|Home | About | Journals | Submit | Contact Us | Français|
During natural infections Chlamydia trachomatis urogenital serovars replicate predominantly in the epithelial cells lining the reproductive tract. This tissue tropism poses a unique challenge to host cellar immunity and future vaccine development. In the experimental mouse model, CD4 T cells are necessary and sufficient to clear Chlamydia muridarum genital tract infections. This implies that resolution of genital tract infection depends on CD4 T-cell interactions with infected epithelial cells. However, no laboratory has shown that Chlamydia-specific CD4 T cells can recognize Chlamydia antigens presented by major histocompatibility complex class II (MHC-I) molecules on epithelial cells. In this report we show that MHC-II-restricted Chlamydia-specific CD4 T-cell clones recognize infected upper reproductive tract epithelial cells as early as 12 h postinfection. The timing of recognition and degree of T-cell activation are dependent on the interferon (IFN) milieu. Beta IFN (IFN-β) and IFN-γ have different effects on T-cell activation, with IFN-β blunting IFN-γ-induced upregulation of epithelial cell surface MHC-II and T-cell activation. Individual CD4 T-cell clones differed in their degrees of dependence on IFN-γ-regulated MHC-II for controlling Chlamydia replication in epithelial cells in vitro. We discuss our data as they relate to published studies with IFN knockout mice, proposing a straightforward interpretation of the existing literature based on CD4 T-cell interactions with the infected reproductive tract epithelium.
Chlamydia trachomatis is the most common bacterial sexually transmitted infection in the developed world, with 2 to 3 million actively infected individuals in the United States (3) and similar numbers in Europe (17). In women C. trachomatis infections can ascend into the upper reproductive tract, causing pelvic inflammatory disease and scarring with resulting infertility and ectopic pregnancies.
Histopathology studies show that C. trachomatis replicates predominantly in the reproductive tract epithelium during natural human infections (16, 36) and experimental murine C. muridarum infections (21). Inclusions are not seen in other cell types even though Chlamydia can undergo limited replication in macrophages and dendritic cells (33). It is unlikely that replication in non-epithelial cell lineages makes a major contribution to genital tract shedding. The mouse model for Chlamydia genital tract infections supports a critical role for CD4 T cells in protective immunity, as mice deficient in major histocompatibility complex class II (MHC-II) cannot control C. muridarum genital tract infections (22), and CD4 T-cell depletion is detrimental to resolution of primary genital tract infections (23). Because C. muridarum replicates in epithelial cells lining the reproductive tract, the most straightforward mechanism for clearing the genital tract would involve Chlamydia-specific CD4 T-cell interactions with infected epithelial cells. However in the absence of any data supporting this specific interaction, other indirect mechanisms based on CD4 T-cell production of gamma interferon (IFN-γ) and provision of help to B cells and CD8 T cells have been proposed as the mechanism for clearance (29).
C. muridarum-specific CD4 T-cell lines protective in adoptive-transfer studies were shown to control C. muridarum replication in polarized epithelial monolayers (14). The mechanism of control was dependent on IFN-γ and physical interaction of T cells with the infected epithelial cells via LFA-1. In the presence of IFN-γ, T-cell engagement of epithelial cells via LFA-1 was shown to augment epithelial nitric oxide production above that induced by IFN-γ alone, and nitric oxide was shown to be the effector molecule responsible for controlling Chlamydia replication (13). This anti-Chlamydia effector mechanism did not require that the CD4 T-cell clone recognize the infected epithelial monolayer in an antigen-specific fashion, as the same preactivated CD4 T-cell clone controlled Chlamydia psittaci replication in polarized epithelial monolayers even though it does not recognize a C. psittaci antigen.
Epithelial cells are semiprofessional antigen-presenting cells (APCs) and, in their unperturbed state, likely play a role in immunotolerance at mucosal surfaces (19). However in inflammatory environments, such as those resulting from transplant rejection and graft-versus-host disease, epithelial cells change their immunophenotype by upregulation of MHC-II (5, 25). In trachoma, an eye infection caused by Chlamydia trachomatis serovars A to C, conjunctival epithelial cells from human clinical specimens showed upregulated cell surface MHC-II and were presumably competent to present antigens to CD4 T cells (11, 12). In vitro studies have shown that rat and murine uterine epithelial cells process and present exogenous ovalbumin to OVA-specific CD4 T cells (28, 37). However in vitro processing and presentation of concentrated extracellular ovalbumin to CD4 T cells by uterine epithelial cells do not directly address whether Chlamydia antigens sequestered in membrane-bound inclusions get processed and presented to Chlamydia-specific CD4 T cells in vivo. The mechanics of CD4 T-cell contributions to resolution of genital tract infections remain unclear.
