Disruption of Cholesterol Efflux by Coxib Medications and Inflammatory Processes: Link to Increased Cardiovascular Risk

Atherosclerosis: A Multifactorial Process

Atherosclerotic disease is the leading cause of morbidity and mortality in the United States and in many other Westernized countries.^24,25 Atherosclerosis can be considered a low-grade chronic inflammatory disease and a lipid disorder.^26,27 Dysfunctional lipid homeostasis plays a central role in the initiation and progression of atherosclerotic lesions, which form in large- and medium-sized muscular arteries. An early pathogenic feature of atherosclerosis is injury to the blood vessel endothelium. After the insult, injured endothelium is activated and exhibits increased expression of a variety of cell adhesion molecules and secretes chemoattractants to recruit lymphocytes/monocytes. ²⁸ Monocytes infiltrate the intima of the vessel wall and upon lipid loading, form a fatty streak. Ultimately, the fatty streak matures as fibrous tissue and smooth muscle cells overgrow and form a plaque containing a necrotic acellular lipid core. The atherosclerotic plaque narrows the arterial lumen and formation of a thrombus at the site of the plaque initiates arterial occlusion and ischemic coronary events.^29,30

Our current view of atherosclerosis as a complex process involving the endothelium as a metabolically active organ and a dynamic participant in the disease is in sharp contrast to an older perspective that looked upon the artery as a passive conduit. ³¹ Healthy endothelium controls vascular integrity and regulates vascular tone.

Cholesterol 27-Hydroxylase, ABCA1, and Reverse Cholesterol Transport

The path toward an understanding of the role of the vessel wall in cholesterol homeostasis has its origins in 1989 with the cloning and sequencing of the gene for cytochrome P450 cholesterol 27-hydroxylase (CYP27A1), a cholesterol metabolizing enzyme, by Dr David Russell's group. ³² Shortly thereafter, in 1991, Cali and Russell ³³ cloned the human 27-hydroxylase complementary DNA. Located in the inner mitochondrial membrane, the 27-hydroxylase catalyzes the first step in the acidic or alternative bile acid biosynthetic pathway in the liver. The enzyme is critical for oxidation of the side chain of cholesterol in connection with bile acid biosynthesis in the liver. Russell detected expression not only in the liver but also in several other organs and tissues, hinting at the possibility that this enzyme might have a role other than in bile acid biosynthesis.

A significant leap forward in our understanding of how the artery itself participates in atherosclerosis was achieved in 1994, when it was reported by our laboratory that the 27-hydroxylase enzyme is expressed in vascular endothelium ³⁴ and by Bjorkhem et al. ³⁵ that it is found in human monocytes and macrophages. This provided evidence that 2 cell types known to be present in the vessel wall were capable of metabolizing cholesterol. In peripheral cells such as endothelium and monocytes/macrophages, the 27-hydroxylase converts cholesterol into 27-hydroxycholesterol and 3β-hydroxy-5-cholestenoic acid (Fig. 1). These metabolites can then be moved out of the cell and transported to the liver for excretion in the bile, a process known as reverse cholesterol transport. ³⁶ Reverse cholesterol transport can remove cholesterol from the artery and from atherosclerotic plaque and is thus protective against atherosclerosis.^37,38 Extrahepatic cells, including those in the blood vessel wall, eliminate intracellular cholesterol via reverse cholesterol transport in a lipoprotein-dependent manner using high density lipoprotein (HDL) and also by an alternative reverse cholesterol transport mechanism involving the cholesterol 27-hydroxylase enzyme.^35,39 Specifically, in the 27-hydroxylase pathway, introduction of an oxygen moiety to the cholesterol molecule gives it a more hydrophilic character, thus enabling it to cross the plasma membrane and exit the cell more readily.^37,40 Unlike neutral sterols, which require extracellular lipoproteins for excretion from extrahepatic cells, cholestenoic acid only requires albumin as an acceptor for cellular export. ⁴¹

OPEN IN VIEWER

Figure 1

Cholesterol 27-hydroxylase catalyzes multiple oxidations at the C-27 position of the sterol side-chain of cholesterol. The cholesterol 27-hydroxylase converts cholesterol to the oxysterol 27-hydroxycholesterol. The cholesterol 27-hydroxylase can further oxidize 27-hydroxycholesterol to form 3β-hydroxy-5-cholestenoic acid.

