Pattern Recognition Molecules Activated by Chlamydia muridarum Infection of Cloned Murine Oviduct Epithelial Cell Lines¹

Urogenital serovars of Chlamydia trachomatis cause ∼3 million sexually transmitted genital tract infections in the United States each year (1), with similar rates of infection in European countries (Ref2 , 〈www.eurosurveillance.org/ew/2004/041007.asp#5.〉). Infections by the urogenital Chlamydia serovars cause urethritis, cervicitis, epididymitis, pelvic inflammatory disease, neonatal conjunctivitis, and neonatal pneumonia. In women, progression of C. trachomatis infection into the upper reproductive tract causes significant inflammation and injury to the Fallopian tubes (3). Chlamydia-induced scarring of the Fallopian tubes is the cause of ∼50% of tubal factor infertility and 25% of ectopic pregnancies in developed countries (4, 5).

The pathophysiology underlying Chlamydia-induced Fallopian tube scarring is unclear. In mice, experimental Chlamydia muridarum genital tract infections are spontaneously cleared by T lymphocytes in a Th1-dependent fashion (6). In humans, the adaptive immune response to C. trachomatis infection is also Th1 mediated (7). It has generally been accepted that immunopathologic T cell responses to the infection cause Fallopian tube scarring (8). However, recent data in mice (9) and humans (10, 11, 12, 13) reveal that Chlamydia infections of the genital tract can persist for long periods of time. Because Chlamydia-infected epithelial cells secrete proinflammatory cytokines (14, 15, 16, 17, 18, 19), it is possible that infected epithelial cells play a dominant role in the pathophysiology of scarring. The hypothesis that scarring of Fallopian tubes is driven by infected epithelial cells, rather than T lymphocytes responding to the infection, has been referred to as the cellular paradigm of Chlamydia pathophysiology (20). Consistent with that paradigm, we recently reported that a cloned oviduct epithelial cell line responded to infection by C. muridarum with a plethora of cytokines (14), including acute phase cytokines IL-1, IL-6, TNF-α, GM-CSF, and TGF-α. These cytokines have been shown to contribute to the pathophysiology of scarring and fibrosis in other systems (21, 22, 23, 24, 25, 26, 27).

Mammalian cells sense the presence of invading microbial pathogens through recognition of pathogen-associated molecular patterns (PAMPs)³ present in microbial structural subunits including microbial cell wall (e.g., peptidoglycan), cell membrane (e.g., LPS), and virulence proteins (e.g., flagellin) (28). TLRs are membrane-bound receptors that bind PAMPs and trigger host defense responses, including cytokine secretion. At least 11 TLRs have been identified in mammals. Humans have TLR1–10; mice have TLR1–9 and 11 (28, 29). In addition to TLRs, mammalian cells have cytosolic surveillance for PAMPs via the nucleotide binding site/leucine-rich repeat (NBS/LRR) proteins, Nod1 and Nod2. Nod1 is activated by a peptidoglycan motif found in Gram-negative bacteria (30, 31), while Nod2 binds muramyl dipeptides (32, 33).

Previous work in humans has shown that primary and immortalized epithelial cells derived from the ectocervix and endocervix express TLRs 1–6, but lack TLR4 (34), while a transformed human uterine epithelial cell line (35) and primary uterine epithelial cells (36) express TLRs 1–9, including TLR4. RT-PCR and Western blot analysis done on homogenates of Fallopian tube tissue show the presence of TLR2 and TLR4 in the upper reproductive tract (37). Similarly, homogenates of murine oviducts have detectable TLR4 mRNA by RT-PCR (38). However, analysis of TLR expression in Fallopian tube and oviduct homogenates includes all cell lineages (epithelium, endothelium, fibroblasts, myeloid, and lymphoid) resident within the organ. To date, the TLR repertoire expressed on upper reproductive tract epithelial cells has not been determined.

Activation of upper reproductive tract epithelial cells via pattern recognition molecules (PRMs) likely plays an important role in the pathophysiology caused by Chlamydia infection. The PRMs used by reproductive tract epithelium to “sense” and respond to invasion by C. muridarum are not known. However, previous work using TLR2- and TLR4-knockout mice has identified TLR2 as a major contributor to infection-induced cytokine secretion and oviduct pathology, without contributing to the clearance of Chlamydia infection from the genital tract (38). In this report, we investigate the PRMs expressed by murine oviduct epithelial cell lines and their roles in C. muridarum-induced secretion of epithelial cytokines.

The following TLR agonists were purchased from InvivoGen: 1) ultrapure LPS of Escherichia coli serotype 0111:B4 (1 × 10⁶ endotoxin units (EU)/mg); 2) Peptidoglycan from E. coli serotype 0111:B4 (125 EU/mg); 3) the synthetic tripalmitoylated lipoprotein analog Pam₃CSK4 (<0.125 EU/mg); 4) flagellin purified from Salmonella typhimurium (125 EU/mg); 5) ODN1826, an oligonucleotide (ODN) containing murine CpGs (<0.125 EU/mg); and 6) ODN1826 control (ODN control), a control ODN without CpGs (<0.125 EU/mg). Poly(I:C), product number P-0913, was purchased from Sigma-Aldrich.

