Long non-coding RNA MIDEAS-AS1 inhibits growth and metastasis of triple-negative breast cancer via transcriptionally activating NCALD

Human breast cancer tissues and adjacent non-tumor tissues were obtained from patients at the Qilu Hospital. Written informed consent was provided by all patients, and the study was achieved approval by the Ethical Committee on Scientific Research of Shandong University Qilu Hospital (IRB number, KYLL-2016(KS)-140).

All cell lines were bought from American Type Culture Collection (ATCC, Manassas, VA). MDA-MB-231, MDA-MB-468 and HEK293T cells were cultured with DMEM. The above media contained 100 U/ml penicillin, 100 μg/ml streptomycin, 10% fetal bovine serum (Cell-Box, HK, China) and were cultured with 5% CO₂ in a humidified cell-culture incubator at 37 °C.

Tissues obtained was minced and then, placed into 50 ml conical tube containing 20 ml Advanced DMEM/F-12 (Gibco, CA, USA) which was supplemented with 1× Glutamax, 10 mM HEPES (Invitrogen, Texas, USA), 1 mg/ml BSA (BasalMedia, Shanghai, China), 1× Primocin (InvivoGen, Texas, USA) and 1 mg/ml collagenase (Sigma-Aldrich, MO, USA), 0.01 mg/ml Hyaluronidase (Sigma-Aldrich, MO, USA), 10 μmol Y-27632 (Sigma-Aldrich, MO, USA). Tubes containing minced tissue and collagenase were digested on a 37 °C constant temperature shaker for about 1.5 h. After digestion, filter with 100 μm cell strainer filtration, cell filtrate was enriched by centrifugation at 300 g, 4 °C, at low-speed horizontal-refrigerated centrifuge and washed with cold Advanced DMEM/F-12 twice. Isolated cells were uniformly mixed with Cultrex RGF Basement Membrane Extract Type 2 (R&D Systems, MN, USA). A 50 μl drop of this suspension was placed in center of a well in an uncoated 48 well plate and allowed to harden for 30 min at 37 °C. Finally, each well was added 350 μl of human organoid medium contained Advanced DMEM/F-12 which was supplemented with 1× Primocin (Invitrogen, Texas, USA), 10 mM HEPES, 1× Glutamax, 1× B27 (Gibco, CA, USA), 100 ng/ml Nrg1 (PeproTech, NJ, USA), 100 ng/ml Noggin (PeproTech, NJ, USA), 50 ng/ml EGF (PeproTech, NJ, USA),500 ng/ml Human R-spondin-1 (PeproTech, NJ, USA), 500 nM A83-01 (Tocris Bioscience, MN, USA),5 μmol Y-27632, 5 mM Nicotinamide (Sigma-Aldrich, MO, USA), 3 μmol SB202190 (Sigma-Aldrich, MO, USA), 10 nM Prostaglandin E2 (Sigma), 1.25 mM n-Acetyl Cysteine (Sigma-Aldrich, MO, USA). Organoids were maintained in 37 °C humidified atmospheres under 5% CO₂. Medium was changed every 3–4 days, and organoids were passaged using TrpLE Express (Invitrogen, Texas, USA) approximately every 2–4 weeks.

Breast cancer gene expression data were downloaded from TCGA (https://​tcga-data.​nci.​nih.​gov/​tcga) and GEO (http://​www.​ncbi.​nlm.​nih.​gov/​geo) database. The data analysis was performed with R software using the DEGseq package. The threshold set for differences was |log2(Fold change) |> 1 and p-value < 0.05.

The small interfering RNAs targeting MIDEAS-AS1, MATR3, NCALD and the scrambled oligonucleotides (NC; 20–50 nM) were purchased from GenePharma (Shanghai, China). Full sequence of MIDEAS-AS1, MATR3, NCALD complementary DNA (cDNA) was amplified by PCR and then, inserted into the vector pcDNA3.1-GFP (abbreviated as “pcDNA3.1”) or pEnter (WZ Biosciences, Jinan, China) to construct pcDNA3.1/MIDEAS-AS1, pcDNA3.1/NCALD, pEnter/MATR3 plasmids. Empty vector (pcDNA3.1, pEnter) was regarded as the control. Using Lipofectamine 2000 (Invitrogen, Texas, USA), all vectors were transfected into breast cancer cells. After 24 h of transfection, cells were collected for following experiments. The sequences that were used are shown in Additional file 1: Table S1.