27-Hydroxycholesterol, an important endogenous oxysterol that functions as a signaling molecule, is produced from the action of cholesterol 27-hydroxylase on cholesterol. 27-Hydroxycholesterol is the most abundant oxysterol in plasma, and it exhibits multiple antiatherogenic properties. ⁴² It is a potent suppressor of cholesterol biosynthesis. It inhibits cholesterol synthesis by blocking proteolytic activation of membrane-bound transcription factors known as sterol regulatory element-binding proteins (SREBPs), which are needed for activation of transcription of genes encoding 3-hydroxy-3-methylglutaryl coenzyme A reductase and multiple other enzymes involved in the cholesterol synthesis. ⁴³ The 27-hydroxycholesterol down regulates cell surface low density lipoprotein (LDL) receptors resulting in diminished receptor-mediated LDL uptake, ⁴⁴ inhibits macrophage foam cell transformation, and impedes smooth muscle proliferation. ⁴⁵ Mutations in the 27-hydroxylase enzyme lead to a rare, severe genetic disorder called cerebrotendinous xanthomatosis, in which there is progressive neurological impairment and premature coronary artery disease despite normal or low plasma cholesterol concentrations. ⁴⁶

In addition to the 27-hydroxylase, adenosine triphosphate-binding cassette transporter A1 (ABCA1), a macrophage plasma membrane protein, plays a key role in reverse cholesterol transport by mediating the cellular efflux of phospholipid and cholesterol to lipid-poor apolipoprotein (apo) A-I that is further lipidated to form HDL. ⁴⁷ The 27-hydroxycholesterol molecule interconnects the functions of 27-hydroxylase and ABCA1 by binding to and activating liver X receptors (LXRs) that in turn induce transcription of ABCA1.^48-50 Adenosine triphosphate-binding cassette transporter A1 in an adenosine triphosphate-dependent process shuttles excess cholesterol and lipids onto an HDL particle. This particle is then taken up by the liver for excretion as bile and other metabolites. Defects in the ABCA1 gene are the underlying cause of Tangier disease, which is characterized by an abnormally low level of HDL and a marked increase risk of coronary artery disease. ⁵¹

Thus, cholesterol 27-hydroxylase and ABCA1 are vital participants in reverse cholesterol transport, a defensive mechanism against cholesterol overload in the artery wall. ⁵²

Exploring Regulation of 27-Hydroxylase and ABCA1

To further characterize and elucidate the mechanisms that control the level of the 27-hydroxylase enzyme expressed in cells of the vessel wall, our laboratory took an innovative approach that applied our knowledge of the immune system as a guide. It is well established that the conditions of individuals diagnosed with autoimmune disorders such as systemic lupus erythematosus (SLE, lupus) and rheumatoid arthritis (RA) have an increased risk of developing atherosclerosis.^53,54 Based on this information, we initiated a study of the expression of 27-hydroxylase under conditions mimicking those found in these autoimmune disease states. Our cell culture models were human aortic endothelial cells and THP-1 monocytes/macrophages. THP-1 is a human monocytic cell line that can be easily differentiated into macrophages and that is commonly used as a model system to study the development of atherosclerosis.^55-57 Building on the knowledge that individuals with lupus and RA have increased levels of circulating cytokines and a high incidence of atherosclerotic disease, ⁵⁸ we investigated the effect of a number of cytokines on the expression of the 27-hydroxylase. We found that exposure of human aortic endothelial and THP-1 cells to interferon-gamma markedly decreased the 27-hydroxylase message and protein level. This was a significant finding, as this was the first report of an effect on the regulation of this gene in cells outside the liver. ⁵⁹ Immune complexes, which are present in excess in persons with autoimmune disorders, are also capable of inhibiting 27-hydroxylase expression by binding to complement component C1q and acting through the C1q receptor. ⁵⁹

Adenosine triphosphate-binding cassette transporter A1 is involved in reverse cholesterol transport, and as a consequence of impaired ABCA1 function, cholesterol may accumulate in macrophages. We found that immune complexes could decrease ABCA1 as well. ⁵⁹ Panousis et al. ⁶⁰ demonstrated the down-regulation of ABCA1 by interferon-gamma in murine peritoneal macrophages.

