Role of C-Reactive Protein in Atherogenesis

From the Central Laboratory Animal Facility (K.R., E.W.) and the Institutes of Pathology (H.-A.L., D.B., S.C.S.), Clinical Chemistry and Laboratory Medicine (M.T., K.J.L.), Medical Biometry, Epidemiology, and Informatics (M.B.), and Medical Microbiology and Hygiene (M.H., S. Bhakdi), Johannes Gutenberg University Mainz, Germany; and the Department of Biochemistry (S. Black, D.S.), Case Western Reserve University, Cleveland, Ohio.

Correspondence to Sucharit Bhakdi, Institute of Medical Microbiology and Hygiene Johannes Gutenberg University Mainz Obere Zahlbacherstr 67 55131 Mainz, Germany. E-mail sbhakdi@uni-mainz.de

    Abstract

Objective— Human C-reactive protein (CRP) was reported to accelerate atherosclerotic lesion development in male but not in female apolipoprotein E (apoE) knockout mice. Here, mice expressing rabbit CRP (rbCRP) were crossbred onto apoE knockout animals, and the effect on atherogenesis was studied.

Methods and Results— Hemolytic complement activity could not be detected in apoE knockout mice. Furthermore, in contrast to human complement, neither rabbit nor human CRP complexed to modified low-density lipoprotein–activated murine complement. At 52 weeks, rbCRP levels were similar in male and female transgenic animals. Serum cholesterol levels were equivalent in female animals irrespective of rbCRP expression, whereas rbCRP–positive males had significantly higher serum cholesterol levels than the rbCRP-negative counterparts. All mice exhibited extensive atherosclerotic lesions, as studied en face, and no differences were noted between rbCRP-negative and rbCRP-positive animals. Atherosclerotic luminal obstruction of aortic arch and first-order neck branches did not differ significantly between rbCRP-positive and rbCRP-negative mice. There was no correlation between rbCRP levels and atherosclerotic lesion formation.

Conclusions— No marked effect of rbCRP on the formation of moderately advanced atherosclerotic lesions could be discerned in the apoE knockout mouse. Because of the oddities of the mouse complement system, however, this may not be a good model to investigate the role of CRP in human atherosclerosis.

Transgenic mice expressing rabbit CRP were crossbred onto apolipoprotein E knockout animals, and the animals were monitored for the development of atherosclerotic lesions. We failed to detect a significant influence of CRP transgenesis on atherogenesis in apoE knockout mice.

Key Words: C-reactive protein ? atherosclerosis ? complement ? transgene murine model

    Introduction

A correlation exists between C-reactive protein (CRP) serum concentrations and the risk of cardiac events in patients with coronary heart disease,1–3 but it is unknown whether the acute-phase protein simply represents a marker of inflammation4 or whether it is causally involved in atherogenesis. CRP is present in early human atheromatous plaques and colocalizes with activated complement.5 The capacity of CRP to bind to enzymatically modified low-density lipoprotein (E-LDL),6 the major target of CRP in the atherosclerotic lesion, and to activate complement7 could imply that it drives inflammation and is, hence, proatherogenic.8,9 Reports that CRP downregulates endothelial biosynthesis of nitric oxide synthase10 and stimulates the expression of endothelin 111 might appear to accord with such a contention. However, reports on proinflammatory cellular effects of CRP require confirmation because uncharacterized commercial CRP preparations were used in past studies. One effect has recently been shown to represent a sodium azide artifact.12

See page 1527

Alternatively, CRP may play a physiological role. Complement activation by CRP/E-LDL halts before the proatherogenic terminal sequence, and incomplete complement activation may be a physiological event enabling LDL to be removed from tissues,13 blunting a detrimental response to LDL retention.13–15

Murine apolipoprotein (apoE)16 and LDL receptor17 knockout strains are widely used to study inflammatory mechanisms in atherosclerosis. However, mouse CRP is a trace protein with a concentration that does not exceed 2 mg/L even after an inflammatory stimulus.18 Hence, research on CRP function in atherosclerosis currently relies on the use of transgenic mice expressing rabbit CRP (rbCRP)19 or human CRP (huCRP).20 In a recent study, introduction of a huCRP transgene into the apoE knockout mouse strain was reported to promote deposition of foam cells in the aortic valve cups in 7-month-old animals.8 The effect was restricted to male animals that expressed far higher CRP levels than did the female mice, because expression of transgenic huCRP in the construct is under strict testosterone control.21 However, data published several months prior using the same mouse strain did not appear to accord with these findings.22