For this study we derived an epithelial cell line from the upper reproductive tract of a female C57BL/6 mouse and a panel of 10 Chlamydia-specific CD4 T-cell clones from immune C57BL/6 mice that previously self-cleared C. muridarum genital tract infections. These reagents gave us the opportunity to directly investigate (i) whether Chlamydia-specific CD4 T cells can recognize C. muridarum-infected reproductive tract epithelial cells, (ii) when during the time course of infection recognition occurs, and (iii) the role of IFNs in modulating epithelial interactions with CD4 T cells. We present the results of those investigations here.
(These data were presented in part at the 2009 Chlamydia Basic Research Conference.)
Female C57BL/6 and BALB/c mice were purchased from Harlan Laboratories (Indianapolis, IN). Female C57BL/6J, B6.C-H2bm12/KhEg, and B6.C-H2bm1/ByJ female mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were housed in Indiana University Purdue University-Indianapolis (IUPUI) specific-pathogen-free facilities. The IUPUI Institutional Animal Care and Utilization Committee approved all experimental protocols.
C57epi.1 is a cloned oviduct epithelial cell line derived from a C57BL/6 mouse (H-2b) using the methodology previously described (9, 10, 15) except for initial ex vivo culture in serum-free medium media supplemented with bovine pituitary gland extract (Gibco/Invitrogen, Carlsbad, CA) for several passages prior to switching to epithelial cell media supplemented as described below. The cultured C57epi.1 cells were grown at 37°C in a 5% CO2 humidified incubator in epithelial cell medium (1:1 Dulbecco's modified Eagle medium-F12K; Sigma, St. Louis, MO), supplemented with 10% characterized fetal bovine serum (HyClone, Logan, UT), 2 mM l-alanyl-l-glutamine (Glutamax I; Gibco/Invitrogen), 5 μg of bovine insulin/ml, and 12.5 ng/ml of recombinant human fibroblast growth factor 7 (keratinocyte growth factor; Sigma). C57epi.1 cells were fixed in 1:1 acetone-methanol and stained with anticytokeratin monoclonal antibody AE1/AE3 (Cappel ICN, Irvine, CA) or control antibody 36-5-7 (anti-H-2Kk; BD Biosciences, San Diego, CA) and counterstained with DAPI (4,6-diamidino-2-phenylindole; nuclear stain) to confirm an epithelial lineage as previously described (15).
Mycoplasma-free Chlamydia muridarum (Nigg isolate), previously known as the C. trachomatis strain mouse pneumonitis biovar (MoPn), was grown in McCoy cells (American Type Culture Collection no. CRL-1696). The titers of mycoplasma-free C. muridarum stocks were determined on McCoy cells with centrifugation as previously described (15). UV-inactivated C. muridarum stocks were made by diluting concentrated stocks in sucrose-phosphate-glutamic acid (SPG) buffer and then exposing 3 to 4 ml of diluted stock in a sterile petri dish to 1,200 J twice in a UV-cross-linking cabinet (Spectralinker; Spectronics Corporation, Westbury, NY). No viable C. muridarum inclusions were detectable in inoculated McCoy cell monolayers after UV inactivation.
Soluble C. muridarum antigen was prepared by infecting C57epi.1 epithelial cell monolayers in four 175-cm2 tissue culture flasks with C. muridarum at 3 inclusion-forming units (IFU) per cell. At 32 h the monolayers were harvested with glass bead agitation in 15 ml of residual medium per flask. Debris was pelleted with a low-speed spin (1,400 rpm [464 × g] for 10 min), and supernatant was collected; then elementary bodies were pelleted out of medium (depleted) with a high-speed spin (16,000 rpm [25,000 × g] for 30 min). The resulting supernatant was concentrated with a 10,000-kDa-molecular-mass-cutoff centrifugal filter (Amicon-15; Millipore, Bilerica, MA), aliquoted, and stored at −80°C.
C57BL/6 mice were treated with 2.5 mg of depoprogesterone (Depo-Provera; Pfizer, New York, NY) injected subdermally 1 week prior to infection. Vaginal infections were accomplished with 5 × 104 IFU of C. muridarum in 10 μl of SPG buffer. Mice were swabbed 7 days later to confirm infection. Vaginal swab IFU were recovered in SPG buffer and quantified using McCoy cell monolayers as previously described (15).
