LINC00662 enhances cell progression and stemness in breast cancer by MiR-144-3p/SOX2 axis

35 paired BC tissues and adjacent non-tumorous tissues were acquired from BC patients in our hospital. The clinicopathological features of patients were shown in Additional file 3: Table S1 and Additional file 4: Table S2. Prior to surgery, no patients had received radiotherapy or chemotherapy. Patients with other kinds of tumors were excluded from this study. The patients put their signatures to the written informed consents prior to sample collection. Subsequent to surgical resection, the excised tissues were snap-frozen in liquid nitrogen and stored at − 80 °C. Kaplan–Meier’s analysis was applied for studying overall survival (OS) with a log-rank test. This study was supported by the ethical committee of our Hospital.

Normal human mammary epithelial cell line (MCF-10A) and human BC cell lines (MDA-MB-453, MDA-MB-231, MCF-7, and MDA-468) were attained from ATCC (Manassas, VA, USA). MCF-10A cells were kept in DMEM/F12 (D9785, Sigma-Aldrich, St. Louis, MO, USA) with the supplementation of 5% HS (23491-45-4, Sigma-Aldrich), 20 ng/mL epidermal growth factor (EGF; PHG0311, Thermo Fisher Scientific, Rockford, IL, USA), 0.5 μg/mL hydrocortisone (211H-500, Sigma-Aldrich), 1% NEAA (M7145, Sigma-Aldrich), 10 μg/mL insulin (12,643, Sigma-Aldrich) and 1% P/S (PB180120, Procell, Wuhan, China). MCF-7 cells were cultivated in DMEM (D9785, Sigma-Aldrich) with 10% fetal bovine serum (FBS; Gibco, Rockville, MD, USA), 10 µg/mL human insulin and 1 µM 4-hydroxytamoxifen (SML1666, Sigma-Aldrich). MDA-MB-231, MDA-468 and MDA-MB-453 cells were cultivated in ATCC-formulated Leibovitz’s L-15 Medium (11,415,049, Thermo Fisher Scientific) with 10% FBS. All the cells were maintained at 37 °C with 5% CO₂. The STR reports were provided in Additional file 5.

BC cells were subjected to transfection with specific short hairpin RNAs (shRNAs) against LINC00662 (sh-LINC00662#1, sh-LINC00662#2, and sh-LINC00662#3) and their corresponding NCs, and the pcDNA3.1 vector against SOX2 or LINC00662 and the empty vector. The plasmids were synthesized by Genechem (Shanghai, China). Besides, miR-144-3p mimics or miR-144-3p inhibitors and corresponding negative controls (miR-NCs) constructed by GenePharma (Shanghai, China) were co-transfected with sh-LINC00662 into BC cells. These transfected cells were seeded in the 24-well plate without FBS. Then, transfection was carried out using 1 μL Lipofectamine 3000 (L3000075, Invitrogen, Carlsbad, CA, USA) in 50 μL serum-free medium (Invitrogen) on the basis of supplier’s suggestions.

QRT-PCR was implemented as per the previous literature [23]. Briefly, Trizol (15596018/15596026, Invitrogen) was used for total RNA extraction from BC cells. Then, extracted RNAs were reverse transcribed to cDNA via PrimeScript™ RT Master Mix (RR036Q, Takara, Tokyo, Japan). Subsequently, qRT-PCR was conducted to investigate gene expression on the LightCycler96 real-time PCR system (Roche, Basel, Switzerland) using the SYBR Green PCR Master mix (4309155, Applied Biosystems, Foster city, CA, USA). 40 cycles of PCR were achieved at 95 °C for 30 s and 57 °C for 30 s. For extension, the temperature was set as 72 °C for 2 min. After the purification, results were calculated in accordance with the 2^−ΔΔCt approach with GAPDH or U6 being the control. This assay was implemented in triplicate and each replicate contained three repetitions.

Soft agar was applied in assays as previously described [27]. Briefly, transfected cells (500 cells/well) were inoculated into 6-well plates. The plates were subjected to incubation at 37 °C with 5% CO₂ for two weeks, with culture medium changed at regular intervals. After that, phosphate-buffered saline (PBS; P4417, Sigma-Aldrich) was utilized to wash cells. Then, cells were fixed by 10% formalin (Z2902, Sigma-Aldrich) and stained with 0.5% crystal violet (V5265, Sigma-Aldrich). For the formation efficiency calculation, 5 fields were randomly selected to be counted.

