Genetic alterations of IL-1 receptor antagonist in mice affect plasma cholesterol level and foam cell lesion size

Abstract

Inflammatory cytokines have been linked to atherosclerosis by using cell culture models and acute inflammation in animals. The goal of this study was to examine lipoprotein levels and early atherosclerosis in chronic animal models of altered IL-1 physiology by using mice with deficient or excess IL-1 receptor antagonist (IL-1ra). IL-1ra knockout C57BL/6J mice fed a cholesterol/cholate diet for 3 mo had a 3-fold decrease in non-high-density lipoprotein cholesterol and a trend toward increased foam-cell lesion area compared to wild-type littermate controls. IL-1ra transgenic/low-density lipoprotein receptor (LDLR) knockout mice fed a cholesterol-saturated fat diet for 10 wk showed a 40% increase in non-high-density lipoprotein cholesterol, consistent with the IL-1ra knockout data, although there was no change in lesion size. When these IL1-ra overexpressing transgenic mice on the LDLR knockout background were fed a high-cholesterol/high-fat diet containing cholate, however, a statistically significant 40% decrease in lesion area was observed compared to LDLR knockout mice lacking the transgene. By immunohistochemistry, IL-1ra was present in C57BL/6J and LDLR knockout aortae, absent in IL-1ra knockout aortae, and present at high levels in LDLR knockout/IL-1ra transgene aortae. In summary, IL-1ra tended to increase plasma lipoprotein levels and, when fed a cholate-containing diet, decrease foam-cell lesion size. These data demonstrate that in selected models of murine atherosclerosis, chronic IL-1ra depletion or overexpression has potentially important effects on lipoprotein metabolism and foam-cell lesion development.

Inflammation is now recognized as a major contributor to atherosclerotic vascular disease through effects on lipoprotein metabolism and arterial wall biology (1–3). IL-1β and tumor necrosis factor α (TNFα) are among the most important cytokine mediators of the inflammatory response (4), and a large number of studies by using cell culture experiments and cytokine-injected animal models have attempted to delineate the role of acute IL-1 exposure in lipoprotein metabolism and atherogenesis (1–3, 5). For example, the increase in plasma triglyceride and very-low-density lipoprotein (VLDL) observed in humans and animals after acute infection can be mimicked in animals by i.v., i.m., or i.p. injection of IL-1β or TNFα (6). Similarly, IL-1 and TNFα injections in rabbits and rodents can acutely increase plasma LDL (6). Interestingly, two previous studies involving continuous administration for 1 wk of either IL-1β or TNFα demonstrated decreased triglyceride levels, and at high doses of IL-1β, decreased plasma cholesterol levels as well (7, 8). There are very few studies on lipoprotein changes in chronic inflammation or chronic cytokine exposure in animal models, and observations in humans with chronic diseases such as rheumatoid arthritis and systemic lupus erythrematosis are complicated by the effects of anti-lipoprotein autoantibodies and anti-inflammatory drugs. Regarding the potentially atherogenic arterial-wall effects of inflammatory cytokines, there has been a multitude of studies examining specific responses of cultured arterial-wall cells to addition of exogenous cytokines, including IL-1 (1–3, 9, 10). Examples of IL-1-mediated responses that have been observed in these studies include induction of clotting factors, increase in transendothelial permeability, expression of adhesion molecules on endothelial cells, induction of monocyte chemotactic protein-1 and monocyte recruitment, expression of macrophage and smooth muscle cell growth factors, and potentially atherogenic molecules like sphingomyelinase, phospholipase A2, and matrix metalloproteinases (11–13).

