MCF-7 was cultured in the MEM supplemented with 10% FBS and antibiotics. ZR-75-1 cells were cultured in the DMEM supplemented with 10% FBS and antibiotics. MDA-MB-231 cells were cultured in the RPMI-1640 medium supplemented with 10% FBS and antibiotics. 293T was cultured in the DMEM supplemented with 10% FBS and antibiotics. All cells were purchased from the Chinese Academy of Sciences Type Culture Collection. Cells were maintained at 37°C in a humidified atmosphere with 5% CO₂.

Human hormone receptor (HR)⁺HER2^- breast cancer tissue samples were obtained from Taizhou Central Hospital with the consent of all patients. The tissue samples were fixed in formalin and embedded in paraffin. Each slide containing both healthy adjacent breast tissue and tumor breast tissue samples was used for immunohistochemistry. The present study was approved by the Medical Ethics Committee of Taizhou Central Hospital (approval no. F-YXLL-004).

Tissue sections were rehydrated using xylene and graded concentrations of ethanol, incubated in sodium citrate (10 mM; pH 6.0) for 10 min, and then cooled down to room temperature. Endogenous peroxidase activity was then blocked with 3% hydrogen peroxide for 30 min at room temperature. The sections were permeabilized with 0.1% Triton and blocked in 10% goat serum for 30 min. The following primary antibodies were used: anti-TM4SF1 (1:400, Abcam), anti–Ki-67 (1:200, Cell Signaling Technology, Inc.), and anti-FLIP (1:400, Cell Signaling Technology, Inc.). All primary antibodies were diluted in 5% goat serum and added to the sections at 4°C overnight. The sections were then incubated with a biotinylated goat anti-rabbit antibody (1:200; BD Pharmingen; BD Biosciences). For detection, streptavidin–horseradish peroxidase and a DAB substrate kit (BD Pharmingen; BD Biosciences) were used. Counterstaining was carried out using hematoxylin.

The pLenti-EF1a-EGFP-P2A-Puro-CMV-TM4SF13-Flag plasmid was designed and synthesized by OBiO Biotechnology Corp., Ltd. The GenBank accession number for the TM4SF1 gene is NM_014220.3. The pMDG2.G and psPAX2 plasmids for the lentivirus assembly were obtained from Addgene, Inc. The 293T cells stably expressing the TM4SF1 plasmid and a control vector were prepared by lentiviral transduction. Briefly, co-transfection was performed by combining the lentiviral plasmid (2.5 μg) with the packaging plasmids (0.75 μg pMD2. and 1.90 μg psPAX2). The 293T cells were then infected with viral supernatants containing 8 μg/ml polybrene (MedChemExpress). Transduced cells were selected using 1 μg/ml puromycin.

Small interfering RNA (siRNA) transfection. Synthetic siRNAs targeting TM4SF1 were purchased from Shanghai GenePharma Co., Ltd. The sequences were as follows: siRNA-1, 5′-GCA​CGA​TGC​ATC​GGA​CAT​TCT-3′; siRNA-2, 5′- GCT​ATG​GGA​AGT​GTG​CAC​GAT-3′; and control-siRNA, 5′-UUC​UCC​GAA​CGU​GUC​ACG​UTT-3′. siRNA was transfected using Lipofectamine®-RNAiMAX (Invitrogen; Thermo Fisher Scientific, Inc.).

Total RNA was extracted using Trizol^® and then reverse-transcribed with the Prime Script™ RT reagent kit (Takara Biotechnology Co., Ltd.). The cDNA templates were amplified by qPCR using the PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific, Inc.). The primer sequences were as follows: TM4SF1 forward, 5′-TGC​AGG​ATC​TGG​CTA​CTG​TG-3′ and reverse, 5′-CAG​AAG​GTA​CTG​GCC​CTC​AG-3′; and GAPDH forward, 5′-GCA​CCG​TCA​AGG​CTG​AGA​AC-3′ and reverse, 5′-GCC​TTC​TCC​ATG​GTG​GTG​AA-3′. The thermocycling conditions were as follows: 1) 50°C for 120 s, 2) 95°C for 120 s, 3) 40 cycles of 95°C for 15 s and 60°C for 60 s, 4) 95°C for 15 s, 5) 60°C for 60 s, and 6) 95°C for 15 s. Gene expression levels were calculated using the ΔCq method. The mRNA levels of TM4SF1 were normalized to those of GAPDH mRNA.

