miR-96-5p antagonizes FOXQ1-driven WNT/β-catenin signaling to inhibit triple-negative breast cancer.

Introduction
Breast cancer is the leading cause of cancer incidence and cancer-related mortality among women, with 2.3 million new cases and 665,000 deaths reported globally in 20221. Triple-negative breast cancer (TNBC), accounting for 15%-20% of all breast cancers2, is distinguished by its highly aggressive nature. Due to the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression, TNBC does not respond to endocrine or HER2-targeted therapies. Moreover, the scarcity of effective molecular targets further restricts treatment options3, leading to a median overall survival of only 12 to 18 months for patients with metastatic TNBC4. Consequently, TNBC is considered one of the subtypes with the poorest prognosis. Identifying viable molecule targets continues to be a critical challenge in the clinical management of TNBC.
FOXQ1, also known as HNF-3/HFH homolog-1 (HFH-1), is a member of the forkhead protein family. Initially identified as a transcription factor, it plays a vital role in the development of hair follicles5,6. Later, FOXQ1 was recognized as an oncogene because numerous studies have demonstrated that FOXQ1 regulates various biological processes, including tumor proliferation, inflammation, epithelial-mesenchymal transition (EMT), migration, invasion, and angiogenesis7–9. In breast cancer, FOXQ1 was required for FGFR1 activation and upregulated proliferation10. It recruits the MLL complex to activate the transcription of genes associated with EMT and promote breast cancer metastasis8. Moreover, FOXQ1 encourages aggressiveness and radioresistance in breast cancer by interacting with a nuclear isoform of RAPH111. Consequently, FOXQ1 induction has been recognized as an independent prognostic factor for gastric and colorectal cancer and a promising therapeutic target for metastatic cancer treatment12.
The Wnt/β-catenin signaling pathway, initially identified by Nusse and Varmus in 198213, is a highly conserved mechanism that plays a critical role in regulating stem cell differentiation14, organ development15, and tissue regeneration16. Hyperactivation of the Wnt/β-catenin signaling pathway is intricately linked to breast cancer initiation, progression, and malignant transformation17. Therefore, the Wnt/β-catenin pathway is acknowledged as a crucial signaling pathway for tumor progression. Our prior research revealed a positive correlation between FOXQ1 expression and the activation of the FGFR1/Wnt/β-catenin signaling pathway in breast cancer18. Moreover, Peng et al.19 illustrated that silencing FOXQ1 in colorectal cancer inhibits the nuclear translocation of β-catenin, consequently attenuating Wnt signaling activity. These findings collectively suggest that FOXQ1 is a critical regulator of the Wnt/β-catenin signaling pathways, which are intricately involved in breast cancer progression. Nevertheless, the exact mechanisms through which FOXQ1 modulates this pathway in TNBC remain largely elusive. Additionally, the potential negative regulators that interact with FOXQ1 and the Wnt/β-catenin signaling pathways warrant further exploration.
MicroRNAs (miR) are a group of non-coding single-stranded RNA molecules composed of 19–25 nucleotides20. It can recognize and directly bind to the complementary site within the 3’-untranslated region (UTR) of target genes’ mRNA, resulting in degradation or transcriptional repression. Consequently, they exhibit potent regulatory capabilities over gene expression. Through their targeted interaction with specific molecules, several miRNAs have been identified as potential suppressors of breast cancer. For instance, miR-142-3p was up-regulated in breast cancer tissue and suppressed breast cancer malignancy by targeting HMGA221. Xu et al.22 reported that miR-193, as an oncogene, enhanced cell EMT and proliferation through the ING5/PI3K/AKT signal pathway in TNBC. These findings highlight the potential of miRNAs for therapeutic interventions in TNBC.
This study systematically investigated the functional interplay among FOXQ1, WNT2, and the Wnt/β-catenin signaling pathway in TNBC cells by integrating population-based data, bioinformatic analyses, and a comprehensive series of in vitro and in vivo experiments. Furthermore, we identified multiple microRNAs (miRNAs) targeting FOXQ1 through bioinformatic prediction and functionally validated the role of miR-96-5p in suppressing FOXQ1-driven tumorigenesis in TNBC. Our findings uncover a novel molecular mechanism underlying TNBC progression and suggest potential therapeutic strategies for its treatment.

