Chlamydia pneumoniae Activates Epithelial Cell Proliferation via NF-κB and the Glucocorticoid Receptor

Cell lines and infection with C. pneumoniae in cell culture.Two human epithelial cell lines were used in these experiments: HL and Calu3. Both cell lines were cultured in minimal essential medium supplemented with 10% fetal bovine serum, 8 mM stabilized l-glutamine, 2% minimal essential medium-vitamin mix, and 20 mM HEPES (Gibco BRL, Paisley, United Kingdom).

The cell lines were checked for Mycoplasma contamination using PCR as described earlier (26). C. pneumoniae (isolate K-6, obtained from the Department of Virology, University of Helsinki, Helsinki, Finland) was grown in cycloheximide-treated HL and Calu3 cells, purified through 30% Urografin (Schering, Berlin, Germany) centrifugation, and resuspended in cycloheximide-free medium. For storage, SPG (250 mM sucrose- 10 mM sodium phosphate- 5 mM l-glutamic acid)-diluted stock was kept frozen at −80°C. The purified C. pneumoniae stock was titrated in both the Calu3 and HL cell lines, and the titer was expressed as inclusion-forming units (IFU) per milliliter or multiplicities of infection (MOI). Purified C. pneumoniae stock was used in the corresponding cells for all experiments at a concentration of 5 × 10⁸ IFU/ml. Mock-infected controls were prepared in the same way as infected cells.

Cells were grown in six-well plates (10⁴ cells per well) and were infected with the C. pneumoniae stock (5 × 10⁸ IFU/ml) by centrifugation (1 h; 3,000 × g; 32°C). Mock-infected controls were prepared by the same procedure. To determine the growth of C. pneumoniae in infected cells, we used fluorescein-labeled monoclonal antibodies (Cellabs, Sydney, Australia).

Preparation of nuclear and cytosolic proteins.Nuclear and cytosolic proteins from infected or mock-infected cells were extracted from six-well plates (10⁴ cells per well) at 0, 0.5, 1, 3, 6, and 24 h postinfection. In brief, the cells were washed twice with 15 ml of ice-cold phosphate-buffered saline (PBS) and collected by centrifugation (5 min; 600 × g). The supernatant was discarded, and the cells were resuspended in 150 μl of low-salt buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM NaVO₄, 1 mM EDTA, 1 mM EGTA, 0.2% Nonidet P-40, 10% glycerol, 1× Boehringer Complete), vortexed (1 min), and incubated on ice (5 min). Following centrifugation (2 min; 13,000 × g), the supernatant was transferred into a new tube and kept as a cytosolic fraction at −70°C. The remaining pellet was resuspended in 60 μl of high-salt buffer B (20 mM HEPES, pH 7.9, 420 mM NaCl, 10 mM KCl, 0.1 mM NaVO₄, 1 mM EDTA, 1 mM EGTA, 20% glycerol, 1× Boehringer Complete), kept on ice (30 min), and vortexed every 5 min. Nuclear membrane fragments were removed by centrifugation (10 min; 13,000 × g; 4°C), and the supernatant was stored as a nuclear fraction at −70°C (7).

Western blotting for transcription factors.Samples of nuclear and cytosolic fractions containing 5 μg of protein (determined by Bradford assay) were size fractionated by denaturing sodium dodecyl sulfate-8% polyacrylamide gel electrophoresis. The proteins were transferred onto a polyvinylidene diflouoride membrane, and protein levels were confirmed by Ponceau staining (28). The membranes were blocked (Tris-buffered saline [TBS] plus 2% skim milk) for 30 min. Protein analysis by Western blotting was performed using antibodies against GR, CCAAT-enhancer binding protein α (C/EBP-α), C/EBP-β, NF-κB, and AP-1 (all from Santa Cruz Biotechnology, Santa Cruz, Calif.) (28). The membranes were incubated with one of the primary antibodies overnight at 4°C and subsequently washed three times with blocking solution before being incubated with peroxidase-labeled sheep anti-rabbit (Chemicon International, Temecula, Calif.) or rabbit anti-goat (Calbiochem, Darmstadt, Germany) secondary antibody for 1 h at room temperature. Protein bands were visualized using a chemiluminescence detection system (ECL-advanced; Amersham) (7).

