Selective Interleukin-12 Synthesis Defect in 12/15-Lipoxygenase-deficient Macrophages Associated with Reduced Atherosclerosis in a Mouse Model of Familial Hypercholesterolemia*

Targeted gene disruption or overexpression of 12/15-lipoxygenase in mice on the genetic background of apolipoprotein E or low density lipoprotein-receptor (LDL-R) deficiency has implicated 12/15-lipoxygenase in atherogenesis. The data support indirectly a role for 12/15-lipoxygenase in the oxidative modification of low density lipoprotein. In this study we set out to explore other potential mechanisms for 12/15-lipoxygenase in atherosclerosis using apolipoprotein B mRNA editing catalytic polypeptide-1/LDL-R double-deficient mice, a model highly related to the human condition of familial hypercholesterolemia. 12/15-Lipoxygenase deficiency in this strain led to ≈50% decrease in aortic lesions in male and female mice at 8 months on a chow diet in the absence of cholesterol differences. While studying 12/15-lipoxygenase-deficient macrophages in culture, we discovered a remarkable selective defect (75–90% decrease) in interleukin-12 production but not in tumor necrosis factor-α or nitric oxide release, in response to lipopolysaccharide in the presence or absence of interferon-γ priming. The lipopolysaccharide/interferon-γ response was associated with a 33–50% decrease in nuclear interferon consensus sequence-binding protein, which is consistent with interferon consensus sequence-binding protein containing protein complex-dependent regulation of the interleukin-12 p40 gene. The decrease in interleukin-12 production was recapitulated in vivo in mouse aortas of the triple knockout group and was reflected in a marked decrease in interferon-γ expression. The data provide support for a novel mechanism linking the 12/15-lipoxygenase pathway to a known immunomodulatory Th1 cytokine in atherogenesis.

formation, monocytes are recruited and transformed into lipidenriched foam cells, and a complex network of regulatory molecules and cellular interactions amplify the pathophysiological process. Substantial evidence supports an active role for Th1derived cytokines like interleukin-12 (IL-12) 1 and interferon-␥ (IFN-␥) in regulating the inflammatory response during early atherogenesis (2)(3)(4). Th1 cells infiltrate lesions and contribute to the autoimmune responses to oxidized LDL in lesions (5).
Several mouse models of atherosclerosis have been used to explore pathogenesis of the disease (6 -8). Recent data have established a role for 12/15-lipoxygenase  in promoting atherogenesis in apolipoprotein E (apoE)-deficient and low density lipoprotein-receptor (LDL-R)-deficient mouse models (9 -14). Whereas LDL oxidation via 12/15-LO may be one potential mechanism for its pro-atherogenic role in mice, we were particularly interested in studying the relationship between this lipoxygenase and cytokine production. To explore further the mechanisms of 12/15-LO pro-atherogenic roles, we have cross-bred 12/15-LO-deficient mice with apolipoprotein B mRNA editing catalytic polypeptide-1 (apobec-1) Ϫ/Ϫ /LDL-R Ϫ/Ϫ mice. Apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mice afford considerable advantage over previous mouse models (15,16) because the hypercholesterolemia results almost entirely from elevated plasma levels of cholesterol-rich apoB-100-containing LDL, and they develop extensive atherosclerosis on a chow diet with strong gender-based differences as observed in humans (15).
Remarkably, we found that 12/15-LO-deficient macrophages have a severe selective defect in IL-12 production that is associated with the attenuation of both IFN consensus sequencebinding protein (ICSBP) in nuclear extracts and atherosclerotic lesions in 12/15-LO-deficient mice on an apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ background. Thus, the data provide strong evidence for a lipoxygenase-Th1 cytokine pathway in atherogenesis that together with enhanced oxidative modification of LDL can contribute to atherogenesis.
Tissue Preparation for Morphometric Determination of Atherosclerotic Lesions-Mice were anesthetized and bled via cardiac puncture. Aorta en face preparations/images were performed as described (9).
Measurement of Urinary Isoprostanes and Plasma Lipids-Urine collection and isoprostane 8,12-iso-iPF 2␣ -VI measurements were performed as described (18). Plasma total cholesterol levels were determined by an automated enzymatic technique on a Cobas Fara II autoanalyzer.