Immunoblotting was conducted using standard procedures. Briefly, the cells were resuspended in the RIPA lysis buffer (Beyotime Institute of Biotechnology). The sample was transferred to a 1.5-ml centrifuge tube and then centrifuged at 10,000 x g for 5 min at 4°C to pellet the cell and tissue debris and collect the protein lysates. Equal masses of protein (50 μg) were analyzed by SDS-PAGE. Antibodies against TM4SF1 (#ab113504, Polyclonal, Abcam, Cambridge, United States), AKT (#9272, Polyclonal), phosphorylated AKT (#9275, Polyclonal), and β-actin (#5125s, Polyclonal) (Cell Signaling Technology, Boston, United States) were obtained. Antibodies against Bax (BA0315), Bcl-2 (BA0412), caspase 3 (BM4340), caspase 9 (BM4619), and FLIP (M01295, Monoclonal, 7F10) were obtained from Boster Biological Technology (Wuhan, China). Antibodies against GAPDH (0411, Polyclonal), LC3 (G-4, sc-398822), and Beclin-1 (E-8, sc-48341) from Santa Cruz Biotechnology (Santa Cruz, CA) were used. All primary antibodies were polyclonal. Proteins were visualized using a peroxidase-conjugated secondary antibody (Sigma-Aldrich; Merck KGaA) using ECL solution (Thermo Fisher Scientific, Inc.) for detection.

Cell proliferation was evaluated using MTT assays. Briefly, 1 × 10³ cells were plated in a 96-well plate using a complete culture medium. After 4, 24, 48, or 72 h, MTT was added to each well at a final concentration of 0.5 mg/ml. The reaction was allowed to proceed for 4 h at 37°C. The formazan crystals were then dissolved in 100 μL DMSO. The plates were incubated in an orbital shaker at room temperature for 15 min, and the absorbance was then read at 490 nm. All assays were performed in triplicates.

Briefly (Chen et al., 2020), a total of 1 × 10³ cells were gently mixed with 30 μL Matrigel and plated as a droplet containing the single cells onto a prewarmed (in a 37°C incubator) 24-well ultra-low attachment plate. The droplets were allowed to solidify at 37°C for 15 min, then covered with 500 μL complete cell culture medium, and incubated at 37°C for 15 days 3D organoid images were examined under a microscope.

For colony formation assays, 1 × 10³ cells were seeded into 6-well plates and incubated for 10 days at 37°C with 5% CO₂. The colonies were fixed in 100% methanol, then stained for 30 min in crystal violet, and washed in ddH₂O. The clones were counted under a microscope.

Cell apoptosis was analyzed by flow cytometry. Following different treatments, cells were trypsinized and washed with PBS twice. The cells were then stained with Annexin V-FITC/propidium iodide (Beyotime Institute of Biotechnology) according to the manufacturer’s instructions. Apoptosis was analyzed by MultiCycle software.

All animal experiments were approved by the Medical Ethics Committee of the Taizhou University College of Medicine (approval no. TZYXY 2020). A total of 24 male BALB/C-nu/nu mice (age, 6-8 weeks; average weight, ∼25 g) were obtained from Changzhou Cavens Animal Co., Ltd. The mice were housed in sterile cages under laminar airflow hoods at 20°C in a specific pathogen-free environment with a 12-h light/dark cycle and provided with autoclaved chow and water ad libitum. A total of 10⁷ parallel-controlled or TM4SF1-overexpressing MCF-7 cells were injected subcutaneously into the flank of the mice, respectively. Tumor dimensions were measured with calipers twice a week, and the volumes were calculated as (width)² x length/2. After 12 weeks, the animals were sacrificed by cervical dissociation, and the tumors were removed and weighed. The largest tumor diameter was 1.434 cm, and none of the animals developed multiple tumors.

The data are presented as the mean ± standard deviation and were analyzed using Student’s unpaired t-test or one way ANOVA by GraphPad Prism software (GraphPad Software). A p-value < 0.05 was considered significant.

The expression of TM4SF1 was first examined in breast cancer (n = 1,097) and normal breast (n = 114) TCGA data. TM4SF1 mRNA was expressed at lower levels in breast cancer than in normal tissue samples (Figure 1A). Moreover, TM4SF1 exhibited a lower expression along with the tumor stage (Figure 1B) and lower in lymph node–positive tissue specimens than that in lymph node–negative tissue specimens (Figure 1C). In addition, the expression levels of TM4SF1 were also analyzed in TNBC subclasses. Basal-like one and mesenchymal TNBC tissue expressed higher levels of TM4SF than the normal breast tissue. However, other types of breast cancer, including luminal androgen receptor–positive TNBC, expressed lower levels of TM4SF (Figure 1D). These findings indicated that the expression profiles of TM4SF1 in breast cancer varied with breast cancer subtypes.