Here, we report our observations on a murine strain transgenic for rbCRP and the CRP-negative apoE knockout control. Expression of the rbCRP transgene is independent of gender, and an additional inflammatory stimulus is also not required. rbCRP and huCRP are similar in structure and function. Both bind phosphocholine, C-polysaccharide, polycations, chromatin, and histones, activate complement, and protect mice from lethal challenges with pneumococci.23–29 On the basis of amino acid similarity, rbCRP30 is 83% and 86% homologous to the murine and human counterparts, respectively. huCRP is only 79% homologous to mouse rbCRP. The methodical drawbacks of transgenic expression of huCRP and rbCRP in the mouse are equivalent. The Genbank identification numbers of the murine and human cDNA used for our inquiry are gi|31982450|ref|NM_007768.2| and gi|17530542|gb|AF449713.1|AF449713, respectively.

Unexpectedly, the interactions between CRP and modified LDL and complement, as have been delineated in the human system,7,13 did not appear to be operative in the apoE knockout mouse. No marked effect of CRP expression on the development of moderately advanced atherosclerotic lesions was detected, and there was no correlation between CRP levels and the extent of atherosclerosis.

    Methods

Mouse Breeds

Mice were maintained at the University of Mainz under specified pathogen free conditions. Lineage PC-12 (internal designation rbCRP) has previously been described.19,28 Briefly, the rbCRP strain carries the coding region of the rbCRP gene cloned downstream of the rat phosphoenolpyruvate carboxykinase (PEPCK) promoter with liver- and kidney-specific activity.31 We used hemizygous rbCRP transgenic mice backcrossed to the C57BL/6 (B6) inbred strain for 5 generations. For rbCRP genotyping, we used primers 5'-TGTGAACATGTGGGACTAT-3' and 5'-TGAAGAAGGCAAAGCTAGG-3'. C57BL/6J-apoEtm1Unc knockout mice (internal designation apoE knockout)16 backcrossed to B6 for 10 generations were obtained from The Jackson Laboratory. The rbCRPxapoE knockout double mutants were generated by outcrossing rbCRP with apoE knockout mice and by backcrossing rbCRP-positive hybrids to the apoE parental strain. The apoE knockout and wild-type alleles could be discriminated as recommended by The Jackson Laboratory.

Diet for Activation of the PEPCK Promoter

The PEPCK promoter of the rbCRP transgene was activated by a protein-rich diet consisting of 60% caseine, 20% cellulose, 11% vegetable oil, 2% brewer’s yeast, and 7% mineral mix with vitamins.31 The mice were fed ad libitum.

Preparation of Aorta and Neck Vessels

Mice were euthanized at the age of 20 and 52 weeks and fixed in 4% buffered formaldehyde. Hearts and aortas were resected en bloc down to the iliac bifurcation. The aortic arch was dissected and 1-mm long rings of the truncus brachiocephalicus, the aortic arch and the left subclavian artery were oriented in paraffin blocks, serially sectioned (approximately two hundred 4-μm sections per sample) and stained with hematoxylin/eosin. The rest of the aortae were stained with Sudan IV as previously described,32 and photographed with a 12 megapixel digital camera (DXM 1200, Nikon, Tokyo, Japan) at a total magnification of 40-fold, thus yielding high resolution images (30 MB; see Figure 4).

   Figure 4. Representative atherosclerotic lesions seen in rbCRP-negative and rbCRP-positive apoE knockout mice 52 weeks of age. A and B, Representative histological sections through the left subclavian artery (A) and the aortic arch (B). C, Sudan IV–stained aorta en face. The histomorphological characteristics of the atherosclerotic lesions were identical in both groups of animals, consisting of foam cells and lipid cores with cholesterol crystals (arrows) and focal mineralizations, separated from the blood stream by fine fibrous caps. None of the lesions exhibited lipid lakes. Likewise, no differences were seen in terms of inflammatory and histiocytic cell infiltrations between lesions of rbCRP-positive and rbCRP-negative mice. Finally, and in agreement with previous observations by our own group and by others,41 we did see a similar frequency and toporegional distribution of chondroid-like changes in the atherosclerotic lesions, the exact nature of which remains unknown. Taken together, virtually all lesions on rbCRP-negative and rbCPR-positive mice corresponded to Stary classification type V. Asterisks demonstrate characteristic chondroid-like lesions found preferably in advanced mouse atherosclerotic lesions.41 The space bars depict distances of 500 μm for the histological sections (A and B) and 0.5 cm for the en face image (C).

