Objective: Inflammation contributes to atherosclerotic plaque initiation and progression. Recent studies suggest that anti-inflammatory drugs such as cyclooxygenase-2 (Cox-2) inhibitors have anti-atherogenic effects. The current study was designed to investigate whether administration of a Cox-2 inhibitor to older apolipoprotein E deficient (apo E−/−) mice with established lesions alters the composition and increases the stability of the lesions. Methods and results: The Cox-2 inhibitor Celecoxib was administered in chow to 26-week-old, male, apo E−/− mice exhibiting advanced, unstable atherosclerotic lesions within the innominate/brachiocephalic artery. Mice administered Celecoxib had no significant changes in serum cholesterol or the average cross sectional area of atherosclerotic lesion in the innominate artery after 15 weeks of treatment in comparison to non-treated control mice. Histological analyses of sections of the innominate artery demonstrated no significant changes in the frequency of markers of advanced and unstable atherosclerotic plaques, including intra-plaque hemorrhage, vascular calcification, thinning of the fibrous cap, size of the necrotic core and macrophage content. There were also no significant differences in the content of Cox-2 within the lesions. Quantitative real time polymerase chain reaction with mRNA isolated from the aorta of each mouse revealed no significant changes in the expression of tissue factor and inducible nitric oxide synthase. However, mRNA levels for MCP-1 were increased fivefold following 15 weeks of treatment with Celecoxib in comparison to non-treated control mice. Conclusions: These data suggest that Celecoxib has no effect on the composition of advanced atherosclerotic lesions in older apo E−/− mice.
The rupture of vulnerable atherosclerotic plaques leads to acute coronary events and stroke [1,2]. The vulnerable plaque is generally composed of an atrophic fibrous cap, a lipid-rich necrotic core, dystrophic mineralization, and the accumulation of inflammatory cells such as macrophages and T-cells [3,4]. Macrophages are capable of degrading extra-cellular matrix by secreting active proteolytic enzymes. This could lead to weakening of the fibrous cap with plaque rupture and intra-plaque hemorrhage, and ultimately formation of occlusive thrombi [5–7]. Macrophage-derived cyclooxygenases (Cox) play a key role in inflammation by converting arachidonic acid into prostaglandins and thromboxanes in response to a large number of pro-inflammatory stimuli [8,9].
Cyclooxygenases exist in several different isoforms. Cox-1 is constitutively expressed in various tissues, including the stomach, kidney and platelets . Recently, a new variant of Cox-1 has been identified that is inhibited by acetaminophen and now referred to as Cox-3 . In contrast, Cox-2 is induced in response to growth factors and cytokines such as interleukin-1β, tumor necrosis factor-α and interferon-γ, suggesting a role for Cox-2 in inflammation and cell growth [8,11]. Cox-2 is expressed in human atherosclerotic lesions [12,13] and there is elevated expression in symptomatic versus asymptomatic plaques, linking Cox-2 to plaque destabilization . These observations have led to the hypothesis that selective Cox-2 inhibition might have beneficial effects on reducing vascular inflammation with subsequent plaque stabilization. However, clinical studies have so far failed to demonstrate any improvement in cardiovascular risk in patients treated with Cox-2 inhibitors [15,16]. In fact, clinical studies such as the VIGOR trial have raised concerns that selective Cox-2 inhibition may actually increase the risk of cardiovascular events by decreasing vasodilatory and anti-aggregatory prostacyclin production . In contrast, in a recent study by Burleigh et al., treatment of LDL receptor deficient mice with a selective Cox-2 inhibitor for 6 weeks reduced the macrophage content and size of early atherosclerotic lesions . The effects of Cox-2 inhibition on plaque stability were not determined in this model of early atherosclerotic lesions development.
We and others have recently shown that older apo E−/− mice develop advanced, unstable atherosclerotic lesions in the brachiocephalic/innominate arteries. These lesions have some of the morphologic features of advanced atherosclerotic lesions in humans, including intra-plaque hemorrhage and calcification [19,20] making this a useful model for investigating interventions that may alter plaque composition and increase plaque stability . In the present study, we investigated the effects of selective Cox-2 inhibition on plaque composition in older apo E−/− mice. Lesion progression, changes in morphologic features related to plaque instability, and inflammatory cell content were evaluated. Furthermore, the effects of inhibition of Cox-2 on the aortic expression of tissue factor (TF), monocyte chemoattractant protein-1 (MCP-1), and inducible nitric oxide synthase (iNOS) were investigated, since these pro-inflammatory factors play important roles in the atherogenic process.
Male apo E−/− mice on a C57BL/6J background (n = 35) were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Mice were fed a chow diet and water ad libitum throughout the study. At 26 weeks of age, Celecoxib (75 mg/kg/day) was added to the chow of the treatment group (n = 16), while the control group continued to receive regular chow (n = 19). This study was approved by the University of Washington Institutional Animal Care and Use Committee.
