A VLP-based vaccine against interleukin-1α protects mice from atherosclerosis


Full Correspondence: Dr. Martin F. Bachmann, Department of Dermatology, University Hospital Zürich, Gloriastr 30, 8091 Zürich, Switzerland

Fax: +41-442551112

e-mail: martin.bachmann@me.com


Interleukin (IL)-1α is a potent proinflammatory cytokine that has been implicated in the development of atherosclerosis. We investigated whether a vaccine inducing IL-1α neutralizing antibodies could interfere with disease progression in a murine model of atherosclerosis. We immunized Apolipoprothin E (ApoE)-deficient mice with a vaccine (IL-1α-C-Qβ) consisting of full-length, native IL-1α chemically conjugated to virus-like particles derived from the bacteriophage Qβ. ApoE−/− mice were administered six injections of IL-1α-C-Qβ or nonconjugated Qβ over a period of 160 days while being maintained on a western diet. Atherosclerosis was measured in the descending aorta and in cross-sections at the aortic root. Macrophage infiltration in the aorta was measured using CD68. Expression levels of VCAM-1, ICAM-1, and MCP-1 were quantified by RT-PCR. Immunization against IL-1α reduced plaque progression in the descending aorta by 50% and at the aortic root by 37%. Macrophage infiltration in the aorta was reduced by 22%. Inflammation was also reduced in the adventitia, with a decrease of 54% in peri-aortic infiltrate score and reduced expression levels of VCAM-1 and ICAM-1. Active immunization targeting IL-1α reduced both the inflammatory reaction in the plaque as well as plaque progression. In summary, vaccination against IL-1α protected ApoE−/− mice against disease, suggesting that this may be a potential treatment option for atherosclerosis.


Atherosclerosis is a disease of the vasculature, causing thickening of the arterial wall, vascular remodeling, obstruction of the vessel lumen, and abnormal blood flow, and eventually diminished oxygen supply to target organs [1, 2]. Injury and/or activation of the endothelium, in conjunction with accumulation of atherogenic lipoproteins, are thought to be initial events leading to plaque formation. Oxidation of lipoproteins and their uncontrolled uptake by macrophages results in foam cell formation. Cholesterol laden foam cells accumulate into a fatty streak, which develops further into an intermediate lesion subsequent to smooth muscle cell, lymphocyte and additional monocyte, and macrophage recruitment [1, 2]. Proinflammatory molecules aggravate the chronic inflammatory reaction developing into the classical atherosclerotic plaque through macrophage and lymphocyte activation, and activation of the endothelium. This process further fosters oxidation of lipoproteins and increases inflammatory cell recruitment [2, 3].

IL-1 is a potent proinflammatory cytokine with pleiotropic functions [4, 5]. There are two subtypes, interleukin-1α (IL-1α) and interleukin-1β (IL-1β). Both proteins lack a signal peptide, and are first synthesized as precursor proteins [4]. The precursor form of IL-1α is biologically active, while the corresponding pro-IL-1β is not. IL-1β is processed into its mature form by caspase-1 and secreted upon inflammatory stimuli, in particular upon activation of the inflammasome [6]. IL-1α is predominantly found in its precursor form in the cytosol, or is membrane-bound [7]. Intracellular IL-1α has been reported to migrate into the nucleus, where it is thought to act as a transcriptional regulator. The prosequence can be cleaved by calpain, a membrane-associated protease, releasing mature IL-1α in the extracellular space [8]. IL-1α can, however, also be released into the extracellular space during apoptotic or necrotic processes. Interestingly, although IL-1α is not a substrate of caspase-1, the presence of the active enzyme is nevertheless required for release of IL-1α from the cytoplasm into the extracellular space [9-11]. In contrast, production of membrane-bound IL-1α occurs independently of inflammasome activation [12]. IL-1α or IL-1β trigger signaling by binding to a complex of IL-1 receptor I and IL-1 receptor accessory protein [4]. Two counter-regulatory mechanisms exist. A decoy receptor (IL-1 receptor II), which is homologous to IL-1RI but lacks the intracellular signaling domain, competes for binding of IL-1 on many cell types. In addition an endogenous soluble antagonist, the IL-1 receptor antagonist (IL-1Ra), binds with high affinity to IL-1 receptor I, without recruiting the accessory protein and thereby inhibits signaling [4]. IL-1β can be detected at low levels in peripheral blood, whereas there appears to be less systemic availability of IL-1α. IL-1α might however be released locally in significant amounts [13], in particular as a consequence of local cell necrosis [14] or inflamasome activation [12, 15, 16]