Total RNA was extracted using the RNA-easy Isolation Reagent Kit (Vazyme, Nanjing, China). Reverse transcription from 1 μg RNA to cDNA was performed using the PrimeScript reverse transcriptase reagent kit (Takara, Shiga, Japan). Real-time PCR was performed with SYBR qPCR SuperMix Plus (novoprotein, Suzhou, China) on a QuantStudio 6 Flex Real-Time PCR System. Results were analyzed using the comparative Ct method normalizing to Actin. The primer sequences are shown in Additional file 1: Table S2.

Separation of nuclear and cytosolic fractions was performed using the PARIS Kit (Invitrogen, Texas, USA) according to the manufacturer’s instructions. Afterward, MIDEAS-AS1, GAPDH (cytoplasmic control) and U6 (nuclear control) in a cytoplasmic fraction or nuclear fraction were detected by qRT-PCR.

The RIP experiments were performed strictly with a Magna RIP RNA-Binding Protein Immunoprecipitation kit (Millipore, Burlington MA, USA) according to the manufacturer’s instruction. Each RIP reaction required 100 μl of 1 × 10⁷ MDA-MB-231 cell lysate, and each immunoprecipitation required 5 μg of antibody. The expression of MIDEAS-AS1 in the precipitated of anti-MATR3 and negative control (IgG) was detected by qPCR, and the content in the IgG precipitate was used as a reference. qRT-PCR was performed as described above.

After transfection treatment for 24 h, MDA-MB-231 and MDA-MB-468 were harvested and resuspended in serum-free DMEM medium and seeded into the upper transwell chambers containing 8 μm pores. As for invasion experiment, the cells were seeded in Matrigel matrix-plated chambers. Culture medium supplemented with 20% FBS was added to the lower chamber. After incubation for 20 h for migration and 24 h for invasion at 37 °C, the chambers were fixed with methanol, stained with 0.5% crystal violet. Then, the cells on the upper surface were wiped off and allowed to dry at room temperature. The migrated and invasive cells were counted and photographed under a light microscopy (200×) (Olympus, Tokyo, Japan).

The cells were seeded in 24-well plates at a density of 3.5 × 10⁵ cells per well for MDA-MB-231 and 5 × 10⁵ cells per well for MDA-MB-468. Then, incubated the wells with cell culture medium at 37 °C overnight. When the cells at approximately 90% confluence, an artificial wound was made with a 10-µl sterile pipette tip. The fragments were washed thoroughly with PBS, and the wells were added the serum-free medium, cultured for various amounts of time. MDA-MB-231 were incubated for 24 h, and MDA-MB-468 cells were incubated for 48 h. We used a microscope to detect cell migration near the wound and obtain images. The images were processed using ImageJ software to quantify the open wound area as average open wound area % and a histogram was drawn.

FISH was performed using the RNA FISH Probe Mix Kit (GenePharma, Shanghai, China) according to the manufacturer’s protocol. Briefly, we placed the cell slides in a 24-well plate, seeded the cells at a density of 1.5 × 10⁵ cells per well, and incubated overnight at 37 °C. The medium was discarded and washed twice with PBS in each well, and 4% paraformaldehyde was added and fixed at the room temperature for 15 min. After blocking, the cells were incubated with the lncRNA probes at 37 °C for 16 h and washed three times with a washing solution for 15 min. Subsequently, we added Hoechst 33342 (Beyotime, Suzhou, China) fluorescent dye solution for 10 min at the room temperature while avoiding light and then, observed the distribution of MIDEAS-AS1 under the Confocal Microscope ZEISS LSM 880 (ZEISS, Berlin, Germany).

ChIP assays were performed using a ChIP assay kit (Cell Signaling Technology, MA, USA) according to the manufacturer’s instructions. Briefly, cells were fixed for 10 min with 1% free formaldehyde and then disrupted in SDS lysis buffer. Chromatin was sonicated by Bioruptor® Pico (Diagenode, Belgium) to shear DNA to an average length ranging from 200 to 1000 bp, as verified by agarose gel electrophoresis. Next, chromatin was immunoprecipitated with anti-Flag (Cell signaling Technology, USA), and normal rabbit IgG was used as the negative control. Final DNA extractions were quantitative-PCR amplified using primer pairs that cover the sequence in the NCALD promoter region (−2000 bp to + 100 bp).

