C. pneumoniae is hypothesized to disseminate systemically from the lungs through infected peripheral blood mononuclear cells and to localize in arteries (17, 41), where it would infect other cell types involved in the atherogenic process and promote vascular inflammation (15).

At present, the role of C. pneumoniae in atherosclerotic vessel walls is still controversial. Few studies failed to detect C. pneumoniae in human atheromata, and those that failed mainly focused on C. pneumoniae DNA detection by PCR. In general, detection of C. pneumoniae DNA has always resulted in a lower prevalence of positive results compared to the prevalence in studies aimed at detecting C. pneumoniae antigens by immunohistochemistry assays (6, 10, 14, 23, 25). This is rather astonishing, since the detection of DNA by PCR is supposed to be more sensitive than that by immunohistochemical staining assays.

We investigated a broad range of human atherosclerotic arteries for the presence of various C. pneumoniae components, DNA and antigens, and assessed whether positive immunohistochemistry assay results for PCR-negative atheromata can be explained by nonspecific reactions with nonchlamydial plaque constituents.

The study was approved by the ethics committee of the Antwerp University Hospital. Written informed consent was obtained from each patient before surgery.

Specimens. One hundred ninety-seven arterial fragments were removed from 165 consecutive patients during surgical revascularization. The specimens were from 28 abdominal aortic aneurysms, 73 carotid endarterectomies, 20 coronary atherectomies, and 76 diseased peripheral arteries (57 femoral, 9 popliteal, 1 tibial, and 9 iliac arteries). Eight mammary arteries obtained during coronary artery bypass grafting and six fetal aortas obtained at autopsy (the deaths were from noncardiovascular, noninfectious causes) were considered not to be prone to atherosclerosis and were included as controls.

DNA extraction. DNA was extracted from 20 mg of tissue with a QIAamp DNA mini kit (4), according to the instructions of the manufacturer (Qiagen, Westburg, The Netherlands). The extracted DNA was processed immediately or stored at −20°C for a maximum of 2 years. The quality of the extracted DNA was evaluated by a PCR specific for the human β-globin gene, as described previously (17).

Conventional PCRs. All 197 atherosclerotic lesions were analyzed by the first PCR (PCR 1) with primer pair HL1-HR1 (9). Use of the HL1-HR1 primer pair was recently recommended by the Centers for Disease Control and Prevention (Atlanta, Ga.) and by the Laboratory Centre for Disease Control (Ottawa, Ontario, Canada) in an effort to standardize the technique (13). Amplification was done as described previously (17). An internal control was added in each amplification reaction (37).

Two additional PCRs were performed with a subgroup of 129 randomly chosen atherosclerotic specimens. Primers 53.1 and 53.2 were used in the second PCR (PCR 2) (22). Amplifications were performed in a final volume of 100 μl by using a Qiagen PCR master mixture with 2.5 mM MgCl₂, 20 pmol of each primer, and 6 μl of extracted DNA. A new, second internal control was added to each reaction mixture and was constructed by amplifying herpes simplex virus DNA with the composite primers 5′-ATGATCGCGGTTTCTGTTGCCAAGTCGTGTGCTGTTTC-3′ and 5′-GAGCGACGTTTTGTTGCATCTCCCCCGCGCGCCCCGAGA-3′, followed by amplification with the 53.1-53.2 primer pair as described previously (37).

A third seminested PCR (PCR 3) was performed in an independent laboratory with the same 129 specimens (8, 29). This PCR was based on the same target used in PCR 1, but with the primers modified to reduce the extent of primer-dimer formation.

For the negative controls, the clinical samples were replaced by distilled water. Amplified products were detected by agarose gel electrophoresis in the presence of separately amplified C. pneumoniae DNA as a positive control.

Amplification products obtained by PCR 1 and PCR 2 were hybridized with probes specific for C. pneumoniae amplicons. Hybridization (Hybridowell Universal kit; Argene Biosoft, Varilhes, France) was performed with probes labeled at their 5′ ends with biotin (Eurogentec, Seraing, Belgium). Probes HM1 (5′-GTGTCATTCGCCAAGGTTAA-3′) (9) and TWAR PROB3 (5′-GCCATATCTCTCTAACGGAG-3′) were used for PCR 1. Probes TWAR PROB1 (5′-GCTGCTGCCGCAACCACGGT-3′) and TWAR PROB2 (5′-AAGCTGTTGTCCAAGCGGTG-3′) were used for PCR 2.

