m6A-modificated Linc00707 promotes the progression of TNBC via U2AF2-mediated alternative splicing of ATG4B Authors: Hongli Li, Yong Li, Qinghua Liu, Chonggao Yin, Yunxiang Zhang, Peng Gao Published: 2026 ABSTRACT BackgroundTriple-negative breast cancer (TNBC) represents a clinically aggressive breast cancer subtype with limited therapeutic options. Emerging evidence suggests that long intergenic noncoding RNA 00707 (Linc00707) plays a role in TNBC; however, the upstream regulators governing Linc00707 expression and the mechanisms by which it contributes to tumor progression remain largely undefined.MethodThe FOXP3-mediated transcriptional regulation of Linc00707 was analyzed using chromatin immunoprecipitation and luciferase reporter assays. The post-transcriptional regulation of Linc00707 was examined by RNA immunoprecipitation (RIP), methylated RIP, and RNA pull-down assays to assess WTAP-dependent m6A modification, IGF2BP2-mediated stabilization, and U2AF2 interaction. RNA sequencing and biochemical analyses were used to identify downstream regulatory pathways. The binding of oligo-anti-Linc00707 to Linc00707 and its ability to disrupt U2AF2 association were confirmed by RIP and RNA fluorescence in situ hybridization (RNA-FISH). Cellular assays and nude mouse xenograft models were employed to evaluate functional and therapeutic effects.ResultsFOXP3 was found to transcriptionally activate Linc00707 through direct promoter binding. The WTAP-mediated m6A modification enhanced the IGF2BP2-dependent stabilization of Linc00707. In the nucleus, Linc00707 interacted with the splicing factor U2AF2, with nucleotides 1–593 identified as a critical interaction region. Building on this observation, we further showed that the Linc00707–U2AF2 interaction functionally controls U2AF2 expression and stability via FOXP3. The Linc00707–U2AF2 complex was associated with altered alternative splicing of the autophagy-related gene ATG4B, contributing to enhanced TNBC cell proliferation, invasion, and autophagy suppression. Importantly, oligonucleotides complementary to the U2AF2-binding region of Linc00707 (oligo-anti-Linc00707) disrupted this interaction, attenuated Linc00707-driven oncogenic phenotypes, and restored autophagic activity. In vivo, oligo-anti-Linc00707 treatment significantly reduced tumor growth, supporting its therapeutic potential in TNBC.ConclusionsFOXP3 activates Linc00707 transcription in TNBC. WTAP-mediated m6A modification enhances Linc00707 stability via IGF2BP2. Linc00707 recruits U2AF2 to drive oncogenic ATG4B splicing, promoting tumor progression. Oligo-anti-Linc00707 specifically blocks Linc00707-U2AF2 complex formation, reversing its tumor-promoting and autophagy-suppressing functions. Thus, targeting Linc00707 represents a promising therapeutic strategy for TNBC.Supplementary InformationThe online version contains supplementary material available at 10.1186/s11658-026-00937-5. Introduction Triple-negative breast cancer (TNBC), defined by the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression, represents one of the most clinically challenging subtypes of breast cancer [1, 2]. Accounting for 15–20% of breast cancer cases, TNBC is distinguished by aggressive tumor biology, early metastasis, and a lack of targeted therapies, resulting in poor patient prognosis [3, 4]. Despite recent advances in chemotherapy and immunotherapy, the molecular mechanisms driving TNBC progression remain incompletely understood, and therapeutic options are limited [5–8]. Emerging evidence suggests that long noncoding RNAs (LncRNAs) serve as critical regulators of tumor growth and metastasis, functioning through diverse mechanisms such as chromatin modification, transcriptional control, and post-transcriptional regulation [9–12]. Linc00707, a long intergenic noncoding RNA, has been implicated in the promotion of TNBC cell proliferation and survival [13]. Linc00707 is upregulated in many tumors such as lung cancer [14], TNBC [15], colorectal cancer (CRC) [16, 17]. However, the upstream factors regulating Linc00707 expression remain poorly characterized. The N6-methyladenosine (m6A) modification is one of the most common RNA modifications. It modulates RNA degradation, RNA stabilization, mRNA translation, mRNA splicing, and other processes. We investigate the upstream regulatory mechanisms and functions of the oncogenic lncRNA Linc00707 in TNBC. We identify forkhead box protein P3 (FOXP3) as a key transcriptional activator of Linc00707 and uncover a Wilms tumor 1-associated protein (WTAP)/insulin-like growth factor 2 mRNA-binding protein 2 (IGF2BP2)-mediated m6A modification that enhances its RNA stability. LncRNAs regulate gene expression by combining with DNA, RNA, or proteins at the epigenetic, transcriptional, and posttranscriptional levels. For example, Linc00707 sponges miR-370-3p to promote osteogenesis of human bone marrow-derived mesenchymal stem cells through upregulating WNT2B [18]. Linc00707 could bind with miR-651-3p, while miR-651-3p bound to the 3′ untranslated region of SP2 mRNA to negatively regulate its expression to regulate the vasculogenic mimicry formation ability of glioma cells [19]. However, it is still unknown whether Linc00707 can promote the progression of TNBC by directly binding to proteins. Herein, we reveal that nuclear Linc00707 binds and stabilizes the splicing factor U2AF2, promoting TNBC progression. Moreover, we demonstrate the therapeutic potential of disrupting the Linc00707-U2AF2 interaction using LNA-modified antisense oligonucleotides. Together, our findings defined a previously unrecognized regulatory axis linking transcriptional activation, epitranscriptomic control, and lncRNA–protein interactions in TNBC. Methods and materials Patient samples and ethical approval We collected 25 pairs of TNBC tissues from the affiliated hospital of Shandong Second Medical University in 2016–2018. All samples were collected for research use only. The research was approved by the ethics committee of Shandong Second Medical University (approval no. 2017050). TCGA database analysis Breast cancer RNA-seq data were downloaded from The Cancer Genome Atlas (TCGA) database. Samples with available expression data and molecular subtype information were included for analysis. Differential expression analyses were performed between normal and tumor tissues, as well as among different breast cancer subtypes, including TNBC, HER2-positive, Luminal A, and Luminal B, using R software (version 4.2.2). Cell culture, transfection, and lentiviral transduction MDA-MB-231 and MDA-MB-468 cells were obtained from Procell Life Science & Technology Co., Ltd (Wuhan, China), and cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂. All plasmids, siRNAs, and lentiviral vectors encoding target genes or shRNAs were synthesized by GenCefe Biotech Co., Ltd. (Jiangsu, China). Transient transfections were carried out using Lipofectamine™ 3000 (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol, and cells were collected 48–72 h post transfection. Stable cell lines were generated by lentiviral transduction following the manufacturer’s instructions provided by GenCefe Biotech Co., Ltd. The siRNA sequences used in this study are listed in Supplementary Table S1. Quantitative real-time PCR (RT–qPCR) Total RNA was isolated using TRIzol reagent. Reverse transcription was performed using a commercial cDNA synthesis kit. RT–qPCR was carried out using SYBR Green Master Mix on a real-time PCR system. All reactions were performed in triplicate. Relative expression levels were normalized to GAPDH and calculated using the 2−ΔΔCt method. Western blotting Total protein was extracted using radioimmunoprecipitation assay (RIPA) lysis buffer containing protease and phosphatase inhibitors. Protein concentration was measured by BCA assay. Proteins were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), transferred to polyvinylidene fluoride (PVDF) membranes, blocked with 5% nonfat milk, and incubated with primary antibodies at 4 °C overnight. After incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies, signals were visualized using an electrochemiluminescence (ECL) detection system. All primary antibody information was shown in Supplementary Table S2. Chromatin immunoprecipitation (ChIP)-qPCR ChIP assays were performed using the ChIP Kit (BersinBio, Guangzhou, China) according to the manufacturer’s instructions. Briefly, breast cancer cells