Author + information
- Received October 11, 2005
- Revision received October 6, 2006
- Accepted November 16, 2006
- Published online August 7, 2007.
- Giuseppina Caligiuri, MD, PhD⁎,⁎ (, )
- Jamila Khallou-Laschet, PhD⁎,
- Marta Vandaele, MSc⁎,
- Anh-Thu Gaston, BSc⁎,
- Sandrine Delignat, BSc⁎,
- Chantal Mandet, BSc†,
- Heinz V. Kohler, MD, PhD‡,
- Srini V. Kaveri, DVM, PhD⁎ and
- Antonino Nicoletti, PhD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Giuseppina Caligiuri, INSERM UMR S 872, équipe 16, Centre de Recherche des Cordeliers, 15, rue de l’Ecole de Médecine, Paris, F-75006, France.
Objectives The present study evaluated the effect of phosphorylcholine (PC) immunization on the extent of experimental atherosclerosis.
Background Immunization against oxidized lipoprotein (oxLDL) or Streptococcus pneumoconiaereduces atherosclerosis. Phosphorylcholine is the main epitope recognized by both antipneumococcus and anti-oxLDL antibodies. Therefore we reasoned that PC-specific antibodies might play an important role in atherogenesis.
Methods Apolipoprotein E knockout mice were immunized with PC every second week over 4 months. At the end of the study, serum antibodies directed to either PC or oxLDL were measured. Splenic and peritoneal B cells were analyzed by flow cytometry. Aortic root atherosclerotic lesions were quantified by morphometry and phenotyped by immunohistochemistry. Immune and control sera were also tested for their effect on foam cell formation in macrophage culture in the presence of oxLDL.
Results The PC-immunized mice showed 3-fold increase in titers of anti-PC and -oxLDL antibodies compared with control mice (p < 0.01). The PC-immunized mice also showed a significant increase in the number of splenic mature B cells. The extent of atherosclerotic aorta root lesions was reduced by >40% in the PC-immunized mice (p < 0.01). Immunohistochemistry showed reduced expression of major histocompatibility complex class II antigens (p < 0.05) and the presence of B-cell clusters in plaques of PC-immunized mice. Finally, PC-immune serum was able to reduce macrophage-derived foam cell formation in the presence of oxLDL in vitro.
Conclusions Phosphorylcholine immunization drives a specific humoral immune response that reduces foam cell formation in vitro and is atheroprotective in vivo.
Antibodies directed to oxidized low-density lipoproteins (oxLDL) are a hallmark of atherosclerosis (1), and immunization to oxLDL reduces experimental atherosclerosis (2,3) independently of CD4+T-cell help (4). Selected peptide sequences from the protein of LDL (apoB-100) induce atheroprotective antibodies and have been proposed as immunization tools for coronary heart disease (5). However, major histocompatibility complex (MHC) polymorphism will impede the use of generic vaccinating peptides. Furthermore, the main epitope of anti-oxLDL antibodies is not contained in the protein portion of the molecule but rather in the phospholipids and has been identified as phosphorylcholine (PC) (6).
In recent years, a large amount of research has focused on the immune-inflammatory response associated with atherosclerosis and its clinical manifestations. Whereas the T-cell–mediated arm of the specific immunity seems to play a proatherogenic role (4,7–13), the B-cell compartment is likely to be atheroprotective. Indeed, B-lymphocyte deficiency increases atherosclerosis in the atherosclerosis-prone LDL receptor knockout (KO) mouse model (14), and we have shown that removal of the spleen, one of the main B-cell organs, leads to a dramatic aggravation of the disease in the apolipoprotein E (apoE) KO mouse model (15). Interestingly, it was shown at the same time that genetically engineered asplenic or splenectomized mice have, respectively, a total absence or a dramatic reduction of the B-1a cell population, which is associated with the inability to mount an immune response to streptococcal infections (16). A recent work by Binder et al. has demonstrated that pneumococcal vaccination decreases atherosclerotic lesion formation via a molecular mimicry between S. pneumoniaeand oxLDL (17). The same group had previously shown that C-reactive protein binds to oxLDL and apoptotic cells through recognition of phosphorylcholine (18). The genes encoding the antigen binding site of anti-oxLDL antibodies in apoE KO mice are structurally indistinguishable from antibodies produced by the T15 B-1 cell clones, which are specific for PC (19) and confer optimal protection against lethal infection with S. pneumoniae(20).
