Background
On a global scale, breast cancer is the most frequent malignancy and the leading cause of cancer death in women, with 24.2% of all cancers and 15% of all cancer death in 2018 [
1,
2]. In China, breast cancer accounts for 12.2% of new cases and 9.6% of cancer deaths [
3]. In the field of therapy, many clinicians are striving for the goal, to identify areas where optimal care may be achieved with ‘escalating’ or ‘de-escalating’ treatment. As ‘de-escalation’ requires more valuable evidence and rigorous judgment, filtering novel biomarkers with prognostic function is considered to be an effective way.
Heterogeneous clusters of tumor cells exist in solid tumors, a special subgroup of which is called cancer stem cells (CSCs), characterized by self-renewal and pluripotency [
4,
5]. Breast cancer stem cells (BCSCs) are regarded as the source of tumor development, differentiation, invasion and metastasis, drug resistance and recurrence in breast cancer [
6‐
8]. Stem cells has shown attractive prospects in therapy. Therefore, studies on stemness regulation of BCSCs are significant in theory and clinical practice.
Long non-coding RNAs (lncRNAs) are a series of transcript RNA longer than 200 nucleotides without potentials of protein-coding [
9]. LncRNAs recruit transcription factors to regulate gene expression, or interact with microRNAs (miRNAs) and influence the stability of mRNAs [
10]. LncRNAs participate in epigenetic regulation in pathophysiology [
11,
12]. LncRNAs are involved in the tumorigenesis and progression and considered to be a potential biomarker for cancer diagnosis, prognosis and therapy [
13] .
In the present study, large scale screening for the differentially expressed lncRNAs between BCCs and BCSCs, is performed by lncRNAs microarray. After filtration by bioinformatics analysis and consultation based on previous studies, LUCAT1(lung cancer associated transcript 1) is selected as the target. LUCAT1 is located at 5q14.3, firstly found in the airway epithelium of smokers [
13,
14]. LUCAT1 is significantly more-expressed in non-small cell lung cancer (NSCLC) tissues than normal tissues. Its high expression is correlated with high TNM staging, positive lymph node metastasis and poor prognosis [
15]. LUCAT1 is related to cisplatin resistance in ovarian cancer and tumorigenesis in colorectal cancer [
16]. Nonetheless, little is known regarding the expression of LUCAT1 in breast cancer and the stemness regulation of LUCAT1 in BCSCs.
The competing endogenous RNA (ceRNA) hypothesis is a classical mode of gene expression regulation. LncRNAs exert a ‘miRNA Sponge’ by competitively binding miRNAs to antagonize the function of miRNAs in inhibiting specific target mRNAs [
17,
18]. Several studies reported that lncRNAs, as ceRNA, affected tumorigenesis and progression, reflecting a new level of post-transcriptional regulation of genes and providing new insights for the molecular mechanism of CSCs [
19,
20].
Herein, we assessed LUCAT1 expression in 151 breast cancer specimens and firstly reported the association of LUCAT1 with clinical pathology factors and prognosis in breast cancer. Based on big data, the relationship between LUCAT1 and stemness marker was analyzed. LUCAT1 functioned as a miR-5582-3p ‘sponge’ and targeted TCF7L2 (transcription factor 7-like 2), mediated Wnt/β-catenin pathway, and regulated self-renewal of BCSCs and proliferation of BCCs, in vitro and in vivo. The LUCAT1/miR-5582-3p/TCF7L2 axis might provide theoretical support for finding new diagnostic markers and therapeutic targets of breast cancer.
Methods
Data extraction and TCGA analysis
Gene expression was downloaded from TCGA-BRCA (
https://cancergenome.nih.gov/). The edgeR package was used to normalize gene expression. Differential expressions of LUCAT1 in CD44
+CD24
− (BCSCs phenotype marker) and non CD44
+CD24
− patients were analyzed [
21]. Expressions of OCT4, ABCC1, Wnt1, HIF-1α in LUCAT1-high and LUCAT1-low patients were analyzed.