Image Analysis

Atherosclerotic lesions en face were quantified on digitized photos using Photoshop-based image analysis (Adobe Systems Inc, San Jose, Calif).32,33 Images of histological sections were taken with a diagnostic microscope and a high-resolution camera (DXM 1200, Nikon) and imported into Photoshop and obstructive plaques quantified as a fraction of the vessel lumen. For that purpose, atherosclerotic lesions were manually marked and the number of pixels quantified. To correct for the noncylindrical geometry of the partially collapsed vessel rings, vessel lumen was calculated from the length of the inner vessel circumference as measured in Photoshop. Plaque area was expressed as percent luminal obstruction.

rbCRP Binding Assay

Serum rbCRP levels were determined by ELISA as described.29,34

Determination of Serum Cholesterol and Fatty Acids

Murine sera diluted 1:3 with physiological saline were subjected to commercial cholesterol (CHOD-PAP, Roche Diagnostics) and triglyceride (GPO-PAP, Roche Diagnostics) assays.

Hemolytic Complement Assay

Hemolysis tests to assess murine complement activity have been notoriously difficult to establish; therefore, we developed a novel assay. Staphylococcal -toxin (SAT) is a protein bacteria toxin that, at concentrations >200 nmol/L, binds nonspecifically, in a nonsaturable manner, and irreversibly to rabbit erythrocytes. A single amino-acid substitution (arginine for histidine at residue 35; H35R) renders the toxin nonlytic, enabling cells to be coated with ample amounts of the mutant toxin without ensuing hemolysis.35 A SAT-specific polyclonal antibody can subsequently be applied and the hypersensitized erythrocytes efficiently lysed by complement of any origin, including the mouse. For this a 5% erythrocyte suspension in veronal-buffered saline (VBS) was incubated with 30 μg/mL H35R SAT for 30 minutes at 24°C. Cells were pelleted, resuspended in VBS to 3%, and 10 μL of anti-SAT serum were added per milliliter. Twenty-five microliters of hypersensitized cells were incubated with 25 μL of mouse serum, and hemolysis was determined after 1 hour (37°C) by reading the absorption of the supernatants at 412 nm.

Complement Consumption Assay

Enzymatic modification of human LDL was performed as described.13,36 The complement consumption assay was carried out as described.13 Briefly, a human serum pool containing <1 μg/mL CRP was diluted 10-fold with VBS, and aliquots of 50 μL were transferred into reaction tubes. Recombinant huCRP (Calbiochem) and human E-LDL were added to the given final concentrations, and the samples were incubated for 60 minutes at 37°C. Subsequently antibody-coated sheep erythrocytes were added, and the degree of hemolysis was determined after 60 minutes by measuring the absorbance of the supernatants at 412 nm. Hemolysis curves were inverted to reflect complement consumption.

For adaptation to the murine system, B6 sera were used instead of human serum samples and, in some of the experiments, rbCRP purified as described29 was used instead of huCRP. These assays were developed using the SAT-hypersensitized rabbit erythrocytes.

Statistical Analysis

Body weights and serum parameters of rbCRPxapoE knockout animals were compared by t tests (Table I, available online at http://atvb.ahajournals.org). Correlation coefficients were calculated between serum rbCRP concentration and atherosclerosis data. When histological data were compared between rbCRP subgroups, parametric t tests and, to compensate for a potential lack of normal distribution, nonparametric Wilcoxon tests were applied. Further, serum cholesterol values were taken into account in a regression model (ANOVA) where the serum cholesterol concentration was used as quantitative variable.

    Results

Lack of Detectable Hemolytic Complement Activity in ApoE Knockout Mice

The apoE knockout strain used in our study had been backcrossed to the B6 strain for 10 generations; therefore, testing the serum hemolytic activity of 2 normocholesterolemic B6 control mice and 2 apoE knockout mice was performed originally just on a precautionary basis. Sera of the B6 control mice had normal hemolytic titers (Figure 1A). However, sera of both apoE knockout mice completely lacked detectable hemolytic complement activity (Figure 1B). We subsequently tested sera from each animal in this study as well as sera from 5 nontransgenic apoE knockout animals and found a hemolytic titer of zero in each case.