In vitro transcription of sense, antisense or truncated MIDEAS-AS1 was achieved by T7 RNA polymerase (Thermo Fisher, MA, USA). Subsequently, the product of in vitro transcription was obtained biotin-labeled with PierceTM RNA 3’ End Biotinylation Kit (Thermo Fisher, MA, USA). Washed streptavidin magnetic beads were incubated with 50 pmol of purified biotinylated transcripts at room temperature for 30 min, followed by addition of the whole-cell lysates (20–200 μg) from MDA-MB-231 cells and incubated for 1 h at 4 °C. The beads containing DNA and proteins were then washed and eluted, then beads were boiled, and precipitated protein was separated by SDS-PAGE and detected by Western blotting analysis.

Protein samples were harvested from cell lysates, and the concentration of total protein was measured with a BCA Protein assay kit (Millipore, Burlington MA, USA). Protein were separated by SDS-PAGE and transferred to a PVDF membrane (Millipore, Burlington MA, USA). The membrane was blocked with 5% nonfat milk at room temperature and incubated overnight with specific primary antibodies at 4 °C: rabbit anti-Fibronectin (Proteintech, Wuhan, China), rabbit anti-N-cadherin (Proteintech, Wuhan, China), rabbit E-cadherin (Proteintech, Wuhan, China), rabbit anti-Vimentin (Proteintech, Wuhan, China), rabbit anti-MMP-9 (Cell Signaling Technology, MA, USA), mouse anti-GAPDH (Servicebio, Wuhan, China), rabbit anti-MATR3 (Proteintech, Wuhan, China), mouse anti-Actin (Servicebio, Wuhan, China), rabbit anti-NCALD (Proteintech, Wuhan, China), rabbit anti-TGF-β (Proteintech, Wuhan, China), rabbit anti-p-p65 (Affinity Biosciences, OH, USA), mouse anti-p65 (Proteintech, Wuhan, China), rabbit anti-p-Erk1/2 (Cell Signaling Technology, MA, USA), rabbit anti-Erk1/2 (Proteintech, Wuhan, China), washed with Tween-20/TBS and incubated with horseradish peroxidase-conjugated secondary antibodies for 2 h, followed by enhanced chemiluminescence detection (Vazyme, Nanjing, China), and protein bands were visualized using a ECL detection system (Tanon, Shanghai, China), band density was determined using the ImageJ analyzer software (version 1.48).

The cell coverslips were placed in a 24-well plate, seeded transfected MDA-MB-231 at a density of 1.5 × 10⁵ cells per well, and incubated overnight at 37 °C. Then, the cells were fixed with 4% paraformaldehyde for 15 min at room temperature and blocked with 10% goat serum for 30 min. The cells were incubated with rabbit anti-NCALD (Proteintech, Wuhan, China) at 4 °C overnight. Cells were washed three times in PBS and incubated for 1 h at room temperature with FITC conjugated-goat anti-rabbit antibody (ZSGB-BIO, Beijing, China). After several washes, the cells were added hoechst 3342 (Beyotime, Suzhou, China) fluorescent dye solution, and coverslips were mounted to the slides using fluorescent mounting medium (PROLONG-GOLD, Thermo Fisher Scientific, MA, USA). Coverslips were imaged on the Nikon Eclipse Ti microscope (Nikon, Tokyo, Japan).

1.5 × 10³ transfected cells were seeded into each well of five 96-well plates. The cells were cultured for five consecutive days and added with 20 μl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Beyotime, Suzhou, China). Afterward, the cells were maintained at 37 °C, 5% CO₂ for another 4 h, and the MTT solution was removed, and 100 μl DMSO was added to each well. The 490-nm optical absorption value of each well was obtained, and the proliferation curves were established accordingly by using the GraphPad Prism 8.3.0 software.

In total, 500 cells were plated in a six-well plate and cultured in DMEM medium containing 10% FBS. Medium was changed every three days. MDA-MB-231 cells were cultured for two weeks, and MDA-MB-468 cells were cultured for four weeks. Then, cells were washed once with PBS solution and fixed with methanol for 15 min, and 500 μl of 0.5% crystal violet (Beyotime, Suzhou, China) was added to each well for 30 min. The colonies were imaged and counted.

Cell cycle analysis was performed using Cell Cycle Staining Buffer (Multi Sciences, Hangzhou, China) following the manufacturer’s protocol. Briefly, transfected cells were digested and washed, and then, resuspended in 500 μl of cell cycle staining buffer for 15 min. The cells were examined on a FACSCalibur (BD Biosciences, CA, USA) within 1 h.