Patient specimens
All patients were diagnosed with infiltrative ductal carcinoma in the First Affiliated Hospital of China Medical University. Prior to operation, patients did not receive chemo- or radiotherapy. Breast cancer specimens (n = 151 cases) were obtained from patients hospitalized from September 2008 to December 2009. The follow-up information was collected from patients or immediate family members through telephone twice a year. Fresh cancer tissues and the matched adjacent normal tissues (n = 26 pairs) were obtained from patients hospitalized from June to July 2018, without follow-up information. Fresh tissues were snap frozen in liquid nitrogen immediately after surgery and stored. This study was approved by Ethics Committee of China Medical University (Approval number: AF-SOP- 07-1.1-01).
In situ hybridization (ISH)
Slides were processed by 1 mL/L DEPC-treated water and APES glue. All liquid and experimental apparatus were treated to remove RNA enzymes. Firstly, the sections were de-waxed by xylene and rehydrated in graded alcohol series. Next, 3% hydrogen peroxide was used to block endogenous peroxidase activity and 3% fresh citric acid diluted pepsin was used to expose mRNA. Then, slides were incubated at 37 °C for 2 h with 20 μL preliminary hybrid liquid, followed by an overnight incubation with 20 μL digoxin-labeled oligonucleotide probe and hybrid liquid at 37 °C. The following step was to add blocking solution, biotinylated rat anti digoxin and SABC, according to protocol in LncRNAs ISH Kit (Boster). DAB staining was evaluated by two pathologists who were blinded to the experiment separately. LUCAT1 expression was estimated by double score semi-quantitative analysis, as previously described [
22]. Patients were categorized into two groups: LUCAT1-high (score > 3) and LUCAT1-low (score ≤ 3). The probe used for LUCAT1 detection was shown in Additional file
1: Table S1.
Cell lines and culture
Normal breast epithelial cell line MCF-10A, breast cancer cell lines MCF-7 and T47D were obtained from ATCC (Manassas, VA, USA). MCF-10A, MCF-7 and T47D cell lines were cultured in high-glucose (4.5 mg/ml) DMEM with 10% (
v/v) FBS (HyClone). MCF-7 CSCs and T47D CSCs were induced and cultured in DMEM-F12 (Gibco, Thermo Fisher Scientific), containing 2% B27 (Gibco), b-FGF 10 μg/L (Promega), EGF 20 μg/L (Promega), as previously described [
23]. All cells were maintained at 37 °C in a 5% CO
2 and 95% air incubator.
Cell transfection and virus infection
MiR-5582 mimic and inhibitor (RIOBOBIO, Guangzhou, China) were transfected into MCF-7 and T47D by Lipofectamine 3000 (Invitrogen), in accordance with protocol. For shRNA knock-down analysis, lentiviral vectors (GV248) were purchased from Genechem Co., Ltd. (Shanghai, China). MCF-7 CSCs and T47D CSCs were transfected with sh-Ctrl or sh-LUCAT1 lentiviral transduction particles (MOI = 20) with 5 μg/mL polybrene (Genechem). For cDNA knock-in analysis, lentiviral vectors (GV502) were purchased from Genechem. MCF7 and T47D were transfected with NC-cDNA or LUCAT1-cDNA (MOI = 20) with polybrene. The medium was replaced with fresh culture medium 24 h after transfection. Stably transfected cells were selected by puromycin (1 μg/ml). The selection was repeated 2–3 times till green fluorescent protein (GFP) was observed in all cells under a fluorescence microscope (Nikon TE 2000-U, Japan). The target siRNAs against LUCAT1 for RNAi, constitution of the vector system were shown in Additional file
1: Table S1.