Accompanying the interference with reverse cholesterol transport resulting from diminution of 27-hydroxylase and/or ABCA1 is induction of lipid overload and foam cell formation. This physiologic manifestation is observed in THP-1 macrophages using oil red to specifically stain intracellular lipids. THP-1 macrophages treated with interferon-gamma or immune complexes exhibited not only compromised expression of the 27-hydroxylase but also manifested enhanced foam cell formation with the appearance of lipid globule-filled cells.

Cyclooxygenase-2 Inhibition Promotes Atherogenesis

The COX-2 inhibitors are highly effective anti-inflammatory analgesic drugs. This class of drugs works by preventing prostanoid formation and are used to relieve symptoms of osteoarthritis and RA and to treat dysmenorrheal. ⁶¹ In September of 2004, rofecoxib was withdrawn from the market by its manufacturer because there was an increased incidence of CV events in the Adenomatous Polyp Prevention on Vioxx trial. ⁶² The Adenomatous Polyp Prevention on Vioxx trial was a randomized, placebo-controlled, double-blind study in which rofecoxib was administered to patients to evaluate its effectiveness in preventing the recurrence of polyps in patients with a history of colorectal adenomas so as to decrease the risk of cancer. Celecoxib is now the only COX-2 inhibitor remaining on the market. As was expected, there has been a decreased use of this highly effective class of pain-relieving medications in clinical practice. ⁶³

Our laboratory has tackled the problem of explaining the elevated myocardial infarction and stroke incidence attributed to COX-2 inhibitors from the perspective of cholesterol metabolism. We postulated that the mechanisms responsible for increased CV risk likely involved lipids in some integral fashion. Cyclooxygenase-2 is widely expressed in atherosclerotic plaque and the arterial wall.^64,65 Cyclooxygenase-2 expression in the artery may play a defensive role against atherosclerosis by stimulating the production of heme oxygenase, an antioxidant, antiapoptotic, and potent anti-inflammatory enzyme that promotes vascular healing, and by limiting cholesterol accumulation.^66,67 A proatherogenic consequence of COX-2 inhibition may be compromised cholesterol balance in the artery wall. Cyclooxygenase-2 inhibitors interfere with the natural process of reverse cholesterol transport that requires products of the COX-2 enzyme for normal regulation. To understand this phenomenon, THP-1 human monocyte/macrophage cells were treated with increasing doses of a COX-2 specific inhibitor, NS398. This treatment resulted in a significant decrease in the 27-hydroxylase message. In THP-1 monocytes, exposure to 50 μmol/L of NS398 for 18 hours decreased 27-hydroxylase messenger RNA to 62.4% (2.2) of control (P < 0.001). Western blot analysis revealed a decrease in the expression of 27-hydroxylase protein as well. NS398 exerted similar effects on ABCA1 message and protein. Oil red O staining revealed extensive lipid deposition in macrophages treated with NS398, and the combination of NS398 and interferon-gamma caused massive lipid uptake (Fig. 2). There was not only a change in the expression of these reverse cholesterol transport proteins but also a compromise in the ability to prevent lipid accumulation within the cells. ⁶⁷ Thus, COX-2 inhibition not only promotes a proaggregatory state but also affects the ability of the cell to rid itself of excess cholesterol. Because arterial cholesterol buildup is a process that occurs for months or years, these events are consistent with the reports that COX-2-related myocardial infarction risk becomes apparent only after at least 9 to 12 months of treatment. ⁶⁸

OPEN IN VIEWER

Figure 2.

