table 1. The dissociated cells from the samples of HNSCC patients were suspended at 1 106 cells/mL in DMEM supplemented with 2% FCS. The identification of aldehyde dehydrogenase 1 (ALDH1) positive HNSCC cells was carried out using the Aldefluor assay (StemCell Technologies, Durham, NC, USA) and fluorescence-activated cell sorting. Cells were suspended in ALDEFLUOR assay buffer containing ALDH substrate (BAAA, 1 mol/l per 1 × 106 cells) and incubated for 40 min at . As a negative control, for each sample of cells, an aliquot was treated with 50 mmol/l diethylaminobenzaldehyde (DEAB), a specific ALDH inhibitor. The sorting gates were established using the cells stained with PI only as a negative control; the ALDEFLUOR-stained cells treated with DEAB and staining with a secondary antibody alone to test for viability. HNSCC-ALDH1+ cells were cultured in a medium consisting of serum-free DMEM/F12 (Gibco-BRL, Gaithersburg, MD), N2 supplement (R and D Systems Inc., Minneapolis), 10 ng/mL bFGF (R and D Systems), and 10 ng/mL EGF (R and D Systems) [13, 32].
2.2. Quantitative Real-Time RT-PCR
Briefly, total RNA (1 g) of each sample was reverse-transcribed using 0.5 g oligo dT and 200 U Superscript II RT (Invitrogen). The primer sequences for real-time RT-PCR were listed in Table 2. The amplification was carried out in a total volume of 20 L containing 0.5 molL-1 of each primer, 4 mmolL-1 MgCl2, 2 L LightCyclerTM-FastStart DNA Master SYBR green I (Roche Molecular Systems, Alameda, CA), and 2 L of 1 : 10 diluted cDNA. PCR reactions were prepared in duplicate and performed using the following program: for 10 min, followed by 40 cycles of denaturation at for 10 sec, annealing at for 5 sec, and extension at for 20 sec. Standard curves (cycle threshold values versus template concentration) were prepared for each target gene and for the endogenous reference gene (GAPDH) for each sample. Quantification of unknown samples was performed using LightCycler Relative Quantification Software version 3.3 (Roche).
2.3. Knockdown and Overexpression of Bmi-1 with Lentivirus
The pLVRNAi vector was purchased from Biosettia Inc. (Biosettia, San Diego CA). The oligonucleotide -AAAACCTAATACTTTCCAGATTGATTTGGAT CCAAATCAATCTGGAAAGTATTAGG- targeting human Bmi-1 (NM_005180, nt 1061–1081) was synthesized and cloned into pLVRNAi to generate the lentiviral expression vector, pLVRNAi/sh-Bmi1. The lentiviral expression vector carrying Bmi-1 full-length cDNA, pLV/Bmi-1 was obtained from Biosettia Inc. pCMVΔR8.9 and pMD.G, expressing GAG-POL and the vesicular stomatitis virus envelope, respectively, were provided by the consortium (Academia Sinica, Taipei, Taiwan). The lentiviruses were generated by cotransfecting 5 106 293FT cells per 10 cm plate with lentiviral vector and packaging plasmids using Lipofectamine 2000 (LF2000, Invitrogen). Supernatants were collected 48 hours after transfection and filtered. The 48-hour posttransduction viral titers were determined by FACS. Subconfluent cells were infected with lentivirus at a multiplicity of infection of 5 in the presence of 8 g/mL polybrene (Sigma-Aldrich) [13, 33].
2.4. Microarray Analysis and Bioinformatics
Total RNA was extracted from cells using Trizol reagent (Life Technologies, Bethesda, MD, USA) and the Qiagen RNAeasy (Qiagen, Valencia, CA, USA) column for purification. Affymetrix HG U133 Plus 2.0 microarrays containing 54,675 probe sets for 47,000 transcripts and variants, including 38,500 human genes. A typical probeset contains eleven 25-mer oligo nucleotide pairs (a perfect match and a mismatch control). For microarray analysis, sample labeling, hybridization, and staining were carried out by Affymetrix standard protocol with affyQCReport. Probeset was normalized with loess method of all microarrays. The average linkage distance was used to assess the similarity between two groups of gene expression profiles as described below. The difference in distance between two groups of sample expression profiles to a third was assessed by comparing the corresponding average linkage distances (the mean of all pairwise distances (linkages) between members of the two groups concerned). The error of such a comparison was estimated by combining the standard errors (the standard deviation of pairwise linkages divided by the square root of the number of linkages) of the average linkage distances involved. Classical multidimensional scaling (MDS) was performed using the standard function of the R program to provide a visual impression of how the various sample groups are related.
