Many studies indicate that Chlamydia pneumoniae infection is a crucial risk factor in atherogenesis. The most relevant cell type for the pathogenesis is the macrophage, which possesses classical scavenger receptors that uptake oxidized low-density lipoprotein (LDL). Here, a direct involvement of vascular endothelial cells in atherogenesis was examined employing in vitro infection of human umbilical vein endothelial cells (HUVEC) with C. pneumoniae. Chlamydia pneumoniae infection greatly enhanced the uptake of oxidized LDL, but not of acetylated LDL, by HUVEC. Among the scavenger receptors analyzed, LOX-1 transcription, which prefers oxidized LDL to acetylated LDL, was significantly amplified.
The possible involvement of Chlamydia pneumoniae infection in atherogenesis has been highlighted in various studies. The presence of the organism has been demonstrated in atherosclerotic lesions by immunohistochemical detection of specific antigens, PCRs of DNA (Kuoet al,1993; Graystonet al,1995; Davidsonet al,1998; Ouchiet al,1998) or recovery of the viable microorganism (Bartelset al,1999). Although the presence of C. pneumoniae in atherotic lesions was first shown in coronary arteries (Kuoet al,1993), it was confirmed in a variety of vascular systems (Rassuet al,2001). A correlation between the severity of atherosclerosis and C. pneumoniae infection has also been suggested (Ericsonet al,2000; Zairiset al,2003). Human endothelial cells, which may be the first to encounter the microorganism during systemic infection, are susceptible to C. pneumoniae (Kaukoranta-Tolvanenet al,1994; Godziket al,1995) and initiate inflammatory reactions in vitro. Namely, they produce chemokines such as MCP-1 or IL-8 (Molestinaet al,1998; Quinn & Gaydos, 1999), and express adhesion molecules (Krullet al,1999) to recruit inflammatory cells. Moreover, growth factors, which are produced by endothelial cells, may induce the proliferation of smooth muscle cells (Coombeset al,2002).
One crucial event for atherogenesis is believed to be foam cell formation, which is the result of lipid accumulation, especially oxidized low-density lipoprotein (LDL), in macrophages. Chlamydia pneumoniae infection was reported to accelerate LDL uptake and foam cell formation in macrophages, in a scavenger receptor- and LDL receptor-independent manner (Kalayoglu & Byrne, 1998; Kalayogluet al,1999). Thus, the recruitment of macrophages by C. pneumoniae-infected endothelial cells may be an important link between C. pneumoniae infection and atherogenesis. Furthermore, enhanced oxidation of LDL was observed in the presence of C. pneumoniae-infected macrophages (Kalayogluet al,1999).
In the present study, the possibility of a direct role of human endothelial cells in lipid accumulation was tested in vitro. In contrast to macrophage lineage cells, endothelial cells do not express significant amounts of classical scavenger receptors such as scavenger receptor types A, BI or CD36. However, several novel scavenger receptors have been discovered on endothelial cells. Among them, SREC (scavenger receptor expressed by endothelial cells) (Adachiet al,1997), FEEL-1 (fasciclin, EGF-like, laminin-type EFG-like, and link domain-containing scavenger receptor-1) (Adachi & Tsujimoto, 2002) and LOX-1 (lectin-like receptor for oxidized LDL) (Sawamuraet al,1997) were analyzed under the infection with C. pneumoniae.
Materials and methods
Collagenase type II, human plasma fibronectin, recombinant human basic fibroblast growth factor (bFGF), 0.53mM EDTA – 0.05% trypsin, MCDB131 medium and antibiotic-antimycotic solution were purchased from Invitrogen Corp. (Carlsbad, CA). Human LDL and acetylated LDL were obtained from Biomedical Technologies Inc. (Stoughton, MA), fetal calf serum (FCS) from Cambrian Chemicals Inc. (Ontario, Canada), and angiotensin receptor inhibitor peptide from Santa Cruz Biotechnology Inc. (Santa Cruz, CA).