The expression of TM4SF1 was subsequently examined in HR⁺HER2^- patients with breast cancer by immunohistochemistry (n = 7; Figure 2A). TM4SF1 staining was semi-quantified by measuring the average optical density by ImageJ software. The results demonstrated that TM4SF1 protein levels were downregulated in the HR⁺HER2^- breast cancer tissue compared with the healthy adjacent breast tissue (Figure 2B) Moreover, total RNA was extracted from those above 7 samples, and the mRNA levels of TM4SF1 were determined using RT-qPCR. The TM4SF1 mRNA expression level was downregulated in the HR⁺HER2^- breast cancer tissue compared with that of the healthy adjacent breast tissue (Figure 2C).

The next experiments were designed to examine whether TM4SF1 acted as a tumor suppressor in HR⁺HER2^- breast cancer. TM4SF1-overexpressing cells were generated using the pLenti-CMV TM4SF1 lentiviral transduction system in the HR⁺HER2^- MCF-7 and ZR-75-1 breast cancer cell lines. As shown in Figures 3A,B, TM4SF1 mRNA and protein levels were markedly increased in MCF-7 and ZR-75-1 cells transfected with pLenti-CMV TM4SF1. Moreover, TM4SF1-overexpressing cells displayed reduced viability in an MTT assay compared with pLenti-CMV vector transfected cells (Figures 3C,D).

As the repopulation of residual tumor cells can lead to cancer recurrence, which depends on the ability of cells to divide (Matsuda et al., 2021), the effect of TM4SF1 on clonogenicity was evaluated using a colony formation assay. As shown in Figure 4A, the TM4SF1 overexpression significantly decreased the plating efficiency (colonies number from 1000 cells) after 10 days of incubation compared with the pLenti-CMV vector group (32 vs. 26% in MCF-7 cells; 27.5 vs. 19.5% in ZR-75-1 cells).

3D organoid models allow the understanding of complex biology in cancer development, whereas 2D models have not proven as successful. When performing 3D cell culture experiments, the cell environment can be manipulated to mimic that of a cell in vivo and provide more accurate data about cell-to-cell interactions, tumor characteristics, drug discovery, metabolic profiling, stem cell research, and other types of diseases (Jensen and Teng, 2020). Thus, the effect of TM4SF1 on breast cancer cell 3D organoid formation was then evaluated. MCF-7 and ZR-75-1 cells transfected with the pLenti-CMV vector could form ∼30 and ∼18 3D organoid structures per 1,000 cells, respectively. By contrast, TM4SF1-overexpressing MCF-7 and ZR-75-1 cells formed only ∼20 and 10 3D organoids per 1,000 cells, respectively (Figure 4B).

We further downregulated the TM4SF1 expression level by siRNA transfection in MCF-7 cells and tested whether disruption of TM4SF1 in HR+/HER2-breast cancer cells affects the cell proliferation. As shown in Figures 5A,B, TM4SF1 mRNA and protein levels decreased obviously in MCF-7 after being transfected with TM4SF1-siRNA. Consequently, cells transfected with TM4SF1-siRNA grow slower than control cells in the MTT assay (Figure 5C). However, TM4SF1-siRNA transfection results in the opposite phenomenon in the triple-negative breast cancer (ER-/PH-/HER2-) cell line MDA-MB-231 (Figures 5D–F).

Because the TM4SF1 expression was downregulated in HR⁺HER2^- breast cancer and the TM4SF1-overexpression inhibited MCF-7 and ZR-75-1 cell viability and proliferation in vitro, the roles of TM4SF1 in the HR⁺HER2^- breast tumor growth were examined in a murine model. As shown in Figure 6A, TM4SF1-overexpressing MCF-7 xenografts grew slower than those in the pLenti-CMV vector control group. Immunohistochemical staining data showed that the frequency of Ki-67⁺ cells was markedly reduced following the TM4SF1 overexpression (Figure 6B), indicating reduced proliferation. Furthermore, TUNEL immunofluorescence staining demonstrated that the TM4SF1 overexpression induced apoptosis in vivo (Figure 6C).

In addition, as shown in Figure 6D, the TM4SF1 overexpression downregulated the levels of phosphorylated AKT and phosphorylated p70s6k1, which indicated that TM4SF1 may inhibit the AKT/mTOR pathway. Consistent with TUNEL staining, cleaved caspase 3/9 and Bax protein levels were upregulated, while those of BCL-2 and FLIP were downregulated following the TM4SF1 overexpression. These results indicated that the TM4SF1 overexpression promoted apoptosis.