Methods

PATIENTS AND SPECIMENS

Carotid artery specimens taken during necropsy of 13 men were obtained from the pathology department of the National Centre for Occupational Health in Johannesburg, South Africa, within one to seven days after death in January 1997 (kindly provided by Dr A Shor) (table 1). The specimens were routinely fixed in 10% buffered formalin for 48 hours, embedded in paraffin wax, and shipped to the National Institute of Public Health and the Environment, Bilthoven, The Netherlands. One 4 μm section of each paraffin wax embedded specimen was stained with haematoxylin and eosin for characterisation of the microscopic pathology, according to the classification of atherosclerosis defined by the American Heart Association.¹²

IMMUNOCYTOCHEMISTRY AND IN SITU HYBRIDISATION

Sections of HEp2 cells (code CCL23; American Type Culture Collection, Manassas, Virginia, USA) infected with C pneumoniae strain TW-183 (Washington Research Foundation (WRF), Seattle, Washington, USA) at a multiplicity of infection of 0.1 inclusion forming units, and of mock infected HEp2 cells, fixed and embedded in paraffin wax similar to the clinical specimens, were used as positive and negative controls, respectively.

Chlamydia pneumoniae antigen was detected in 4 μm sections by immunocytochemistry using an indirect immunoperoxidase method, as described and evaluated previously.^6,13 A species specific monoclonal antibody, RR-402 (WRF), against a C pneumoniae membrane protein was used.¹⁴ An irrelevant mouse monoclonal antibody of the same isotype as monoclonal antibody RR-402 (IgG3) was used as a negative control antibody. For the identification of macrophages and smooth muscle cells, monoclonal antibodies PG-M1, directed against the macrophage marker CD68, and 1A4, directed against smooth muscle cell actin (DAKO A/S, ITK Diagnostics BV, Uithoorn, The Netherlands), were used.

The detection of C pneumoniae DNA by in situ DNA hybridisation (ISDH) and of Chlamydiales specific 16S rRNA by in situ RNA hybridisation (ISRH) on 4 μm sections adjacent to the sections used for ICC was carried out as described and evaluated previously.^6,13 ISDH was carried out with a digoxigenin (DIG) labelled C pneumoniae major outer membrane protein (MOMP) gene fragment as a probe and with a DIG labelled minute virus of mouse (MVM) genome fragment as a negative control probe. ISRH was carried out with the following DIG labelled oligonucleotide probes: (1) an antisense probe specific for 16S rRNA of the order Chlamydiales; (2) a sense probe to assess the specificity of the antisense probe for rRNA; and (3) a nonsense probe composed of the same nucleotides as the antisense probe but in a different sequence as a negative control probe. A 3′ and 5′ DIG labelled probe of 33 nucleotides, specific for eukaryotic 28S rRNA, was used to assess the presence of hybridisable RNA in the carotid artery sections.¹⁵

Specimens were considered positive for C pneumoniae antigen or DNA, or Chlamydiales specific 16S rRNA when clear cell associated staining was observed.

PCR ASSAYS

Before DNA extraction, five 4 μm sections of each specimen were dewaxed using UltraClear (Mallinckrodt Baker BV, Deventer, The Netherlands), washed with ethanol and acetone, and dried. As a positive control, one 4 μm section of the lung of a mouse two days post intranasal infection with C pneumoniae¹³ was used. DNA isolation using the Easy DNA kit (Invitrogen BV, Leek, The Netherlands) and PCR assays for C pneumoniae 16S rDNA (165 bp fragment) and for the human β-globin gene (110 bp fragment) were carried out as described previously.^16,17

TERMINAL DEOXYNUCLEOTIDYL TRANSFERASE (TDt) mediated fluorescein dutp nick end labelling (tunel) assay

