Cancer stem cell-suppressing effect of nonsteroidal anti-inflammatory drugs via cyclooxygenase-2-dependent and -independent pathways in colorectal cancer

Abstract

Cancer stem cell (CSC) model assumes that a small subset of cells in tumors have the ability to initiate and sustain tumor growth. CSCs are resistant to many current chemotherapeutic agents and play a pivotal role in cancer relapse. In this study, we aimed to identify the effective agents to increase the sensitivity to chemotherapeutic agents by suppressing CSCs in human colorectal cancer (CRC).Colosphere forming assay and flow cytometric analysis of CSC markers (CD133 and CD44) were performed to investigate the CSC suppressing effect of nonsteroidal anti-inflammatory drugs (NSAIDs), which are known having the activities of peroxisome proliferator-activated receptor γ (PPARγ) agonist and γ-secretase inhibitor as well as cyclooxygenase (COX) inhibitor. In vitro experiments using SW620 cells, CSC markers and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) assay were carried out after treatment of control, indomethacin, 5- fluorouracil (5-FU), and combination of 5-FU and indomethacin. We also carried out nude mice xenograft experiments using 5-FU resistant SW620 cells and the same drug combination with in vitro cell line experiments. To investigate the underlying mechanisms, we measured changes of CSC population after treatment of COX-2 selective inhibitor (celecoxib), other NSAIDs (sulindac and aspirin) and combination of indomethacin and prostaglandin E2 (PGE2), and performed reporter assay using PPAR-responsive element (PPRE)-Luc. In xenograft experiments, the expressions of HES1 (Notch signaling marker), PPARγ and COX-2 as well as CD133 and CD44 were evaluated by immunohistochemical (IHC) stain.As a result, NSAIDs, including indomethacin, sulindac and aspirin, celecoxib, γ-secretase inhibitor (DAPT), and PPARγ agonist (rosiglitazone) significantly decreased CD133+CD44+ cells and induced over 50% decrease in the number of colospheres

compared to control of SW620 cells. Compared to the control (100%), the treatment of low dose indomethacin (12.5 μM) for 4 days significantly decreased CD133+CD44+ cells (72.1%, P = 0.014), treatment of low dose 5-FU (2.0 μM) for 4 days led to the significant increases of CD133+CD44+ cells (228.2%, P = 0.014), and this 5-FU induced increase of CD133+CD44+ cells was inhibited by combination with indomethacin for the same period (133.1%, P = 0.021). In MTT assay, there was no significant difference in cell survival between groups, and these CSC-inhibitory effects of indomethacin was reversed by PGE2 in a dose-dependent manner. Indomethacin treatment, as well as rosiglitazone, significantly increased PPRE transcriptional activity. In xenograft experiments, 5-FU treatment combined with indomethacin significantly reduced tumor growth compared to 5-FU alone treated group. In addition, the treatment of indomethacin alone and combination of 5-FU and indomethacin decreased the expression of CD133, CD44, COX-2 and HES-1, and increased PPARγ expression, compared to control and 5-FU alone treated mice, respectively.In conclusion, NSAIDs could selectively reduce the colon CSCs and suppress 5-FU induced increase of CSCs through COX-2-dependent and -independent pathways, such as PPARγ and Notch pathway. These suggest that NSAIDs could play an important role of adjunctive treatment with conventional chemotherapy in CRC.