Effect of the combination of interferon-gamma and NS398 on foam cell transformation in lipid-loaded THP-1 macrophages. Representative photomicrographs at ×40 magnification of lipid-laden macrophages stained with oil red O. Control macrophages were exposed to acetylated LDL (50 μg/mL). B, Treated macrophages were exposed to acetylated LDL (50 μg/ml), interferon-gamma (500 U/mL), and NS398 (50 μmol/L). Interferon-gamma and NS398 induced a marked increase in foam cell formation in THP-1 macrophages after 48 hours.

Our study also revealed a concentration-dependent reduction in 27-hydroxylase and ABCA1 messages in THP-1 monocytes exposed to the nonselective COX inhibitor indomethacin. ⁶⁷ Indomethacin is associated with an increased risk of CV events. ⁶⁹

Mason et al.^70,71 have found a prooxidant effect of the sulfone COX-2 inhibitor rofecoxib that may contribute to the atherogenicity of that specific drug. They determined that rofecoxib alters the structure of lipid molecules, including human LDL, in a way that increases their susceptibility to oxidative modification, a key event leading to atheroma formation. The effect was blocked by an antioxidant and was not seen with celecoxib or NSAIDs.

Abrogating the Effect of COX-2 Inhibition on Cholesterol Metabolism

Once the effect of COX-2 inhibitors on reverse cholesterol transport had been defined, the next step was to neutralize the effect without losing analgesic or anti-inflammatory potency. The COX-2 pathway leads to the formation of multiple PGs and lipid mediators including PGE₁, PGE₂, PGD₂ and thromboxane A₂ (Fig. 3). ⁷² Our simple approach was to expose THP-1 cells to NS398, then reconstitute PGE₁, PGE₂, PGD₂ or thromboxane A₂, each individually, then analyze cellular expressions of 27-hydroxylase and ABCA1. Cholesterol transport was restored when cells were treated with either PGE₂ or PGD₂. It is well documented that PGD₂ has anti-inflammatory and proinflammatory effects ⁷³ ; thus, it may be possible to achieve COX-2 inhibition while sparing PGD₂ activity and retaining anti-inflammatory properties.

OPEN IN VIEWER

Figure 3.

Prostaglandin species produced via the PG synthesis pathway with their cognate cell surface receptors. Cyclooxygenase-1 and -2 isoenzymes act as cyclooxygenase and peroxidase. The product, an endoperoxide (PGH₂), is converted to PGD₂ and PGE₂ and to a thromboxane (TXA₂) and prostacyclin (PGI₂). Prostaglandin D₂ acts via 2 receptors, the prostanoid DP receptor and chemoattractant receptor homologous molecule expressed on T_H2 cells (CRT_H2). Prostaglandin D₂ readily undergoes dehydration to biologically active 15-deoxy-Δ^12,14-PGJ₂ (15d-PGJ₂), a natural peroxisome proliferator-activated receptor (PPAR)-γ ligand. Endopeptidases-1 to -4 (EP1-4) are the 4 subtypes of receptors for PGE₂. Prostacyclin and TXA₂ act on their corresponding receptors.

A second approach to counteracting the COX-2 inhibitory effects on cholesterol balance was to evaluate methotrexate (MTX), an anti-inflammatory agent commonly used in RA that is known to reduce CV risk. ⁷⁴ Methotrexate may provide a substantial survival benefit in RA, largely by reducing CV mortality. ⁷⁵ Many of the anti-inflammatory effects of MTX are mediated by the endogenous autocoid adenosine.^76,77 Our previous studies have shown that adenosine A_2A receptor (A_2AR) ligation increases expressions of both 27-hydroxylase and ABCA1. ⁷⁸ Our most recent work has revealed that in THP-1 monocytes, both MTX and CGS-21680 (a specific A_2AR ligand), acting through activation of the adenosine A_2AR, were able to overcome the reduction in 27-hydroxylase and ABCA1 caused by exposure to the selective COX-2 inhibitor NS398. Methotrexate also ablated the NS398-induced increase in foam cell transformation in THP-1 macrophages that had been cholesterol loaded with acetylated LDL. Although 72.7% (4.9) of THP-1 macrophages became foam cells after NS398 treatment, combined treatment with NS398 and MTX yielded only 36.3% (3.2) foam cells (n = 3, P < 0.001). ⁷⁹ Reversal of inhibitory effects on 27-hydroxylase and ABCA1 in the presence of MTX further support the antiatherogenicity of this drug. Because the phenomenon is mediated via the adenosine A_2AR, A_2AR ligation may provide a suitable mechanism for the development of a promising treatment paradigm for the in vivo reversal of atherogenic COX-2 effects.