   Figure 1. Lack of hemolytic activity of apoE knockout serum. Hypersensitized rabbit erythrocytes were incubated with various concentrations of C57BL/6 (A) and apoE–/– knockout (B) sera at the given concentrations. The complement system is intact in the C57BL/6 strain,47 and total hemolysis was observed with undiluted serum (A). In contrast, no hemolytic activity could be observed with the apoE knockout serum (B). Data shown represent the means of 2 independent determinations. Similar absence of hemolytic activity was confirmed in all other mice entered in this study.

Human and Rabbit CRPs Fail to Augment E-LDL–Dependent Activation of Murine Complement

Exploratory mixing experiments indicated that lack of hemolytic complement activity in apoE knockout mice was attributable to the presence of an inhibitor. Because the inhibitor might not necessarily be present in tissues, we tested whether CRP would augment E-LDL–dependent complement activation in normal mouse serum. E-LDL activates human complement via a highly efficient CRP-dependent (Figure 2A, open circles) and a less-efficient CRP-independent mechanism (Figure 2A, closed circles). The same experiment was conducted with mouse B6 serum. High concentrations of E-LDL alone activated murine complement (Figure 2, closed circles). However, neither huCRP (Figure 2B, open circles) nor rbCRP (not shown) had any augmenting influence on murine complement consumption. The experiment was reproduced with 3 different batches of recombinant huCRP.

   Figure 2. Lack of huCRP to augment activation of mouse complement. Dose-dependent human (A) and murine (B) complement consumption induced by E-LDL in the presence or absence of 10 μg/mL huCRP. In human serum (A) huCRP-dependent complement consumption could be detected at E-LDL concentrations of >6 μg/mL (), whereas E-LDL alone was only capable of activating complement at concentrations >50 μg/mL (). In contrast, huCRP did not augment activation of mouse complement by E-LDL (B).

Induction of rbCRP Transgene Expression by a Protein-Rich Diet

Starting at the age of 8 weeks, rbCRPxapoE knockout offspring mice were fed a protein-rich diet to drive rbCRP transgene expression.31 Body weights of rbCRPxapoE knockout mice exhibited significant gender differences (Table I) and hence rbCRP serum concentrations and all further data were not only assorted according to genotype but also to gender. rbCRP was absent in the rbCRP negative mice but expressed at high serum levels independent of gender in the transgenic animals (Table I). Two rbCRP-positive apoE–/– mice did not express the transgene and were excluded from the study. Overall, expression of rbCRP observed here, including the occasional lack of expression, was in accord with the literature.19,28

Development of Hypercholesterolemia in ApoE Knockout Mice

At 52 weeks, serum triglyceride levels of apoE knockout mice barely exceeded physiological murine values. In contrast, serum cholesterol concentrations were increased over those in sex-matched apoE+/– controls of the same rbCRP genotype to a similar degree as that of apoE knockout mice on a balanced diet.37,38 Whereas cholesterol levels in all female animals were similar (Table I), cholesterol and triglyceride concentrations of rbCRP-positive males were significantly (t test) higher than those of rbCRP-negative males (Table I). A similar trend was noted after introduction of a huCRP transgene into apoE knockout mice,8 but the differences did not attain statistical significance, presumably because of the shorter duration of the experiment (6 months versus 12 months in this study).

rbCRP Transgenesis Does Not Augment Atherogenesis in ApoE Knockout Mice at 20 Weeks

Five rbCRP-negative and 5 rbCRP-positive male and 7 rbCRP-negative and 4 rbCRP-positive apoE knockout female mice were euthanized at 20 weeks. In each group, only 1 to 2 animals presented with minimal atherosclerotic changes (<3% of aortic surface and <5% luminal obstruction), whereas the remaining animals showed no atherosclerotic lesions at all. No statistically significant differences were noted between any groups.

rbCRP Transgenesis Does Not Augment Plaque Development in the Aorta of 52-Week-Old ApoE Knockout Mice As Studied En Face

No atherosclerosis was detected in apoE+/– mice. In contrast, all apoE knockout mice exhibited extensive aortic lesions at 52 weeks, and the mean lesion size was significantly more extensive in males compared with females (Figure 3A). However, neither by using t test nor Wilcoxon test nor after adjusting for serum cholesterol in the regression analysis (ANOVA) were any differences in plaque area found between rbCRP-negative and rbCRP-positive mice of either sex (Figure 3A).