Peritoneal Macrophages Cultures-C57BL/6 and 12/15-LO Ϫ/Ϫ mice were injected intraperitoneally with 2 ml of sterile 3% Brewer's thioglycollate broth. After 4 days, macrophages were harvested by peritoneal lavage with Ca 2ϩ /Mg 2ϩ -free phosphate-buffered saline, adherencepurified, and cultured in RPMI supplemented with 10% fetal calf serum, 50 M ␤-mercaptoethanol, 10 g/ml polymyxin B, 1% penicillin, streptomycin, and fungizone with 5% CO 2 . Cells were treated with 10 ng/ml lipopolysaccharide (LPS) (Sigma) for 24 h, and supernatants were stored at Ϫ20°C for future analysis. In some experiments cells were pretreated with 100 units/ml IFN-␥ (Genzyme) for 16 h prior to stimulation with LPS. For RNA analysis, cells were stimulated with LPS for 6 h prior to RNA extraction. For nuclear transcription factor analysis, cells were treated with IFN-␥ (100 units/ml) and LPS (1 g/ml) for 4 h. All cultures were analyzed for viability by the metabolic MTT assay as described previously (19,20).
RT-PCR and Southern Blot Hybridization of PCR Products-Total RNA was extracted from mouse whole aorta using Trizol Reagent (Invitrogen). cDNA was synthesized from 2 g of total RNA by reverse transcription using the Superscript TM First-strand Synthesis System for RT-PCR (Invitrogen), and defined aliquots were used for PCR. The primers and reaction conditions for analysis of ␤-actin, IL-12 p40, and IFN-␥ were as described (2) except the number of PCR cycles used for ␤-actin and IFN-␥ were 25 and 35, respectively, and conditions for IL-12 p40 proceeded for 40 cycles at 94°C for 30 s, 62°C for 30 s, and 72°C for 45 s, followed by an extension at 72°C for 7 min using Advantage TM cDNA PCR kit (CLONTECH, Palo Alto, CA). PCR products were electrophoresed and transferred to Hybond-N nylon membranes (Amersham Biosciences). Specific oligonucleotides internal to the PCR amplification primers were used as probe for hybridization. The sequences of the oligonucleotides are as follows: ␤-actin, 5Ј-CCATGTACCCAGGCA-TTGCTGACAGGATGC-3Ј; IL-12 p40, 5Ј-CTGTGACACGCCTGAAGA-AGATGACATCAC-3Ј; and IFN-␥, 5Ј-CAGCGCTTTAACAGCAGGCCA-GACAGCACT-3Ј. The probe was 5Ј-end-labeled using T4 polynucleotide kinase (New England Biolabs) and [␥-32 P]ATP (3,000 Ci/mmol) (PerkinElmer Life Sciences). Blots were pre-hybridized in Rapid-Hyb buffer (Amersham Biosciences) at 42°C for 1 h, followed by probe hybridization for 4 h. Blots were washed in 2ϫ SSC and 0.1% SDS at room temperature for 2-5 min and then exposed to film.
Cytokine Analysis-Macrophage supernatants were analyzed for IL-12 p40 production by radioimmunoassay as described previously (19,20). Tumor necrosis factor-␣ (TNF-␣) levels were measured by ELISA kit (BD PharMingen) and nitric oxide with Greiss reagents and spectrophotometric assay with sodium nitrite as standard (21). All values were normalized for cell viability by the MTT assay. Aorta IL-12 p40 levels were measured by ELISA. Monoclonal antibody C17.8 (BD PharMingen) was used as a capture antibody and biotinylated C15.6 (BD PharMingen) as a detecting antibody.
RNase Protection Assay (RPA)-RNA was extracted from peritoneal macrophage cultures using Trizol reagent. mRNA levels of IL-12 p40 and MIF, as well as the housekeeping genes L32 and glyceraldehyde-3-phosphate dehydrogenase were analyzed using BD RiboQuant TM Multiprobe Template Sets (BD PharMingen) and RPA Assay kit (Ambion). Relative levels of mRNA expression were quantified using a PhosphorImager. Data were normalized to the level of L32 detected in each sample as control for variation in sample loading.
Western Blot Analysis-Nuclear proteins were extracted with NE-PER TM nuclear and cytoplasmic extraction reagents (Pierce). Protein concentrations were determined by the Bradford method (Bio-Rad). 5 g of nuclear extract was separated by 12% BisTris-NuPAGE gel (Invitrogen) under reducing conditions and transferred to Hybond ECL nitrocellulose membranes (Amersham Biosciences). The blots were blocked in 3% bovine serum albumin in TPBS (phosphate-buffered saline, 0.1% Tween 20) for 1 h. Primary antibodies (Santa Cruz Biotechnology) were used at the concentration of 0.4 g/ml in TPBS containing 1% bovine serum albumin for 1 h. After washing 3 times in TPBS, the blot was incubated with TPBS containing 1% bovine serum albumin and 1:40,000 dilution of secondary antibodies (Jackson Immu-noResearch) for 1 h. ECL detection (Amersham Biosciences) was used for the visualization of targeted protein.