Several studies have addressed the role of IL-1 in atherosclerosis. IL-1RI deficiency has been shown to decrease the severity of atherosclerosis in homozygous Apolipoprothin E (ApoE)-deficient mice fed a high fat “western” diet [17] or in heterozygous ApoE+/− mice fed a high fat diet and/or exposed to bacterial challenge [18]. Likewise, lack of IL-1β led to a reduction in atherosclerosis in ApoE-deficient mice fed normal chow [19]. Moreover, heterozygous deficiency in IL-1Ra in ApoE-deficient mice fed normal chow was proatherogenic and led to increased accumulation of macrophages in the lesions, while transgenic overexpression of the antagonist in ApoE KO mice fed a high cholesterol diet reduced the extent of atherosclerosis [20]. Interestingly, homozygous deficiency in IL-1Ra in ApoE KO mice fed cholesterol-rich diet led to massive inflammatory damage and infiltration in the absence of atherosclerosis [20]. Finally, a study specifically looking at the role of IL-1α in atherosclerosis has shown that IL-1α deficiency leads to a significant reduction in plaque progression in C57/BL6 mice fed a high fat diet [21].

The low systemic availability of IL-1α, and the possibility that significant amounts of this cytokine are released locally during inflammatory and/or necrotic processes occurring in the atherosclerotic plaque, render it an attractive target for a selective prevention or treatment of atherosclerotic diseases [22]. The chronic and inflammatory nature of atherosclerosis, the need for long-term treatment and the high prevalence of atherosclerotic diseases make immunotherapy targeting IL-1α an attractive treatment option [23]. In the present study, we investigated whether neutralization of IL-1α could reduce atherosclerosis and vascular inflammation in ApoE KO mice fed a high fat diet. We immunized mice with a vaccine against IL-1α we have recently described [24, 25], which is based on virus-like particles derived from the RNA phage Qβ chemically conjugated to IL-1α (IL-1α-C-Qβ). This vaccine induces antibodies that specifically neutralize the biological activity of IL-1α without cross-reacting with either IL-1β or IL-1Ra [24, 25].


Vaccination against IL-1α results in sustained antibody titers

Although an important role for IL-1α in atherosclerosis is becoming increasingly clear, a demonstration that blocking IL-1α can inhibit disease progression is still lacking. To this end, and in order to test the potential of vaccination as a therapy for atherosclerosis, we immunized ApoE KO mice twice (day 0, 14) with the recently described IL-1α-C-Qβ vaccine or with nonconjugated control Qβ virus-like particle (VLP) [24, 25]. IL-1α-C-Qβ has been shown to generate a high-titer and neutralizing anti-IL-1α response [24]. High IL-1α-specific antibody titers were achieved by day 21 (Fig. 1A), at which time the diet was switched from normal chow to a cholesterol-rich diet. The antibody response was further increased by a third injection on day 28 (Fig. 1A). Thereafter the animals received three more injections to maintain a high titer at throughout the remainder of the study until day 159.

Figure 1.

Treatment of ApoE−/− mice with IL-1α-C-Qβ or Qβ. (A) Anti-IL-1α antibody titer measured by ELISA as OD 50%, the serum or plasma dilution giving half-maximal binding in the assay. (B) Plaque load in the descending aorta quantified by image analysis in aortas prepared “en face” is shown for ApoE−/− mice treated with IL-1α-C-Qβ or Qβ. Data are shown as median with box and whisker plots representing interquartile range and maximum and minimal values (whiskers). Significance was evaluated in comparison to controls p = 0.008, n = 13 for IL-1α-C-Qβ and n = 10 for Qβ, Mann–Whitney test. (C) The plaque area was quantified by image analysis in cryosections stained with oil red O and compared to controls p = 0.0005, t-test, n = 13 for IL-1α-C-Qβ, and n = 11 for Qβ. (D) Total cholesterol and triglyceride levels in IL-1α-C-Qβ-treated or Qβ-treated mice (IL-1α-C-Qβ (n = 13) and Qβ (n = 12)). (A, C and D) Data are shown as mean + SEM. Significant differences are labeled with an asterisk.

