THSG was obtained from Chengdu Herbpurify CO., LTD (Cat#E−022-160,001, Chengdu, China) and dissolved in Roswell Park Memorial Institute 1640 (RPMI 1640, Cat#12633020, Gibco, United States) with 2% fetal bovine serum (FBS) (Cat#FSD500, Excell, Jiangsu, China) at a concentration of 10 mmol/L stock solution. The concentrations used in this study were 0, 100, 200, 300, 400, and 500 μmol/L, freshly diluted in RPMI 1640 medium before use. The concentration of 0 μmol/L THSG was set as control.

T47D cells (Cat#BFN60805678, ATCC) cultured in RPMI 1640 supplemented with 10% FBS and 1% penicillin as well as streptomycin (PS, Cat#G4003, Servicebio, Wuhan, China). The cells were cultured at 37 °C with 5% CO₂ in humidified conditions. Transfection was performed with Lipofectamine 2000 (Cat#11668030, Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions. For RNA interference, cells were transfected with appropriate siRNAs using Lipofectamine RNAiMAX (Cat#13778075, Invitrogen, Carlsbad, CA, United States) and harvested 48 h later for analyses. With subsequent examinations, T47D cells was transfected with scramble siRNA (Cat#A06002, Shanghai GenePharma Co., Ltd. Shanghai, China) and Ncapd3 siRNA. (Ncapd3 siRNA sense, 5-GUG​CUG​CCU​UUC​ACU​UUA​ATT-3, and its antisense siRNA was 5-UUA​AAG​UGA​AAG​GCA​GCA​CTT-3, Cat#A03001, Shanghai GenePharma Co., Ltd. Shanghai, China). Then, T47D cells line was collected and the mRNA expression of Ncapd3 were measured using quantitative polymerase chain reaction (qPCR) in 48 h later.

The proliferation and cytotoxicity of T47D cells was assessed using the methyl thiazolyl tetrazolium (MTT) assay (Cat#C0009S, Beyotime Biotechnology, Shanghai, China). T47D cells were seeded at a cell density of 5 × 10³ cells/well in 96-well plates and cultured with different concentrations (0, 100, 200, 300, 400, and 500 μmol/L) of THSG were added for 24, 48, and 72 h. The cells were cultured for 4 h after the addition of MTT solution (10 μL per well). Subsequently, 100 μL of dimethyl sulfoxide (DMSO, Cat#ST038, Beyotime Biotechnology, Shanghai, China) per well was added, and the mixture was incubated for 10 min while the cells were shaken. Relative absorbance was determined at 570 nm by subtracting the absorbance at 630 nm. Inhibition rate (%) = 100% × (control cell OD -dosing cell OD)/(control cell OD-blank OD).

T47D cells (1 × 10²) with 100–200 μmol/L THSG or Phosphate Buffered Saline (PBS) (Cat#G4202, Servicebio, Wuhan, China) were inoculated and incubated in 6-well plates for 24 h. The cells were subsequently cultivated in mediums lacking THSG for 2 weeks. All of the dishes were subsequently rinsed multiple times with ddH2O to eliminate unattached crystal violet prior to enumerating the colonies employing ImageJ 1.53a software.

Cell cycle was evaluated using a related Kit (Cat#C1052, Beyotime Biotechnology, Shanghai, China) as follows: T47D cells were treated to various doses of THSG (0, 200, and 250 μmol/L) for a duration of 24 h. Subsequently, the cells were treated with a concentration of 200 μmol/L THSG for 24 and 48 h. Afterwards, the cells were subjected to centrifugation and re-suspended in PBS. The cells were stained with propidium iodide (PI) following the instructions provided by the manufacturer. The data were acquired using flow cytometry equipment (LSRFortessa, United States) and processed using FlowJo-V10 software (Tree Star Inc.).

Third-Generation Sequencing (TGS) was commercialized by Oxford Nanopore Technologies (ONT) Minions. T47D cells including control group (T47DC) and THSG-treated group (T47DT) pretreated with THSG (0 or 200 μmol/L) for 24 h, with three replicates per group. Total cellular RNA was extracted and mailed to Beijing Biomarker Technology Co., Ltd. for sequencing analysis. The raw data format used by Oxford Nanopore Technologies (ONT) is the second-generation fast5 format, which stores the original sequencing signals generated during the sequencing process. The data in fast5 format was transformed to fastq format using the Guppy software included in the MinKNOW 2.2 package. The consistent sequences of each sample were aligned with the reference genome by minimap2. Perform redundant-remove for alignment result, filter out sequences with identity lower than 0.9 and coverage lower than 0.85, consistent sequences of each sample can be used for alternative splicing analysis after redundant-remove. finally 44,416 redundant-removed transcript sequences can be obtained. It's recommended to use IGV (Integrative Genomics Viewer) to open alignment result file between transcriptome sequencing Reads and reference genome sequence (usually in BAM format), species reference genome sequence and the annotation file for visual browsing.