Chlamydia strain and propagation
Chlamydia pneumoniae strain TW-183 was obtained from ATCC (Manassas, VA) and was propagated in Hela229 cells in the presence of 1μgmL−1 cycloheximide. The inoculation of the microorganism was performed by centrifugation at 500g at 25°C for 50min and subsequent incubation for 1h at 37°C in sucrose-phosphate-glutamate (SPG) buffer (Warfordet al,1984). The medium was replaced with MCDB131 (Invitrogen Corp.) supplemented with 10% FCS (HyClone Laboratories Incet al, Logan, UT) and 10μgmL−1 gentamicin, and cells were cultured for 2 days. The inoculated cells were scraped and sonicated for 10s in SPG buffer supplemented with 10% FCS. After the removal of cell debris by centrifugation at 500g for 10min, the supernatant was aliquoted and stored at −80°C. The titer was determined as inclusion forming units (IFU) on Hela229 cells using a fluorescein-labeled monoclonal antibody (DENKA SEIKEN Co. Ltdet al, Tokyo, Japan). Fixed C. pneumoniae were prepared by incubating for 1h in 3% formalin at 25°C. The preparation was washed twice with PBS with centrifugation at 50000g (Baeret al,2003).
Human umbilical vein endothelial cells (HUVEC) were prepared and cultured as previously described (Yoshidaet al,1999). Briefly, HUVEC were isolated with 0.25% collagenase type II, fractionated with Percoll (Amersham Biosciences Corp. Piskataway, NJ) at a density of 1.04, and were cultured in MCDB131 medium supplemented with 15% FCS, 2ngmL−1 bFGF and 1% antibiotic-antimycotic solution (Sigma-Aldrich, Corp. St Louis, MO). The purity of endothelial cells was over 98% as determined by the expression of factor VIII-related antigen on laser flow cytometry. All experiments were performed from day 5 to day 20 after cell collection. After day 5 of cell collection, no macrophages or monocytes were detected with anti-CD68 monoclonal antibody (Beckman Coulter Incet al, Fullerton, CA). Human monocytic cell line U937 was cultured in MCDB13 medium supplemented with 10% FCS and 1% antibiotic-antimycotic solution, and was stimulated by 1ngmL−1 of phorbol myristate acetate for 24h before use.
Chlamydia pneumoniae infection and cell viability analysis
Endothelial cells were plated to fibronectin-coated 12-, 24- or 96-well plates (Falcon, Cockeysville, MD), such that confluency was achieved. The monolayers were infected with C. pneumoniae at the indicated multiplicity of infection (MOI) in 10% FCS SPG buffer. The infection protocol was the same as used for propagation in Hela229, except that the cells were cultured in the absence of cycloheximide. The viability of cells was determined by uptake and digestion of calcein-AM (Molecular Probes, Eugene, OR). Cells were pulsed for 1h with the calcein-AM solution (2.5μM) in 10mM Hepes buffered saline pH 7.4 and the amount of intracellular calcein was measured as the fluorescence intensity of 520nm exited at 490nm. The relative viability was determined as the percentage of the ingested dye amount to the uninfected control, after subtracting the background value. The uninfected control was treated with an equivalent amount of Hela229 cell homogenate. Microscopic observations of endothelial cells were performed after 3 days of infection at MOI of 2. Cells were fixed with acetone and stained with fluorescein-labeled monoclonal antibody for fluorescence microscopy.
RNA extraction and semi-quantitative reverse transcriptase (RT)-PCR
After 2 days of infection with C. pneumoniae, total cellular RNA was extracted with TRIZOL (Invitrogen Corp.) according to the manufacturer's instruction and stored in the presence of rRNasin (Promega Coet al, Madison, WI) until analysis. Semi-quantitative reverse transcriptase (RT)-PCR was performed using the Titan one-tube RT-PCR system (Roche Diagnostics GmbH, Mannheim, Germany). Primer sequences were as follows: SREC (forward) 5′-CTGCTCAGTTCCTTGTGAATGC-3′, SREC (reverse) 5′-ATCGGATGAGAAGGAGTCATCAG, FEEL-1 (forward) 5′-AGCTTGCCTAGAGCTCAT, FEEL-1 (reverse) 5′-CAGCCGCTCATGGACACC, LOX-1 (forward) 5′-CAGCTGATCTGGACTTCATCCA, LOX-1 (reverse) 5′-TTGGCACCCAAGTGACAAAG, G3PDH (forward) 5′-ACCACAGTCCATGCCATCAC, G3PDH (reverse) 5′-TCCACCACCCTGTTGCTGTA. All primers were synthesized and purified by Rikaken Co. Ltd (Nagoya, Japan). Following reverse transcription at 50°C for 30min and denaturation at 95°C for 2min, the thermal cycling consisted of denaturing at 95°C for 30s, annealing for 30s and extension at 68°C for 1min. The annealing temperatures were 50°C for FEEL-1 and 58°C for other genes. The products were electrophoresed in 2% Seekem GTG agarose (Cambrian Chemicals Incet al, Oakville, Ontario, Canada) gel in TAE buffer and visualized with SyBR Gold (Invitrogen Corp.) fluorescence dye. The DNA band densities were quantified with Cool Saver (Atto Coet al, Tokyo, Japan). All RT-PCR experiments were repeated at least three times and representative data are shown.