2.2. Animal sacrifice and preparation of arterial tissues
Following 15 weeks of Celecoxib treatment, the mice were sedated (Avertin; Aldrich, Milwaukee, WI, USA), blood was collected from the inferior vena cava, and the animals were sacrificed by exsanguination. The animals were perfused with 10 ml ice cold phosphate-buffered saline (PBS) at physiological pressure via the left ventricle. A suture was tied around each aorta distal to the branch of the left subclavian artery. The aortas were collected beginning from the suture and extending to the bifurcation of the femoral arteries and the surrounding fat and connective tissue were removed. The dissected tissues were immediately snap frozen in liquid nitrogen and stored at −80°C until further use. The animals were then perfused with 10% buffered formalin via the left ventricle for 4 min. The brachiocephalic artery from each animal (referred to as the innominate artery) was dissected out, embedded in paraffin and serially sectioned (5 μm). Every fifth section was stained with a modified Movat's pentachrome stain . To identify vascular calcification, adjacent slides were stained with the von Kossa stain . Total serum cholesterol was measured colorimetrically using a commercially available cholesterol oxidase enzymatic kit.
2.3. Real time reverse transcriptase polymerase chain reaction (RT-PCR)
Total RNA was extracted from each aorta into Tripure reagent (Boehringer Mannheim, Mannheim, Germany). Reverse transcription was performed with 1 μg total RNA isolated from each separate aorta. To determine the levels of expression of TF, MCP-1, iNOS and β-actin, real time PCR quantification (TaqMan) was performed (ABI Prism 7700 Sequence Detection System; PE Biosystems, Foster City, CA, USA). The primer sequences for TF, MCP-1, iNOS and β-actin were as follows. TF forward: 5′-CACTCATCATTGTGGGAGCAGTG-3′ and reverse: 5′-CGCGACGGGGTGTTCTT-3′; for MCP-1 forward: 5′-TCTGGGCCTGCTGTTCAC-3′ and reverse: GTGAATGAGTAGCAGCAGGTGAGT-3′, for iNOS forward: 5′-CAGCAGCGGCTCCATGAC-3′ and reverse: 5′-CGGCACCCAAACACCAAG-3′ and for β-actin forward: 5′-CCCTAAGGCCAACCGTGAAA-3′ and reverse 5′-ACGACCAGAGGCATACAGGGA-3′. The sequences for the probes were TF: 5′-CATATCTCTGTGCAAGCGC-3′, MCP-1: 5′-CTCAGCCAGATGCAGTTAA-3′, iNOS: 5′-TCATGCGGCCTCCTTTGAGCCCT-3′ and for β-actin: 5′-ATGACCCAGATCATGTTTGAGACCTTCAACAC-3′. The transcript for the constitutive gene product β-actin was used for data normalization.
Tissue sections of the innominate artery (5 μm) were stained with an anti-mouse macrophage antibody (Mac-3; Pharmingen, San Diego, CA, USA), anti-smooth muscle actin antibody (1A4; Dako, Carpenteria, CA, USA) and anti-Cox-2 (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturers’ protocols. The extent and distribution of the reaction products within the lesions were determined using computer-assisted morphometry (Image Pro; Media Cybernetics, Silver Spring, MD, USA).
2.5. Evaluation of plaque composition and plaque size
Two independent investigators blinded to the study protocol, evaluated each section for the frequency of features characteristic of plaque instability. These included: thickness of the fibrous cap (thin fibrous cap was defined as three or fewer cell layers), size of the necrotic core (a large necrotic core was defined as occupying more than 50% of the volume of the plaque), intra-plaque hemorrhage (defined as the presence of red blood cells independent of microvessels), calcification (defined on the basis of positive staining with the von Kossa stain), medial erosion (defined as enlargement of the media caused by infiltration of plaque components) and lateral xanthomas (defined as the presence of aggregates of macrophage-derived foam cells situated on the lateral margins of the plaques). These were recorded as binary outcomes and the frequency for each animal was determined. The cross sectional area of lesion in each section was determined using computer-assisted morphometry (Image Pro; Media Cybernetics) and is reported as maximum plaque area per animal.
2.6. Statistical analyses
All data were expressed as mean±S.E. Significant differences between means in serum cholesterol and lesion size were determined with the Student's two-tailed t-test. For analysis of gene expression and plaque morphology, groups were compared using the Mann–Whitney U-test.