Materials and methods

Cell culture
Human TNBC cell lines (MDA-MB-231 and Hs578T) were purchased from the Shanghai Cell Bank(Shanghai, China) MDA-MB-453, BT474, MCF-7, T47D, and HEK293T cells were provided by the Institute for Cancer Medicine, School of Basic Medical Sciences, Southwest Medical University. The cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Cytiva HyClone, USA) supplemented with 10% fetal bovine serum (FBS, PAN, Germany) at 37 °C in a tissue culture incubator with 5% CO2.

Construction of cell lines that stably express FOXQ1
The FOXQ1-expressing plasmid pEZ-LV201 (EX-Y5225-LV201) and its empty control vector LV201CT (EX-NEG-LV201) were purchased from GeneCopoeia (Rockville, USA). The cells were transfected with each plasmid DNA using Lipofectamine 8000 (c0533, Beyotime Biotechnology, Shanghai, China) at a DNA-to-transfection reagent ratio of 1:1.25. Five hours after transfection, the medium was replaced with fresh medium containing 10% FBS. Once significant green fluorescence was observed in the cells, the culture medium was changed to a growth selection medium containing 2.25 µg/mL puromycin. After 2 weeks of selection, the enhanced green fluorescent protein (eGFP)-positive cells were isolated from the surviving population by flow cytometry and expanded. Then, the FOXQ1 messenger RNA (mRNA) and protein levels were analyzed using real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) and western blotting, respectively.

Oligonucleotide transfection
Two small interfering RNAs (siRNAs; si-1# and si-2#) targeting FOXQ1 and a negative control were obtained from Sangon Bioengineering Co., Ltd. (Shanghai, China). The miR-96-5p mimic and its negative control vector were sourced from General Bio Co (Anhui, China). These oligonucleotides were transfected into cells using Lipofectamine 8000 at an oligonucleotide to transfection reagent ratio of 1:1.25. The knockdown efficiency of the FOXQ1 protein was evaluated with western blotting, and the levels of miR-96-5p in cells were assessed with RT-qPCR.

Western blotting
Protein was extracted with radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime Biotechnology) containing 1% ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor cocktail (MedChemExpress, USA), and quantified with a BCA protein assay kit (Beyotime Biotechnology). Forty micrograms of protein were separated by 10% or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (P-N66485, PALL, USA). After incubating with 5% bovine serum albumin (BIO FROXX, USA) dissolved in Tris-buffered saline (pH 8.3) and 0.1% Tween-20 (TBS-T) at room temperature for 1 h to block non-specific protein binding, the membranes were incubated with primary antibodies specific to FOXQ1 (1:250,sc-166265, Santa-Cruz, USA), WNT2 (1:1000, 27214-1-AP, Proteintech, USA), phosphorylated GSK-3β (1:1000, AF5830, Beyotime Biotechnology), GSK-3β (1:1000, AG751, Beyotime Biotechnology), β-catenin (1:5000,51067-2-AP, Proteintech), and GAPDH (1:5000, 60004-1-Ig, Proteintech) overnight at 4 °C. Subsequently, the membranes were incubated for 2 h at room temperature with a fluorescence-labelled secondary antibody: DyLight 800-conjugated anti-rabbit IgG (5151P, Cell Signaling, USA) or DyLight 800-conjugated anti-mouse IgG (5257P, Cell Signaling, USA). After this incubation, the membranes were washed three times with TBS-T. The Odyssey Imaging System (LI-COR, NE, USA) was utilized to monitor the fluorescence intensity of the bands. Band intensity was analyzed using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). GAPDH served as a loading control for total proteins.

