Select Dietary Phytochemicals Function as Inhibitors of COX-1 but Not COX-2

Recent clinical trials raised concerns regarding the cardiovascular toxicity of selective cyclooxygenase-2 (COX-2) inhibitors. Many active dietary factors are reported to suppress carcinogenesis by targeting COX-2. A major question was accordingly raised: why has the lifelong use of phytochemicals that likely inhibit COX-2 presumably not been associated with adverse cardiovascular side effects. To answer this question, we selected a library of dietary-derived phytochemicals and evaluated their potential cardiovascular toxicity in human umbilical vein endothelial cells. Our data indicated that the possibility of cardiovascular toxicity of these dietary phytochemicals was low. Further mechanistic studies revealed that the actions of these phytochemicals were similar to aspirin in that they mainly inhibited COX-1 rather than COX-2, especially at low doses.


Introduction
Consistent clinical studies have indicated that long-term administration of COX-2 inhibitors is associated with an enhanced risk of experiencing adverse cardiovascular events [1,2]. Although the exact mechanism still remains unclear, accumulating evidence supports the idea that COX-2 plays a cardioprotective role after cardiac injury [3][4][5]. Functional recovery after induced cardiac injury was improved in COX-2 transgenic mouse, but was greatly reduced by deficiency of COX-2. In the search to identify promising cancer chemopreventive agents, dietary phytochemicals have emerged as potential agents based on their observed anticancer activities as well as perceived safety [6]. Some mechanistic studies revealed that active dietary factors, such as EGCG, curcumin or resveratrol, might suppress carcinogenesis by targeting COX-2 [7][8][9][10][11][12][13]. One question that has been raised is why the lifelong use of phytochemicals that likely inhibit COX-2 has reportedly not been associated with adverse cardiovascular side effects. We hypothesize that those naturally occurring compounds might share a similar mechanism of action with aspirin, and might preferentially target COX-1 rather than COX-2. To test this idea, we selected a library of well-known active dietary factors and evaluated used in this study were obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained following their instructions. Cells were cytogenetically tested and authenticated before the cells were frozen. The passage number was routinely limited to approximately 20 and morphology monitored with each passage.

Cell viability assay
Cells were seeded in 96-well-plate at a density of 5000 cells per well and allowed to incubate at 37 °C for 24 h for attachment. After drug treatment for 8 h, 20 µL CellTiter96 Aqueous One Solution (Invitrogen, Carlsbad, CA) were added, and cells were further incubated for 1 h at 37 °C. Finally, the optical density was determined at 492 nm.

Measurement of NO
HUVEC cells (6×10 5 ) were seeded in a six-well plate in the presence of 10% FBS. At 70-80% confluence, cells were pretreated with vehicle or individual compounds in 1 mL fresh medium for 2 h. After that, cells were incubated or not with IL-1β (17.5 ng/mL) for another 8 h. Supernatant fractions were collected for measurement of nitric oxide (NO). NO concentration was measured as nitrite using the Nitrate/Nitrite colorimetric assay kit (Cayman Chemicals, Ann Arbor, MI).

In vitro COX enzyme assay
COX activity was evaluated using a COX Inhibitor Screening Kit from Cayman Chemical Company (Ann Arbor, MI) according to the manufacturer's instructions.
Thromboxane B 2 (TXB 2 ) and 6-keto prostaglandin F 1α (6keto PGF 1α ) assay Cells (6×10 5 ) were seeded in a six-well plate in the presence of 10% FBS. At 70-80% confluence, cells were pretreated with vehicle or individual compounds in 1 mL fresh medium for 2 h. After that, cells were or were not incubated with IL-1β (17.5 ng/mL) for another 8 h. Supernatant fractions were collected for prostaglandin measurement using enzyme immunoassay kits (Cayman Chemical Company).

Western blot analysis
Protein samples (20 µg) were resolved by SDS-PAGE and transferred to Hybond C nitrocellulose membranes (Amersham Corporation, Arlington Heights, IL). After blocking, the membranes were probed with primary antibodies (1:1000) overnight at 4 °C. The targeted protein bands were visualized using an enhanced chemiluminescence reagent (Amersham Corporation) after hybridization with a secondary antibody conjugated with horseradish peroxidase.

Statistical analysis
All experiments were performed at least three times independently. Statistical analysis was performed using the Prism statistical package. Turkey's t-test was used to compare data between two groups. One-way ANOVA and the Bonferroni correction were used to compare data between three or more groups. Values are expressed as means ± S.E.M. and a p < 0.05 was considered statistically significant.