Statistical Analysis-Initial analyses were performed by the Student's t test or one-way analysis of variance test. p Ͻ 0.05 was considered significant. Correlations between isoprostane 8,12-iso-iPF 2␣ -VI levels and the extent of en face lesions were determined by linear regression analysis. Prism 3.0 software (GraphPad) was used for all calculations. All data are presented as mean Ϯ S.E.  (Table I) (Fig. 1, A and B). Clear gender-based differences in lesion development were evident at 8 months (Fig. 1).
The Defect of IL-12 Production by 12/15-LO-deficient Macrophages Is Not Mediated by Soluble Factors-Several soluble factors have been identified that can negatively regulate macrophage IL-12 production including IL-10, prostaglandin E 2 , and TNF-␣ (24 -26). To determine whether 12/15-LO regulates IL-12 production through one of these or other unidentified soluble factor(s), we performed a series of supernatant transfer experiments. Comparable levels of IL-12 were produced by 12/15-LO Ϫ/Ϫ macrophages regardless of the presence of fresh media or conditioned media from stimulated macrophages from either 12/15-LO ϩ/ϩ or 12/15-LO Ϫ/Ϫ mice. Furthermore, the higher IL-12 levels detected in 12/15-LO ϩ/ϩ macrophage cultures were not affected by transfer of conditioned media from 12/15-LO Ϫ/Ϫ cells (Fig. 5). Together these data suggest that the regulation of IL-12 by 12/15-LO is not mediated through a stable soluble factor. However, these experiments do not eliminate the possibility that the regulation of IL-12 expression by 12/15-LO may be mediated through a soluble mediator with a short half-life or one that is consumed quickly by the macrophage cultures. To address this question, we also set up co-cultures that contained 50% each of 12/15-LO Ϫ/Ϫ and 12/15-LO ϩ/ϩ macrophages. These incubations produced an intermediate level of IL-12 providing further evidence that the nature of the defect in IL-12 production is intrinsic to the metabolism of the 12/15-LO Ϫ/Ϫ macrophages rather than mediated by production of a soluble transferable mediator (Fig. 5).
12/15-LO Ϫ/Ϫ Macrophages Exhibit Defective Nuclear Expression of ICSBP-ICSBP acts as a principal activator of IL-12 p40 transcription in RAW 264.7 cells (27). Together with c-Rel and PU.1, ICSBP has been shown to form a multiprotein complex, which binds to the Ets-2 site (5Ј-TTTCCT-3Ј; Ϫ210 to Ϫ205 for human and Ϫ218 to Ϫ213 for murine) on the IL-12 p40 promoter. The inhibition of this complex is responsible for the selective decrease of IL-12 p40 in macrophages following Fc␥ receptor ligation (28). To examine further the mechanism involved in the selective decrease of IL-12 p40 in 12/15-LOdeficient macrophages, we performed Western blot analysis of several transcription factors that have been reported to interact with the Ets-2 site on the IL-12 p40 gene promoter. In the nuclei of unstimulated wild type macrophages, minimal levels of Ets-2, ICSBP, IRF-1, and c-Rel were detected. Stimulation of the macrophages with IFN-␥ ϩ LPS significantly induced Ets-2, ICSBP, IRF-1, and c-Rel. ICSBP induction was markedly diminished, whereas no obvious changes were observed in Ets-2, IRF-1, and c-Rel in 12/15-LO Ϫ/Ϫ peritoneal macrophages compared with 12/15-LO ϩ/ϩ macrophages (Fig. 6). DISCUSSION Our studies conclusively demonstrate that 12/15-LO gene disruption markedly attenuates atherosclerotic lesion development in 8-month-old apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mice on a normal chow diet. These observations are consistent with studies performed in apoE Ϫ/Ϫ mice (9, 10) and LDL-R Ϫ/Ϫ mice (11,12). The apoE-deficient mouse develops typical lesions on a normal chow diet that faithfully mimics human disease progression from monocyte adhesion to foamy