Selective inhibition of cyclooxygenase-2 inhibits colon cancer cell adhesion to extracellular matrix by decreased expression of beta1 integrin

Abstract

High level expression of cyclooxygenase (COX)-2 is reported in 80-90% of colorectal adenocarcinomas. In the recent years, selective inhibitors of COX-2 have been developed, and are shown to effectively protect against cancer development and progression. Colon cancer cells, as well as the epithelial cells in general, are dependent on appropriate interactions with the extracellular matrix (ECM) proteins to achieve a number of important functions, such as proliferation, differentiation, invasion and survival. These interactions are mediated via a family of cell-surface receptors called integrins, which interact with cytoskeletal proteins on the cytoplasmic side of the plasma membrane and thereby provide a link between the ECM and the cytoskeleton. In the present study, a high-COX-2 (high level COX-2 expression) colon cancer cell line, HT-29, and a low-COX-2 (low level COX-2 expression), DLD-1, were used to investigate the anticolon cancer effect of the selective COX-2 inhibitor, JTE-522. Moreover, to clarify its mechanisms of action, we focused especially on the ability to adhere to and to migrate on ECM. We could clearly demonstrate that, in addition to the decrease of the proliferative activity, JTE-522 caused a dose-dependent decrease in both the ability of colon cancer cells to adhere to and to migrate on ECM. These effects were, at least in part, dependent on the down-regulation of beta1-integrin expression, which was evident in HT-29, the high-COX-2 colon cancer cells, but not the low-COX-2, DLD-1. In addition, prostaglandin E2 almost completely reversed the effect of JTE-522, strongly suggesting the involvement of a COX-2-dependent pathway. In conclusion, for the first time, we could demonstrate the down-regulation of beta1 integrin caused by COX-2 inhibition, with consequent impairment of the ability of cancer cells to adhere to and to migrate on ECM, which are crucial steps for cancer metastases to develop.

Figure 1

Western blotting analysis of cyclooxygenase‐2 (COX)‐2 expression levels in HT‐29 and DLD‐1 colon cancer cell lines. β‐Actin was used as an indicator for equality of lane loading. HT‐29 expressed high levels, and DLD‐1expressed low levels of COX‐2.

Figure 2

Proliferative activities of colon cancer cell lines assessed by MTS assay. (○) HT‐29, and (•) DLD‐1 were cultured for 24 h in the absence or presence of various concentrations of JTE‐522, and the proliferative activity assessed as the dehydrogenase activity in the cell culture medium. Results are expressed as the proliferation rate in relation to control cells, and expressed as the mean ± standard deviation of triplicate wells.

Figure 3

Detection of apoptosis in colon cancer cells by the AnnexinV‐ fluorescein isothiocyanate (FITC)/propidium iodide (PI) double staining apoptosis detection assay (AnnexinV:FITC Apoptosis Detection Kit; BD Pharmingen, San Jose, CA, USA). Data are expressed as the percentage of viable cells (Annexin‐V (–)/PI (–) cell population), after treatment with vehicle or JTE‐522 (25 µM or 50 µM) for 24 h. Data shown are the means ± standard deviation of three independent experiments.

Figure 4

Effect of JTE‐522 on the cell cycle distribution of colon cancer cell lines, HT‐29 and DLD‐1, determined by dual parameter flow‐cytometry. The percentage of cells in each phase of the cell cycle is represented. Compared to control (closed bars), a significant increase in the percentage of cells in G1 phase as well as a significant decrease of cells in S phase was observed in JTE‐522‐treated ([25 µM, open bars][50 µM, shaded bars]) cells. The level of G1 arrest is more pronounced in HT‐29 than in DLD‐1. Data represent the mean ± standard deviation of three independent experiments. *P < 0.05 versus control.

Figure 5

Effect of JTE‐522 on the adhesion of (a) HT‐29 and (b) DLD‐1 to the various extracellular matrix (ECM). Bars represent untreated (closed bar), and JTE‐522‐treated cells (25 µM (open bars) or 50 µM (shaded bars)). JTE‐522 dose‐dependently inhibited the adhesion of HT‐29 to all ECM studied. JTE‐522 also dose‐dependently inhibited the adhesion of DLD‐1, but the effect was rather weaker than that on HT‐29. (c) On HT‐29, prostaglandin E₂ (PGE₂)‐treatment could almost completely reverse the effect of JTE‐522 (net‐like bars). PGE2‐treatment alone (open bars) had no significant effect. Data represent the mean ± standard deviation of triplicate wells. *P < 0.05 versus control.

Figure 6

Effects of JTE‐522 on the surface expression of β1, α2, α6 integrins of colon cancer cell lines. On HT‐29, JTE‐522 decreased the expressions of β1, α2 and α6 integrins, and prostaglandin E₂‐treatment could almost completely reverse the effect. However, the integrin expressions were almost not affected in DLD‐1. The filled histograms represent the negative control. (a–d) represent the histogram of cells stained with the specific monoclonal antibodies.

Figure 7

Immunofluorescent micrographs of β1‐integrin expression on HT‐29 and DLD‐1. Cells were allowed to adhere to type‐I collagen‐coated glass coverslip and treated with vehicle or JTE‐522 (50 µM) in culture medium for 24 h. Then cells were fixed with 3.7% paraformaldehyde and stained with fluorescein isothiocyanate‐conjugated anti‐β1‐integrin antibody. The cells were observed in a confocal microscopy (magnification ×400). Although the β1‐integrin expression of low‐cyclooxygenase‐2 (COX)‐2 (DLD‐1) cells was not affected, that of high‐COX‐2 (HT‐29) cells was significantly decreased by JTE‐522.

Figure 8

Flow cytometric analysis of β1, α2 and α6 integrins on HT‐29. Cells were stained unfixed (left column) or after fixation and permeabilization (right column). The expression on fixed and permeabilized cells represents the total amount of the integrins, that is, surface plus intracellular. Although JTE‐522 caused a significant decrease in the surface expression of β1 integrins, the total amount of expression was not affected.

Figure 9

The effect of JTE‐522 on the migratory activity of HT‐29 and DLD‐1 evaluated by the wounded monolayer repair assay. (a) Representative microphotographs of wounds at times 0, 36 and 72 h after treatment with JTE‐522, obtained in a confocal microscopy (magnification × 40), (b) the percentage closure of the wounds. Each data point represents the mean of nine measures at different points of the wounds ± standard deviation, and results were reported as the percentage of closure relative to values obtained at time 0. Cyclooxygenase‐2 (COX)‐2 inhibition significantly impaired the migration of high‐COX‐2 cells, HT‐29, whereas that of low‐COX‐2, DLD‐1, was only slightly affected.