The “in situ cell death detection kit, POD” (Boehringer Mannheim, Mannheim, Germany) was used to detect free 3′ ends of DNA. Sections (4 μm thick) adjacent to those used for ICC were dewaxed, rehydrated, and treated with 2.5 mg/litre proteinase K in 50 mmol/litre Tris/HCl (pH 7.6) containing 2 mmol/litre CaCl₂ for 15 minutes at 37°C. They were then washed twice for two minutes with PBS, blocked for endogenous peroxidase by incubation in 0.3% H₂O₂ in methanol for 30 minutes, and washed twice for five minutes with PBS. Optimal staining was achieved by using a dilution of the TdT enzyme (1/2 in 30 mmol/litre Tris/HCl (pH 7.2) containing 140 mmol/litre sodium cacodylate and 1 mmol/litre CoCl₂), and of the peroxidase labelled antifluorescein antibody (1/2 in 100 mmol/litre Tris/HCl (pH 7.5) containing 150 mmol/litre NaCl and 30 g/litre bovine serum albumin (BSA) fraction V), according to the manufacturer's instructions. Sections were incubated for 60 minutes at 37°C with TUNEL reaction mixture (5 μl 1/2 diluted TdT enzyme, 45 μl labelling solution containing fluorescein labelled dUTP and unlabelled dNTPs), washed three times for one minute with PBS, and incubated for 30 minutes at 37°C with 1% BSA fraction V in PBS. This was followed by 30 minutes at 37°C with 1/2 diluted peroxidase labelled antifluorescein antibody, after which the sections were washed three times for two minutes with PBS. Peroxidase was visualised with diaminobenzidine (Sigma-Aldrich Chemie BV, Zwijndrecht, the Netherlands). The sections were counterstained with Mayer's haematoxylin. An artificial positive control of a carotid artery section was prepared by treatment of the section with 1 × 10⁶ U/litre DNAse I (Boehringer Mannheim) in 50 mmol/litre Tris/HCl (pH 7.5) containing 1 mmol/litre MgCl₂ and 1 g/litre BSA fraction V for 10 minutes at room temperature. For each carotid artery specimen, a background staining control was included, which was incubated with the TUNEL reaction mixture without the TdT enzyme. Sections that showed staining of matrix vesicles were pretreated with 10% EDTA in PBS for one hour, as described previously.¹⁸

Sections of C pneumoniae infected HEp2 cells that were incubated with the complete TUNEL mixture, with and without pretreatment of the sections with DNAse I, or incubated with the TUNEL mixture without the TdT enzyme, were used as controls.

Results

HISTOLOGY

Of the 13 carotid artery specimens, five were classified as type I (normal artery/initial lesion), two as type II (fatty streak), and six as advanced lesions of type IV to V/VI (table 1).

IMMUNOCYTOCHEMISTRY AND IN SITU HYBRIDISATION

Inclusions in HEp2 cells infected with C pneumoniae stained strongly with the C pneumoniae membrane protein specific monoclonal antibody RR-402, the C pneumoniae specific MOMP probe, and the Chlamydiales 16S rRNA specific antisense oligonucleotide probe, whereas no staining was seen in mock infected HEp2 cells. No staining of inclusions was observed with the MVM probe, the sense oligonucleotide probe, or the nonsense oligonucleotide probe.

Immunoreactivity with the C pneumoniae membrane protein specific monoclonal antibody was detected in all type IV–VI lesions (fig 1A), whereas it was not detected in type I and II lesions (table 1). Immunoreactivity was always seen in cells within or in close proximity to the lesion, but not in adjacent normal artery tissue. No staining was observed with the control monoclonal antibody. Cells immunoreactive for C pneumoniae membrane protein appeared to be macrophages and were found in areas in which cells showed immunoreactivity for the macrophage marker CD68 (fig 1B) in adjacent sections. Cells immunoreactive for smooth muscle cell actin were not seen in these areas. The specimens gave strong signals in ISRH with the eukaryotic 28S rRNA probe (fig 1C), demonstrating the suitability of the specimens for hybridisation experiments. None of the specimens from the 13 cases was positive in ISDH for C pneumoniae DNA (fig 1D). Sections from cases 2, 4, 5, and 9 detached from the slides during ISRH and were not included further in the ISRH experiments. Chlamydiales 16S rRNA was not detected in any of the other specimens.

PCR ASSAYS

The DNA extracted from the positive control showed a strong C pneumoniae specific band upon PCR analysis. The 110 bp β-globin fragment was amplified from the DNA extracted from the 13 patient specimens, indicating the presence of PCR amplifiable DNA. Nonetheless, the C pneumoniae specific 16S rDNA fragment could not be amplified from any of the DNA extracts from these patients.