Matrigel-uncoated or Matrigel pre-coated transwell inserts (Millipore, Bedford, MA, USA) were employed for evaluation of cell migration or invasion. Transfected cells were inoculated to the upper chambers in medium with no FBS, while 600 μL of 100% FBS was added to the lower chambers. After 12 or 16 h, cells in the upper chambers were slightly wiped using a cotton swab. Migrated or invaded cells into the lower chambers were subjected to fixation with methanol (34,860, Sigma-Aldrich) and staining with crystal violet, followed by observation under an Olympus inverted microscope (Tokyo, Japan, 10 × 10). Lastly, the cells were counted by MShot Image Analysis System and 5 fields per chamber were observed.

Sphere formation assay was carried out as reported previously [24]. Concisely, transfected MDA-MB-231 or MCF-7 cells were inoculated in ultra-low attachment 6-well plates (Corning, NY, USA; 5 × 10³ cells per well) for 10–14 days. Cells were maintained in DMEM/F12 serum-free medium (PM150312A, Procell) added with 5 μg/mL insulin (Sigma-Aldrich), 0.4% bovine serum albumin (BSA; A1933, Sigma-Aldrich), 2% B27 (MAB1285, Sigma-Aldrich), 20 ng/mL basic fibroblast growth factor (bFGF; PHG0369, Thermo Fisher Scientific), and 20 ng/mL EGF (Peprotech, Rocky Hill, NJ, USA). Generated spheres were photographed and then counted using a light microscope (Zeiss, Oberkochen, Germany). Sphere formation efficiency is calculated as (number of spheres/number of inoculated cells) × 100%, using MShot Image Analysis System. Only spheres with diameters greater than 75 μm were counted.

Western blot were carried out in line with standard protocols as previously described [25]. Anti-OCT4 (ab181557), anti-SOX2 (ab97959), anti-Nanog (ab109250) and anti-GAPDH (ab8245) were commercially attained from Abcam (Cambridge, USA). Briefly, total proteins isolated from BC cells were processed with SDS-PAGE for separation and transferred to PVDF membranes, followed by being blocked with 5% skim milk. Afterwards, the membranes were subjected to incubation with primary antibodies overnight 4 °C. The membranes were then washed in TBST, followed by incubation with secondary antibodies for 1 h at room temperature. The blots were then visualized using ECL. GAPDH was applied as the internal control.

Nuclear and Cytoplasmic Extraction Reagents (78,833, Thermo Fisher Scientific, Waltham, MA, USA) was employed to perform the subcellular fractionation as per the supplier’s protocol. The cells were placed on ice and resuspended in the buffer solution. After 10 min of incubation, the cells were centrifuged for the separation of cytoplasm and nucleus. The RNAs were extracted using Trizol. QRT-PCR was implemented to measure the extracted RNAs. U6 or GAPDH acted as the nuclear or cytoplasmic control.

RIP assays were carried out as described previously [23]. Briefly, anti-Ago2 antibodies (TS-10X10ML-U, Millipore) or anti-IgG antibodies (MABE-253, Sigma-Aldrich) were incubated with protein A/G agarose beads (78,610, Thermo Fisher Scientific) at 4 °C for the whole night. BC cells were washed in ice-cold PBS, followed by being lysed. Afterwards, cell lysates were incubated with the A/G agarose beads conjugated with anti-Ago2 antibodies or anti-IgG antibodies overnight at 4 °C. When the immunoprecipitation was fully achieved, RNAs were extracted via Trizol from A/G agarose beads and antibodies. Finally, qRT-PCR was applied to measure the relative expression of RNAs (LINC00662, miR-144-3p or SOX2).

