Stage-dependent differential effects of interleukin-1 isoforms on experimental atherosclerosis

Abstract

Aims

Targeting interleukin-1 (IL-1) represents a novel therapeutic approach to atherosclerosis. CANTOS demonstrated the benefits of IL-1β neutralization in patients post-myocardial infarction with residual inflammatory risk. Yet, some mouse data have shown a prominent role of IL-1α rather than IL-1β in atherosclerosis, or even a deleterious effect of IL-1 on outward arterial remodelling in atherosclerosis-susceptible mice. To shed light on these disparate results, this study investigated the effect of neutralizing IL-1α or/and IL-1β isoforms starting either early in atherogenesis or later in ApoE^–/– mice with established atheroma.

Methods and results

The neutralization of IL-1α or of both IL-1 isoforms impaired outward remodelling during early atherogenesis as assessed by micro-computed tomographic and histologic assessment. In contrast, the neutralization of IL-1β did not impair outward remodelling either during early atherogenesis or in mice with established lesions. Interleukin-1β inhibition promoted a slant of blood monocytes towards a less inflammatory state during atherogenesis, reduced the size of established atheromata, and increased plasma levels of IL-10 without limiting outward remodelling of brachiocephalic arteries.

Conclusion

This study established a pivotal role for IL-1α in the remodelling of arteries during early experimental atherogenesis, whereas IL-1β drives inflammation during atherogenesis and the evolution of advanced atheroma in mice.

Translational perspective

These experiments demonstrate a stage-dependent role of interleukin-1 (IL-1)α or IL-1β in mouse atherosclerosis. We evaluated in mice the effect of neutralization of IL-1α, IL-1β or both isoforms on early atherogenesis, or on the evolution of established atheroma, an approach that mimics secondary prevention in humans. Interleukin-1α blockade affected early atherosclerosis, whereas anti-mouse IL-1β treatment, but not IL-1α neutralization, limited progression and inflammation in established lesions. These results inform the interpretation of CANTOS as well as disparate experimental studies, and guide future anti-inflammatory interventions on patients with cardiovascular disease.

Introduction

Atherosclerosis remains a leading cause of death worldwide. Numerous human and animal studies have shown a prominent role for inflammation in atherogenesis and its complications.¹ The intrinsic vascular wall cells, endothelium and smooth muscle, can express the genes that encode both isoforms of the potent pro-inflammatory cytokine interleukin-1 (IL-1).^2,3 Thus, IL-1 could participate in the earliest steps of atherogenesis and help initiate leucocyte recruitment to the nascent atheroma.⁴ IL-1 can promote the proliferation of smooth muscle cells (SMCs) and their production of IL-6 and prostaglandins.^5,6 Interleukin-1 can also promote vascular cell production of matrix metalloproteases (MMPs), enzymes that participate critically in extracellular matrix degradation, and vascular remodelling during atherogenesis.^7,8

Interleukin-1 cloning identified two isoforms: IL-1α and IL-1β.⁹ These cytokines signal through the IL-1 Type I receptor and can induce each other’s expression, leading to an auto-amplification loop in vascular and mononuclear cells.^10–12 The IL-1 receptor antagonist (IL-1Ra), a member of the IL-1 family that acts as an endogenous competitive inhibitor, modulates cytokine-receptor interactions and downstream signalling. Variations in plasma levels of IL-1Ra caused by human genetic variations as well as overexpression or genetic deletion of IL-1Ra in mice have implicated IL-1 signalling in atherogenesis, generally without identifying the involvement of individual members of the IL-1 family in the process.^13–18 Multiple studies over the past two decades have shown that selective IL-1β inhibition decreases atherogenesis in mice.^19–21 Yet, recent animal studies have questioned the benefits of targeting the IL-1β isoform in atherosclerosis and suggested that IL-1α predominates over IL-1β in atherogenesis.^22,23 For instance, IL-1α but not IL-1β deficiency in haematopoietic cells decreased the atherosclerotic lesion area in the aortic roots of LDL-R^–/– mice.²³ Yet, IL-1 type 1 Receptor (IL-1R) deficiency—which blunts the response to both IL-1 isoforms—reduced MMP-3 expression in the brachiocephalic artery and consequently impaired beneficial outward remodelling in advanced atherosclerotic lesions.¹³ These studies, together with our observation that either IL-1 isoform can promote the expression of MMP-3 and other remodelling-related proteases in human atherosclerotic lesions and vascular cells,⁸ suggested that selective IL-1 isoform neutralization may avoid the potentially adverse effects of abrogating the action of both IL-1 isoforms.