   Figure 3. Quantitative analyses of atherosclerotic lesions in rbCRP-negative and rbCRP-positive apoE knockout mice 52 weeks of age. A, Aorta en face. The aortic plaque area of apoE knockout mice was analyzed by computer-based image analysis of aortic preparations en face. Irrespective of gender, no statistically significant (ANOVA) differences in the size of aortic lesions were found between the rbCRP-positive and the rbCRP-negative apoE knockout mice. B, Aortic arch obstruction. Obstruction was calculated in the murine aortic arch using morphometric analysis of serial histological sections. The degree of obstruction was increased in the rbCRP-positive males and females as compared with the rbCRP-negative controls; however, the values did not attain statistical significance (ANOVA). C, Neck vessel obstruction. In the first-order branches of the aortic arch (right, brachiocephalic arch; left, subclavian artery), the rbCRP-positive males and females exhibited a higher degree of luminal obstruction as compared with the rbCRP-negative controls of the same sex. However, the differences observed were not statistically significant (ANOVA). Data are means±SD. n.s. indicates not significant.

Quantification of Luminal Vessel Obstruction

The results of these analyses are shown in Figure 3B and 3C. No histomorphological differences were detected between atherosclerotic lesions of rbCRP-negative and rbCRP-positive mice (Figure 4). By using parametric t test, the differences of aortic and neck vessel obstruction between rbCRP-positive and rbCRP-negative mice were found to be statistically significant for the males but not for the females. These results were confirmed by data analysis with the nonparametric Wilcoxon test. However, when adjusting for the influence of serum cholesterol in the regression analysis (ANOVA), the results between the rbCRP-positive and rbCRP-negative mice were no longer significant, indicating that cholesterol was an important confounding factor in our study. The trend toward increased atherosclerosis in rbCRP-positive animals was more pronounced in males, where the probability values (ANOVA) reached 0.07 and 0.10 for aortic and neck vessel obstruction, respectively. When a regression analysis of CRP levels and vessel obstruction was performed, no correlation was observed in male animals (Figure I, available online at http://atvb.ahajournals.org) or female animals; in the latter, correlation coefficients between CRP levels and obstruction of aortic arch and neck vessel were –0.024 and –0.002, respectively.

    Discussion

Two studies have previously addressed the possible significance of CRP in atherogenesis by expressing huCRP in apoE knockout mice.8,22 Whereas Hirschfield et al22 detected no proatherogenic effect, Paul et al8 reported an increase in lesion size in the aortic arch of male mice and concluded that CRP is atherogenic. In that study,8 only transgenic huCRP male mice had increased atherosclerotic lesions, and the effect was observed in naive animals as well as in mice subjected to a strong inflammatory stimulus (turpentine injections). For unknown reasons, the CRP levels were extraordinarily high as compared with other studies using the same transgenic strain,21,39,40 suggesting a concomitant inflammatory stimulus of unknown origin. In this context, inflammatory stimuli themselves accelerate atherogenesis,32 and conclusions regarding the specific role of CRP in turpentine-treated animals cannot be drawn if appropriate controls are lacking. Of note, although female transgenic mice expressed low CRP levels, these were nevertheless in the range that is considered a risk factor in humans; however, no differences between the CRP-positive and -negative animals were seen.8

In addition to these caveats, animals in the study by Paul et al8 were euthanized at an early age (29 weeks), where genuine atherosclerotic lesions had not yet been formed, and only microscopic histiocytic deposits in the aortic valve cups were taken as the basis for quantification of atherosclerosis. Marginal differences in lesion size were noted only in the aortic route, whereas lesions in the remainder of the aorta exhibited no differences. The finding that male CRP-transgenic mice had more C3 deposits in the lesions8 is also open to interpretation, because current immunohistochemical staining methods permit no differentiation between activated and nonactivated murine complement components. For the same reason, mere detection of CRP in the murine lesions does not demonstrate its complement-activating capacity. In contrast, positive stainings for C3dg and C5b-9-neoantigens constitute unequivocal evidence for complement activation in human lesions.7,13

Experiments with mice expressing rbCRP do not harbor the problem of testosterone and inflammation-dependency of the construct. Atherosclerotic lesion development is slow in the apoE knockout mouse, and only very small lesions begin to appear after 20 weeks. After 52 weeks, moderately advanced lesions developed that corresponded to stage V of the Stary classification. We discerned no proatherogenic effect of CRP at this stage. For unknown reasons, male CRP transgenic mice had significantly higher cholesterol levels than all other groups, which could explain the increase in lesion size that was detected when vascular obstruction was quantified in first-order branches as proposed by Rosenfeld et al.41 This observation would have been missed by conventional en face plaque measurement alone. The group size in our study would have permitted only detection of strong CRP effects, and small differences at this or at earlier stages of atherogenesis might have been missed. This potential shortcoming is common to most animal experiments. Moreover, current discussions on the possible atherogenic effects of CRP in humans relate to consequences in late-stage disease that might be mediated by far lower CRP levels than expressed here.