Female B6.C-H2^bm12/KhEg and B6.C-H2^bm1/ByJ mice were purchased from The Jackson Laboratory and housed in Indiana University-Purdue University specific pathogen-free facilities. The Institutional Animal Care and Utilization Committee approved all experimental protocols.

RAW264.7 cells (American Type Culture Collection) were cultured at 37°C in a 5% CO₂ humidified incubator and maintained in DMEM supplemented with 10% (v/v) Fetalclone III (HyClone). The cloned oviduct epithelial cell line Bm1.11 has been described previously (14), and a new cloned epithelial cell line described in this report, designated Bm12.4, was derived using a similar methodology. Reproductive tract tissues from a B6.C-H2^bm12/KhEg female mouse encompassing the proximal half of the uterine horn to the oviduct tissue immediately adjacent the ovary were harvested. Lumenal epithelial cells were released with PHC mixture as previously described for the Bm1.11 cell line (14). The resulting epithelial cell-enriched cells were expanded in vitro and cloned by limiting dilution after passage through a 0.40-μM filter to remove cell clumps. Resulting clones were screened for IFN-γ-inducible MHC class II (feature of murine epithelial cells but not generally murine fibroblasts). A selected clone with inducible MHC class II expression and epithelial morphology was limiting-diluted a second time to ensure clonality and the resulting clone designated Bm12.4. The epithelial lineage of Bm12.4 was confirmed by immunohistochemical staining for cytokeratin intermediate filaments. The upper reproductive tract epithelial cells were grown at 37°C in a 5% CO₂ humidified incubator and maintained in epithelial-cell medium: 1:1 DMEM:F12K (Sigma-Aldrich), supplemented with 10% characterized FBS (HyClone), 2 mM l-alanyl-l-glutamine (Glutamax I; Invitrogen Life Technologies), 5 μg of bovine insulin/ml, and 12.5 ng/ml recombinant human fibroblast growth factor-7 (keratinocyte growth factor; Sigma-Aldrich).

The dominant-negative (DN) MyD88 expression plasmid pIRES2-EGFP-DN-MyD88 (39) was generously provided by Dr. S.-C. Hong (Indiana University School of Medicine, Indianapolis, IN).

Mycoplasma-free C. muridarum, previously known as C. trachomatis strain MoPn, was grown in McCoy cells (American Type Culture Collection). The titers of Mycoplasma-free C. muridarum stocks were determined on McCoy cells with centrifugation as described previously (14, 40).

Bm12.4 cells grown on glass chamber slides (Labtek) were fixed for 2 min with ice-cold methanol, blocked with 5% normal goat serum/PBS for 45 min, then stained with mAb mixture AE1/AE3 specific for acidic and basic cytokeratins (Cappel/ICN) or 36-7-5, a control mAb specific for H-2K^k (epitope not present on Bm12.4 cells) (BD Pharmingen). Detection was accomplished with a FITC-coupled goat anti-mouse IgG F(ab′)₂ affinity-purified antiserum (Cappel/ICN).

For flow cytometry, Bm12.4 cells were dislodged from tissue culture plastic using a Hank’s salt-based, enzyme-free cell dissociation buffer (Sigma-Aldrich). Cells were stained for 20 min on ice in PBS/2%BSA with PE-coupled M5/114.15.2 (I-A^b) (eBioscience). IFN-γ induction was accomplished by addition of 10 ηg/ml murine rIFN-γ (Sigma-Aldrich) to the culture medium for 14 h before staining. Cells were analyzed with a FACSCalibur cytometer (BD Biosciences).

Bm1.11, Bm12.4, and RAW 264.7 cells were plated in 48-well tissue culture plates and used when confluent. For all experiments, the cells were infected with 10 inclusion-forming-units (IFU) of C. muridarum/cell in 900 μl of culture medium. The 48-well plates were centrifuged at 1000 rpm (300 × g) in a table-top centrifuge for 30 min then incubated at 37°C in a 5% CO₂ humidified incubator without subsequent change of medium for 24–30 h, depending upon the assay. Mock-infected wells received an equivalent volume of sucrose-phosphate-glutamic acid buffer lacking C. muridarum.

To generate the DN MyD88-expressing cell line designated DN-MyD88, Bm1.11 cells were transfected with 5 μg of the pIRES2-EGFP-DN-MyD88 plasmid in 75% confluent 6-well plates using LipofectAMINE 2000 reagent (Invitrogen Life Technologies). After 48 h, the cells were removed from the plate using enzyme-free cell dissociation buffer (Sigma-Aldrich) and passed through 40-μM nylon cell mesh to remove clumps of cells. The cells were suspended in cell-sorting buffer (PBS + 2% FCS) for screening GFP expression (incorporated in the plasmid), using the FACSVantage SE cell sorter (BD Bioscience). Clones were obtained by limiting dilution in epithelial-cell medium supplemented with 800 μg/ml G418. Selected clones were screened for expression of the DN MyD88 protein by Western blot analyses. The DN-MyD88 cell line was maintained in epithelial-cell medium supplemented with 400 μg/ml G418.