CCK-8 (Dojindo, Kumamoto, Japan) was performed for assessing the proliferative ability of BC cells. The transfected BC cells were planted into 96-well plates, followed by the addition of CCK-8 solution. Next, the cells underwent 24, 48, 72 and 96 h of incubation, followed by addition of 100 μL CCK-8 solution to each well. Afterwards, the incubation lasted for another 4 h. The absorbance (OD) was examined at 450 nm through a microplate reader (Bio-Rad, Hercules, CA, USA).

The sequences of wild type or mutant type SOX2/LINC00662 were sub-cloned into pmirGLO dual-luciferase vectors (Promega, Madison, WI, USA). Then, the luciferase reporter vectors were transfected into BC cells along with miR-144-3p mimics or miR-144-3p mimics + pcDNA3.1/LINC00662 or miR-NC via Lipofectamine 3000. Firefly and Renilla luciferase activities were analyzed by Dual-luciferase reporter assay system (Promega), normalized to the empty vectors. The signals from luciferase were detected by use of fluorescence microscope (DMI8, Leica, Wetzlar, Germany).

Twenty nude mice (6-week-old, female) procured from Shanghai Laboratory Animal Co., Ltd. (SLAC; Shanghai, China) were adopted for the in vivo assay, which was permitted by the Institutional Animal Care and Use Committee of our institution (2021-AE027). The transfected MDA-MB-231 cells were subcutaneously injected into the right flank of mice. Then, the nude mice were put into the euthanasia cage and exposed to carbon dioxide (CO₂) for 5 min. The CO₂ flow rate were displaced 10% of the cage volume per minute. Tumor growth was monitored every 4 days and tumor volume was estimated by length and width. 4 weeks post injection, the nude mice were sacrificed. Tumors resected from the nude mice were weighed with an electronic scale.

Ki67 antibodies (ab16667, 1/200, Abcam, Cambridge, MA, USA) and caspase 3 antibodies (ab32351, 1/100, Abcam) were used for immunohistochemical analysis. The tissues from xenograft tumor treated with hydrogen peroxide were incubated with primary antibodies against Ki67 and caspase 3 overnight at 4 °C. Next, secondary antibodies were added to tissues for incubation at 37 °C for 1 h. Subsequently, tissues were stained with DAB and the expression of Ki67 or caspase 3 was measured through a light microscope (Zeiss, Oberkochen, Germany).

All assays were carried out in triplicate. Student’s t-test or one-way ANOVA was utilized for the comparison between groups. R square analysis was applied for analyzing the correlation between SOX2, LINC00662 and miR-144-3p expressions. The statistical analyses in this study were conducted with the application of GraphPad Prism (GraphPad, La Jolla, CA, USA). P < 0.05 was considered as the threshold for statistical significance.

We applied qRT-PCR analysis to evaluate LINC00662 expression in breast tumor tissues and corresponding adjacent normal tissues. It was uncovered that expression of LINC00662 was overtly elevated in tumor tissues relative to that in normal tissues (Fig. 1A). 35 patients were split into two groups, one with high expression of LINC00662, the other with low expression. As shown in Additional file 1: Fig. S1A, statistical analyses indicated that LINC00662 expression was negatively linked to OS. We detected LINC00662 expression respectively in normal breast cell line MCF-10A and cancerous cell lines (MDA-MB-231, MCF-7, MDA-468, and MDA-MB-453) as well, finding LINC00662 expression was evidently up-regulated in cancerous cell lines than that in normal cell line. Based on the above-mentioned results and the previous study [14], we conjectured that LINC00662 is an oncogene in BC cells (Fig. 1B). To further study the association between LINC00662 and biological behaviors of BC cells, we firstly selected two cancerous cell lines MDA-MB-231 and MCF-7 for our further studies due to the prominent expression of LINC00662 they presented. Next, we transfected sh-LINC00662#1/2/3 into MDA-MB-231 and MCF-7 cell lines, finding that LINC00662 expression was significantly silenced after the transfection. We chose sh-LINC00662#1 and sh-LINC00662#2 for follow-up experiments, due to the higher efficiency (Fig. 1C). Subsequently, cell colony formation assay was implemented in BC cells to detect cell proliferation. The results showed that cell colonies were significantly decreased after knockdown of LINC00662, indicating that LINC00662 promotes BC cell proliferation (Fig. 1D). Afterwards, western blot was conducted to detect the protein level of apoptosis marker, cleaved caspase 3 in BC cells after LINC00662 ablation. It was shown that cleaved caspase 3 level was highly increased, indicating that LINC00662 inhibits BC cell apoptosis (Additional file 1: Fig. S1B). Transwell assays were carried out next to observe migration and invasion capacity in BC cells. As evidenced by the decrease in migrated and invaded cells, cell migratory and invasive abilities were attenuated after knockdown of LINC00662 (Fig. 1E). Furthermore, we performed sphere formation assay in BC cells to investigate the effects of LINC00662 on stemness ability. We examined cell sphere numbers after knockdown of LINC00662 and found the number was decreased sharply, suggesting a decline in sphere formation efficiency (Fig. 1F). Also, we evaluated the expression level of several stemness related proteins, OCT4, SOX2 and Nanog after knockdown of LINC00662, finding the levels of these proteins were all dropped after LINC00662 knockdown, further confirming LINC00662 up-regulation promotes stemness of cancerous cells (Fig. 1G). For further verification, we performed gain-of-function experiments in MDA-MB-453 cells using pcDNA3.1/LINC00662. It was shown by the results that cell proliferation, migration, invasion and stemness were facilitated and cell apoptosis was inhibited after LINC00662 overexpression (Additional file 1: Fig. S1C–F). Taken together, LINC00662 plays an oncogenic role in BC cells.