Amplicons from mammary arteries analyzed by PCR 1 were examined only by agarose gel electrophoresis.

Real-time PCR. All 197 atherosclerotic fragments plus the 8 mammary arteries were examined by an inhibitor-controlled real-time PCR assay. The primers described by Tondella et al. (36) and two fluorescent probes targeting the VD4 domain of the ompA gene were used. The probe sequences were as follows: probe VD4FL, CTGCATGGAACCCTTCTTTACTAGGAA-fluorescein; probeVD4LC,LightCyclerRed640-TGCCACAGCATTGTCTACTACTGATTC.Four microliters of extracted DNA was analyzed with 16 μl of a PCR hot-start reaction mixture, consisting of 1× LightCycler-FastStart DNA Master Hybridization Probes mixture (Roche Diagnostics, Vilvoorde, Belgium) containing 5 mM MgCl₂, 500 nM each primer, and 200 nM each fluorescent probe. A third internal control was constructed by amplification of phage λ DNA with the composite primers 5′-TCCGCATTGCTCAGCCGGACGCGGGCGCTG-3′ and 5′-AAACAATTTGCATGAAGTCTGAGAAGCGCCCATACCGGTTTTATG-3′, followed by amplification with primers VD4F and VD4R, as described previously (37). Amplification and detection of PCR products were performed with a LightCycler system (Roche Molecular Biochemicals). The thermal cycling conditions were 95°C for 10 min; 20 cycles of denaturation at 95°C for 15 s, annealing at 65°C for 10 s, and elongation at 72°C for 10 s; and 4 cycles in which the annealing temperature was lowered in steps of 1 to 61°C, followed by 16 identical cycles in which the annealing temperature was lowered to 60°C. Each run contained a negative control, in which sterile water replaced the specimen, and a positive control consisting of C. pneumoniae DNA.

Histology and immunohistochemistry assay. Eighty atherosclerotic specimens were randomly selected. Sections were prepared from frozen or formalin-fixed tissue, which was decalcified in FE10 (100 g of EDTA, 12 g of NaOH, 100 ml of formalin, 900 ml of distilled water) and embedded in paraffin. Each paraffin block contained two pieces of plaque material taken from different parts of the lesion.

One section of each specimen was stained with hematoxylin-eosin. The extent of atherosclerosis was determined according to the Stary classification (34, 35). The immunohistochemistry assay was performed by the streptavidin-biotin complex-horseradish peroxidase technique (19). The following primary monoclonal antibodies (MAbs) were used: a C. pneumoniae species-specific antimembrane protein MAb, RR402 (Washington Research Foundation, Seattle, Wash.); a Chlamydia genus-specific antilipopolysaccharide (anti-LPS) MAb, CF2 (Washington Research Foundation); and a Chlamydia genus-specific anti-heat shock protein 60 (anti-hsp60) MAb (Affinity Bioreagents Inc., Golden, Colo.). Macrophages and smooth muscle cells were identified with an anti-CD68 MAb (DakoCytomation, Carpinteria, Calif.) and anti-human muscle actin MAb, clone HHF35 (DakoCytomation). Positive control slides contained acetone-fixed HEp-2 cells infected with C. pneumoniae. For the negative controls, the primary antibody was omitted and an unrelated antibody of the same isotype was substituted for it: monoclonal mouse antibody to synaptophysin, isotype immunoglobulin G1 (IgG1; Biogenex, San Ramon, Calif.); monoclonal mouse anti-human Ki-1 antibody, isotype IgG3 (DakoCytomation); or monoclonal mouse anti-human cytokeratin 20 antibody, isotype IgG2a (DakoCytomation). The slides were scored as follows (16): 0, absence of positive cells; 1+, one to five positive cells per tissue section; 2+, more than five positive cells per tissue section; 3+, foci with many C. pneumoniae-positive cells per tissue section; 4+, many foci with abundant C. pneumoniae-positive cells per tissue section.