DNA mobility shift assay for activated glucocorticoid receptor and C/EBPs.DNA mobility shift analysis for each of the above-mentioned transcription factors was performed using commercially available DNA consensus sequences (Santa Cruz Biotechnology) as described earlier (7). Equal amounts of either cytosolic or nuclear protein preparations were incubated with 10 fM radioactively labeled DNA consensus sequences: GR, 5′-AGA GGA TCT GTA CAG GAT GTT CTA GAT-3′; CCAAT, 5′-TGC AGA TTG CGC AAT CTG CA-3′; NF-κB, 5′-AGT TGA GGG GAC TTT CCC AGG C-3′ (all from Santa Cruz Biotechnology). The mixtures were incubated for 30 min at room temperature and were then loaded onto a nondenaturing 10% polyacrylamide gel and size fractionated by gel electrophoresis for 2 h (50 mA; 4°C). The gels were dried for 1 h, and the bands of DNA-protein complexes were visualized using a phosphorimager system.

Unlabeled DNA consensus fragments were used in 10-fold excess to specify protein-DNA bands, and specific antibodies were used for supershift characterization of DNA-protein bands as described earlier (7).

Luciferase reporter gene constructs.Two different human luciferase reporter gene constructs were used to determine the functionality of the C. pneumoniae-induced transcription factors: (i) interleukin-6 (IL-6) promoter and (ii) p21^(Waf1/Cip1) promoter. Subconfluent cell layers (10⁴ cells per well) were transiently transfected with one of the two reporter gene constructs, and gene activation was assessed 24 h postinfection as described earlier (7, 28).

Antisense oligonucleotides.To down regulate transcription factors, the cells were incubated with thioated antisense oligonucleotides (1 μMol) for 24 h prior to infection with C. pneumoniae, and 0.5 μmol of antisense oligonucleotides was added to the cultures every 24 h for the duration of the experiments. The oligonucleotide sequences were as follows: C/EBP-α, 5′-GAA GGC CGC GGC GCT GCT GGG CGC GT-3′; scrambled C/EBP-α, 5′-AGC TCG GAT GCA TGG AGG AG-3′; C/EBP-β, 5′-GTG GTC GGG GGT GTG ATT AT-3′; scrambled C/EBP-β, 5′-GTG GGG GTC GTG GGT TTA AT-3′; NF-κB, 5′-GAA ACA GAT CGT CCA GGT-3′; and scrambled NF-κB probe, 5′-GAA GAT ACA CCA CGT GGT-3′ (MWG, Ebersberg, Germany).

DNA synthesis.Cells (10⁴ per well) were seeded in a 24-well tissue culture plate, grown overnight, and serum deprived (0.1% serum) for 24 h. The cells were infected with C. pneumoniae or with heat-inactivated (65°C for 60 min) C. pneumoniae elementary bodies at MOI of 2,500, 1,250, 625, and 312 per well and were cultured for 24 or 48 h. [³H]thymidine (1 μCi/ml) was added for the final 5 h of culture. The cells were washed twice with ice-cold PBS and fixed with methanol-acetic acid (3:1) for 5 min. After an additional wash with PBS, the cells were lysed in 100 μl of 0.5 M NaOH, and the incorporated [³H]thymidine was measured after the addition of scintillation liquid (7). [³H]thymidine incorporation data were confirmed by counting infected and uninfected cells 3 days postinfection in a Neubauer chamber.

