Human HER2+/HR+ breast cancer cell lines: BT474 was purchased from the American Type Culture Collection (ATCC, Manassas, VA, United States). EFM-192A was purchased from Shanghai Cell Bank (Shanghai, China). EFM-192A cells were maintained in DMEM medium supplemented with 10% FBS. BT474 was cultured in RPMI 1640 supplemented with 10% FBS.

Pyrotinib (SHR 1258) and SHR6390 were obtained from Jiangsu Hengrui Medicine (Jiangsu, China). Agents were dissolved in dimethylsulfoxide (DMSO) at a concentration of 20 mM and were then kept at −20°C until further use. Antibodies were purchased from Bioss (Beijing, China): GSK-3β, GSK-3β (Ser9), NF-κB, pNF-κB (Ser468), FOXM1, RB, pRb (Ser 780); pFOXM1 was purchased (Thr600) from Affinity Biosciences. Antibodies against GAPDH were purchased from Cell Signaling Technology (Beverly, MA, United States).

We determined the 50% inhibitory concentration (IC50 value) of different cell lines through MTT assay. The cells are seeded into a 96-well plate at a density of 3,000–10,000 cells/well. The next day, we treated the cells with pyrotinib, SHR6390, the combination of the two drugs, or DMSO; the treatment continued for 72 h. The microplate reader (BioTek, Norcross, GA, United States) was operated to automatically detect the optical density value (OD value) of each well. We took the average of the OD values (absorbance) of each group of eight wells and calculated the cell inhibition rate of each group. IR (inhibition rate) = (1-OD value of the experimental group/OD value of the negative control group) × 100% (Van Meerloo et al., 2011). GraphPad software (GraphPad Prism 5) was used to calculate the IC50 value, after the cell inhibition rate was obtained. CompuSyn (ComboSyn Inc.) was used to calculate the Combination Index (CI) value (Chou, 2018). The CI value can show the effect of synergistic (<1), additive (=1), or antagonistic (>1) effect of the two-drug combination.

The pyrotinib-resistant HER2+/HR+ breast cancer cell line was inducted by gradually increasing cytotoxic drugs. The EFM-192A cell line was inducted from 10% IC50 under the action of pyrotinib. The cells were initially seeded in a Petri dish so that the confluence was about 20%. After 24 h, pyrotinib was added to the culture medium in the universal container (initial concentration is 10% IC50 value), and then the culture medium was added to the cells in the Petri dish under standard aseptic conditions. When the cells were confluent, the cells can be subcultured in the usual way. We performed two subcultures at this concentration, and we increased the drug concentration if the cell viability was normal (note: freeze a batch of cells for each subculture). We stopped to induce drug resistance until the drug concentration significantly slowed down cell growth.

Genetic sequencing on drug-resistant cell lines and parental cell lines was performed. Statistical software R (version 3.6.1, https://www.r-project.org/) and packages of Bioconductor (http://www.bioconductor.org/) were applied for the significant analysis of differentially expressed genes (DEGs) between the two cell lines. The “limma” (linear models for microarray data) R package was used to screen the DEGs between parental EFM-192A cell line and pyrotinib-resistant EFM-192A cell line. The SAM (significance analysis of microarrays) with FDR (false discovery rate) < 0.01 and |log2 fold change (FC)| ≥ 2 were chosen as the cutoff criteria. KEGG pathway enrichment analysis of DEGs was performed to explore the critical pathways closely related to pyrotinib-resistant HER2+/HR+ breast cancer initiation. The “ggplot2” package and “pathview” package (version 1.24.0), which were based on Bioconductor, were used to make the KEGG pathway enrichment analysis visualization. The PPI network was used to identify the key genes which are involved in HER2+/HR+ breast cancer development at the interaction level. The screened DEGs were submitted to the STRING database (version 11). Then, Cytoscape software (version 3.7.1) was used for construction of the PPI network. MCODE, a plug in Cytoscape, was used for the visual analysis with “Combined score > 0.2 and MCODE > 11” as the threshold.

The experimental groups were added with pyrotinib (20% IC50, 50% IC50), SHR6390 (20% IC50, 50% IC50), and the combination of the two drugs (50% IC50 pyrotinib and 50% IC50 SHR6390), and the no-drug group was the blank control group. The six groups of breast cancer cell lines were cultured in six-well plates. When the cell fusion degree was 80%, the cells were scratched with a 200-μl sterile pipette tip, washed twice with PBS, and added with serum-free medium for culture. Then, we used a Nikon TS100 microscope (Nikon, Tokyo, Japan) and randomly selected in each well three fields of view (×100) in the scratched area to take pictures at 0 h. The medium containing the drug and 10% fetal bovine serum was added, and after 24 h, each well was photographed in the abovementioned field of view. Measure the width of the scratch at different times, compare the difference in scratch healing between the groups, perform the test at least three times, and take the average value (Justus et al., 2014).

