Stromal POSTN Enhances Motility of Both Cancer and Stromal Cells and Predicts Poor Survival in Colorectal Cancer

Figure 1. Expression and survival significance of periostin (POSTN) in colorectal cancer (CRC). (a) POSTN was weakly expressed on mesenchymal cells of non-neoplastic colonic mucosa. Note that non-neoplastic epithelial cells were negative for POSTN. Cancer-associated fibroblasts, but not CRC cells, expressed POSTN. ITGA6 and ITGB4, the canonical receptors for POSTN, were expressed in CRC cells. Bar, 100 µm. (b) ROC curves for POSTN expression on the patient survival at 5 years. (c) Kaplan–Meier curves for patients classified by stromal POSTN expression.

Figure 2. Correlation between stromal POSTN expression and cellular proliferation markers in CRC cells. Cellular proliferation marker expression in CRC cells classified by stromal POSTN expression. (a) PHH3, (b) CCNA, (c) GMNN, and (d) Ki-67. POSTN-positive tumors contained a significantly lower number of PHH3-, CCNA-, and GMNN-positive cells than negative tumors. The circles indicate outliers.

Figure 3. Association between gene mutation status and stromal POSTN expression. (a) Association between KRAS/BRAF mutation and stromal POSTN expression. (b) Associations between p53 expression patterns and POSTN expression. CRCs showing p53 complete loss expressed significantly higher levels of stromal POSTN than p53 wild-type tumors. The circles indicate outliers. *, p < 0.05.

Figure 4. Association between stromal POSTN and CAF marker expression. (a–c) Association between POSTN and CAF marker expression. (a) DCN, (b) FAP, and (c) α-SMA. POSTN was significantly associated with DCN and FAP but not α-SMA. The circles indicate outliers.

Figure 5. Effects of POSTN on CRC cells. (a) Immunoblot analysis showing POSTN and ITGs in CRC and NIH/3T3 cells. (b) Results for FACS analyses showing ITGA6 and ITGB4 expression on CW-2 cells. (c–e) Results of migration assays showing that co-culture with NIH/3T3POSTN or addition of rPOSTN in the lower chamber accelerated the migration of CRC cells. Bar = 100 µm. Assays were performed in quadruplicate. Data are shown as mean ± S.D. *, p < 0.05, **, p < 0.01. (f) Addition of rPOSTN to the culture medium enhanced the proliferation of CRC cells. Assays were performed in triplicate. Data are shown as mean ± S.D. *, p < 0.05. (g) Immunoblot analysis showing upregulated P-FAK and P-AKT in rPOSTN-stimulated CRC cells. The uncropped bolts are shown in Supplementary Material File S1.

Figure 6. POSTN enhanced the motility but suppressed the proliferation of NIH/3T3 cells. (a,b) Exogenous POSTN enhanced the migration of NIH/3T3 cells. Bar = 100 µm. Assays were performed in quadruplicate. Data are shown as mean ± S.D. **, p < 0.01. (c) POSTN suppressed the cellular proliferation of NIH/3T3 cells. Assays were performed in triplicate. Data are shown as mean ± S.D. *, p < 0.05. (d) Immunoblot analysis showing upregulated STAT3 and P-STAT3 in NIH/3T3 cells. Exogeneous POSTN upregulated P-STAT3 at S727. Co-culture with CRC cells upregulated basal STAT3 and P-STAT3 at Y705 in NIH/3T3. (e) Stattic, a selective inhibitor for STAT3, significantly downregulated the migration of NIH/3T3 cells at 10 µM. Assays were performed in quadruplicate. Data are shown as mean ± S.D. **, p < 0.01. The uncropped bolts are shown in Supplementary Material File S1.