3.1. Effects of Celecoxib on drug levels, serum cholesterol and maximum lesion area
Treatment with Celecoxib (75 mg/kg/day) led to a serum drug level of 537.6±185.1 ng/ml (Table 1). There were no significant differences in serum cholesterol levels between groups following 15 weeks of the drug treatment (Table 1). Furthermore, no significant differences were observed in the average maximum cross sectional area of atherosclerotic lesions in the innmominate arteries of the treated animals versus controls (Table 1).
|Treatment||Drug level||Total cholesterol||Lesion area|
|15 weeks Celecoxib||537.6±185.1||455.8±83.2||145,848±56,148|
|15 weeks control||n.d.||499.7±96.9||140,357±51,897|
|Treatment||Drug level||Total cholesterol||Lesion area|
|15 weeks Celecoxib||537.6±185.1||455.8±83.2||145,848±56,148|
|15 weeks control||n.d.||499.7±96.9||140,357±51,897|
* Values shown are the means±S.E.
3.2. Effects of Celecoxib on the aortic expression of tissue factor, MCP-1 and iNOS
Real time RT-PCR was performed with total mRNA isolated from the aortas of each of the apo E−/− mice. Following 15 weeks of Celecoxib treatment, there were no significant changes in expression of TF or iNOS (Fig. 1). However, a significant increase was observed in expression of MCP-1 in the aortas of the Celecoxib-treated mice relative to control mice (approximate 5.2-fold increase, P<0.05, Fig. 1).
3.3. Effects of Celecoxib on plaque composition
No significant differences in plaque composition were observed following 15 weeks of treatment with Celecoxib (Fig. 2). There were no significant differences in the frequencies of lateral xanthomas and intra-plaque hemorrhage. Hemorrhage was most frequently observed within the necrotic zones, in the lateral xanthomas and underneath thin fibrous caps in the lateral margins of the lesions in all of the animals (Fig. 3, panels A&B). There were also no significant differences in the frequencies of intra-plaque calcification, thin fibrous caps (Fig. 3, panel C), medial erosion, size of the necrotic cores or presence of perivascular inflammation (Fig. 3, panel D). Immunostaining for macrophages also demonstrated no significant differences in the macrophage content in the innominate arteries of the Celecoxib-treated and control mice (13±3.9% of lesion area in the control animals versus 19.2±5.2% in the Celecoxib-treated animals). In both the treated and control mice, macrophages were predominantly located in lateral xanthomas and diffusely distributed throughout the lipid-rich necrotic cores (data not shown). Cox-2 was expressed primarily by macrophages situated adjacent to the necrotic core of the lesions, in the lateral margins of the lesions as well as along the lumen of the lesions (Fig. 4). Again, there were no significant differences in the percentage of the lesion area immunostained with an antibody for Cox-2 within the lesions of the treated and control mice (23.75±7.9% of lesion area for the control animals versus 17.5±2.9% for the treated animals). No significant changes in smooth muscle cell content were observed.
The results of the present study suggest that chronic administration of Celecoxib to older apo E−/− mice does not inhibit the progression of advanced lesions in the innominate artery or alter the composition of the plaques. These data are consistent with the observations of Duffy et al. , who found that treatment with a selective Cox-2 inhibitor, SC-236, also had no effect on lesion development in younger apo E−/− mice. Furthermore, in a study from the same group inhibition of Cox-2 in a rat balloon angioplasty model had no effect on intimal hyperplasia . In contrast, Burleigh et al. reported that treatment of LDL receptor deficient mice with Rofecoxib inhibited lesion progression . The differences between the present observations and those of other groups may be related to the differences in mouse strain, to differences in the age of the animals and to differences in the site of the lesions evaluated and composition of the lesions at the time of administration of the Cox-2 inhibitors. It is likely that the anti-inflammatory effects of the drugs may be more evident in early lesions composed predominantly of leukocytes.
In comparison to early lesions at all vascular sites, there are reduced numbers of macrophages within the advanced lesions in the innominate arteries of the apo E−/− mice. In the older mice, intact macrophages are located predominantly within small lateral xanthomas . Furthermore, lesions in the innominate arteries of the older apo E−/− mice do not appreciably increase in size after about 36–42 weeks of age and this may also help explain the failure of the Celecoxib to inhibit progression in the present study . The differences in the effects of Cox-2 inhibitors on early versus established lesions may also be due to the progressive buildup and release of oxidation products in the advanced lesions. Eligini et al., have recently reported that oxidized phospholipids inhibit expression of Cox-2 in macrophages . Thus, Cox-2 expression may have already been reduced independently of the inhibitor and this would explain the similar levels of Cox-2 protein observed within the plaques of the treated and control mice and the failure of the inhibitor to alter the properties of the lesions.
