Abstract

COX-2 has been implicated in the development of many epithelial cancers, as well as in tumor angiogenesis. COX-2 inhibitors have been shown to have anti-tumor activity in experimental cancer. Little information exists, however, on the expression or role of COX-2 in hematologic malignancies. We have use a variety of immunochemical assays to document expression of COX-2 in human and murine leukemias and hematopoietic cells. The factor-dependent murine cell lines FDCP1 and 32D expressed COX-2 when growing continuously in the presence of IL-3; expression declined markedly when growth factor was removed. FDCP1 cells constitutively expressing bcl-2, pim-1, or bcr-abl had markedly elevated levels of COX-2, and continued to express this enzyme even after removal of growth factor. To assess COX-2 expression in human hematopoietic cells we developed a flow cytometry assay using a FITC-labelled anti-COX-2 MoAb (Cayman). Cells were washed once in serum-free medium, fixed briefly in 1% paraformaldehyde, permeabilized with PBS/0.2% Triton X100, then stained with antibody. Negative control samples were processed similarly but stained with antibody that had been preincubated with immunizing peptide. Specific COX-2 staining was interpreted as the difference between the histograms from blocked versus unblocked anti-COX-2 antibody, as determined by Kolmogorov-Smirnoff analysis. In buffy coat preparations from normal donors, we found constitutive expression of COX-2 in lymphocytes (both B-cells and T-cells). In contrast little or no COX-2 was detected in unstimulated neutrophils or monocytes. In human acute myelogenous leukemia (AML) cell lines we found COX-2 expression to be universal and easily detected. In several cell lines we confirmed the results of our flow cytometry assay with immunoblotting. We further examined 25 cryopreserved samples of human acute leukemia blasts obtained from peripheral blood. COX-2 expression was variable, but universal. Levels generally were less than those seen in immortalized cell lines, and did not correlate with blasts morphology (AML, ALL, APL, AMoL, CML-BT). To determine if COX-2 inhibitors could play a role in the treatment of acute leukemias, we performed cytotoxicity assays using the COX-2 specific inhibitors, celecoxib and NS398. Survival and growth of human AML cell lines were inhibited by both agents. These data demonstrate that 1) a variety of oncogenes can induce expression of COX-2 in hematopoietic cells, 2) clinical human acute leukemias uniformly express COX-2 in circulating blasts, and 3) COX-2 inhibitors are cytotoxic for human leukemia cells. Combination therapies for acute leukemias may evaluate the incorporation of COX-2 inhibitors for added cytotoxic effects or angiogenesis inhibition.

Author notes

Corresponding author