A suspension of 1 × 10⁵ HER2+/HR+ breast cancer cell lines was added to Transwell chambers (Corning Inc., Corning, NY, United States) with 50 µl of Matrigel (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China). The experimental group was added with different concentrations of pyrotinib, SHR6390, and the combination of the two drugs, respectively. It was the vertical group without drugs. After 24 h of starvation, six groups of breast cancer cell lines were respectively added into the abovementioned 300-μl medium of different drug concentrations into the upper chamber of the Transwell chamber. The lower chamber is usually added with 500 μl medium containing 10% fetal bovine serum. After being cultured for 48 h, the cells in the upper chamber were gently scraped off with a sterilized cotton swab and stained with 0.1% crystal violet for 30 min. Then, the cells were taken out and rinsed in distilled water. The number of cells passing through the basement membrane of the ventricle was observed and counted under inverted microscopy (100× and 200× magnification) (Justus et al., 2014). The experiment was performed at least three times.

Different concentrations of pyrotinib, SHR6390, and the combination of the two drugs were put to the experimental group, and the no-drug group was the blank control group. After 24 h of starvation, cells were trypsinized in the logarithmic growth phase, washed with cold PBS, fixed with absolute ethanol, and stored 1 h at −20°C. Then, the fixed cells were collected, washed, and resuspended with cold PBS. Then, the cell solution was adjusted to 5 × 10⁵ cells/ml, soaked in the RNase and propidium iodide (PI, Sigma-Aldrich, St. Louis, MO, United States) incubation solution for 15 min in the dark at 37°C, and finally analyzed by the flow method according to the manufacturer’s instructions. All experiments were performed three times independently (Nunez, 2001). Analysis was done by flow cytometry using FlowJo 10 software (Ashland, OR, United States).

The FITC Annexin V/PI assay was used to detect apoptosis. The experimental group was added with different concentrations of pyrotinib, SHR6390, and the combination of the two drugs, respectively. After 48 h of treatment, the cells were washed with PBS, adjusted to 5 × 10⁵ cells/ml, resuspended in binding buffer, and stained with FITC Annexin V (5 µl) (Annexin V-FITC Apoptosis Detection Kit, NJ, United States) and propidium iodide (5 µl). Then, the stained cells were incubated in the dark at 37°C for 30 min. Finally, the flow cytometer (BD FACSCalibur, BD Biosciences, San Jose, CA, United States) was used to evaluate cell counts to determine apoptosis, according to the manufacturer’s instructions (Riccardi and Nicoletti, 2006; Crowley et al., 2016). All analyses were in triplicate, independently.

Cells were treated with different concentrations of drugs. Cells were lysed with cell lysis buffer, and PMSF, protease, and phosphatase inhibitors were added into the buffer. The proteins from the cells were separated by 10% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membrane by electrophoresis. Proteins were visualized using a Western blot imaging system (Bio-Rad, Hercules, CA, United States) and then quantified using ImageJ software (National Institutes of Health, Bethesda, MD, United States).

All experiments were conducted in accordance with the guidelines of the Regulations on the Management of Experimental Animals and were approved by the Experimental Animal Ethics Committee of Henan University of Science and Technology. Four-to five-week-old female BALB/c nu-mice were raised in the SPF laboratory breeding facility. The suspension of EFM-192A cells (1 × 10⁷ cells/100 μl) was injected into the skin of the left back of each mouse, 2 μl per mouse. The nude mice were fed routinely and observed every day. When the tumor reached 100–200 mm³, they were taken as the 0th day of medication. Mice were randomly assigned to four groups (seven in each group) and were treated by oral gavage with pyrotinib, SHR6390, and pyrotinib combined with SHR6390 vehicle (PBS) once a day. It was continuous for 25 days. During this period, tumors and weights of each mouse were measured twice a week. The tumor volume was calculated as V = 1/2 (length×width²). After the experiment was completed, the mice were executed humanitarianly. We collected tumor samples, recorded the tumor size, and calculated the multiplication rate. Samples of tumors were subjected to immunohistochemistry.

Tumors were fixed in 4% formalin at room temperature, and then they were embedded in paraffin. The embedded paraffin was sectioned. Sections were stained with p-GSK3β, pNFκB, pFOXM1, and pRB antibodies.