Vaccination against IL-1α reduces plaque load in ApoE−/− mice

The effect of inhibiting IL-1α on atherosclerosis was assessed by analyzing the plaque burden in both the descending aorta and at the root of the aorta. We observed a 50% reduction in plaque load in the descending aorta prepared “en face” (p = 0.008, Fig. 1B) in one experiment, and 31% in another experiment. For the latter, at the root of the aorta, the plaque area in cryo-sections stained by oil red O was reduced by 37% (p = 0.0005, Fig. 1C) in animals vaccinated against IL-1α compared to Qβ vaccinated controls. The reduction in the extent of atherosclerosis was not due to an imbalance in total cholesterol or triglyceride levels, which did not differ significantly between the groups (Fig. 1D).

Vaccination against IL-1α results in reduced expression of inflammatory marker

One way in which neutralization of IL-1α could affect atherosclerotic plaque growth is through inhibition of endothelial cell activation. This would reduce expression of cellular adhesion molecules and hence lead to decreased leukocyte recruitment to the plaques. We quantified the expression of VCAM-1 and ICAM-1 in the aortic arch by RT-PCR, and found a statistically significant reduction of 42% for VCAM-1 (p = 0.01), and 32% for ICAM-1 (p = 0.01) compared with Qβ controls (Fig. 2). In addition to cellular adhesion molecules, the chemokine MCP-1 plays a major role in monocyte recruitment and their differentiation to macrophages. We observed a trend toward reduced expression of MCP-1 in IL-1α-C-Qβ-immunized animals as compared with those of controls (−30%, p = 0.09, Fig. 2). There was no significant reduction in expression of TNF-α (Fig. 2).

Figure 2.

Quantification of expression of TNF-α, ICAM-1, MCP-1, and VCAM-1 in IL-1α-C-Qβ and Qβ-treated ApoE−/− mice by RT-PCR. Expression levels are expressed in arbitrary units, relative to β-actin expression. Expression of VCAM-1 (p = 0.01), ICAM-1 (p = 0.01), MCP-1 (p = 0.09), and TNF-α was compared with those of controls. Data are shown as mean + SEM (n = 6 per group). Significant differences evaluated with t-test are labeled with an asterisk.

Vaccination against IL-1α reduces cellular infiltration of plaques

To investigate whether a reduction in cellular adhesion molecules and possibly in MCP-1 expression translated into a reduction of the macrophage content of the plaques, we stained aortic root cross-sections for CD68, a lysosomal glycoprotein present in macrophages and foam cells. The extent of staining was quantified in the intimal and medial layers of the vasculature. We observed a reduction in the CD68 stained average area of 22% in animals vaccinated against IL-1α compared with that of controls (p = 0.04, Fig. 3A). Interestingly, the average total aortic cross-sectional area was also decreased upon vaccination (−16%), however this reduction was not statistically significant. The reduction in macrophage content of the plaques was further investigated by assessing the F4/80 expression in the aortic arch by RT-PCR. A reduction of 41%, (p = 0.02, Fig. 3B) was observed, confirming the result obtained by immunohistochemistry.

Figure 3.

Quantification of macrophage infiltration by histology (CD68 stained area) and gene expression (F4/80) in ApoE−/− mice treated with IL-1α-C-Qβ and Qβ. (A) The CD68-positive area was calculated in sections from IL-1α-C-Qβ treated mice and Qβ controls as described under Methods. Data are shown as mean + SEM of n = 12 (IL-1α-C-Qβ treated) and n = 8 (Qβ controls) mice (p = 0.04, t-test). (B) Quantification of F4/80 expression in the aortic arch by RT-PCR (p = 0.02) is shown. Significant differences evaluated with t-test are labeled with an asterisk.

Our initial investigation of the gross pathology of the lesions was performed using hematoxylin and eosin stained sections. A striking feature observed in control animals were strong peri-adventitial inflammatory infiltrates (Fig. 4A), which appeared reduced in vaccinated animals (Fig. 4B). We therefore scored the infiltrates, and observed a reduction in median score of 54% (from 1.3 to 0.6, p = 0.0003, Fig. 4C) between controls and IL-1α vaccinated animals. Vaccination against IL-1α therefore reduced the extent of vascular inflammation both in the neointima and the adventitia.

Figure 4.

Hematoxylin and eosin staining of cryosections cut at the root of the aorta. (A) Cryosections from the aortic root of (A) a Qβ-treated animal and (B) a IL-1α-C-Qβ vaccinated animal are shown. Original micrograph with objective 20×. (C) Median inflammatory scores for IL-1α-C-Qβ vaccinated animals (n = 13) and Qβ controls (n = 12) are shown in a box and whiskers plot (median, interquartile range, and whiskers as minima and maxima; p = 0.0003, Mann–Whitney test). Significant difference is indicated by an asterisk in the graph.