Early passage (<32) Bm1.11 and Bm12.4 cells were set up in RPMI 1640 medium lacking phenol red and supplemented with 10% characterized FBS (HyClone), 2 mM l-alanyl-l-glutamine (Glutamax I; Invitrogen Life Technologies), 5 μg of bovine insulin/ml, 12.5 ηg/ml recombinant human fibroblast growth factor-7 (keratinocyte growth factor; Sigma-Aldrich) and 10⁻⁸ M 17-β-estradiol (Sigma-Aldrich) for 48 h before isolating total RNA for RT-PCR analysis.

Control and DN-MyD88 Bm1.11 cells were grown in monolayers in a 6-well plate to confluence. After removal of the growth medium, the monolayers were gently washed with PBS, and cytosolic proteins were recovered in the cell fractionation buffer provided in the PARIS kit (Ambion). Cytosolic proteins were quantified using the Micro BCA Protein Assay kit (Pierce). Twenty-five micrograms of lysate from either the DN-MyD88 or nontransfected Bm1.11 cell line were boiled in 5× Immunopure Reducing Sample Buffer (Pierce) before SDS PAGE. After separation by 10% SDS-PAGE, the proteins were transferred to Immobilon-P (Millipore) transfer membranes. Transfer membranes were blocked in 5% nonfat dry milk and subsequent immunoblotting was performed using a 1/5000 dilution of the MyD88-specific polyclonal Ab HFL-296 (Santa Cruz Biotechnology) in TBST buffer (20 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 0.05% Tween 20). The primary Ab incubation, performed for 1 h at room temperature, was followed by three TBST washes. Membranes were then incubated with 1/10,000 HRP-conjugated goat anti-rabbit polyclonal Ab (Amersham Biosciences) for 1 h, followed by three washes in TBST. Proteins were visualized via chemiluminescence using the ECL plus Western blotting detection system (Amersham Biosciences) as described in the manufacturer’s protocol.

MyD88-specific small interfering (si)RNA, MyD88-specific primers, scrambled control siRNA, siRNA transfection reagent, and siRNA transfection medium were purchased from Santa Cruz Biotechnology. siRNA transfections targeting endogenous MyD88 and scrambled controls were conducted in 48-well plates seeded with 4 × 10⁴ Bm1.11 cells/well using the manufacturer’s protocol. The siRNA mixture was mixed with the transfection reagent mix and allowed to form liposome-siRNA complexes for 30 min. After initial 5 h of incubation, the transfection medium was replaced with fresh epithelial medium, and the cells were returned to the 37°C CO₂ humidified incubator for 30 h before using them in experiments.

Cytoplasmic RNA was purified from the siRNA transfected Bm1.11 cells using the PARIS kit (Ambion). The RNA was quantified by spectrophotometric analysis, and RNA integrity was confirmed by agarose gel electrophoresis. cDNA synthesis was performed using 1 μg of the cytoplasmic RNA with the iScript cDNA synthesis kit (Bio-Rad). The cDNA product was diluted 1/50, while the MyD88-specific and β-actin control primers (Table I) were adjusted to 10 pmol/μl working stock. Real-time PCR was conducted with the diluted cDNA and primers as per the protocol outlined in the iTaq SYBR Green Supermix with ROX kit (Bio-Rad). Real-time PCR was performed with an ABI Prism 7700 machine (Applied Biosystems): 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, and 2 min at 60°C. Cycle threshold (C_T) values were determined by automated threshold analysis with ABI Prism version 1.0 software. The amplification efficiencies were determined by serial dilution and calculated as E = exp^−/m, where E is the amplification efficiency and m is the slope of the dilution curve (41). Dissociation curves were recorded after each run to ensure primer specificity.

Total RNA was isolated from oviduct epithelial cell lines using RNeasy mini columns (Qiagen). Total RNA enriched in luminal epithelial cell RNA was isolated from the upper reproductive tract and terminal ileum of a female B6.C-H2^bm1/ByJ mouse. A segment of the upper reproductive tract extending from the mid-uterine horn to immediately adjacent to the ovary was resected, as well as a segment of terminal ileum of equivalent length. The lumen of the terminal ileum was cannulated and gently flushed with 2 cc of PBS. Total RNA was extracted from the lumens of the upper reproductive tract and terminal ileum by cannulating each with a 28-gauge needle attached to a 0.5-cc syringe and running 0.5 ml of cell lysing solution containing guanidine thiocyanate (Qiagen RLT buffer) through the lumen of each organ to preferentially lyse the luminal epithelial cells while preserving the organs general architecture. Total RNA within the luminal flow-through was isolated using RNeasy mini columns (Qiagen). During purification, all RNA samples were treated with RNase-free DNase I (Qiagen) to remove genomic DNA contamination. The RNA was quantified by spectrophotometric analysis. For TLR analyses, RNA integrity was confirmed by agarose gel electrophoresis. Optimized primer pairs were designed using Vector NTI Suite (Infomax). Specific primer pairs (Amitof) are listed in Table I. Using 1 μg of total RNA as the template for each reaction, RT-PCR was accomplished by using a single-tube avian myeloblastosis virus RT-Tfl polymerase kit (Access RT-PCR; Promega). Cycling conditions were as follows: 90 s of initial denaturation at 95°C, followed by eight cycles, each of which consisted of 30 s at 95°C, 15 s at 56°C, and 30 s at 72°C. After the initial eight cycles, the PCR continued for 32 additional cycles of 30 s at 95°C, 15 s at 56°C, and 30 s + 3 s/cycle at 72°C, followed by a final extension at 72°C for 3 min. Reactions were also amplified in the absence of reverse transcriptase as negative controls. The semiquantitative RT-PCR technique used to generate qualitative data from the siRNA experiments used a two-step PCR event using a proprietary nested-primer set for MyD88 and the β-actin control per the manufacturer’s protocol (Santa Cruz Biotechnology).