T-cell cultures were grown in RPMI 1640 (Sigma) supplemented with 10% characterized fetal bovine serum (HyClone), 2 mM l-alanyl-l-glutamine (Glutamax I; Gibco/Invitrogen), 25 μg/ml gentamicin (Sigma), and 5 × 10−5 M 2-mercaptoethanol (Sigma). This supplemented medium is referred to as T-cell medium hereafter. Secondary mixed lymphocyte culture (MLC) supernatants were prepared by combining 25 × 106 C57BL/6 splenocytes with 25 × 106 UV-irradiated (1,500 rads) BALB/c splenocytes in an upright 25-cm2 flask containing 20 ml of Dulbecco's modified Eagle medium supplemented with 10 mM HEPES, 10% characterized fetal bovine serum (HyClone), 2 mM l-alanyl-l-glutamine (Gibco/Invitrogen), 25 μg/ml gentamicin (Sigma), and 5 × 10−5 M 2-mercaptoethanol (Sigma; DME CM). Ten days later the viable C57BL/6 T cells were recovered and stimulated with irradiated BALB/c splenocytes (10 × 106 C57BL/6 T cells plus 25 × 106 irradiated BALB/c splenocytes in 20 ml of DME CM in an upright 25-cm2 flask) for 20 h. Supernatants were collected, filtered through 0.22-μm filters, aliquoted, and stored at −80°C until use.
Chlamydia-specific CD4 T-cell clones were derived from immune C57BL/6 (H-2b) female mice that had cleared a primary C. muridarum genital tract infection and were 7 days into clearing a secondary vaginal challenge with C. muridarum. Immune splenocytes harvested from mice were plated at 12.5 × 106 cells per well in tissue culture-treated 12-well plates, in T-cell medium containing murine recombinant interleukin-1α (IL-1α; 2 ng/ml), IL-6 (2 ng/ml), IL-7 (3 ng/ml), IL-15 (4 ng/ml), human recombinant IL-2 (100 units/ml), 20% 2° MLC, and 10 μg of UV-inactivated C. muridarum (~2.5 IFU equivalents per splenocyte) or 15 μl of soluble C. muridarum antigen (~1.5 cm2 infected monolayer equivalents). The resulting polyclonal T-cell populations were serially passaged and limiting diluted to obtain CD4 T-cell clones. CD4 T-cell clones designated uvmo-1, uvmo-2, uvmo-3, and uvmo-4 were derived from four independent polyclonal T-cell lines originating from immune splenocytes of four different mice using UV-inactivated C. muridarum as the antigen. These T-cell clones also recognize irradiated splenocytes pulsed with UV-inactivated C. muridarum grown in C57epi.1 (H-2b) epithelial cells, ruling out specificity for McCoy alloantigens originating from the McCoy fibroblasts used to propagate C. muridarum (data not shown). CD4 clones designated LN4-10, LN4-11, LN4-12, and LN4-13 were derived from the lymph nodes draining the reproductive tract (inguinal, iliac, and para-aortic), and Spl4-10 and Spl-11 were from immune splenocytes of a fifth mouse using soluble C. muridarum antigen. With the exception of IL-2, the recombinant T-cell growth factors used reflect those secreted by infected epithelial cells (15) and bone marrow-derived dendritic cells pulsed with heat-killed C. muridarum (32), which are remarkably similar. All T-cell clones listed above were CD4+ CD8− by flow cytometry (data not shown).
For routine passage of clones uvmo-1, -2, and -3, 1 × 105 CD4 clone cells were plated in 24-well tissue culture-treated wells containing 1.5 ml of T-cell medium-15% MLC supernatant supplemented with murine IL-1α (2 ng/ml), IL-6 (2 ng/ml), IL-7 (2 ng/ml), IL-15 (4 ng/ml), and human recombinant IL-2 (75 units/ml) plus 5 × 106 gamma-irradiated C57BL/6 splenocytes (1,200 rads) that had been prepulsed at 37°C with 2.5 IFU equivalent of UV-irradiated C. muridarum per splenocyte for 30 min. The remaining clones (uvmo-4; LN4-10, -11, -12, and -13; and Spl4-10 and -11) were passaged using the same conditions except that the irradiated splenocytes came from female mice that had self-cleared a C. muridarum genital tract infection (i.e., immune-irradiated splenocytes) and the antigen was 1.5 cm2 equivalent soluble C. muridarum antigen per 5 × 106 irradiated splenocytes. Immune-irradiated splenocytes are likely better able to efficiently process antigens present in low concentrations in the soluble antigen preparation (30). Chlamydia-specific CD4 T-cell clones were passaged every 6 to 8 days under these conditions. For the experiments in this study T cells were used on day 7 of their culture cycle. Recombinant murine cytokines were purchased from a commercial vendor (R&D Systems, Minneapolis, MN). Human recombinant IL-2 was obtained from Chiron Corporation (Emeryville, CA).
