High levels of serum cholesterol and low density lipoprotein (LDL) have been associated with the genesis of atherosclerosis, a leading cause of cardiovascular diseases that affect a large number of people all over the world. It has become clear that the pathological conditions associated with an excess level of LDL are actually caused by the oxidized products of LDL 1. More specifically, cholesterol oxides (also termed oxysterols), the oxygenated derivatives of cholesterol, appear to be the major cytotoxic components in oxidized LDL 23. These compounds have a hydroxyl- or a keto- group on the cholesterol molecule. Examples of cholesterol oxides include: 7-OH-, 7-keto-, 19-OH-, 22-OH- and 25-OH-cholesterol (see 4 for a recent review).

Tissue damage caused by cholesterol oxides has been the subject of many studies. Cholesterol oxides are cytotoxic to a variety of cell types 5. These compounds can damage endothelial cells, smooth muscle cells and fibroblasts, all of which are major components of the arterial wall. Cholesterol oxides are also toxic to cells derived from the immune system, such as macrophages, thymocytes, lymphoma cells and leukemic T-cells. Pharmacological agents capable of reducing cholesterol oxide toxicity are yet to be discovered.

The mode of cell death caused by cholesterol oxides has generated much interest recently. There is evidence showing that programmed cell death (PCD, also known as apoptosis) occurs in some cells treated with cholesterol oxides 678. PCD is a special type of cell death that can be induced by growth factor deprivation or toxins 91011. By using the PCD induced by nerve growth factor deprivation in sympathetic neurons as an example, the cell death process begins with the generation of reactive oxygen species followed by a significant decrease in glucose uptake, RNA and protein synthesis, activation of the immediate early protein, c-jun, and release of cytochrome c from mitochondria. Cell death is executed by the activation of a family of proteases termed caspases. Morphologically, cells dying of PCD appear atrophic and exhibit condensed nuclei, which can be stained with nuclear stains (bisbenzimide or propidium iodide), or by the TUNEL assay. DNA extracted from cells dying of PCD is often fragmented and shows a characteristic ladder-type pattern on the agarose gel upon electrophoresis.

MAP kinases (Mitogen-Activated Protein kinases) have very important roles in PCD. The MAP kinase pathway consists of 3 major parallel pathways designated as the ERK (p42/44, Extracellular signal-Regulated Kinase), SAPK/JNK (p46, Stress-Activated Protein Kinase/ c-Jun NH2-terminal Kinase) and p38 pathways 121314. Activation of each pathway above involves phosphorylation of a number of upstream and downstream family members of the pathway. During growth factor deprivation induced PCD in neuronal PC12 cells, there is a sustained activation of the JNK and p38 pathways while the activities of ERK pathway is inhibited. It was proposed that a dynamic balance of JNK-p38 and ERK can determine if the neuronal cell will live or die 15. Similar interplay of the MAP kinases was shown in the PCD of non-neuronal cells 16. It is known that oxidized LDL can cause activation of MAP kinases 1718. Whether cholesterol oxides, the main toxic components in oxidized LDL, can induce MAP kinase activation, and whether MAP kinase activation has a role in cholesterol oxide induced PCD remain to be investigated, and is the subject of this study.

Peroxisome proliferator-activated receptors (PPARs) belong to a group of nuclear receptors which includes steroid, retinoid, thyroid hormone receptors and others 192021. There are three types of PPARs: PPARα is found predominantly in the liver, heart, kidney, brown adipose and stomach mucosa, and is important for lipid catabolism. PPARγ is found in adipose tissues, and is important for adipogenesis. PPARβ is found in most tissues, but its role is less well-defined.

PPARγ is expressed in atherosclerotic lesions 22 and is important in the process of atherogenesis. Depending on the experimental system and the particular question asked, PPARγ can be involved with pro-atherogenesis or anti-atherogenesis. For example, PPARγ is involved in a positive feedback mechanism that induces the formation of foam cells 23. On the other hand, activation of PPARγ was shown to inhibit monocyte 24 and macrophage 25 inflammatory responses by preventing the activation of nuclear transcription factors, such as NF-kB (nuclear factor-kappa B), AP-1 (activator protein-1) and STAT 1 (signal transducer and activator of transcription 1). Since inflammation plays an important role in atherogenesis, this anti-inflammatory effect of PPARγ should help to reduce the risk of atherogenesis. Development of PPAR agonists that retain the antiatherogenic activity but dismiss the proatherogenic activities will be of great value 26.

