Surgically treated female breast cancer patients (n = 180) with confirmed pathology of invasive ductal carcinoma were included, and their clinicopathological characteristics are summarized in Table 1. Breast cancer tissues were obtained from the patients when undergoing surgical treatment at the Department of Surgery, Cancer Hospital affiliated to Shantou University Medical College, during the period from 2010 to 2013. All patients received primary treatment by surgery followed by adjuvant radiotherapy, chemotherapy, or hormone therapy. The mean age of the patients was 50 years (range = 19–78 years). The histological types and grades of the primary tumors were determined according to a system modified from the WHO classification. Tumor staging was defined according to the 7th Edition of the American Joint Committee on Cancer TNM Staging manual. For all cases, representative parts of the primary invasive ductal carcinoma, along with adjacent noncancerous tissue were frozen in liquid nitrogen-cooled isopentane at the time of diagnosis and stored at a temperature below -80°C until use. The rest of the tissue was fixed in buffered formalin, embedded in paraffin, and used for routine histology. Clinical research protocols of this study were reviewed and approved by the Ethics Committee of Shantou University Medical College (IRB serial number: # 04–070). Written informed consents were obtained from patients in accordance with principles expressed in the Declaration of Helsinki.

Four-μm-thick sections from representative tumor areas and normal mammary gland tissue from surgical specimen tissue blocks were processed according to standard immunohistochemistry (IHC) protocols (Dong et al., 2017b; Wang et al., 2019; Wang et al., 2020; Wang et al., 2021a) and stained with antibodies against PTPRO (Cat. No. HPA034525, Sigma-Aldrich, St. Louis, United States), and Ki67 (Cat. No. #9449, Cell Signaling Technology, Danvers, MA, United States). Sections incubated with immunoglobulin (Ig)G of appropriate species (rabbit or mouse) as the primary antibody were used as negative controls.

The percentage of positively stained cells was scored using the following scales: 0, no staining of cells in any field; 1, ≤10%; 2, 11–50%; 3, 51–75%; 4, >75%. The intensity of staining was scored using the following scales: 1+, weak staining; 2+, moderate staining; 3+, strong staining. Percentage (P) and intensity (I) of nuclear or cytoplasm or membrane expression were multiplied to generate a numerical score (S = P*I). The immunoreactivity was evaluated by two different pathologists with no prior knowledge of patient data. When the opinions of the two evaluators were different, the agreement was then compared and discussed.

Raw data from 1,079 breast cancer gene expression profiles from 5 independent datasets (GSE3494, GSE7390, GSE6532, GSE1456, and GSE2034) were collected from NCBI Gene Expression Omnibus (GEO) (Pawitan et al., 2005; Wang et al., 2005; Sotiriou et al., 2006; Desmedt et al., 2007; Loi et al., 2007). Each dataset selected for this study was based on Affymetrix U133A, and all of the datasets contained clinical outcome data. The raw expression data were normalized by Robust multi-array average (RMA) (Irizarry et al., 2003) and compiled. The batch effect was removed by using Bayesian Factor Regression Modeling (BFRM) (Carvalho et al., 2008). Following this procedure, a normalized gene expression dataset compiling 1,079 breast cancer samples was generated.

Human breast cancer cell lines were obtained from American Type Culture Collection (ATCC, USA). SKBR3 and BT474 cells were cultured in DMEM/F12 (GIBCO/Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS). All cells were maintained at 37°C in an incubator containing 5% CO₂. Primary mouse embryonic fibroblasts (MEFs) were derived from Ptpro ^+/+ and Ptpro ^−/− mice. In brief, uteri firstly isolated from 13.5-days-pregnant mice, then washed uteri with phosphate-buffered saline (PBS) and removed the head and visceral tissues. After washed with fresh PBS, the remaining bodies were minced using a pair of scissors, transferred into a 0.25% trypsin/EDTA solution (1 mL per embryo) at 37°C for 20 min. The trypsin was inactivated by DMEM with 10% FBS. The cells were cultured for 24 h at 37°C. Adherent cells were used as MEF cells.

The full-length sequence encoding PTPRO was inserted into a pCR3.1 expression vector. Lentiviral pGIPZ shRNA vectors targeting human PTPRO (pGIPZ-shPTPRO, V2LHS_226,171) and non-targeting pGIPZ control vector (pGIPZ-shCtrl) were obtained from Open Biosystems (Huntsville, AL, United States). Transfection of plasmids was performed using Lipofectamine 3,000 (Cat. No. L3000015, Thermo Fisher Scientific) according to the manufacturer’s instructions. To overexpress mouse Erbb2 in Ptpro ^+/+ MEFs and Ptpro ^−/− MEFs, the plasmid FUGW-Erbb2 (a generous gift from Prof. Yi Li, Baylor College of Medicine) was transfected into the cells, which expresses mouse Erbb2 under control of a ubiquitin promoter, using the Calcium Phosphate Transfection Kit (Invitrogen) based on the manufacturer’s instructions.