Thus, PC is the common target of the atheroprotective antibodies directed to oxLDL or S. pneumoniae. We therefore hypothesized that anti-PC humoral immunity might directly impact atherogenesis.
The apoE KO mice from our breeding facility were maintained on a regular chow diet and kept in standard conditions. All experiments conformed to “Position of the American Heart Association on Research Animal Use,” adopted on November 11, 1984.
The immunization protocol started in 8-week-old female apoE KO mice (n = 8/group) and was scheduled every second week. As an adjuvant, we chose the oligonucleotide CpG, because we have recently shown that among several adjuvants tested (Alum, CpG, complete and incomplete Freund’s adjuvants), CpG is the only one that does not possess intrinsic atheromodulating properties (21). Mice received IP injections (100 μl/injection) of either phosphate-buffered saline (PBS) or the 1826 CpG oligonucleotide (synthesized with a nuclease-resistant phosphorothioate backbone; 500 μg/mouse/injection; Operon Quiagen, Alameda, California) coupled to keyhole limpet hemocyanin (KLH; Biosearch Technologies, Novato, California) or the 1826 CpG oligonucleotide coupled to PC-KLH (22). Mice were euthanized after 4 months of treatment.
Blood was collected either by a retro-orbital puncture to follow the antibody titers during the experiment or by cardiac puncture at death. At death, peritoneal cells were harvested by a lavage with 2 ml RPMI 1640. Peritoneal cells were centrifuged at 300 gfor 5 min at 4°C. Erythrocytes were lysed with ammonium chloride/potassium carbonate (ACK) buffer, and the remaining leukocytes were washed before staining for flow cytometry. The aorta and the spleen were dissected. Blood was centrifuged at 4°C for 5 min at 12,000 g, and serum was stored at −20°C until analysis. Total serum cholesterol (TC) level was measured using a commercial kit (méthode “CHOD-PAP;” Boehringer, Mannheim, Germany). The aortic root was dissected under a microscope and embedded in optimum cut temperature medium, snap frozen, and sectioned.
Quantification of atherosclerotic lesions
The aortic root lesions were quantified as previously described (23). Briefly, 10-μm-thick serial cryostat sections were cut from the proximal 1 mm of the aortic root. The mean lesion size in each animal was determined by computer-assisted morphometry (Leica Qwin customized program) on 4 hematoxylin/oil red-O–stained sections cut at 200, 400, 600, and 800 μm from the cusp origin.
Immunohistochemistry and lesion morphology analysis were performed on serial aortic cryosections at 50-μm intervals between 500 μm and 700 μm from the cusp origin. Slides were air dried, fixed in cold acetone, and washed in water and PBS. Biotinylated primary antibodies were revealed with avidin alkalin-phosphatase (Vectastain ABC kit; Vector Laboratories, Burlingame, California) and Fast Red substrate (Dako, Glostrup, Denmark). Anti-T15 antibodies were kindly provided by John Kearney, University of Alabama, Birmingham. The other primary antibodies were purchased from BD Biosciences (San Jose, California) and directed against IAb(MHC II), CD3 (T cells), and CD19 (B cells). Primary antibodies were revealed by alkaline phosphatase-conjugated secondary antibodies and Fast Red substrate. Slides were counterstained by hematoxylin. Positive (red) staining was quantified by computer-assisted image analysis. The cellular/extracellular components of plaques was analyzed on Masson’s trichrome-stained cryosections.