2.5. In Vivo Tumor Growth and Metastasis
All procedures involving animals were in accordance with the institutional animal welfare guidelines of Taipei Veterans General Hospital. Eight-week-old SCID mice and/or nude mice (BALB/c strain) were injected with 105 cells orthotopically. In vivo GFP imaging was performed using an illuminating device (LT-9500 Illumatool/TLS equipped with an excitation source (470 nm) and filter plate (515 nm)). Tumor size was measured with calipers and the tumor volume was calculated using the formula (Length Width2)/2. The integrated optical density of green fluorescence intensity was captured and analyzed using Image Pro-plus software [33, 34].
2.6. Statistical Analysis
The Statistical Package of Social Sciences software (SPSS, Inc., Chicago, IL) was used for statistical analysis. An independent Student’s -test was used to compare the continuous variables between groups. The Kaplan-Meier procedure was used to calculate survival probability estimates. A log-rank test was used to compare the cumulative survival durations in different patient groups. The statistical significance level was set at 0.05 for all tests.
3. Results
3.1. HNSCC-Derived ALDH1-Positive Cells Displayed Tumorigenic and Stemness Properties
Initially, parental, isolated ALDH1+, and ALDH1− cells were isolated from tissue samples of six HNSCC patients using the Aldefluor assay and the fluorescence-activated cell sorting (FACS) analysis (Figure 1(a) and Table 1) [13, 35]. It has been reported that cancer stem-like cells can be cultured in suspension to generate floating spheroid-like bodies (SB) under serum-free medium with bFGF and EGF [36]. Interestingly, ALDH1+ increased higher tumor spheres-forming capability than that of ALDH1- (Figure 1(b)). Furthermore, ALDH1+-derived spheres with regular 10% serum cultivation increased epithelial-attached cells and differentiation marker (CK18)(See Figure (a) in supplementary material available online at doi: 10.1155/2011/609259).To evaluate the enhancement of tumorigenicity of HNSCC-ALDH1+ cells, soft agar colony formation assays and Matrigel/Transwell-invasion and were examined. Compared with parental and ALDH1−, ALDH1+ derived from HNSCC Patients no. 1 and no. 2 showed colony-forming ability and higher invasion activity (Figures 1(c) and 1(d)). To evaluate the in vivo tumor initiating capability of ALDH1+ and ALDH1−, we injected 1000, 3000, and 104 cells into the neck of SCID mice. The results showed that 104 ALDH1− did not induce tumor formation but 3,000 ALDH1+ from the HNSCC tissues of six patients in xenotransplanted mice all resulted in the generation of visible tumors 6 weeks after injection (Table 1).The results of xenotransplanted analysis further showed that ALDH1+ demonstrated higher abilities to induce tumor growth (Figure 1(e)). Lastly, serial xenotransplanted analysis suggested that ALDH1+ had in vivo self-renewal ability (Supplementary Figure (b)). Based on these findings, the ALDH1+-lineage cells isolated from HNSCC patients presented the significant tumor-initiating abilities, especially in ALDH1+ cells from patients no. 1 and no. 2. Real-time RT-PCR data demonstrated that the stemness and EMT-related genes (especially in Bmi-1 and Snail) were significantly activated in HNSCC ALDH1+ (Table 2 and data not shown).
