Gastric mucosal cell model for estimating relative gastrointestinal toxicity of non-steroidal anti-inflammatory drugs

Abstract

The study objective was to characterize the AGS human gastric mucosal cell line as a model for estimating gastrointestinal toxicity of COX-inhibiting compounds. Rofecoxib, celecoxib, nimesulide, ibuprofen, indomethacin, aspirin, salicylic acid, naproxen and acetaminophen were tested for inhibition of COX-2-mediated prostaglandin E₂ synthesis in A549 and AGS cells. The IC₅₀ ratio AGS/A549 was calculated as an estimate of the therapeutic index (TI) for gastrointestinal toxicity. Calculated IC₅₀ values of non-steroidal anti-inflammatory drugs (NSAIDs) in A549 cells were in excellent agreement with published values (r=0.996; P<0.005). Calcium ionophore induction of arachidonic acid release in AGS cells provided TI similar to those using platelets and A549 cells (r=0.918; P<0.01). The AGS/A549 model exhibited lower TI than the platelet/A549 model. Spearman ranking correlated clinical NSAID gastropathy with lower AGS TI values. The AGS cell line has excellent potential to serve as a model for assessing the gastrointestinal effects of COX-inhibiting compounds.

Introduction

The discovery of an inflammation-inducible cyclooxygenase isozyme (COX-2), which is regulated differently than the constitutively expressed COX-1 isozyme, prompted the identification of compounds that could selectively inhibit COX-2 without affecting COX-1. The newer NSAIDs developed as COX-2-specific inhibitors have generally demonstrated fewer gastrointestinal side effects in short-term studies as compared to the older NSAIDs that inhibit both COX-1 and COX-2 (e.g., aspirin, ibuprofen). But surprisingly, long-term clinical experience has revealed that these COX-2-specific agents do show adverse gastrointestinal effects, which includes induction of spontaneous bleeding and delayed ulcer healing. Currently, there exists no validated in vitro assay to evaluate potential gastrointestinal effects of natural COX-2 inhibitors or NSAIDs.

The generally held concept (COX dogma) is that COX-1 is expressed constitutively in most tissues, whereas COX-2 is the inducible enzyme triggered by pro-inflammatory stimuli including mitogens [1], cytokines [2] and bacterial lipopolysaccharide (LPS)[3], [4] in cells in vitro and at inflamed sites in vivo [5]. Based primarily on such differences in expression, COX-1 has been characterized as a housekeeping enzyme and is thought to be involved in maintaining physiological functions, such as cytoprotection of the gastric mucosa, regulation of renal blood flow, and control of platelet aggregation [6], [7]. COX-2 is considered to mainly mediate inflammation, although constitutive expression is found in the brain, kidney, gastrointestinal tract and blood vessels [7], [8], [9].

Depicted in Table 1 are the categories of the numerous in vitro assays that have been developed for testing and comparing the relative inhibitory activities of NSAIDs and natural compounds against COX-1 and COX-2. As reviewed by Pairet and van Ryn [10], these test systems can be classified into three groups: (1) systems using animal enzymes, animal cells, or cell lines; (2) assays using human cell lines, or human platelets and monocytes; and (3) currently evolving models using human cells that are representative of the target cells for the anti-inflammatory and adverse effects of NSAID and dietary supplements. Generally speaking, models using human cell lines or human platelets and monocytes are the current standard, and validated target cell models have not been forthcoming [13].

Prostaglandins (PG) are believed to play an important role in maintenance of human gastric mucosal homeostasis [11]. COX-1 appears to be responsible for PG synthesis in normal gastric mucosa in order to maintain mucosal homeostasis. The classical COX-2 hypothesis, however, has downplayed the role of COX-2 expression in the gastrointestinal mucosa. There is increasing evidence that a detectable amount of COX-2 mRNA and protein are both constitutively expressed and inducible at specific locations of the gastric mucosa in both animals and humans [12], [13]. COX-2 is expressed by normal gastric mucosa at low levels with induction of expression during ulcer healing, following endotoxin exposure, and after cytokine stimulation [13], [14]. Recent studies in rats have shown that, whereas selective inhibition of COX-1 or COX-2 is not ulcerogenic, combined inhibition of both COX-1 and COX-2 induces severe lesions in the stomach and small intestine comparable with the effects of NSAIDs [15], [16]. This observation suggests an important contribution of COX-2 to the maintenance of gastrointestinal mucosal integrity.

Furthermore, up-regulation of COX-2 expression in the gastric mucosa can be induced by various growth factors and cytokines. Conditions in which significant over-expression of COX-2 occurs in the gastric mucosa are Helicobacter pylori infection, stress damage to the gastric mucosa, ischemia/reperfusion, and gastric ulcer healing [12], [14], [17]. Therefore, it now appears that both COX-1 and COX-2 have important physiological roles in the normal gastric mucosa.

These observations indicate an appropriate in vitro cellular model for assessing the gastrointestinal effects of cyclooxygenase inhibitors should constitutively express both COX-1 and COX-2 in appropriately relative amounts and should express COX-2 in response to physiological stressors. Previously published research indicated that the AGS human gastric mucosal cell line had an essential characteristic—specifically constitutive expression of both COX-1, COX-2, and additional COX-2 induced by IL-1β—to potentially make it an excellent model for assessing gastrointestinal toxicity of COX-2 inhibitors [14]. The objective of this study was to characterize the AGS cell line as a model for estimating relative gastrointestinal toxicity of prescription and over-the-counter COX-inhibiting compounds.