RT-qPCR
Total RNA was extracted from cells using Trizol (vazyme, China), and HiScript III All-in-one RT SuperMix Perfect for qPCR(R333-01, Vazyme, China) and miRNA 1st Strand cDNA Synthesis kit(MR201-, Vazyme, China) were used to reversely transcribe RNA into cDNA for mRNA and miRNA, respectively. Real-time PCR was performed with cDNA, primers, and the Taq Pro Universal SYBR qPCR Master Mix (Vazyme). β-actin or U6 served as the internal reference. The relative levels of FOXQ1 and miRNAs were calculated according to the 2−ΔΔCt method. The following primers were used: FOXQ1: forward, 5′-GCGGACTTTGCACTTTGAA-3′, and reverse, 5′-TTTAAGGCACGTTTGATGGA-3′; Vimentin: forward,5′-GAGAACTTTGCCGTTGAAGC-3′, and reverse, 5′-GCTTCCTGTAGGTGGCAATC-3′; miR-96-5p: forward, 5′-GCTTTGGCACTAGCACATTTTTGCT-3′, and reverse, the Novozyme qPCR Universal Primer Q(Vazyme); U6: forward 5′-CTCGCTTCGGCAGCACA-3′, and reverse, 5′-AACGCTTCACAATTTGCGT-3′; and β-actin: forward, 5′-CCTAGAAGCATTTGCGGTGG-3′, and reverse 5′-GAGCTACGAGCTGCCTGACG-3′.

Transwell migration and invasion assay
A transwell assay was used to measure the migration and invasion ability of TNBC cells. For this assay, a 24-well chamber (Corelle, Kenny Bunker) was pre-coated with Matrigel (BD, Franklin Lakes, NJ, USA) diluted in serum-free medium at a ratio of 1:8 (for the invasion assay) or without Matrigel (for the migration assay). Approximately 30,000 cells were mixed with 300 µL of serum-free DMEM and placed in the upper chamber; 500 µL of DMEM containing 20% serum was added to the lower chamber. After culturing for 24 h, the migratory or invasive cells were washed with phosphate-buffered saline (PBS), fixed with methanol, stained with 0.1% crystal violet solution, and counted under a microscope.

Luciferase reporter assay
Pmir-GLO (Hebio, Shanghai, China) is a plasmid containing both firefly luciferase and Renilla luciferase. The 3′ untranslated region (UTR) of FOXQ1 mRNA was amplified with PCR and inserted into the downstream area of the SV40 promoter-driven Renilla luciferase cassette in the Pmir-GLO plasmid. The FOXQ1 binding sequence region was antisense-mutated, and the mutant Pmir-GLO plasmid (MUT) was constructed in the same way. Next, HEK293T cells were transfected with miR-96-5p mimics and NC, as well as Pmir-GLO. After 48 h, the cells were lysed and collected for measurement of luciferase activity by the dual-luciferase assay system (Beyotime Biotechnology), following the manufacturer’s protocols.

TOP/FOP-Flash reporter assay
To evaluate the transcriptional activation of the β-catenin-mediated Wnt signaling pathway, HEK293T cells were co-transfected with either the TOP-Flash or FOP-Flash firefly luciferase reporter plasmid, along with the SV40-Renilla luciferase plasmid (Beyotime Biotechnology). The TOP-Flash reporter system, which contains functional TCF/LEF binding sites, was used to assess β-catenin transcriptional activity. As a negative control, the FOP-Flash construct, containing mutated TCF/LEF binding sites, was used in parallel experiments. Following transfection, cells were treated for 24 h with one of the following: vehicle control, the Wnt inhibitor IWP-223 (8 µM), FOXQ1 overexpression, or FOXQ1 combined with IWP-2. Luciferase activity was measured using the Dual Luciferase Assay Reporter System (Beyotime Biotechnology). In all assays, Renilla luciferase activity was measured concurrently and used for normalization to account for variations in transfection efficiency and cell viability. The concentration of IWP-2 (8 µM) was selected based on a CCK-8 cytotoxicity assay, which confirmed minimal cytotoxicity under the experimental conditions. Additionally, the efficacy of IWP-2 at this concentration in suppressing Wnt/β-catenin signaling was independently validated using the TOP/FOP-Flash reporter assay (see Supplementary Figure S1).