were crosslinked in 1% formaldehyde solution for 10 min, followed by the addition of 125 mM of glycine for 5 min at room temperature. Cell lysates were then sonicated to generate DNA fragments ranging from 200 to 500 bp. Antibodies, including anti-FOXP3 (ABclonal Technology, Wuhan, China) and normal IgG, were used for each immunoprecipitation. The extracted DNA was separated and purified and the enrichment fold was measured by qPCR. The Linc00707 promotor primer sequences are listed in Supplementary Table S3. Luciferase reporter assay Luciferase activity was measured using the Dual-Luciferase® Reporter Assay System (Promega, Madison, WI, USA) following the manufacturer’s instructions. Proliferation, colony formation, Transwell, and wound healing assay assays Cell proliferation was assessed using the EdU assay according to the manufacturer’s instructions. Colony formation assays were performed by seeding cells at low density and culturing for 10–14 days, followed by crystal violet staining. Cell migration and invasion were evaluated using Transwell chambers, with Matrigel coating applied for invasion assays. Wound healing assays were performed by scratching confluent cell monolayers with a sterile pipette tip and monitoring wound closure at the indicated time points. Methylated RNA immunoprecipitation (MeRIP) enrichment and RT–qPCR The Magna m6A MeRIP Kit (BersinBio, Guangzhou, China) was used to determine the quantification of m6A-modified Linc00707 levels according to the manufacturer’s instructions. Total RNA was isolated from MDA-MB-231 cells by Trizol extraction. Then, RNA was fragmented and immunoprecipitated with anti-m6A antibody in buffer containing RNase inhibitor. The complex was combined with magnetic beads for isolation and purification. The m6A enrichment was analyzed via RT–qPCR analysis. The m6A primer sequences were listed in Supplementary Table S4. RNA stability assay Cells were transiently transfected with siRNA using Lipofectamine™ 3000. After 48 h, cells were treated with actinomycin D and collected at multiple time points. RNA stability was analyzed by RT–qPCR. Immunohistochemistry (IHC) Ki67 expression in tumor tissues was examined by IHC using a universal two-step detection kit (PV-9000, ZSGB-BIO, Beijing, China) according to the manufacturer’s protocol. Sections were visualized with DAB and counterstained with hematoxylin. Protein stability analysis The stability of the target protein was assessed over a 24-h period. Protein samples were incubated under the specified conditions, and aliquots were collected at designated time points (0.5, 1, 4, 6, 12, and 24 h). Following collection, samples were immediately frozen and stored at −80 °C to prevent further degradation until batch processing. Protein stability was evaluated by western blotting analysis. RNA pulldown assay and mass spectrometry Linc00707 were in vitro-transcribed and biotin-labeled using a RNAmax-T7 Biotin Labeling Transcription Kit (RiboBio, Guangzhou, China). RNA pulldown was performed using a RNA pulldown kit (BersinBio, Guangzhou, China) according to the manufacturer’s instructions. The RNA–protein complexes were washed, degenerated, eluted, and used in mass spectrometry and western blotting. RNA immunoprecipitation (RIP) assay RIP assays were conducted using the PureBinding RNA Immunoprecipitation Kit (Geneseed, Guangzhou, China) according to the manufacturer’s instructions. Input RNA was processed in parallel. RIP-enriched RNA was analyzed by RT–qPCR to detect U2AF2-associated Linc00707 and pre-mRNAs. The primers used for RIP-qPCR analysis were listed in Supplementary Table S5. RNA sequencing (RNA-seq) RNA-seq of oe-con or oe-U2AF2 MDA-MB-231 cells were performed using the platform Illumina NovaSeq 6000. The data have been deposited at the Genome Sequence Archive for human (GSA-Human) website with series numbers of HRA011703. RNA fluorescence in situ hybridization (RNA-FISH) The colocalization of Linc00707 and oligo-anti-Linc00707 in TNBC cells were detected by RNA FISH kit (RiboBio, Guangzhou, China). FAM-labeled oligo-anti-Linc00707 sequence is as follows: AGGGGAAAGATGAAGGUCAG. NC sequence is as follows: GCGUATTATAGCCGAUUAAC. Representative images were collected and analyzed under a confocal laser scanning microscope (TCS SP8, Leica, Germany). All experiments were independently repeated three times. Tumor xenograft model Animal experiments were approved by the Animal Experimental Research Ethics Committee of Shandong Second Medical University and conducted according to its guidelines (number: 2023SDL311). BALB/C nude mice (female, 4 weeks) were obtained from PengYue Laboratory Animal Technology Co. Ltd (Jinan, China). MDA-MB-231 cells with lentivirus vector LV-oe-Linc00707 or LV-oe-NC were injected subcutaneously into the mammary fat pads of the mice. When the tumor grew to a diameter of about 0.5 cm, we injected oligo-anti-Linc00707 and negative control nucleotides (NC) into the tumor at multiple points every 3 days. Tumor volumes (V) were calculated using the formula V = (tumor length × width2)/2. At 8 weeks after injection, the mice were sacrificed, and the tumors were collected. Statistical analysis Differences between two groups were assessed by Student’s t-test, and one-way analysis of variance (ANOVA) was used to compare three or more groups. P-values lower than 0.05 were considered statistically significant. Correlations were analyzed by Spearman rank correlation. All data were statistically analyzed using GraphPad Prism 8.0. Results FOXP3 activates transcription of Linc00707 in TNBC cells Previous research has shown that Linc00707 is upregulated in TNBC [15], and many studies have demonstrated that transcription factors (TFs) can be involved in the activation of some LncRNAs [20, 21]. According to the promoter sequence, the hTFtarget, PROMO, and JASPAR prediction websites were used to predict the TFs that may bind to the promoter region of Linc00707. All the predicted TFs were intersected (Supplementary Fig. 1A). The binding sites between FOXP3, HOXD9, VDR, TBP, YY1, and SPY and the promoter sequence of Linc00707 (Supplementary Fig. S1B). We also downloaded the TCGA database to analyze the expression of six TFs in breast cancer and different subtypes of breast cancer. The results showed in Supplementary Fig. S1C–N. We selected FOXP3 and VDR, which have significantly increased expression in breast cancer tumors and TNBC and are positively correlated with Linc00707 expression (Fig. 1A–F), for verification.Fig. 1FOXP3 activates transcription of Linc00707 in TNBC cells. A The TCGA data analysis showed that the expression of FOXP3 in breast cancer tumors (Tumor) was significantly higher than that in normal breast tissue (Normal). B The TCGA data analysis showed that the expression of FOXP3 in TNBC was significantly higher than that in Normal, Her2+, LumA, and LumB. C The TCGA data analysis showed a positive correlation between FOXP3 and the expression of Linc00707 in breast cancer. D The TCGA data analysis showed that the expression of VDR was significantly higher in Tumor than in Normal. E The TCGA data analysis showed that the expression of VDR in TNBC was significantly higher than that in Normal, Her2+, and LumA, but there was no statistically significant difference from that in LumB. F The TCGA data analysis showed a positive correlation between VDR and the expression of Linc00707 in breast cancer. G The expression of Linc00707 in TNBC tissues was positively correlated with the expression of FOXP3 mRNA. H The expression of Linc00707 in TNBC tissues was positively correlated with the expression of VDR mRNA. I RT–qPCR results showed that the expression of Linc00707 in MDA-MB-231 and MDA-MB-468 cells was significantly reduced after knockdown of FOXP3. J RT–qPCR results showed that the expression of Linc00707 was significantly increased after overexpressing FOXP3 in MDA-MB-231 and MDA-MB-468 cells. K ChIP results showed that FOXP3 endogenously binds to the Linc00707 promoter region, where the second binding site is the most enriched; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. L ChIP results showed that the enrichment of FOXP3 in MDA-MB-231 cells with downregulated FOXP3 was reduced compared with that in control cells; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. M ChIP results showed increased enrichment of FOXP3 in the Linc00707 promoter region in cells overexpressing FOXP3; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. N Dual-luciferase results showed that the second binding site is the core promoter region of FOXP3 transcriptional activation Linc00707. The above is the deletion fragment of the upstream promoter region of Linc00707 and the complete sequence constructed on the pGL3 luciferase reporter vector. Below are the results of statistical analysis of luciferase activity. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation. ns P ≥ 0.05, ** P < 0.01, *** P < 0.001 FOXP3 activates transcription of Linc00707 in TNBC cells. A The TCGA data analysis showed that the expression of FOXP3 in breast cancer tumors (Tumor) was significantly higher than that in normal breast tissue (Normal). B The TCGA data analysis showed that the expression of FOXP3 in TNBC was significantly higher than that in Normal, Her2+, LumA, and LumB. C The TCGA data analysis showed a positive correlation between FOXP3 and the expression of Linc00707 in breast cancer. D The TCGA data analysis showed that the expression of VDR was significantly higher in Tumor than in Normal. E The TCGA data analysis showed that the expression of VDR in TNBC was significantly higher than that in Normal, Her2+, and LumA, but there was no statistically significant difference from that in LumB. F The TCGA data analysis showed a positive correlation between VDR and the expression of Linc00707 in breast cancer. G The expression of Linc00707 in TNBC tissues was positively correlated with the expression of FOXP3 mRNA. H The expression of Linc00707 in TNBC tissues was positively correlated with the expression of VDR mRNA. I RT–qPCR results showed that the expression of Linc00707 in MDA-MB-231 and MDA-MB-468 cells was significantly reduced after knockdown of FOXP3. J RT–qPCR results showed that the expression of Linc00707 was significantly increased after overexpressing FOXP3 in MDA-MB-231 and MDA-MB-468 cells. K ChIP results showed that FOXP3 endogenously binds to the Linc00707 promoter region, where the second binding site is the most enriched; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. L ChIP results showed that the enrichment of FOXP3 in MDA-MB-231 cells with downregulated FOXP3 was reduced compared with that in control cells; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. M ChIP results showed increased enrichment of FOXP3 in the Linc00707 promoter region in cells overexpressing FOXP3; 1, 2, and 3 respectively represent the three primers corresponding to the three binding sites between the promoter region of Linc00707 and FOXP3. N Dual-luciferase results showed that the second binding site is the core promoter region of FOXP3 transcriptional activation Linc00707. The above is the deletion fragment of the upstream promoter region of Linc00707 and the complete sequence constructed on the pGL3 luciferase reporter vector. Below are the results of statistical analysis of luciferase activity. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation. ns P ≥ 0.05, ** P < 0.01, *** P < 0.001 First, to verify the accuracy of the TCGA data, we detected the expression levels of FOXP3 and VDR mRNA in 25 TNBC tissues using RT–qPCR. The results showed that the expression of Linc00707 in TNBC tissues was positively correlated with the expression of FOXP3 and VDR (Fig. 1G, H). Next, to verify whether FOXP3 and VDR can transcriptionally activate the expression of Linc00707 in TNBC, we transfected FOXP3 and VDR knockdown or overexpression plasmids into MDA-MB-231 and MDA-MB-468 cells. RT–qPCR was used to detect the expression of FOXP3 and VDR mRNA and Linc00707, and western blotting was used to detect the protein expression of FOXP3 and VDR. The results showed that the expression of FOXP3 and Linc00707 was significantly reduced after the knockdown of FOXP3 in MDA-MB-231 and MDA-MB-468 cells (Fig. 1I, Supplementary Fig. S1O, P), and overexpression of FOXP3 could upregulate the expression of FOXP3 and Linc00707 (Fig. 1J, Supplementary Fig. S2P, Q). After knockdown or overexpression of VDR in both MDA-MB-231 and MDA-MB-468 cells, VDR was downregulated or upregulated, whereas there was no significant change in expression of Linc00707 (Supplementary Fig. S2R–V). Therefore, we conjecture that FOXP3 can activate the transcription of Linc00707. To verify this hypothesis, we designed three pairs of primers for ChIP experiments against FOXP3 according to the three binding sites predicted by JASPAR database. The results showed that the Linc00707 promoter region was more enriched in the anti-FOXP3 group, and the second binding site was highest enriched for Linc00707, about 12-fold in MDA-MB-231 cells and about 8-fold in MDA-MB-468 cells (Fig. 1K). Knockdown or overexpression of FOXP3 decreased or increased the enrichment of FOXP3 in the Linc00707 promoter region, respectively (Fig. 1L, M). To verify the above results, we also designed a luciferase reporter assay. The results showed that luciferase activity was significantly increased in cells comprising the full-length sequence of the promoter region of Linc00707 (Linc00707 WT) and significantly decreased in cells with deletion of three binding sites, namely Linc00707 Mut2 (Fig. 1N). Collectively, these findings confirm that FOXP3 directly activates Linc00707 transcription in TNBC by binding to its promoter, particularly through a core regulatory site located at the second predicted binding motif. WTAP-mediated m6A modification promotes Linc00707 expression via IGF2BP2-dependent stabilization of RNA N6-methyladenosine (m6A) modification is the most prevalent internal RNA modification and plays key roles in regulating RNA degradation, stability, and splicing [22–24]. To investigate whether m6A contributes to the post-transcriptional regulation of Linc00707, we used the sequence-based RNA adenosine methylation site predictor (SRAMP) tool to predict potential m6A modification sites. A total of 21 putative m6A sites were identified in Linc00707 (Supplementary Fig. S2A). Among them, one extremely high-confidence site and six high-confidence sites were selected for further analysis. The predicted methylation sites and their surrounding secondary structures are shown in Supplementary Fig. S2B. We designed six primer pairs for MeRIP assays, in which the nearby 531nt and 536nt sites were combined into one primer set (site 2). MeRIP-qPCR results revealed that site 2 exhibited the highest m6A enrichment, showing approximately a 5.22-fold increase compared with the IgG control (Fig. 2A). These findings suggest that Linc00707 undergoes m6A methylation at specific nucleotide positions.Fig. 2WTAP-mediated m6A modification promotes Linc00707 expression through IGF2BP2-dependent RNA stabilization. A MeRIP-qPCR analysis in MDA-MB-231 cells revealed m6A enrichment of Linc00707 at multiple predicted sites, with the highest enrichment observed at site2. B RT–qPCR analysis showed that overexpression of WTAP significantly increased the expression levels of WTAP and Linc00707 in MDA-MB-231 cells. C RT–qPCR analysis showed that overexpression of WTAP significantly increased the expression levels of WTAP and Linc00707 in MDA-MB-468 cells. D RT–qPCR analysis showed that knockdown of WTAP significantly reduced the expression levels of WTAP and Linc00707 in MDA-MB-231 cells. E RT–qPCR analysis showed that overexpression of WTAP significantly reduced the expression levels of WTAP and Linc00707 in MDA-MB-468 cells. F Analysis of the TCGA database demonstrated that WTAP expression was significantly higher in triple-negative breast cancer (TNBC) tissues compared with Normal, HER2-positive, Luminal A, and Luminal B subtypes. G MeRIP-qPCR results showed that WTAP overexpression enhanced the m6A enrichment of Linc00707 at site2 compared with the control group. H RT–qPCR analysis showed that knockdown of YTHDC2 reduced Linc00707 expression in MDA-MB-231 cells. I RT–qPCR analysis showed that knockdown of IGF2BP2 reduced Linc00707 expression in MDA-MB-231 cells. J RT–qPCR analysis showed that knockdown of HUR reduced Linc00707 expression in MDA-MB-231 cells. K RT–qPCR analysis showed that knockdown of YTHDC2 reduced Linc00707 expression in MDA-MB-468 cells. L RT–qPCR analysis showed that knockdown of IGF2BP2 reduced Linc00707 expression in MDA-MB-468 cells. M RT–qPCR analysis showed that knockdown of HUR reduced Linc00707 expression in MDA-MB-468 cells. N RT–qPCR analysis showed that IGF2BP2 knockdown reversed WTAP-induced upregulation of Linc00707 in MDA-MB-231 cells. O RT–qPCR analysis showed that IGF2BP2 knockdown reversed WTAP-induced upregulation of Linc00707 in MDA-MB-468 cells. P RNA stability assays demonstrated that knockdown of IGF2BP2 significantly shortened the half-life of Linc00707 in MDA-MB-231 cells following actinomycin D (ActD) treatment. Q RNA stability assays demonstrated that knockdown of IGF2BP2 significantly shortened the half-life of Linc00707 in MDA-MB-468 cells following ActD treatment. R RIP-qPCR analysis showed significant enrichment of Linc00707 in anti-IGF2BP2 immunoprecipitates compared with IgG controls in MDA-MB-231 cells, with the strongest binding observed at site2. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA as appropriate. ns P ≥ 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 WTAP-mediated m6A modification promotes Linc00707 expression through IGF2BP2-dependent RNA stabilization. A MeRIP-qPCR analysis in MDA-MB-231 cells revealed m6A enrichment of Linc00707 at multiple predicted sites, with the highest enrichment observed at site2. B RT–qPCR analysis showed that overexpression of WTAP significantly increased the expression levels of WTAP and Linc00707 in MDA-MB-231 cells. C RT–qPCR analysis showed that overexpression of WTAP significantly increased the expression levels of WTAP and Linc00707 in MDA-MB-468 cells. D RT–qPCR analysis showed that knockdown of WTAP significantly reduced the expression levels of WTAP and Linc00707 in MDA-MB-231 cells. E RT–qPCR analysis showed that overexpression of WTAP significantly reduced the expression levels of WTAP and Linc00707 in MDA-MB-468 cells. F Analysis of the TCGA database demonstrated that WTAP expression was significantly higher in triple-negative breast cancer (TNBC) tissues compared with Normal, HER2-positive, Luminal A, and Luminal B subtypes. G MeRIP-qPCR results showed that WTAP overexpression enhanced the m6A enrichment of Linc00707 at site2 compared with the control group. H RT–qPCR analysis showed that knockdown of YTHDC2 reduced Linc00707 expression in MDA-MB-231 cells. I RT–qPCR analysis showed that knockdown of IGF2BP2 reduced Linc00707 expression in MDA-MB-231 cells. J RT–qPCR analysis showed that knockdown of HUR reduced Linc00707 expression in MDA-MB-231 cells. K RT–qPCR analysis showed that knockdown of YTHDC2 reduced Linc00707 expression in MDA-MB-468 cells. L RT–qPCR analysis showed that knockdown of IGF2BP2 reduced Linc00707 expression in MDA-MB-468 cells. M RT–qPCR analysis showed that knockdown of HUR reduced Linc00707 expression in MDA-MB-468 cells. N RT–qPCR analysis showed that IGF2BP2 knockdown reversed WTAP-induced upregulation of Linc00707 in MDA-MB-231 cells. O RT–qPCR analysis showed that IGF2BP2 knockdown reversed WTAP-induced upregulation of Linc00707 in MDA-MB-468 cells. P RNA stability assays demonstrated that knockdown of IGF2BP2 significantly shortened the half-life of Linc00707 in MDA-MB-231 cells following actinomycin D (ActD) treatment. Q RNA stability assays demonstrated that knockdown of IGF2BP2 significantly shortened the half-life of Linc00707 in MDA-MB-468 cells following ActD treatment. R RIP-qPCR analysis showed significant enrichment of Linc00707 in anti-IGF2BP2 immunoprecipitates compared with IgG controls in MDA-MB-231 cells, with the strongest binding observed at site2. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA as appropriate. ns P ≥ 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 m6A modification is dynamically regulated by methyltransferases (METTL3, METTL14, WTAP) [25–27], demethylases (FTO, ALKBH5) [28, 29], and m6A-binding proteins (readers) such as the YTHDF family and IGF2BPs [30, 31]. To determine which m6A regulators are involved in the regulation of Linc00707, we first overexpressed METTL3, METTL14, WTAP, FTO, and ALKBH5 in MDA-MB-231 and MDA-MB-468 cells. RT–qPCR analysis showed that WTAP overexpression significantly increased Linc00707 expression in both cell lines, whereas overexpression of the other m6A regulators did not produce a significant effect (Fig. 2B, C). Conversely, WTAP knockdown markedly reduced Linc00707 expression (Fig. 2D, E). However, overexpression and silencing of METTL14, METTL3, ALKBH5, or FTO did not show a consistent or significant impact (Supplementary Fig. S2C–J). These gain- and loss-of-function experiments collectively indicate that WTAP is the primary m6A regulator affecting Linc00707 expression in breast cancer cells. In support of this observation, analysis of TCGA dataset revealed that WTAP expression was significantly elevated in TNBC compared with normal breast tissues and other breast cancer subtypes (Fig. 2F). In addition, the expression levels of METTL14, METTL3, ALKBH5, and FTO did not exhibit significant subtype-specific differences among breast cancer subtypes, including TNBC (Supplementary Fig. S2K–N), suggesting that the dysregulation of Linc00707 in TNBC is unlikely to be driven by global alterations in m6A writer or eraser expression. To further confirm the role of WTAP in mediating the m6A modification of Linc00707, MeRIP assays were conducted in MDA-MB-231 cells following WTAP overexpression. The results revealed a marked increase in m6A enrichment across multiple sites of Linc00707, with site 2 showing the most pronounced effect-an approximately 15.69-fold enrichment compared with the control (Fig. 2G). These findings strongly suggest that WTAP acts as a key methyltransferase responsible for the m6A modification of Linc00707. We next aimed to identify the m6A readers responsible for stabilizing Linc00707 upon WTAP-mediated methylation. Small interfering RNAs (siRNAs) targeting six candidate m6A readers (YTHDC2, hnRNPA2B1, IGF2BP1, IGF2BP2, IGF2BP3, and HUR) were transfected into TNBC cells. RT–qPCR analysis revealed that knockdown of YTHDC2, IGF2BP2, and HUR significantly reduced Linc00707 expression (Fig. 2H–M), whereas silencing of hnRNPA2B1, IGF2BP1, or IGF2BP3 had no notable effect (Supplementary Fig. S2O-T). Notably, IGF2BP2 knockdown completely abolished the Linc00707 upregulation induced by WTAP overexpression (Fig. 2N, O), suggesting a direct regulatory role of IGF2BP2 in this pathway. To determine whether IGF2BP2 stabilizes Linc00707 transcripts, an actinomycin D (ActD) assay was performed to measure RNA half-life. The results showed that knockdown of IGF2BP2 significantly reduced the stability of Linc00707 RNA in TNBC cells (Fig. 2P, Q). Consistently, RIP assays revealed that IGF2BP2 was significantly enriched on Linc00707 transcripts at specific binding regions, further supporting a direct interaction between IGF2BP2 and Linc00707 in TNBC cells (Fig. 2R). Taken together, these data indicate that WTAP enhances Linc00707 expression by promoting m6A methylation at key sites, which in turn stabilizes Linc00707 transcripts through IGF2BP2-dependent binding. Linc00707 interacts with U2AF2 in the nucleus and enhances its protein stability in TNBC cells Linc00707 is localized in both the nucleus and cytoplasm of TNBC cells, suggesting potential interactions with nuclear proteins [15]. RNA pull-down assays using biotin-labeled full-length Linc00707 identified specific protein bands at 40–55 kDa that were absent in the LacZ control (Fig. 3A). Mass spectrometry identified 74 candidate Linc00707-interacting proteins (Supplementary Fig. S3A). On the basis of molecular weight, peptide score, reported relevance to tumor progression, autophagy association (Supplementary Fig. S3B), survival analysis (Supplementary Fig. S3C–H), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway (Supplementary Fig. S3I, J), U2AF2 and HNRNPK were selected for further validation. Western blotting analysis confirmed specific enrichment of U2AF2 in Linc00707 pull-down complexes, whereas HNRNPK showed no enrichment (Fig. 3B). Analysis of TCGA dataset showed that U2AF2 expression was elevated in TNBC tissues and positively correlated with Linc00707 expression (Fig. 3C, D). This correlation was confirmed in 25 samples from patients with TNBC (Fig. 3E). RIP further demonstrated enrichment of Linc00707 in U2AF2 immunoprecipitates, confirming their interaction in TNBC cells (Fig. 3F). RNA pull-down assays using truncated Linc00707 fragments showed that U2AF2 bound specifically to the 5′ region of Linc00707 (1–593 bp) but not to the 594–3099-bp fragment (Fig. 3G, H). RT–qPCR and ActD chase assays showed that altering Linc00707 and FOXP3 expression did not affect U2AF2 mRNA levels or stability (Supplementary Fig. S3K, L).Fig. 3Linc00707 interacts with U2AF2 in the nucleus and enhances its protein stability in TNBC cells. A Silver staining analysis of proteins pulled down by biotin-labeled Linc00707 sense or LacZ in MDA-MB-231 cells, with the differential band indicated by a red box. B Western blotting analysis confirming the presence of U2AF2 and HNRNPK in the Linc00707 sense RNA pull-down complex. C Analysis of TCGA database showing significantly higher U2AF2 expression in TNBC tissues compared with Normal. D Pearson correlation analysis demonstrating a positive correlation between U2AF2 and Linc00707 expression levels in TNBC samples. E Pearson correlation analysis demonstrating a positive correlation between U2AF2 and Linc00707 expression levels in TNBC samples. F RIP-qPCR analysis showing significant enrichment of Linc00707 in anti-U2AF2 immunoprecipitates compared with IgG controls. G Schematic representation of full-length and truncated Linc00707 fragments used for RNA pull-down assays, followed by western blotting analysis showing the binding of U2AF2 to different Linc00707 fragments. H Western blotting analysis identifying the specific Linc00707 fragment responsible for binding to U2AF2. I Western blotting analysis showing U2AF2 and FOXP3 protein levels in MDA-MB-231 cells following different experimental treatments. J Western blotting analysis showing U2AF2 and FOXP3 protein levels in MDA-MB-468 cells following following different experimental treatments. K CHX chase assays showing that Linc00707 overexpression prolonged the protein half-life of U2AF2 in MDA-MB-231 cells. L CHX chase assays showing that Linc00707 overexpression prolonged the protein half-life of U2AF2 in MDA-MB-468 cells. M CHX chase assays showing that FOXP3 overexpression rescued the reduced stability of U2AF2 caused by Linc00707 knockdown in MDA-MB-231 cells, with corresponding quantitative analysis. N CHX chase assays showing that FOXP3 overexpression rescued the reduced stability of U2AF2 caused by Linc00707 knockdown in MDA-MB-468 cells, with corresponding quantitative analysis. O Western blotting analysis of U2AF2 protein levels in MDA-MB-231 and MDA-MB-468 cells under different treatment conditions, including sh-Linc00707 or sh-control, in the presence or absence of 3-MA or MG132, with GAPDH used as a loading control. Quantification of relative U2AF2 expression is shown on the right. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA as appropriate.ns P ≥ 0.05, * P < 0.05, *** P < 0.001, **** P < 0.0001 Linc00707 interacts with U2AF2 in the nucleus and enhances its protein stability in TNBC cells. A Silver staining analysis of proteins pulled down by biotin-labeled Linc00707 sense or LacZ in MDA-MB-231 cells, with the differential band indicated by a red box. B Western blotting analysis confirming the presence of U2AF2 and HNRNPK in the Linc00707 sense RNA pull-down complex. C Analysis of TCGA database showing significantly higher U2AF2 expression in TNBC tissues compared with Normal. D Pearson correlation analysis demonstrating a positive correlation between U2AF2 and Linc00707 expression levels in TNBC samples. E Pearson correlation analysis demonstrating a positive correlation between U2AF2 and Linc00707 expression levels in TNBC samples. F RIP-qPCR analysis showing significant enrichment of Linc00707 in anti-U2AF2 immunoprecipitates compared with IgG controls. G Schematic representation of full-length and truncated Linc00707 fragments used for RNA pull-down assays, followed by western blotting analysis showing the binding of U2AF2 to different Linc00707 fragments. H Western blotting analysis identifying the specific Linc00707 fragment responsible for binding to U2AF2. I Western blotting analysis showing U2AF2 and FOXP3 protein levels in MDA-MB-231 cells following different experimental treatments. J Western blotting analysis showing U2AF2 and FOXP3 protein levels in MDA-MB-468 cells following following different experimental treatments. K CHX chase assays showing that Linc00707 overexpression prolonged the protein half-life of U2AF2 in MDA-MB-231 cells. L CHX chase assays showing that Linc00707 overexpression prolonged the protein half-life of U2AF2 in MDA-MB-468 cells. M CHX chase assays showing that FOXP3 overexpression rescued the reduced stability of U2AF2 caused by Linc00707 knockdown in MDA-MB-231 cells, with corresponding quantitative analysis. N CHX chase assays showing that FOXP3 overexpression rescued the reduced stability of U2AF2 caused by Linc00707 knockdown in MDA-MB-468 cells, with corresponding quantitative analysis. O Western blotting analysis of U2AF2 protein levels in MDA-MB-231 and MDA-MB-468 cells under different treatment conditions, including sh-Linc00707 or sh-control, in the presence or absence of 3-MA or MG132, with GAPDH used as a loading control. Quantification of relative U2AF2 expression is shown on the right. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA as appropriate.ns P ≥ 0.05, * P < 0.05, *** P < 0.001, **** P < 0.0001 Western blotting analysis showed that overexpression of either Linc00707 or FOXP3 increased U2AF2 expression. Notably, FOXP3 overexpression was sufficient to rescue the reduction of U2AF2 caused by Linc00707 knockdown, highlighting a dominant role for FOXP3 in regulating U2AF2 expression (Fig. 3I,J, Supplementary Fig. S3M). Notably, western blotting analysis revealed no significant changes in FOXP3 expression upon overexpression or knockdown of Linc00707, consistent with FOXP3 acting upstream of Linc00707 in this regulatory axis (Fig. 3I,J, Supplementary Fig. S3N). Cycloheximide (CHX) chase assays further showed that Linc00707 overexpression significantly prolonged the half-life of U2AF2 protein in both cell lines (Fig. 3K, L, Supplementary Fig. S3O, P). FOXP3 overexpression also increased U2AF2 protein stability, whereas this effect was largely abolished by Linc00707 knockdown in both TNBC cell lines (Fig. 3M, N). To determine whether the reduction of U2AF2 induced by Linc00707 knockdown occurs through proteasomal or autophagy–lysosomal degradation, pharmacological inhibition assays were performed. The proteasome inhibitor MG132 markedly rescued the reduction of U2AF2 caused by Linc00707 knockdown, whereas the autophagy inhibitor 3-methyladenine (3-MA) had no such effect, indicating that U2AF2 degradation is mediated through the proteasomal pathway rather than the autophagy–lysosomal pathway (Fig. 3O). Changing U2AF2 expression levels altered U2AF2 mRNA abundance while having no effect on Linc00707 expression in either MDA-MB-231 or MDA-MB-468 cells (Supplementary Fig. S3Q-T). Together, these results indicate that Linc00707 stabilizes U2AF2 protein predominantly by preventing proteasomal pathway. Linc00707 promotes proliferation, invasion and migration of TNBC cells via U2AF2 protein To investigate whether the oncogenic functions of Linc00707 are mediated through its interaction with U2AF2, a series of rescue experiments were performed in TNBC cell lines MDA-MB-231 and MDA-MB-468. EdU assays and colony formation assays revealed that U2AF2 overexpression significantly promoted cell proliferation and partially rescued the proliferation defects caused by Linc00707 knockdown (Fig. 4A-F). Similarly, Transwell migration and invasion assays demonstrated that U2AF2 overexpression enhanced the migratory and invasive capabilities of TNBC cells, and could partially reverse the suppressive effects induced by Linc00707 silencing (Fig. 4G-I). Wound healing assays further confirmed that U2AF2 restoration mitigated the impaired cell mobility caused by Linc00707 depletion (Fig. 4K, L), as reflected by the quantification of cell motility ratios (Fig. 4J).Fig. 4Linc00707 promotes proliferation, invasion and migration of TNBC cells via U2AF2 protein. A, B Typical images of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 by EdU assay. Scale bar: 100 μm. C EdU statistical analysis of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells. D Colony formation statistical analysis of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells. E, F Typical images of the proliferation abilities of MDA-MB-231 and MDA-MB-468 cells by colony formation. G Typical images of Transwell assay in MDA-MB-231 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. H Typical images of Transwell assay in MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. I Transwell statistical analysis of the