RAW264.7 cells and confluent Bm1.11 and Bm12.4 monolayers grown in 48-well tissue culture-treated plates were either infected with 10 IFU of C. muridarum/cell or treated with the appropriate TLR ligand at the concentrations specified in the text. Supernatants were harvested at either the 24- or 36-h time points and analyzed for cytokine content using ELISAs for IL-6 and GM-CSF according to the manufacturer’s protocol. The anti-mouse IL-6 mAb 32C11, biotin-labeled anti-mouse IL-6 Ab 20F3, recombinant murine IL-6, and recombinant murine GM-CSF were purchased from Endogen. The anti-mouse GM-CSF mAb MP1-22E9, biotin-labeled anti-mouse GM-CSF mAb MP1-31G6, and the streptavidin-HRP conjugate were purchased from BD Pharmingen. For the IFN-β ELISA, the rabbit polyclonal anti-IFN-β capture Ab (R&D Systems) was used at a 1/250 dilution in bicarbonate buffer; the anti-IFN-β detection rat monoclonal 8.S.415 (U.S. Biological) was used at 1 μg/ml; the rabbit anti-rat HRP conjugate (Jackson ImmunoResearch Laboratories) was used at 1/10,000. All standards and experimental samples were done in triplicate. The lower range of assay sensitivity for individual cytokines was 50 pg/ml for IL-6 and 10 pg/ml for GM-CSF and IFN-β. All experiments were repeated at least three times, and significance was determined using Student’s t test.

We previously published the derivation of a cloned oviduct epithelial cell line without viral or chemical transformation, designated Bm1.11, from a B6.C-H2^bm1/ByJ mouse (14). We have derived a second cloned oviduct epithelial cell line without viral or chemical transformation, designated Bm12.4, from a B6.C-H2^bm12/KhEg mouse using a similar methodology. Characteristic of an epithelial lineage, Bm12.4 cells express cytokeratins (42) and have IFN-γ-inducible MHC class II expression (Fig. 1).

To confirm their anatomical site of origin, RT-PCR analysis for oviduct-specific mRNAs helix factor hepatocyte NF/forkhead homologue-4 (hfh-4) and mammalian oviduct-specific glycoprotein (mogp-1) were performed on Bm1.11 and Bm12.4 cell lines ∼30 passages after clonal derivation. Hfh-4 and mogp-1 are expressed by oviduct but not uterine epithelium (43). hfh-4 is expressed by ciliated oviduct epithelial cells (44), while mogp-1 is expressed by secretory oviduct epithelial cells (45). The expression of hfh-4 and mopg-1 is mutually exclusive within individual cells. No mRNA could be detected for mogp-1 in either cell line under optimal conditions for its expression (10⁻⁸ M 17-β-estradiol in phenol red-free medium), while hfh-4 mRNA was detectable in Bm1.11 and Bm12.4 cells (Fig. 2). Detection of hfh-4 mRNA but not mogp-1 mRNA is consistent with Bm1.11 and Bm12.4 cells being of ciliated oviduct epithelial origin. Low levels of hfh-4 mRNA are not anomalous as cultured oviduct ciliated epithelial cells dedifferentiate and rapidly loose their cilia ex vivo (46). Absence of mRNA for the polymeric Ig receptor indicates that ex vivo culturing of these cell lines under the conditions used has altered some aspects of their biology compared with oviduct epithelial cells in vivo.

The mock-infected and C. muridarum-infected (10 IFU/cell) mouse oviduct epithelial cell lines Bm1.11 and Bm12.4 were examined by RT-PCR for expression of TLRs 1–9 and the NBS/LLR proteins Nod1 and Nod2. Recently discovered murine TLR11 (29) was not investigated, and human TLR10 does not have a murine homologue. For expression analysis, 1 μg of total cellular RNA was subjected to a 40-cycle RT-PCR using the primers listed in Table I. All primer pairs were shown to generate specific PCR products using total RNA from peritoneal macrophages or the RAW264.7 macrophage cell line (data not shown). As shown in Fig. 3,A, both uninfected cell lines had detectable mRNA for TLRs 1, 2, 3, 5, and 6 and Nod1. The faint PCR products for TLRs 1, 3, and 5 in Bm12.4 cells were readily demonstrable by increasing the cycle number to 48 (Fig. 3 B). There were qualitative differences in the mRNA expression levels for TLRs 1, 3, and 5 between the Bm1.11 and Bm12.4 cell lines. Expression of TLR4 mRNA was undetectable in both cell lines even with a 48-cycle RT-PCR.