Animal experiments
All animal experimental procedures adhered to the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. The experimental protocol was reviewed and approved by the Experimental Animal Ethics Committee of Southwest Medical University under protocol No. 20220624-010. Female BALB/c-nu mice (3–4 weeks old) were purchased from Chengdu Pharmachem Biotechnology Co., Ltd. (Chengdu, China). On day 1, before surgery, mice were anesthetized via intraperitoneal injection with 3% sodium pentobarbital at a dose of 1.5 mg/kg body weight. Ten million Hs578T cells that stably overexpress FOXQ1 or control cells were injected into the mammary fat pad of each mouse. Each group consisted of 10 mice. After 9 days, when most tumors had reached approximately 50 mm3 in volume, the tumor formation rate was recorded. Tumor volume was calculated using a caliper with the modified ellipsoidal formula: volume (mm3) = (width×width×length)/2. Mice bearing FOXQ1-overexpressing tumors were then randomly divided into two groups and received either miR-96-5p agonist or saline injection intratumorally every 2 days. Control mice that developed tumors (n = 5) received intratumoral saline injection every 2 days, while those that did not develop tumors were given saline injections in the mammary fat pad. On day 20, all mice were humanely euthanized via excessive CO2 inhalation according to the approved protocol No. 20220624-010. Tumor size and weight were subsequently measured. All procedures, including anesthesia, tumor inoculation, and euthanasia, were designed to minimize animal suffering.

Immunohistochemistry
Fresh tumor samples were fixed, embedded with paraffin, and sectioned, and the endogenous peroxidase activity was inactivated, as previously described10. The sections were incubated with goat-serum-containing blocking solution (Cat #abs933, Absinand) and then with the following antibodies overnight at 4 °C: rabbit anti-FOXQ1 antibody (1:250, sc-166265, Santa-Cruz), rabbit anti-KI67 antibody (1:1000, GB111499, Servicebio, China), and rabbit anti-vimentin (1:800, GB11192, Servicebio). After three washes with PBS, the cells were incubated with horseradish peroxidase-labelled secondary antibodies of the corresponding species for 50 min. After incubation with 3,3’-diaminobenzidine (DAB) to develop a colored product, the sections were stained with hematoxylin for 3 min, dehydrated, sealed, and examined with a microscope. The FOXQ1 and vimentin immunoreactivity scores, as well as the percentage of Ki67-positive cells, were calculated as described previously10.

Software and databases
Population data were sourced from The Cancer Genome Atlas (TCGA) (https://www.cancer.gov/ccg/research/genome-sequencing/tcga). The Kaplan-Meier Plotter (https://kmplot.com/analysis/index.php?p=service) was applied to assess the association between FOXQ1 expression levels and both overall survival and recurrence-free survival. Potential downstream targets of FOXQ1 were predicted using TRANSFAC (https://genexplain.com/transfac/) and TRRUST (https://www.grnpedia.org/trrust/). Gene Ontology (GO) enrichment analysis was performed using R software (version 4.3.1) with the clusterProfiler package (v4.10.0) to identify biologically relevant pathways. Data visualization was conducted using the ggplot2 package (v3.4.4). The Starbase database (https://starbase.sysu.edu.cn/) was used to evaluate the correlation between FOXQ1 and WNT2 mRNA expression. At the same time, the Cancer Cell Line Encyclopedia (CCLE) (https://sites.broadinstitute.org/ccle) was utilized to investigate the relationship between FOXQ1 expression and WNT2 copy number variations. Putative microRNAs targeting FOXQ1 were identified through TargetScan (https://www.targetscan.org/vert_80/) and miRDB (https://mirdb.org/).

Statistical analysis
Each experiment was repeated more than three times, and the graphs in the figures show the representative results of these experiments. Statistical analysis was performed using GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA). The data are presented as the mean ± standard deviation. The difference between the two groups was assessed using an unpaired Student’s t-test, while the difference among multiple groups was evaluated using a one-way analysis of variance (ANOVA).

Results

FOXQ1 is significantly upregulated in TNBC and is associated with an unfavorable prognosis
To investigate the expression pattern of FOXQ1, we conducted RT-qPCR to examine the FOXQ1 mRNA levels across various breast cancer cell lines. There was a significantly higher average level in three TNBC cell lines, specifically MDA-MB-231, MDA-MB-453, and Hs578T, compared with non-TNBC cell lines, such as BT474, MCF-7, and T47D, all of which were cultured in our laboratory (Fig. 1a). Additionally, we analyzed the FOXQ1 mRNA levels in 1079 breast cancer samples from the TCGA database. The FOXQ1 mRNA levels were notably higher in the basal-like subtype compared with the HER2, luminal A, and luminal B subtypes, as well as normal tissue samples (P < 0.05) (Fig. 1b). Furthermore, we investigated the correlation between FOXQ1 expression and patient outcomes using data from the TCGA database. Kaplan-Meier Plotter analysis demonstrated a significant correlation between increased FOXQ1 expression and both unfavorable overall survival (HR = 1.34, P = 0.041) and poorer recurrence-free survival (HR = 1.19, P = 0.037) (Fig. 1c and d). These findings suggest that FOXQ1 expression is significantly upregulated in TNBC and is associated with a poorer prognosis.