Despite the tremendous interest in the effects of inflammatory cytokines on lipoprotein metabolism and atherogenesis, there have been very few in vivo studies examining chronic effects of cytokine alterations on these processes. In mice lacking TNF receptor p55, Schreyer et al. (14) demonstrated no significant change in plasma lipid levels and an unexpected increase in atherosclerosis, possibly related to increased macrophage scavenger receptor activity. Gupta et al. (15) showed that apolipoprotein E knockout mice lacking the interferon-γ receptor had a reduction in lesion size that might have resulted from an increase in potentially atheroprotective apolipoprotein A-IV-containing particles and possibly increased ABCA1-mediated cholesterol efflux (16). Elhage et al. (17) used a s.c. osmotic minipump or repeated s.c. injections of IL-1 receptor antagonist (IL-1ra) or TNF-binding protein in apolipoprotein E knockout mice. Neither treatment affected serum total cholesterol, and the IL-1ra treatment, but not TNF-binding protein, was associated with a decrease in lesion size. In this study, however, the duration of cytokine antagonist treatment was relatively short (1 mo), the delivery of the cytokine antagonists was by an unnatural route (i.e., s.c.), the IL-1ra component of the study lacked proper controls (e.g., no sham minipump control), the study was not conducted in a barrier facility, and apolipoprotein E depletion itself may have confounding effects on cytokine alterations.

Thus, a properly conducted study on the effects of IL-1 on lipoprotein metabolism and atherogenesis in a chronic animal model is lacking. In this context, we chose to examine the effects of IL-1ra depletion and endogenous IL-1ra overexpression by using IL-1ra knockout and transgenic mice, respectively. IL-1ra is an endogenous competitive inhibitor of the types I and II IL-1 receptors and thus blunts both IL-1α and IL-1β responses (11, 18, 19). Importantly, IL-1ra has recently been shown to be expressed in human endothelial cells and atherosclerotic lesions (20, 21), and a variable number tandem repeat polymorphism in the IL-1ra gene in humans is associated with coronary artery disease (22). IL-1ra knockout mice on the C57BL/6J genetic background demonstrate decreased weight and increased susceptibility to lethal endotoxinemia, decreased serum levels of IL-1 after endotoxin exposure, decreased susceptibility to infection with Listeria monocytogenes (23, 24), and an increased hepatic acute phase response to a turpentine abscess (25). Interestingly, IL-1ra knockout mice on the BALB/cA background develop a chronic inflammatory polyarthropathy (26), those on the DBA1 background demonstrate increased susceptibility to collagen-induced arthritis (27), and those on a 129/Ola × MF1 background develop arterial inflammation (28); neither arthrophathy nor arteritis is found in the IL-1ra knockout C57BL/6J mice (23, 26, 28). IL-1ra overexpressing transgenic mice (C57BL/6J background) have normal weight gain and demonstrate effects opposite to those of the knockout mice, namely, decreased mortality and increased serum IL-1 after endotoxin challenge and increased susceptibility to infection with Listeria monocytogenes (23, 24). In the current study, we report that chronic alterations in IL-1ra expression in selected models of murine atherosclerosis affect lipoprotein levels and/or foam cell lesion development.

Materials and Methods

Mice and Genotyping.

Offspring of heterozygote breeding pairs of IL-1ra knockout mice on the C57BL/6J background (16 backcrosses) were genotyped by PCR analysis to distinguish the endogenous from the disrupted IL-1ra locus (23). Beginning at 1 mo of age and continuing for 3 mo, the mice were fed a high fat/cholesterol/bile salt diet containing 1.25% cholesterol, 7.5% cocoa butter, 7.5% casein, and 0.5% sodium cholate (TD 88051 from Harlan Teklad, Madison, WI), hereafter referred to as the cholate/cholesterol diet (29) To test the health status of IL-1ra knockout and wild-type mice, we analyzed serum from nonfasted mice after 7 wk on the cholate/cholesterol diet. The serum levels of total protein, total bilirubin, alkaline phosphatase, aspartate aminotransferase, alanine aminotransferase, and glucocorticoids were assayed by VetPath Laboratory (Tulsa, OK). The serum level of IL-1β protein was assayed by ELISA (BioSource International, Camarillo, CA).