For apoptosis analysis, FITC-Annexin-V/7-AAD double staining method was used (BD Biosciences, CA, USA). Transfected cells were collected, washed twice with cold PBS, centrifuged at 1000 rpm 5 min, and the supernatant discarded. Then, resuspended cells in 1× Binding buffer at a concentration of 1 × 10⁶ cell/ml. Transfer 100 μl of solution to a 1.5 ml culture tube, added 5 μl FITC-Annexin V, 5 μl 7-AAD to resuspend the cells, reacted for 15 min at the room temperature in the dark. After staining, added 400 μl of 1× Binding buffer to each tube, the apoptosis percentage were analyzed by FACSCalibur flow cytometer (BD) within 1 h.

Sequences containing all length or truncated of NCALD promoter were subcloned into pGL4.26-control vector (Addgene, MA, USA). For the reporter assay, HEK-293 T cells were plated onto 48-well plates, and the pGL4.26-NCALD or pGL4.26-truncated-NCALD, the Renilla luciferase plasmid (pRL-TK) were contransfected with pcDNA3.1, MIDEAS-AS1 overexpression plasmids, si-NC, si-MATR3 using Lipofectamine 2000. Finally, each group of cells was transfected with. After transfection for 48 h, the cells were harvested and assayed with a luciferase reporter assay system (Promega, Madison, WI, USA) according to the manufacturer’s instructions. The results were normalized against Renilla luciferase activity.

BALB/c nude mice (female, 4-week-old) with weights of about 20 g were procured from GemPharmatech (Jiangsu, China) for in vivo study. The transfected cells were injected subcutaneously into mice, and the width and length of formed tumors were monitored every 5 days after injection. Tumor volumes were calculated using the following formula: tumor volume = length × width² × 0.5. Following 25 days, the nude mice were sacrificed, and the collected tumors were weighed and then, fixed in formalin for immunohistochemical (IHC) staining and Hematoxylin and eosin (H&E) staining. To evaluate the influence of MIDEAS-AS1 on metastasis, the transfected cells were injected into the lateral tail veins of nude female mice (six mice per group). After about 2 months, the mice were sacrificed, and the lungs were collected to evaluate the number of pulmonary metastatic lesions and then, also were fixed for HE staining. The animal experiments were approved by the Ethical Committee on Scientific Research of Shandong University Qilu Hospital.

Paraffin-embedded tissues sections were deparaffinized and rehydrated with xylene, gradient ethanol (70, 80, 95, 100%) and incubated for 20 min in 3% H₂O₂ to block endogenous peroxidase activity. After blocking with 10% goat serum and then incubated with 50 μl rabbit anti-Ki67 (Proteintech, Wuhan, China), rabbit anti-MATR3 (Proteintech, Wuhan, China), rabbit anti-NCALD (Proteintech, Wuhan, China) at 4 °C overnight. Subsequently, sections were incubated with an appropriate secondary antibody for 20 min at 37 °C and developed with diaminobenzidine and stained with hematoxylin for 5 min. After each treatment, the sections were dehydrated, cleared, mounted, and viewed under the microscope.

Statistical test in this study was performed using GraphPad Prism software 8.3.0 software. Results were expressed as the mean ± standard deviation (SD). The Kaplan–Meier method and log-rank test were used to analyze survival rate. Differences of two or more groups were analyzed using the student’s t-test or one-way/two-way ANOVA. Statistical significance was determined as p < 0.05. All tests were conducted in triplicates.

