Online database analysis

The expression levels of SERTAD4 mRNA were assessed using the GEPIA2 database (http://gepia2.cancer-pku.cn/) and the TNMplot database (http://www.tnmplot.com/). The protein expression levels and localization of SERTAD4 were sourced from The Human Protein Atlas (HPA,https://www.proteinatlas.org/). A survival map for SERTAD4 was generated utilizing the GEPIA2 database. Additionally, the Kaplan–Meier Plotter (https://kmplot.com/) was used to generate the survival curve for SERTAD4.

The Cancer Genome Atlas (TCGA) database analysis

The seq-RNA count data from the TCGA-BRCA dataset, along with corresponding clinical information from BC patients, were acquired for analysis. Clinical correlation analyses and visualizations were conducted utilizing the R packages “limma” and “ggpubr”.

Single-Cell RNA sequencing (scRNA-seq) analysis

The GSE161529 dataset was acquired from the GEO database (https://www.ncbi.nlm.nih.gov/geo/), and scRNA-seq data from six primary tumors were extracted. This dataset comprises two ER + BC cases, two HER2 + BC cases, and two triple-negative BC (TNBC) cases. The scRNA-seq data were initially subjected to quality control using the following criteria: a minimum of 300 and a maximum of 5,000 genes per cell, with UMI counts ranging from 1000 to 25,000. Additionally, cells exhibiting mitochondrial gene content exceeding 5% or hemoglobin gene content exceeding 1% were excluded from subsequent analysis. The “Harmony” algorithm was employed to conduct batch effect correction across samples. The data were subsequently normalized and scaled using the “NormalizeData” and “ScaleData” functions, respectively. The top 2,000 highly variable genes were identified using the “FindVariableFeatures” function. UMAP dimensionality reduction and graph-based clustering were performed using the top 40 principal components (PCA) with a clustering resolution parameter of 1.0. Cell populations were subsequently manually annotated through marker gene expression profiling. Cell–cell communication was assessed using the R package “CellChat.” The “infercnv” R package was utilized to conduct copy number variation (CNV) scoring to evaluate the malignancy of cells. Additionally, signaling pathway scoring was performed using the “ssGSEA” algorithm.

Enrichment analysis

The samples were stratified into high and low expression groups based on the median expression levels of SERTAD4. Differentially expressed genes were subsequently identified utilizing a logFC ≥ 0.25 and an adjusted p value (p_val_adj) of 0.05. For additional analysis, differentially expressed genes satisfying the criteria of logFC ≥ 1 and p_val_adj ≤ 0.05 were incorporated into Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) [10] enrichment analysis to investigate the biological functions and signaling pathways associated with SERTAD4.

Molecular docking

The molecular structure of SERTAD4 protein was retrieved from the AlphaFold database (https://alphafold.ebi.ac.uk/), while the PI3K structure was acquired from the RCSB Protein Data Bank (PDB, https://www.rcsb.org/). Subsequent molecular docking analysis was performed using PyMOL software.

Clinical samples

This study utilized samples obtained from patients presenting at the Fourth Hospital of Hebei Medical University between January 2019 and December 2021. The cohort comprised 6 normal breast tissue specimens and 81 BC tissue specimens. All samples were collected from treatment-naive patients who had received no prior systemic therapies, including chemotherapy, radiotherapy, targeted therapy, or immunotherapy. Participants with complete medical records were included, while those with concurrent malignancies, severe psychiatric disorders, or autoimmune diseases were excluded.

This retrospective study, approved by the Ethics Committee of the Fourth Hospital of Hebei Medical University, was granted a waiver of written informed consent due to its non-interventional nature. Patient data were anonymized in accordance with the Declaration of Helsinki. The research did not impact clinical management and maintained strict confidentiality of all medical records.

Immunohistochemistry (IHC)

After standard deparaffinization in xylene/ethanol gradients, endogenous peroxidase was blocked with 3% H2O2 (37 °C, 10 min). Heat-induced epitope retrieval was conducted in pH 8.0 EDTA buffer (95°C, 10 min). Primary antibody incubation (1:150 dilution) proceeded at 4 °C for 8 h, followed by 30-min secondary antibody application at room temperature. DAB development was microscopically monitored (40–60s) prior to quenching. SERTAD4 expression was evaluated by two blinded senior pathologists, with discrepancies adjudicated by joint review.

Cell culture

The MDA-MB-231 and MCF-7 cell lines were acquired from Pricella. The growth medium employed was DMEM high glucose supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin (P/S). The culture conditions were maintained at 95% O2, 5% CO2, and a temperature of 37 °C.