(a)
(b)
(c)
(d)
(e)
3.2. Knockdown of Bmi-1 in HNSCC-ALDH1+ Cells Down-Regulates Snail and Lessens in vitro Tumorigenicity
To further investigate the role of Bmi-1 in maintaining the biological properties of HNSCC-ALDH1+, we used a loss-of-function approach, in which Bmi-1 was knocked down by small hairpin RNA (shRNA) in HNSCC-ALDH1+ cells. Stable knockdown of Bmi-1 in HNSCC-ALDH1+ cells was achieved by transduction with lentivirus that expressed shRNA targeting Bmi-1 (sh-Bmi-1). Lentivirus that expressed shRNA targeted against luciferase (sh-Luc.) was used as a control. Western blot analysis confirmed that shBmi-1 repressed Bmi-1 protein expression in HNSCC-ALDH1+ cells (Figure 2(a)). Importantly, silencing Bmi-1 expression led to downregulation of Snail and ALDH1 expression (Figure 2(a)). Additionally, our results showed that silencing of Bmi-1 in HNSCC-ALDH1+ cells inhibited the ability of the cells to form colonies on soft agar (Figure 2(b)) and migrate/invade (Figure 2(c)).
(a)
(b)
(c)
(d)
(e)
(f)
3.3. Overexpression of Bmi-1 in HNSCC-ALDH1− Cells Enhances Tumorigenic Properties by Upregulating Snail
To evaluate whether overexpression of Bmi-1 could enhance the tumorigenic properties of HNSCC-ALDH1- cells, we generated stable Bmi-1-overexpressing (Bmi-1Over) HNSCCs using lentiviral transduction (Figure 2(d)). Total proteins from HNSCC-ALDH1- overexpressing Bmi-1 exhibited elevated expression of Snail and ALDH1 (Figure 2(d)). In addition, overexpression of Bmi-1 significantly increased soft agar colony formation (Figure 2(e)), and migration/invasion of HNSCC-ALDH- cells (Figure 2(f)). Taken together, our results suggest that Bmi-1 modulates the in vitro tumorigenic properties in HNSCC-ALDH1+ or ALDH1− cells by regulating Snail.
3.4. Overexpression of Bmi-1 in HNSCC-ALDH1- Cells Promotes Stemness Properties
To explore molecules governing stemness and tumorigenicity in HNSCC-CD44-ALDH1- cells treated with Bmi1-overexpressing lentivirus, we examined their transcriptome profile using gene expression microarray analysis (Figure 3(a)). Principle component analysis (PCA) further showed that the transcriptome profile of HNSCC-ALDH1− cells overexpressing Bmi-1 demonstrated higher expression levels of embryonic stem cells (ESCs) transcriptomes (Table 3 and Figure 3(b)). Multidimensional scaling analysis further demonstrated that HNSCC-ALDH1+ cells and HNSCC-ALDH1− cells overexpressing Bmi-1 are more similar to ESCs than HNSCC-ALDH1− cells (; Figure 3(c)). To validate the microarray analysis results, real-time PCR was performed to confirm that the mRNA expression levels of the embryonic genes (Oct-4, Nanog, Sox2, KLF4, and Lin28), EMT-related genes (Snail and Slug), and drug-resistant-related genes (MDR-1 and ABCG2) in Bmi-1-overexpressing ALDH1- cells were significantly higher than those in ALDH1− cells ; Table 2 and Figure 3(d)).