To explore the potential role of lncRNAs in TNBC, we first analyzed the differentially expressed lncRNAs using GEO (Fig. 1A) and TCGA databases (Additional file 1: Fig. S1A). Then, integrated analysis was performed on the lowly expressed lncRNAs at the GSE60689 and TCGA lncRNA databases (Fig. 1B). Among them, the newly named lncRNA MIDEAS-AS1 attracted our attention. A search of the UCSC database (http://​genome.​ucsc.​edu/​) and CPC2 database (http://​cpc2.​gao-lab.​org) revealed that MIDEAS-AS1 was located on the human chromosome 14q24.3 and had three exons, without protein-coding capabilities (Additional file 1: Fig. S1B-C). Based on TCGA datasets, we validated that MIDEAS-AS1 was markedly low expressed in breast cancer tissues compared to adjacent normal tissues (Fig. 1C). Additionally, we also analyzed the expression levels of MIDEAS-AS1 in different subtypes of breast cancer. Among patients with breast cancer, the expression of MIDEAS-AS1 was lower in TNBC than in other subtypes through the GSE21653 database (Additional file 1: Fig. S1D). To further validate the decrease in MIDEAS-AS1 in breast cancer, we examined the expression of MIDEAS-AS1 using qRT-PCR in breast cancer tissues and their adjacent normal breast tissues, as well as, TNBC and non-TNBC tissues. Compared to the adjacent normal tissues, the expression of MIDEAS-AS1 was significantly downregulated in breast cancer tissues (Fig. 1D). Meanwhile, we also confirmed the results that patients with TNBC had a significantly lower expression of MIDEA-AS1, compared with non-TNBC (Fig. 1E). Moreover, we further examined the MIDEAS-AS1 expression level in breast cancer organoids and normal organoids, and the result revealed that MIDEAS-AS1 was significantly low expressed in the breast cancer organoids (Fig. 1F). In addition, fluorescence in situ hybridization (FISH) assay revealed similar results, and downregulated expression of MIDEAS-AS1 was identified in breast cancer organoids compared to the normal organoid (Fig. 1G). There were some literatures reported that MIDEAS-AS1 was associated with tumor stage and tumor-node-metastasis (TNM) stage [26]. To assess the clinical significance of MIDEAS-AS1 in TNBC, we analyzed the relationship between MIDEAS-AS1 expression level and clinicopathological characteristics. We found that the low expression level of MIDEAS-AS1 was significantly correlated with larger tumor size and higher pathological grade (Table1). Moreover, Kaplan–Meier survival analysis showed that lower expression level of MIDEAS-AS1 was correlated with significantly poorer overall survival (OS) rate (Fig. 1H) and progression-free survival (PFS) rate of patients with TNBC (Fig. 1H-I). Together, these data indicated that MIDEAS-AS1 was downregulated in TNBC and low expression of MIDEAS-AS1 was associated with poor prognosis of TNBC patients.

To explore the potential roles of MIDEAS-AS1, we transfected the MIDEAS-AS1 overexpression plasmids into MDA-MB-231 and MDA-MB-468 cells, and the overexpression efficiency was determined by qRT-PCR (Additional file 1: Fig. S2A, S2B). Results from MTT assay indicated that the proliferation ability of MDA-MB-231 and MDA-MB-468 cells with MIDEAS-AS1 overexpression was reduced compared to control group (Fig. 2A). Meanwhile, overexpression of MIDEAS-AS1 also inhibited the proliferation of breast cancer organoids (Fig. 2B). Moreover, overexpression of MIDEAS-AS1 repressed cell colony-forming activity (Fig. 2C-D). Additionally, flow cytometry was used to detect the apoptosis rate and cell cycle after different treatments. Our results indicated that MIDEAS-AS1 overexpression led to remarkably increased apoptosis rate (Fig. 2E) and cell number in G1 phase (Additional file 1: Fig. S2C). On the other hand, MIDEAS-AS1 knockdown promoted the proliferation and colony-formation abilities of TNBC cells (Fig. 2A, D). Furthermore, the apoptosis rate was obviously decreased and the number of cells in G1 phase was less in MIDEAS-AS1 knockdown group compared to that in control group (Fig. 2E, Additional file 1: Fig. S2C). We then investigated the effect of MIDEAS-AS1 on TNBC cell migration and invasion using transwell assays and wound healing assay in MDA-MB-231 and MDA-MB-468 cells. The results of transwell assays showed a significant reduction in the migratory and invasive abilities of MDA-MB-231 and MDA-MB-468 cells after MIDEAS-AS1 overexpression (Fig. 2F). Furthermore, wound healing assay indicated that the wound closure area of MDA-MB-231 and MDA-MB-468 cells was dramatic decreased following MIDEAS-AS1 overexpression (Additional file 1: Fig. S2D). Consistently, MIDEAS-AS1 knockdown led to increased migratory and invasive abilities of TNBC cells (Fig. 2G). Given that epithelial-mesenchymal transition (EMT) is one of the major mechanisms for cancer metastasis, we further examined the effect of MIDEAS-AS1 on the expression of EMT-related marker proteins by Western blot. The results showed that N-cadherin, vimentin, fibronectin and MMP9 proteins were downregulated, while the E-cadherin protein was upregulated in MDA-MB-231 and MDA-MB-468 cells after overexpression of MIDEAS-AS1 (Fig. 2H). On the contrary, N-cadherin, vimentin, fibronectin and MMP9 proteins were upregulated, and the E-cadherin protein was downregulated in TNBC cells after MIDEAS-AS1 knockdown (Fig. 2I). Therefore, MIDEAS-AS1 could regulate the EMT process to modulate TNBC metastasis. Collectively, these results indicated that MIDEAS-AS1 played critical roles in suppressing cell proliferation and mobility of TNBC cells.