Clarification of the effects of the two IL-1 isoforms participate in atherosclerosis has critical clinical importance, given the availability of selective anti-IL-1 isoform antibodies for clinical use. For example, the Canakinumab Antiinflammatory Thrombosis Outcome Study (CANTOS) showed that selective neutralization of IL-1β with a monoclonal antibody (canakinumab) significantly reduced recurrent cardiovascular events in patients with residual inflammatory risk despite guideline-directed therapy.^24–26 An anti-IL-1α antibody has undergone clinical evaluation.²⁷ Anakinra, the recombinant IL-1Ra, blocks the action of both isoforms. Thus, therapies available for clinical use can target either IL-1α, IL-1β or the two sibling cytokines.

This study aimed to dissect the relative contribution of IL-1 isoforms to mouse atherosclerosis by examining the effect of selective neutralization of IL-1α, IL-1β, or both on the early phase of atherogenesis and on the evolution of already established atheromata. The experiments addressed particularly the roles of the isoforms on arterial remodelling, as some data show an adverse effect of IL-1 interruption on this process. We evaluated both the aortic root, to represent an elastic artery, and the more muscular brachiocephalic artery. The results demonstrated that neutralization of IL-1α or of both isoforms reduces lesion area in the aortic roots, but impairs the outward remodelling of the brachiocephalic artery during atherogenesis. Furthermore, IL-1β neutralization promoted a monocyte slant towards a less pro-inflammatory state in atherogenesis, decreased the size of established atheromata, and increased plasma levels of IL-10. Interleukin-1 neutralization did not affect arterial remodelling in established atheromata. These data establish distinct roles of the two IL-1 isoforms depending on the stage and location of experimental lesions, help resolve some of the apparent disparities in the literature, and can inform the design of future clinical trials with different strategies of IL-1 inhibition.

Methods

For a full description of the methods used, see Supplementary material online.

Mouse experiments

All experiments were approved by the Institutional Animal Care and Use Committees and were performed in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care and the National Institutes of Health. These experiments used male mice to reduce variability and to conserve animal use. Mice were assigned randomly to experimental groups and results were analysed in a blinded fashion. Genotyping was performed by Transnetyx using polymerase chain reaction (PCR).

Effect of interleukin-1 isoform neutralization on atherogenesis

Eight-weeks-old male ApoE^–/– mice in a C57BL/6 background (Jackson Lab) consumed a Western diet (TD.88137, Harlan Teklad) and received mouse anti-mouse monoclonal antibodies (IgG2a) that selectively neutralize IL-1α (Flo1-2a, Xbiotech), and/or IL-1β (01BSUR, Novartis), or isotype-matched control IgG2a (Novartis) (Supplementary material online, Figure S1A and Methods). These antibodies were administered subcutaneously at a dose of 10 mg/kg. Based on pharmacokinetic studies performed by the manufacturers, anti-IL-1β and control IgG2a were dispensed weekly, whereas anti-IL-1α was administered twice a week due to more rapid clearance. Mice were sacrificed (n = 15 per group) after 28 weeks of treatment (at 36 weeks of age) to evaluate the contribution of IL-1 isoforms to atherogenesis.

Effect of interleukin-1 isoform neutralization on the evolution of advanced-stage atheromata

Male ApoE^–/– mice in a C57BL/6 background (Jackson Lab) consumed a western diet (TD.88137, Harlan Teklad) from 8 weeks to 22 weeks of age to develop advanced-stage atherosclerotic lesions, followed by the administration of the different antibodies as described above (Supplementary material online, Figure S1B). After 14 weeks of treatment (36 weeks old), mice were sacrificed (n = 15–17 per group) to evaluate the effect of IL-1 isoform neutralization on already established advanced lesions.

Statistical analysis

The results are presented as mean ± standard deviation. The D’Agostino–Pearson test verified the normality of the distribution of each variable. Statistical significance was assessed by the ordinary one-way analysis of variance (ANOVA) for normally distributed variables and by the Kruskal–Wallis test for multiple comparisons for non-normally-distributed variables (P < 0.05 considered significant). Statistical testing used Graphpad Prism7 software.