FOXQ1 plays a crucial role in enhancing the migration and invasion capabilities of TNBC cells
Having previously demonstrated that elevated FOXQ1 levels promote breast cancer cell proliferation and growth10, we further explored the functional role of FOXQ1 in TNBC progression. To evaluate its influence on cell migration and invasion, we selected two TNBC cell lines—MDA-MB-231 and Hs578T—that exhibit relatively high endogenous FOXQ1 expression. Efficient depletion of FOXQ1 mRNA was achieved by transfecting cells with two distinct siRNAs (si-1# and si-2#), as validated by qRT-PCR and immunoblotting (Fig. 2a and d). Knockdown of FOXQ1 significantly impaired the migratory (Fig. 2e and f) and invasive (Fig. 2g and h) abilities of both cell lines. In contrast, ectopic overexpression of FOXQ1 (Fig. 2i and l) markedly enhanced migration (Fig. 2m and n) and invasion (Fig. 2o and p) in MDA-MB-231 and Hs578T cells. Collectively, these findings demonstrate that FOXQ1 plays a critical role in driving the migratory and invasive phenotypes of TNBC cells.

FOXQ1 enhances activation of the Wnt/β-catenin signaling pathway by upregulating WNT2 expression
To elucidate the molecular mechanism underlying the impact of FOXQ1 on TNBC migration and invasion, we conducted a bioinformatic analysis to predict the downstream targets of FOXQ1 by using the TRANSFAC (https://genexplain.com/transfac/) and TRRUST (https://www.grnpedia.org/trrust/) databases. We employed the overlap dataset derived from the two databases for GO enrichment analysis. FOXQ1 target genes are implicated in various biological processes, including embryonic organ development, axonogenesis, forebrain development, and gland development (Fig. 3a). Notably, the category of embryonic organ development was enriched with the highest number of genes (24 genes). Importantly, this category includes two key members of the Wnt family, WNT2 and WNT3A. We concentrated our subsequent investigation on the impact of FOXQ1 on WNT2 expression rather than on all WNT proteins. This decision was motivated by the critical role of WNT2 as a ligand for the frizzled receptor in initiating the canonical Wnt/β-catenin signaling pathway.

We initially examined the correlation between FOXQ1 and WNT2 expression. According to the Starbase database (https://starbase.sysu.edu.cn/), FOXQ1 and WNT2 expression levels had a positive correlation in 1104 breast cancer samples (r = 0.367, p < 0.05) (Fig. 3b). Furthermore, data from the Cancer Cell Line Encyclopedia (CCLE) database (https://sites.broadinstitute.org/ccle) also demonstrated a significant correlation between the copy numbers of these two genes (r = 0.6655, p = 0.0019) (Fig. 3c).
Subsequently, we investigated the impact of FOXQ1 on WNT2 protein expression and the activity of the Wnt/β-catenin signaling pathway. Consistent with our hypothesis, overexpression of FOXQ1 in MDA-MB-231 cells resulted in an approximately 1.8-fold increase in WNT2 expression, a two-fold increase in phosphorylation of GSK-3β (Fig. 3d), and enhanced nuclear localization of β-catenin (Fig. 3e). Notably, treatment with the WNT2 inhibitor IWP-2 abolished the FOXQ1-induced increase in GSK-3β phosphorylation and nuclear β-catenin localization without altering FOXQ1 overexpression-induced WNT2 upregulation (Fig. 3d and e). Similar results were observed in Hs578T cells (Fig. 3f and g). These findings suggest that FOXQ1 can upregulate WNT2 independently of Wnt/β-catenin signaling pathway activity.
Next, we examined the impact of FOXQ1 on the activity of the Wnt/β-catening pathway. The TOP/FOP Flash assay, a well-established method for assessing the activation of the Wnt/β-catenin signaling pathway, revealed that β-catenin-mediated transcriptional activity was nearly doubled in the FOXQ1-overexpressing cells compared with the control cells (Fig. 3h). Conversely when Wnt protein activity was inhibited by IWP-2, the overexpression of FOXQ1 no longer enhanced β-catenin-mediated transcriptional activity (Fig. 3h). These findings further confirmed that FOXQ1 enhances Wnt/β-catenin signaling pathway activity by upregulating WNT2 expression.