C57epi.1 cells were plated in 6-, 12-, 24-, or 48-well tissue culture plates and were used when confluent. Cells were infected with 3 IFU of C. muridarum per cell in 0.25 to 2 ml of culture medium depending on the culture plate format. The plates were centrifuged at 1,200 rpm (300 × g) in a tabletop centrifuge for 30 min and then incubated at 37°C in a 5% CO2 humidified incubator without change of medium for 3 to 21 h, depending upon the assay. Mock-infected wells received an equivalent volume of SPG buffer lacking C. muridarum.
T cells were dislodged from tissue culture plastic by removal of medium and incubation for 5 min in phosphate-buffered saline-EDTA. C57epi.1 cells were dislodged from tissue culture plastic using an EDTA wash followed by Hanks' salt-based enzyme-free cell dissociation buffer (Sigma). Cells were stained for 20 min on ice in phosphate-buffered saline-2% bovine serum albumin with phycoerythrin (PE)-coupled 53-5.8 (CD8β), PE-coupled YTS191.1 (CD4) (Cedarlane Laboratories, Burlington, NC), fluorescein isothiocyanate-coupled mouse immunoglobulin G2a (IgG2a) (control antibody), PE-coupled rat IgG2b (control antibody), PE-coupled M5/114.15.2 (MHC-II), GK1.5 (CD4; low endotoxin/no azide), and rat IgG2a (control; low endotoxin/no azide) (Ebioscience, San Diego, CA). Cells were fixed with 1% paraformaldehyde after staining. Cells were analyzed at the Indiana University Cancer Center Flow Cytometry Facility using a FACScan cytometer (BD Biosciences).
Relative IFN-γ levels were determined by enzyme-linked immunosorbent assay (ELISA) using monoclonal antibody XMG1.2, according to the manufacturer's protocol (Pierce-Endogen, Rockford, IL). Recombinant murine IFN-γ (R&D Systems) was used as the standard.
Epithelial cell targets were treated with 50 μg/ml of mitomycin C for 20 min at 37°C, washed twice with EDTA, dislodged with enzyme-free cell dissociation buffer, filtered through a 40-μM nylon filter, and counted. T cells (5 × 104) with 5 × 104 epithelial cells were cocultured in 200 μl of T-cell medium. At 36 h culture supernatants were harvested (50 μl) for cytokine analysis and wells were pulsed with 0.5 μCi of [3H]thymidine per well for 12 h. Proliferation assay products were harvested on glass fiber filters and counted using a Packard Matrix 9600 direct beta counter.
To test whether the CD4 clones could control Chlamydia replication in vitro, C57epi.1 monolayers in 48-well plates were untreated or treated with IFN-γ (10 ng/ml) for 14 h prior to infection or at the time of infection with 3 IFU of C. muridarum per cell. After addition of C. muridarum the plates were spun at 1,200 rpm (300 × g) for 30 min. Four hours after infection the inoculum was removed and CD4 T-cell clones were added in T-cell medium. Thirty-six hours postinfection, the cells and medium in each well were harvested by scraping and stored at −80°C until C. muridarum titers were determined on McCoy monolayers as previously described (15). Recombinant murine IFN-γ at all concentrations tested (up to 1,000 pg/ml) had no effect on C. muridarum titrations done on the McCoy monolayers; maximum IFN-γ carryover in dilutions used for quantifying C. muridarum was <50 pg/ml.
Summary figures for each experimental investigation are presented as “pooled” means with their associated standard errors of the means (SEM). Figure legends indicate the number of independent experiments pooled to generate each figure. Student's two-tailed t test was used to assess significance of pooled experimental data. P values that were <0.05 were considered statistically significant.
An upper reproductive tract epithelial cell line was derived from a C57BL/6 (H-2b) mouse by limiting dilution cloning as described in Materials and Methods. C57epi.1 epithelial cell monolayers in chamber slides were fixed and stained with control antibody (Fig. (Fig.1A)1A) and antibody specific for cytokeratins (Fig. (Fig.1B).1B). C57epi.1 cells express cytokeratins, consistent with an epithelial lineage. Also consistent with an epithelial lineage, they also have IFN-γ-inducible MHC-II expression (Fig. (Fig.1C1C).