The cell viability was examined by MTT assay. MTT and colony formation assays were performed as described previously (Feng et al., 2014; Gan et al., 2016). For MTT assay, cells were inoculated in a 96-well plate and incubated with 10-fold serial dilutions of lapatinib ranging from 0.001 to 10 μM for 72 h at 37°C. The optical density (OD) of each well was measured with a microplate reader set at 570 nm. Cell inhibition ratio (%) = [OD_control-OD_{lapatinib treatment}]/OD_control ×100%.

For colony formation assay, cells were trypsinized and 1,000 viable cells (depending on the experiment) were inoculated into six-well plates (in triplicate). Cells were allowed to adhere and colonize for 2 weeks. To visualize colonies, media was removed and cells were fixed in 100% methanol for 15 min and stained with crystal violet (0.1%, Sigma) for 20 min.

The cell death was confirmed by a photometric enzyme-immunoassay (Cat. No. 11774425001, cell death detection ELISA^PLUS kit, Roche Diagnostics, Mannheim, Germany) according to manufacturer’s protocol as previously described (Chandrasekar et al., 2008). In brief, the cells (1×10⁵ cells/mL) were placed in a streptavidin-coated microtiter plate with 10-fold serial dilutions of lapatinib ranging from 0.001 to 10 μM for 72 h at 37°C. After that, cells were incubated with a mixture of biotin-labeled monoclonal histone antibody and peroxidase-conjugated monoclonal DNA antibody at 37°C for 2 h. The amount of mono- and oligonucleosomes was measured as the absorbance at 405 nm with a microplate spectrophotometer (Bio-Rad, United States) after washing to remove unbound antibodies.

Total RNA was extracted from cells using TRIzol (Invitrogen). Four-μg RNA was reverse transcribed using M-MLV Reverse Transcriptase (Invitrogen) with oligo-(dT)20 primer (Invitrogen) according to the manufacturer’s instructions. Then equal amounts of cDNA were amplified using SYBR Green PCR amplification kit (Invitrogen) with the Applied Biosystems 7,500 Real-Time PCR system (Applied Biosystems, Foster City, CA, United States), as described previously (Zhang et al., 2013; Lin et al., 2019). The results were normalized to GAPDH expression. All reactions were run in triplicate. Western blot was performed as described previously (Du et al., 2019; Xiong et al., 2020; Wang et al., 2021b). Whole-cell extracts from breast cancer cell lines. The following antibodies were used: anti-PTPRO (Cat. No. 12161-1-AP; Proteintech Group Inc., Chicago, IL, United States); anti-β-actin (Cat. No. #4967, Cell Signaling Technology, Beverly, MA, United States). Specially, the membranes of PTPRO proteins were incubated with horseradish enzyme-labeled secondary antibody for 1 h at room temperature. An enhanced chemiluminescence (IBright FL1000, Life Technologies, United States) detection system was applied to detect protein blots.

All statistical analyses were performed using the SPSS 19.0 statistical software package (SPSS Inc., IL, United States). The differences in PTPRO expression immunoreactivity score were compared by the Mann-Whitney U test. The statistical significance of differences was evaluated with a t-test, and means, standard error, and 95% confidence intervals were calculated using GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA). Correlations between PTPRO expression and clinicopathological factors were investigated using the χ² test. The correlation coefficient between variables was performed using Spearman’s rank test. Survival curves were generated using the Kaplan-Meier estimates, and the statistical significance of differences between curves was evaluated by the log-rank test. Furthermore, hazard ratios (HR) and 95% confidence intervals (CI), which were computed from univariate and multivariable Cox proportional hazards regression models (approximate proportionality was verified by visual examination of the Kaplan–Meier estimates), were used to assess associations between relapse-free survival and clinicopathological characteristics. A p-value of less than 0.05 was considered to be significant, and all tests were 2-sided.