WNT2 mediates EMT, migration, and invasion promoted by FOXQ1 in TNBC
Subsequently, we investigated the role of WNT2 in FOXQ1-regulated migration and invasion of TNBC cells. Based on western blotting, MDA-MB-231 cells exhibited a lack of E-cadherin and N-cadherin protein expression. Overexpression of FOXQ1 resulted in elevated expression of vimentin, a marker of EMT. This effect was abrogated upon inhibition of WNT2 activity by IWP-2 treatment (Fig. 4a). We observed consistent results in Hs578T cells (Fig. 4b). A transwell assay further demonstrated that suppression of WNT2 by IWP-2 significantly diminished the migration and invasion of MDA-MB-231 (Fig. 4e and f) and Hs578T (Fig. 4g and h) cells. Collectively, these findings suggest that WNT2 mediates EMT, migration, and invasion, which FOXQ1 promotes in TNBC.

miR-96-5p negatively regulates FOXQ1 expression by directly targeting the 3′UTR of its mRNA
To identify specific endogenous molecules that inhibit FOXQ1, we performed a comprehensive analysis leveraging population and bioinformatics data. Initially, we screened for differentially expressed miRNAs in cancerous versus non-cancerous tissues from the TCGA database, identifying 137 miRNAs with at least a twofold change in expression (Fig. 5a). Then, we utilized the Targetscan (https://www.targetscan.org/vert_80/) and miRDB (https://mirdb.org/) databases to predict potential miRNAs that target FOXQ1, which yielded 69 and 59 candidates, respectively (Fig. 5a). By cross-referencing these datasets, we identified seven miRNAs present across all three sources: miR-96-5p, miR-182-5p, miR-378a-3p, miR-133a-3p, miR-140-3p, miR-1277-5p, and miR-342-3p (Fig. 5a and b). Among these, we selected miR-96-5p due to its high overall score (Fig. 5b) and its established role in regulating breast cancer growth, migration24, and invasion25.
We first validated the association between miR-96-5p expression and patient outcomes by using population data from the TCGA database. Kaplan-Meier survival analysis (https://kmplot.com/analysis/) revealed that patients with higher miR-96-5p expression exhibited significantly improved overall survival compared with those with lower miR-96-5p expression (HR = 0.82, P < 0.05) (Fig. 5c).
We then investigated the association between miR-96-5p levels and FOXQ1 expression by using data from the TCGA database. Pearson correlation analysis revealed a negative correlation between FOXQ1 mRNA expression and miR-96-5p expression in 1,085 breast cancer samples from the TCGA database (r = -0.101, p < 0.001) (Fig. 5d).
Based on the findings above, we investigated how miR-96-5p regulates the expression of FOXQ1. We increased miR-96-5p levels in FOXQ1-overexpressing MDA-MB-231 cells by transfecting them with miR-96-5p mimics (Fig. 5e). As a result, the FOXQ1 mRNA (Fig. 5f) and protein levels (Fig. 5g) were reduced to 10% and 70% of the control levels, respectively. Similarly, in FOXQ1-overexpressing Hs578T cells transfected with miR-96-5p mimics (Fig. 5h), the upregulation of miR-96-5p resulted in a decrease in FOXQ1 mRNA (Fig. 5i) and protein expression (Fig. 5j) to 20% and 50%, respectively. A dual luciferase reporter assay demonstrated that treatment with miR-96-5p mimics led to a more than 2-fold reduction in fluorescence intensity of the luciferase construct containing the wild-type FOXQ1 3′UTR compared with the control group (Fig. 5k). In contrast, we observed no significant effect for the construct harboring the mutant-type FOXQ1 3′UTR (Fig. 5k). These results robustly demonstrated that miR-96-5p specifically inhibits FOXQ1 expression by targeting its 3′UTR.

