Human breast cancer cell line T47D and MDA-MB-231 were purchased from American Type Culture Collection (ATCC). T47D cells were cultured in high glucose Dulbecco’s Modified Eagle’s Medium (Gibco, Waltham, MA, United States) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 0.1 mg/mL streptomycin and 1% non-essential amino acid (NEAA) solution (Gibco). And the cells were maintained in the incubator at 37°C and in the presence of 5% CO₂. MDA-MB-231 cells were cultured in L-15 medium (Gibco, Waltham, MA, United States) supplemented with 10% FBS, 100 U/mL penicillin, 0.1 mg/mL streptomycin. The cells were maintained in the incubator at 37°C.

In order to better analyze the correlation between PPA1 and breast cancer progression, we hope to choose the cell line with relatively lower expression level at basal status for overexpression assays and with relatively higher expression level for knockdown assays. Meanwhile, we would like to select the cell line with high malignancy and aggressively metastatic characteristics for overexpression, and study the role of PPA1 in different molecular subtypes of breast cancer. For above considerations, we choose T47D for knockdown and MDA-MB-231 for overexpression assays.

To stably knockdown PPA1 in breast cancer cells, T47D cells were infected with lentivirus carrying pLV-H1-shPPA1-puro or pLV-H1-shRNA-scramble-puro plasmid, then treated with puromycin to obtain the stable cell line with PPA1 silencing (shPPA1) and the shRNA control. The sequences of shRNAs were: shPPA1#1: GCTACTGTGGACTGGTTTA; shPPA1#2: GGAATCAGTTGCATGAATA; shRNA control: GCTACACTATCGAGCAATT.

For the stable overexpression of PPA1 in mammalian cells, the cDNA of PPA1 was inserted into the pLV-EF1α-MCS-IRES-Bsd plasmid. MDA-MB-231 cells were infected with lentivirus carrying the recombinant plasmid and empty plasmid which was used as the control. Cells were selected using blasticidin to generate stable cell lines with PPA1 overexpression and their control.

Cell lysates were prepared from T47D and MDA-MB-231 with RIPA buffer containing protease inhibitor cocktail, phosphatase inhibitor cocktails 2 and 3 (Sigma-Aldrich, St Louis, MO, United States). Proteins (20–40 μg) were loaded onto 10–15% Tris-Acrylamide gels and blotted with primary antibodies that included: anti-PPA1 (Cat. # HPA019878, Sigma, St. Louis MO), anti-β-actin (sc-47778, Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-E-cadherin, N-cadherin, Vimentin, ZO-1, Claudin-1, Slug, PI3K, p-PI3K (Tyr458), AKT, p-AKT (Ser473), p-GSK3β (Ser9), P38, p-P38 (Thr180/Tyr182), JNK, p-JNK (Thr183/Tyr185), (Cat. # 3195, Cat. # 13116, Cat. # 5741, Cat. # 8193, Cat. # 13255, Cat. # 9585, Cat. # 4292, Cat. # 4228, Cat. # 4691, Cat. # 4060, Cat. # 5558, Cat. # 8690, Cat. # 9216, Cat. # 9252, Cat. # 4668, Cell Signal Technology Inc., Danvers, MA, United States), anti-Ki67 (Cat. # ab16667, Abcam Inc., Cambridge, United Kingdom), and followed by incubation with horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Images were acquired using the Amersham Imager 600RGB detection system (GE Healthcare).

Total RNA was extracted from cells using the TRIZOL reagent and reverse transcription was performed using the M-MLV reverse transcriptase (Promega, Madison, MI). Real-time PCR was performed using the TransStart Green qPCR Super Mix Kit (TransGen Biotech, Beijing, China). The 2^{–ΔΔ Ct} method was used to obtain the relative fold changes. The primers for human GAPDH were CTCTGATTTGGTCGTATTGGG and TGGAAGATGGTGATGGGATT. The primers for human PPA1 were CGCTATGTTGCGAATTTGTTC and CCAGTATGTTTATCATTGTGCC.

Transwell chambers were pre-coated with Matrigel. The bottom chamber was filled with culture medium containing 10% FBS. Equal amounts of cells were plated in the upper chamber in serum-free medium. After 12–24 h, the invasive cells were fixed and stained with 0.5% crystal violet. Then, we used microscope to observe the invasive cells.

For each test, T47D or MDA-MB-231 cells were seeded in plates. When the cells reached full confluence, a “wound” was created in the middle of the culture plate and the concentration of the serum in culture medium was changed from 10 to 1% to avoid the influence of cell growth rate on wound healing. The wound healing process was recorded at 0, 12, and 24 h after wound creation. The wound healing rate was quantified as the distance between the wound recovered compared to that of the original wound.

Cell proliferation was measured using the Cell Counting Kit-8 (CCK-8). Equal amounts of cells were seeded in 96-well plates. At designated time points (0, 24, 48, 72, and 96 h), the CCK-8 reagent was added to each well and then incubated for 2 h. Finally, the absorbance was measured at 450 nm.