Western blotting. Serial frozen sections (n = 5; sections 1 to 5, respectively) from 17 randomly selected human atheromata and 6 consecutive fetal aortas were examined by Western blotting. Equivalent amounts of protein (according to a β-actin blot) from each specimen were loaded onto a sodium dodecyl sulfate-12.5% polyacrylamide gel. After electrophoresis, samples were transferred to Hybond nitrocellulose membranes (Amersham Pharmacia Biotech, Uppsala, Sweden). The signals were visualized and quantified with a Lumi-imager (Roche Diagnostics). The following primary mouse MAbs were used: a Chlamydia genus-specific anti-hsp60 MAb (Affinity Bioreagents Inc.); an anti-C. pneumoniae MAb, clone M73066 (IMGEN); and an anti-C. pneumoniae MAb, VM1691 (IMGEN).

Localization of autofluorescent deposits in the same atheromata. A sequentially cut frozen section (section 6) from each of the 17 atheromata was examined by UV light excitation with a fluorescence microscope equipped with an Olympus U-MNU filter set (excitation filter BP 360 to 370 nm, dichroic mirror DM 400, and emission filter BA 420). An immunohistostaining assay for C. pneumoniae antigens was done with the next three slides (sections 7 to 9, respectively). The autofluorescent and C. pneumoniae-stained sections were projected on a screen, and the images were mapped.

Statistical analysis. Comparisons of the rates of positivity for C. pneumoniae components in different vascular regions and in relation to the Stary classification were done by the Pearson chi-square test. The number of positive samples was expressed as a percentage of the total number of specimens examined. A P value <0.05 was considered significant.

Stary classification. Histopathologic examination revealed moderate to severe atherosclerotic disease, Stary grades 4 through 7: 23 arterial segments were classified as atheromata, 4 were classified as fibroatheromata, 16 were classified as complicated fibroatheromata, and 37 were classified as calcified lesions. Mammary arteries were classified as Stary grade 1. No microscopic evidence of atherosclerosis was seen in any of the six fetal aortic arteries.

Detection of C. pneumoniae DNA by PCR. C. pneumoniae DNA was not detected in any of the arterial plaque fragments by any of the four PCRs applied. Positive and negative controls reacted appropriately. PCR inhibition was sometimes observed but could be eliminated by diluting the sample 1:2 or, on some occasions, 1:5. The sensitivity of PCR 1 was approximately 40 molecules per reaction mixture, and that of PCR 2 was 4 molecules per reaction mixture. The real-time PCR detected as little as 0.02 inclusion-forming units (IFU) of C. pneumoniae per reaction mixture, and the seminested PCR detected 0.001 IFU/reaction mixture. A PCR for β-globin was positive for all specimens tested.

Detection of chlamydial antigens by immunohistochemistry analysis. C. pneumoniae species-specific membrane protein RR402 was detected in 63 of 80 (79%) of the atherosclerotic specimens tested; equal proportions of all arteries were positive: 7 of 11 (64%) aortas, 25 of 31 (81%) carotid arteries, and 31 of 38 (82%) peripheral arteries (P value = 0.417). There was no association between C. pneumoniae membrane protein positivity and the extent of atherosclerosis (P = 0.580). Serial sections demonstrated that areas rich in smooth muscle cells were predominantly positive for RR402.

Genus-specific reactivity was much less prevalent: 9 of 80 (11%) atherosclerotic specimens were positive for chlamydial LPS, and 13 (16%) of the 80 specimens stained positive for chlamydial hsp60. The prevalences of chlamydial hsp60 and LPS within the different vascular regions did not differ significantly (P = 0.289 and 0.731, respectively). The presence of genus-specific antigens was not related to the extent of disease: chlamydial hsp60 was present in 15% of the atheromata and fibroatheromata, 19% of the complicated fibroatheromata, and 16% of calcified lesions (P = 0.944), while chlamydial LPS was detected in 7, 19, and 11% of those specimens, respectively (P = 0.520). Serial sections showed that chlamydial antigen-positive cells were predominant in areas rich in macrophages and smooth muscle cells.

Specimens positive for chlamydial LPS or hsp60 also showed 2+ to 4+ staining for the species-specific membrane protein RR402 (Table 1).

TABLE 1. Semiquantitative data for positivity for chlamydial hsp60, chlamydial LPS, and C. pneumoniae RR402 in atherosclerotic lesionsa

All eight mammary arteries scored positive for C. pneumoniae: eight with C. pneumoniae species-specific antibody RR402, three with the genus-specific anti-LPS antibody, and two with the genus-specific anti-hsp60 antibody. Many foci contained numerous cells that stained 4+ with the RR402 and CF2 MAbs. A lower score (3+) was obtained for tissues positive with the anti-hsp60 MAb and one specimen positive for chlamydial LPS. No immunoreactivity with the anti-C. pneumoniae antibodies was observed in the fetal aortas.