miR-96-5p suppresses FOXQ1-mediated migration and invasion of TNBC
The introduction of miR-96-5p into FOXQ1-overexpressing MDA-MB-231 cells significantly inhibited cell migration and invasion (Fig. 6a and b). However, these inhibitory effects were reversed upon restoration of FOXQ1 expression (Fig. 6a and b). Similar results were observed in Hs578T cells (Fig. 6c and d). Collectively, these findings indicate that miR-96-5p suppresses TNBC cell migration and invasion by targeting FOXQ1.

miR-96-5p inhibits FOXQ1-mediated malignant progression of TNBC in vivo
We conducted animal experiments to investigate the in vivo impact of miR-96-5p on the malignant progression of TNBC. FOXQ1-overexpressing Hs578T cells and control cells were inoculated into the breast fat pad of BALB/C-nu mice, with 10 mice per group. Nine days post-inoculation, 50% of the mice (5 out of 10) in the control group had developed tumors. In contrast, all mice (10 out of 10) in the FOXQ1-overexpressed group developed tumors, resulting in a 100% tumorigenicity rate. These results provide robust evidence that overexpression of FOXQ1 promotes tumorigenesis in this experimental model (Fig. 7a).
Subsequently, we randomly and equally divided these mice harboring FOXQ1-overexpressing tumors into two groups: the saline control group (OE + saline) and the miR-96-5p agomir treatment group (OE + agomir), with five mice in each group. Meanwhile, the mice injected with control cells continued to receive saline treatments (Fig. 7a). After five times of intratumoral injections, the OE group exhibited larger tumors (Fig. 7b), both in terms of size (Fig. 7c) and weight (Fig. 7d), compared with the control group. In contrast, the OE + agomir group demonstrated markedly smaller tumors compared with the OE group (Fig. 7c and d) but did not exhibit a significant difference compared with the control group (Fig. 7d). Immunohistochemistry revealed a substantial increase in FOXQ1 levels in the OE group, approximately double compared with the control group (Fig. 7e). Concurrently, the OE group exhibited a 20-fold increase in Ki67 positivity (Fig. 7f) and a 3-fold increase in vimentin expression compared with the control group (Fig. 7g). Notably, the introduction of miR-96-5p agomir markedly reduced FOXQ1 levels (Fig. 7e) and reversed the elevated Ki67 (Fig. 7f) and vimentin expression (Fig. 7g). These results suggest that FOXQ1 overexpression significantly enhances the growth and EMT of TNBC, while miR-96-5p can inhibit the FOXQ1-driven malignant progression of TNBC in vivo.

Discussion
Triple-negative breast cancer (TNBC) is characterized by aggressive clinical behavior, including rapid proliferation, epithelial–mesenchymal transition (EMT), and a high propensity for metastasis3,26–28. Although FOXQ1 has been linked to the activation of Wnt/β-catenin signaling in cancer progression, the precise molecular mechanisms underlying this relationship remain incompletely understood. In this study, we demonstrate that FOXQ1 enhances Wnt/β-catenin signaling by upregulating the Wnt ligand WNT2, thereby promoting the aggressiveness of TNBC. Our findings reveal a previously unrecognized regulatory axis between FOXQ1 and WNT2 that drives Wnt pathway activation, providing new mechanistic insights into the progression of TNBC (Fig. 8).

