Kaplan-Meier survival curves were plotted, and log rank test was done. 1 High COX-2 expression in breast cancer TAMsA. The relative COX-2 mRNA expression in different monocytes/macrophages. Mean SD, = 9, **< 0.01. B. PGE2 amount in supernatants of MDMs or MDMs-derived TAMs was measured by CIA assay. Mean SD, = 9, **< 0.01. C. The representative double immunofluorescence staining of CD163 (green) and COX-2 (red) in breast cancer tissues (Left) or pericarcinoma tissues (Right) (original magnification, 400). D. Clasto-Lactacystin b-lactone Correlation of COX-2+ TAMs and Ki67 in breast cancer tissues (= 160) was analyzed by Pearson's correlation analysis. E. Correlation of COX-2+ TAMs and COX-2 in breast cancer cells (= 160) was analyzed by Pearson's correlation analysis. F. Kaplan-Meier 10-years OS curves for breast cancer patients according to COX-2+ TAMs density (= 160). High COX-2 expression in TAMs correlates with poor prognosis in breast cancer patients In order to determine the role of COX-2 in breast TAMs, a double immunofluorescent staining of COX-2 and CD163 (a specific marker for TAMs) was performed in a breast tissue array containing 160 human breast cancer tissue specimens and 10 pericarcinoma tissue controls. A greater number of COX-2+ macrophages were found in cancer samples than that in nonmalignant pericarcinoma samples (< 0.001, Figure ?Figure1C).1C). The number of COX-2+ TAMs was associated with increased clinical staging (= 0.024) and aggressive tumor biology by advanced histopathological grading (< 0.001) and lymph node metastasis (= 0.021) (Table ?(Table1).1). Furthermore, there was a significant positive correlation between COX-2+ TAMs and the cell proliferation marker Ki67 (= 0.449, < 0.001, Figure ?Figure1D)1D) or Clasto-Lactacystin b-lactone COX-2 expression (= 0.888, < 0.001, Figure ?Figure1E)1E) in breast cancer cells. However, there was no association between COX-2+ TAM counts and other clinical parameters including Clasto-Lactacystin b-lactone patient age and molecular subtypes (> 0.05). Kaplan-Meier survival curve with a median follow-up period of 118 months demonstrated that a significantly higher overall survival (OS) rate was observed in patients with low COX-2+ TAM counts than those with high COX-2+ TAM counts (< 0.01, Figure ?Figure1F).1F). In a multivariate Cox regression analysis, COX-2+ TAM counts were associated with poor survival prognosis of breast cancer patients (HR = 2.085, = 0.036), independent of other clinical covariates (Table ?(Table2),2), indicating that COX-2+ TAM is an independent prognostic biomarker for breast cancer outcome, and COX-2 in TAMs may play an important role in breast cancer progression. Table 1 Correlation of COX-2 Expressing TAM Counts with Clinicopathological Status in 160 Cases of Patients with Breast Cancer = 57)= 103)Valuevalues> 601.6940.902C3.1810.101Tumor size(>2 cm)0.6480.314C1.3370.240TNM stage (III)2.0151.032C3.9330.040Histological Grade (>II)2.9251.096C7.8020.032ER0.7980.418C1.5320.494PR0.6910.369C1.2950.249HER21.7950.939C3.4330.077Ki670.9020.488C1.6680.743Density of COX-2+ TAM (>13.59 mm?2)2.0851.050C4.1400.036 Open in a separate window Over-expression of COX-2 in TAMs promotes breast cancer cell proliferation and survival In order to elucidate the tumor-promoting role of COX-2 in breast TAMs, TAMs were first transfected with adenoviral COX-2 or siRNA COX-2 (Supplementary Figure S2), and then co-cultured with different breast cancer cell lines (MCF-7 and MDA-MB-231) for 7 days. Cancer cell proliferation, viability or apoptosis induced by various cytotoxic drugs were measured by CCK-8 or PI staining assays, respectively. We found that TAMs promoted proliferation and resistance to drugs-induced apoptosis in breast cancer cells, which was enhanced by COX-2 over-expression but attenuated by COX-2 knockdown in TAMs (Figure ?(Figure2A2AC2B and Supplementary Figure S3). Consistent with these findings, higher mammary tumor weight/volume was observed in NOD/SCID mice injected with 4T1 murine breast cancer cells/RAW 264.7-derived TAMs, compared with that in FGF5 mice injected with 4T1 cells only. Clasto-Lactacystin b-lactone Tumor weight/volume was much higher in mice injected with 4T1/COX-2+ TAMs, while lower in mice injected with 4T1/COX-2? TAMs than that.