Control slides containing acetone-fixed HEp-2 cells infected with C. pneumoniae reacted with all three antibodies to Chlamydia. Vascular tissue regions that scored positive for C. pneumoniae did not show immunoreactivity with irrelevant isotype-matched primary antibodies.

Detection of C. pneumoniae proteins by Western blotting. No specific signals were detected by Western blotting assays performed with protein extracts derived from frozen sections of 17 randomly selected atherosclerotic arteries. Controls consisting of C. pneumoniae proteins consistently gave positive reactions, the detection limit being 490 IFU/reaction mixture. Fetal aortic vessels were negative.

Colocalization study by UV light excitation. Of the 17 atheromata examined in serial sections by Western blotting, immunohistostaining, and UV light excitation assays, 16 stained abundantly with the RR402 MAb, and 14 and 15 showed identically located but less massive staining with the anti-chlamydial hsp60 MAb and the VM1691 MAb, respectively. The sites of autofluorescence perfectly matched the immunoreactive sites (Fig. 1). Fetal vessels showed no reactivity.

FIG. 1. Perfect matching of sites of C. pneumoniae immunoreactivity with sites with autofluorescent ceroid deposits. (A) Human atherosclerotic plaque stained by the C. pneumoniae species-specific anti-membrane protein MAb RR402; (B) detailed view of the boxed (more ...)

In this study we aimed to investigate the controversial detection of C. pneumoniae in human atherosclerotic plaques by means of conventional and real-time PCR formats and by immunohistochemistry, Western blotting, and UV light excitation assays.

Four PCRs targeting three different C. pneumoniae genes performed with 197 atherosclerotic lesions were negative.

Immunohistochemistry assays with three different MAbs to C. pneumoniae on 80 of these 197 atherosclerotic specimens detected a C. pneumoniae species-specific membrane protein (MAb RR402) significantly more frequently than chlamydial hsp60 or chlamydial LPS (79% versus 16 and 11%, respectively). No association was found between this immunoreactivity for chlamydial antigens and the extent of atherosclerotic disease, nor was a predisposition for any particular vascular region present.

Western blotting for C. pneumoniae proteins in adjacent atherosclerotic sections was negative. The immunoreactive vascular sites detected with the anti-C. pneumoniae MAbs perfectly matched the sites with UV light-induced autofluorescence of ceroid, an insoluble lipid that is commonly present in the diseased vasculature.

Published studies on the presence of C. pneumoniae DNA in human atherosclerotic plaques resulted in prevalences ranging from 0 to 83% (32). However, Apfalter et al. (1) reported major problems with both inter- and intralaboratory reproducibilities. In a multicenter study with nine laboratories for the detection of C. pneumoniae DNA by PCR with aliquots of homogenized tissue of the same atherosclerotic specimens, positivity rates varied from 0 to 60%. The maximum rate of concordant results for positivity among the various laboratories was only 25% for one carotid artery. In addition, no correlation between the detection rates and the sensitivity of the methods used was found (1). In a second multicenter study with C. pneumoniae, Apfalter et al. (2) demonstrated that even experienced laboratories suffer from major contamination problems with nested PCR, pointing to the possibility that the literature in the field might be biased by false-positive PCR results. In this study, no C. pneumoniae-specific DNA was detected in 197 different atheromatous human arteries (carotid, coronary, aortic, popliteal, femoral, and iliac arteries) by the application of three conventional gel-based PCRs and one real-time PCR targeting three different DNA fragments of the organism: a C. pneumoniae PstI fragment, a C. pneumoniae 53-kDa protein gene, and the VD4 domain of the C. pneumoniae ompA gene. These negative PCR results are in accordance with the results of a growing number of studies aimed at the detection of C. pneumoniae-specific DNA in human atherosclerotic lesions and abdominal aortic aneurysms (6, 10, 20, 21, 23, 24, 25, 26, 28, 31, 39), suggesting either strong DNA degradation or the complete absence of C. pneumoniae DNA in atherosclerotic lesions.

As in other studies, we also observed immunoreactivity for C. pneumoniae in the macrophages and the smooth muscle cells of atherosclerotic plaques (42). The higher prevalence of lesions reacting with the C. pneumoniae species-specific MAb (RR402) rather than with the genus-specific MAbs was not unprecedented either. Meijer et al. (26) examined specimens from 19 patients with abdominal aortic aneurysms and detected the C. pneumoniae-specific membrane protein in all specimens, while Chlamydia genus-specific antibody CF2 produced a positive result with only 32% of the specimens.