(a)
(b)
(c)
(d)
3.5. Elevation of In Vivo Tumor Growth, Metastatic Activity, and Radioresistance in HNSCC-ALDH1- Cells by Overexpression of Bmi-1
We next sought to determine if Bmi-1 expression could modulate the in vivo tumor initiating activity in immunocompromised nude mice. To monitor the in vivo growth of ALDH1+, ALDH1−, and Bmi-1-overexpressing ALDH1− cells, these cells were transfected using a lentivector combined with the green fluorescent protein gene (GFP) and followed by in vivo GFP imaging system. Firstly, the results showed that 1 104 ALDH1− cells did not induce tumor formation in nude mice, but 1000 ALDH1+ cells generated visible tumors 6 weeks after injection (Table 1). In contrast to ALDH1− cells, one of three (33.3%) nude mice was detected with the tumor formation after 6-week transplantation of 3000 Bmi-1-overexpressing ALDH1− cells. Furthermore, tumor volumes in HNSCC-ALDH1+ transplanted mice were significantly decreased when mice were treated with sh-Bmi-1 (Table 1; Figure 4(a)). Overexpression of Bmi-1 enhanced in vivo tumor growth in HNSCC-ALDH1− (Table 1; Figure 4(a)). Furthermore, we investigated the role of Bmi-1 in the radio sensitivity of HNSCC-ALDH1− and HNSCC-ALDH1+ treated with sh-Bmi-1 and Bmi-1 overexpressing. An ionizing radiation (IR) dose of 0 to 10 Gy was applied to these cells, and HNSCC-ALDH1+ cells showed greater radioresistance than the ALDH1− cells (; Figure 4(b)). Knockdown of BMI-1 in ALDH1+ cells results in significant inhibition of radioresistance while overexpression of BMI-1 in ALDH- cells promotes radioresistant properties (; Figure 4(b)). Moreover, to confirm that Bmi-1 is crucial for metastasis in vivo, mice were injected with different numbers of ALDH1+, ALDH1+/sh-Bmi-1, ALDH1−/Bmi-1over or control GFP-expressing ALDH1− cells. 5x105 Bmi-1-overexpressing ALDH1− cells significantly increased local invasion, distant metastasis to the lungs and tumor size compared with control ALDH1− cells (Figures 5(a) and 5(b)). In addition, silencing Bmi-1 in ALDH1+ cells effectively reduced the number of lung metastases and tumor size in vivo (Figures 5(a) and 5(b)). Taken together, our results reveal a crucial role for Bmi-1 signaling in the maintenance of in vivo tumorigenicity and metastasis of HNSCC-ALDH1+ and -ALDH1- cells.
(a)
(b)
(a)
(b)
(c)
3.6. Coexpression of Bmi-1, Snail, and ALDH1 in HNSCC Tissues Correlates with Poor Overall Survival Rate of HNSCC Patients
Elevated Snail protein expression in HNSCC is correlated with the development of metastasis and poor survival [37]. Elevated expression of ALDH1 also correlates with poor prognosis for HNSCC patients [13]. To investigate whether there is a positive correlation between Bmi-1, Snail, and ALDH1 in head and neck cancers, we studied the expression of Bmi-1, Snail, and ALDH1 by immunohistochemical (IHC) staining of a panel of specimens array from 93 HNSCC patients. The IHC results showed that elevated expression of Bmi-1, Snail, and ALDH1 was positively associated with high-grade, poorly differentiated HNSCC (Figure 6(a)). Our results also showed a significant positive correlation between ALDH-1, Bmi-1 (Figure 6(b)); ALDH-1 and Snail (Figure 6(c)); Bmi-1 and Snail (Figure 6(d)) in HNSCC tissues. This is consistent with previous studies that reported that HNSCC-ALDH1+ cells have elevated Bmi-1 and Snail expression [13, 38]. To determine the prognostic significance of Bmi-1, Snail, and ALDH1 coexpression in patients with HNSCC, Kaplan-Meier survival analysis was performed. Patients who were triple positive for Bmi-1, Snail, and ALDH1 were predicted to have the worst survival rate compared with other head and neck cancer patients (Figure 6(e); Bmi-1+/Snail+/ALDH1+ versus other groups). Overall, these data indicate that expression of Bmi-1, Snail, and ALDH1 in HNSCC patients could be a critical factor in predicting disease progression and clinical outcomes.