The oddities of mouse complement, particularly with regard to CRP-dependent activation by lipoproteins, surfaced when we tested whether murine complement would be activated by modified LDL and whether this process would be amplified by CRP as found previously with human complement.13 By using erythrocytes hypersensitized with a bacterial protein toxin, a hemolysis target was obtained that was readily lysed by mouse complement. To our surprise, the apoE knockout mice completely lacked hemolytic complement activity. Earlier work has uncovered an inhibitory effect of mouse fibronectin42 and the existence of an inhibitor of the terminal complement sequence,43 which partially explains the notorious inefficiency of murine complement.44 Pilot mixing experiments indicated that lipemic sera from apoE knockout mice contain either larger amounts of these inhibitors or additional inhibitory factors that await future identification.

Because the putative inhibitor might not diffuse into atherosclerotic lesions, we determined whether the pattern of complement activation by E-LDL±CRP as found in human serum13 could be reproduced in normal murine serum. At high concentrations, modified LDL was indeed capable of consuming murine complement, but complement activation was not augmented by CRP.

All of our present findings thus uncover the disturbing facts that the interactions among CRP, complement, and LDL, as have been delineated in humans, may not exist similarly in mice. It can naturally not be excluded that mouse CRP might be active, but the incapacity of transgenically expressed CRP to execute 1 of its primary functions places obvious constraints on the validity of this animal model. Hence, any effect of CRP in the apoE knockout mouse is probably independent of the complement system. Vice versa, it must be conceded that any lack of effect as observed here might be attributable to this shortcoming.

It has been reported that C3 deficiency protects LDL receptor knockout mice against atherosclerosis.45 This finding would be in accord with the earlier observation that C6 deficiency in rabbits is also protective46 and need not stand in conflict with the present work. Preliminary data from this laboratory indicate that hemolytic complement activity in LDL receptor knockout mice is not nil and, as shown above, higher concentrations of E-LDL do activate mouse complement independent of CRP. In summary, caution should be exercised when extrapolating observations in genetically engineered mouse models with incompletely characterized physiological alterations to the situation in human disease, and we believe that a definitive statement regarding the role of CRP in human atherogenesis should be deferred at this stage.

    Acknowledgments

This work was supported by the Deutsche Forschungsgemeinschaft (Bh 2/2-3) and by grant 8312-38 62 61/527 of the "Stiftung Rheinland-Pfalz für Innovation." The excellent technical assistance of Antonietta Valentino, Claudia Braun, Petra Nusser, and Sybille Ott is gratefully acknowledged.

References

Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. Multiple Risk Factor Intervention Trial. Am J Epidemiol. 1996; 144: 537–547.

Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Eng J Med. 1997; 336: 973–979.

Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med. 2004; 350: 1387–1397.

Hackam DG, Anand SS. Emerging risk factors for atherosclerotic vascular disease: a critical review of the evidence. JAMA. 2003; 290: 932–940.

Torzewski J, Torzewski M, Bowyer DE, Frohlich M, Koenig W, Waltenberger J, Fitzsimmons C, Hombach V. C-reactive protein frequently colocalizes with the terminal complement complex in the intima of early atherosclerotic lesions of human coronary arteries. Arterioscler Thromb Vasc Biol. 1998; 18: 1386–1392.

Bhakdi S, Dorweiler B, Kirchmann R, Torzewski J, Weise E, Tranum-Jensen J, Walev I, Wieland E. On the pathogenesis of atherosclerosis: enzymatic transformation of human low density lipoprotein to an atherogenic moiety. J Exp Med. 1995; 182: 1959–1971.

Bhakdi S, Torzewski M, Klouche M, Hemmes M. Complement and atherogenesis: binding of CRP to degraded, nonoxidized LDL enhances complement activation. Arterioscler Thromb Vasc Biol. 1999; 19: 2348–2354.