IL-1ra transgenic mice (line T14) were bred for over ten generations into the C57BL/6J background; the T14 locus has approximately six copies of the IL-1ra gene in tandem array (23). As described by Hirsch et al. (23), the transgenic construct used to make these mice contained the promoter, exons, and introns for only the secreted form of IL-1ra (i.e., not the intracellular form). These mice were crossed with low-density lipoprotein receptor (LDLR) knockout mice (30), also on the C57BL/6J background, to generate LDLR knockout mice that either lacked the IL-1ra transgene (Tg) or were homozygous for the transgenic locus. Transgene status was determined by quantitative PCR by using IL-NEST (5′-GTC AGT CCA ACA CAT CCA AGG-3′) and MRA 547 (5′-AAA CCC CAG TTC TTA TGG CAC-3′) (23). The PCR conditions were 30 cycles of 94°C, 1 min; 60°C, 2 min; and 72°C, 2 min. Samples were taken after 15, 18, 21, or 30 cycles and electrophoresed in an agarose gel. The intensity of the 677-bp PCR product was compared to samples from mice previously confirmed to have zero, one, or two copies of the IL-1ra transgene by breeding with transgene null mice. The mice were fed either the cholate/cholesterol diet described above for 1 mo or a so-called “Western” diet containing 21% anhydrous milk fat and 0.15% cholesterol (TD 88137 from Harlan Teklad) for 10 wk.

All mice were maintained in a specific pathogen-free facility and handled in accordance with institutional guidelines.

Plasma Lipoprotein Analysis and Aortic Root Assay.

On the day of analysis, food was removed from the cages in the morning, and the mice were fasted for 7 h. The animals were then anesthetized, blood was withdrawn by cardiac puncture, the heart was perfused with PBS, and the heart and proximal aorta were harvested. The heart and aorta were perfused ex vivo with PBS and then 10% buffered formalin, embedded in OCT compound (Sakura Finetek, Torrance, CA), snap-frozen in an ethanol-dry ice bath, and stored at −70°C. Sections (10-μm-thick) were cut at −20°C by using a Microm (Walldorf, Germany) microtome cryostat HM 505 E. Starting from the atrial leaflets, every eighth section was retained for analysis for a total of ten sections. To evaluate the fatty streak lesions, the sections were stained with Oil Red O for neutral lipid and with Harris hematoxylin for nuclei. The lesion area was quantified by using a Nikon Labophot-2 microscope equipped with an Sony CCD-Iris/RGB color video camera attached to a computerized imaging system using IMAGE-PRO PLUS 3.0 software. The mean area of intimal lipid accumulation per section from six sections was determined in a blinded fashion for individual animals.

Total plasma cholesterol and triglyceride was determined by using commercial enzymatic kits (Wako Chemicals GmbH). Plasma high-density lipoprotein (HDL) cholesterol was determined after dextran sulfate-Mg²⁺ precipitation of apoB-containing lipoproteins. Plasma lipoproteins were analyzed by an FPLC system consisting of two Superose 6 columns connected in series (Amersham Pharmacia). The cholesterol content of each fraction was measured by enzymatic assay and calculated to indicate the cholesterol concentration in plasma (mg/ml).

Immunohistochemistry.

Sections (10-μm-thick) of the proximal aortae were cut at −20°C on a cryostat, placed on poly-l-lysine-coated glass slides, and fixed in ice-cold acetone for 10 min. The sections were air-dried for 5 min, washed in PBS containing 0.1% Triton X-100 for 10 min, and rinsed in multiple changes of PBS for 10 min at room temperature. The sections were then preincubated with 1.5% normal donkey serum in PBS for 1 h at room temperature. Next, the sections were incubated with 1.5% donkey serum containing 2 ng/μl of goat polyclonal anti-IL-1ra N-terminal peptide IgG (Q-19 from Santa Cruz Biotechnology) for 16 h at 4°C. In some control experiments, the IgG was preabsorbed with a 5-fold molar excess of the N-terminal peptide before addition to the slides. After the sections were washed in PBS for 10 min, the bound primary Ab was visualized by using 4 ng/μl biotinylated anti-goat IgG (Santa Cruz Biotechnology), followed by streptavidin peroxidase (Vectastain Elite ABC peroxidase kit, Vector Laboratories) and 3,3′-diaminobenzidine. The sections were counterstained with hematoxylin, rinsed, mounted in Permount, and viewed with a Nikon microscope with a ×40 objective. No reactivity was observed when the primary Ab was omitted.