The data were downloaded from the BMK Cloud (https://international.biocloud.net). To detect valid alternative splicing events, those with a P < 0.05 and |△PSI| >10% were categorized as differential alternative splicing (diAS) events. For exons in alternative splicing, percent-spliced-in (PSI) was calculated as PSI = Splice in/(Splice in+ Splice out). “Splice-in” and “splice-out” represent the number of reads that corroborate the occurrence of splice-in and splice-out, respectively, in the RNA-seq data. All pairwise comparisons were assessed using the DESeq2 R package (1.6.3) to DEGs and DETs. In the DESeq2 analysis, differentially regulated genes were defined as those with a two-fold change, with an adjusted P < 0.05. Following that, the data that had been chosen were used for visualisation by being loaded into SRplot. Finally, the overlapping genes of DETs and diAS were imported into Venn 2.1.0 and histograms were generated with the ggplot2 R package (v3.3.6).

T47D cells RNA in each group was extracted following the protocol of Trizol reagent kit (Cat#15596018CN, Invitrogen, Carlsbad, CA, United States), The RNA content was assessed using an ultra-microspectrophotomete. Then, complementary DNA (cDNA) was obtained from the RNA through the process of reverse transcription, utilizing a HiScript II Q RT SuperMix (Cat#R222-01; Epizyme, shanghai, China). Afterwards, SYBR-Green method was used for real-time quantitative cDNA amplification (Cat#Q711-02; Epizyme, shanghai, China). Finally, the relative mRNA levels have been determined to use the 2^−ΔΔCT method and standardized against GAPDH. RT-PCR primers that are intended to amplify two or more isoforms of various sizes are illustrated. The primers sequences of the genes for qPCR/RT-PCR are shown in Table 1. The ImageJ software was used to quantify the PCR results.

The Western blot analysis was carried out in accordance with the previously reported techniques (Wang P et al., 2024; Wang PY et al., 2024). In brief, protein samples ranging from 20 to 40 μg were analysed using a 10%–12% (w/v) SDS-PAGE gel (Cat#PG212/PG213; Epizyme, shanghai, China). The proteins that had been isolated were subjected to electroblotting and subsequently deposited onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked using TBST solution that consisted of 5% nonfat milk or bovine serum albumin. Subsequently, the membranes were subjected to an overnight at 4°C with primary antibodies (xCT (Cat#26864-1-AP), GPX4 (Cat#30388-1-AP), HO-1 (Cat#10701-1-AP), NCAPD3 (Cat#16828-1-AP), 1:1000, ProteinTech Group, Chicago, United States). And its was subjected to incubation with secondary antibodies (Cat#L3032/L3012, 1:10,000, Signalway Antibody, Greenbelt, MD, United States) at ambient temperature. The visualisation of protein blots was achieved by the utilisation of an ECL system and the Image Lab detection system, manufactured by BioRad in Hercules, CA. GAPDH, β-actin, or α-tubulin (Cat#WL01114/WL01372/WL02296, 1:1000, Wanlei Biological Technology Co., Ltd., Shanghai, China) were employed to normalize the protein bands and examine them using ImageJ.

Statistical analysis was conducted using the GraphPad Prism (version 9.00). The experiments were repeated at least three times, and the data are expressed as the mean ± standard error of the mean (SEM). Statistical differences between two groups were analyzed using Student’s t-test, One-way ANOVA and Two-way ANOVA were used for comparison among multiple groups. P < 0.05 was considered to indicate statistical significance.

Alternative splicing is an important factor in the development of protein variety. The different types of alternative splicing include exon skipping (ES), alternative 3′splice site (A3SS), mutually exclusive exon (MEX), alternative 5′splice site (A5SS), and intron retention (IR) (Figure 1A). As illustrated in Figure 1B, the classification of alternative splicing events in T47D cells following THSG treatment (T47DT) was shown. We observed that over half alternative splicing events were ES events. Furthermore, GO and KEGG enrichment analysis were performed to explore the differential alternative splicing (diAS). As shown in Figure 1C, it shown that the BP was mainly concentrated on the mRNA metabolic process and CC was mainly located in the mitochondrial envelope and mitochondrial membrane. Meanwhile, the Spliceosome was the most valuable pathway among the top 10 KEGG pathways (Figure 1D). Taken together, it indicated that THSG may inhibit T47D cells proliferation by regulating mRNA metabolic process and spliceosome. Nevertheless, the specific transcripts by which THSG exerts its effects have not been identified.

The overlapping transcripts (40) between the differentially-expressed transcripts (DETs) and differential alternative splicing events (diAS) related genes were shown in Figure 2A. Moreover, volcano plot was revealed in downregulated transcripts for Tead2-202, Cenpx-202, Ncaph-201, Ncapd3-203, and Paqr4-201 or upregulated transcripts for Unp35-206, Pcdh1-201, Clk-1-201, Mdm4-203, and Ccnt2-204 in THSG-treated groups (Figure 2B). Then, GO enrichment of overlapped transcripts (top 10) was performed in Figure 2C, and the top 3 pathways were cell cycle, cell division, and M phase. Meanwhile, the ∆PSI as the indicator ranked the downregulated transcripts (∆PSI > 0) as shown in Figure 2D.