Oxidation of LDL and ingestion assay
To produce oxidized LDL, LDL (100μgmL−1) was reacted with 1μM CuSO4 at 37°C for 16h after repeated dialysis against PBS. The reaction was terminated with 1mM EDTA and the solution was concentrated using Centri-Plus (YM-10, Millipore Corpet al, Bedford, MA). The oxidized LDL, acetylated LDL and LDL (80μg) were labeled with 18.5 MBq of Na125I (ICN Biochemicals Incet al, Irvine, CA) by reacting for 3min at 26°C in 100μL 0.25M sodium phosphate buffer (pH 7.4), in the presence of Iodo-beads (Bio-Rad, Hercules, CA). The labeled material was purified with PD-10 column (Amersham Biosciences Corp.), which was eluted with PBS. After 2 days of infection with C. pneumoniae, HUVEC in 24-well plates were incubated with the 125I-labeled LDL, oxidized LDL or acetylated LDL (2μgmL−1) for 2 days. The cells were washed three times with Hepes-buffered saline and lyzed with 0.1mL of 0.1N NaOH. The radioactivity was measured with a γ-counter and normalized to the protein content, which was determined with Coomasie Plus-200 protein assay reagent (Pierce Biotechnology Incet al, Rockford, IL). The data obtained in the presence of 50μgmL−1 cold counterparts were assumed as the background level.
Chlamydia pneumoniae infection enhanced oxidized LDL uptake by HUVEC
The primary culture of HUVEC could be infected with C. pneumoniae through a conventional method (Fig. 1), which was in agreement with previous studies (Kaukoranta-Tolvanenet al,1994; Godziket al,1995). Most of the cells were positive to the specific antibody at MOI of 2 (Fig. 1). Cell viability was reduced by approximately 20–30%, corresponding with the MOI ratios (Fig. 2). The amounts of total cellular protein were also decreased in parallel with cell viability (data not shown). Thus, the effect of C. pneumoniae infection on the uptake of LDL, oxidized LDL and acetylated LDL was assessed after normalization to the total cellular protein content. The uptake of oxidized LDL by HUVEC was increased by about 2.2-fold with C. pneumoniae infection (Fig. 2). The increment was statistically significant but less than that observed for the U937 monocyte cell line (2.6-fold increase, data not shown). The uptake of acetylated LDL was also increased to a limited degree (Fig. 2). In contrast, LDL uptake, of which the basal amount was much smaller, was not significantly affected. The enhanced uptake of oxidized LDL and acetylated LDL prompted us to analyze the expression of lipoprotein receptors after C. pneumoniae infection.