There are no published murine CD4 or CD8 Chlamydia-specific T-cell clones derived from mice that self-cleared primary genital tract infections. We derived a panel of 10 C. muridarum-specific CD4 T-cell clones from five C57BL/6 (H-2b) female mice that cleared primary genital infections. Immune lymphocytes were harvested from spleens and lymph nodes draining the genital tract 1 week into a second vaginal challenge. The Chlamydia antigens used to activate T cells in vitro were crude preparations of UV-irradiated C. muridarum elementary bodies and soluble C. muridarum antigens; the APCs for routine passage were naive irradiated C57BL/6 splenocytes or immune-irradiated splenocytes. C57epi.1 cells were pretreated with IFN-γ and then mock infected or infected with C. muridarum for 12 and 18 h prior to harvest for use as targets. The 10 C. muridarum-specific CD4 T-cell clones were tested for their ability to recognize mock-infected versus C. muridarum-infected C57epi.1 epithelial cells at 12 and 18 h postinfection. They were also tested for their ability to recognize mock-pulsed syngeneic irradiated naive splenocytes (autoreactivity control) versus immune syngeneic irradiated splenocytes pulsed with UV-inactivated C. muridarum (specific antigen) (Table (Table1).1). The T-cell assays were done in the presence of 10 μg/ml tetracycline to block synthesis of additional Chlamydia polypeptides and progression of infection. Immune irradiated splenocytes pulsed with C. muridarum secreted a modest amount of IFN-γ without proliferating, while naive irradiated splenocytes make no detectable IFN-γ under identical conditions (data not shown). For the splenocyte APC data in Table Table1,1, the IFN-γ produced by immune irradiated splenocytes pulsed with UV-inactivated C. muridarum in control wells lacking CD4 T-cell clones was subtracted from IFN-γ produced in experimental wells containing antigen-pulsed immune irradiated splenocytes plus CD4 T-cell clones. This IFN-γ accounting procedure had no effect on the experimental conclusions.
As seen in Table Table1,1, all CD4 T-cell clones, regardless of derivation strategy, were able to recognize infected epithelial cells and immune splenocytes pulsed with UV-inactivated C. muridarum. CD4 T-cell clones differed in their abilities to recognize infected epithelial cells at 12 h and 18 h postinfection, and T-cell activation as determined by measuring IFN-γ production was significantly less for infected epithelial cells than for antigen-pulsed immune splenocytes for all CD4 T-cell clones. The limited numbers of T-cell clones derived using the different strategies are too small to draw conclusions about derivation-specific differences in relative activation by antigen-pulsed splenocytes versus infected epithelial cells. However, it is clear from comparing each clone's activation by infected epithelial cells to its activation by antigen-pulsed irradiated splenocytes that Chlamydia-specific CD4 T-cell activation by infected epithelial cells was submaximal for all clones tested.
Igietseme et al. (14) showed that T-cell lines that protected mice from vaginal infections with C. muridarum in adoptive-transfer experiments were also able to control C. muridarum replication in a polarized epithelial tumor cell line in vitro. We tested the ability of our panel of CD4 T-cell clones to control C. muridarum replication in C57epi.1 epithelial cells (Fig. (Fig.2).2). Monolayers of C57epi.1 cells in 48-well plates (~200,000 epithelial cells per well) were untreated or treated with IFN-γ, either 14 h prior to infection or at the time of infection, and then infected with C. muridarum. Four hours postinfection the inoculating medium was replaced with T-cell medium containing CD4 T-cell clones; 150,000 T cells were added per well, for an effector-to-target ratio of ~0.75:1. Thirty-two hours later the wells were harvested with additional SPG buffer and recovered C. muridarum was titered on McCoy monolayers to score replication. Pretreatment of C57epi.1 cells with IFN-γ had a modest effect on C. muridarum replication in the control wells (medium, [15 ± 10] × 106 IFU/well; IFN-γ treatment, [4 ± 2] × 106 IFU/well; pooled means from two experiments; P < 0.001). IFU recovered from experimental wells were compared with IFU from identically treated (untreated or IFN-γ-treated) parallel control wells (no T cells) to calculate % control replication. This normalization controls for the difference in C. muridarum replication in the untreated versus IFN-γ-treated C57epi.1 cells.