Quantitative real-time polymerase chain reaction (qRT-PCR)
Trizol reagent kit (CWBIO, China) was used to extract total RNA. RNA concentration was detected by NanoDrop 2000 spectrophotometer (Termo Scientific, USA). cDNA was compounded using the PrimeScript™ RT reagent Kit with gDNA Eraser (Takara, Japan). The miRNA qRT-PCR Starter kit (RIOBOBIO) was used for reverse transcription of miRNA. The primers of U6 and miR-5582 were purchased from RIOBOBIO. SYBR® Green Realtime PCR Msater Mix (TOYOBO, Japan) was used. With 2
-ΔΔCT method (β-actin as the reference), concentration error could be eliminated. The primer sequences were presented in Additional file
1: Table S1.
RNA isolation of nuclear and cytoplasmic fractions
The NE-PER™ Nuclear and Cytoplasmic Extraction Reagents Kit (Thermo Scientific, USA) was applied to isolate and collect cytosolic and nuclear fractions. The experiment generally followed manufacturer’s protocol. RNA levels of LUCAT1, U6 (nuclear control transcript), and GAPDH (cytoplasmic control transcript) were analyzed by qRT-PCR.
Western blot
Cells or tissues were washed with cold PBS and lysed in RIPA buffer containing 1% proteinase inhibitor cocktail solution (Sigma-Aldrich) on ice for 30 min. The NE-PER™ Nuclear and Cytoplasmic Extraction Reagents Kit (Thermo Scientific, USA) was applied to isolate nuclear fractions. Lysates were centrifuged, and protein was quantified using the BCA assay kit (Beyotime, Jiangsu, China). SDS-PAGE electrophoresis was performed on proteins from lysates and transferred them to PVDF membranes (Millipore, Bedford, MA). Membranes were incubated with primary antibody overnight at 4 °C, and then detected by Enhanced Chemiluminescence Detection Kit (BOSTER, USA). Lamin B1 was used as an endogenous control of cell nuclear fraction. All antibodies used were shown in Additional file
2: Table S2.
Immunohistochemistry (IHC)
IHC was performed as previously described [
22] . Ultra-sensitive™ S-P Kit (Maixin-Bio, China) was used. Briefly, sections from paraffin-embedded tumor tissues from transplanted nude mice were incubated with primary antibodies. Results were evaluated by two pathologists who were blinded to the experiment separately.
Flow cytometry assay
All cell lines were digested with 0.25% trypsin and washed with PBS three times. Then cells were resuspended in 100 μL PBS and incubated with anti-CD44-APC (1.25 μl/ test) and anti-CD24-PE (5 μl/ test) (Bio legend, San Diego, USA), or with their controls at 4 °C for 30 min. After incubation, the cells were washed two times with PBS and suspended in 300 μL PBS. Analysis was performed on a MACSQuantTM Flow Cytometer (Miltenyi Biotec).
Sphere-formation assay was performed as previously described [
24]. Briefly, MCF-7 CSCs and T47D CSCs suspension (1 × 10
3/well) was plated in ultra-low adhesion plates (Corning, Kraemer, CA). The cells grown in 2 ml serum-free DMEM-F12 with 10 μg/L bFGF, 20 μg/L EGF and 2% B27. For Colony formation assay, 2 × 10
3 MCF-7 and T47D cells were seeded into 6 cm petri dishes and cultured at 37 °C in 5% CO
2. After 14 days, cells were washed with PBS and fixed in paraformaldehyde for 15 min, stained with 0.5% crystal violet for 15 min.