Immunohistochemistry.HL cells were grown overnight to confluency of 10⁴ per well in 24-well plates and were infected with C. pneumoniae at an MOI of 2,500 per well. Mock-infected cells were used as a control. At appropriate times, the cells were fixed by adding 1 volume of 3:1 methanol-glacial acetic acid to 1 volume of growth medium. After incubation at room temperature (5 min), the cells were washed with the fixative and fixed for an additional 10 min, rinsed with TBS, and blocked for 30 min with 1% bovine serum albumin in TBS. The cells were stained with a mixture of 1:500-diluted anti-Chlamydia monoclonal antibody (Research Diagnostics Inc., Flanders, N.J.) and 1:400-diluted anti-GR rabbit polyclonal antibody (Santa Cruz Biotechnology Inc.) or a mixture of anti-Chlamydia antibody and 1:400-diluted anti-NF-κB p65 rabbit polyclonal antibody (Santa Cruz Biotechnology Inc.). Endogenous peroxidase activity was removed with 0.1% NaN₃ and 0.3% H₂O₂ (30 min). Samples were incubated with primary antibodies at +4°C overnight, washed, and stained using a StreptABComplex/HRP Duet Mouse/Rabbit kit (Dako, Glostrup, Denmark) and then developed with a DAB substrate kit for peroxidase (Vector Laboratories, Burlingame, Calif.). All wells were examined using an Olympus microscope, and the cells with black nuclear (GR and NF-κB) or cytoplasmic (C. pneumoniae) staining were accepted as positive.

The Chlamydia infection rate in HL cells was assessed by staining with anti-Chlamydia monoclonal antibody (Research Diagnostics Inc.) and alkaline phosphatase-conjugated goat anti-mouse antibody (Zymed Laboratories, San Francisco, Calif.), and the red color was developed with a DAKO-APAAP kit for alkaline phosphatase conjugates (Dako). The cells were counterstained with hematoxylin.

Effect of dexamethasone treatment on C. pneumoniae infection rate.HL cells (5 × 10⁴ per well) were grown overnight in 24-well plates, infected with C. pneumoniae at an MOI of 625 (3 × 10⁷ IFU) per well, and grown for 72 h in antibiotic- and cycloheximide-free medium. The cells were pretreated with 10⁻⁸, 10⁻¹⁰, and 10⁻¹² M dexamethasone (Calbiochem) for 30 min before infection. C. pneumoniae was prepared in 10% FCS RPMI medium containing 10⁻⁸, 10⁻¹⁰, or 10⁻¹² M dexamethasone. Infected cells were grown for 72 h in cycloheximide under antibiotic-free conditions and stained using an alkaline phosphatase-based detection method as previously described.

Statistics.The null hypothesis was that C. pneumoniae does not affect the growth of epithelial cells or the activation of human transcription factors. We tested the null hypothesis by analysis of variance (ANOVA) and Student's paired t test.

C. pneumoniae infection in vitro.Infection of a human lung epithelial cell line (HL) showed fast attachment (within 30 min) of the microorganism to the cells (Fig. 1A, I). Alkaline phosphatase-based immunohistochemistry staining for C. pneumoniae infection in vitro showed a constant increase of infected cells within the first 6 h, and after 3 h, C. pneumoniae had clearly invaded the host cell cytoplasm (Fig. 1A, III and IV).

• Open in new tab
• Download powerpoint

• Open in new tab
• Download powerpoint

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Kinetics of C. pneumoniae infection in human epithelial cells (HL). C. pneumoniae was stained with a red substrate, and the cells were counterstained with hematoxylin. (B) GR location and C. pneumoniae infection by double immunohistochemical staining. Shown is the time course of C. pneumoniae-induced translocation and activation of the GR in infected cells (I to IV) and in uninfected control cells (V to VIII). Inactive GR is visible as a black-stained region around the nuclear membrane, and active GR is indicated by dark-grey to black nuclei (arrows). Some of the immunohistochemically stained C. pneumoniae in cells are circled in image I. (C) Representative Western blot assay of C. pneumoniae (C. pneu)-induced translocation of GR in HL cells. Similar results were obtained in Calu3 cells, and each experiment was repeated three times. (D) Functional activation of GR by C. pneumoniae is depicted as a representative EMSA over a period of 6 h. GR binding was monitored by a double-stranded GRE DNA consensus sequence.

C. pneumoniae activates transcription factors.We tested the hypothesis that infection of human lung epithelial cells with C. pneumoniae modulates the activity of the host transcription factors that are related to proliferation in the host cells.

As shown in Fig. 1B, infection of HL cells with C. pneumoniae resulted in rapid activation and translocation of the GR into the nucleus, which started at 30 min and became maximal after 3 h (Fig. 1B, I to III). In uninfected cells, the GR was located in a zone around the nucleus (Fig. 1B, V to VIII).

