ApoE−/− PGC-1α−/− Mice Display Reduced IL-18 Levels and Do Not Develop Enhanced Atherosclerosis

Background Atherosclerosis is a chronic inflammatory disease that evolves from the interaction of activated endothelial cells, macrophages, lymphocytes and modified lipoproteins (LDLs). In the last years many molecules with crucial metabolic functions have been shown to prevent important steps in the progression of atherogenesis, including peroxisome proliferator activated receptors (PPARs) and the class III histone deacetylase (HDAC) SIRT1. The PPARγ coactivator 1 alpha (Ppargc1a or PGC-1α) was identified as an important transcriptional cofactor of PPARγ and is activated by SIRT1. The aim of this study was to analyze total PGC-1α deficiency in an atherosclerotic mouse model. Methodology/Principal Findings To investigate if total PGC-1α deficiency affects atherosclerosis, we compared ApoE−/− PGC-1α−/− and ApoE−/− PGC-1α+/+ mice kept on a high cholesterol diet. Despite having more macrophages and a higher ICAM-1 expression in plaques, ApoE−/− PGC-1α−/− did not display more or larger atherosclerotic plaques than their ApoE−/− PGC-1α+/+ littermates. In line with the previously published phenotype of PGC-1α−/− mice, ApoE−/− PGC-1α−/− mice had marked reduced body, liver and epididymal white adipose tissue (WAT) weight. VLDL/LDL-cholesterol and triglyceride contents were also reduced. Aortic expression of PPARα and PPARγ, two crucial regulators for adipocyte differentitation and glucose and lipid metabolism, as well as the expression of some PPAR target genes was significantly reduced in ApoE−/− PGC-1α−/− mice. Importantly, the epididymal WAT and aortic expression of IL-18 and IL-18 plasma levels, a pro-atherosclerotic cytokine, was markedly reduced in ApoE−/− PGC-1α−/− mice. Conclusions/Significance ApoE−/− PGC-1α−/− mice, similar as PGC-1α−/− mice exhibit markedly reduced total body and visceral fat weight. Since inflammation of visceral fat is a crucial trigger of atherogenesis, decreased visceral fat in PGC-1α-deficient mice may explain why these mice do not develop enhanced atherosclerosis.


Introduction
Atherosclerosis is a chronic inflammatory disease that results from interaction between activated endothelial cells, modified lowdensity lipoproteins (LDL), monocyte-derived macrophages, T cells, and the vessel wall.Activated endothelial cells express adhesion molecules that attract and recruit blood monocytes and lymphocytes.Upon binding to the endothelial layer, these monocytes transmigrate into the subintimal space, and differentiate into macrophages.Plaque macrophages interact with lymphatic cells, mainly T cells, ingest modified LDL via scavenger receptors and become foam cells, thereby promoting plaque formation [1].
The phenotype of PGC-1a knock-out mice underlines the central role of this transcription cofactor in homeostatic control of metabolism: they are leaner than wild-type (WT) littermates, have markedly reduced body fat content, and are resistant to dietinduced obesity, hence protected from developing insulin resistance and impaired glucose tolerance [11].This difference is explained by their CNS-linked hyperactivity and is not a consequence of altered food intake [11].
Overexpression of PGC-1a in human aortic smooth muscle and endothelial cells in vitro has been shown to prevent reactive oxygen species (ROS) production and NAD(P)H oxidase activity, with subsequently reduced NF-kB activity and lower expression levels of MCP-1 and VCAM-1 [12], which are important triggers of inflammation and atherosclerosis.Moreover, PGC-1a overexpression in endothelial cells prevented alpha-linoleic acid-induced ROS formation in vitro and improved endothelial dysfunction in aortic rings ex vivo [13].
The following studies suggest a link between PGC-1a and atherogenesis at the clinical level: Xie et al. reported a correlation between PGC-1a polymorphism and hypertension [14], and Zhang et al. showed an association between PGC-1a polymorphism and the prevalence of coronary artery disease [15].