Results

Neutralization of interleukin-1α or of both interleukin-1 isoforms limits early atherogenesis in the mouse aortic root

To determine whether IL-1 isoform neutralization affects the development of early-stage atheromata, 8-week-old male ApoE^–/– mice consumed a western diet and received treatment with either anti-IL-1α, anti-IL-1β, the combination of both antibodies, or control IgG from diet initiation as described in Methods section and Supplementary material online, Figure S1A. These treatments did not alter body or spleen weight; total, high density lipoprotein (HDL)- or low density lipoprotein (LDL)-cholesterol (Supplementary material online, Figure S2). Yet, neutralization of IL-1α or of both IL-1 isoforms increased plasma triglycerides (Supplementary material online, Figure S2C) as observed in humans with genetic gain of function of the IL-1Ra.¹⁴

Neutralization of IL-1α or of both IL-1 isoforms, but not of IL-1β alone, decreased the lesion area in the aortic roots compared with the control group (P = 0.021 and P = 0.002, respectively, Figure 1A and B). Glagovian remodelling affects the aorta,²⁸ and Alexander et al.¹³ demonstrated that total absence of IL-1 receptor 1 signalling impairs compensatory arterial enlargement in ApoE^–/– mice. Therefore, our study also evaluated the aortic roots by morphometry. Treatment with anti-IL-1α or with anti-IL-1α and anti-IL-1β reduced slightly the area within the internal elastica lamina (IEL, P = 0.039 and P = 0.017, respectively, Figure 1C), but did not alter the lumen area (Figure 1D). Collectively, these data demonstrate that neutralization of IL-1α alone or of both isoforms limits atheroma development and modestly impairs the compensatory outward remodelling of the aortic root during early atherogenesis.

Figure 1.

Neutralization of interleukin-1α or of both interleukin-1 isoforms during mouse atherogenesis reduces the lesion area in aortic root. (A) Elastica staining of the mouse aortic root, ×20, scale bars: 200 μm, (B) atherosclerotic lesion area, (C) area within the internal elastica lamina, and (D) lumen area. The results are expressed as mean ± standard deviation of n = 13–15 mice per group.

Neutralization of interleukin-1α or of both interleukin-1 isoforms impairs outward remodelling of the brachiocephalic artery during early atherogenesis

Selective or simultaneous neutralization of IL-1 isoforms did not modify atherosclerotic lesion area in the brachiocephalic artery (Figure 2A and B). Yet, the inhibition of IL-1α or of both isoforms decreased the area within the IEL (P = 0.005 and P = 0.020, respectively, Figure 2C) and the lumen area of the brachiocephalic artery compared with the control group (P = 0.001 and P = 0.002, respectively, Figure 2D). Micro-computed tomography (microCT) imaging after preserving the brachiocephalic artery caliber with a polymerized silicone cast deployed at physiologic pressure (Supplementary material online, Methods) verified that the observed lumen area reduction after neutralization of IL-1α or of both IL-1 isoforms did not result from post-mortem tissue shrinkage (Figure 2E and F). These results corroborated the immunohistochemical data and demonstrated that inhibition of IL-1α or of both IL-1 isoforms—but not of IL-1β alone—impairs the Glagovian outward remodelling of the mouse brachiocephalic artery.

Figure 2.

Early neutralization of interleukin-1α or of both isoforms impairs the compensatory outward remodelling of mouse brachiocephalic artery during atherogenesis. (A) Elastica staining of the brachiocephalic arteries, ×20, scale bars: 200 μm, (B) atherosclerotic lesion area, (C) area within the internal elastica lamina, (D) lumen area, (E) micro-computed tomography images of brachiocephalic artery, three-dimensional rendering on left panel, transverse view on the top right panel, scale bars: 500 μm, and longitudinal view on the bottom right panel, scale bars: 100 μm, for each group, and (F) lumen area measured using micro-computed tomography images from 0 mm (aortic arch) to 2 mm up of the brachiocephalic artery. Data represent mean ± standard deviation of n = 13–15 mice per group.