Luciferase reporter and TOP/FOP-flash reporter assays
Full-length LUCAT1 sequence of wild-type (WT) and mutant-type (MUT) miRNA binding site was constructed in Genechem Co., Ltd. (Shanghai, China). LUCAT1 WT/MUT were transfected into MCF-7 with miR-5582-3p mimic/NC mimic. Similarly, the binding sites for miR-5582-3p in the 3′-untranslated region (3′-UTR) sequence of TCF7L2 were obtained from Genechem. TCF7L2 3′-UTR WT/MUT were transfected into MCF-7 with miR-5582-3p mimic/NC mimic. The Dual-Luciferase Reporter Assay (Promega) was applied according to the manufacturer’s instructions. For the TOP/FOP-Flash assay, TOP/FOP-Flash (Genechem) was co-transfected into cells along with LUCAT1 silence or overexpression vector, miR-5582-5p mimic or inhibitor, and/or the miRNA control. The TOP/FOP-Flash values were normalized to the
Renilla reniformis (Promega) reading and the TOP/FOP ratio was measured, as previously described [
25]. Experiments were performed in triplicate.
RNA immunoprecipitation (RIP)
The Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, Bedford, MA, USA) was applied according to the manufacturer’s protocol [
26]. MCF7 cells were transfected with NC-cDNA or LUCAT1-cDNA. Complete RIP lysis buffer was used to lyse cells. Magnetic beads conjugated with anti-Argonaute 2 (AGO2) (Millipore) or control anti-immunoglobulin G (IgG) antibody were used to incubate the cell extract. The cell extract was incubated for 6 h at 4 °C. As the protein beads were removed, qRT-PCR was conducted for the purification of RNA.
RNA pull-down assay
The pull-down assay was performed as previously described [
26]. Briefly, purified biotinylated LUCAT1 and LUCAT1-mut transcripts were purchased from Sangon Biotech (Shanghai, China). One milligram of MCF-7 cell lysates was incubated with three micrograms of purified biotinylated transcripts for 1 h at 25 °C. The biotin-coupled RNA complexes were isolated by streptavidin agarose beads (Invitrogen) and microRNAs abundance in the pull-down material was analyzed by qRT-PCR.
Xenograft model
1.0 × 10
6 of MCF-7 cells stably transfected with NC-cDNA and LUCAT1-cDNA were suspended in 100 μl PBS and injected into each mammary fat pad of 3/4-week-old female BALB/c (nu/nu) mice (Hua Fukang Biological Technologies Inc., Beijing). The mice were randomized into four groups (
n = 3 per group). Once the tumors had formed, we injected 1.5 nmol miR-5582-3p agomir or agomir NC (RIOBOBIO) at multiple points in tumors, at an interval of 2 days between each injection (i.e., day 10, 13, 16 …) [
26] . The tumor diameter and weights were measured every other day. The tumors were removed and weighed until the largest tumor was close to 2 cm. Tumor volume (mm
3) was measured by a digital caliper and calculated as (width)
2 × (length/2). All mice were bred at pathogen-free conditions in the Animal Center. All animal experiments were approved by the Institutional Animal Care and Use Committee of China Medical University (Approval number: 2017007 M).
Statistical analysis
Statistical analyses were conducted in GraphPad Prism 6.0 (La Jolla, CA, USA) and SPSS 24.0 (Chicago, IL, USA). Results were presented as the mean ± standard deviation for at least three experiments. Student’s independent t test was used between two groups. The relationship between LUCAT1 and clinical pathology factors was examined by Pearson chi-square tests, Fisher’s exact tests and logistic regression analyses. Survival probabilities were judged by the Kaplan-Meier and assessed by a log-rank test. Probability values less than 0.05 were considered statistically significant.
Discussion
Large-scale researches have indicated the essential role of lncRNAs in stemness of CSCs in various cancer. LncRNA XIST promoted glioblastoma CSCs stemness [
30]. LncRNA DGCR promoted cancer cell stem-like properties by targeting miR-330-5p/CD44 axis in NSCLC [
31]. LncRNA Gata6 increased cancer cell stem-like properties and promoted occurrence in colorectal cancer [
32]. LncRNA n339260 promoted CSCs development in hepatocellular carcinoma [
33]. Function of lncRNA to CSCs and the pathway to block its regulation is a new task and hotspot in cancer research [
34,
35]. Similarly, in breast cancer, new lncRNAs are unceasingly discovered to be involved in regulation of stemness. LncRNA HOTAIR (HOX tran antisense RNA), a 2.2 kilobase ncRNA residing in the
HOXC locus, indirectly inhibited MiR-7 and contributed to the EMT of BCSCs by activating the STAT3 pathway. LncRNA FEZF1-AS1 regulated BCSCs stemness by sponging miR-30a targeting Nanog [
8,
24,
36‐
38]. In the present study, LUCAT1 had a possible role in promoting BCCs stem-like properties and BCSCs stemness in vitro
and in vivo
.