Using Western blot analysis, we confirmed that infection of HL cells and Calu3 cells with C. pneumoniae caused translocation of the GR from the cytosol into the nucleus, which was nearly complete at 3 h (Fig. 1C). Functional activation of the GR by C. pneumoniae infection was shown by using a DNA electrophoretic mobility shift assay (EMSA) demonstrating that the GR recognized and bound to its specific DNA consensus sequence. Figure 1D shows that C. pneumoniae induced binding of a protein to the glucocorticoid response element (GRE) within 3 h after infection, further increasing until 6 h in both cell lines. The specificity of the GR signal in EMSA was confirmed using HeLa cell extract (Fig. 1D, second lane from left) and a 10-fold excess of unlabeled GRE (fourth lane from left). We tested for nonspecific DNA interactions using Oct-1 (fifth lane from left) oligonucleotides as a negative control.

Infection with C. pneumoniae also activated C/EBP-α and C/EBP-β, two members of the C/EBPs (Fig. 2A). The activation of C/EBP-α by C. pneumoniae infection was not very strong, since we could not observe a significant decrease in the C/EBP-α signal in the cytosol and observed only a marginal increase of the signal in the nuclear protein samples (Fig. 2A). In contrast, the activation of C/EBP-β was significant within 1 h after infection with C. pneumoniae (Fig. 2A). Using a C/EBP binding DNA consensus sequence, we observed a protein-DNA complex 3 h after infection in HL cells; the signal reached maximal level at 6 h and was not detectable after 24 h (Fig. 2B). In Calu3 cells, the C/EBP signal appeared 1 h after C. pneumoniae infection, with maximal activity at 3 h and declining thereafter (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Representative Western blot of C. pneumoniae-induced activation of C/EBP-α and C/EBP-β in HL cells. Similar results were obtained in triplicate experiments and in Calu3 cells. (B) Representative EMSA of C. pneumoniae (C. pneu)-activated CCAAT oligonucleotide binding nuclear proteins in HL and Calu3 cells. Similar results were obtained in triplicate experiments. The specificity of C/EBP proteins was demonstrated by 10-fold cold CCAAT oligonucleotide (oligo) and anti-C/EBP-α or C/EBP-β antibodies.

In addition, the translocation of NF-κB was induced by C. pneumoniae infection in the two epithelial cell lines. As shown by immunohistochemistry (Fig. 3A), the activation of NF-κB by C. pneumoniae did not occur before 1 h after infection, and maximal activation was observed after 6 h (Fig. 3A, I to IV). In uninfected cells, NF-κB was located in the cytosol close to the nuclear membrane. The activation of NF-κB by C. pneumoniae was confirmed by Western blotting, which revealed a time course similar to that shown by immunohistochemistry (Fig. 3B). NF-κB signal decreased in the cytosol, increased after 1 h in the nucleus, and was maximal at 3 h (Fig. 3B). Furthermore, the functional activation of NF-κB was demonstrated by EMSA, which also showed a slow increase in DNA binding of a protein to the NF-κB oligonucleotide (Fig. 3C). We observed a slight difference when the binding kinetics of NF-κB in the two cell lines were compared. While significant binding was observed in HL cells at 3 h, which increased until 6 h, in Calu 3 cells NF-κB binding occurred at 30 min, peaked at 3 h, and then declined (Fig. 3C).

• Open in new tab
• Download powerpoint

• Open in new tab
• Download powerpoint

FIG. 3.

(A) Cell compartment localization of NF-κB in C. pneumoniae-infected HL cells (I to IV) and in uninfected HL cells (V to VIII) by double immunohistochemical staining. Inactive NF-κB is visible as a black-stained region around the nuclear membrane, and active NF-κB is indicated by dark-grey to black nuclei (arrows). Some of the immunohistochemically stained C. pneumoniae cells are circled in image I. (B) Representative Western blot assay of C. pneumoniae-induced translocation of NF-κB in HL cells. Similar results were obtained in Calu3 cells, and each experiment was repeated three times. (C) Functional activation of NF-κB by C. pneumoniae (C. pneu) depicted as a representative EMSA over a period of 6 h. Oligo, oligonucleotide.

