Cyclooxygenase-2 (COX-2) Is Frequently Expressed in Multiple Myeloma (MM) and Is Associated with Poor Outcome.

Introduction. A perturbed microenvironment with secretion of inflammatory cytokines is typical of MM. Prostaglandins (pg) are implicated in inflammation and angiogenesis and play a role in the pathogenesis of several solid malignancies. Expression of COX-2, the key enzyme of pg synthesis in inflamed tissues, is common in many of these cancers and plays a major role in their development. Moreover, it often acts as a poor prognostic indicator. Despite a large amount of data concerning COX-2 expression in solid tumors, few data are currently available in hematological malignancies. In MM there are several biological, epidemiological and clinical considerations suggesting a potential involvement of the pg pathway. Aim of this study is to verify the involvement of COX-2 in MM and to assess its prognostic role.

Patients and methods. COX-2 expression has been assessed by western blotting (WB) as previously described (Du Bois RN, et al, Gastroenterology, 1996). Our positive control was the COX-2 positive cell line HT-29, while bone marrows (BM) from 15 healthy donors were our negative controls. We assessed a panel of 124 samples obtained by 113 patients with plasma cell dyscrasias. Sixteen samples belonged to subjects with MGUS, 80 to patients with MM at diagnosis, and 28 to patients with relapsed/refractory MM. In 11 patients, samples taken at different treatment phases were available. To confirm WB findings and to demonstrate that COX-2 expression occurs in malignant plasma cells immunohistochemistry (IC), and flow cytometry for COX-2 were also performed in 31 and four patients, respectively. Finally, COX-2 expression has been assessed in BM cells from four COX-2 positive patients following selection for the CD138 antigen using the Miltenyi cell separation system. COX-2 expression at the mRNA level has also been assessed by real time quantitative PCR.

Results. A dilution test showed that our technique is sensitive enough to detect 2% HT-29 cells in a background of COX-2 negative cells. The 15 normal BM were COX-2 negative. In contrast, COX-2 expression was noticed in 12.5% of MGUS, 34.6% of MM at diagnosis and 56% of MM at relapse. COX-2 positivity at diagnosis and relapse was unrelated to disease stage, BM plasmacytosis, creatinine, Hb levels and ß2 microglobulin. COX-2 expression appeared to be of prognostic relevance: at diagnosis the median time to progression was 14 months in COX-2 positive and 40 months in COX-2 negative subjects (p<0.001). At relapse, of 14 patients showing COX-2 expression, 10 have already died of MM, and four are still alive. In contrast, of 11 COX-2 negative patients only one patient died while 10 are still alive (p<0.001). IC, cell separation and flow cytometry studies indicate that COX-2 expression is related to the malignant plasma cell population. COX-2 mRNA was overexpressed in patients showing increased COX-2 protein expression.

Conclusions: a) COX-2 is frequently expressed in plasma cell dyscrasias; b) COX-2 expression is more frequent in advanced disease phases; c) COX-2 expression correlates to a worse outcome. Future studies are required to verify whether COX-2 might be clinically useful as a prognostic marker and/or therapeutic target.