Statistical Analysis.

Data are reported as means ± SE. For data that was normally distributed, the unpaired, two-tailed t test was used for comparisons between a single experimental group and a control. For comparisons among three groups, ANOVA was used initially. When ANOVA indicated differences among the groups, pairwise comparisons between groups (α = 0.05) were performed by using the Student-Newman-Keuls posthoc test (SPSS 10.0 software, SPSS, Chicago). Data that was not normally distributed were square-root transformed to achieve a normal distribution (31) and then statistically analyzed as above. Square-root transformation is the best method for normalization of murine lesional data and some plasma cholesterol data, specifically when most of the data points are in a cluster while a few points have much larger values, skewing the distribution to the right (the mean much larger than the median) (31).

Results

IL-1ra Knockout Mice.

By setting up IL-1ra ± breedings, we obtained mice that were wild-type, heterozygous, or homozygous for the dysfunctional IL-1ra allele. Starting at 4 wk of age, these mice were fed a cholate-containing cholesterol-enriched diet (29), and 12 wk later the mice were analyzed for total and HDL plasma cholesterol and for the size of foam cell lesions in the proximal aorta. The death rate during the 12-wk feeding period was 8/32 for wild-type, 8/45 for heterozygotes, and 9/28 for knockouts. The weights of the surviving mice, however, were not statistically different (16.6 ± 0.6 g, 17.7 ± 0.5 g, and 16.9 ± 0.7 g for wild-type, heterozygous, and homozygous knockout mice, respectively). Moreover, none of the surviving mice in any of the groups showed signs of severe illness, although a small number (3 of 19) had hunched backs. Additionally, we assayed the serum of several mice in each group for total protein, total bilirubin, alkaline phosphatase, aspartate transferase, alanine transferase, and glucocorticoids. The protein, bilirubin, and glucocorticoid values were nearly identical between the two groups of mice. There was a trend toward lower alkaline phosphatase, aspartate transferase, and alanine transferase levels in the knockout mice, but none of these differences reached statistical significance (data not displayed).

As shown in Fig. 1A, there was a decrease in total cholesterol levels that correlated with the number of dysfunctional IL-1ra alleles; the difference between the homozygous knockout value and the wild-type value reached statistical significance by ANOVA (P < 0.05). HDL cholesterol represented a very small proportion of total cholesterol and did not change substantially, indicating that IL-1ra deficiency leads to a decrease in non-HDL cholesterol. The atherosclerosis data in Fig. 1B showed a high level of variability and did not reach statistical difference among the three groups of mice, but there was a trend toward increased lesion size that correlated positively with the number of dysfunctional IL-1ra alleles. In conclusion, in the cholate/fat-fed C57BL/6J model, IL-1ra deficiency is associated with a decrease in non-HDL cholesterol and a trend toward an increase in atherosclerotic lesion size.

Figure 1.

IL-1ra deficiency is associated with decreased non-HDL plasma cholesterol and increased foam cell lesion size. Wild-type (n = 24), IL-1ra ± (Heterozyg. IL-1ra KO; n = 37), and IL-1ra −/− (Homozyg. IL-1ra KO; n = 19) mice on the C57BL/6J background were fed a cholate-containing cholesterol-enriched diet for 3 mo and then analyzed for total plasma and HDL cholesterol concentrations (A) and foam cell lesion area in the proximal aorta (B). (A) The homozygous IL-1ra knockout total plasma cholesterol value was statistically different from the wild-type knockout values (P < 0.05). (B) The square root-transformed values for wild-type, heterozygous knockout, and homozygous knockout lesion sizes were 59.9 ± 5.8, 71.9 ± 5.2, and 77.2 ± 17.3 μm, respectively (P > 0.05). For an explanation of the statistical analysis of these data, see Statistical Analysis.