To examine whether infection with Chlamydia alters PRM expression, both cell lines were infected with C. muridarum at 10 IFU/cell, and RT-PCR analysis was performed at 6, 12, 18, and 24 h postinfection. Infection-induced changes in TLR expression were first noted at 18 h postinfection and increased further by 24 h postinfection. The only qualitative changes seen during the above time course were increases in TLR mRNAs. No TLR message significantly decreased or vacillated up and down during the time course examined. Representative data from the 24 h time point are presented in Fig. 3 C. Interestingly, both cell lines exhibited strong induction of Nod2 24 h postinfection (indicated by the arrow). Nod2 mRNA is known to be up-regulated by TNF-α (47). We have previously shown the TNF-α mRNA is induced in infected Bm1.11 cells and that TNF-α protein is detectable in infected-Bm1.11 culture medium (14). There was no detectable TLR4 mRNA at any time point postinfection (or in uninfected cells), suggesting TLR4 does not play a role in oviduct epithelial activation by Chlamydia infection.

To assess the functional status of the identified TLRs, uninfected oviduct epithelial cell lines were challenged with highly purified TLR-specific agonists. TLR activation was determined by measurement of epithelial cytokines in cell culture supernatants. IL-6 and GM-CSF were chosen as representative acute phase cytokines and IFN-β as a representative immunomodulatory cytokine, based on our previous work in Bm1.11 cells (14). Even though Chlamydiae do not have a TLR5 PAMP (flagellin), we chose to characterize TLR5 because pelvic inflammatory disease often includes bacterial species from the lower reproductive tract. TLR9 was included in the functional analysis because Chlamydia DNA contains TLR9 PAMPs, unmethylated cytosines within CpG motifs. Because mRNAs for TLRs 7 and 8 were not detected with RT-PCR analysis, and Chlamydia lack their relevant PAMPs (viral ssRNA; Refs.48 and 49), we chose not to address those PRMs further in this study.

TLR5 mRNA was detected in both epithelial cell lines. The PAMP recognized by TLR5 is flagellin, a component of the flagellar motility apparatus used by many pathogenic bacteria. Because C. trachomatis and C. muridarum genomes do not encode a flagellin (50), TLR5 is not likely to play a role in epithelial activation during uncomplicated C. trachomatis genital tract infections but may play a role in polymicrobial pelvic inflammatory disease. Exposure of the oviduct epithelial cell lines to purified E. coli flagellin caused a modest dose-dependent secretion of IL-6 and GM-CSF in both cell lines consistent with functional TLR5 molecules (Fig. 4). Low levels of LPS contaminating the purified flagellin (∼12.5 pg/ml LPS in 100 ng/ml flagellin) are unlikely to account for the IL-6 production seen in Fig. 4 based on the poor IL-6 response of the oviduct epithelial cell lines to μg/ml quantities of LPS (upcoming section and see Fig. 6). Unmethylated CpG dinucleotide motifs within consensus sequences in bacterial and viral DNA are PAMPs that bind and trigger TLR9. Chlamydia genomic DNA contains unmethylated cytosines within CpG motifs. The synthetic ODN ODN1826 is a TLR9-specific agonist. As shown in Fig. 4, ODN1826 was unable to induce secretion of IL-6 or GM-CSF in either cell line, which is consistent with the absence of TLR9 mRNA in the RT-PCR analyses. TLR9 is not likely to be an important PRM for epithelial cells during Chlamydia upper reproductive tract infections, at least in the murine C. muridarum experimental model.

RT-PCR analysis shows that both oviduct epithelial cell lines express TLR2. TLR2 exists on cell surfaces as heterodimeric complexes with either TLR1 or TLR6. TLR1 and TLR6 do not have known PAMP recognition functions independent of TLR2 (51). TLR2 heterodimers are triggered by a wide variety of PAMPs, including peptidoglycan, lipopeptides, and fungal cell wall components. Exposure of the oviduct epithelial cell lines to the TLR2 agonist Pam₃CSK4, a lipopeptide analog, caused modest secretion of IL-6 and GM-CSF. The cytokine responses were saturated at 100 ηg/ml Pam₃CSK4 in both cell lines (Fig. 5). Pam₃CSK4 is a weak activator of TLR2 without simultaneous engagement of other signaling molecules such as dectin-1 (52). Conversely, exposure of the oviduct epithelial cell lines to E. coli peptidoglycan (PGN-EC) caused marked secretion of IL-6 and GM-CSF at levels similar to that seen with C. muridarum infection (Fig. 5).