Flow cytometry analysis
Spleen cell suspensions were prepared by meshing the tissue on 100-μm nylon filters and by lysing erythrocytes with the ACK buffer. Nonsplenic (control) leukocytes were obtained by peritoneal lavage. Splenocytes and peritoneal leukocytes were stained for 30 min at 4°C with 2 combinations of monoclonal antibodies to identify B-cell populations: IgD-FITC/CD5-CyChrome/IgM-biotinylated+Avidin-PE and IgM-FITC/CD5-PE/CD45RB220-CyChrome. All antibodies were from BD Biosciences. Cells were analyzed with a FACSCalibur (BD Biosciences) flow cytometer.
The LDL (1.019 < d < 1.063) was isolated by ultracentrifugation from pooled untouched apoE KO mouse plasma collected on EDTA. Protein content was determined by the Bradford method with bovine serum albumin (BSA) as standard. Copper-oxidized LDL was prepared by incubation of LDL (1 mg/ml) with 5 μmol/l Cu2+for 24 h at 37°C. Before incubation with cells, oxLDL was dialyzed against PBS 3 times.
Total IgG and IgM and specific antibodies to oxLDL, PC, and KLH were quantified using customized enzyme-linked immunosorbent assay (ELISA). For determination of total Ig, F(ab′)2fragments of antimouse IgG (Pierce, Rockford, Illinois) or antimouse IgM (BD Biosciences) were used as capture antibodies. For determination of specific antibodies, oxLDL (10 μg/ml), PC-BSA (2 μg/ml), and KLH (10 μg/ml) were coated onto the ELISA plate wells. For competitive ELISA, PC-BSA (Biosearch Technologies) was plated onto the 96-well enzyme immunoassay plate and oxLDL was used as competitive antigen. Increasing concentrations of oxLDL were coincubated with the sera for 1 h at room temperature. Preliminary experiments indicated that 2 μg/ml was the optimal concentration of coated PC-BSA. Increasing dilutions of individual sera were tested for anti-PC IgG, and the dilution corresponding to 50% of the maximal IgG binding was used for the competitive ELISA (1:20). Total and antigen-specific IgG were revealed using antimouse IgG (Calbiochem, Darmstadt, Germany) and antimouse IgM (BD Biosciences) alkaline phosphatase-conjugated secondary antibodies and p-nitrophenylphosphate disodic salt substrate. Plates were read at 405 nm. Data were read in duplicates; intra- and interassay variability were <5% and <20%, respectively.
Isolation of peritoneal macrophages
Six- to 8-week-old C57Bl/6 female mice (Charles River) received an IP injection of 2 ml thioglycollate medium (3%) 4 days before peritoneal macrophage harvesting by lavage. Cells were resuspended in complete DMEM supplemented with 10% fetal calf serum (FCS) and 20% of L-929 conditioned medium (source of CSF-1) and seeded at 0.8 × 106/well in 8-well Lab-Tek chamber slides. After 2 h of incubation at 37°C in a 5% CO2incubator, nonadherent cells were removed by gentle washing and adherent cells were incubated in complete DMEM 10% FCS + 20% L929-conditioned medium for 24 h. Cells were further cultured for 12 h before experiments in the same conditions but without FCS. With these conditions, >80% cells were identified as MAC-3+macrophages by flow cytometry and immunofluorescent microscopy (data not shown). Macrophages were then washed and incubated with or without freshly prepared oxLDL (50 μg/ml) and with or without 50 μg/ml of serum from PBS or CpG-PC–treated mice. After 15 h of incubation, macrophages were washed, fixed with 4% paraformaldehyde, stained with oil red O, and counterstained with hematoxylin. Lab-Tek slides were mounted under coverslip, and internalized LDL was analyzed on 3 random fields per condition captured with a CCD camera connected to a phase-contrast light microscope and a customized program written with Qwin (Leica, Wetzlar, Germany).
Results are expressed as mean ± SEM. One-way (group effect) or 2-way (group and distance effects) analysis of variance was performed using Statview 5.0 software (SAS Institute, Cary, North Carolina). Bonferroni post hoc test was used to evaluate the differences between each group. Differences between groups were considered to be significant if p < 0.05.