Eight of the 10 CD4 T-cell clones were able to block >90% of C. muridarum replication when epithelial cells were treated with IFN-γ prior to infection (Fig. (Fig.2A).2A). The two clones that could not were LN4-11 and uvmo-4. Control of replication only loosely correlated with each CD4 T-cell clone's ability to make IFN-γ when activated by infected epithelial cells (Table (Table1).1). The “ineffective” clone uvmo-4 (77 pg/ml) was one of the T-cell clones that was least activated by infected epithelial cells, while the other “ineffective” clone, LN4-11 (335 pg/ml), was in the middle. CD4 T-cell clones that made less IFN-γ than LN4-11 (LN4-10 and -13 and Spl4-10 and -11) were still able to control C. muridarum replication. Seven of the 10 clones showed improved control of C. muridarum replication with IFN-γ pretreatment of the epithelial monolayers (Fig. (Fig.2A).2A). Of note, three CD4 T-cell clones (uvmo-1, uvmo-2, and uvmo-3) were able to block ≥80% C. muridarum replication without IFN-γ pretreatment of the epithelial monolayers.
A previous study with human epithelial tumor cell lines and C. trachomatis serovar L2 showed that Chlamydia infection prior to IFN-γ exposure blocked IFN-γ-mediated upregulation of epithelial MHC-II by degrading an MHC-II transcription factor (38). To test whether C. muridarum could avoid cell-mediated immunity via this mechanism in vitro, the experiments shown in Fig. Fig.2A2A were repeated except that IFN-γ was added at the time of C. muridarum infection. Addition of IFN-γ at the time of infection had no effect on C. muridarum replication (medium, [5.6 ± 0.8] × 106 IFU/well; IFN-γ treatment, [5 ± 1] × 106 IFU/well). When IFN-γ was added at the time of infection only three CD4 clones were able to block ≥90% of C. muridarum replication (uvmo-1, uvmo-2, and uvmo-3) and only 4 of the 10 clones showed improved control of C. muridarum replication with IFN-γ treatment of the epithelial monolayers (Fig. (Fig.2B2B).
In summary we found that 2 of 10 Chlamydia-specific CD4 T-cell clones were ineffective even though they recognized infected epithelial cells, 5 clones effectively controlled C. muridarum in an IFN-γ-dependent fashion, and 3 clones efficiently controlled C. muridarum replication even without exogenous IFN-γ treatment of the epithelial monolayers. In addition, we found that C. muridarum infection interfered with the ability of the five IFN-γ-dependent CD4 clones to control replication when IFN-γ was added at the time of infection but not when it was added 14 h prior to infection. Based on the existing literature, these data strongly suggested that C. muridarum infection had a negative effect on IFN-γ-mediated upregulation of MHC-II and that the level of epithelial MHC-II expression was a limiting factor for the IFN-γ-dependent CD4 T-cell clones (LN4-10, -12, and -13 and Spl4-10 and -11). There was no clear correlation between lymphoid organ of origin (spleen versus draining lymph node) or antigen used ex vivo to activate T-cell lines (UV-inactivated C. muridarum versus soluble antigen) and the ability of the resulting T-cell clones to control in vitro C. muridarum replication.
We investigated the effect of C. muridarum on inducible epithelial cell surface MHC-II expression using the same experimental protocol used for the replication control experiments. C57epi.1 cells pretreated with IFN-γ for 14 h prior to infection (Fig. (Fig.2C)2C) were compared to C57epi.1 cells treated with IFN-γ at the time of infection (Fig. (Fig.2D).2D). Eighteen hours postinfection the epithelial monolayers were harvested, stained for MHC-II, and analyzed by flow cytometry. Consistent with the C. trachomatis serovar L2 data and the hypothesis that the role of IFN-γ for the IFN-γ-dependent CD4 T-cell clones is to upregulate epithelial MHC-II, C. muridarum infection modestly but reproducibly blocked MHC-II upregulation by IFN-γ added at the time of infection (Fig. (Fig.2D)2D) but not 14 h prior to infection (Fig. (Fig.2C).2C). Because of the difference in duration of IFN-γ exposure, the absolute amount of cell surface MHC-II was higher with 14-h IFN-γ pretreatment (Fig. (Fig.2C)2C) than with IFN-γ addition at the time of infection (Fig. (Fig.2D).2D). The three most effective CD4 clones, uvmo-1, -2, and -3, achieved nearly maximal inhibition of replication with the modest increase in MHC-II induced by addition of IFN-γ at the time of infection, while the IFN-γ-dependent CD4 clones (LN4-10, -12, and -13 and Spl4-10 and -11) could not control C. muridarum with this lower level of MHC-II. These data, combined with the data in Table Table11 showing that uvmo-1, -2, and -3 are better activated by infected epithelial cells as measured by IFN-γ production, are consistent with the hypothesis that the IFN-γ-independent CD4 clones are better able to control Chlamydia replication because they are better activated at lower levels of epithelial MHC-II. At the higher levels of epithelial MHC-II induced by 14 h of IFN-γ pretreatment, there was little difference between uvmo-1, -2, and -3, LN4-10, -12, and -13, and Spl4-10 and -11 in their abilities to control C. muridarum replication (Fig. (Fig.2A2A).