Paul A, Ko KW, Li L, Yechoor V, McCrory MA, Szalai AJ, Chan L. C-reactive protein accelerates the progression of atherosclerosis in apolipoprotein E-deficient mice. Circulation. 2004; 109: 647–655.

Pepys MB, Hirschfield GM. C-reactive protein and atherothrombosis. Ital Heart J. 2001; 2: 196–199.

Venugopal SK, Devaraj S, Yuhanna I, Shaul P, Jialal I. Demonstration that C-reactive protein decreases eNOS expression and bioactivity in human aortic endothelial cells. Circulation. 2002; 106: 1439–1441.

Verma S, Li SH, Badiwala MV, Weisel RD, Fedak PW, Li RK, Dhillon B, Mickle DA. Endothelin antagonism and interleukin-6 inhibition attenuate the proatherogenic effects of C-reactive protein. Circulation. 2002; 105: 1890–1896.

van den Berg CW, Taylor KE, Lang D. C-reactive protein-induced in vitro vasorelaxation is an artefact caused by the presence of sodium azide in commercial preparations. Arterioscler Thromb Vasc Biol. 2004; 24: e168–e171.

Bhakdi S, Torzewski M, Paprotka K, Schmitt S, Barsoom H, Suriyaphol P, Han SR, Lackner KJ, Husmann M. Possible protective role for C-reactive protein in atherogenesis: complement activation by modified lipoproteins halts before detrimental terminal sequence. Circulation. 2004; 109: 1870–1876.

Bhakdi S, Lackner KJ, Han SR, Torzewski M, Husmann M. Beyond cholesterol: the enigma of atherosclerosis revisited. Thromb Haemost. 2004; 91: 639–645.

Williams KJ, Tabas I. The response-to-retention hypothesis of early atherogenesis. Arterioscler Thromb Vasc Biol. 1995; 15: 551–561.

Piedrahita JA, Zhang SH, Hagaman JR, Oliver PM, Maeda N. Generation of mice carrying a mutant apolipoprotein E gene inactivated by gene targeting in embryonic stem cells. Proc Natl Acad Sci U S A. 1992; 89: 4471–4475.

Ishibashi S, Brown M, Goldstein J, Gerard RD, Hammer RE, Herz J. Hypercholesterolemia in LDL receptor knockout mice and its reversal by adenovirus-mediated gene delivery. J Clin Invest. 1993; 92: 883–893.

Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest. 2003; 111: 1805–1812.

Xia D, Samols D. Transgenic mice expressing rabbit C-reactive protein are resistant to endotoxemia. Proc Natl Acad Sci U S A. 1997; 94: 2575–2580.

Ciliberto G, Arcone R, Wagner EF, Ruther U. Inducible and tissue-specific expression of human C-reactive protein in transgenic mice. EMBO J. 1987; 6: 4017–4022.

Szalai AJ, van Ginkel FW, Dalrymple SA, Murray R, McGhee JR, Volanakis JE. Testosterone and IL-6 requirements for human C-reactive protein gene expression in transgenic mice. J Immunol. 1998; 160: 5294–5299.

Hirschfield GM, Gallimore R, Gilbertson JA, Benson M, Pepys MB. Systemic inflammation and atherogenesis in apoE knockout mice expressing transgenic human C-reactive protein. Circulation. 2003; 108 (suppl IV): IV–41Abstract.

Bach BA, Gewurz H, Osmand AP. C-reative protein in the rabbit: isolation, characterization and binding affinity to phosphocholine. Immunochemistry. 1977; 14: 215–219.

Oliveira EB, Gotschlich EC, Liu TY. Comparative studies on the binding properties of human and rabbit C- reactive proteins. J Immunol. 1980; 124: 1396–1402.

Robey FA, Jones KD, Tanaka T, Liu TY. Binding of C-reactive protein to chromatin and nucleosome core particles. A possible physiological role of C-reactive protein. J Biol Chem. 1984; 259: 7311–7316.

Horowitz J, Volanakis JE, Briles DE. Blood clearance of Streptococcus pneumoniae by C-reactive protein. J Immunol. 1987; 138: 2598–2603.

Dougherty TJ, Gewurz H, Siegel JN. Preferential binding and aggregation of rabbit C-reactive protein with arginine-rich proteins. Mol Immunol. 1991; 28: 1113–1120.