Interleukin-1β neutralization augments Ly-6C^low monocyte count in mouse plasma during atherogenesis

Evaluation of the effect of IL-1 isoform neutralization on systemic inflammation involved quantification of total leucocytes, neutrophils, eosinophils, and Ly-6C^high (‘classical’) and Ly-6C^low (‘patrolling’) monocyte subsets in mouse plasma by flow cytometry. The selective inhibition of IL-1β augmented the Ly-6C^low monocyte count (P = 0.005, Figure 3A and E) and showed a trend towards an increased eosinophil count in plasma (Figure 3D). None of the antibody treatments modified significantly the total counts of leucocytes, neutrophils, or Ly-6C^high monocytes. These results indicate that IL-1β neutralization shifts blood monocytes towards a less inflammatory state in hypercholesterolemic ApoE^–/– mice.

Figure 3.

Neutralization of interleukin-1β during atherogenesis increases the circulating Ly-6C^low monocyte count and reduces the immune activation status of brachiocephalic artery lesions. (A) Flow cytometry of blood. Viable neutrophils were identified as CD90⁻ CD19^– CD115^– CD11b⁺ LY6G⁺ cells and monocytes as CD90^– CD19^– LY6G^– CD11b⁺ CD115⁺ cells. Quantification by flow cytometry of blood (B) leucocytes, (C) neutrophils, (D) eosinophils, (E) Ly-6C^high monocytes, and (F) Ly-6C^low monocytes. (G) Plasma interleukin-10 quantification by ELISA, (H) MHC II staining of the brachiocephalic arteries, ×20, scale bars: 100 μm, and (I) quantification of MHC II staining in the brachiocephalic arteries. The results expressed as are mean ± standard deviation of n = 11–15 mice per group.

The selective neutralization of interleukin-1β decreases MHC Class II expression in the brachiocephalic artery during early atherogenesis

The profile shift from pro-inflammatory to reparative monocytes elicited by anti-IL1β treatment prompted the measurement of plasma levels of the anti-inflammatory cytokine IL-10 and the examination of major histocompatibility complex (MHC) Class II expression in brachiocephalic artery plaques. The selective neutralization of IL-1β decreased MHC Class II expression in atherosclerotic lesions (P = 0.036, Figure 3H and I) and tended to increase plasma IL-10 levels (Figure 3G). The selective neutralization of IL-1α tended to increase plasma IL-10 (Figure 3G) and to decrease MHC Class II, but these trends did not reach statistical significance (Figure 3H and I). Together, these results indicate that IL-1β inhibition decreases immune activation during atherogenesis in ApoE^–/– mice.

Interleukin-1β neutralization decreases the area of established atheromata in the aortic root

The administration of anti-IL-1 isoform antibodies to mice after 22 weeks of western diet consumption enabled evaluation of the effect of these treatments on the evolution of established atheromata, which mimics the situation in secondary prevention (Supplementary material online, Figure S1B). These treatments did not influence body or spleen weights, plasma triglycerides, total-, HDL-, or LDL-cholesterol (Supplementary material online, Figure S3). The selective neutralization of IL-1β but not of IL-1α decreased the area of established aortic root plaques compared with the control group (P = 0.037, Figure 4B). Neutralization of both isoforms showed a trend towards decreased lesion area, but this reduction did not reach statistical significance (Figure 4B).

Figure 4.

Neutralization of interleukin-1β reduces the lesion area of established atheromata in aortic root. (A) Elastica staining of the mouse aortic root, ×20, scale bars: 200 μm, (B) atherosclerotic lesion area, (C) area within the internal elastica lamina, and (D) lumen area. The results are expressed as mean ± standard deviation of n = 16–17 mice per group.

Interleukin-1 isoform neutralization does not alter Glagovian remodelling of brachiocephalic arteries or aortic roots in mice with established atheromata

None of the antibody treatments administered during advanced disease impaired outward remodelling of the aortic root or the brachiocephalic artery in these experimental settings, as assessed by measurement of IEL and lumen areas (Figure 4C, D, 5C, and D). microCT imaging verified that IL-1 isoform neutralization at an advanced stage did not modify the lumen area of the brachiocephalic artery (Figure 5E and F). Neutralization of both IL-1 isoforms together but not individually increased the counts of total leucocytes and Ly-6C^low monocytes (P = 0.043 and P = 0.044, respectively, Figure 6A and E) in mice with established atheromata.

Figure 5.