cell migration and invasion ability in MDA-MB-231 and MDA-MB-468 cells. J Wound healing assay analysis of the cell migration ability of MDA-MB-231 and MDA-MB-468 cells. K, L Typical images of Wound healing assay in MDA-MB-231 and MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test and one-way ANOVA was used to compare three or more groups. ns P ≥ 0.05, ** P < 0.01, *** P < 0.001 Linc00707 promotes proliferation, invasion and migration of TNBC cells via U2AF2 protein. A, B Typical images of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 by EdU assay. Scale bar: 100 μm. C EdU statistical analysis of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells. D Colony formation statistical analysis of the cell proliferation ability in MDA-MB-231 and MDA-MB-468 cells. E, F Typical images of the proliferation abilities of MDA-MB-231 and MDA-MB-468 cells by colony formation. G Typical images of Transwell assay in MDA-MB-231 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. H Typical images of Transwell assay in MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. I Transwell statistical analysis of the cell migration and invasion ability in MDA-MB-231 and MDA-MB-468 cells. J Wound healing assay analysis of the cell migration ability of MDA-MB-231 and MDA-MB-468 cells. K, L Typical images of Wound healing assay in MDA-MB-231 and MDA-MB-468 cells transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2. Scale bar: 100 μm. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test and one-way ANOVA was used to compare three or more groups. ns P ≥ 0.05, ** P < 0.01, *** P < 0.001 Taken together, these results indicate that U2AF2 acts as a functional mediator of Linc00707-driven tumor progression, supporting the conclusion that Linc00707 promotes TNBC cell proliferation, invasion, and migration, at least in part through its interaction with U2AF2. Linc00707 inhibits autophagy in TNBC cells through the U2AF2 protein Many studies have showed that autophagy played a significant role and LncRNA can affect autophagy in human cancers including TNBC [15, 32, 33]. To determine whether Linc00707 regulates autophagy through U2AF2, rescue experiments were performed in TNBC cells. Immunofluorescence analysis of LC3 puncta and autophagic flux assays showed that the overexpression of U2AF2 significantly reduced autophagosome formation and autophagic flux in both MDA-MB-231 and MDA-MB-468 cells compared with controls, indicating that U2AF2 inhibits autophagy. Moreover, U2AF2 overexpression reversed the increase in autophagic flux induced by Linc00707 knockdown (Fig. 5A, Supplementary Fig. S4A). Mitochondrial membrane potential (MMP) was further assessed using JC-1 staining. The results showed that U2AF2 overexpression increased MMP and counteracted the reduction in MMP caused by Linc00707 knockdown in both TNBC cell lines (Fig. 5B, Supplementary Fig. S4B). LC3-II/I ratio and SQSTM1/p62 levels are commonly used to evaluate autophagic activity. Western blotting analysis showed that U2AF2 overexpression decreased the LC3-II/I ratio and increased SQSTM1/p62 levels, whereas U2AF2 overexpression reversed the changes in LC3-II/I ratio and SQSTM1/p62 induced by Linc00707 knockdown in both MDA-MB-231 and MDA-MB-468 cells (Supplementary Fig. S4C). The results indicate that Linc00707 inhibits autophagy in TNBC cells by regulating U2AF2.Fig. 5Linc00707 inhibits autophagy in TNBC cells through the U2AF2 protein. A The MDA-MB-231 and MDA-MB-468 cesll lines stably expressing stubRFP-sensGFP-LC3 were transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 and StubRFP-SensGFP-LC3 fluorescent spots were observed by the fluorescence microscope. GFP represented autophagosomes, RFP represented the total number of autophagosomes and autolysosomes, and yellow puncta in the merged image represented the flow rate of autophagic flow from autophagosomes to autolysosomes. Scale bar: 50 μm. B Changes in mitochondrial membrane potential of MDA-MB-231 and MDA-MB-468 were transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 were monitored by JC-1. Green fluorescence: JC-1 monomer; red fluorescence: JC-1 aggregates. Scale bar: 50 μm. All experiments were repeated independently three times. Data are presented as means ± standard deviation Linc00707 inhibits autophagy in TNBC cells through the U2AF2 protein. A The MDA-MB-231 and MDA-MB-468 cesll lines stably expressing stubRFP-sensGFP-LC3 were transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 and StubRFP-SensGFP-LC3 fluorescent spots were observed by the fluorescence microscope. GFP represented autophagosomes, RFP represented the total number of autophagosomes and autolysosomes, and yellow puncta in the merged image represented the flow rate of autophagic flow from autophagosomes to autolysosomes. Scale bar: 50 μm. B Changes in mitochondrial membrane potential of MDA-MB-231 and MDA-MB-468 were transfected with sh-con + oe-NC, sh-con + oe-U2AF2, sh-Linc00707 #1 + oe-U2AF2, sh-Linc00707 #2 + oe-U2AF2 were monitored by JC-1. Green fluorescence: JC-1 monomer; red fluorescence: JC-1 aggregates. Scale bar: 50 μm. All experiments were repeated independently three times. Data are presented as means ± standard deviation Linc00707/U2AF2 complex regulates variable splicing of autophagy related protein ATG4B Previous studies have established that U2AF2 plays a critical role in regulating alternative splicing [34–36]. To investigate the role of the Linc00707/U2AF2 complex in regulating alternative splicing of autophagy-related genes in TNBC RNA-seq analysis was performed in MDA-MB-231 cells overexpressing U2AF2. A total of 5306 differential alternative splicing events (DASEs) involving 3211 genes were identified, with exon skipping (ES) being the most predominant event (Fig. 6A). Similarly, analysis of TCGA TNBC datasets comparing Linc00707-high and Linc00707-low expression groups revealed 1698 DASEs in 1244 genes (Fig. 6B). Integration of both datasets identified 197 overlapping DASEs corresponding to 195 genes, primarily involving ES, alternative 5′ splice sites (A5SS), and 3′ splice sites (A3SS) (Fig. 6C, D). Protein–protein interaction (PPI) network analysis of these 195 genes (Fig. 6E), followed by Gene Ontology (GO) and KEGG enrichment analysis of high-degree nodes, showed enrichment in mitochondrial membrane components, RNA binding, and autophagy-related biological processes (Fig. 6F), with KEGG pathways including autophagy (animal and other), mitophagy, and FoxO signaling (Fig. 6G). Among the enriched autophagy-related genes, BECN1, ATG12, and ATG4B were identified as shared targets (Fig. 6H). RIP assays demonstrated that U2AF2 binds specifically to the pre-mRNA of ATG4B but not ATG12 or BECN1 in both MDA-MB-231 and MDA-MB-468 cells (Fig. 6I–N). U2AF2 knockdown significantly reduced the expression of the alternatively spliced isoform ATG4Bs (Fig. 6O, P). Overexpression of Linc00707 or FOXP3 significantly increased ATG4Bs expression, whereas this increase was abolished by U2AF2 knockdown or Linc00707 knockdown, respectively, in both MDA-MB-231 and MDA-MB-468 cells (Fig. 6Q, R). These results indicate that the Linc00707-U2AF2 complex regulates autophagy in TNBC cells by modulating ATG4B alternative splicing. To further examine the functional impact of ATG4B alternative splicing on autophagy, ATG4B or ATG4Bs was overexpressed in TNBC cells. The overexpression of full-length ATG4B increased the LC3-II/I ratio and reduced SQSTM1/p62 levels, indicating enhanced autophagy, whereas overexpression of ATG4Bs showed the opposite effect in both MDA-MB-231 and MDA-MB-468 cells (Supplementary Fig. S5A, B).Fig. 6The Linc00707/U2AF2 complex regulates alternative splicing of autophagy-related gene ATG4B in TNBC cells. A, B Summary of differential alternative splicing events (DASEs) and genes in (A) U2AF2-overexpressing MDA-MB-231 cells (RNA-seq) and (B) TCGA TNBC dataset comparing Linc00707-high versus -low expression groups. ES, exon skipping; RI, retained intron; ME, mutually exclusive exons; A5SS, alternative 5′ splice site; A3SS, alternative 3′ splice site. C Venn diagram showing 197 overlapping DASEs shared between the two datasets. D Breakdown of DASE types among overlapping events, with ES being predominant. E Protein–protein interaction (PPI) network of 195 overlapping genes. Key autophagy-related genes (BECN1, ATG12, ATG4B) and high-degree nodes (e.g., STAT3) are highlighted. F GO enrichment analysis of high-degree PPI nodes, indicating enrichment in RNA splicing regulation, autophagy, and mitochondrial function. G KEGG pathway enrichment showing significant involvement in autophagy (animal and other), mitophagy, and FoxO signaling pathways. H Venn diagram categorizing overlapping autophagy-related genes into “animal” and “other” subgroups. I-N RIP-qPCR assays showing that U2AF2 specifically binds to the pre-mRNA of ATG4B but not ATG12 or BECN1 in MDA-MB-231 and MDA-MB-468 cells. O-P qRT–PCR showing that U2AF2 knockdown significantly reduces expression of the ATG4Bs splice isoform. Q RT–qPCR analysis showing that Linc00707 overexpression increased ATG4B expression, which was attenuated by U2AF2 knockdown or FOXP3 overexpression in MDA-MB-231 cells. R RT–qPCR analysis showing that Linc00707 overexpression increased ATG4B expression, which was attenuated by U2AF2 knockdown or FOXP3 overexpression in MDA-MB-468 cells. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation. ns P ≥ 0.05, ***P < 0.001, ****P < 0.0001 The Linc00707/U2AF2 complex regulates alternative splicing of autophagy-related gene ATG4B in TNBC cells. A, B Summary of differential alternative splicing events (DASEs) and genes in (A) U2AF2-overexpressing MDA-MB-231 cells (RNA-seq) and (B) TCGA TNBC dataset comparing Linc00707-high versus -low expression groups. ES, exon skipping; RI, retained intron; ME, mutually exclusive exons; A5SS, alternative 5′ splice site; A3SS, alternative 3′ splice site. C Venn diagram showing 197 overlapping DASEs shared between the two datasets. D Breakdown of DASE types among overlapping events, with ES being predominant. E Protein–protein interaction (PPI) network of 195 overlapping genes. Key autophagy-related genes (BECN1, ATG12, ATG4B) and high-degree nodes (e.g., STAT3) are highlighted. F GO enrichment analysis of high-degree PPI nodes, indicating enrichment in RNA splicing regulation, autophagy, and mitochondrial function. G KEGG pathway enrichment showing significant involvement in autophagy (animal and other), mitophagy, and FoxO signaling pathways. H Venn diagram categorizing overlapping autophagy-related genes into “animal” and “other” subgroups. I-N RIP-qPCR assays showing that U2AF2 specifically binds to the pre-mRNA of ATG4B but not ATG12 or BECN1 in MDA-MB-231 and MDA-MB-468 cells. O-P qRT–PCR showing that U2AF2 knockdown significantly reduces expression of the ATG4Bs splice isoform. Q RT–qPCR analysis showing that Linc00707 overexpression increased ATG4B expression, which was attenuated by U2AF2 knockdown or FOXP3 overexpression in MDA-MB-231 cells. R RT–qPCR analysis showing that Linc00707 overexpression increased ATG4B expression, which was attenuated by U2AF2 knockdown or FOXP3 overexpression in MDA-MB-468 cells. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation. ns P ≥ 0.05, ***P < 0.001, ****P < 0.0001 Collectively, these findings identify ATG4B alternative splicing as a key downstream effector of the FOXP3–Linc00707–U2AF2 axis in regulating autophagy in TNBC cells. Correlation of Linc00707, FOXP3, and U2AF2 expression with clinical parameters in patients with TNBC To explore the clinical relevance of Linc00707 in TNBC, the association between Linc00707 expression and clinicopathological characteristics was analyzed using TCGA data. A total of 93 patients with TNBC were stratified into low (n = 46) and high (n = 47) Linc00707 expression groups. As shown in Supplementary Table S6, no significant differences were observed between the two groups with respect to age, age distribution (≤ 60 vs. > 60), tumor stage, T stage, N stage, or M stage (all P > 0.05). Next, the clinical significance of the FOXP3–Linc00707–U2AF2 pathway was evaluated. Patients were divided into low activated (n = 27) and high activated (n = 66) pathway expression groups. As presented in Supplementary Table S7, a significant difference in tumor stage distribution was observed between the two groups (χ2 = 8.827, P = 0.032). However, no significant associations were detected between pathway expression and age, T stage, N stage, or M stage (all P > 0.05). These results indicate that the FOXP3–Linc00707–U2AF2 pathway is clinically relevant and may be associated with tumor progression in TNBC. Linc00707 is a potential therapeutic target for TNBC Having identified the 5′ region (1–593 bp) of Linc00707 as critical for its interaction with U2AF2, we next sought to define the precise binding interface. Molecular docking analysis revealed six putative interaction sites between Linc00707 and U2AF2 (Fig. 7A). On the basis of these predicted binding motifs, a short antisense oligonucleotide targeting the Linc00707–U2AF2 interaction region (oligo-anti-Linc00707), modified with locked nucleic acid (LNA) and 6-carboxyfluorescein (FAM), was designed to disrupt this interaction.Fig. 7Linc00707 is a potential therapeutic target for TNBC. A Molecular docking showed that the binding sites of Linc00707 and U2AF2. B RNA-FISH indicated the colocalization of Linc00707 and oligo-anti-Linc00707. C, D RIP assay showed that the interaction between Linc00707 and U2AF2 could be suppressed by oligo-anti-Linc00707. E The growth curve of mice subcutaneously injected with LV-oe-Linc00707 were treated with the oligo-anti-Linc00707 was shown. F The xenograft tumors were harvested and observed by an imaging system. G The weights of the xenograft tumors were measured. H Immunohistochemistry for Ki-67 detection in LV-oe-Linc00707 + NC and LV-oe-Linc00707 + oligo-anti-Linc00707 group. I Schematic diagram of Linc00707-mediated progression in TNBC. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation.** P < 0.01, *** P < 0.001 Linc00707 is a potential therapeutic target for TNBC. A Molecular docking showed that the binding sites of Linc00707 and U2AF2. B RNA-FISH indicated the colocalization of Linc00707 and oligo-anti-Linc00707. C, D RIP assay showed that the interaction between Linc00707 and U2AF2 could be suppressed by oligo-anti-Linc00707. E The growth curve of mice subcutaneously injected with LV-oe-Linc00707 were treated with the oligo-anti-Linc00707 was shown. F The xenograft tumors were harvested and observed by an imaging system. G The weights of the xenograft tumors were measured. H Immunohistochemistry for Ki-67 detection in LV-oe-Linc00707 + NC and LV-oe-Linc00707 + oligo-anti-Linc00707 group. I Schematic diagram of Linc00707-mediated progression in TNBC. All experiments were repeated independently three times. Data are presented as means ± standard deviation. Differences between two groups were assessed by Student’s t-test, one-way ANOVA was used to compare three or more groups, and correlations were analyzed by Spearman rank correlation.** P < 0.01, *** P < 0.001 RNA fluorescence in situ hybridization demonstrated effective colocalization of oligo-anti-Linc00707 with endogenous Linc00707 in TNBC cells (Fig. 7B). RIP assays further showed that oligo-anti-Linc00707 markedly reduced the binding of Linc00707 to U2AF2 in both MDA-MB-231 and MDA-MB-468 cells (Fig. 7C, D), indicating the effective disruption of the Linc00707–U2AF2 complex. In addition, overexpression of Linc00707 could increase the proliferation, invasion and migration abilities of MDA-MB-231 and MDA-MB-468 cells. However, after co-transfection with oligo-anti-Linc00707 (oe-Linc00707 + oligo-anti-Linc00707), compared with the negative control of transfecting oligo-anti-Linc00707 (oe-Linc00707 + NC), it can reverse the promoting effect of Linc00707 on the proliferation, invasion and migration of MDA-MB-231 and MDA-MB-468 cells (Supplementary Fig. S6A–L). To further confirm that disruption of the Linc00707–U2AF2 interaction restores autophagic activity, autophagy-related markers were examined following oligo-anti-Linc00707 treatment in Linc00707-overexpressing TNBC cells. Quantitative analyses of the mRFP-GFP-LC3 reporter and JC-1 staining further confirmed that oligo-anti-Linc00707 significantly enhanced autophagic flux and reduced mitochondrial membrane potential in Linc00707-overexpressing MDA-MB-231 and MDA-MB-468 cells (Supplementary Fig. S6M–O). Western blotting analysis further demonstrated increased LC3-II/I ratios and decreased SQSTM1/p62 levels following oligo-anti-Linc00707 administration (Supplementary Fig. S6P), confirming restoration of autophagy suppressed by Linc00707 overexpression. To evaluate the therapeutic potential of targeting Linc00707 in vivo, xenograft tumor models were established