Lin CS, Xia D, Yun JS, Wagner T, Magnuson T, Mold C, Samols D. Expression of rabbit C-reactive protein in transgenic mice. Immunol Cell Biol. 1995; 73: 521–531.

Black S, Agrawal A, Samols D. The phosphocholine and the polycation-binding sites on rabbit C-reactive protein are structurally and functionally distinct. Mol Immunol. 2003; 39: 1045–1054.

Hu S, Miller S, Samols D. Cloning and characterization of the gene for rabbit C-reactive protein. Biochemistry. 1986; 25: 7834–7839.

McGrane MM, de Vente J, Yun J, Bloom J, Park E, Wynshaw-Boris A, Wagner T, Rottman FM, Hanson RW. Tissue-specific expression and dietary regulation of a chimeric phosphoenolpyruvate carboxykinase/bovine growth hormone gene in transgenic mice. J Biol Chem. 1988; 263: 11443–11451.

Lehr HA, Sagban TA, Ihling C, Zahringer U, Hungerer KD, Blumrich M, Reifenberg K, Bhakdi S. Immunopathogenesis of atherosclerosis: endotoxin accelerates atherosclerosis in rabbits on hypercholesterolemic diet. Circulation. 2001; 104: 914–920.

Lehr HA, van der Loos CM, Teeling P, Gown AM. Complete chromogen separation and analysis in double immunohistochemical stains using Photoshop-based image analysis. J Histochem Cytochem. 1999; 47: 119–126.

Agrawal A, Shrive AK, Greenhough TJ, Volanakis JE. Topology and structure of the C1q-binding site on C-reactive protein. J Immunol. 2001; 166: 3998–4004.

Jursch R, Hildebrand A, Hobom G, Tranum-Jensen J, Ward R, Kehoe M, Bhakdi S. Histidine residues near the N terminus of staphylococcal alpha-toxin as reporters of regions that are critical for oligomerization and pore formation. Infect Immun. 1994; 62: 2249–2256.

Klouche M, Gottschling S, Gerl V, Hell W, Husmann M, Dorweiler B, Messner M, Bhakdi S. Atherogenic properties of enzymatically degraded LDL: selective induction of MCP-1 and cytotoxic effects on human macrophages. Arterioscler Thromb Vasc Biol. 1998; 18: 1376–1385.

Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science. 1992; 258: 468–471.

Zhang SH, Reddick RL, Burkey B, Maeda N. Diet-induced atherosclerosis in mice heterozygous and homozygous for apolipoprotein E gene disruption. J Clin Invest. 1994; 94: 937–945.

Szalai AJ, Briles DE, Volanakis JE. Human C-reactive protein is protective against fatal Streptococcus pneumoniae infection in transgenic mice. J Immunol. 1995; 155: 2557–2563.

Hirschfield GM, Herbert J, Kahan MC, Pepys MB. Human C-reactive protein does not protect against acute lipopolysaccharide challenge in mice. J Immunol. 2003; 171: 6046–6051.

Rosenfeld ME, Polinsky P, Virmani R, Kauser K, Rubanyi G, Schwartz SM. Advanced atherosclerotic lesions in the innominate artery of the ApoE knockout mouse. Arterioscler Thromb Vasc Biol. 2000; 20: 2587–2592.

Hitsumoto Y, Okada M, Makino H. Inhibition of human and mouse complement-dependent hemolytic activity by mouse fibronectin. Immunopharmacology. 1999; 42: 203–208.

Sassi F, Hugo F, Muhly M, Khaled A, Bhakdi S. A reason for the cytolytic inefficiency of murine serum. Immunology. 1987; 62: 145–147.

Rosenberg LT, Tachibana DK. Activity of mouse complement. J Immunol. 1962; 89: 861–867.

Buono C, Come CE, Witztum JL, Maguire GF, Connelly PW, Carroll M, Lichtman AH. Influence of C3 deficiency on atherosclerosis. Circulation. 2002; 105: 3025–3031.

Schmiedt W, Kinscherf R, Deigner HP, Kamencic H, Nauen O, Kilo J, Oelert H, Metz J, Bhakdi S. Complement C6 deficiency protects against diet-induced atherosclerosis in rabbits. Arterioscler Thromb Vasc Biol. 1998; 18: 1790–1795.

D’Eustachio P, Kristensen T, Wetsel RA, Riblet R, Taylor BA, Tack BF. Chromosomal location of the genes encoding complement components C5 and factor H in the mouse. J Immunol. 1986; 137: 3990–3995.