Interleukin-1 isoform neutralization does not influence the compensatory outward remodelling the brachiocephalic artery during the evolution of established atheroma. (A) Elastica staining of the brachiocephalic arteries, ×20, scale bars: 200 μm, (B) atherosclerotic lesion area, (C) area within the internal elastica lamina, (D) lumen area, (E) micro-computed tomography images of brachiocephalic artery, three-dimensional rendering on left panel, transverse view on the top right panel, scale bars: 500 μm, and longitudinal view on the bottom right panel, scale bars: 100 μm, for each group, and (F) lumen area measured using micro-computed tomography images from 0 mm (aortic arch) to 2 mm up of the brachiocephalic artery. Data represent mean ± standard deviation of n = 14–17 mice per group.

Figure 6.

Effect of interleukin-1 isoform neutralization on circulating cells and on the immune activation status of brachiocephalic established atheroma. (A) Flow cytometry of blood. Viable neutrophils were identified as CD90^– CD19^– CD115^– CD11b⁺ LY6G⁺ cells and monocytes as CD90^– CD19^– LY6G^– CD11b⁺ CD115⁺ cells. Quantification by flow cytometry of blood (B) leucocytes, (C) neutrophils, (D) eosinophils, (E) Ly-6C^high monocytes, and (F) Ly-6C^low monocytes. (G) Plasma interleukin-10 quantification by ELISA, (H) MHC II staining of the brachiocephalic arteries, ×20, scale bars: 100 μm, and (I) quantification of MHC II staining in the brachiocephalic arteries. The results are expressed as mean ± SD of n = 12–15 mice per group.

Interleukin-1β neutralization in mice with established atheroma increases plasma interleukin-10 levels

The neutralization of IL-1β at advanced stages of atherosclerosis increased IL-10 plasma levels compared with the control group (P = 0.029, Figure 6G). None of the antibody treatments administered at an advanced stage of disease decreased MHC Class II expression in brachiocephalic artery lesions (Figure 6H and I).

Discussion

Despite a decades long recognition of IL-1 as a modulator of atherosclerosis, recent studies have engendered debate about the functions of the two isoforms of this pro-inflammatory cytokine. These unsettled issues have importance for understanding the mechanisms of inflammation in atherosclerosis, and also from a therapeutic perspective regarding which isoform(s) to target when treating atherosclerosis and its complications. Moreover, the Owens group has indicated that total blockade of IL-1 signalling¹³ or selective neutralization of IL-1β²⁹ can limit outward remodelling of arteries during mouse atherosclerosis, and even appears to confer acquisition by plaques of rupture-prone characteristics. Yet, CANTOS established that the selective neutralization of the IL-1β isoform decreases recurrent cardiovascular events in patients with a previous myocardial infarction and residual inflammatory risk, as indicated by above median plasma C-reactive protein concentration despite standard care including statin treatment.^24–26 The present study aimed to aid the interpretation of the results of CANTOS and guide future human trials that translate inflammation biology to patients with or at risk for atherosclerosis. The selective IL-1 isoform neutralization used here mimicked the human therapeutic situation, and obviated the limitations of prior germline modification experiments due to possible compensatory adjustments to congenital absence of IL-1 signalling.^13,23

The present study implicated IL-1α as a participant in the outward arterial remodelling of the brachiocephalic artery during the initial formation of mouse atheromata. This finding highlights the distinct biochemical, cell biological, functional, and pathophysiologic properties of IL-1α and IL-1β in vivo, despite their redundant action on several aspects of vascular cell activation in vitro.^8,30 Arterial shear stress, one of the main factors inducing outward remodelling,^28,31 elicits IL-1α production by endothelial cells, leading to SMC chemotaxis, and the release of MMP-3, a key proteinase in the process of outward remodelling.^8,13,32 The early neutralization of IL-1α also decreased plaque SMC content (Supplementary material online, Figure S4A). These observations support the hypothesis that IL-1α neutralization impedes SMC chemotaxis and production of MMP-3, effects that could impair outward remodelling.

The present study also showed an increase of plasma triglycerides associated with neutralization of IL-1α or of both IL-1 isoforms. This observation agrees with previous studies that demonstrated that anti-IL-1α decreases leptin expression, triggering an increase of oleate incorporation into triglycerides, and increase of plasma triglycerides.^33,34 In CANTOS, IL-1β inhibition produced a negligible increase in triglycerides.