TUNEL ASSAY

A carotid artery section pretreated with DNAse I and subsequently processed by the TUNEL assay showed partial to total staining of the nuclei of most cells, indicating the ability of the procedure to label fragmented DNA. To investigate whether fragmented chlamydial DNA could be visualised, the TUNEL assay was applied to HEp2 cells infected with C pneumoniae. No staining for fragmented DNA was seen when these cells were incubated with the TUNEL reaction mixture, with or without TdT enzyme. However, stained fragmented chlamydial DNA could be seen in clearly recognisable inclusions in these cells with the TUNEL assay after DNAse I pretreatment of the cells (fig 1E). The staining pattern of fragmented chlamydial DNA appeared as small dense dots of uniform size. Therefore, this pattern of TUNEL staining was analysed in the carotid artery specimens and could easily be distinguished from the typical staining pattern of fragmented eukaryotic DNA from apoptotic nuclei and apoptotic bodies (not uniform in size) in the cytoplasm of cells. The TUNEL assay without the TdT enzyme in the reaction mixture applied to sections of cases 2, 3, 4, 11, 12, and 13 showed staining of clearly recognisable calcium deposits and faint staining of some other structures. This type of staining was abolished by EDTA pretreatment of the sections. Sections from the other patients showed no staining in the TUNEL assay without the TdT enzyme in the reaction mixture. Table 1 summarises the results of the TUNEL assay on the carotid artery specimens. Small uniformly sized dots stained by the TUNEL assay were observed in the cytoplasm of cells in the pathological areas of the arteries (fig 1F and G). Without the TdT enzyme in the reaction mixture, this type of staining was not seen (fig 1H). The cells with cytoplasmic TUNEL staining were found in areas where cells positive for the macrophage marker CD68 but not smooth muscle cell actin were found in adjacent sections. This cytoplasmic TUNEL staining was not caused by calcium deposits, which can cause non-specific TUNEL staining,¹⁸ because EDTA pretreatment of the sections did not affect this type of staining. In four of the six cases that showed cells immunoreactive for C pneumoniae membrane protein, cells with cytoplasmic TUNEL staining were also seen (table 1). Cells immunoreactive for C pneumoniae membrane protein were morphologically similar to those cells in which the cytoplasm stained positive in the TUNEL assay in adjacent sections, and the stained cells showed a similar distribution pattern in both assays.

Discussion

We have demonstrated the presence of immunoreactivity for C pneumoniae membrane protein in macrophages in advanced atherosclerotic lesions (type IV–VI) of carotid arteries, but not in carotid arteries at lesion prone sites (type I) or in fatty streaks (type II). Genomic C pneumoniae DNA and chlamydial 16S rRNA could not be detected in any of the carotid arteries by in situ hybridisation, and neither could C pneumoniae 16S rDNA by PCR. Fragmented chlamydial DNA, demonstrated by the TUNEL assay, was present as uniformly sized dots in cultured cells infected with C pneumoniae and pretreated with DNAse I. This type of TUNEL staining was also detected in carotid artery specimens, in the cytoplasm of cells that were also immunoreactive for C pneumoniae membrane protein in adjacent sections.

First, we analysed the presence of C pneumoniae in carotid arteries by ICC for membrane protein and by ISDH and PCR. Immunoreactivity for C pneumoniae membrane protein was detected in six carotid arteries, whereas C pneumoniae specific DNA could not be demonstrated by ISDH or PCR, in agreement with our previous results using specimens from abdominal aortic aneurysms.⁶ The presence of immunoreactivity for C pneumoniae membrane protein in advanced lesions was in agreement with the results of Kuo et al.¹⁹ We did not see immunoreactivity for C pneumoniae membrane protein in fatty streaks from two patients, although Kuo et al found the antigen in fatty streaks of seven of 16 patients.¹⁹ This discrepancy probably results from the small number of specimens in our study. Chlamydia pneumoniae membrane protein was found in cells of the macrophages population, as determined by their morphology and immunoreactivity for the macrophage marker CD68 but not for smooth muscle cell actin. These findings are in agreement with results from coronary atherectomy specimens, presented previously.²⁰ However, in another study C pneumoniae antigens were found in macrophages as well as in smooth muscle cells in aortic lesions.²¹ It should be noted, however, that atherectomy specimens do not always contain the deeper layers in which the positive smooth muscle cells were observed in the aorta biopsies.²² Our carotid artery specimens included all layers.

Thus, we were not able to demonstrate the presence of C pneumoniae DNA, whereas antigens reactive with the monoclonal antibody RR-402 were found regularly. Therefore, we evaluated the possible involvement of other members of the order Chlamydiales. No evidence was obtained for the presence of 16S rRNA from any member of the order Chlamydiales by ISRH. This observation is consistent with the absence of any hybridisable C pneumoniae DNA, supporting the hypothesis that only antigens persist after infection of the vessel wall.