For logistical reasons we chose to focus on three CD4 T-cell clones (uvmo-1, uvmo-2, and uvmo-3) to further investigate CD4 T-cell interactions with infected reproductive tract epithelial cells. These three CD4 T-cell clones were derived from independent mice and were the most effective at controlling C. muridarum in vitro.
We investigated the role of the CD4 coreceptor during activation by infected epithelial cells because this directly addresses the role of MHC-II in T-cell activation. All the CD4 T-cell clones were CD4+ CD8−. A representative CD4 staining for T-cell clone uvmo-2 is shown in Fig. Fig.3A.3A. We mapped the MHC restriction element for uvmo-1, -2, and -3 using C57BL/6J (H-2b), bm1 (H-2IabKbm1), and bm12 (H-2Iabm12) mouse naive splenocytes mock pulsed or pulsed with UV-inactivated C. muridarum. These C57BL/6-derived mouse strains have a single MHC-II αβ heterodimer; C57BL/6J splenocytes are syngeneic with the CD4 T-cell clones, bm1 splenocytes are mismatched at the MHC class I K locus, and bm12 splenocytes are mismatched at the MHC-II locus. All three clones recognized C57BL/6J splenocytes and bm1 splenocytes pulsed with UV-inactivated C. muridarum, but their recognition of bm12 splenocytes (MHC-II mismatch) pulsed with C. muridarum was negligible or markedly attenuated (Fig. (Fig.3B).3B). The bm12 MHC-II heterodimer differs from that of the C57BL/6J heterodimer by 3 amino acids in the beta chain (20). This small change likely accounts for the residual partial activation of the uvmo-3 clone by UV-inactivated C. muridarum-pulsed bm12 splenocytes. With that small caveat, all three CD4 T-cell clones clearly recognize Chlamydia antigens presented by an MHC-II molecule.
Having demonstrated that the CD4 T-cell clones were MHC-II restricted, we asked whether the CD4 coreceptor was important during activation of CD4 T cells by infected epithelial cells. Infected epithelial cell targets were prepared by pretreating C57epi.1 monolayers with IFN-γ and then mock infecting them or infecting them with C. muridarum for 18 h. Eighteen hours postinfection the cell monolayers were harvested and cocultured with the CD4 T-cell clones in the presence of an anti-CD4 monoclonal antibody (GK1.5) or control antibody. Twenty-four hours later culture supernatants were harvested and assayed for IFN-γ content by ELISA to measure T-cell activation. Engagement of the CD4 coreceptor by epithelial MHC-II was critical for activation of all three CD4 T-cell clones (Fig. (Fig.44).
The timing of CD4 T-cell recognition of infected epithelial cells during the course of infection is unknown. Bone marrow-derived macrophages pulsed with heat-killed C. muridarum and fixed at ~2-h intervals showed nearly maximal activation of Chlamydia-specific CD4 T cells by 4 h post-antigenic pulse (34). To determine when CD4 T cells could recognize infected epithelial cells over the time course of infection, C57epi.1 monolayers were untreated or pretreated with IFN-β plus IFN-γ prior to infection (mock infection was time zero). C57epi.1 cells were pretreated with both IFN-β and IFN-γ because the reproductive tract epithelium of a wild-type mouse is likely exposed to IFN-β and IFN-γ during C. muridarum genital tract infections (addressed in more detail in next section). IFN treatments and infections were staggered such that all targets were ready at the same time. Monolayers were harvested and cocultured with CD4 T-cell clones in the presence of tetracycline. Tetracycline served to block progression of the C. muridarum infection and additional protein synthesis, the source of the Chlamydia polypeptides that serve as T-cell antigens. Thirty-six hours later culture supernatants were harvested to measure IFN-γ, and experimental wells were pulsed with [3H]thymidine to measure proliferation. CD4 T-cell clones were not stimulated to proliferate by infected epithelial cells at any time point during the course of infection; all clones were stimulated to proliferate by Chlamydia-pulsed irradiated splenocytes (data not shown). uvmo-1, -2, and -3 were activated by infected epithelial cells, as determined by measuring production of IFN-γ (Fig. (Fig.55).