using MDA-MB-231 cells overexpressing Linc00707. Orthotopic administration of oligo-anti-Linc00707 significantly suppressed tumor growth compared with control treatment, as reflected by reduced tumor growth rates, tumor volumes, and tumor weights (Fig. 7E–G). Immunohistochemical analysis further revealed a marked decrease in Ki-67-positive tumor cells following oligo-anti-Linc00707 treatment (Fig. 7H). The results suggest that oligo-anti-Linc00707 can effectively inhibit the growth of MDA-MB-231 tumors and Linc00707 is a potential therapeutic target and anti-Linc00707 treatment. Discussion Transcriptional and post-transcriptional regulation are fundamental mechanisms that orchestrate gene expression programs underlying cancer progression [37–40]. In this study, we delineate a multifaceted regulatory mechanism by which the long noncoding RNA Linc00707 promotes TNBC progression through both transcriptional and post-transcriptional modulation. We identify FOXP3 as a key transcriptional activator of Linc00707 and demonstrate that WTAP-mediated m6A modification, recognized by IGF2BP2, enhances the stability of Linc00707 RNA. Furthermore, we reveal that nuclear Linc00707 exerts oncogenic effects by directly binding to and stabilizing the splicing factor U2AF2, thereby promoting cell proliferation, metastasis, and autophagy suppression in TNBC cells. Our findings provide new insights into the complex regulatory networks that govern TNBC progression, particularly the interplay between transcription factors, RNA modifications, and lncRNA–protein interactions. While FOXP3 has been previously implicated in immune regulation [41–43], our data uncover a noncanonical role for FOXP3 in driving oncogenic lncRNA expression in TNBC. This highlights the broader relevance of immune-related transcription factors in tumor intrinsic pathways. The identification of WTAP and IGF2BP2 as key modulators of Linc00707 stability extends the understanding of m6A modifications in TNBC. WTAP, together with METTL3 and METTL14, forms the core m6A methyltransferase complex and has been shown to regulate RNA stability and oncogenic transcriptomes in multiple cancer types [27, 44, 45]. Although m6A-dependent regulation of mRNA has been widely reported [46–48], our study underscores the importance of lncRNA m6A modification in maintaining oncogenic RNA function, adding to the growing evidence that epitranscriptomic control is central to cancer biology. Mounting studies have demonstrated LncRNAs exert their function by interacting with proteins and regulated expression of targets by means of specific protein [49, 50]. Our results of RNA pull-down assay followed by mass spectrometry and RIP assays indicated that U2AF2, an important RNA-binding protein, was directed binding protein. LncRNA can bind to target proteins and affect their stability, thereby affecting the expression of target proteins. For example, Linc00707 impairs the natural killer cell antitumour activity in hepatocellular carcinoma through decreasing YTHDF2 stability [51]. Previous studies, including our own, have established that Linc00707 promotes TNBC progression through activation of the PI3K–AKT–mTOR signaling axis, a well-characterized negative regulator of autophagy [15]. Given the central role of mTOR in autophagy control, it is likely that multiple downstream mechanisms cooperate to shape the autophagic and proliferative phenotypes driven by Linc00707. In the present study, we identify U2AF2-mediated alternative splicing of ATG4B as an additional layer of regulation contributing to autophagy suppression. U2AF2 is a well-characterized splicing regulator [52, 53], and while it has not traditionally been associated with autophagy suppression in TNBC, recent evidence indicates a broader oncogenic role. Rather than functioning independently of mTOR signaling, this splicing-based mechanism may operate in parallel with, or converge upon, canonical kinase pathways at the level of autophagy execution. This interpretation is supported by emerging evidence indicating bidirectional interactions between ATG4B activity and mTORC1 signaling, underscoring the complexity of autophagy regulation. Notably, the manipulation of ATG4B splicing alone was sufficient to alter autophagic flux, mitochondrial homeostasis, and proliferative capacity, suggesting that RNA splicing regulation can exert a direct and functionally significant impact on autophagy, beyond upstream signaling events. These findings expand the current understanding of how lncRNAs coordinate oncogenic processes by integrating transcriptional, post-transcriptional, and signaling-based regulatory mechanisms. Although our data support a functional role for WTAP-mediated m6A modification and IGF2BP2 in stabilizing Linc00707, the present study does not resolve the precise m6A sites required for this effect, nor does it define IGF2BP2 binding at nucleotide resolution. MeRIP assays provide transcript-level enrichment without distinguishing individual methylation sites, and depletion of m6A readers establishes functional relevance rather than direct physical occupancy. Thus, while our findings support a WTAP–m6A–IGF2BP2 regulatory axis, high-resolution mapping will be required to delineate the exact molecular determinants underlying this interaction. Multiple m6A readers and RNA-binding proteins have been reported to influence RNA stability in a context-dependent manner. Consistent with this, knockdown of additional readers, including YTHDC2 and HUR, also affected Linc00707 abundance. Notably, IGF2BP2 exerted the strongest and most consistent effects on Linc00707 steady-state levels, RNA decay kinetics, and WTAP-dependent regulation, supporting a predominant role in this cellular context rather than exclusive specificity. Future studies employing IGF2BP2 RIP or eCLIP approaches, combined with site-directed mutagenesis of candidate m6A motifs, will be required to determine whether specific methylation sites are necessary for reader recruitment and to dissect potential cooperative or competitive interactions among m6A readers on Linc00707. Together, these findings establish WTAP–m6A–IGF2BP2 as a functionally relevant regulatory axis controlling Linc00707 stability, while highlighting important avenues for further mechanistic refinement. Therapeutically, we demonstrate that LNA-modified antisense oligonucleotides targeting the U2AF2-binding motif of Linc00707 effectively disrupt its function and suppress TNBC tumor growth both in vitro and in vivo. These results support the potential of Linc00707 as a therapeutic target in TNBC and establish a proof-of-concept for lncRNA-targeted intervention based on functional interaction domains. Given the challenges associated with targeting lncRNAs, identification of such regulatory interfaces may provide a tractable strategy for therapeutic development. Despite the mechanistic insights provided, this study has several limitations. The broader regulatory network of Linc00707, including potential cofactors such as FOXP3 and the full spectrum of m6A modification dynamics, remains to be characterized. In addition, while we confirmed a functional role for U2AF2 in TNBC cells, the complete set of downstream splicing targets regulated by U2AF2 has yet to be defined. Finally, although our in vivo findings based on xenograft models support a tumor-promoting role for Linc00707, validation in more clinically relevant systems and optimization of delivery strategies will be required for translational application. Conclusions Our study defines Linc00707 as a critical nexus linking transcriptional regulation, m6A-dependent RNA stability, and splicing factor modulation in TNBC. Targeting the Linc00707-U2AF2 axis offers a promising strategy for combating TNBC and potentially other malignancies driven by similar lncRNA-mediated mechanisms. Supplementary information Additional file 1. Additional file 2. Additional file 3. Additional file 4. Additional file 5. Additional file 6. Additional file 7. Additional file 8. Additional file 9. Additional file 10. Additional file 11. Additional file 12. Additional file 13. Additional file 14. Additional file 15. Additional file 1. Additional file 2. Additional file 3. Additional file 4. Additional file 5. Additional file 6. Additional file 7. Additional file 8. Additional file 9. Additional file 10. Additional file 11. Additional file 12. Additional file 13. Additional file 14. Additional file 15.