The in situ hybridisation results can be considered reliable because control sections of HEp2 cells infected with C pneumoniae consistently showed staining of inclusions with the MOMP and the antisense 16S rRNA probes, and these methods have been evaluated and used successfully in studying C pneumoniae infected mice.¹³ Furthermore, sections of all carotid artery specimens that could be analysed showed staining for eukaryotic 28S rRNA, indicating the suitability of these specimens for in situ hybridisation experiments.

Finally, we evaluated the possibility of the rapid degradation of DNA and the persistence of membrane protein to explain our results. Fragmented chlamydial DNA was present as uniformly sized dots in cultured cells infected with C pneumoniae and treated with DNAse I. The absence of TUNEL staining without DNAse I pretreatment showed that viable C pneumoniae does not contain fragmented DNA. Cells with uniformly sized dots in their cytoplasm stained by the TUNEL assay were found in four of six cases with cells immunoreactive for C pneumoniae membrane protein. The distribution of TUNEL positive cells was similar to that of cells positive for C pneumoniae membrane protein. This suggests that fragmented DNA in these cells might originate from C pneumoniae. The bacterial DNA might have been fragmented in the natural process of clearing an infection by phagocytosis of bacteria and subsequent degradation in lysosomes, or as a consequence of conditions in the infected cell initiating a process of DNA fragmentation. The first process is most likely because C pneumoniae specific membrane protein (our study), chlamydial specific LPS, and heat shock protein 60 (hsp60) have been found in macrophages (professional phagocytes).^6,20,21,23 However, limited replication of C pneumoniae in in vitro experiments using the monocytic cell line U-937, or fresh human peripheral blood mononuclear cells, suggests that the second option is also possible.^24,25

An analogy of C pneumoniae infection in atherosclerosis with chronic infections caused by Chlamydia trachomatis has been suggested.²⁶ These chronic infections are characterised by the presence of persistent forms of C trachomatis. A hallmark of persistence of C trachomatis in trachoma, tubal occlusion, and Reiter's syndrome is the presence of nucleic acids, whereas only small amounts of antigens are detected in a small fraction of specimens positive for nucleic acids.^27–30 In contrast, in this and a previous study,⁶ we noted an abundance of C pneumoniae antigens in the complete absence of hybridisable DNA and 16S rRNA in atherosclerotic lesions. Thus, it is more likely that antigens of C pneumoniae, rather than viable bacteria, persist in atherosclerotic lesions after infection. Alternatively, the presence of C pneumoniae antigens in atherosclerotic lesions in the absence of its DNA might be explained by the “travelling” of macrophages, which have ingested and degraded bacteria at other sites in the body infected by C pneumoniae (for example, the lung), to the area of inflammation in the atherosclerotic lesion, as suggested by Capron.³¹ Nevertheless, this might have important implications for prevention and treatment strategies, as has already been demonstrated in rabbits infected with C pneumoniae.³² A seven week course of azithromycin had no effect on the presence of antigens in aortic tissue compared with untreated rabbits. Furthermore, persistent chlamydial antigens, such as LPS and hsp60, but not necessarily viable bacteria, might very well contribute to the pathogenesis of atherosclerosis through the induction of macrophage foam cell formation and the activation of macrophages.^23,25,33,34

In contrast to our findings, viable C pneumoniae has been cultured more^3,4,35 or less⁸ successfully from atherosclerotic tissue. An explanation for this discrepancy might be that specimens were collected during episodes of circulating C pneumoniae in the blood, and cultures might become positive as a result of bacteria or infected blood cells trapped in the vasa vasorum. In a large proportion of patients with cardiovascular disease, and in blood donors, C pneumoniae DNA can be detected in peripheral blood mononuclear cells.³⁶ In situ techniques, as used in our present study, are the only way to locate viable C pneumoniae organisms and their components in cells in the lesion of the vessel wall.

In conclusion, evidence has been obtained for the presence of fragmented DNA in cells positive for C pneumoniae membrane protein in advanced atherosclerotic lesions. No evidence was found for the involvement of other members of the order Chlamydiales in atherosclerosis. The demonstration of the presence of C pneumoniae membrane protein in the absence of DNA and 16S rRNA supports the hypothesis of the persistence of antigens, rather than of viable bacteria, in atherosclerotic lesions.