Abstract
In comparison to other cancers, advanced prostate cancer is resistant to chemotherapy. There is a need to understand the mechanisms which are responsible for this resistance and find better treatments for this disease or methods to increase the efficacy of current treatments. Cancer cells often evade apoptosis. Cyclooxygenase-2 (COX-2) is an enzyme reported to be elevated in prostate cancer, and has oncogenic properties, including apoptosis attenuation. Because of this, COX-2 inhibition could be beneficial for both prevention and treatment of cancer. This study showed that COX-2 protein was not detected in LNCaP, PC-3 or DU145 cells. To assess the effect COX-2 has on the apoptotic sensitivity of prostate cancer cells, we transfected two cell lines (stable transfection in LNCaP, transient in PC-3) with the human COX-2 gene or empty control vector. We measured the effect on cell viability of COX-2 after treatment with a diverse set of agents e.g. etoposide, carboplatin, Fas, TRAIL, celecoxib and sulindac, using the MTT assay. We observed COX-2 dependent resistance to carboplatin, etoposide and celecoxib in LNCaP but not PC-3. Carboplatin mediated reduction in cell viability was due to an S phase block and induction of apoptosis. COX-2 transfection in LNCaP cells attenuated both the cell cycle block and apoptosis. There was reduced p53 and p27KIP1 induction following carboplatin treatment in LNCaP-COX-2, compared to LNCaP-Neo. COX-2 transfection caused elevated cellular- levels of anti-apoptotic proteins Bcl-2, BC1-XL and survivin. Summary Celecoxib could not reverse the resistance seen in LNCaP-COX-2 to carboplatin, and PGE2 could not increase the resistance in LNCaP-Neo cells, suggesting that COX-2 mediates an apoptotic resistance which is COX-2 enzymatic activity independent. Other mechanisms were sought to reverse the carboplatin resistance. PI3K inhibitors wortmannin and LY294002 partially reversed the LNCaP-COX-2 resistance. These data suggests that COX-2 acts on the PI3K signalling pathway to mediate resistance in LNCaP. This was confirmed by Western blot findings of elevated levels of in P-Aktser473 LNCaP-COX-2 compared with LNCaP-Neo. Celecoxib decreased levels of P-Aktser473 inn a manner similai- to PI3K inhibitors, indicating that the PI3K inhibitors and celecoxib act in different ways on PI3K signalling. Because complete reversal of resistance does not occur with PI3K inhibition, COX-2 must be acting on other pathways also. However, unlike the PI3K inhibitors, celecoxib could not reverse resistance to carboplatin. Wortmannin and LY294002 decreased levels of Bcl-2, BC1-XL, survivin and, surprisingly, COX-2 protein. In contrast, celecoxib had little effect on Bcl-2 and increased COX-2 levels. Cyclin B1 levels were higher in LNCaP-COX-2 than LNCaP-Neo. Gene microarray analysis confirmed the up-regulation of survivin in LNCaP-COX-2 and also showed elevation of cyclin I and fatty acid synthase and down-regulation of glutathione-S-transferase in LNCaP-COX-2. In conclusion, stable COX-2 transfection causes a selective resistance to cytotoxic agents which is mediated via activation of the PI3K pathway and attenuation of p53 induction in LNCaP. COX-2 also causes the up-regulation of a number of anti- apoptotic factors and PI3K inhibition results in down-regulation of these proteins.