It had been suggested that the regulatory mechanisms of lncRNAs were closed associated with their subcellular localization [25]. Therefore, we firstly detected the subcellular localization of MIDEAS-AS1 in TNBC cells. Following isolation of the nuclear and cytoplasmic RNA in MDA-MB-231 and MDA-MB-468 cells, qRT-PCR analysis demonstrated that MIDEAS-AS1 was mainly located in the nucleus (Fig. 3A). Furthermore, FISH assay also obtained the similar results (Fig. 3B), indicating the potential role of MIDEAS-AS1 in transcriptional regulation through acting as a scaffold for TFs [25, 27]. To identify proteins interacted with MIDEAS-AS1, RNA pull-down assay was performed. We incubated MDA-MB-231 cell lysates with biotinylated MIDEAS-AS1 or its antisense RNA transcribed in vitro, and the silver staining and mass spectrometry (MS) identified MATR3 as one of the major proteins in the MIDEAS-AS1 pull-down precipitations (Fig. 3C). Moreover, mass spectrometry analysis found that MIDEAS-AS1 interacted with MATR3 at “DLSAAGIGLLAAATQSLSMPASLGR” and “YQLLQLVEPFGVISNHLILNK” peptide sequences (Fig. 3D). The specific binding between MATR3 protein and MIDEAS-AS1 was further confirmed by RNA pull down following Western blot (Fig. 3E). Then, we performed RNA immunoprotein (RIP) assay using flag antibodies and revealed that MIDEAS-AS1 was significantly enriched in MATR3 immunoprecipitations (Fig. 3F). Meanwhile, RNA FISH technology combined with immunofluorescence analysis demonstrated the co-localization of MIDEAS-AS1 and MATR3 protein in MDA-MB-231 cells (Fig. 3G). To identify the specific binding regions between MIDEAS-AS1 and MATR3, we constructed vectors containing full-length MIDEAS-AS1 or three truncated sequences (Fig. 3H). The RNA pull down assay indicated that deleting the sequence of 217–439 bp (Δ2 vector) could abolish the MIDEAS-AS1-MATR3 interaction, which was consist with the prediction results obtained from catRAPID database (Additional file 1: Fig. S3A). Furthermore, we wonder whether MIDEAS-AS1 could regulate the expression of MATR3. Significantly, overexpression or knockdown of MIDEAS-AS1 shows no effect on the RNA and protein expression levels of MATR3 (Additional file 1: Fig. S3B), indicating that MIDEAS-AS1 was not involved in the post-transcriptional regulation of MATR3. Collectively, these results revealed that MIDEAS-AS1 specifically interacted with MATR3 in TNBC cells.

