Effects of selective cyclooxygenase-2 and nonselective cyclooxygenase inhibition on myocardial function and perfusion

Abstract

Nonselective nonsteroidal anti-inflammatory drugs and selective cyclooxygenase-2 (COX-2) inhibitors are purported to increase adverse cardiovascular events. We hypothesized that COX-2 inhibitors would alter myocardial blood flow, microvascular reactivity, oxidative stress, and prostaglandin levels. Adult Yorkshire swine were divided into 3 groups: no drug (control, n = 7), a nonselective COX inhibitor (naproxen 400 mg daily, NAP, n = 7), or a selective COX-2 inhibitor (celecoxib 200 mg daily, CBX, n = 7). After 7 weeks, physiologic measurements were taken and tissue harvested. Animals in the CBX group demonstrated significantly higher blood pressure and rate-pressure product. The NAP and CBX groups demonstrated an increased microvascular contraction response to serotonin. The NAP group showed increased myocardial levels of thromboxane and lower levels of prostacyclin. Levels of protein oxidative stress were increased in the CBX group. Myocardial apoptosis was lowest in the NAP group. Immunoblotting demonstrated decreased vascular endothelial growth factor and phosphorylated endothelial nitric oxide synthase expression in the NAP and CBX groups. Myocardial tumor necrosis factor-α was increased in both treated groups. Immunostaining for thromboxane A2 synthase and receptor demonstrated expression within the vascular smooth muscle and no observable differences between groups. Nonselective and selective COX inhibition does not alter myocardial perfusion but results in altered myocardial and vascular physiology that may have implications regarding cardiovascular risk.

Figure 1

Microvessel reactivity. A) Microvessels harvested from the NAP and CBX group demonstrated increased contraction responses to serotonin as compared to the control (* p= 0.003 for control v. NAP and CBX). B) There was no difference in the microvascular responses to ADP, an endothelial dependent vasorelaxation agent.

Figure 2

Prostaglandin levels. A) Animals in the NAP group demonstrated increased levels of thromboxane A₂ (TXA₂) as compared to the control and CBX groups (* p= 0.003). The NAP group was found to have lower levels of prostacyclin I₂ (PGI₂) as compared to the control and CBX groups († p< 0.001). There was no difference between the control and CBX groups in TXA₂ or PGI₂ levels (p> 0.05 for both).

Figure 3

Arteriogenesis and Angiogensis. The control group demonstrated increased arteriole density by co-staining for CD-31 (an endothelial marker- green), smooth muscle actin (SMA- red), and DAPI (nuclear stain- blue) as compared to the NAP and CBX groups. A) Control group. B) NAP group. C) CBX group. All images are at 160×. D) Quantification of arteriolar density in the myocardium (p< 0.001 for control v. NAP and CBX). E) There was a trend toward decreased capillary density by CD-31 staining in the NAP and CBX groups (p= 0.12). F) Endothelial proliferation, assessed by co-staining with Ki-67 (proliferation marker) and CD-31, was significantly decreased in the NAP and CBX groups as compared to the control (p< 0.001).

Figure 4

Immunoblotting. A) Expression of the angiogenic protein VEGF was highest in the control group, and significantly decreased in both the NAP and CBX groups (* p= 0.03). B) The expression of the vasodilatory, pro-angiogenic protein phospho-eNOS was highest in the control group, and significantly downregulated in the drug treatment groups († p= 0.03). C) Myocardial TNFα levels were upregulated in the NAP and CBX groups (‡ p= 0.02). D and E) Levels of activated Akt and ERK, normalized to total levels, were increased in the NAP group (§p= 0.004 and ‖ < 0.001, respectively).

Figure 5

Levels of Oxidative stress and Apoptosis. Protein peroxidation was increased in the CBX group as compared to the control and NAP groups (* p< 0.001). Apoptotic cell death was decreased in the NAP groups vs. control and CBX groups († p= 0.007).

Figure 6

Immunostaining for thromboxane synthase (TS) and receptor (TR). Co-staining for TS and SMA (A) demonstrated localization in the vascular smooth muscle, and no obvious differences between groups; 1) control, 2) NAP, and 3) CBX. Immunostaining for TR and SMA (B) also showed localization in the vascular smooth muscle and no major differences in expression levels between groups; 1) control, 2) NAP, and 3) CBX.