macrophages, fatty streaks, and advanced fibrosis (6,7). The LDL-R-deficient mouse model (8) is also well established but requires a high fat diet to induce elevated LDL cholesterol levels and lesion development. It should be noted that the current model, chow-fed apobec-1 Ϫ/Ϫ / LDL-R Ϫ/Ϫ mice, is the most similar to human lipoprotein physiology and is an excellent model for familial hypercholesterolemia. The advantage of the apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mouse model over the apoE Ϫ/Ϫ and LDL-R Ϫ/Ϫ models lies in two points: 1) atherosclerosis in the apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mouse model is much more pronounced in males than in females, whereas neither apoE Ϫ/Ϫ nor LDL-R Ϫ/Ϫ mouse model exhibits the clear gender-based distinction in expression of the disease so evident in humans (men are far more susceptible to atherosclerosis than premenopausal women); 2) familial hypercholesterolemia in humans is characterized by increased LDL cholesterol levels. The apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mouse model exhibits elevated plasma levels of LDL cholesterol. In the apoE Ϫ/Ϫ mouse model, the elevated cholesterol is in very low density lipoprotein which is not the case in humans. Thus, a pro-atherogenic role for 12/15-LO is consistently detected in three different atherosclerotic mouse models. In each, a Ϸ50% decrease in lesion size is observed at varying time points throughout the lifespan of the mice. At early time points (10 and 15 weeks) in apoE Ϫ/Ϫ mice there is a remarkable lag in lesion initiation (9,10). In the current study, decreased lesions were evident in male apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ 12/15-LO Ϫ/Ϫ aortas at 15 weeks but not in females. The average lesion size in females is always lower than males in this model (15) and, due to normal variation and limitations of the assay, may have been too early for detection of significant differences in females. Alternatively, hormonal status may be an important factor in early atherogenesis in apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ female mice. The paradoxical results showing that overexpression of 15-LO from the lysozyme promoter in transgenic rabbits decreased aortic lesion formation remains an enigma, potentially indicating strong species differences in the roles of 12/15-LO (29).
In addition to the effects of 12/15-LO gene disruption on IL-12 production, we also observed a decrease of oxidative stress as detected by urinary isoprostane measurement. Isoprostanes are chemically stable prostaglandin isomers that result from oxidative modification of arachidonic acid through a mechanism catalyzed by free radicals (30). They are associated with atherosclerotic lesion development and are regarded as reliable markers of oxidative stress (31). A strong correlation of urinary 8, 12-iso-iPF 2␣ -VI levels with lesion formation in apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ mice was observed. Despite no significant difference in total cholesterol levels between age-and gender-matched apobec-1 Ϫ/Ϫ /LDL-R Ϫ/Ϫ and apobec-1 Ϫ/Ϫ / LDL-R Ϫ/Ϫ /12/15-LO Ϫ/Ϫ mice, disruption of 12/15-LO in the double knockout mice caused a parallel decrease of 8,12-iso-iPF 2␣ -VI levels and lesion formation. These results are consistent with our recent report (10) that 8,12-iso-iPF 2␣ -VI levels decreased in parallel with decreased lesion size in apoE Ϫ/Ϫ /12/15-LO Ϫ/Ϫ mice. Although the mechanism involved in this pathway still needs to be investigated, one could speculate that 12/15-LO enzymatic activity on macrophage membranes may initiate a subsequent series of nonenzymatic lipid peroxidations that lead to the formation of isoprostanes in oxidized LDL.