Evaluation of potential cardiovascular toxicity
The imbalance between COX-1-derived pro-thrombotic thromboxane A 2 (TXA 2 ) and COX-2-relateded anti-thrombotic prostacyclin (PGI 2 ) production has long been suspected to contribute to cardiovascular side effects of COX-2 inhibitors [16,17]. COX-2 -/-mice are more prone to cardiovascular risk than wild type mice, evidenced by increased cardiac ischemia and/or reperfusion injury [18]. Therfore, we firstly examined this idea in wildtype (COX2 +/+ ) or knockout (COX -/-) mouse embryonic fibroblasts (MEFs, Figure 1A). COX-2 deficiency enhanced the ratio of thromboxane B 2 (TXB 2 , the stable breakdown product of TXA 2 ) to 6-keto prostaglandin F 1α (6-keto PGF 1α , the hydrolysis product of PGI 2 ) by 22-fold. We further tested such hypothesis in an in vitro model using human umbilical vein endothelial cells (HUVECs). Again, the ratio of TXB 2 /6-keto-PGF1 α was dramatically increased by the selective COX-2 inhibitor celecoxib, but not by aspirin, which is known to target COX-1 rather than COX-2 ( Figure 1B). All of these findings support the reliability of the ratio of thromboxane B 2 to 6-keto prostaglandin F 1α as a biomarker for COX-2 inhibitionrelated cardiovascular toxicity.
Considering the fact that people usually take selective COX-2 inhibitors to relieve pain and reduce inflammation, HUVECs were treated with IL-1β, an inflammatory cytokine implicated in vascular diseases, to mimic pro-inflammatory conditions. Stimulation of cells with IL-1β resulted in a remarkable increase in COX-2 expression as well as 6-keto-PGF1 α synthesis. Similar to physiological conditions, the ratio of TXB 2 /6-keto-PGF1 α was significantly enhanced by celecoxib, whereas was only weakly affected by aspirin as well as by dietary factors (Figure 3).
In addition to COX-2-related PGI 2 , endothelial-derived nitric oxide (NO) also acted as an endogenous vasodilator and protected the blood vessel wall by inhibiting platelet aggregation. In this study, we observed that IL-1β treatment caused a 2.6 fold increase in NO production compared with the control group. More importantly, most dietary phytochemicals had no effect on NO release ( Figure 4A). We also excluded generalized cytotoxicity by examining the effects of dietary phytochemicals on cell viability and found that they had little effect on HUVEC cell viability after 8 hours of treatment ( Figure  4B). Taken together, these findings suggested that the possible cardiovascular toxicity of dietary phytochemicals is low.

Determination of the effect of phytochemicals on COX activity in vitro
The differential effects between dietary factors and celecoxib on cardiovascular toxicity biomarkers indicated that these dietary phytochemicals might not be specific COX-2 inhibitors. Accordingly, we determined their effect on COXs activity using a COX inhibitor screening assay kit. Results clearly indicated that most of the compounds are COX-1 inhibitors with a mild to moderate COX-1 selectivity index (Table 1). Among eight dietary phytochemicals, six selectively inhibited COX-1 activity rather than COX-2. Another two flavonols (naringenin and quercetin) are likely not COX1/2 inhibitors because their 50% inhibitory concentrations (IC 50 ) against COX1/2 activity was higher than 400 µM.

Discussion
In this study, the cardiovascular safety of selected dietary factors was systemically evaluated for the first time. Our data indicated that the possible cardiovascular toxicity of dietary phytochemicals was low because the compounds tested might share a mechanism of action similar to aspirin and most appeared to preferentially target COX-1 rather than COX-2. During the course of this study, EGCG, an active ingredient in green tea, exhibited an unexpected cardioprotective property and might merit further investigation. Among dietary factors studied, EGCG exhibited the most potent inhibitory effect against the ratio of TXB 2 /6-keto-PGF1 α either under physiological or pathological conditions. This finding was consistent with several recent epidemiologic studies, which suggested regular consumption of green tea might provide cardioprotective effects [28,29]. This unanticipated finding provides critical insight into the potential application of green tea for cardioprotection.
Aspirin at low dose (81 mg per day) is widely accepted to be able to provide both cardioprotective and chemopreventive effects [17,23,30]. However, pharmacokinetic data analysis revealed that at this dose, aspirin might mainly targets COX-1 rather than COX-2, because the maximal serum concentration achieved was well below the reported whole blood COX-2 IC 50 values [31,32]. In this study, we confirmed that most natural product-based compounds were COX-1, rather than COX-2 selective inhibitors. This raised the question of whether those natural occurring compounds exert their chemopreventive activity, at least in part, by targeting COX-1. Although no conclusion can be drawn due to insufficient data at this time, accumulating evidence suggests that COX-1 is involved in carcinogenesis [33][34][35][36]. For example, overexpression of COX-1 leads to tumorigenic transformation, whereas genetic disruption of ptgs-1 greatly reduced cancer incidence both in skin and colon. Although COX-1 is now becoming a target to be reconsidered for cancer prevention or treatment, selective COX-1 inhibition is still a controversial issue. For example, inhibition COX-1 has been strongly implicated in the gastric ulceration and bleeding induced by non-steroidal antiinflammatory drugs (NSAIDs) because people believe that COX-1 is responsible for the prostaglandins essential for normal mucosal physiology in gut. As no gastrointestinal toxicity data were collected in this study, whether these phytochemicals cause gastrointestinal bleeding is still unknown and further study in these areas is required.

Author Contributions
Conceived and designed the experiments: HL FZ RAL AMB ZD. Performed the experiments: HL NO HC YS BL. Analyzed the data: HL FZ RAL ZD. Contributed reagents/materials/ analysis tools: HC RAL ZD. Wrote the manuscript: HL AMB.