Esophageal cancer is the eighth most common cancer and ranks as the fourth highest cause of cancer-related mortality worldwide, with 456,000 new cases and 400,000 deaths in 2012 [1
]. Esophageal squamous cell carcinoma (ESCC) is the principal histological type of esophageal cancer, with high incidence and mortality in China, Korea, Japan, and Africa. ESCC patients are more likely to be diagnosed at a later stage because there is a lack of specific biomarkers for diagnosis. The 5-year survival rate of esophageal cancer patients is no more than 20% [2
]. The poor overall survival of patients with ESCC is mainly responsible for therapy resistance and recurrence. Finding novel, sensitive, and specific biomarkers for early diagnosis and elimination of therapy resistance has great potential to improve the outcomes of ESCC patients.
Cancer stem cells (CSCs) are considered to be closely related to the origin of cancer. Principle research for CSCs proposed that tumors are composed of a few CSCs, which have a strong capacity to self-renew, and a large proportion of common tumor cells that are derived from CSCs [3
]. CSCs are believed to originate from stem cells with high tumorigenic properties. A large number of studies and clinical trials indicated CSCs are responsible for recurrence of ESCC [3
]. Unfortunately, CSCs are not sensitive to chemotherapy and radiotherapy because abnormal modulating signaling pathways exist, such as Notch, Wnt, and Hedgehog [7
]. Therefore, the study of CSCs in ESCC will provide a new window for ESCC research.
Because there is a lack of specific methods for identifying and isolating CSCs, research on CSCs in ESCC has progressed slowly. Currently, CSCs in ESCC are mainly isolated with flow cytometry using multiple protein markers and fluorescent probes such as CD44, Side-population (SP), CD271, Aldehyde dehydrogenase (ALDH), and CD90 [10
]. However, these biomarkers are not specific for ESCC, which have been reported to be expressed in other tumor CSCs; there are no unique biomarkers for CSCs in ESCC. Even the expression of these biomarkers varies considerably in different kinds of CSCs. A combination of multiple biomarkers can greatly improve the specificity for stem cell sorting. In this study, various combinations of CD71, CD338, CD271, and CD49f have been considered and tested in ECa9706, ECa109, KYSE50, and CAES17. Finally, CD71−
were verified to be positive biomarkers for identifying CSCs in ESCC.
The exact mechanisms for CSCs involved in ESCC tumor genesis remain largely unknown. MicroRNAs (miRNAs) are important regulators in CSCs by binding to the 3′UTR region of target-mRNAs. miRNA networks in CSCs play an important role in the maintenance of stemness, which was considered to be a potential target in ESCC therapy [14
]. In this study, we discovered the expression of hsa-miR-21-3p is up-regulated in CSCs of ESCC, and hsa-miR-21-3p can promote cell proliferation but suppress cell apoptosis. We further identified TRAF4 as the direct target of hsa-miR-21-3p.
CSCs were mainly isolated using fluorescence-activated cell sorting (FACS) or immune-magnetic beads. (Antibody-mediated cell sorting using FACS is more suitable and straightforward to purify rare populations of CSCs in tumors.) Different panels of biomolecules were identified to detect and isolate these CSCs in various cancers. However, research progress on CSC sorting in ESCC has been hampered by the lack of suitable biomarkers for prospective isolation. It has been reported that CD44, CD133, ALDH (Aldehyde dehydrogenase), CD271, CD90, and Side-population (Hoechst 33,342 dye exclusion) are potential biomarkers used to identify cancer stem cells in ESCC. Still, there are lots of challenges and limitations. Usually these biomarkers have low specificity for rendered populations, and many biomarkers have no direct evidence in demonstrating the stemness. Multiple biomarkers could improve the specificity of cell sorting, and purer CSCs could be obtained with a combined selection rather than a single selection.