This study documented that neutralization of IL-1α or of both IL-1 isoforms, but not of IL-1β alone, reduces the lesion area in the aortic roots but does not affect the atheroma size in the brachiocephalic arteries during mouse atherogenesis. These results conform with previous publications that demonstrated that deficiency of IL-1α—but not of IL-1β—in haematopoietic cells decreases atherogenesis in mouse aortic roots,^22,23 and that IL-1β neutralization does not modify the lesion area in the brachiocephalic artery of atherosclerosis-susceptible mice consuming a western diet.^19,20 The present data demonstrate that IL-1β neutralization augments Ly-6C^low monocytes in blood, increases IL-10 plasma levels, and reduces MHC Class II expression in atherosclerotic lesions, indicating a muting of both innate and adaptive immunity implicated in atherogenesis. The selective neutralization of IL-1α also tended to increase IL-10 plasma levels and to decrease MHC Class II expression. Previous in vitro studies have shown that IL-1α- or IL-1β-deficient macrophages increase release of IL-10.²²

This study also examined the effect of IL-1 isoform neutralization on the evolution of established mouse atheromata, an approach that approximates more closely the clinical application of anti-IL-1 therapies in humans with cardiovascular diseases. Neutralization of IL-1β decreased the area of established lesions in aortic roots and increased plasma levels of IL-10, whereas IL-1α neutralization had no effect on the characteristics of established atheromata. The different results of IL-10/MHC-II between the two distinct studies reflect the complexity of immune responses. Indeed, IL-10 inhibits MHC Class II expression in macrophages, and promotes recruitment of reparative monocytes, which in turn produce IL-10.³⁵ Consistent with published clinical data showing that anti-IL-1β treatment with canakinumab did not change arterial structure in patients with atherosclerosis,³⁶ IL-1β neutralization did not affect the morphometric parameters of the aortic roots or the brachiocephalic arteries in mice with established plaques. These mechanistic differences can result from the distinct signalling mechanisms of the two isoforms. Indeed, IL-1α is mainly cell membrane associated, and operates primarily via juxtacrine signalling, whereas IL-1β is released and acts through paracrine or endocrine signalling.³⁷

Advantages of the current study include the use of perfusion fixation at physiologic pressure and injection of a radiopaque resin into the arterial system, procedures intended to avoid artefactual alterations in arterial anatomy post-mortem such as recoil or compression during cutting. The use of two independent methods for morphometric analysis of arterial tissues, classical histologic, and microCT evaluation enhanced confidence in our conclusions regarding arterial remodelling and lumen caliber. We also went to lengths to study both early atherogenesis and initiation of anti-IL-1 treatment after establishment of atherosclerotic plaques. Yet, as with all experimental studies, this investigation has limitations. We used an atherogenic diet that produces cholesterol concentrations that far exceed those encountered in contemporary patients.³⁸ These experimental conditions permitted us to compare our results with those of other laboratories. Yet, they may have contributed to some seemingly paradoxical findings such as the increase in leucocyte count with IL-1 inhibition. Hypercholesterolaemia can drive leucopoiesis,³⁹ and the altered baseline under the experimental conditions of these experiments might explain why IL-1 inhibition raised leucocyte counts in the mice with prolonged elevation in lipid levels, while white cell numbers fall slightly in humans treated with anti-IL-1 therapies.²⁴ The use of perfusion fixation, mandated by the key goal of rigorous study of arterial remodelling, precluded certain analyses that could deepen mechanistic understanding of the clear distinctions in IL-1 isoform effects demonstrated here, a subject for future study in unfixed specimens.

Collectively, the results of this study provide insight into the mechanisms whereby the selective neutralization of IL-1β in CANTOS reduced atherosclerotic complications. These mouse experiments provide evidence that IL-1β blockade shifts blood monocytes towards a less inflammatory state during atherogenesis, reduces the size of established atheromata, and increases plasma levels of IL-10 without affecting the expansive remodelling of brachiocephalic arteries. Inhibition of IL-1α or a total blockade of IL-1 signalling did not offer advantages in these respects over selective targeting of IL-1β. These results shed light into some apparent discrepancies between mouse and human studies, and help inform further translation of IL-1 treatments.

Take home figure.

Stage-dependent differential effects of interleukin-1 isoforms on experimental atherosclerosis.

See page 2492 for the editorial comment on this article (doi: 10.1093/eurheartj/ehz133)