To determine whether IL-1ra was present in the proximal aorta of wild-type mice but absent in IL-1ra knockout mice, sections from mice fed the cholate/cholesterol diet for 7 wk were subjected to immunohistochemistry by using two different anti-IL-1ra Abs. As shown in Fig. 2A, IL-1ra protein was present in the endothelium of the proximal aorta. The specificity of this Ab is demonstrated by the complete absence of reaction product in the proximal aorta of IL-1ra knockout mice (Fig. 2B). A study of human atherosclerotic arteries also has shown that IL-1ra was localized in the endothelium (20). Thus, IL-1ra is present in the endothelium of the proximal aorta of cholate/cholesterol-fed C57BL/6J mice.

Figure 2.

IL-1ra protein is present in the endothelium of the proximal aorta of cholate/cholesterol-fed C57BL/6J mice. C57BL/6J wild-type (WT; A) and IL-1ra −/− (IL-1ra KO; B) mice were fed the cholate/cholesterol diet for 7 wk. Sections of proximal aorta from these mice were then subjected to immunohistochemical analysis by using an anti-IL-1ra Ab. (Insets) Enlargements of the sections outlined by the dotted box in each A and B.

IL-1ra Transgenic Mice.

To further test the role of IL-1 responses in lipoprotein metabolism and atherosclerosis, we studied transgenic mice that overexpress only the secreted form of IL-1ra and have a blunted response to inflammatory stimuli (23). In this set of experiments, the LDLR knockout model of murine atherosclerosis was used (30). Thus, wild-type mice or mice homozygous for the IL-1ra transgene were bred into the LDLR knockout background; all mice were C57BL/6J. At 4 wk of age, the mice were fed a cholesterol-enriched “Western” diet (23), and 10 wk later the mice were analyzed for total plasma cholesterol levels, plasma lipoprotein profile, and atherosclerotic lesion size in the proximal aorta. There was only 3% mortality in the LDLR knockout mice and none in the transgenic mice, and the average weights of the LDLR knockout mice with or without the IL-1ra transgene were 31.5 ± 1.7 and 30.5 ± 0.8 g, respectively (difference not significant). As shown in Fig. 3A, the transgenic mice had statistically significant 40% increases in both total plasma cholesterol and triglyceride. The lipoprotein profile in Fig. 3B shows that each type of mouse had three major peaks: a rapidly eluting VLDL peak, a middle LDL peak, and a slowly eluting HDL peak, which was much smaller than the first two peaks. Most importantly, the VLDL and LDL peaks were substantially larger in the IL-1ra transgenic mice, whereas the relatively small HDL peaks were similar. Thus, consistent with our previous finding that deficiency of IL-1ra was associated with a decrease in non-HDL cholesterol, IL-1ra over-expression was associated with an increase in non-HDL cholesterol.

Figure 3.

IL-1ra over-expression is associated with increased non-HDL plasma cholesterol in LDLR knockout mice fed a cholesterol-fat diet. LDLR knockout mice (LDLR KO/No Tg; n = 31) and LDLR knockout mice homozygous for the IL-1ra transgene (LDLR KO/IL-1ra Tg; n = 18) were fed a cholesterol-fat enriched “Western” diet for 10 wk and then analyzed for total plasma cholesterol concentrations (A), plasma cholesterol FPLC profile (B), and lesion area in the proximal aorta (C). (A) IL-1ra transgenic total plasma cholesterol and triglyceride values were statistically different from the wild-type values (P = 0.001 and 0.003, respectively). (C) Square root-transformed values for wild-type and transgenic lesion sizes were 455.8 ± 17.6 and 450.8 ± 23.1 μm, respectively (P > 0.05).

The data in Fig. 3C demonstrate that there was no significant difference in proximal aorta lesion size between LDLR knockout mice with or without the IL-1ra transgene. There are several possibilities to explain why lesion size was not decreased in the transgenic mice, which might have been predicted given that overexpression of IL-1ra blunts IL-1-induced inflammatory responses (23). First, the increased levels of non-HDL cholesterol in the plasma of the IL-1ra transgenic mice could have counteracted a trend toward decreased lesion size. Second, the level of endogenous IL-1ra in the arterial wall may be “saturating” and so increasing expression above this level would have little effect. Third, the model chosen for the IL-1ra transgenic study had relatively large lesions, and the biological effects of IL-1ra on atherosclerosis may only be apparent in earlier, smaller lesions. Fourth, the underlying inflammatory milieu of the cholate-containing diet (32) may be necessary for IL-1ra transgene expression and/or to observe the biological effects of IL-1ra overexpression (23).