In this study we identified three potential biomarkers for CSC sorting including CD271, CD338, and CD71. CD271 is also called p75 neurotmphin receptor (p75NTR), and in ESCC, cells expressing CD271 were reported to have higher self-amplifying and self-renewal capacities than cells not expressing CD271 [26
]. The expression of CD271 was reported to be closely related with the survival and maintenance of cancer. In addition, the expression of stem cell-associated genes was dependent on CD271 [27
]. Kojima discovered that CD271-positive cells have enhanced CSC properties that are mitotically quiescent in ESCC. CD338, also called ABCG2, is an isoform of an ATP-binding cassette transporter. The overexpression of ABCG2 was reported to be correlated with lymph node metastasis in ESCC patients. ABCG2 is considered to be a potential biomarker for CSCs in ESCC, and ABCG2-positive cancer seemed to produce more stemness [28
]. CD71 is also known as transferrin receptor protein 1(TRf1), and CD71 has been used as a surface biomarker to isolate hemopoietic stem cells. In ESCC, CD71 is reported to be correlated with tumorigenic properties. The combination of CD71−
can be satisfactorily used to isolate CSCs in ESCC with a high specificity and efficiency, which can provide us with new strategies for further research on CSCs in ESCC.
It is widely believed that miRNAs are critical during stem cell epigenesis. It has been shown that the expression patterns of miRNAs changes during the differentiation of embryonic stem cells, which suggests miRNAs may play important roles in maintaining the pluripotency and self-renewal capacities of ES cells, and these miRNAs may also serve as molecular markers for ES cells. In addition, miRNAs played a key role in the process of proliferation and differentiation of hematopoietic cells, fats, nerves, muscles, and cardiomyocytes. Similarly, in CSCs, miRNA expression profiles were different from normal cancer cells. miR-135a can inhibit the development of Cancer Stem Cell-Driven Medulloblastoma by repressing Arhgef6 Expression [29
]. Peng et al. [17
] demonstrated that the miRNA-103/107 family can promote stem cell phenotypes by targeting ribosomal kinase p90RSK2. Liu et al. [30
] discovered that miRNA-148b suppressed CSCs by targeting neuropilin-1 in hepatocellular carcinoma. However, there are few studies on the expression of miRNA in CSCs of ESCC. In this study, we detected miRNA and mRNA expression profiles in positive and negative cell subpopulations of CSCs. Fifty-four differently expressed miRNAs and 303 differently expressed mRNAs were discovered. Biological analyses revealed that differential expression of miRNAs and mRNAs are involved in transcriptional regulation, cell cycle regulation, cell differentiation regulation, and RNA splicing, which are closely with the maintenance of CSCs. miRNAs are potentially critical for the maintenance of CSCs in ESCC.
It has been reported that hsa-miR-21-3p can inhibit proliferation and invasion in ovarian cancer cells. In hepatocellular carcinoma, hsa-miR-21-3p inhibited tumor cell growth and promoted apoptosis [31
]. However, in colorectal cancer hsa-miR-21-3p was upregulated and promoted cell migration and invasion [34
], which revealed hsa-miR-21-3p played different roles in different tumors. In bone marrow mesenchymal stem cells, hsa-miR-21-3p had an abnormal expression [35
]. It seems that hsa-miR-21-3p is potentially related with the maintenance of stemness. We detected that hsa-miR-21-3p was differently expressed between two subpopulations of cells. Then, we demonstrated hsa-miR-21-3p could promote cell proliferation, migration, and invasion in ESCC. We further identified TRAF4 as a direct target of has-miR-21-3p using a Dual-Luciferase Reporter assay and Western blot assay. TRAF4, as a strong evolutionary conservation gene, reinforced the idea that it exerted important biological functions [35
]. The subcellular localization of TRAF4 has been controversial for years. Indeed, TRAF4 has been detected at the cell membrane, in the cytoplasm, and in the nucleus [36
]. Several reports discovered that TRAF4 might be a regulated gene of p53 (mediating cell cycle arrest), DNA repair, and apoptosis of cells, and it was associated with the ability of responding to cellular stress [37
], colony formation [38
], and squamous cell carcinoma of the head and neck differentiation [39
]. Kedlinger et al. [40
] implicated TRAF4, in one of the emerging TJ-dependent signaling pathways, responded to cell polarity by regulating the cell proliferation/differentiation balance and, subsequently, epithelium homeostasis in TRAF4-deficient mice and drosophila. Research also suggested that TRAF4, as a mediator in the TNF-induced signaling pathway leading to activation of p70S6K, inhibited Fas-induced apoptosis [41
]. Xin et al. [42
] reported that TRAF4 can directly act on p75 NTR, and in this way the activation of NF-kB was inhibited. Interestingly, the expression of p75 NTR(CD271) was positive in our isolated stem cells. The hsa-miR-21-3p/TRAF4 axis potentially promoted cell proliferation by acting on p75 NTR. At the same time, by regulating signal transduction of NF-kB, cell apoptosis was inhibited. Our results supplied understanding of CSCs in ESCC, that the hsa-miR-21-3p/TRAF4 axis may potentially be a new target for inhibiting ESCC.