PC immunization reduces atherogenesis in apoE KO mice
Seven-week-old apoE KO mice were immunized or not with PC coupled to the KLH hapten carrier (22). The TC levels and body weights were equivalent in all groups. As shown in Figure 1,the control groups of mice that received the KLH alone or PBS developed lesions of similar size. Mice immunized with PC-KLH displayed lesions >40% smaller in size than control mice. Proportionally to the decrease in lesion size, lesions in PC-KLH immunized mice contained less extracellular matrix and fewer cells than control mice (Table 1).
PC immunization raises both anti-PC and anti-oxLDL antibody titers
Anti-PC immunization induces both IgM and IgG production (20,24,25). Total circulating IgG or IgM were not significantly affected by the immunization protocol (Fig. 2).As expected, anti-KLH antibodies were increased in both PC-KLH– and KLH-treated mice (data not shown). However, only PC-KLH–immunized mice showed a significant increase in the serum titers of anti-PC IgG and IgM (Fig. 2), substantiating the efficiency of the immunization protocol. Importantly, serum IgG and IgM from PC-KLH–immunized mice displayed also a much stronger reactivity against oxLDL compared with sera from control apoE KO mice (Fig. 3),and competitive ELISA showed that increasing doses of oxLDL compete with anti-PC IgG in apoE KO mouse sera (Fig. 3), in agreement with earlier findings that anti-oxLDL antibodies in apoE KO mice are indeed targeted to phosphorylcholine (26).
PC-immune apoE KO mouse serum reduces oxLDL uptake by cultured macrophages
The incubation of serum-deprived macrophages with oxLDL increased by 4.7 times the intracellular oil red O staining (Fig. 4).Addition of serum from nonimmunized apoE KO mice significantly increased the oxLDL uptake (Fig. 4), possibly because of the excess of lipoproteins contained in the hypercholesterolemic mouse serum, which is spontaneously oxidized in the culture conditions (27). In contrast, oxLDL uptake was similar to the one observed in the absence of mouse serum in PC-immunized mice in the presence of similar hypercholesterolemia but with 3-fold higher titers of anti-oxLDL and anti-PC antibodies compared with nonimmunized mice (Fig. 3).
B-cell activity and homing to plaques in PC-immunized mice
The evaluation of B-cell populations by flow cytometry (IgMhighIgD−Transitional T1 B cells, IgMhighIgDdullmarginal zone T1 B cells, IgMhighIgDhighT2 B cells) revealed that they were not affected in the spleen or the peritoneum (data not shown). On the contrary, mature B cells analyzed either as IgMdullIgDhighor as IgMhighB220+cells were significantly increased in the spleen of PC-KLH immunized mice but not in the peritoneum (mean of IgMhighB220+mature B cells in spleen ± SEM: KLH 9.3 ± 2.9%; PC-KLH 16.1 ± 2.6% [p < 0.05 vs. KLH and PBS]; PBS 10.6 ± 0.7%).
Immunohistochemistry of atherosclerotic lesions showed the presence of T15 antibodies, which were enriched in the lesions of PC-immunized mice. Furthermore, B-cell clusters were observed in 7 of 9 mice immunized with PC-KLH (Figs. 5 and 6).⇓⇓In contrast, fewer I-Ab+cells (activated leukocytes) were detected in the lesions of PC-immunized mice (Figs. 5 and 6). Although enriched in the adventitia, CD3+cells (T lymphocytes) were only occasionally observed within the atherosclerotic plaques (Fig. 5) regardless of the treatment.
A growing body of evidence suggests that the spontaneous humoral immune response developed in parallel with atherosclerosis is tentatively protective. In particular, it appears that antibodies recognizing PC, such as the germline-encoded B-1 cell natural antibodies, T15 (19), the antipneumococcus antibodies (17), and anti-oxLDL antibodies (2,3), reduce atherosclerosis. Thus, natural or induced antibodies reacting with PC might dynamically orchestrate protective host responses in infection, autoimmunity, and atherosclerosis (28).