RT-PCR analysis did not detect TLR4 mRNA in either oviduct epithelial cell line (Fig. 3). To further investigate any possible role of TLR4 in epithelial detection of C. muridarum, we exposed the cell lines to “ultrapure” E. coli LPS. Standard preparations of LPS are commonly contaminated by other microbial TLR2 PAMPs such as lipopeptides (53). Exposure of the oviduct epithelial cell lines to 1 μg/ml ultrapure LPS caused little or no secretion of IL-6. In the same experiments, exposure of RAW 264.7 macrophage cells to 1 μg/ml ultrapure LPS triggered marked secretion of IL-6 (Fig. 6). RAW 264.7 cells express significant levels of TLR4 mRNA in RT-PCR analyses (data not shown). The modest amount of IL-6 released by the Bm1.11 cell line in response to ultrapure LPS is 10- to 50-fold less than the same cell line releases in response to either PGN-EC or C. muridarum infection (Fig. 5 A). The IL-6 secreted by Bm1.11 cell in response to ultrapure LPS may be due to minor residual contaminants in the ultrapure LPS preparation signaling via TLR2 or Nod1 signaling pathways, or an alternative LPS signaling pathway, such as CD180 present in B cells and macrophages (54), that could exist in some reproductive tract epithelial cells.

With the exception of TLR3, all TLRs trigger secretion of acute phase cytokines such as IL-6 and GM-CSF via a MyD88-dependent signaling pathway. The NBS/LRR proteins Nod1 and Nod2 trigger acute phase cytokine secretion via the RICK/RIP2/CARDIAK adaptor, a MyD88-independent signaling pathway (55). To determine whether epithelial secretion of acute phase cytokines is dependent on MyD88 signaling, Bm1.11 cells were transfected with either siRNA specific for the MyD88 gene (siMyD88), a control scrambled RNA sequence siRNA (siSCR), or a DN MyD88 expression plasmid. To demonstrate the efficiency and specificity of the siRNA used in these experiments, real-time RT-PCR analyses were done with cytoplasmic RNA from siMyD88 and siSCR transfectants. Calculations from the SYBR-green readout of multiple real-time PCR assays demonstrated a 7.5- to 9.6-fold reduction of MyD88 mRNA in the siMyD88 transfections. Fig. 7,A shows typical results of siRNA experiments analyzed by real-time PCR (first panel) and a semiquantitative two-step PCR using MyD88 specific primers (second panel). Fig. 7 B shows a Western blot demonstrating expression of native MyD88 and truncated (DN) MyD88 proteins in the DN-MyD88 cell line and native MyD88 protein in Bm1.11 cells.

siMyD88-transfected, siSCR-transfected, and DN-MyD88 cells were stimulated with purified PGN-EC, infected with C. muridarum, or left untreated. Relative TLR activation via the MyD88 pathway was determined by measuring IL-6 released into the medium (Fig. 7 C). As shown, transfection of the Bm1.11 cells with the siRNA specific for the MyD88 gene, or expression of the dominant negative MyD88 protein, greatly reduced the IL-6 response to PGN-EC compared with Bm1.11 cells that were transfected with the siSCR or untreated Bm1.11 cells, respectively. Additionally, Bm1.11 cells transfected with the MyD88-specific siRNA, or expressing the DN MyD88 construct, exhibited marked reductions in IL-6 secretion when infected with C. muridarum

The data above are consistent with a dominant role for TLR2 in epithelial secretion of acute phase cytokines in response to C. muridarum infection. MyD88-dependent signaling through TLRs 4, 7, 8, and 9 is improbable because their mRNAs were not detected by RT-PCR in the oviduct epithelial cell lines. TLR5 is an unlikely trigger for acute phase cytokine secretion because Chlamydiae lack flagellin. The marked reduction in IL-6 secretion seen by blocking the MyD88 signaling pathway suggests that Nod1 and Nod2 are not major PRM for Chlamydia-infected epithelial cells because they trigger acute phase responses through a MyD88-independent signaling pathway (55).

RT-PCR analysis (Fig. 3) detected mRNA for TLR3 in both oviduct epithelial cell lines. The natural PAMPs for TLR3 molecules are dsRNA molecules associated with RNA virus replication and possibly cellular mRNA from necrotic cells (56). TLR3 is unique among the TLRs in that it is an important TLR for IFN-β induction, and its induction of IFN-β is MyD88 independent (57). PAMP-activated TLR3 signals through an adaptor protein called Toll/IL-1R domain-containing adaptor-inducing IFN-β (TRIF)/Toll-IL-1R domain-containing adaptor molecule (TICAM)-1 to activate IFN regulatory factor (IRF)3 that drives transcription of IFN-β. The other major TLR signaling pathways that lead to IFN-β transcription are TLRs 4, 7, 8, and 9. TLR4 activates IRF3 via the TICAM2 and TRIF/TICAM-1 adaptor proteins (51, 58). TLR 7, 8 and 9 induce IFN-β via MyD88-dependent pathway(s) where IFN-β transcription is driven by IRF5 or IRF7 rather than IRF3 (59, 60).

We previously reported that C. muridarum-infected Bm1.11 cells expressed moderate levels of IFN-β mRNA in RNase protection assays as early as 12 h postinfection. In multiple experiments, we consistently detect IFN-β in culture supernatants of infected Bm1.11 and Bm12.4 cells. Chlamydia species have no obvious PAMP to serve as a ligand for TLR3.