Many naturally occurring or synthetic compounds are agonists for the PPARs 2021. For example, a number of prostaglandins are agonists for PPARγ . The prostaglandin, 15-deoxy-delta 12, 14-PGJ₂ (15d-PGJ₂), is the most potent endogenous PPARγ agonist known. On the other hand, many synthetic compounds such as the non-steroid anti-inflammatory drugs (NSAIDs) are PPARα and PPARγ agonists 27. Since these are agents used clinically for other purposes, it is important to determine whether interaction of these agents with PPARs could modulate atherogenesis in a positive or negative manner.

While most research regarding cholesterol oxides are focused on the cardiovascular system, it should be noted that the nervous tissue has the ability to generate cholesterol oxides, which may exert toxic effect on cells and tissues of the nervous system. For example, it has been shown that synaptosomes and mitochondria prepared from CNS tissue can convert cholesterol to cholesterol oxides, which may play an important role in brain tissue damage during oxidative stress 28. Consistent with this notion, we have shown cholesterol oxides are toxic to neuronal PC12 cells 2930 and neurons derived from sympathetic ganglia 31, retina 32 and cerebellum 33.

Microglia in the central nervous system express scavenger receptors and are involved in the oxidized LDL uptake and metabolism 34. These cells are, thus, the macrophage-equivalent in the central nervous system 35. Results from this laboratory showed that cholesterol oxides can cause programmed cell death in microglia 7, which suggests that microglia, similar to macrophages, can be a potential target for oxidized LDL toxicity.

Based on the observations that: 1. cholesterol oxides cause PCD but the involvement of MAP kinases remains to be investigated; 2. PPARs are important in atherosclerosis but the effect of PPAR activation on cholesterol oxide induced PCD has not been determined, this study was designed to test if a dysregulation of MAP kinases was associated with the cholesterol oxide induced PCD, and to determine if PPAR activation could modulate the cholesterol oxide induced PCD.

Cholesterol oxides are formed in cholesterol-containing food (powdered milk, cheese, egg products) during storage 3839 or during cholesterol catabolism in animals 40. The level of cholesterol oxide in meat product increases significantly during the cooking process 39. Animals can take up cholesterol oxides in a way similar to the absorption of cholesterol. Cholesterol oxides given exogenously can be cleared from plasma rapidly and be widely redistributed in different parts of the body 41. Cholesterol oxides and authentic cholesterol are transported within the cell by a similar mechanism, but the transport is more efficient for cholesterol oxides 42. High levels of cholesterol oxides have been observed in hypercholesterolemic animals. For example, the plasma level of 7-OH-cholesterol increases from 15 μM to approximately 200 μM in rabbits fed with high cholesterol diet for 6 weeks 43.

Cholesterol oxide treatment can lead to alterations of cellular functions. These compounds can activate some nuclear receptors including steroidogenic factor 1 44 and LXR, which may be partially responsible for their biological activities 4546. Cholesterol oxides are inhibitors of HMG-CoA reductase, the key enzyme of cholesterol biosynthesis pathway 47. The ability to bind to an intracellular receptor, oxysterol binding protein, may be responsible for their inhibition of cholesterol biosynthesis 4849. It should be noted that addition of cholesterol does not prevent cholesterol oxide induced cell death, which leads to the suggestion that inhibition of cholesterol synthesis is not the cause of cholesterol oxide induced PCD 5051. The critical events that result in cholesterol oxide induced cytotoxicity remain to be determined.

Results from this study indicated that 25-OH-cholesterol was very cytotoxic to microglia, such that a 2-day treatment with 5 μM 25-OH-cholesterol reduced cell viability to 5–10% of controls. In contrast, 7-β-OH-cholesterol was much less toxic to microglia (Fig. 1,2). Selective toxicity of cholesterol oxides toward a particular cell type was also observed in our previous studies 29313233. Further experiments with Western blot analyses indicated that the cytotoxicity of cholesterol oxides was associated with their abilities to cause c-jun activation, as evidenced by the increase in activated c-jun, i.e., phospho-c-jun, in 25-OH-cholesterol treated cells but not in 7-β-OH-cholesterol treated cells (Fig. 3). There was a dose-dependent increase in c-jun activation in cells treated with 25-OH-cholesterol (Fig. 4).