The DNA binding proteins were further characterized by an EMSA either in the presence of the GR inhibitor RU486 or after pretreatment (24 h) of the cells with antisense oligonucleotides (1 μM) targeting C/EBP-α, C/EBP-β, or NF-κB. As shown in Fig. 4, the signal for the GR-GRE complex was significantly reduced when the cells were preincubated for 2 h with the GR inhibitor RU486 (10⁻⁶ M) prior to infection with C. pneumoniae (Fig. 4, second lane from left). In the presence of an antisense oligonucleotide to NF-κB, the C. pneumoniae-dependent NF-κB-DNA complex was abolished in supershift EMSA (Fig. 4, fourth lane from left). In the presence of an antisense oligonucleotide specific for C/EBP-α, the C/EBP band was slightly reduced (Fig. 4, seventh lane from left), and in the presence of an antisense oligonucleotide specific for C/EBP-β, the signal of the protein-DNA complex was clearly reduced (Fig. 4, right lane).

• Open in new tab
• Download powerpoint

FIG. 4.

Characterization of transcription factors activated by C. pneumoniae infection in HL cells shown as a representative EMSA supershift assay. Nuclear extracts from HL cells infected with C. pneumoniae for 3 h were used for the analysis. To characterize GR, cells were pretreated with RU486 (10⁻⁶ M) for 2 h prior to infection (two left lanes). For NF-κB characterization, we incubated the nuclear extract that was isolated 3 h after C. pneumoniae infection overnight in the presence or absence of an NF-κB binding DNA oligonucleotide (lanes 3 and 4 from left). Lane 5 from the left is empty, and C/EBP-α and -β were characterized using CCAAT DNA sequence without (lane 6) or with a specific anti-C/EBP-α or anti-C/EBP-β antibody (two right lanes). Similar results were obtained in Calu3 cells.

C. pneumoniae activates a glucocorticoid-inducible promoter.To further assess the functions of transcription factors activated by C. pneumoniae, we transfected HL and Calu3 cells either with a luciferase-coupled human IL-6 promoter or with the human GR-dependent p21^(Waf/Cip1) promoter.

The human IL-6 promoter was not activated by C. pneumoniae in any cell line within 24 h. In contrast, the GR-inducible p21^(Waf1/Cip1) promoter was activated in a dose-dependent manner 24 h after C. pneumoniae infection (Fig. 5A). In HL cells, we observed a significant increase in luciferase activity with C. pneumoniae at an MOI of 1,250 per well (ANOVA, P ≤ 0.05) (Fig. 5A). In Calu3 cells, C. pneumoniae infection activated the p21^(Waf1/Cip1) promoter at an MOI of 625 (ANOVA, P ≤ 0.05) (Fig. 5A).

• Open in new tab
• Download powerpoint

FIG. 5.

(A) C. pneumoniae stimulates the human promoter for p21^(Waf1/Cip1) in epithelial cells in a dose-dependent manner. Each bar represents the mean + standard error of three independent experiments. *, P ≤ 0.05 (ANOVA); **, P ≤ 0.01 (ANOVA). FBS, fetal bovine serum. The numbers below the bars are MOI. (B) Involvement of transcription factors in C. pneumoniae-induced p21^(Waf1/Cip1) promoter activity. Each bar represents the mean + standard error of three independent experiments. **, P ≤ 0.01 (ANOVA). +, present; −, absent; scramb, scrambled.

Figure 5B shows that infection with C. pneumoniae induced the activity of the p21^(Waf1/Cip1) promoter luciferase construct in both cell lines. The GR antagonist RU486 had no effect on luciferase activity in uninfected cells and reduced the C. pneumoniae-induced luciferase activity by 56.1% ± 5.2% in HL cells (ANOVA, P ≤ 0.01) and by 63.8% ± 5.2% in Calu3 cells (ANOVA, P ≤ 0.05). When the cells were pretreated with antisense oligonucleotides targeting NF-κB p65, we observed no inhibition of C. pneumoniae-induced luciferase activity. Similarly, antisense oligonucleotides inhibiting C/EBP-β did not affect C. pneumoniae-inducible luciferase activity. The control oligonucleotides for either NF-κB or C/EBP-β did not inhibit the activation of the p21^(Waf1/Cip1) promoter luciferase construct by C. pneumoniae (Fig. 5B).