Given the critical role of nuclear matrix-associated protein MATR3 in transcriptional regulation, we wonder whether MIDEAS-AS1 could regulate the expression of downstream genes through interacting with MATR3 to further affect the progression and metastasis of TNBC. We performed an RNA-sequencing analysis on MIDEAS-AS1-overexpressing and control cells, and 471 differentially expressed genes were identified (Fig. 4A-B), including 252 up-regulated and 219 down-regulated genes. Meanwhile, the top 20 up-regulated genes are shown in Fig. 4C. Due to the tumor-suppressive role of MIDEAS-AS1 in TNBC, we integrated the down-regulated genes in breast cancer tissues from TCGA database and the up-regulated genes after MIDEAS-AS1 overexpression in TNBC cells, and then, 9 candidate genes were selected as potential downstream targets of MIDEAS-AS1 (Fig. 4D). It has been reported in the literature that antisense lncRNA may affect the expression of sense gene [28], so we also examined the influence of MIDEAS-AS1 on its sense gene (MIDEAS) by qRT-PCR. However, the results showed that MIDEAS-AS1 did not affect the expression of MIDEAS (Additional file 1: Fig. S4B). Meanwhile, the qRT-PCR results revealed that the expression of AF131215.5, CXCL1, and NCALD was remarkably increased after MIDEAS-AS1 overexpression and reduced after MIDEAS-AS1 knockdown (Additional file 1: Fig. S4A). CXCL1 (C-X-C motif chemokine ligand 1) was the most abundant chemokine secreted by TAMs, and CXCL1 could promote migration, invasion ability, and EMT in breast cancer [29, 30]. AF131215.5 was a lncRNA whose function remained unclear in cancer [31, 32]. Interestingly, NCALD (neurocalcin delta) expression was lower in lung adenocarcinoma tissues [33], and patients with higher NCALD levels exhibited a higher survival rate [34, 35]. Among these genes, we hypothesized that NCALD, the tumor suppressor gene, might be a potential target of MIDEAS-AS1 in breast cancer. Significantly, NCALD expression was markedly downregulated in breast cancer tissues compared with normal tissues (Additional file 1: Fig. S4C-D). Moreover, immunohistochemistry (IHC) showed higher expression of NCALD in normal tissues than breast cancer tissues according to Human Protein Atlas database (Additional file 1: Fig. S4E). These results indicated that NCALD might play a significant tumor suppressor role in breast cancer. Therefore, NCALD was selected as the functional downstream mediator for MIDEAS-AS1-mediated cell migration and metastasis in TNBC. Significantly, the qRT-PCR and Western blot results revealed that the expression of NCALD was remarkably increased after MIDEAS-AS1 or MATR3 overexpression and reduced after MIDEAS-AS1 or MATR3 knockdown in MDA-MB-231 and MDA-MB-468 cells (Fig. 4E-F, Additional file 1: Fig. S4E), indicating the regulatory effect of MIDEAS-AS1 and MATR3 on the expression of NCALD. Furthermore, rescue experiments were performed to demonstrate the causal link between MIDEAS-AS1, MATR3 and NCALD. The qRT-PCR analysis indicated that MIDEAS-AS1 overexpression promoted the expression of NCALD, while MATR3 knockdown could weaken the increased tendency of NCALD expression caused by overexpression of MIDEAS-AS1 in MDA-MB-231 and MDA-MB-468 cells (Fig. 4G). Furthermore, we found that the increasing trend of NCALD protein level caused by MIDEAS-AS1 overexpression could be attenuated by MATR3 knockdown in MDA-MB-231 and MDA-MB-468 cells (Fig. 4H). Then, we further investigated the transcriptional regulation effect of MIDEAS-AS1 and MATR3 on the expression of NCALD. The dual luciferase reporter assay showed that overexpression of MIDEAS-AS1 or MATR3 substantially increased the luciferase activity of NCALD vectors, while knockdown of MIDEAS-AS1 or MATR3 significantly inhibited the luciferase activity of NCALD vectors in HK293T cells (Fig. 4I). Furthermore, rescue experiments revealed that the promoter activity of NCALD was activated by MIDEAS-AS1 overexpression, which was inhibited after co-transfection with si-MATR3 (Fig. 4J). These results indicated that MIDEAS-AS1 could regulate the expression of NCALD through modulating the function of MATR3. To identify the specific binding regions of MIDEAS-AS1 and MATR3 on the NCALD promoter, we ectopically expressed full-length NCALD promoter (−2000 to + 100 bp) as well as five mutants: the truncated NCALD-1 (−2000 to −1564 bp) mutant, the truncated NCALD-2 (−1563 to −1094 bp) mutant, the truncated NCALD-3 (−1093 to −654 bp) mutant, the truncated NCALD-4 (−653 to −269 bp) mutant, and the truncated NCALD-5 (−268 to + 100 bp) mutant (Fig. 4K), and a luciferase assay was performed. We found that the luciferase activity was most significantly increased in −1563 bp to −1094 bp, indicating the presence of positive regulatory elements, which enhanced NCALD transcription in this region (Fig. 4L). In addition, the ChIP followed by qPCR assays was subsequently performed in Flag-MATR3 transfected breast cancer cells with or without MIDEAS-AS1 overexpression to determine the effect of MIDEAS-AS1 on MATR3 recruitment to the NCALD promoter. Consistent with the luciferase assay, MIDEAS-AS1-MATR3 complex was identified to significantly bound to −1563 bp to −1094 bp sites of the NCALD promoter region in MDA-MB-231 and MDA-MB-468 cells (Fig. 4M). These findings suggested that MIDEAS-AS1-MATR3 complex could enhance NCALD transcription by directly binding to its promoter.