Although the importance of the IL-1 pathway for the development of atherosclerosis has been inferred from previous studies with genetically engineered mouse strains [3-7, 15], our study is the first to demonstrate the potential of pharmacological inhibition of IL-1α for the treatment of atherosclerosis. We used a vaccine to raise neutralizing antibodies to inhibit IL-1α and the titers raised were of high magnitude, and could be maintained throughout the study by booster administrations. As has been shown before [24, 25], linkage of IL-1 to the virus-like particle derived from the RNA phage Qβ thus allowed to overcome B-cell unresponsiveness and induced an antibody response against the self-antigen. Vaccination against IL-1α resulted in a consistent and robust reduction in plaque load both at the root of the aorta, where atherosclerosis develops early due to disturbed flow patterns, as well as in the descending aorta, where lesions may develop later. Neutralization of IL-1α led to a reduction in expression of cellular adhesion molecules and recruitment of monocytes and possibly other leukocytes to the plaque. Inhibition of IL-1α therefore targets the key processes in atherogenesis. Moreover we observed reduced expression of VCAM-1 and ICAM-1 in the aorta. The interaction between VCAM-1 and its ligand VLA-4, expressed on leukocytes, has been shown to play an important role in atherosclerosis. Treatment of ApoE−/− mice with an antibody against VLA-4 has been shown to lead to reduce recruitment of macrophages to the plaques [26]. Furthermore, LDLr−/− mice with a genetically modified form of VCAM-1 which exhibit drastically reduced levels of VCAM-1 mRNA and protein expression, display reduced lesion size compared with controls [27]. Deficiency in ICAM-1 also appears to protect against atherosclerosis, and antibody treatment against ICAM-1 has been shown to reduce short-term recruitment of macrophages to the plaque [26]. Both expression of VCAM-1 and ICAM-1 is induced by oxidized lipids such as oxLDL. By neutralizing IL-1α, antibodies induced by IL-1α-C-Qβ significantly reduced the expression of both adhesion molecules. Although the reduction in MCP-1 was a trend, it is consistent with a mechanism of action whereby targeting IL-1α prevents recruitment of leukocytes and in particular monocytes to the plaque by reducing expression of VCAM-1, ICAM-1, and MCP-1. MCP-1 plays a key role in the recruitment of monocytes, which differentiate to macrophages in the neointima. The reduction in macrophage or foam cell content of the plaques was shown by immunohistochemistry using an antibody against CD68 and by RT-PCR using F4/80-specific primers. The reduction in cell adhesion molecule expression therefore correlated with a reduction of the recruitment of monocytes and ultimately macrophages to the plaque. Neutralization of IL-1α did not have an effect on total cholesterol, providing the opportunity to investigate the role of IL-1α without perturbing cholesterol levels.

T and B lymphocytes do not appear to play a major role in the development of plaque pathology under high cholesterol western diet in ApoE−/− mice [28]. The reduction in plaque pathology we observed under high hypercholesteremia therefore suggests a proximal role for IL-1α in the vascular innate inflammatory and remodeling process triggered by cholesterol. Together with macrophages, T and B cells reside in the adventitia of healthy vessels [29, 30]. Moreover, IL-1α has been implicated in the induction of T-cell responses [13, 28]. Despite the minor role postulated for T and B lymphocytes on plaque pathology in the high fat ApoE−/− model [28], it is interesting to note that we observed a very prominent reduction in peri-adventitial inflammation upon neutralization of IL-1α. Peri-adventitial T lymphocytes have been implicated in the pathology of hypertension [31], and it is an intriguing possibility that modulation of peri-adventitial T-cell responses could play a significant role under more modest hypercholesteremia than in our experimental model, such as is the case in human disease. In particular, IL-1α could trigger plaque instability by its proinflammatory action. Thus, blocking IL-1α may not only halt or delay plaque formation but may also prevent cardiovascular complications by promoting plaque stability.

Materials and methods

Vaccine preparation

IL-1α-C-Qβ was prepared as described [24], taking particular care to remove free IL-1α by tangential flow filtration.

