Requirement of matrix metalloproteinase-9 for the transformation of human mammary epithelial cells by Stat3-C

Abstract

Persistently activated Stat3 is found in many different cancers, including approximately 60% of breast tumors. Here, we demonstrate that a constitutively activated Stat3 transforms immortalized human mammary epithelial cells and that this oncogenic event requires the activity of matrix metalloproteinase-9 (MMP-9). By immunohistochemical analysis, we observe a positive correlation between strong MMP-9 expression and tyrosine phosphorylated Stat3 in primary breast cancer specimens. These results demonstrate a relationship between activated Stat3 and MMP-9 in breast oncogenesis.

Fig. 1.

Stat3-C induces tumorigenesis of HMLHT and MCF-10A cells in a dose-dependent manner. (A) Anti-Flag Western blot showing Stat3-C expression in MCF-10A and HMLHT cells expressing pBabe control vector (pB) and pBabe-Stat3-C (pB-3C). (B) EMSA performed with nuclear extracts from cell lines described in A. Stat3-C DNA binding was supershifted with anti-Flag antibody, indicated with a +. (C) Colony formation in soft agar of empty retroviral control (pB) and Stat3-C infected (pB-3C) MCF-10A and HMLHT cells (mean ± SD). (D) Tumor growth in nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice when using the HMLHT pBabe control cell line and subclones with high Stat3-C expression levels (no. 16 and no. 24) or low Stat3-C expression (3CL). Results are expressed as the mean of 4–10 tumors ± SD at the indicated times after injection. (E) Nuclear extracts from a Stat3-C-derived tumor (tu), normal murine breast tissue (nbr), and cell line no. 16 (+con) were analyzed for the presence of Stat3-C by Flag immunoblot.

Fig. 2.

Stat-3C-dependent induction of MMP-9 mRNA, luciferase activity, and protein. (A) Induction of MMP-9 mRNA in pBabe-Stat3-C (pB-3C)- and pBabe (pB)-infected MCF-10A and HMLHT cells determined by RT-PCR (Upper), normalized to GAPDH (Lower). (B) HMLHT cells were transfected with an MMP-9 promoter luciferase construct in conjunction with either pBabe (pB)- or pB Stat3-C (3-C)-expressing plasmids. Luciferase activities are shown as the mean ± SD of three experiments performed in duplicate. (C) MMP-9 protein expression in cell culture medium from pBabe (pB)- and pBabe-Stat3-C (pB-3C)-infected MCF-10A and HMLHT cells shown by gelatin zymography.

Fig. 3.

Active MMP-9 is localized to the cell surface. (A and B) An ELISA specific for enzymatically active MMP-9 was performed on cell culture medium (A) and cell extracts (B) from pBabe (pB)- and pBabe-Stat3-C (pB-3C)-expressing cells. MMP-9 activity from cell culture medium and cell extracts was measured without (black columns) and with pretreatment with APMA (gray columns). Results are shown as the mean ± SD of three experiments performed in duplicate. (C) In situ zymography of HMLHT cells expressing pBabe and pBabe-Stat3-C (pB-3-C) treated with DMSO (Upper) or 1.5 μM MMP-2/9 inhibitor (Lower). The cells were then overlayed with DQ gelatin. Green staining indicates MMP-9-digested gelatin whereas blue indicates nuclear staining [4′,6-diamidino-2-phenylindole (DAPI)]. (D) MMP-9 expression shown by immunofluorescence in the cell lines described in C.

Fig. 4.

MMP-9 activity is required for Stat-3C-dependent anchorage-independent growth. (A) Anchorage-independent growth of pBabe-Stat3-C cells (pB-3C)-, pBabe v-src (pB-src)-, and pBabe H-ras V12 (pB-ras)-expressing HMLHT cells. DMSO (D) control and increasing concentrations of MMP-2/9 inhibitor in μM were added to the soft agar assay every other day (mean ± SD). (B) Gelatin zymography of supernatants derived from HMLHT cells expressing either pBabe (pB), pBabe-Stat3-C (3C), pBabe v-src (src), or pBabe H-ras V12 (ras) and 0.5 ng of recombinant MMP-9 as a loading control.

Fig. 5.

Persistently phosphorylated Stat3 correlates with MMP-9 expression in primary breast cancer samples. Immunohistochemistry was performed on sequential sections of 34 primary breast cancer microtissue arrays with antiphospho-Stat3 (pStat3) and anti-MMP-9 antibodies. A schematic overview of the tissue arrays and a summary of the immunohistochemistry results are shown. Representative sections of strong staining are indicated as +++ and shaded in black, moderate staining as ++ and shaded in gray, and weak to no staining as 0/+ and shaded in white. Normal breast had 0/+ staining for both pStat3 and MMP-9. A positive correlation was observed between (+++/++) staining for pStat3 and MMP-9 (P < 0.001) by χ² test.