As expected, blocking the MyD88 signaling pathway with MyD88-specific siRNA had minimal affects (∼10% decrease) on IFN-β secretion by infected Bm1.11 cells (infected-Bm1.11: 197 ± 4 pg/ml; siMyD88 treated-then-infected-Bm1.11: 174 ± 3 pg/ml). Poly(I:C) is a synthetic nucleoside polymer that mimics dsRNA and is a potent TLR3 agonist. Exposure of the oviduct epithelial cells to 75 μg/ml poly(I:C) (Fig. 8), and in concentrations as high as 200 μg/ml (data not shown), did not trigger IFN-β secretion by the oviduct epithelial cell lines. In contrast, RAW 264.7 cells treated with poly(I:C) released substantial amounts of IFN-β into the medium in the same experiments demonstrating that the poly(I:C) stock was bioactive. C. muridarum-infected Bm1.11 and Bm12.4 epithelial cells, and RAW264.7 cells exposed to C. muridarum, secreted IFN-β (Fig. 8). Treatment of oviduct epithelial cells with the TLR3 agonist poly(I:C) also did not trigger detectable IL-6 secretion (data not shown).

TLR3 is localized on cell surfaces in fibroblasts and intracellularly in dendritic cells. Dendritic cells with intracellular TLR3 receptors respond to poly(I:C) exposure by secreting IFN-β (61). TLR3 localization in epithelial cells is not known. It is not clear from our data whether TLR3 is the PRM responsible for IFN-β secretion by C. muridarum-infected epithelial cells.

Understanding Chlamydia-induced pathophysiology is critical for developing interventions to prevent tubal infertility and ectopic pregnancies. Fallopian tube epithelium is the logical epicenter for the relevant pathophysiology because 1) Chlamydia replication is restricted almost exclusively to reproductive tract epithelial cells, and 2) Fallopian tube epithelium lines the lumen of the reproductive organ that becomes distorted and dysfunctional as a consequence of the infection. Our work and that of others have previously shown that Chlamydia-infected epithelial cells secrete proinflammatory cytokines (14, 15, 16, 17, 18, 19), including cytokines that have roles in leukocyte recruitment, leukocyte activation, fibrosis, and scarring. In this report, we address the PRMs that trigger cytokine release by infected oviduct epithelial cells. To our knowledge this is the first investigation focused on oviduct epithelial cells.

Our murine oviduct epithelial cell lines express TLRs 1, 2, 3, 5, and 6 as determined by RT-PCR analysis. With respect to TLR4, this expression pattern is similar to that described for primary and immortalized human lower reproductive tract epithelial cell lines derived from ectocervix and endocervix that lack TLR4. Similar to the primary and immortalized cervical epithelial cell lines, our oviduct epithelial cell lines were not significantly activated by E. coli LPS, a potent TLR4 ligand (34). Conversely, our murine oviduct epithelial lines differ significantly from a human uterine epithelial cell line and primary uterine epithelial cell lines that express TLR4 in addition to TLRs 1, 2, 3, 5, 6, 7, 8, and 9 (35, 37). The discordant TLR4 findings may represent intrinsic differences between epithelial cells lining the uterus vs those lining the cervix and oviducts. The absence of TLRs 7–9 in our mouse oviduct epithelial cell lines may represent intrinsic differences related to localization within the reproductive tract or differences between mice and humans.

At least in the murine system, TLR4 does not appear to play a major role in oviduct epithelial secretion of acute phase cytokines, such as IL-6 and GM-CSF, in response to C. muridarum infection. This conclusion is consistent with data from experiments done in TLR4-knockout mice. Mice deficient in TLR4 are not compromised in their ability to clear C. muridarum infections and are not spared the deleterious scarring of the oviducts caused by infection. TNF-α, IL-6, IFN-γ, and MIP-2 levels in genital secretions of infected TLR4-deficient mice are nearly identical to levels seen in wild-type mice (38).

Based on RT-PCR expression analysis, TLRs 2, 3, and 5 and the NBS/LRR proteins Nod1 and Nod2 were candidate PRMs for oviduct epithelial cell activation by Chlamydia infection. The inability to detect mRNA for TLRs 4, 7, 8, and 9 by RT-PCR analysis eliminated those TLRs as likely Chlamydia PRM for oviduct epithelial cells. However, while the Bm1.11 and Bm12.4 epithelial cell lines express TLR5, E. coli flagellin only weakly activated cytokine secretion and the Chlamydia genome does not encode a flagellin. TLR5 is not likely to be an important TLR for oviduct epithelial activation by Chlamydia infection, although it may play a role in polymicrobial pelvic inflammatory disease.

TLR3 is expressed by the epithelial cell lines and has a MyD88-independent pathway using TRIF/TICAM-1 instead of MyD88 to activate TRAF6 leading to NF-κB activation and acute phase cytokine secretion (62). Treatment of the oviduct epithelial cell lines with poly(I:C) did not trigger detectable IL-6 secretion. Our data differ from primary human polarized uterine epithelial cells that secreted significant amounts of IL-6, GM-CSF, TNF-α, and G-CSF in response to poly(I:C) (36). It is not clear whether the difference observed is due to a species difference, polarization vs monolayer culture, or intrinsic differences in epithelial cells lining the uterus vs those lining the oviducts. At least in murine oviduct epithelial cells, the dependence of IL-6 secretion on MyD88 and the lack of IL-6 secretion in response to the TLR3 agonist poly(I:C) make it unlikely that TLR3 makes a major contribution toward secretion of acute phase cytokines by infected oviduct epithelial cells.