We also assessed the transcription factor AP-1, but we did not observe its activation by C. pneumoniae infection under any culture conditions (data not shown).

C. pneumoniae increases epithelial cell proliferation via GR and NF-κB activation.Infection with C. pneumoniae elementary bodies stimulated the proliferation of Calu3 and HL epithelial cells compared to uninfected cells in 0.1% fetal bovine serum (Fig. 6A). The proliferative effect of C. pneumoniae was determined by both thymidine incorporation and direct cell counting. Both methods showed a significant increase of proliferation in the presence of C. pneumoniae at an MOI of 1,250 (ANOVA, P ≤ 0.01) (Fig. 6A).

• Open in new tab
• Download powerpoint

FIG. 6.

(A) Effect of diluted C. pneumoniae stock on epithelial cell proliferation. Proliferation was determined by direct cell counting after 3 days of culture of infected and uninfected cells. Each bar represents the mean + standard error of three independent experiments. **, P ≤ 0.01 (ANOVA). FBS, fetal bovine serum. The numbers below the bars are MOI. (B) Involvement of transcription factors in C. pneumoniae-dependent epithelial cell proliferation determined by [³H]thymidine incorporation after 24 h of culture. Each bar represents the mean + standard error of three independent experiments. *, P ≤ 0.05 (ANOVA); **, P ≤ 0.01 (ANOVA). +, present; −, absent; scramb, scrambled.

To assess whether one or more of the above-described host transcription factors is involved in C. pneumoniae-induced cell proliferation, we pretreated the cells for 2 h with RU486 (10⁻⁶ M). As shown in Fig. 6B, Ru486 had no significant effect on uninfected cells but reduced the proliferation of C. pneumoniae-infected HL cells by 70.1% ± 8.2% (ANOVA, P ≤ 0.01) and that of Calu3 cells by 62.5% ± 8.2% (ANOVA, P ≤ 0.01).

When cells were treated for 24 h with NF-κB antisense oligonucleotides (10⁻⁶ M) prior to infection with C. pneumoniae, proliferation was reduced by 19.2% ± 2.5% in HL cells (ANOVA, P ≤ 0.05) and by 34.9% ± 5.0% in Calu3 cells (ANOVA, P ≤ 0.01) (Fig. 5.B). In contrast, C/EBP-β antisense oligonucleotides (10⁻⁶ M) did not alter the proliferation of C. pneumoniae-infected epithelial cells (Fig. 5B). The control oligonucleotides for either NF-κB or C/EBP-β did not inhibit the proproliferative effect of C. pneumoniae (Fig. 5B).

To prove that live C. pneumoniae is essential for the observed proliferation of the host cell, we incubated HL cells with heat-inactivated and undiluted C. pneumoniae at 5 × 10⁸ IFU/well. While C. pneumoniae at an MOI of 2,500 induced a 2.61-fold increase in cell proliferation, heat-inactivated C. pneumoniae and undiluted C. pneumoniae stock caused only a 1.47-fold increase.

Effect of glucocorticoids on C. pneumoniae infection.In addition to the modulation of the host transcription factors by C. pneumoniae, we investigated whether the GR may also affect the intracellular growth of C. pneumoniae. To test this, we infected HL cells (5 × 10⁴ per well) with C. pneumoniae at an MOI of 625 in the presence and absence of various concentrations of dexamethasone (10⁻¹², 10⁻¹⁰, and 10⁻⁸ M). The infected cells were further cultured for 72 h and then stained for C. pneumoniae as shown in Fig. 7. Dexamethasone at concentrations of 10⁻¹² or 10⁻¹⁰ M did not increase staining for C. pneumoniae inclusion bodies, while 10⁻⁸ M dexamethasone significantly enhanced C. pneumoniae staining in infected cells (Fig. 7, I to IV).

• Open in new tab
• Download powerpoint

FIG. 7.

Dose-dependent support of dexamethasone for C. pneumoniae propagation in infected in HL cells. C. pneumoniae was stained with a red substrate, and immunohistochemical analysis was performed 72 h after infection with C. pneumoniae at an MOI of 625.