The tissue microarray used for the analysis of PPA1 expression in breast cancer were purchased from Biomax Inc. (Cat. # BR2086, Rockville, United States). Immunohistochemistry (IHC) was performed on tissue microarray using antibodies against PPA1 at 1:200 dilution. The expression level of PPA1 was evaluated according to the percentage of positive cells and its staining intensity in each tumor tissue. The percentage of positive cells was separated into < 10%, 10%–29%, 30%–49%, and ≥ 50% subgroups, which were scored as 1, 2, 3, and 4, respectively. The staining intensities were evaluated as negative, weak, moderate, and strong, which were scored as 1, 2, 3, and 4, respectively. The cell percentage score and staining intensity score were multiplied to obtain the IHC score.

Cells grown on glass slides were fixed in 4% paraformaldehyde and labeled with primary antibodies overnight at 4°C, followed by incubation with Alexa Fluor 488 goat anti-mouse IgG (Invitrogen) or Alexa Fluor 594 goat anti-rabbit IgG (Invitrogen) at RT for 1 h. The cells were then subjected to confocal imaging (Leica, Germany).

All in vivo mouse experiments were approved by the Ethics Committee of Xinxiang Medical University. Female nude mice at 5–6 weeks old were separated randomly into four groups (n = 5 for each group). Cells were subcutaneously inoculated into the fourth mammary fat pad of each mouse. Tumor volume was measured using calipers and calculated using the formula: length × width²/2. Fifty days after inoculation, tumor and lung tissues were collected and subjected to IHC or hematoxylin and eosin (H&E) staining. For the inhibitor treatment assay, 10 days after tumor cells injection, mice were treated with PI3K inhibitor (LY294002, 70 mg/kg) every 2 days, while dimethyl sulfoxide (DMSO) was used as the control.

Survival analyses were conducted using an online tool.¹ Patients with breast cancer (n = 1,070) were selected for the overall survival assay. The log-rank test was computed automatically.

Values were expressed as the mean ± SEM. Statistical significance was determined using the Student’s t-test. A value of p < 0.05 was considered statistically significant. ^∗ indicates significant difference with p < 0.05, ^∗∗ indicates significant difference with p < 0.01, ^∗∗∗ indicates significant difference with p < 0.001.

In order to verify the clinical significance of PPA1 in breast cancer, we first examined PPA1 expression in tissue microarray containing human normal/para-carcinoma breast tissues and breast tumors with different TNM stage and histological grade by IHC. The results indicated that PPA1 was highly expressed in tumor tissues compared to normal/para-carcinoma tissues (Figures 1A,B). In addition, since the pathological grade and clinical stage of tumor were closely correlate with tumor malignancy and progression, we investigated the associations between PPA1 expression and clinicopathological characteristics. We found that PPA1 expression was positively correlated with TNM stage and histological grade of human breast cancer (Figures 1C,D). Moreover, the increased expression of PPA1 was also observed in breast cancer cell lines compared to that in the normal breast cell line (Figures 1E,F). Furthermore, we analyzed the correlation between the expression of PPA1 and patient overall survival (OS), and found that high PPA1 expression was associated with poor clinical outcomes in patients with breast cancer (Figure 1G). Taken together, these results indicate that PPA1 may play a vital role in breast cancer development, and that it may be a diagnostic marker for tumor progression.

To elucidate the function of PPA1 in breast cancer, we first silencing PPA1 in T47D cells by stable expression of control or two PPA1-targeting shRNAs. Meanwhile, we established stable MDA-MB-231 cell lines that overexpressed PPA1.

Wound healing assays were performed to detect the role of PPA1 in migration. As shown in Figure 2A, although gap-filling was significantly retarded in T47D-shPPA1 cells, the control cells migrated and almost filled the gap at 24 h after wounding (Figures 2A,B), indicating that silencing PPA1 suppressed the migration properties of T47D cells. Moreover, the transwell assay also demonstrated that PPA1 knockdown reduced the cell invasion properties of T47D-shPPA1 cells compared to that in the control group (Figures 2C,D). The CCK-8 results revealed that silencing PPA1 inhibited cell proliferation (Figure 2E).

To strengthen our conclusions, we examined the effect of PPA1 overexpression in MDA-MB-231 cells. Consistently, wound healing and transwell assay confirmed that ectopic PPA1 facilitated breast cancer cell migration and invasion (Figures 2F–I). Furthermore, PPA1 overexpression increased cell proliferation (Figure 2J). Taken together, these findings verified that PPA1 promoted proliferation, migration, and invasion of breast cancer cells.