We have unveiled a previously unrecognized connection between a lipoxygenase pathway and IL-12 production in macrophages stimulated in vitro and in a chow-fed mouse model of hypercholesterolemia and atherosclerosis in vivo. Lipid-laden monocyte/macrophages are the predominant cell type in atherosclerotic lesions. IL-12 is a heterodimeric cytokine mainly produced by monocytes/macrophages. The production of IL-12 heterodimer usually depends on the inductive transcription of the p40 subunit; however, pre-formed stores can be mobilized quickly in some cases of parasite infection (22,32). We detected a decreased production of IL-12 that was associated with decreased steady-state IL-12 mRNA levels suggesting that 12/15-LO regulates directly or indirectly with the transcription of the IL-12 p40 gene or the stability of IL-12 p40 mRNA. This reduction in IL-12 production was evident in 12/15-LO-deficient macrophages that were stimulated through the LPS-Toll receptor pathway. IL-12 promotes the generation of Th1 type responses by playing an essential role in the induction of IFN-␥ production by Th1 cells. Both IL-12 and IFN-␥ are present in human and murine atherosclerotic lesions, and recombinant IL-12 administration accelerates progression of atherosclerotic disease in these mice (2,3). Furthermore, atherosclerotic lesion area is reduced in IFN-␥ receptor null mice (23). Together these findings support an active role for Th1 cytokines in modulating atherogenesis in apoE-deficient mice (2-4). The decreased IL-12 production in atherosclerotic lesions in turn resulted in a marked reduction of IFN-␥ mRNA expression in apobec-1 Ϫ/Ϫ / LDL-R Ϫ/Ϫ 12/15-LO Ϫ/Ϫ mouse aorta compared with apobec-1 Ϫ/Ϫ / LDL-R Ϫ/Ϫ aortas. The lower expression of IFN-␥ implies a weaker Th1-mediated immune response at the aortic macrophage level in 12/15-LO Ϫ/Ϫ mice. A complete block in Th1mediated responses as a result of 12/15-LO deficiency is not evident because the mice still mount normal responses to infection with Listeria monocytogenes and Toxoplasma gondii (17). 2 Taken together, these data suggest that the proatherogenic contribution of 12/15-LO may function partly via an effect on the IL-12/IFN-␥ pathway.
Data from RPA provide evidence that 12/15-LO deficiency may contribute to IL-12 p40 gene down-regulation at the transcription level. Reportedly, IL-12 p40 gene expression is regulated by soluble factors and transcription factors that bind to regulatory elements on the IL-12 p40 promoter including Ets-2, NF-B, C/EBP, and GA-12 sites (24 -27, 33-35). The possibility of IL-12 p40 dysregulation by soluble transferable mediators was ruled out by supernatant transfer and co-culture experiments. Should any of the transcription factors be responsible for the selective IL-12 p40 gene dysregulation in 12/15-LOdeficient macrophages, then it should reveal the following features: 1) functional uniqueness to the IL-12 p40 gene (i.e. responsible for selective alterations for IL-12 p40 gene and not for genes controlling TNF-␣ and NO synthesis that were unchanged in 12/15-LO-deficient macrophages), and 2) should be induced by LPS and its transcriptional regulatory effect should be enhanced by IFN-␥ priming. NF-B is required for the induction of both TNF-␣ and inducible NO synthase genes. Furthermore, NF-B was shown not to account for the selective suppression of IL-12 transcription in macrophages following Fc␥ receptor ligation (28). Thus, although additional experiments will be required to formally exclude a role for NF-B, we do not consider it likely that NF-B is a critical factor in regulating the IL-12 p40-selective decrease in 12/15-LO-deficient macrophages.
A multiprotein complex that could bind to the Ets-2 site fits the two criteria mentioned above. It has been shown to be responsible for the selective IL-12 down-regulation following Fc␥ receptor ligation (28), and its components, including IRF-1, c-Rel, and ICSBP, are highly induced by either IFN-␥ or LPS. Based on the reported difficulties in quantitating protein complexes in primary macrophages (26), we instead examined the nuclear levels of the individual protein complex components. Markedly attenuated ICSBP nuclear levels were detected in 12/15-LO Ϫ/Ϫ macrophages in response to IFN-␥ ϩ LPS. This observation is in accordance with previous findings (36) that ICSBP knockout mice display defective IL-12 production. Our data provide evidence that the defect of the ICSBP-containing protein complex correlates with the selective IL-12 defect in LPS ϩ IFN-␥-stimulated 12/15-LO Ϫ/Ϫ macrophages. The nature of the decreased ICSBP nuclear accumulation in LPS ϩ IFN-␥-stimulated 12/15-LO Ϫ/Ϫ macrophages remains unknown. Several factors may be responsible for this ICSBP attenuation, and these factors include the following: 1) the products of 12/15-LO, such as 12-hydroxyeicosatetraenoic acid, 15-hydroxyeicosatetraenoic acid, 13-hydroxyoctadecadienoic acid, 12-hydroperoxyeicosatetraenoic acid, 15-hydroperoxyeicosatetraenoic acid, and 13-hydroperoxyoctadecadienoic acid; 2) an undetermined protein that is down-regulated or functionally deficient in 12/15-LO-disrupted macrophages. The function of this protein may be closely related to the ICSBP nuclear expression; 3) oxidative stress-related factors that influence gene expression, such as reactive oxygen species (37)(38). Further study to examine the contribution of these factors on ICSBP attenuation is needed for a complete understanding of the mechanism of IL-12 synthesis defect in 12/15-LO-deficient macrophages.