Bioinformatics analysis indicates LUCAT1 may bind with multiple miRNAs. Among them, mir-5582-3p arouses our attention. The precursor of human mir-5582 is located on chromosome 11p11.2, with a length of 68 bp. Previous studies found that mir-5582-3p leaded to apoptosis and cell cycle arrest of tumor cells and its expression in colorectal cancer was significantly lower than that of adjacent normal tissues, with an obvious anti-cancer effect [
39]. This study illustrated that mir-5582-3p was a novel suppressor-miR in BCCs stem-like properties and BCSCs stemness via directly targeting the TCF7L2 gene. Further investigations are required to explore the role of mir-5582-3p.
The Wnt/ β-catenin pathway is constitutively active in most tumor cells [
40]. It is a recognized pathway in regulating the self-renewal of BCSCs [
41]. TCF7L2 is a key effector that formed a TCF7L2/β-catenin complex which transcriptionally activated downstream factors in the Wnt pathway [
42]. TCF7L2 represents a central factor in metabolism, stress responses, cell differentiation/proliferation, and cell death [
43]. Dysregulation of Wnt/β-catenin pathway and TCF7L2 target genes play a role in carcinogenesis and is especially well documented for breast cancer [
44]. The present study indicated that the expression level of nuclear β-catenin and Wnt/β-catenin signaling activity were significantly decreased by the down-regulation of LUCAT1. As the screened lncRNAs participate in many classical pathways such as Hedgehog, Wnt, MAPK and TGF-β pathway, future studies can focus on other signaling pathways in stemness regulation of LUCAT1 [
45].
How to convert basic research into clinical practice is a concern to the clinicians. In the tumor cells from breast cancer survivors undergoing chemotherapy, the proportion of BCSCs with CD44
+CD24
− phenotype after chemotherapy was significantly increased compared with that before chemotherapy, and the ability of BCCs to form microspheres was also improved. However, in the tumor cells from Her-2 positive breast cancer patients after treatment with lapatinib, there was no significant increase of CD44
+CD24
− phenotype cells [
46]. LUCAT1 and other potential specific therapeutic targets may suppress malignant tumor by inhibiting BCSCs stem phenotype conversion. Traditional chemotherapy drugs and new drugs targeting BCSCs should be combined to reduce the resistance of BCSCs and improve breast cancer treatment.
Our study has some limitations. We selected MCF-7 CSCs and T47D CSCs because changes of stem phenotype in these two cell lines were relatively obvious and the results were stable. More cell lines or primary cell can be used to further verify whether LUCAT1 exerted similar function in cells with various kinds of molecular typing. Although the induction and culture of BCSCs in our research group was approved by previous reports, selection and identification of BCSCs can be improved with more advanced technologies. The effect of BCSCs heterogeneity on the results cannot be neglected.
In-depth study of the molecular regulation mechanism of lncRNAs in BCSCs is expected to enrich the regulation network of gene expression in breast cancer, and explore new diagnostic and specific therapeutic targets for breast cancer.
Acknowledgements
We thank Prof. DeShu Shang (Department of Cellular Biology, CMU) for providing valuable suggestions on our study design and manuscript. We also thank Zhipeng Cao, Feng Li, Miao He, Xinmiao Yu, Yuying Wang, Ming Zhang and Xinnan Wang for their kind help in the course of the experiment.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.