GraphPad Prism Version 7 software (GraphPad software, San Diego, CA, United States) and IBM SPSS version 22 (SPSS, Armonk, NY, United States) were used for statistical analysis. The half-maximal inhibitory concentration (IC50) values were calculated by nonlinear regression analysis of the dose–response curve. We used ComboSyn Inc. software to calculate the CI values. The results were expressed as the means ± SD (standard deviation) of three independent tests. The analysis of variance model and Student’s t test were used to compare the differences between test groups. The statistical significance of the difference between the test sample and the control sample was evaluated when the significance threshold was *p < 0.05, **p < 0.01, ***p < 0.001.

Firstly, the inhibitory ability of SHR6390 and pyrotinib on EFM-192A and BT474 HER2+/HR+ breast cancer cell lines was evaluated. The results of the study showed that both SHR6390 and pyrotinib inhibited the tumor proliferation of HER2+/HR+ cell lines (pyrotinib IC50 0.8 ± 0.1 μg/ml in EFM-192A, 4.6 ± 0.2 μg/ml in BT474; SHR6390 IC50 3 ± 0.2 μg/ml in EFM-192A, 6.4 ± 0.2 μg/ml in BT474). Next, we determined whether the combination of SHR6390 and pyrotinib had a synergistic effect on HER2+/HR+ breast cancer cells. We evaluated the effect of different concentrations of pyrotinib and SHR6390 in inhibiting the proliferation of EFM-192A and BT474 cells (Figures 1A–E,G–K). The extension of the median effect equation (MEE) proves the combination index equation (CIE) (CI < 1) that defines the synergy, additive effect (CI = 1), and antagonism (CI > 1), and the data can be automatically simulated by the CompuSyn software, after it is entered into the software (Chou and Talalay, 1984). In order to determine whether the antitumor effects were synergistic, CompuSyn was used to calculate the CI value. The results showed that pyrotinib combined with SHR6390 significantly inhibited the growth of HER2+/HR+ breast cancer cell lines. In both cell lines, the CI value was less than 1, indicating that the two drugs had a synergistic effect (Figures 1F,L).

In order to explore the molecular mechanism of the combination SHR6390 and pyrotinib in synergistically inhibiting breast cancer cells, we induced EFM-192A cells with a low dose of pyrotinib. After 8 months of continuous induction, we obtained pyrotinib-resistant EFM-192A cell lines with 10 times the IC50 value. Then, next-generation RNA sequencing for the parental EFM-192A cells and the pyrotinib-resistant EFM-192A cells was performed.

After removing batch differences and data normalization, a total of 1,579 DEGs were obtained, including 819 upregulated genes and 760 downregulated genes based on cutoff criteria (p < 0.01 and |logFC| ≥ 2). All abnormally expressed genes with the log2 FC score and −log10 p value were used to generate a Volcano diagram in R, which is an intuitive tool to show the overall gene expression level of the DEGs (Figure 2A).

GO term enrichment analysis results varied from GO classification and expression change of DEGs. These significantly enriched pathways and terms could help us to further understand the role of DEGs in pyrotinib-resistant breast cancer occurrence and progress (Figure 2B). In addition, KEGG pathway enrichment analysis was used to examine the function of the DEGs. The result of KEGG analysis showed that the upregulated genes were significantly enriched in the PI3K-Akt signaling pathway (Figure 2C).

Using the STRING online database, 819 upregulated genes were filtered into the DEG PPI network complex containing nodes and edges with a combined score of >0.2. The Cytoscape software was employed to analyze the interactive relationship between the candidate proteins, and then a cluster containing 12 nodes was selected with a cutoff k-score = 11 by the MCODE scoring system. The results showed that the hub gene was the intersections between the initial key genes. FOXM1 enriched was associated with inducted resistance to anti-HER2 therapies (Figure 2D).

The migration of HR+/HER2+ breast cancer cells decreased at 24 h in the pyrotinib or SHR6390 groups compared with the control group, and the migration rate of pyrotinib combined with SHR6390 was significantly reduced after 24 h. These results were consistent in both EFM-192A and BT474 cell lines. BT474 was a cell line with slow migration, and the control group could not close the wound after 24 h. A significant reduction in migration was shown after treatment with different concentrations of pyrotinib or SHR6390, while almost no migration was observed in the combination group. EFM-192A, a highly migratory cell line, almost completely closed the wound in the control group after 24 h. Different concentrations of pyrotinib or SHR6390 induced reduced wound healing at 24 h compared with untreated cells but were significant in the combination group (Figures 3A–D).