CSCs are considered to be closely related to tumor genesis and tumor recrudescence, which undergoes changes in early stages of tumor development. Differently expressed molecules may be used as biomarkers for early diagnosis and treatment. To detect the possibility of has-miR-21-3p as a biomarker for ESCC, we analyzed the relationship between the expression of has-miR-21-3p and the risk for ESCC in tissues collected during surgery. Our findings indicated miR-21-3p might serve as a biomarker for the diagnosis of ESCC.
4. Materials and Methods
Esophageal carcinoma tissues and their corresponding normal non-tumor tissues (from adjacent 3 cm) were surgically collected between 2009 and 2010, and stored in tubes at −80 °C. A total of 137 cases of esophageal carcinoma patients, ranging from 43 to 80 years old, including 94 males and 43 females, were collected. All patients had not been treated with chemoradiation. Written informed consent was obtained from all subjects prior to recruitment to the study. Ethical approval was provided by the Institutional Review Board of the Southeast University-Affiliated Zhongda Hospital (Nanjing, China) (Approval no: 2011ZDL002.0, 24 February 2011).
4.2. Animal and Cell Lines
NOD/SCID mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The mice were female, aged 3 to 4 weeks, and weighed 21–25 g. The animals were housed and maintained in specific pathogen-free (SPF) shelves with a constant temperature (20–26 °C) and constant humidity (50–56%). Human ESCC cell lines ECa9706, ECa109, KYSE150, and CAES17 were provided by Key Laboratory of Environmental Medicine Engineering, Ministry of Education, School of Public Health, Southeast University. Cells not sorted were grown in RPMI-1640 containing 10% fetal bovine serum (Gibico, Grand Island, NY, USA), 100 U/mL penicillin–streptomycin solution (Gibico), and 200 mM L-glutamine (Invitrogen) at 37 °C in an incubator containing a 5% CO2 humidified atmosphere. The sorted cells were cultured in SSM and SFM with 10 ng/mL of EGF and bEGF (Invitrogen, Carlsbad, CA, USA). EC9706 cells over-expressing and under-expressing hsa-miR-21-3p were obtained using Micron™ miRNA mimic and inhibitor (RiboBio, Guangzhou, China). siRNAs used for decreasing TRAF4 were synthesized by RiboBio (forward, ATCCGAAAGCAGTGTGAACACTCCTTTCTTTCGTTAGGCTTGAATGAAGAACGAG; reverse, AGCAATAGTCGGTTCTGATTTCCAGTCTTACCAAAGCGTTAGGAACCGCGAAATTC). Lipofectamine® RNAiMAX reagent (Thermo Fisher, Waltham, MA, USA) was used for transfection according to the manufacturer’s instructions.
4.3. Antibodies and Reagents
Anti-cytokeratin 13 antibody and anti-pan cytokeratin antibody (AE1 + AE3) were obtained from Abcam (Cambridgeshire, UK). GAPDH, β-actin, and TRAF4 antibodies were obtained from Cell Signaling Technology (CST, Danvers, MA, USA). The secondary anti-rabbit IgG1 HRP-conjugated and anti-mouse IgG HRP-conjugated antibodies were purchased from Abcam and CST, respectively.
4.4. Flow Cytometry and Cell Sorting
PE labeled anti-ABCG2 (CD338), FITC labeled anti-TfR1 (CD71), APC labeled anti-p75NTR (CD271), ALXA FLOUR® 647 labeled anti-p75NTR (CD271), and PE-CY5 labeled anti-Integrinα6 (CD49f) were purchased from BD Biosciences (San Jose, CA, USA). Cells were marked according to the manual of regents. Live cells were counted using Trypan blue exclusion. Fluorescence was analyzed and sorted on a total of 1 × 104 cells per sample using a flow cytometer (Facs Aria II; Becton Dickinson, Mountain View, CA, USA). The number of cells sorted depended on the requirements of the assay.