Furthermore, Reverse Transcription-Polymerase Chain Reaction (RT-PCR) was performed to validate the AS changes of main transcripts (Figure 3). The inclusion reads of Ncapd3 and Cenpx2 were increased in THSG treatment, with the same as the results of RNA-seq data (Figures 3B, C). However, the inclusion reads of Tead2 and Ncaph has no significant difference after THSG intervention (Figures 3A, D). Notably, the mRNA expression of Ncapd3-203 and Cenpx-201 were decreased in treatment of THSG (Figures 4B, D). Therefore, these results confirmed that the overlapping transcripts of DETs and diAS play an important roles in the cell cycle, and it was worthwhile to explore the transcripts of Ncapd3 and Cenpx in further study.

Given the critical role of protein-level analyses in understanding biological processes, we predicted the protein structure of NCAPD3 (NCAPD3-203) and CENPX (CENPX-202). Compared with NCAPD3, the altered region of NCAPD3-203 is located within the domain (Figures 4A, B), whereas the change of CENPX-202 is located within the non-structural domain (Figures 4C, D). Therefore, it will indicate that the alternative splicing event of Ncapd3 will be more valuable for research.

As shown in Figure 4E, Ncapd3 is highly expressed in breast cancer, and the protein levels of NCAPD3 and the mRNA expression of Ncapd3 were decreased in THSG group (Figure 4F). Collectively, the findings corroborated the elevated expression of NCAPD3 in T47D cells. And THSG could inhibit the protein (mRNA) expression level of NCAPD3 (Ncapd3).

To investigate the potential mechanism of THSG suppress T47D cells proliferation, differentially-expressed genes (DEGs) and differentially-expressed transcripts (DETs) were selected to analysis. Compared with the KEGG enrichment of upregulated genes (transcripts) in DEGs and DETs after THSG treatment of T47D cells (Figures 5A,B), ferroptosis was the common pathway and was in the first place ranked in rich factor. However, the P value of the ferroptosis pathway in the upregulated genes of DETs is smaller, which will indicate that the ferroptosis pathway is more likely to be enriched in DETs.

To further explore the specific genes in ferroptosis pathway, the ferroptosis marker genes were selected in FerrDb V1 (http://www.zhounan.org/ferrdb/legacy/) database. It was worth noting that 3 overlapping genes (Slc7a11, Gpx4, Hmox1) were detected in ferroptosis genes in KEGG and ferroptosis marker genes (Figure 5C). In addition, the protein level of xCT, GPx4, and HO-1 was decreased in THSG treatment (Figure 5D). Altogether, it was determined that THSG treatment triggers ferroptosis in T47D cells.

Given that abnormalities in the cell cycle are a principal cause of cancer development and progression. Consistent with the reports, cell cycle played an vital roles in enriched for downregulated genes of DEGs and DETs (Figures 6A, B). Further, we conducted flow cytometry analysis to investigate the specific mechanisms about cell cycle in T47D cell death. As shown in Figures 6C–F, cell cycle arrest was notably induced by THSG treatment at various doses and time points, especially in the G2/M phase. As the concentration of THSG increased for 24h, the percentage of cells in the G2/M phase was increased significantly, whereas the percentage of cells in the S and G0/G1 phases was decreased (Figures 6C, D). Furthermore, time is a crucial determinant of the cell cycle. With time, there was also a significant increase in the percentage of cells in the G2/M stage in 24h and 48 h (Figures 6E, F). Notably, when 200 μM THSG was used to treat T47D cells for 24 h, THSG had the most significant effect on the cell cycle arrest.

Collectively, the KEGG enrichment results of DEGs and DETs preliminarily showed that THSG could promote the occurrence of cell cycle arrest in T47D cells.

To explore the effects of THSG on T47D cells, we observed a dose-time independent increasing in the inhibitory rate of THSG. It is worth noting that the T47D cells is significantly inhibited when 200 μmol/L THSG is used (Figures 7A, B). Moreover, the clone formation experiment showed that 200 μmol/L THSG treatment has a significant inhibitory effect on the proliferation of T47D cells, and this concentration was used in subsequent experiments (Figures 7C, D).

We further investigated the ferroptosis-related proteins to determine the role of NCAPD3 in ferroptosis in T47D cells. Notably, the expression levels of xCT and GPx4 proteins were reduced after knockdown of NCAPD3 (Figures 7E, F). When THSG was administered, the survival rate of T47D cells was decreased significantly (Figure 7G). Therefore, it would indicated that THSG could accelerate ferroptosis of T47D cells and thereby inhibited cell proliferation.