We used lnclocator (http://​www.​csbio.​sjtu.​edu.​cn/​bioinf/​lncLocator/​) for prediction of the subcellular localization of LINC00662. The results of prediction disclosed that LINC00662 is prominently located in the cytoplasm (Additional file 2: Fig. S2A). As shown in subcellular fractionation assay, the cytoplasmic fraction of LINC00662 was more enriched than its nuclear fraction, verifying a possibility of ceRNA mechanism (Fig. 2A). To seek the most matched miRNA of LINC00662, we retrieved starBase V2.0 (http://​starbase.​sysu.​edu.​cn/​) and found four miRNAs (miR-144-3p, miR-15a-5p, miR-16-5p, and miR-34a-5p) which share binding sites with LINC00662 (Fig. 2B). We conducted qRT-PCR in MDA-MB-231 cells to select the one that may be prominently modulated by LINC00662. It was unmasked that miR-144-3p expression was strikingly elevated relative to other candidate miRNAs after knockdown of LINC00662 (Fig. 2C). To further confirm the mechanism between LINC00662 and miR-144-3p, we performed RIP assay in BC cells. It was disclosed that miR-144-3p and LINC00662 expressions were enriched in Anti-Ago2 group, confirming that miR-144-3p and LINC00662 coexists in RNA-induced silencing complex (Fig. 2D). The binding sites between LINC00662 and miR-144-3p and the base sequence of LINC00662 mutant type were predicted by starBase V2.0. It was disclosed by luciferase reporter assay that the luciferase activity of WT group was significantly attenuated after miR-144-3p overexpression while that of Mut group had no marked change, verifying that LINC00662 binds with miR-144-3p (Fig. 2E). Next, a series of rescue experiments were conducted in BC cells transfected with sh-NC, sh-LINC00662#1 or sh-LINC00662#1 + miR-144-3p inhibitor. First, CCK-8 results demonstrated that cell proliferation was decreased after knockdown of LINC00662 and was then resumed after co-transfection with miR-144-3p inhibitor (Fig. 2F). As exhibited in Transwell assays, cell migratory and invasive capacities were impaired after knockdown of LINC00662 and then reversed after co-transfection with miR-144-3p inhibitor (Fig. 2G). The results of western blot showed that cleaved caspase 3 level was increased by LINC00662 depletion and then counteracted by miR-144-3p inhibition (Additional file 2: Fig. S2B). In addition, sphere formation efficiency which was declined after LINC00662 knockdown, was then reversed by miR-144-3p knockdown (Fig. 2H). Furthermore, we performed western blot to detect stemness-related protein level and found levels of OCT4, SOX2 and Nanog shared the same changes as the above results (Fig. 2I). Taken together, LINC00662 regulates BC cell progression by competitively binding to miR-144-3p.