In general, C. pneumoniae antigens are more frequently detected than C. pneumoniae DNA, or, as stated in other terms, immunohistostaining procedures yield more positive results for C. pneumoniae than PCR techniques. Boman and Hammerschlag (7) commented extensively in their review on the poor correlation between antigen and DNA detection methods in the same atherosclerotic plaque. One of the reasons for this discrepancy has been attributed to the presence of large quantities of inhibitors in the extracted DNA (40). In a study by Thomas et al. (35), tissues with severe lesions tended to contain more inhibitors than tissues with mild lesions. In this study, the negative PCR results cannot be explained by the presence of inhibitors, since an internal control was included in each amplification reaction to monitor possible inhibition. In addition, the amount of internal control added to each reaction mixture was minimal in order to avoid interference with possible small concentrations of target DNA present in the sample. Furthermore, no internal control was used in the seminested PCR. No false-positive amplifications were detected. Finally, the amplification of the human β-globin gene was positive for all specimens tested, indicating that the DNA extraction was efficient and contained amplifiable DNA.

Western blotting confirmed the negative findings for C. pneumoniae. The immunoblotting technique might not be able to notice very small amounts of the organism's proteins in the sample, but it should have had a positive result, given the consistent and abundant positive immunohistostaining assay results for the sequentially cut arterial sections. The negative results obtained by the DNA detection efforts, as well as the absence of C. pneumoniae proteins in atherosclerotic lesions, as revealed by Western blotting, also mentioned previously by Vammen et al. (38) for abdominal aortic aneurysms, strongly question the validity of the immunohistochemistry assay results, pointing to the possibility of nonspecific reactions. Fluorescence under UV light identified these immunoreactive sites as ceroid deposits.

Ceroid is an autofluorescent insoluble complex of oxidized lipid and protein derived from oxidation and peroxidation of phospholipids and unsatured fatty acids (11, 27). Ceroid can be abundantly present as tiny granules in both early and advanced atherosclerotic lesions. It is found within the cytoplasms of lipid-loaded macrophages and foam cell-like smooth muscle cells, and it is also found extracellularly as granular or ring-shaped deposits in patients with necrosis (3). Immunohistochemical localization of antigens in the human vasculature can be complicated by the presence of complex molecules such as ceroid, which may cross-react with antibodies used in immunohistochemistry analyses. In the present study, sites of immunoreactivity for C. pneumoniae in atherosclerotic lesions matched precisely the sites of autofluorescent ceroid deposits. Interestingly, Binder et al (5). showed molecular mimicry between epitopes of oxidized low-density lipoproteins and those of Streptococcus pneumoniae. Ceroid in the human atherosclerotic plaque is partly composed of such oxidized low-density lipoproteins. In analogy with the findings of Binder and colleagues (5), epitopes of oxidized low-density lipoproteins may also share high degrees of homology with epitopes of C. pneumoniae, resulting in nonspecific staining reactions of the antichlamydial antibodies in atherosclerotic lesions.

Conclusion. The negative results for amplification of C. pneumoniae DNA from human atherosclerotic lesions, together with negative results by Western blotting for detection of C. pneumoniae proteins, and the exact matching of positive C. pneumoniae histological staining sites with sites of autofluorescent ceroid deposits suggest cross-reactivity of antichlamydial MAbs with nonchlamydial plaque constituents. On the basis of the uncertainties of the seroepidemiologic evidence (18) and the present results, there are no arguments to attribute a direct role for C. pneumoniae in the pathogenesis of atherosclerosis.

This work was supported by the Fund for Scientific Research—Flanders (project no. 7.0010.98).

We thank Inez Rodrigus, Department of Cardiac Surgery, University Hospital Antwerp, for providing the mammary arteries from patients undergoing coronary bypass surgery. We also thank Sonia Hirt, Department of Medical Microbiology, University of Zurich, and Lieve Crauwels, Hilde Wouters, and Kristof Bergs, Department of Microbiology, University Hospital Antwerp, for excellent assistance with the PCRs. We thank Paul Van Schil, Department of Thoracic and Vascular Surgery, University Hospital Antwerp, for the atherosclerotic specimens.