Dependence of acute phase cytokine secretion on MyD88 signaling identifies TLR2 heterodimeric complexes as the principal TLR activating oviduct epithelial cell lines during C. muridarum infection. siRNA transfection targeting MyD88 and expression of a truncated DN MyD88 protein had dramatic effects on reducing IL-6 secretion triggered by the TLR2 agonist peptidoglycan or by Chlamydia infection. TLR2 signaling is MyD88 dependent (51). Because IL-6 secretion was MyD88 dependent and Nod1/Nod2 signaling is MyD88 independent, Nod1 and Nod2 are not likely to play major roles in oviduct epithelial secretion of acute phase cytokines in response to infection.

The conclusion that TLR2 heterodimers are the principal TLR for epithelial acute phase cytokine responses to infection is consistent with data from TLR2-knockout mice. Like TLR4-deficient mice, TLR2-deficient mice clear Chlamydia genital infections with kinetics similar to wild-type controls. Unlike TLR4-deficient mice, TLR2-deficient mice infected by C. muridarum had significantly lower levels of TNF-α, IL-6, IFN-γ, and MIP-2 in genital secretions compared with wild-type controls, implying a major role for TLR2 in releasing cytokines during infections of the genital tract. In addition, TLR2-deficient mice develop less oviduct scarring than is seen in infected wild-type and TLR4-deficient mice. TLR2 appears to contribute to oviduct pathology without contributing to host defense (38). Our data complement the in vivo data of Darville et al. (38) and suggest that TLR2 on oviduct epithelial cells contributes to acute phase cytokine secretion and scarring.

We have not investigated the Chlamydia PAMPs responsible for epithelial activation via TLR2. The recent finding that C. trachomatis LPS activates NF-κB by signaling through TLR2 rather than TLR4 (63) and that C. trachomatis has a potential peptidoglycan synthesis pathway (64, 65) suggest several possible Chlamydia PAMPs for TLR2.

We attempted to determine the PRM responsible for oviduct epithelial secretion of the immunomodulatory cytokine IFN-β. IFN-β plays an important role in the transition from innate to adaptive immunity (66). We previously showed that IFN-β mRNA is up-regulated in oviduct epithelial cells infected with C. muridarum (14). We document that infected oviduct epithelial cells secrete IFN-β in this report. Engagement of TLRs 3, 4, 5, 7, 8, and 9 up-regulates transcription of IFN-β via several different signaling pathways (51, 57, 58, 59). However, IFN-β induction via TLR5 is TLR4 dependent (67), and TLRs 4, 7, 8, and 9 were not expressed by our oviduct epithelial lines. Although TLR3 is the most logical TLR candidate for IFN-β secretion by infected-oviduct epithelial cells, we were not able to demonstrate that TLR3 was functional in the oviduct epithelial cell lines using the TLR3 agonist poly(I:C). This result differs from data with polarized human primary uterine epithelial cell lines that up-regulate IFN-β mRNA in response to poly(I:C) (36). It is unclear whether this observed difference is due to species, culture conditions, or differences between uterine and oviduct epithelium. The above issue aside, Chlamydiae do not have a known dsRNA type PAMP recognizable by TLR3. Cellular mRNA associated with necrotic cells has been shown to be a TLR3 ligand in dendritic cells (56); however, cellular mRNA is a less potent TLR3 agonist than poly(I:C), and IFN-β is up-regulated before cell death occurs in the oviduct epithelial cell cultures (data not shown). Currently, it is not clear whether TLR3 plays a significant role in IFN-β secretion by Chlamydia-infected oviduct epithelial cells. It is possible that C. muridarum induces IFN-β through an alternative pathway such as the IRF-3-dependent, TLR-independent, and Nod2-independent pathway described in macrophages infected by Listeria monocytogenes (68).

In vitro experiments using the cloned oviduct epithelial cell lines Bm1.1 and Bm12.4 implicate TLR2 as the principal PRM responsible for acute phase cytokine secretion triggered by C. muridarum infection. These cell lines should serve as useful reagents for determining the PRM responsible for infection-induced secretion of immunomodulatory cytokines such as IFN-β. Based on agreement with previously published in vivo data in TLR2- and TLR4-knockout mice, the oviduct epithelial cell lines appear to be useful reagents for studying innate epithelial responses to Chlamydia infection in the murine model.

We gratefully acknowledge Dr. Soon-Cheol Hong for providing the DN MyD88 expression plasmid pIRES2-EGFP-DN-MyD88 and thoughtful comments during evolution of this manuscript. We gratefully acknowledge Dr. Howard Edenberg at the Indiana University Center for Medical Genomics for assistance with real-time PCR. We also acknowledge the thoughtful manuscript critiques provided by Drs. Stanley Spinola, Soon-Cheol Hong, Byron Batteiger, and Margaret Bauer.

The authors have no financial conflict of interest.