Background: Cyclooxygenase-2 (COX-2) and its principle enzymatic metabolite, prostaglandin E2 (PGE2), are implicated in cancer progression. Based upon immunohistochemical (IHC) evidence that several tumor types in animals overexpress COX-2 protein, COX-2 inhibitors are used as anticancer agents in dogs and cats.
Hypothesis: IHC is inaccurate for assessing tumor-associated COX-2 protein and enzymatic activity.
Methods: Five mammalian cell lines were assessed for COX-2 protein expression by IHC and Western blot analysis (WB), and functional COX-2 activity was based upon PGE2 production.
Results: Detection of COX-2 protein by IHC and WB were in agreement in 4 of 5 cell lines. In 1 cell line that lacked COX-2 gene transcription because of promoter hypermethylation (HCT-116), IHC produced false-positive staining for COX-2 protein expression. Functional COX-2 enzymatic activity was dissociated from relative IHC-based COX-2 protein expression in 2 cell lines (RPMI 2650 and SCCF1). The RPMI 2650 cell line demonstrated strong COX-2 protein expression but minimal PGE2 production.
Conclusions and clinical importance: Western blot is more accurate than IHC for the detection of COX-2 protein in the cell lines studied. Furthermore, the semiquantitative identification of COX-2 protein by IHC or WB does not necessarily correlate with enzymatic activity. Based upon the potential inaccuracy of IHC and dissociation of COX-2 protein expression from enzymatic activity, the practice of instituting treatment of tumors with COX-2 inhibitors based solely on IHC results should be reconsidered.