It has been suggested that a dysregulation of MAP kinases is responsible for the growth factor deprivation 15 or chemotherapeutic agent 16 induced PCD. Consistent with this idea, results from this study suggest that an unfavorable balance of the JNK and ERK causes the cell death observed in cholesterol oxide treated microglia. Cells treated with 25-OH-cholesterol showed a time-dependent increase in c-jun protein as well as phospho-c-jun (Fig. 5). This c-jun activation was apparent at 6 hours after treatment. In contrast, 25-OH-cholesterol caused a slight, initial increase in phospho-ERK, which decreased to a non-detectable level 6 hours after treatment. The change in p38 and phospho-p38, on the other hand, was not as apparent. It was reported that rabbits fed with high-cholesterol diet exhibited high JNK protein levels and JNK activity in their atherosclerotic lesions, and oxidized LDL caused JNK activation in cultured smooth muscle cells 18. Based on this study, there is a possibility that cholesterol oxides in oxidized LDL were responsible for those previous observations.

Activation of the JNK pathway can lead to the phosphorylation of downstream members including c-jun that belongs to the AP-1 transcription factor family 1252. Because results from this study indicated that JNK pathway was activated by 25-OH-cholesterol, one could expect that AP-1 would be activated in cells treated with cholesterol oxides. Consistent with this idea, it was shown that LDL could induce AP-1 activity in endothelial cells 53 and cholesterol oxide treatment could lead to increased AP-1 DNA binding activity and AP-1 dependent transcription in other cell types 5455. Results from this study, thus, complemented and extended the findings of those reports.

The PPARγ agonist, 15d-PGJ₂, functions as a general inhibitor of microglial functions 565758. For example, this agent inhibits the expression of inducible nitric oxide synthase, tumor necrosis factor-α, interleukin-1 β and major histocompatibility complex class II in activated microglia. Inhibition of the signal transducer and activator of transcription 1 (STAT-1) and the nuclear factor kB (NF-kB) may be partially responsible for the inhibitory effects. Results from this study further indicated that PPAR agonists could reduce c-jun activation (Fig. 6). Given the suggestion that an up-regulation of c-jun is a cause of PCD 1516, a reduction in c-jun activation by PPAR agonists should attenuate 25-OH-cholesterol induced cytotoxicity. Consistent with this idea, results from this study indicated that 15d-PGJ₂, indomethacin and WY14643 could attenuate the cytotoxicity caused by 25-OH-cholesterol to various degrees (Fig. 7,8,10). Since the protective effect of WY14643 could be observed only at concentrations higher than 200 μM, there is a possibility that this agent acted on the PPARγ receptor to achieve the protection. Similar protective roles of PPAR agonists have been shown in other nervous tissues. For example, PPARγ agonists were shown to protect cytokine-induced PCD in cerebellar granule cells 59. These agents could also block the neurotoxic effects caused by β-amyloid activated microglia, and thus might be beneficial in the treatment of Alzheimer's disease 60. In non-neuronal tissues, PPARα activation was shown to suppress both spontaneous rat hepatocyte PCD and the PCD induced by transforming growth factor-β1 61.

Treatment of microglia with 25-OH-cholesterol caused an induction of c-jun, phospho-c-jun and a reduction in cellular viability. PPAR agonists reduced the 25-OH-cholesterol induced c-jun activation and attenuated the cytotoxicity. This group of compounds may be useful in future development of pharmacological agents against cholesterol oxide induced cytotoxicity.

15d-PGJ₂, 15-deoxy-delta 12, 14-PGJ₂; AP-1, activator protein-1; ERK, p42/44, Extracellular signal-Regulated Kinase; LDL, low density lipoprotein; MAP kinases, Mitogen-Activated Protein kinases; MTT (3-(4,5-dimethylthiazole-2yl)-2,5-diphenyltetrazolium bromide); PCD, programmed cell death; PPARs, peroxisome proliferator-activated receptors; SAPK/JNK, Stress-Activated Protein Kinase/ c-Jun NH2-terminal Kinase.