Toll-like receptor 9 expression in breast and ovarian cancer is associated with poorly differentiated tumors

Abstract

Toll-like receptor 9 (TLR9) activates the innate immune response when exposed to non-methylated CpG-DNA. TLR9 was recently shown to be expressed by cancer cells which have been previously characterized by global hypomethylation. We set out to examine the expression and molecular activity of TLR9 in breast and ovarian cancer cells. Firstly, we confirmed higher levels of hypomethylated DNA in the serum of patients with metastatic breast cancer (n = 18) versus age-matched tumor-free women (n = 18). In breast cancer cell lines and tissues, TLR9 mRNA expression was associated with estrogen-receptor (ER) status (n = 124, P = 0.005). Expression also correlated with increasing tumor grade in both breast (P = 0.03) and ovarian cancer specimens (n = 138, P = 0.04). Immunohistochemical analysis of formalin-fixed paraffin-embedded (FFPE) breast cancer tissues revealed higher TLR9 protein expression in hormone-receptor (HR)-negative specimens (n = 116, P < 0.001). Using an in vitro scratch assay, we observed that cell lines transfected to overexpress TLR9 demonstrated increased cellular migration when stimulated with CpG-DNA. When assessing the molecular activity of TLR9 in breast cancer, we found a strong positive correlation of nuclear factor-kappa B (NF-kappaB) activity with TLR9 mRNA expression (correlation coefficient r = 0.7, P < 0.001). Finally, immunofluorescence analysis of BT-20 and Hs578T breast cancer cell lines showed partial colocalizations of CpG-DNA with TLR9, which diminished when the cells were exposed to methylated CpG-DNA (mCpG-DNA) or control GpC-DNA. In summary we demonstrate that TLR9 expression is associated with poor differentiation in breast and ovarian cancer specimens, and that TLR9 overexpression and stimulation with hypomethylated DNA augments the migratory capacity of cancer cell lines.

Figure 1

LINE‐1 DNA hypomethylation analysis of circulating DNA in sera of 18 breast cancer patients and 18 healthy control patients.

Figure 2

Toll‐like receptor 9 (TLR9) mRNA expression in breast and ovarian cancer. (a) analysis in eight breast cancer cell lines; (b) analysis in 10 non‐neoplastic breast tissues and 124 breast cancer samples in association with estrogen receptor (ER) status. (c) Correlation of TLR9 expression with tumor grade in 10 non‐neoplastic breast tissues and 124 breast cancer samples (d) and in 138 ovarian tumor samples and 30 non‐neoplastic ovarian tissues. Tumor grade information was not available for two patients.

Figure 3

Toll‐like receptor 9 (TLR9) expression in breast cancer. (a) Results of immunohistochemical TLR9 staining performed in 116 formalin‐fixed paraffin‐embedded (FFPE) tissues. Association of TLR9 protein expression with hormone receptor (HR)‐status is shown. (b) TLR9 protein expression in untransfected cells, TLR9‐overexpressing cells (pLIB‐TLR9‐iresNeo), and mock‐transfected cells (pLIB‐iresNeo). (c) In vitro scratch assay. Untransfected cells, TLR9‐overexpressing cells, and mock‐transfected cells were treated with CpG‐DNA, GpC‐DNA, or were left untreated. The cell monolayer was then wounded. The same fields were photographed immediately (0 h), and 24, 48, and 55 h later. Results were plotted as percentage of wound closure relative to hour 0. (d) Proliferation measurement of untransfected cells, TLR9‐overexpressing cells, and mock‐transfected cells. Results of six experiments are shown.

Figure 4

Toll‐like receptor 9 (TLR9) and nuclear factor‐kappa B (NF‐κB). (a) DNA binding activity of NF‐κB, assessed by p50 electrophoretic mobility‐shift assay (EMSA) and TLR9 mRNA expression in 23 tumor specimens. (b) Exemplary EMSA picture of nine tumor specimens. Arrows indicate the position of p50/ p65, or p50/ p50 complexes respectively.

Figure 5

Colocalization analysis in BT‐20 human breast cancer cells. (a) Unstimulated cells expressing eCFP‐tagged TLR9 were stained with endoplasmatic reticulum tracker dye (blue, blue‐white tracker). Untransfected cells were stimulated with (b) CpG‐DNA, (c) GpC‐DNA, or (d) mCpG‐DNA. Cells were stained with (b–d) anti‐TLR9 (red, Alexa 568) and (b,c) DAPI nucleic‐acid stain. (e) Stimulated, untransfected cells were stained with anti‐TLR9 (red, Alexa 568), DAPI nucleic‐acid stain, and anti‐EEA1 an early endosome marker (green, FITC). (f) Unstimulated, untransfected cells were stained with anti‐TLR9 (red, Alexa 568) and anti‐EEA1 an early endosome marker (green, FITC).

Figure 6

Colocalization analysis in Hs578T human breast cancer cells. (a) Unstimulated cells expressing eCFP tagged TLR9 were stained with endoplasmatic reticulum tracker dye (blue, blue‐white tracker). Untransfected cells were stimulated with (b) CpG‐DNA, (c) GpC‐DNA, or (d) mCpG‐DNA. Cells were stained with (b–d) anti‐TLR9 (red, Alexa 568) and (b,c) DAPI nucleic‐acid stain. (e) Stimulated, untransfected cells were stained with anti‐TLR9 (red, Alexa 568), DAPI nucleic‐acid stain, and anti‐EEA1 an early endosome marker (green, FITC). (f) Unstimulated, untransfected cells were stained with anti‐TLR9 (red, Alexa 568) and anti‐EEA1 an early endosome marker (green, FITC).