Roles of COX, NOS, and COX/NOS interactions in breast cancer angiogenesis and lymphoangiogenesis

4941

We had shown that an aberrant expression of COX-2 or NOS promotes breast cancer progression and metastasis; tumor-derived PGE2 and NO independently stimulated breast cancer cell migration, invasiveness and angiogenesis. Utilizing highly metastatic, COX-2 and eNOS expressing murine (C3L5) and human (MDA-MB-231) breast cancer cell lines, PGE2-mediated migratory responses were shown to involve EP4 receptors, adenylate cyclase (AC) and PKA. NO-mediated migratory responses required guanylate cyclase and MAP kinase. While NO had no influence on PGE2 production by C3L5 cells, endogenous PGE2 upregulated iNOS under inducible conditions, further stimulating cellular invasiveness. This action of PGE2 also utilized EP4/AC axis. Thus both COX-2 and EP4 present as important therapeutic targets for blocking invasion and metastasis. In the present study we examined COX/NOS interactions in promoting angiogenesis and lymphoangiogenesis, utilizing C3L5 cells in vivo and MDA-MB-231 cells in vitro. Angiogenesis assay (Jadeski and Lala, Am J Pathol 1999, 155:1381-1390) utilized s.c. implants of growth factor-reduced matrigel inclusive of C3L5 cells in syngenic C3H/HeJ mice (15 mice/treatment group), killed on day 14. Mice received COX inhibitor indomethacin (12 μg/ml in 0.2% ethanol in drinking water) or vehicle alone (control), NOS inhibitor L-NAME (given s.c. via osmotic minipumps 25 μg/200 μl saline; 0.5 ml/hour) or D-NAME (control), a combination of indo+L-NAME, or vehicle+D-NAME (control). The mid-cross sectional area of the implants provided a measure of tumor growth, whereas the mean number of blood vessels per unit area provided a measure of angiogenesis. Both indo and L-NAME as well as their combination caused a significant but similar decrease in tumor size as well as angiogenesis. MDA-MB-231 cells expressed mRNA for the angiogenic factors VEGF A and B, and lymphoangiogenic factors VEGF C and D, as tested with RT-PCR. Production of VEGF-C as measured by ELISA in 24 hr serum-free culture supernatants was high, but VEGF-D was scanty. Cells were treated with the nonselective COX inhibitor indo (20 μM), selective COX-2 inhibitor NS-398 (50 μM), NOS inhibitor L-NAME (1 mM), or its inactive enantiomer D-NAME (1 mM), or a combination of indo+L-NAME, or EP1 receptor antagonist SC-51322 (10 μM), or EP4 receptor antagonist AH-23848B (10 μM). A significant and similar reduction of VEGF-C secretion was noted with indo, NS-398, indo+L-NAME, but not L-NAME alone. Both EP1 and EP4 antagonists also reduced VEGF-C secretion. Thus COX-2 and possibly EP1/EP4 are important therapeutic targets for blocking angiogenesis/lymphoangiogenesis in breast cancer. (Supported by the CBCRA grant # 012312 to PKL).