(a)
(b)
(c)
(d)
(e)
4. Discussion
A recent study demonstrated that Bmi-1 mRNA and protein overexpressed in a subpopulation of tumor initiating cells in CD44+ HNSCC, which possessed self-renewal and tumor formation ability [39]. Zhang et al. also reported that there are side populations of oral squamous cell carcinomas that express high levels of ABCG2, ABCB1, CD44, Oct-4, Bmi-1, NSPc1, and CK19 [28]. Our previous work showed that HNSCC-ALDH1+ cells have high levels of Bmi-1. The ability to self-renew and radiochemoresistance were significantly suppressed in Bmi-1-silenced HNSCC-ALDH1+ cells [38]. Using microarray, western-blotting, and immunofluorescent assays, Chen et al. confirmed that ALDH1+-lineage cells underwent epithelial-mesenchymal transition (EMT) and endogenously co-expressed Snail [13]. In the current study, our data demonstrated that HNSCC-ALDH1+ cells had high levels of Bmi-1, at both the mRNA and protein levels (Figure 2). Using a lentiviral vector expressing shRNA targeting Bmi-1, we observed that the level of ALDH1 expression and tumorigenic properties of HNSCC-ALDH1+ could be down-regulated by knockdown of Bmi-1 (Figure 2). Importantly, overexpression of Bmi-1 could turn HNSCC-ALDH1− into cancer stem cell-like HNSCC-ALDH1+ cells (Figure 3). Consistent with these findings, the immunohistochemical survey of 93 HNSCC patient tissues showed a positive correlation between expression of Bmi-1, Snail, or ALDH1 and tumor stage (Figure 6). Similar results were noted in other malignancies [40]. Kaplan-Meier analysis demonstrated that patients expressing Bmi-1, Snail, and ALDH1 were predicted to have the worst survival prognosis of HNSCC patients (Figure 6(e)). However, a recent study showed a significant correlation between negative Bmi-1 protein expression and the recurrence of tongue cancer. Their results showed Snail and c-myc expression did not correlate with prognosis [41]. The divergence from our results may be due to the different pathophysiology of HNSCC. Most HNSCC patients in Taiwan consume alcohol, chew betel quid and smoke cigarettes. Tongue cancer patients, especially female tongue cancer patients, usually do not have these habits [3]. The close relationship between tongue cancer and human papillomavirus has been explored by many researchers [42–45]. The correlation between cancer stem cells and the virus infection remains to be discovered.
The prognosis of HNSCC patients with distant metastases in the lung, liver, and bone is very poor [3, 46]. In this study, we found that Bmi-1 can regulate Snail and ALDH1; change the EMT-related genotypes of the ALDH1- cells; and modulate distant lung metastases (Figure 5). Distant metastases have been reported to be associated with Bmi-1 expression in breast cancer [47–49], melanoma [50], gastric cancer [51], and colon cancer [30]. Microarray analysis revealed that eleven gene signatures, which were correlated to the Bmi-1-driven pathway, were closely related to distant lung metastases [40]. Bmi-1 is the target gene of SALL4 in human hematopoietic as well as leukemic cells and is down-regulated if SALL4 is knocked down by the siRNA in the HL-60 leukemia cell line [52, 53]. Recently, researchers employed microRNA profiling to gain insight into the role of Bmi-1 in regulating EMT. Overexpression of miR-200c decreased Bmi-1 expression in breast cancer stem cells (BCSCs) and inhibited the formation of mammary ducts as well as tumors by normal mammary stem cells and BCSCs [54]. Bhattacharya et al. found that miR-15a and miR-16 directly targeted the Bmi-1 untranslated region and correlated with Bmi-1 protein levels in ovarian cancer patients and cell lines [55]. Further research effort is needed in this area. Together, our research shows that the Bmi-1 signaling pathways play a major role in the maintenance of stemness and the metastatic ability of HNSCC-CSC by regulating of Snail expression. Additionally, we demonstrate coexpression of Bmi-1, Snail, and ALDH1 in HNSCC patients was positively correlated with tumor grade and the worst prognosis.
Acknowledgments
This study was supported by research Grants from the National Science Council (NSC-97-3111-B-075-001-MY3/98-2314-B-075-008-MY3), the Taipei Veterans General Hospital (V97B1-006/E1-008/F-001/F-010), the National Yang-Ming University (Ministry of Education, Aim for the Top University Plan), the Chung Shan Medical University Hospital (CSH-2010-C-025), the Technology Development Program for Academia (98-EC-17-A-19-S2-0107), and Department of Industrial Technology, Ministry of Economic Affairs, Taiwan.
Supplementary Materials
Figure 1. (a) Cell morphology in ALDH1+ HNSCC cells under specific serum free medium and 10% serum (right panel). Epithelial differentiation marker, CK18 positive cells in ALDH1+ HNSCC cells under specific serum free medium and 10% serum (right panel). (b) In vivo self-renewal ability of HNSCC-ALDH1+ cells.