To tested the regulatory effect of PPA1 on EMT, we first examined the expression of EMT-related proteins by western blot. We observed that PPA1 deficiency contributed to an epithelial phenotype as the expression levels of E-cadherin, ZO-1, and Claudin-1 were elevated, and those of the mesenchymal marker proteins Vimentin and N-cadherin were diminished in PPA1-deficient cells compared to those of the control group (Figure 3A), while ectopic PPA1 contribute to mesenchymal characteristic by attenuating expression levels of E-cadherin, ZO-1 and Claudin-1, and augmenting levels of Vimentin and N-cadherin (Figure 3C), indicating the EMT-promoting effects of PPA1. Consistently, this was further supported by the immunofluorescence results, which revealed an elevation in the intensity of the E-cadherin and ZO-1 signals in T47D-shPPA1 cells, and a decrease in the intensity of these signals in MDA-MB-231-PPA1 cells. Meanwhile, the expression of N-cadherin and Vimentin was attenuated in T47D-shPPA1 cells, and enhanced in MDA-MB-231-PPA1 cells (Figures 3B,D).

In addition, several transcriptional factors, including Slug, Twist, and ZEB1, function as molecular switches in the EMT program. Especially Slug, a major EMT inducer, is closely associated with tumor metastasis. To investigate which transcriptional factors mediate the regulatory effects of PPA1 on EMT, we used western blotting and found that the protein expression levels of Slug decreased in two T47D-shPPA1 cells, but increased in MDA-MB-231-PPA1 cells (Figures 3A,C). Meanwhile, PPA1 knockdown or overexpression elicited no change in ZEB1, Twist, and Snail expression. These results verified that PPA1 played a vital role in triggering EMT and revealed the crucial function of Slug in mediating EMT promotion effect of PPA1.

To further investigate the mechanism underlying the regulatory effects of PPA1 on breast cancer progression and EMT, we first performed western blot and the results demonstrated that silencing PPA1 restrained the phosphorylation levels of PI3K, AKT, and GSK3β. In contrast, ectopic PPA1 augmented the expression of p-PI3K (Tyr458), p-AKT (Ser473), and p-GSK3β (Ser9) (Figures 4A,B). PPA1 also could induce more phosphorylated GSK3β accumulation in the cytoplasm and less in the nucleus (Supplementary Figures 1A,B). However, significantly changes in JNK, p-JNK (Thr183/Tyr185), P38, and p-P38 (Thr180/Tyr182) were not as evident (Figures 4A,B). Consistently, this was further supported by immunofluorescence assay results (Figures 4C,D). Collectively, these findings indicated that PPA1 facilitated breast cancer progression and EMT via activating PI3K/AKT/GSK3β signaling pathway.

To gain insights into the importance of PPA1 mediated signaling pathway in promoting the breast cancer progression and EMT, we investigated the effects of PI3K inhibitor (LY294002) in MDA-MB-231-PPA1 cells. We verified that the PI3K inhibitor reversed the proliferation, migration and invasion promoting effects of PPA1 in MDA-MB-231 cells (Figures 5A–E). Consistently, the treatment of the cells with PI3K inhibitor rescued expression levels of ZO-1, attenuated Slug activity, and blunted PPA1-induced PI3K/AKT/GSK3β signaling pathway activation (Figure 5F). Taken together, these results confirmed that inhibitors targeted PPA1 mediated signaling pathway suppressed breast cancer progression and EMT in vitro.

To evaluate the function of PPA1 in breast cancer progression in vivo, we performed xenograft experiments using MDA-MB-231 cells. To this end, stable MDA-MB-231-control and MDA-MB-231-PPA1 cells were injected into the fourth mammary fat pad of mice. We verified that ectopic PPA1 significantly promoted tumor growth, which could be reversed by the PI3K inhibitor (Figures 6A–C). In addition, we observed that more PPA1-upregulated cells metastasized to the lung and initiated secondary tumor, which was also rescued by the PI3K inhibitor (Figures 6D–F). Furthermore, immunohistochemical staining demonstrated that PPA1 upregulated boosted the expression of mesenchymal effect (N-cadherin, Vimentin) and reduced epithelial protein expression (E-cadherin). Meanwhile, the levels of Ki-67, Slug, p-PI3K (Tyr458), p-AKT (Ser473) and p-GSK3β (Ser9) were elevated in tumor tissues (Figures 6G,H). However, the PI3K inhibitor reversed the expression levels of EMT markers and modulated the activation of PI3K/AKT/GSK3β pathway (Figures 6G,H), demonstrating that inhibitors targeted PPA1 mediated signaling pathways suppressed breast cancer progression. Taken together, these results confirmed that PPA1 promoted tumor growth and metastasis via the PI3K/AKT/GSK3β signaling in vivo.

In summary, our study demonstrated a critical role of PPA1 in mediating PI3K/AKT/GSK3β signaling-induced tumor progression, which could be a useful target to prevent breast cancer.

Based on the totality of our findings, we propose the following model: PPA1 activates PI3K/AKT signaling, which enhances GSK3β phosphorylation, thereby maintains Slug stability and promotes Slug nuclear translocation. Slug then leads to the repression of E-cadherin expression, which promotes the EMT process and tumor metastasis (Figure 7).