The canonical Wnt/β-catenin pathway, a key regulator of cell fate and EMT, is activated by Wnt ligands that stabilize β-catenin, enabling its nuclear translocation to drive the expression of pro-tumorigenic genes like SNAIL and c-MYC29,30. Among these ligands, WNT2 acts as a key activator of the canonical pathway. Although essential in embryonic development, WNT2 is often silenced in adult tissues but is reactivated in multiple malignancies31–33. For instance, in gastric cancer, WNT2 cooperates with SOX4 to form a positive feedback loop that sustains the self-renewal of cancer stem cells34. In oesophageal cancer, WNT2 promotes cancer cell proliferation by activating the Wnt/β-catenin signalling pathway, leading to the upregulation of cyclin D1 and c-myc expression32. Thus, WNT2 functions not merely as a binary Wnt switch but as a dynamic signaling hub embedded within broader oncogenic networks35.
Previous studies, including one on colorectal cancer, have identified FOXQ1 as a transcriptional target of the Wnt/β-catenin signaling pathway36. Here, we uncover a reverse regulatory relationship in TNBC: bioinformatic analysis revealed a conserved FOXQ1-binding motif in the WNT2 promoter, and functional studies confirmed that FOXQ1 overexpression upregulates WNT2 expression, stimulates β-catenin nuclear translocation, and enhances downstream transcriptional activity. Critically, inhibition of Wnt ligand palmitoylation and secretion using IWP-2 abolished β-catenin activation but did not suppress FOXQ1-induced WNT2 upregulation, indicating that FOXQ1 transcriptionally controls WNT2 independently of canonical Wnt/β-catenin signaling. Interestingly, IWP-2 treatment partially reduced FOXQ1 protein levels, suggesting that FOXQ1 expression remains influenced by Wnt activity, which is consistent with earlier reports in colorectal cancer37. Together, these data support a model in which FOXQ1, WNT2, and β-catenin form a positive feedback loop that sustains constitutive Wnt pathway activation in TNBC. This self-reinforcing circuit may underlie the persistent aggressiveness of this breast cancer subtype (Fig. 8).
Based on the established role of this pathogenic feedback loop, we further explored its therapeutic potential. We identified miR-96-5p as a microRNA that directly targets FOXQ1 and effectively suppresses its expression. Although miR-96-5p is known for its context-dependent oncogenic or tumor-suppressive roles across various cancers38–41, our data firmly establish its tumor-suppressive function in TNBC. Analysis of TCGA data revealed that high expression of miR-96-5p is significantly associated with improved survival in a cohort of 3,033 breast cancer patients, consistent with a previous report that it suppresses breast cancer metastasis through CTNND142. Mechanistically, we confirmed that miR-96-5p directly binds to the 3’UTR of FOXQ1, leading to the suppression of FOXQ1 expression and subsequent inhibition of downstream Wnt/β‑catenin signaling. This interaction was shown to be specific, as mutating the miR-96-5p binding site abolished its regulatory effect. Functionally, overexpression of miR-96-5p effectively counteracted FOXQ1-driven proliferation, migration, and invasion in both in vitro and in vivo models, underscoring its therapeutic promise.
Collectively, our work delineates a previously unrecognized pathogenic circuit in TNBC, wherein a positive feedback loop connecting FOXQ1, WNT2, and β-catenin drives and sustains Wnt pathway hyperactivation. The identification of miR-96-5p as a direct and potent suppressor of FOXQ1 not only provides mechanistic insight but also unveils a promising therapeutic vulnerability.
Several considerations for future work emerge from our findings. First, while our data firmly establish the core FOXQ1-WNT2-β-catenin axis, the potential involvement of additional parallel effectors downstream of FOXQ1 remains to be fully elucidated. Second, the generalizability of this signaling module across the spectrum of TNBC heterogeneity, beyond the mesenchymal-subtype models primarily used here, awaits validation in a broader panel of models, including basal-like subtypes and patient-derived xenografts. Addressing these questions will be crucial for defining the patient population most likely to benefit from this strategy.
Notwithstanding these limitations, the therapeutic implications are compelling. The efficacy of miR-96-5p in suppressing FOXQ1-driven oncogenicity, coupled with the maturity of oligonucleotide delivery platforms such as lipid nanoparticles, positions the restoration of miR-96-5p function as a viable and attractive strategy. Future efforts should focus on advancing delivery efficacy and exploring rational combination regimens, such as those with Wnt inhibitors, to achieve durable pathway suppression and circumvent potential resistance. Ultimately, targeting the FOXQ1-miR-96-5p node offers a novel avenue to disrupt a key self-reinforcing signaling circuit in TNBC.

Conclusion
In conclusion, we define a self-reinforcing FOXQ1-WNT2-β-catenin feedback loop as a key driver of Wnt pathway hyperactivation in TNBC, and identify its potent suppression by miR-96-5p as a promising therapeutic strategy.

Supplementary Information
Below is the link to the electronic supplementary material.