The primary focus of the present study was to determine whether the Cox-2 inhibitor altered plaque composition and contributed to the stability of advanced lesions. The data clearly indicate that Celecoxib does not have any significant effects on altering the composition or stability of advanced lesions in the apo E−/− mice. One must be very cautious in equating these observations with mouse models of advanced atherosclerosis with the potential effects of Cox-2 inhibitors on advanced human atherosclerosis. However, as noted, recent human trials such as the VIGOR study suggest that Cox-2 inhibitors may also not alter advanced plaque composition in humans .
TF, MCP-1 and iNOS are key factors involved in the atherogenic process. It is likely that MCP-1 promotes the infiltration of monocytes into the vascular wall and their transformation into macrophage-derived foam cells at all stages of lesion development [28,29]. TF is the initiating factor in the extrinsic pathway of coagulation, can be activated by a variety of inflammatory stimuli [30,31], and appears to contribute to immediate thrombus formation after plaque disruption . Production of nitric oxide by iNOS expressed by macrophages and smooth muscle cells has an important bactericidal function but may also contribute to oxidant stress. It has also been shown to reduce the collagen content in advanced lesions and thus, could contribute to plaque instability [33,34].
The data from the current study demonstrate that chronic Celecoxib treatment of older apo E−/− mice does not reduce the aortic expression of these pro-inflammatory genes. In fact, expression of MCP-1 in the aorta was increased after treatment with the Cox-2 inhibitor. This may be due to the fact that cyclooxygenase inhibitors have been shown to attenuate the production of prostaglandin E2 (PGE2), a potent suppressor of MCP-1 expression both in vivo and in vitro [35–38]. In a recent study by Takayama et al. using cDNA microarrays, PGE2 inhibited lipopolysaccharide induced expression of pro-inflammatory cytokines, such as MCP-1 and interleukin-8 in human monocytes. Treatment of the cells with a selective Cox-2 inhibitor suppressed PGE2 production and restored the production of the pro-inflammatory cytokines . It is unclear why the increased expression of MCP-1 in the treated mice did not lead to an increase in the macrophage content of the lesions in the innominate arteries. However, hypercholesterolemia itself is a potent inducer of macrophage recruitment and there are additional mechanisms that are redundant with MCP-1 that may have accounted for our failure to observe an additive effect of the increased MCP-1 expression. It is also possible, although unlikely, that the increase in MCP-1 mRNA did not translate into an increase in MCP-1 protein.
We did not observe a significant difference in the content of Cox-2 within the lesions, despite the relatively high dosage of Celecoxib and sufficient plasma levels of the drug. In contrast, Burleigh et al., found reduced expression of Cox-2 in early lesions of young LDL receptor deficient mice after treatment with Rofecoxib . Differences in the strain and age of the mice as well as use of a different Cox-2 inhibitor may again account for the differences between these studies. However, as noted, a buildup of oxidation products in the advanced lesions as compared to early lesions may have already reduced expression of Cox-2 . It is also feasible that the capacity of the inhibitor to block Cox-2 expression was overwhelmed by the stimulatory effects of the many pro-inflammatory factors released into the necrotic zones of the advanced lesions. These include; IL-1, TNF-alpha, MIF-1, HSP-60, ceramide, peroxynitrite, and proteases, all previously shown to stimulate Cox-2 expression in vitro [40–44].
The current paradigm for the formation of occlusive thrombi and subsequent coronary events and stroke involves rupture of the fibrous cap, intra-plaque hemorrhage, and exposure of blood to thrombogenic plaque components including the lipid-rich necrotic core, inflammatory cells, and/or vascular calcification . In this regard, there is evidence that suggests that products of the cyclooxygenases may promote expression of matrix metalloproteinases (MMPs) in macrophages [14,46]. The proteolytic activity of these enzymes may contribute to plaque instability in advanced atherosclerotic plaques . In the current study however, Cox-2 inhibition resulted in no change in the frequency of intra-plaque hemorrhage suggesting that inhibition of cyclooxygenase 2 does not attenuate the proteolytic function of the MMPs in this model or that increased expression and activation of MMPs may not play a significant role in stimulating plaque rupture and fissures in the innominate arteries of older apo E−/− mice.
This is the first report of the effects of the inhibition of cyclooxygenase-2 on atherosclerotic plaque composition and plaque stability in older apo E−/− mice. Celecoxib had no effects on plaque composition in this mouse model of advanced atherosclerotic lesions suggesting that chronic inhibition of cyclooxygenase-2 may not have an anti-inflammatory effect on advanced atherosclerotic lesions.
We would like to thank Jerry Ricks, Carl P. Storey, Elizabeth P. Wallace, Monica I. Shelley and Annette Buttler for expert technical assistance. We would also like to thank Daniel Donnelly, Brad Keller, and Elaine Krul for help in the design and implementation of the project and for critical feedback. This study was supported by a non-restricted grant from Pharmacia, Inc. (now a division of Pfizer, Inc.) and by NIH HL 58954.