Quantitative analysis showed that the 10% FBS chemical attraction of cells in the vertical group resulted in the highest cell invasion rate. To evaluate the effect of pyrotinib or SHR6390 treatment, we compared the HER2+/HR+ breast cancer cell lines treated with different concentrations of SHR6390 and pyrotinib with the vertical group. We observed that the invasion rate decreased when cells were treated with pyrotinib or SHR6390; the higher the concentration, the less the cell migration. In the combination group of SHR6390 and pyrotinib, cell invasion decreased significantly. Similar results could be gotten in either the EFM-192A cell line or the BT474 cell line. All these results indicated that FBS chemotactic agents increased cell invasion ability, but this effect could be reduced by pyrotinib or SHR6390 treatment. In the combination of the two drugs, the invasion ability was significantly reduced. Therefore, pyrotinib combined with SHR6390 can weaken the invasion ability of HER2+/HR+ breast cancer cells (Figures 3E–H).

The synergistic effect of SHR6390 and pyrotinib combined inhibition on cell viability could be mediated by changes in proliferation, apoptosis, or both. To explore this, the effect of SHR6390 and pyrotinib or combination therapy on cell-cycle progression was evaluated. It was found that in HER2+/HR+ breast cancer, both SHR6390 and pyrotinib could induce G1/S arrest. The combination of the two drugs could produce the greatest reduction in the percentage of G2 phase cells. The same conclusion was observed in both EFM-192A and BT474 cell lines (Figures 4A,B,E,F).

We also sought evidence of increased tumor cell apoptosis after combination therapy. For the two cell lines, we treated the cells with different concentrations of drugs (pyrotinib: 20% IC50, 50% IC50; SHR6390: 20% IC50, 50% IC50). It was found that both SHR6390 and pyrotinib could increase cell apoptosis. The greater the drug concentration we treated, the higher the apoptosis rate. Then, the combination of SHR6390 and pyrotinib was used to treat the cells, and the apoptosis rate of the two cell lines was further increased (Figures 4C,D,G,H).

Therefore, the enhanced inhibition of cell viability observed when combining SHR6390 and pyrotinib (compared to either drug alone) appears to be mediated by both increased suppression of cell proliferation and increased apoptosis.

Western blot was used to assess the protein content of cell lines treated with SHR6390, pyrotinib, or the combination of the two drugs. We found that both SHR6390 and pyrotinib could reduce the phosphorylation of FOXM1, and the two drugs combined with FOXM1 significantly reduce the phosphorylation. It showed that the cyclin D‐CDK4/6 complex phosphorylates retinoblastoma proteins (RB) and dissociates them from the E2F transcription factors, which were ultimately responsible for cell-cycle progression (Pandey et al., 2019). Besides, both SHR6390 and pyrotinib could inhibit the phosphorylation of RB. The phosphorylation of RB dropped sharply after the combined treatment.

In the PI3K-AKT pathway (Figure 2C), we found that GSK3 was related to CCND1 and the cell cycle. Studies showed that treatment with CDK4/6 inhibitors (such as palbociclib) can prevent RB phosphorylation and preserve the activity of NF-κB transcription (Kim et al., 2019). After treatment with SHR6390 and pyrotinib, through the detection of p-GSKβ protein expression, we found that pyrotinib significantly reduced the phosphorylation of GSKβ, but SHR6390 had no significant effect on it. The same conclusion can be obtained when detecting the phosphorylation of NFκB. In summary, we speculated that SHR6390-combined pyrotinib inhibited the proliferation of HER2+/HR+ breast cancer through FOXM1 (Figures 5A–D).

For evaluating the effect of pyrotinib and SHR6390 in HER2+/HR+ breast cancer in vivo, we tested whether these two drugs alone or in combination could inhibit tumor growth in a xenograft mouse model. The mice were transplanted with BT474, and they were randomly divided into four groups. These groups include the vehicle group, pyrotinib (10 mg/kg/day) (Jianing et al., 2021), SHR6390 (75 mg/kg/day) (Long et al., 2019), or the simultaneous administration of pyrotinib and SHR6390. We found that pyrotinib was more effective than SHR6390 in inhibiting the growth of BT474 tumors. However, compared with the two drugs alone, the combined use of pyrotinib and SHR6390 has a much stronger inhibitory effect on the growth of BT474 xenografts (Figures 5E,F). The weight of the combined treatment group was better than that of the single-drug group (Figures 5G,H). These data indicated that the combination therapy including SHR6390 and pyrotinib could enhance anticancer activity.

Immunohistochemical results showed that the expression levels of pFOXM1, pGSK-3β, pNF-κB, and pRB were low in both the control group and the experimental group. This result is inconsistent with that at the cellular level (Supplementary Figure S1). In our analysis, it might be due to the inactivation of phosphorylated protein during tumor tissue embedding and paraffin section processing, so no positive results were obtained. We need to further investigate the effect of FOXM1 on HER2+/HR+ breast cancer at the tissue level.