In the context of these issues, LDLR knockout mice with or without the IL-1ra transgene were fed the cholate/cholesterol diet for 4 wk. There was no mortality in either group of mice, and the average weights (17.2 ± 0.4 g for wild-type and 16.5 ± 0.1 g for transgenics) were statistically identical. As shown in Fig. 4 A and B, the plasma cholesterol level was only slightly higher in the transgenic mouse (P > 0.05), and the triglyceride levels and lipoprotein profiles, in which particles in the VLDL size range predominated, were essentially identical. Most importantly, there was a statistically significant decrease in lesion size between LDLR knockout mice with and without the IL-1ra transgene (Fig. 4C). Note that the range of lesions sizes in this model were ≈10-fold larger than those of the LDLR-positive C57BL/6J model (compare y axes of Figs. 1C and 4C). Thus, in a model in which the IL-1ra transgene had no significant effect of plasma cholesterol or lipoproteins, in which the overall degree of atherosclerosis was moderately advanced, and in which inflammation may have been increased by the presence of cholate in the diet (32), overexpression of the IL-1ra transgene was associated with smaller lesion size.

Figure 4.

IL-1ra over-expression is associated with decreased foam cell lesion size in LDLR knockout mice fed a cholate/cholesterol diet. LDLR knockout mice (LDLR KO/No Tg; n = 36) and LDLR knockout mice homozygous for the IL-1ra transgene (LDLR KO/IL-1ra Tg; n = 39) were fed a cholate-containing cholesterol-enriched diet for 4 wk and then analyzed for total plasma cholesterol concentrations (A), plasma cholesterol FPLC profile (B), and lesion area in the proximal aorta (C). In A, the IL-1ra transgenic total plasma cholesterol value was statistically different from the wild-type value (P = 0.01). In C, the square root-transformed values for wild-type and transgenic lesion sizes were 222.7 ± 8.8 and ± 199.4 ± 7.9 μm, respectively (P = 0.027).

These data suggest that overexpression of IL-1ra can have effects on the arterial wall. In an attempt to correlate this finding with the presence of IL-1ra in the proximal aorta, sections from LDLR knockout with or without the IL-1ra transgene, fed the cholate/cholesterol diet for 4 wk, were probed with the N-terminal anti-IL-1ra Ab mentioned above. Proximal aorta from transgenic mice demonstrated a very strong brown reaction product in the endothelial and subendothelial areas (Fig. 5A). Specificity was demonstrated by the complete absence of reaction product in sections exposed to either a control IgG (not shown) or to the immune IgG preabsorbed with the N-terminal IL-1ra peptide antigen (Fig. 5C). In LDLR knockout mice not expressing the transgene, endogenous IL-1ra was evident by the presence of a diffuse brown reaction product in the endothelial and subendothelial areas (Fig. 5B; control in Fig. 5D), but the intensity of the staining was substantially less than that seen in the IL-1ra transgenic mice (compare A and B in Fig. 5. The differential amount of staining seen in the wild-type vs. the IL-1ra transgenic mice agrees with the previous finding that the IL-1ra gene copy number determines the level of IL-1ra protein (23). Thus, consistent with an anti-atherogenic effect of IL-1ra, levels of this cytokine regulator in the proximal aorta are higher in LDLR knockout carrying the IL-1ra transgene compared to nontransgenic LDLR knockout mice.

Figure 5.