Animal experiments

Male ApoE−/− mice, approximately 8 weeks of age, were immunized subcutaneously on days 0, 14, 21, 28, 56, 94, and 122 or 0, 14, 28, 56, 84, 112, and 140 with 50 μg of IL-1α-C-Qβ (n = 13) or Qβ-VLP (n = 12 or 10). Mice were fed a normal chow between day 0 and 20, and a western diet (20% fat, 0.15% cholesterol, no cholate, Provimi Kliba AG, Switzerland, or 21.2% fat, 0.2% cholesterol, Harlan, Netherlands) from day 21 to the end of the experiment. Mice were bled at regular intervals throughout the experiment and the antibody response against IL-1α was measured in plasma. Sacrifice was on day 159, and the aorta were isolated and prepared essentially as described [32]. The animals were bled by cardiac puncture and perfused with cold PBS. The aorta was sectioned 2 mm from the heart and further processed. In addition, hearts were removed and snap-frozen in liquid nitrogen for subsequent histologic preparation. The aortic arch was sectioned 5 mm down from the left subclavian artery and immediately frozen in liquid nitrogen. The heart was sectioned in the middle, and the upper part was immediately frozen in Hank's balanced salt solution in a plastic tube in liquid nitrogen for histological analysis essentially as described [33, 34]. The serial sections (7 μm thickness) were cut through the origin of the aorta and harvested upon appearance of at least two valve cusps, until disappearance of the last valve cusps. Sections were fixed in formalin, stained with oil red O, and plaque load was evaluated in 4–7 sections per mouse (for one animal of the Qβ group only three sections were available) by quantitative image analysis (Image J). An average plaque area was computed for each animal, and an average group plaque area was computed for the IL-1α-C-Qβ and Qβ group, respectively. Statistical analysis was performed with a Student's t-test.

For the evaluation of atherosclerosis in the descending aorta, these were further cleaned from residual adventitia on a glass petri dish filled with cold PBS. The aorta were cut longitudinally, pinned out on a black wax surface, and fixed overnight in 4% formalin. They were then stained overnight in oil red O. The plaques were quantified with imaging software on digital photographs. The plaque load was expressed as the sum of the surface of all plaques of the aorta taken up to the iliac bifurcation, divided by the total surface of the aorta measured up to the iliac bifurcation, in percentage. The difference in median plaque load between the IL-1α-C-Qβ and Qβ groups was analyzed using a Mann–Whitney test.

Inflammatory infiltrates in the adventitia were scored by a pathologist, using a scoring scheme ranging from 0 to 4 in 0.5 increments. Two rounds of scoring were performed, each with the sections blinded and ordered in a different sequence. Three to four sections per animal were quantified, with the exception of one animal in each group where only two sections were available. An average score was computed, and the data analyzed with a Mann–Whitney test.

CD-68 staining on cryosections was performed by shortly fixing the sections in formalin. Peroxidase activity was blocked by incubation with H2O2. The CD68-positive area was quantified in the intima and media of three to four sections per animal (for two animals of the IL-1α-C-Qβ group two sections only were available). Average plaque area was computed and statistical analysis performed with a Student's t-test.


The antibody titer against IL-1α was measured by ELISA as described [24]. Recombinant IL-1α [24] was coated on Maxisorp plates at 1 μg/mL, and serial dilutions of serum from IL-1α-C-Qβ-immunized mice were applied to the plates. Detection was with a horseradish peroxidase-conjugated goat anti-mouse IgG secondary antibody (Jackson Immuno Research Laboratories) at a dilution of 1:1000 in PBS/0.05% Tween/2% BSA. The color reaction was developed with 1,2-phenylenediamine dihydrochloride (0.4 mg/mL in 0.066 M Na2HPO4, 0.035 M citric acid, 0.01% H2O2, pH 5.0). OD450 was determined using an ELISA reader (Bio-Rad). Titers were expressed as those serum dilutions that lead to half-maximal OD450 (OD50%).

Cholesterol and triglyceride measurements

Cholesterol and triglycerides were analyzed as described [35].


RT-PCR was performed using the Stratagene Brilliant SybrGreen kit according to the manufacturer instructions. Aortic arches were lysed in lysis buffer and reverse transcribed usingprimer pairs described at http://pga.mgh.harvard.edu/primerbank/. A standard was generated for each tested cDNA, including β-actin, and used for quantification in every PCR reaction. Two independent PCR reactions were run for each cDNA. Gene expression was quantified relative to β-actin expression, and expressed in arbitrary units.

Statistical analysis

p < 0.05 was considered statistically significant. Data not normally distributed were analyzed by Mann–Whitney test, normally distributed data by t-tests.


This work was supported by grants from by Cytos Biotechnology AG.

Conflict of interest

ACT, GS, GJT, AS, RW, MM, MV, and MFB are former employees of Cytos Biotechnology AG and used to or currently have shares or rights to shares of the company.