Next, we probed into the downstream mRNA of miR-144-3p. Previous studies verified that SOX2, SOX4, OCT4, and Nanog are involved in cell migration, invasion and stemness [26, 27]. Thus, we conjectured that miR-144-3p may mediate cell migration, invasion and stemness via targeting the above genes. Subsequently, we detected the expressions of these genes in MCF-10A and BC cells using qRT-PCR, finding that their expression all rose significantly after knockdown of LINC00662 (Fig. 3A). According to the prediction by starBase V2.0, only SOX2 was found sharing binding sites with miR-144-3p (Fig. 3B). Moreover, SOX2 expression level in BC patients was claimed to be positively related to OS [28, 29]. Thus, SOX2 was supposed as the downstream target of miR-144-3p. Subsequently, several functional rescue assays were conducted in BC cells after transfection of sh-NC, sh-LINC00662#1 or sh-LINC00662#1 + pcDNA3.1/SOX2. It was displayed by CCK-8 and Transwell assays that cell proliferative, migratory and invasive abilities were suppressed by knockdown of LINC00662, and was then resumed by overexpression of SOX2 (Fig. 3C, D). As shown in Additional file 2: Fig. S2C, cleaved caspase 3 level was increased by LINC00662 knockdown and then counteracted by SOX2 overexpression. Similarly, cell stemness ability and stemness-related protein level were both decreased markedly after the silencing of LINC00662 and recovered after co-transfection with pcDNA3.1/SOX2 (Fig. 3E, F). Moreover, RIP and luciferase reporter assays were carried out to further confirm the regulatory mechanism among LINC00662, miR-144-3p and SOX2. The results of RIP suggested that SOX2, LINC00662 and miR-144-3p co-exists in the RISC (Fig. 3G). It was unmasked by luciferase reporter assay that luciferase activity reduced by miR-144-3p mimics was reversed by LINC00662 (Fig. 3H). Figure 3G, H confirmed the post-transcriptional regulation character and ceRNA mechanism among these three. To sum up, LINC00662 targets miR-144-3p/SOX2 axis to modulate BC cell progression.

It was detected through qRT-PCR that miR-144-3p was notably overexpressed in tumor tissues compared with that in normal tissues while SOX2 expression was restrained in BC tissues (Fig. 4A, B). Subsequently, correlation analysis using R square graphs indicated that LINC00662 expression is negatively associated with miR-144-3p expression and is positively linked to SOX2 expression; and SOX2 expression is negatively related to miR-144-3p expression in BC tissues (Fig. 4C−E). Taken together, relative expressions of LINC00662 and SOX2 are negatively linked to miR-144-3p expression that is inhibited in BC tissues.

Ten nude mice transplanted with subcutaneous sarcoma were chosen as the vivo specimens. Tumors were monitored through the whole assay and changes were observed after the transfection of sh-LINC00662 into MDA-MB-231 cell line (Fig. 5A). Compared with sh-NC group, an obvious decrease of tumor growth was observed (Fig. 5B). Tumor weight and volume were showed with a significant decrease after knockdown of LINC00662 as well (Fig. 5C, D). Moreover, the levels of stemness-related proteins were decreased as well after silencing LINC00662 (Fig. 5E). In addition, IHC staining was used to detect the levels of proliferation marker Ki67 and cleaved caspase 3 to respectively analyze the proliferation and apoptosis of BC tissues excised from nude mice. The result showed that proliferation was sharply shaved while the apoptosis was overtly promoted in sh-LINC00662 group compared with the control group (Fig. 5F). Correlation analysis using R square graphs were utilized to detect the relationship between LINC00662, miR-144-3p and SOX2. The graphs showed a negative relationship between LINC00662 and miR-144-3p, a positive relationship between LINC00662 and SOX2 and a negative relationship between miR-144-3p and SOX2 in vivo (Fig. 5G−I). So far, ceRNA mode of LINC00662/miR-144-3p/SOX2 axis and its influence on BC progression is clear as shown in Graphical abstract. Thus, we may conclude LINC00662 promotes migration, invasion and stemness in BC by targeting miR-144-3p/SOX2 axis, in vitro and vivo.