4.5. Apoptosis and Cell Cycle
Cell apoptosis was quantified using the Annexin V-FITC Apoptosis Detection Kit (KGA107, Keygen Biotech, Nanjing, China) according to the manufacture’s protocol. Flow cytometry (PI staining) was used to detect cell cycle using the PI cell cycle Detection Kit (KGA107, KeyGEN Biotech) according to the manufacture’s protocol.
4.6. Cell Proliferation Assay
Cells with a density of 1 × 104 cells/well on 96-well plates were quantified using a 5-ethynyl-2′-deoxyuridine (EdU) labeling/detection kit (Ribobio, Guangzhou, China) to detect proliferation. Firstly, 50 mM EdU was applied to the cultures, and the cells were grown for an additional 2 h. Then, the cells were fixed with 4% formaldehyde in PBS for 30 min and incubated with glycine for 5 min. After washing with PBS and 0.5% TritonX-100 in PBS, the cells were incubated with 1× Apollo dye at room temperature in darkness for 30 min. Lastly, the cells were washed with 0.5% TritonX-100 in PBS and methanol, and they were incubated with 1× Hoechst 33342 dye at room temperature in darkness for 30 min. After labeling, cells were preserved with 100 μL PBS. Analyses of cell proliferation (ratio of EdU + to the total) were performed using images of five randomly selected fields obtained on a fluorescence microscope. Assays were performed in five parallels.
4.7. Plate Cloning Assay
The cells were suspended and cultured for sorting with SFM, and 1 × 102 cells/well were planted on 96-well plates. After 14 d of culture, colony forming efficiency was calculated only for clones containing more than 50 cells.
4.8. Soft Agar Cloning Assay
Agaropectin was prepared, containing 0.6% and 0.2% of low melting point agar for the bottom and upper layers, respectively, using SFM. A total of 5 × 102 cells/well were planted in 6-well plates. Clones were counted after 3 weeks under a microscope (Olympus, Tokyo, Japan).
4.9. Scratch-Healing Experiment
Sorted cells were planted in 6-well plates with SFM. When cells reached 90% confluence, the cells were scratched with a standard 10 µL pipette tip. Then, the plate was washed to remove cell debris, freshened with medium, and cultured for 48 h. After 48 h the size of wound was observed and measured under a microscope.
4.10. Invasive Experiment
Cell migration assays were performed using 8.0 μm Transwell chamber (Corning, Corelle, NY, USA). We first set a layer of Matrigel (Corning), and 5 × 104 cells were seeded into the upper chamber in SFM, while the lower chamber was filled with DMEM-F12 containing 50% fetal bovine serum. Cells were then cultured for 24 h, colored, and the number of invasive cells were counted.
4.11. RNA Extraction and Genome-Wide mRNA Microarray
Total RNA was isolated by Trizol reagent (Invitrogen) using the standard method. The RNA samples were quantified with a Nanodrop spectrophotometer (Thermo). Human miRNA V16.0 (Agilent, Santa Clara, CA, USA) was used to detect miRNA profiles in positive (CD71−/CD271+/CD338+) and negative (CD71+/CD271−/CD338−) cells. A Roche NimbleGen expression chip was used to detect mRNA profiles in positive (CD71−/CD271+/CD338+) and negative (CD71+/CD271−/CD338−) cells. A HumanHT-12 v4 Expression Bead Chip Kit (Illumina, Santiago, CA, USA) was used to detect samples transfected with miR-21-3p and negative controls.
4.12. Bioinformatics Analysis of Microarray and Target Prediction
Gene ontology (GO) hierarchy analyses were carried out on the differentially expressed genes using the Gene Ontology Enrichment Analysis Software Toolkit (Version 1.30, Beijing, China). GO was organized into three partially overlapping categories: biological processes, molecular functions, and cellular components. Pathway enrichment analyses of gene expression were obtained using web gestalt WEB-based Gene Set Analysis Toolkit software, which involved databases from the Kyoto Encyclopedia of Genes and Genomes (KEGG). A p value reflecting the importance of GO or the pathway results value was used to identify the significant GO terms and pathways. miRWalk, miRanda, miRDB, RNA22, and Targetscan were used for target prediction of miRNAs. For predicting the miRNA-associated mRNA network, we first predicted target mRNAs of abnormally expressed miRNA. Targets identified by more than three prediction tools were selected for further analysis. We then selected common mRNAs with mRNA expression profiles from the predicted mRNAs to visualize a miRNA-associated mRNA network using the Cytoscape (Denver, CO, USA).