Quantitative Real-time PCR(qRT-PCR)

Total RNA was extracted from the MCF7 and MDA-MB-231 cell lines using TRIzol reagent (Takara, Dalian, China). The extracted RNA was subjected to reverse transcription to generate complementary DNA (cDNA), in accordance with the protocols outlined in the reverse transcription kit (Takara). Specific primers were utilized for amplification, employing cDNA as the template. The primers utilized were as follows: Forward: AGCAGAGCAGAGGACAACTT; Reverse: CGTTCTTCAGCCTGTCACAG. The expression levels of the target gene were quantified using the 2 − ΔΔCt method based on the obtained Ct values.

Western Blot

Cells were lysed on ice in radio-immunoprecipitation assay (RIPA) buffer. Equal quantities of proteins were separated via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) utilizing a 10% gel. The separated proteins were transferred onto polyvinylidene difluoride (PVDF) membranes with pore sizes of 0.2 µm and 0.45 µm. The membranes were blocked with 5% skim milk for 1 h and subsequently incubated with primary antibodies diluted 1:1,000 overnight at 4 °C. Following three washes with Tris-buffered saline containing 0.05% Tween-20 (TBST) (10 min each), immunoreactive signals were developed using a chemiluminescence solution. The antibodies utilized in the procedure include: PI3K (Phosphoinositide 3-Kinase, Huabio, ET1608-70), AKT (Protein Kinase B, Huabio, ET1609-51), p-AKT (Phospho-Protein Kinase B, Huabio, ET1609-51), and p-PI3K (Phospho-Phosphoinositide 3-Kinase, Huabio, HA721672).

RNA interference

Cells were seeded into 6-well plates and permitted to grow until they reached approximately 40% confluence. At this juncture, the cells were serum-starved using high-glucose DMEM for a duration of 1 day. Following serum starvation, siRNA and Lipofectamine 3000 (Lipo3000) were introduced to facilitate RNA interference. The knockdown efficiency of SERTAD4 was evaluated using qRT-PCR and Western blot analysis. The knockdown efficiency of SERTAD4 mRNA reaching 70% is regarded as a success. The SERTAD4-specific siRNA sequences were designed as follows: The sense strand (S) sequence is 5'-GGAACGAGCCCACAUCUUTT-3', and the antisense strand (AS) sequence is 5'-AAGGAUGUGGCGCUCUUCUTT-3'. The negative control siRNA is as follows: The S sequence is 5'- UUCUCCGAACGUGUCACGUTT-3', and the AS sequence is 5'- ACGUGACACGUUCGGAGAATT-3'.

Wound healing assay

Following the preparation of the cell suspension, the cells were seeded into 6-well plates at a density of approximately 50%. Once the cells adhered and attained over 90% confluence, a 1,000 µL pipette tip was employed to scratch the cell surface, thereby creating a wound. Photographs were captured at 0 h and 24 h post-scratch to analyze the width of the wound and evaluate cell migration capability.

Transwell

Following serum starvation, the cells were prepared as a cell suspension. A layer of Matrigel was applied to the upper chamber, and the cells were seeded into this chamber with 600 µL of complete medium containing 10% fetal bovine serum (FBS), ensuring approximately 20,000 cells per chamber. The lower chamber was filled with complete medium containing 20% fetal bovine serum (FBS). After 24 h of incubation, the upper chamber was removed, and the cells were fixed with 4% paraformaldehyde for 20 min, followed by staining with crystal violet for an additional 20 min. Cell invasion was subsequently analyzed by capturing images of the cells under a microscope.

Clone formation assay

Cells were seeded into 12-well plates and thoroughly mixed, ensuring approximately 100 cells per well. The culture medium was gently replaced every 3 days throughout the incubation period. After approximately 10 days, when visible colonies had formed, the cell culture was terminated. The cells were subsequently fixed with 4% paraformaldehyde for 20 min and stained with crystal violet for an additional 20 min. Images were captured, and the number of colonies per well was quantified to assess cell proliferation and tumorigenic potential.

Co-immunoprecipitation (Co-IP)

Co-IP was performed using the rProtein A/G Magnetic IP/Co-IP Kit (ACE Biotechnology, China). Cell lysates were incubated overnight at 4°C with target-specific antibodies conjugated to magnetic beads. Following incubation, the resin was washed after supernatant removal. Captured protein complexes were then eluted and analyzed by western blotting.

Statistical analysis

The comparison of means between two groups was conducted using the t test, whereas comparisons involving more than two groups were performed using the Wilcoxon rank-sum test and the Kruskal–Wallis test. Univariate Cox regression analysis and chi-squared tests were utilized to evaluate differences in categorical data. The log-rank test was utilized to evaluate cancer prognosis. Statistical significance was defined as p < 0.05, with * indicating p < 0.05, ** indicating p < 0.01, *** indicating p < 0.001 and **** denoting p < 0.0001. All statistical analyses were conducted using SPSS version 22.0 software.