To determine if PTPRO plays any role in breast cancer, we first conducted reverse transcription-quantitative PCR (RT-qPCR) to estimate the mRNA levels of PTPRO in 37 breast tumor tissues and found that compared with their adjacent normal tissues, the mRNA level of PTPRO is significantly lower in 67.6% of cancer tissues (Figures 1A,B). We then conducted IHC analysis of PTPRO on a tissue microarray (TMA) containing 180 human breast cancer samples and 18 normal breast tissues. Figure 1C showed the representative photomicrographs with PTPRO IHC scores of 0, 9, and 12 along with normal human kidneys which expresses a high level of PTPRO as a positive reference (Dong et al., 2017a). Compared with that of the normal breast tissues, PTPRO IHC scores in breast cancer samples are significantly lower (p < 0.001) (Figure 1D). The association between the levels of PTPRO and the clinicopathological parameters in the cohort of 180 breast cancer specimens used in the TMA was also examined (Table 1). The levels of PTPRO in poorly and moderately differentiated cancer specimens were significantly lower than that of well-differentiated cancer (p = 0.031). In addition, the levels of PTPRO were consistently lower in larger (>2cm, p = 0.013), more invasive (T3+T4 vs. T1+T2, p = 0.003), and ERBB2-positive tumors (p = 0.027). Of note, the levels of PTPRO were inversely correlated with the levels of Ki67 (p < 0.001) and the representative cases shown in Figure 1E (Left panel). This finding is consistent with the relationship between PTPRO expression and the Ki67 index of the breast cancer specimens used in the TMA (r = - 0.559; p < 0.001) (Figure 1E, right panel). These data altogether suggest that reduced PTPRO expression is associated with a wide spectrum of clinicopathological parameters in breast cancer.

Since PTPRO expression is closely related to ERBB2 status (Dong et al., 2017a), we thus further investigated the relationship between PTPRO expression and clinical outcomes in ERBB2-stratified breast cancer patients in a large dataset. In patients with lower levels of PTPRO, the relapse-free survival is significantly shorter than those with higher levels of PTPRO (p = 0.007) (Figure 1F). Multivariate analyses suggest that PTPRO can serve as an independent prognostic factor for patients with high levels of ERBB2 (HR = 0.512, 95% CI = 0.324 to 0.807, p = 0.004) (Table 2).

To investigate whether the levels of PTPRO affect lapatinib sensitivity, we generated mouse MEFs derived from two genotypic mice, Ptpro ^+/+ and Ptpro ^−/− , and stably expressed ERBB2 by transfecting the cells with the plasmid FUGW-Erbb2. When the cells were treated with either 1 μM or 10 μM lapatinib, the survival rate of Erbb2/Ptpro ^−/− MEFs is significantly higher than that of Erbb2/Ptpro ^+/+ MEFs (p < 0.01; Figure 2A). Colony formation assay also showed that more colonies were formed by Erbb2/Ptpro ^−/− MEFs than Erbb2/Ptpro ^+/+ MEFs in the presence of 1 μM lapatinib (p < 0.01; Figure 2B). Cell death assay suggested that Erbb2/Ptpro ^−/− MEFs can attenuate apoptotic cell death than Erbb2/Ptpro ^+/+ MEFs in the presence of 0.1–10 μM lapatinib (p < 0.01; Figure 2C). Furthermore, depletion of PTPRO from ERBB2-positive breast cancer cells SKBR3 and BT474 also reduced their lapatinib sensitivity (Figures 2D–F). These findings altogether suggest the importance of PTPRO in lapatinib sensitivity. Next, we analyzed two public available datasets GSE38376 (Komurov et al., 2012) and GSE16179 (Liu et al., 2009) and found that the PTPRO mRNA level is lower in lapatinib-resistant breast cancer cells (SKBR3-lapR and BT474-lapR) than that of their parental cells (SKBR3-P and BT474-P) (Figure 3A). We then established two lapatinib-resistant ERBB2-positive breast cancer cell lines (SKBR3-lapR and BT474-lapR) in our lab following the published protocol (Bianco et al., 2008). As expected, the SKBR3-lapR and BT474-lapR cells showed their resistance to the lapatinib treatment (Figures 3B,C). More importantly, we found that the levels of PTPRO in SKBR3-lapR and BT474-lapR are significantly lower than that of their parental cells (Figure 3D). These findings demonstrated that downregulated PTPRO is correlated with lapatinib resistance.

We next assessed the impact of PTPRO on the lapatinib resistance in ERBB2-positive breast cancer cells. Two established lapatinib resistant ERBB2-positive breast cancer cells, SKBR3-lapR and BT474-lapR, were transfected with the plasmid overexpressing PTPRO. The mRNA and protein levels of PTPRO were then estimated by RT-qPCR and Western blot (Figures 4A,B respectively). As shown in Figure 4C, cells were treated with 10-fold serial dilutions of lapatinib from 0.001 to 10 μM for 72 h, and the inhibition ratio exhibited a dose-dependent increase in the dose range of 0.1–10 μM. Of note, both SKBR3-lapR and BT474-lapR cells become more sensitive to lapatinib when PTPRO is overexpressed (Figure 4C). Altogether, our data strongly suggest that upregulated PTPRO in ERBB2-positive breast cancer cells can enhance lapatinib sensitivity.