IL-1ra protein is present in increased amounts in the proximal aorta of IL-1ra transgenic LDLR knockout mice. LDLR knockout mice homozygous for the IL-1ra transgene (IL-1ra Tg; A and C) and LDLR knockout mice without the transgene (no Tg; B and D) were fed the cholate/cholesterol diet for 4 wk. Sections of proximal aorta from these mice were then subjected to immunohistochemical analysis by using an Ab directed against IL-1ra. The sections in C and D were probed with the anti-IL-1ra Ab that had been preabsorbed with the IL-1ra peptide antigen (Preabs Ab). Note that staining above background is present in the section from the nontransgenic mouse (B), but the staining is much more intense in the section from the transgenic mouse (A).

Discussion

The overall goal of this study was to initiate an exploration of the role of IL-1 physiology in lipoprotein metabolism and atherosclerosis in vivo by using a genetically engineered chronic animal model. As described in detail previously, studies of this nature are scarce despite a vast literature on the acute effects of IL-1 on lipoprotein metabolism in animal models and on responses of cultured arterial wall cells to exogenous IL-1. Previous studies of chronic administration of cytokines were conducted for only 1 wk (7, 8), whereas in the current study, the time period was 1–3 mo. In addition, this study has shown that the IL1-ra protein is present in the arterial wall of murine atherosclerosis models, consistent with recent findings in human lesions (20) and supportive of the hypothesis that cytokine influences on atherosclerosis are subject to local regulation by this molecule. The results of the in vivo study reported here raise many mechanistic questions that will require further probing of the above models as well as the creation of new in vivo models.

Among the outstanding issues that need to be studied further in the current models include the mechanisms of the decrease of non-HDL lipoproteins in the IL-1ra knockout mice and of the increase of these lipoproteins in one of our models of IL-1ra transgenic mice. Although ill health of the mutant mice could theoretically affect lipoprotein levels, our data and observations do not support this mechanism. Almost all of the mutant mice looked healthy, their mortality rates and weights were similar to those of wild-type mice, and there were no abnormalities in liver function tests. Therefore, we speculate that inflammatory cytokine signaling in hepatocytes directly affects changes in lipoprotein synthesis, lipoprotein catabolism, or both. In future studies, this idea can be tested both in whole animals and in isolated hepatocytes.

The statistically significant decrease in lesion size in LDLR knockout/IL-1ra transgenic mice fed a cholate-containing diet (Fig. 4D) supports the hypothesis that IL-1 responses are atherogenic in an inflammatory model. A trend supporting this hypothesis also was seen with cholate-fed IL-1ra knockout mice (Fig. 1B), but there was tremendous variation in lesion size, and statistical significance was not reached. Because the mice were housed in a barrier facility, fed identical synthetic, irradiated diets, and backcrossed 16 generations into the C57BL/6J background, the explanation for the wide variation in lesion size is not obvious. Of note, this trend toward smaller lesion size in the cholate-fed IL-1ra knockout mice was seen despite an increase in non-HDL cholesterol, suggesting an anti-atherogenic effect of IL-1ra deficiency in this model that is more potent than was actually observed. In reality, one might not predict extremely large effects on atherosclerosis because there are compensatory responses that should partially balance excess IL-1β signaling, and there are overlapping cytokine pathways that should partially counteract defective IL-1β signaling. For example, there is a delicate balance between IL-1 and IL-1ra secretion (11, 18, 19). Indeed, consistent with the data of Hirsch et al. (23), we found that cholate-fed IL-1ra knockout mice had lower levels of serum IL-1β than cholate-fed wild-type mice (data not shown). Furthermore, there are other mechanisms in addition to IL-1ra to keep IL-1 responses in check (33, 34), and there are other cytokines, notably TNFα, that undoubtedly add a degree of redundancy to these in vivo models.

Despite these complexities, the current work has uncovered potentially important lipoprotein and lesional consequences of chronic IL-1ra perturbation in vivo. These findings, which may be related to those showing an association between an IL-1ra gene polymorphism and coronary artery disease in humans (22), provide the basis for more detailed explorations of this important area of vascular disease research.

Abbreviations

ra

receptor antagonist

LDLR

low-density lipoprotein receptor

VLDL

very-low-density lipoprotein

TNF

tumor necrosis factor

HDL

high-density lipoprotein

KO

knockout

Tg

transgene