Fig 2: Chemotherapy upregulated periostin expression in chemoresistant triple-negative breast cancer xenografts and human TNBC patient samples. (A) Tumor growth curve in response to paclitaxel (PTX), doxorubicin (DXR) and cyclophosphamide (CPA) in MCF10DCIS xenografts. Arrows indicate times of drug administration. Data are presented as mean ± standard error (n = 6). (B) Quantitative PCR analysis of indicated genes in tumors harvested at the end of the experiment in A. Relative values are presented as mean ± standard error (n = 3). *P < 0.05 versus vehicle-treated control, Student’s t test. (C) Correlation analysis of POSTN expression in 59 human breast cancer cell lines. Heatmap of gene expression and Pearson product-moment correlation coefficient (r value) are shown. (D) Patient survival analysis in basal-like breast cancer patients (n = 580) according to POSTN expression (low: 66.6%, high: 33.3%). Hazard ratio (HR), 95% confidence intervals. P value, log rank test. (E) Immunohistochemistry staining of periostin with human TNBC samples. Scale bar, 100 µm. (F) Periostin score before and after chemotherapy in 26 matched pair TNBC samples which showed residual disease after chemotherapy. ***P < 0.001, Wilcoxon signed-rank test.

Fig 3: Comparison of gene expression between PDL tissue and PDLFs by quantitative PCR in four PDL marker genes (OPN, ASPN, POSTN, and SPARC). PCR, polymerase chain reaction; PDL, periodontal ligament; PDLF, PDL fibroblast

Fig 4: An antibody targeting cancer-specific periostin variants inhibited mesenchymal cell expansion and tumor invasion after chemotherapy and halted tumor progression. (A) Diagram of splicing variants of human and mouse POSTN. Arrows indicate the location of designed primers. (B) Quantitative PCR analysis of POSTN splicing variants in tumor with or without PTX treatment. The region over exon 20 and 21 (hPNex20-21) was also amplified. Relative values against human total POSTN expression in vehicle-treated tumor are presented. Relative values are presented as mean ± standard error, *P < 0.02, Student’s t test. (C) Immunofluorescence staining of periostin variants containing exon 21 in tumor upon PTX treatment and periostin knockdown. Scale bar, 100 µm. (D) Tumor growth curve in MCF10DCIS xenografts treated with PTX and/or PN21-Ab. Upper arrows indicate times of PTX administration, and lower arrows indicate times of PN21-Ab administration. Tumor volumes were measured on the indicated days and are presented as mean ± standard error (n = 10). **P < 0.01, Student’s t test. (E) Flow cytometry analysis of tumors at Day 24 using CD44 and CD24 antibodies. Samples were prepared from a mixture of 10 tumors. Ratio of CD44high/CD24low to CD24high tumor cells is presented on the plots. (F) The ratio of CD44high/Ki67+ cancer cells against total Ki67+ cells was quantified in tumors collected at Day 24 (n = 15). Bars represent mean, and error bars represent standard deviation. ****P < 0.0001, Mann-Whitney’s test. (G) Immunofluorescence staining of CD44 and Ki67 in tumors collected at Day 24. Scale bar, 100 µm. (H) Immunofluorescence staining of laminin-5 and periostin in tumors collected at Day 24. Scale bar, 100 µm.

Fig 5: SPOCK1 was a TGF-ß-induced myoepithelial marker and SPOCK1 overexpression enhanced invasiveness.(A) Immunohistochemistry demonstrated that SPOCK1 was the only myoepithelial marker among the evaluated TGF-ß–upregulated gene products in MCF10A cells. CXCR7 staining was observed in luminal epithelia but not myoepithelia whereas FBN1 staining was observed diffusely in the stroma but not in the epithelia. IGFL2, PRRX2, PLAT, PCDH9, POSTN and NOV staining was not found in the ductolobular units (left and middle panels). By contrast, SPOCK1 was immunolocalized within (arrow) or beneath (arrowhead) the myoepithelia (right panel). (Magnification × 400). (B) Upregulation of SPOCK1 at mRNA levels (upper panels) and protein levels (lower panels) were observed in MCF10A, M10, and MCF12A cells, 3 days after treatment with TGF-ß. S26 was used as an mRNA loading control and GAPDH was used as a protein loading control. (C) Western blotting was used to detect the expression of SPOCK1 in MCF10A, T47D, and MB231 cells 24 h after transient transfection with pCDH-SPOCK1 or control vector pCDH. (D) The invasive capability and proliferation were measured in the cells shown in (C). Data from invasion assay are shown as the mean ± SD of 3 fields. Data from MTT assay are shown as mean ± SD of 3 independent experiments. These results are presented as the percentage relative to their control cells (*, P < 0.05; N.S., nonsignificant).