Previous studies had found that NCALD was not only involved in cell apoptosis, cell cycle progression and other biological processes in several cancers [36], but also associated with the prognosis of cancers [34, 35]. However, there is no report about the function of NCALD in breast cancer. Our above results revealed downregulated expression of NCALD in breast cancer tissues and breast cancer cells, indicating the tumor-suppressive role of NCALD in breast cancer. To confirm the function of NCALD, we transfected NCALD overexpressing vectors and si-NCALD into MDA-MB-231 and MDA-MB-468 cells, and the overexpression and knockdown efficiency were determined by qRT-PCR and Western blot assays (Fig. 5A-B). The MTT and colony formation assay results indicated that NCALD overexpression reduced TNBC cell proliferation and colony formation abilities (Fig. 5C, D). Additionally, the apoptosis rate was remarkably increased and the number of cells in G1 phase was increased after NCALD overexpression (Fig. 5F, Additional file 1: Fig. S5A). On the other hand, NCALD knockdown led to significantly increased proliferation, colony formation abilities of TNBC cells (Fig. 5C, E). Moreover, the apoptosis rate was decreased and the number of the cells in G1 phase was less in NCALD knockdown cells than that in control cells (Fig. 5F, Additional file 1: Fig. S5A). Additionally, transwell assay and wound healing assay revealed that NCALD overexpression significantly reduced TNBC cell migration and invasion abilities (Fig. 5G, Additional file 1: Fig. S5B), while NCALD knockdown led to increased migration and invasion abilities in TNBC cells (Fig. 5H, Additional file 1: Fig. S5B). These data demonstrated that NCALD served as a suppressive functional factor in TNBC progression.

Previous literature reported that NCALD might be related to ERK1/2 signaling pathway, NF-κB signaling pathway, TGF-β signaling pathway and immune response pathway in ovarian cancer [35]. We further investigated the potential molecular mechanism caused by the change of NCALD in TNBC cells. Western blot analysis indicated that overexpression of NCALD would significantly inhibited phosphorylate p-65, but not TGF-β and, p-ERK1/2 in MDA-MB-231 and MDA-MB-468 cells (Fig. 5I), while NCALD knockdown brought about opposite results (Fig. 5J). Collectively, our results demonstrated that NCALD inhibited TNBC cell proliferation, migration, and invasion by suppressing NF-κB signaling pathways.

Our previous results have showed that the association between MIDEAS-AS1 and MATR3 played significant role in initiating NCALD transcription. To further confirm whether MIDEAS-AS1 exerts tumor-suppressive functions via modulating NCALD, we co-transfected si-NCALD and MIDEAS-AS1-overexpressed plasmid in TNBC cell lines. The MTT and colony formation experiments indicated that NCALD knockdown remarkably rescued the proliferation ability of MDA-MB-231 and MDA-MB-468 cells inhibited by the overexpression of MIDEAS-AS1 (Fig. 6A-B). Moreover, the transwell and wound healing assays showed that NCALD knockdown could partially recover the cell migration and invasion abilities reduced by MIDEAS-AS1-overexpression (Fig. 6C-D). Moreover, Western blot analysis revealed that NF-κB pathway-related proteins were decreased after overexpression of MIDEAS-AS1, and that was increased after co-transfection of MIDEAS-AS1 overexpression plasmid and si-MATR3 (Fig. 6E). Taken together, MIDEAS-AS1 associates with MATR3 to initiate NCALD transcription and inhibits NF-κB signaling pathway, which further affects the TNBC progression.

In order to further evaluate the biological function of MIDEAS-AS1 in vivo, a subcutaneous xenograft model was first constructed. The MDA-MB-231 cells stably transfected with MIDEAS-AS1 overexpressing vectors or control vectors were subcutaneously injected to the flanks of nude mice. The result showed that the tumor weight and tumor volume were significantly inhibited in xenografts of MIDEAS-AS1 overexpressing group compared with those in the control group (Fig. 7A–C). Furthermore, immunohistochemistry (IHC) assays confirmed that MIDEAS-AS1 overexpression caused decreased Ki67 expression and increased NCALD expression, but did not affect MATR3 expression (Fig. 7D-E). In addition, we further constructed pulmonary metastasis model through intravenously injecting MDA-MB-231 cells stably expressing MIDEAS-AS1 or control vectors to compare the metastatic abilities in vivo. The results showed that mice injected with MIDEAS-AS1-overexpressing TNBC cells had no or fewer metastatic foci compared with the control group (Fig. 7F–H). Together, these results indicated that MIDEAS-AS1 played a critical role in inhibiting TNBC progression and metastasis.