The present study provides direct support to this hypothesis by showing that active PC immunization succeeds in reducing lesion size in the atherosclerosis-prone apoE KO mice. The intensity of protection conferred by active immunization to PC is in the range of the atheroprotection conferred by oxLDL (3) or pneumococcal vaccination (17). Of note, in parallel with anti-PC specific antibodies, PC immunization was able to raise the titers of serum anti-oxLDL antibodies to 3-fold higher than those of control-immunized apoE KO mice, further supporting the hypothesis that anti-oxLDL antibodies are indeed elicited by and directed to the PC on oxLDL (6). The rise in anti-oxLDL antibodies did not alter the TC levels, in agreement with the recent finding that such antibodies fail to alter the clearance of oxLDL in apoE KO mice (29).
Active PC immunization resulted in activation of the B-cell compartment in parallel with enrichment of B cells and T15-like antibodies within the atherosclerotic plaques where their protective function is likely optimized while being restricted topologically. Among the various mechanisms by which T15 antibodies can exert atheroprotection, the blockade of oxLDL uptake by macrophages, and thus of foam cell formation, has been proposed (19). Indeed, anti-PC IgM could mask the very epitopes on oxLDL that are recognized by macrophage scavenger receptors (19). We found that oxLDL uptake by serum-deprived cultured macrophages significantly increased in the presence of serum from nonimmunized apoE KO mice possibly owing to its abundance in lipoproteins spontaneously oxidized in the culture conditions (27). In spite of similar cholesterol content, we did not observe any increase of oxLDL uptake in the presence of serum from PC-immunized mice, suggesting that the increase of anti-PC antibodies in the serum of the latter blocked, at least partially, the oxLDL uptake. This is likely an important mechanism that explains the atheroprotection conferred by PC immunization.
The present findings show that directly targeted immunization to the PC epitope reproduces the beneficial effect observed in pneumococcal and oxLDL antiatherosclerotic vaccination protocols. Restricting the immune response to the minimal epitope while achieving equivalent therapeutic effect represents a major advance in the search for a straightforward, safe, and comprehensive vaccination against atherosclerosis.
Supported by Université Pierre et Marie Curie-Paris 6, INSERM, and the Fondation de France. The first two authors contributed equally to this work.
- Abbreviations and Acronyms
- apolipoprotein E
- keyhole limpet hemocyanin
- oxidized low-density lipoprotein
- phosphate-buffered saline
- total cholesterol
- Received October 11, 2005.
- Revision received October 6, 2006.
- Accepted November 16, 2006.
- American College of Cardiology Foundation
- Steinberg D.,
- Lewis A.
- Palinski W.,
- Miller E.,
- Witztum J.L.
- Zhou X.,
- Caligiuri G.,
- Hamsten A.,
- Lefvert A.K.,
- Hansson G.K.
- Zhou X.,
- Robertson A.K.,
- Rudling M.,
- Parini P.,
- Hansson G.K.
- Fredrikson G.N.,
- Soderberg I.,
- Lindholm M.,
- et al.
- Laurat E.,
- Poirier B.,
- Tupin E.,
- et al.
- Lee T.S.,
- Yen H.C.,
- Pan C.C.,
- Chau L.Y.
- Tupin E.,
- Nicoletti A.,
- Elhage R.,
- et al.
- Zhou X.,
- Nicoletti A.,
- Elhage R.,
- Hansson G.K.
- Major A.S.,
- Fazio S.,
- Linton M.F.
- Wardemann H.,
- Boehm T.,
- Dear N.,
- Carsetti R.
- Chang M.K.,
- Binder C.J.,
- Torzewski M.,
- Witztum J.L.
- Briles D.E.,
- Forman C.,
- Hudak S.,
- Claflin J.L.
- Khallou-Laschet J.,
- Tupin E.,
- Caligiuri G.,
- et al.
- Briles D.E.,
- Nahm M.,
- Schroer K.,
- et al.
- Friedman P.,
- Horkko S.,
- Steinberg D.,
- Witztum J.L.,
- Dennis E.A.
- Reardon C.A.,
- Miller E.R.,
- Blachowicz L.,
- et al.