In summary, cellular cholesterol homeostasis is maintained by feedback repression of genes involved in cholesterol synthesis and supply. Macrophage 27-hydroxylase and ABCA1 counteract intracellular cholesterol accumulation by promoting cholesterol efflux from peripheral cells. We demonstrated that COX inhibition (selective or nonselective) reduced these reverse cholesterol transport proteins. We further demonstrated that addition of PGE₂ or PGD₂ to cells treated with a COX-2 inhibitor restored normal levels of 27-hydroxylase and ABCA1. Adenosine A_2AR ligation up regulated both 27-hydroxylase and ABCA1 expressions and prevented COX-2 inhibitor-mediated down-regulation of transcription for both these proteins (Fig. 4).

OPEN IN VIEWER

Figure 4.

Proposed effects of COX inhibitors and adenosine A_2AR agonists on cholesterol efflux via reverse cholesterol transport pathways. Cholesterol 27-hydroxylase catalyzes the conversion of cholesterol to the polar oxysterol 27-hydroxycholesterol. 27-Hydroxcholesterol exits the cell more easily than cholesterol either as a component of HDL or, after further oxidation to 3β-hydroxy-5-cholestenoic acid by the 27-hydroxylase, without lipoprotein. 27-Hydroxycholesterol is a potent endogenous ligand for the nuclear receptor LXR that in turn activates transcription of ABCA1. Adenosine triphosphate-binding cassette transporter A1 transfers 27-hydroxycholesterol and cholesterol to lipid-poor acceptor apolipoproteins, such as apoA-I and eventually to HDL. High density lipoprotein delivers cholesterol to the liver for conversion to bile acids leading to biliary excretion from the body. Prostaglandins E₂ and D₂, acting via 1 or more specific receptors, maintain normal homeostatic levels of 27-hydroxylase and ABCA1 in monocytes/macrophages. When production of PGs is reduced by COX inhibitors, 27-hydroxylase and ABCA1 levels fall, and cholesterol accumulation increases. The addition of exogenous PGE₂ or PGD₂ restores 27-hydroxylase and ABCA1 expressions. Specific activation of the adenosine A_2AR subtype by the purine nucleoside adenosine enhances expression of both 27-hydroxylase and ABCA1 via mechanisms that have not yet been elucidated.

Cyclooxygenase-2 Effects on Foam Cell Formation Amplified by Lupus Plasma

The population most likely to need coxibs for pain relief on an ongoing basis may ironically consist of those most at risk for CV disease: the elderly, and those with autoimmune inflammatory conditions.^61,80 To determine whether the combination of a coxib and plasma derived from patients with SLE would impact foam cell transformation, we exposed THP-1 macrophages to 50% SLE patient plasma with and without the addition of the COX-2 selective inhibitor NS398 (10 μmol/L) then incubated the cells for 48 hours under cholesterol-loading conditions (acetylated LDL, 50 μg/mL). Cyclooxygenase-2 inhibition in the presence of 50% lupus plasma increased foam cell transformation almost 2-fold versus lupus plasma alone (42.8% [5.3] for lupus plasma vs 82.7% [2.8] for lupus plasma combined with NS398, P < 0.001). Enhanced foam cell formation in the presence of SLE plasma and NS398 is reversed by PGD₂ or PGE₂, but not by PGI₂ (unpublished results). The atherogenicity of COX-2 inhibition may result from both blockade of antiaggregatory PG production and concomitant dysregulation of cholesterol efflux mechanisms. This may be further augmented by inflammatory mediators in SLE plasma, posing an exaggerated risk to patients with SLE treated with COX-2 inhibitors. Notably, disrupted cholesterol balance may be corrected by specific PGs, suggesting possible future therapeutic modalities for COX inhibition that would support an antiatherogenic PG state.