4.13. Western Blot Analysis
Cellular protein was extracted with cold RIPA buffer containing protease inhibitors (Beyotime, Shanghai, China). Lysates were cleared by centrifugation at 14,000 rpm at 4 °C for 15 min. Protein concentrations were determined using the BCA assay (Thermo Scientific). Aliquots of protein (20 μg) were separated by 10% SDS-PAGE, and the separated proteins were transferred to PVDF membrane. Membranes were blocked with 5% (w/v) non-fat milk in Tris-HCl buffered saline (pH 7.4) with Tween-20 and incubated with the primary, monoclonal antibody overnight at 4 °C. Subsequently, membranes were washed with Tris-HCl buffered saline and incubated with secondary antibodies conjugated to horseradish peroxidase, diluted to 1:3000 (CST) and 1:5000 (Abcam), at room temperature for 1 h. Membranes were washed in Tris-HCl buffered saline, and bounds were detected with SuperSignal West Femto/Pico Kit (Thermo Scientific). Blots were visualized and quantified using a Tanon-5200 Imaging System (Tanon, Shanghai, China).
4.14. Luciferase Reporter Assay
Plasmids containing the mutant and non-mutant sequence of 3′UTR of TRAF4 were structured. A Luciferase reporter gene assay was performed using the Dual-Luciferase Reporter assay system (Ribobio) according to the manufacturer’s instructions. Cells of 90% confluence were seeded in 96-well plates with a concentration of 1 × 104/well and incubated for 24 h. Cells were co-transfected with miR-21-3p or negative control and a Luciferase Reporter plasmid. A total of 100 ng of Luciferase Reporter plasmid was mixed with 1.5 p mol of miRNA mimic or negative control in 10 μL of opti-MEM. A total of 0.25 μL of lipofectamine2000 was diluted into 10 μL of opti-MEM and added into the former mixture after incubation for 5 min. When incubated for another 20 min, 20 μL of the transfection mixture and 80 μL of antibiotic-free RPMI1640 media were added into the 96-well plate and incubated at 37 °C and 5% CO2. Reporter gene assays were performed 48 h post-transfection using the Dual-Luciferase assay system (Promega, Madison, WI, USA). Firefly luciferase activity was normalized for transfection efficiency using the corresponding Renilla luciferase activity. All experiments were performed at least three times.
All primers (Bulge-Loop™ miRNA RT-qPCR Primer kits) for miRNA were purchased from Guangzhou RiboBio Co., Ltd. (Guangzhou, China) All primers for mRNA were synthetized from the GenScript Corporation. Detailed sequences are shown in Table S4
Total RNA (~2 μg) was extracted using Trizol regent (Invitrogen). cDNA was synthesized using Moloney Murine Leukemia Virus (MMLV) reverse transcriptase (Promega) and ribonuclease inhibitor (Fementas, Madison, WI, USA). SYBR Green mastermix was purchased from Toyobo Technologies (Osaka, Japan). QPCR reactions were run using the StepOnePlus system (Applied Biosystems, Carlsbad, CA, USA). The data for miRNA and mRNA were normalized to U6 and β-actin, respectively. The expressions of miRNA and mRNA were presented as relative RNA expression using ΔΔCq formula (the fold change in target gene expression was equal to 2−ΔΔCq). All results were presented as the mean of triplicates ± SD from three independent experiments.
4.16. Tumor Xenograft in Nude Mice
Female NOD/SCID mice (6 per group) were subcutaneously injected with 5 × 103 of cells of positive (CD71−/CD271+/CD338+) and negative (CD71+/CD271−/CD338−) cells into the upper limb. The xenografts were monitor for 8 weeks, then the mice were sacrificed by cervical dislocation. Subcutaneous implanted tumors were collected and stained with HE and AE1/AE3 antibody. All animal experiments were conducted in accordance with the protocols approved by the Laboratory Animal Centre of Southeast University (20110226006, 26 February 2011).
4.17. Statistical Analysis
Statistical analysis was performed using SPSS 17.0. (Armonk, NY, USA). A p value < 0.05 was considered to be statistically significant. Wherever stated, one asterisk denotes p < 0.05, two asterisks denote p < 0.01, three asterisks denote p < 0.001, and NS denotes p > 0.05.