Chlamydia pneumoniae infection enhanced LOX-1 expression
The expressions of lipoprotein receptors were examined by comparing the amount of mRNA by RT-PCR. Although human endothelial cells do not express remarkable amounts of classical scavenger receptors, such as scavenger receptor type A, BI or CD36, they express particular classes of scavenger receptors that have been shown to participate in lipoprotein uptake. Among them, the expression of LOX-1 (Sawamuraet al,1997) was highly upregulated by C. pneumoniae infection for 2 days (Fig. 3). The degree of increment was 4- to 20-fold depending on the experimental conditions. In contrast, the expressions of SREC-1 (Adachiet al,1997) or FEEL-1 (Adachi & Tsujimoto, 2002) were decreased by C. pneumoniae infection (Fig. 3). A significant induction of LOX-1 transcription was observed as early as 24h after infection with C. pneumoniae (data not shown). In accordance with the previous reports, both SREC-1 and FEEL-1 were constitutively expressed on endothelial cells (Fig. 3). In contrast, LOX-1 is known to be inducible (Sawamuraet al,1997) with tumour necrosis factor α (TNF-α), phorbol 12-myristate 13-acetate (Kumeet al,1998) and angiotensin II (Liet al,1999; Morawietzet al,1999). In the present experimental system, however, the concentrations of TNF-α in the culture supernatant were below the detection limit (1pgmL−1) and the presence of anti-TNF-α antibody during infection did not reduce the LOX-1 induction (data not shown). Furthermore, the addition of an angiotensin II receptor inhibitor peptide was also ineffective at 50μM, which was far above the effective dose (Dillonet al,1998) (Fig. 5). Interestingly, the induction of LOX-1 was observed with the inoculation of formaldehyde-fixed C. pneumoniae, as well as with the infection of the live microorganism (Fig. 6).
Many studies have proved the importance of inflammatory reactions in the development of atherosclerotic lesions. The crucial roles played by inflammatory cytokines have been confirmed by the data from cytokine- or receptor-gene deficient mice. For example, CCR2 gene deficiency, which is the receptor for monocyte chemoattractant protein-1, reduced the atherosclerotic lesions in vivo (Boringet al,1998). Similar observations have been reported for other inflammatory cytokines, such as gamma interferon (IFN-γ) (Tellideset al,2000) and interleukin-18 (IL-18) (Elhageet al,2003), even in the absence of particular infectious agents. On the other hand, it is well documented that vascular infection of C. pneumoniae could initiate an inflammatory reaction, such as macrophage chemotaxis and activation (Molestinaet al,2000; Niessneret al,2003), which would result in atherogenesis through the uptake of degenerated LDL by macrophages. Our preliminary data also indicated the production of monocyte chemoattractant protein-1 by endothelial cells after C. pneumoniae infection (data not shown). In the scheme of inflammatory origin of atherosclerosis, the most important cell type responsible for degenerated LDL accumulation is the macrophage.
In the present study, we investigated whether there was a direct contribution of endothelial cells to atherogenesis after C. pneumoniae infection. Interestingly, in vitro infection with C. pneumoniae increased the uptake of oxidized LDL and acetylated LDL, but not of LDL by human vascular endothelial cells (Fig. 3). Accordingly, the expression of inducible lipoprotein receptor, LOX-1, was enhanced remarkably by C. pneumoniae infection, as indicated by semi-quantitative RT-PCR. The greater degree of uptake enhancement in oxidized LDL compared to acetylated LDL corresponds well with the report that LOX-1 prefers oxidized LDL to acetylated LDL (Kumeet al,1998). In contrast, SREC and FEEL-1, which are constitutively expressed on endothelial cells, were decreased by C. pneumoniae infection. This result corresponds with the fact that the major ligand of SREC and FEEL-1 is acetylated LDL (Adachiet al,1997; Adachi & Tsujimoto, 2002). The induction of LOX-1 is known to be dependent on TNF-α or angiotensin-II. In the present study, however, the apparent involvement of either mediator could not be confirmed. On the other hand, the fact that MOI 0.2 of C. pneumoniae infection caused a significant effect on LOX-1 expression and oxidized LDL uptake, as did an excess dose of MOI 2, might suggest an involvement of such intercellular signaling mediators. Our observation of LOX-1 induction by formaldehyde-fixed C. pneumoniae indicated that the live organism is not required, but that certain products of the microorganism might be responsible for the induction. Indeed, HUVEC are known to express toll-like receptor-2 and -4, and respond to bacterial LPS or lipoproteins (Faureet al,2001). Collectively, we demonstrated that human endothelial cells might contribute directly to atherogenesis through augmented expression of LOX-1 scavenger receptor, when infected with C. pneumoniae.
The authors are grateful to Dr Kazuko Takahashi and Ms Akiko Morikawa for technical assistance. This work was partly supported by Ministry of Education, Culture, Sports Science and Technology, Japan.