Discussion
Our data show that ApoE 2/2 PGC-1a 2/2 and ApoE 2/2 PGC-1a +/+ mice do not differ with regard to atherosclerosis, features of plaque vulnerability, expression of VCAM-1, and T cells number.Increased expression of ICAM-1 or CD68-positive cells in plaques of ApoE 2/2 PGC-1a 2/2 do not appear to play a substantial role as they do not affect plaque size.Importantly, the double knockout mice are leaner, have lighter liver and epididymal fat, and less cholesterol and triglycerides in VLDL and LDL subfractions.In addition, aortic expression of PPARa and PPARc as well as some of their target genes is reduced in ApoE 2/2 PGC-1a 2/2 mice.This phenotype is in line with the first study that described the phenotype of PGC-1a 2/2 mice, which also have markedly reduced body fat content [11].Because visceral (epididymal) WAT inflammation contributes to disease progression [18], it is not astonishing that we observed no difference in plaque lesions between ApoE 2/2 PGC-1a 2/2 and ApoE 2/2 PGC-1a +/+ mice.Beyond this notion, our data propose that total PGC-1a deficiency may rescue an increased atherosclerotic phenotype because of the reduced paracrine effects mediated by the visceral fat.
The lower aortic expression of PPARa and PPARc as well as of PPAR target genes proposes that the function of these two PPARs is suppressed in ApoE 2/2 PGC-1a 2/2 mice.Interestingly, both PPARa and PPARc can exert anti-atherogenic functions in the  [23], and transplantation of PPARc-deficient bone marrow into recipient LDL-R knockout mice enhanced atherosclerosis [24].One of the main atherogenic targets of PPARc is LXRa [24,25], whose expression was not changed between ApoE 2/2 PGC-1a 2/2 and ApoE 2/2 PGC-1a +/+ aortic lysates.
Reduced expression of Rarres2 (chemerin), Serpine1 (PAI-1), and IL-18 in visceral adipose tissue could be sufficient to avoid increased atherogenesis.Rarres2 is associated with white adipose tissue inflammation and promotes mobilization and chemotaxis of dendritic cells and macrophages [26,27].While its correlates with inflammatory markers, such as C-reactive protein, it does not predict atherosclerosis in humans [28].Nevertheless, an atherogenic contribution of Rarres2 cannot be excluded.
PAI-1 is an anti-fibrinolytic enzyme and has beneficial and deleterious effects in atherogenesis.For example, PAI-1-deficient mice showed attenuated neointima formation after perivascular cuff-induced injury [29], and local PAI-1 overexpression prevented the development of abdominal aortic aneurysm [30].On the other hand, PAI-1 levels are elevated in various cardiovascular diseases and associated with atherothrombosis [31].
The lowest expression of the tested cytokines in the visceral WAT of ApoE 2/2 PGC-1a 2/2 mice was observed for IL-18.IL-18 is a pro-atherogenic cytokine: Overexpression of IL-18 binding protein and direct injection of recombinant IL-18 accelerate atherogenesis, whereas IL-18 deficiency diminishes plaque formation in ApoE 2/2 mice [21,22,32,33].Furthermore, elevated levels of plasma IL-18 are observed in patients with previous myocardial infarction and are associated with the extent of coronary atherosclerosis [34,35].We did not only observe a reduced expression of IL-18 in epididymal WAT, but also in aortic tissue and plasma samples of ApoE 2/2 PGC-1a 2/2 mice.It is conceivable that the lower expression of IL-18 alone is sufficient to avoid an acceleration of atherogenesis in our ApoE 2/2 PGC-1a 2/2 mouse model.
Cfd encodes adipsin, the mouse homolog of human complement factor D, which is a serine protease that cleaves factor B in the alternative complement pathway, and it is secreted at high levels in adipose tissue [43,44,45].While adipsin expression is increased in catabolic conditions such as fasting, it is down-regulated in different models of genetic and acquired obesity [46].In line with these observations, epididymal WAT expression of adipsin was higher in ApoE 2/2 PGC-1a 2/2 compared to ApoE 2/2 PGC-1a +/+ mice.Expression of adipsin and other components of the alternative complement pathway correlate with atherosclerosis [47], suggesting that the elevation of adipsin in ApoE 2/2 PGC-1a 2/2 provides a pro-atherogenic contribution.
Atherosclerosis is a disease combining the complexity of lipid/ lipoprotein and inflammatory/immune disorders [48].Since PGC-1a is affecting these two important atherogenic systems, it is difficult to dissect the functions of this enzyme in the chosen animal model.For example, the reduced body weight and VLDL/ LDL-cholesterol and triglyceride contents as well as the diminished expression of IL-18 are certainly anti-atherogenic, whereas the increased expression of adipsin may play a pro-atherogenic role in ApoE 2/2 PGC-1a 2/2 mice.Further studies using tissuespecific PGC-1a knockout or overexpression will be necessary to address these questions in more detail.

Ethics Statement
All animal procedures were approved by the local animal committee (Kantonales Veterina ¨ramt Zu ¨rich, protocol no.171/ 2006) and performed in accordance with our institutional guidelines.Immunohistochemistry 5 mm serial cryosections from the aortic sinus were stained with rat anti-CD68, rat anti-CD3 (Abcam), rat anti-VCAM-1 (BD Biosciences), rat anti-ICAM-1 (Serotec), or oil-red O (ORO).Thoraco-abdominal aortae were fixed with 4% paraformaldehyde and plaques stained with ORO for en face analysis.Collagen, fibrous cap thickness, and necrotic core size were analyzed on Elastica van Gieson (EVG)-stained cryosections of the aortic sinus as described [50,51].Means were taken from n = 10 different mice evaluating 6 serial cryosections/tissue from each mouse.

RNA and protein analysis
Total RNA isolated from proximal aortae was extracted with TRIZOL (Invitrogen), reverse transcribed with Ready-To-Go You-Prime First-Strand Beads (GE Healthcare), and the cDNA (n$9 per genotype) quantified by qPCR using SYBR Green JumpStart Taq ReadyMix (Sigma-Aldrich).Primer sequences can be found in the supplemental Table S1.

IL-18 and CXCL16 ELISA
Quantification of IL-18 and CXCL16 in plasma of mice where performed with Mouse IL-18 Platinum ELISA kit (Bender MedSystems) and Mouse CXCL16 ELISA kit (RayBiotech) according to the manufacturers instructions.Plasma was diluted 1:2 for the IL-18, and 1:32 for the CXCL16 ELISA assay.

Cholesterol, triglycerides, and lipoprotein subfractioning
Total plasma cholesterol and triglycerides were quantified using Infinity Cholesterol TR13421 and Infinity Triglycerides TR22421 (Thermo Electron Cooperation), respectively.The lipid distribution in plasma lipoprotein fractions was assessed by fastperformance liquid chromatography gel filtration with a Tricorn Superose 6 10/300 GL column (GE Healthcare) [52].

Statistical analyses
Data are presented as mean 6 SEM.The en face ORO quantification was analyzed using a non-parametric Mann-Whitney U t-test.Statistical significance of differences of all other experiments was calculated using an unpaired Student's t-test.
Significance was accepted at the level of p,0.05.