For two of the three CD4 clones, pretreatment of the epithelial monolayer with IFN-β/γ improved T-cell activation compared to that in untreated monolayers, as determined by measuring IFN-γ production. Improved T-cell activation could represent either increased engagement of the T-cell receptor (TCR) due to more MHC-II antigen complexes on the epithelial cell surface or changes in epithelial accessory molecules that augment the TCR signal. Plotting the data with a smaller IFN-γ scale allows visualization of the earliest recognition events (Fig. (Fig.5).5). For CD4 T-cell clone uvmo-1, pretreatment of the epithelial monolayers with IFN-β/γ moved recognition from ~18 h in the untreated state to ~12 h with IFN pretreatment; for CD4 clones uvmo-2 and uvmo-3, recognition advanced from ~18 h to ~15 h postinfection. These experiments show that input antigen alone was not sufficient for CD4 T-cell recognition, as infection had to progress for at least 12 h (before addition of tetracycline) to generate a CD4 T-cell-recognizable target. IFN pretreatment improved both CD4 T-cell activation and recognition, though the magnitude of these effects varied by CD4 T-cell clone.
The Chlamydia pathogenesis knockout mouse literature shows contrasting roles for type 1 and type 2 IFNs during clearance of genital tract infections. As a broad generalization, type 2 IFN (IFN-γ) makes a positive contribution to clearance (7, 8), while type 1 IFNs (IFN-α/β) have a negative effect, recently documented in the IFNAR1 knockout mouse (24). Physiologic levels of IFN-γ have been documented in the genital secretions of Chlamydia-infected humans (2) and mice (7). IFN-β is secreted by infected epithelial cells (9, 10, 15) and is detectable in the genital secretions of mice infected vaginally with C. muridarum (W. A. Derbigny, personal communication). The presence of type 1 and type 2 IFNs in genital secretions during infection makes it likely that many reproductive tract epithelial cells are exposed to IFNs prior to becoming infected with Chlamydia. We examined the role of type 1 and type 2 IFNs in vitro using our reproductive tract epithelial cell line and CD4 T-cell clones. We propose that untreated epithelial cells mimic the IFN milieu of early infection (low IFN levels), IFN-β-pretreated cells mimic IFN-γ knockout mice (IFN-β with no IFN-γ), IFN-γ-pretreated cells mimic IFNAR1 knockout mice (IFN-γ with no functional IFN-α/β), and IFN-β/γ-pretreated cells mimic wild-type mice.
C57epi.1 cells were untreated or pretreated for 14 h with IFN-β, IFN-γ, or IFN-β/γ and then infected with C. muridarum. Eighteen hours postinfection the epithelial monolayers were harvested and cocultured with CD4 T-cell clones in the presence of tetracycline. Twenty-four hours later culture supernatants were harvested and analyzed for IFN-γ to score T-cell activation (Fig. (Fig.66).
For all three CD4 T-cell clones, IFN-β pretreatment of epithelial cells prior to infection augmented CD4 T-cell activation compared with that in untreated infected epithelial cells, with the caveat that the comparison for uvmo-3 had a P value of 0.050 (Fig. (Fig.6A).6A). IFN-γ pretreatment impressively augmented T-cell activation beyond that seen with IFN-β for all three CD4 clones; interestingly, copretreatment with IFN-γ and IFN-β blunted IFN-γ augmentation, though copretreatment was still better at activating the CD4 T-cell clones than was IFN-β pretreatment alone. We hypothesized that these results could be explained by effects of type 1 and type 2 IFNs on expression of epithelial MHC-II.
To look at the effects of type 1 and 2 IFNs on inducible epithelial cell surface MHC-II expression in the setting of C. muridarum infection, we used the same protocol that was used for preparing the infected epithelial targets described above. C57epi.1 cells were untreated or pretreated with IFN-β, IFN-γ, or IFN-β/γ prior to infection. Eighteen hours postinfection the epithelial monolayers were harvested, stained for MHC-II, and analyzed by flow cytometry (Fig. (Fig.6B).6B). As hypothesized, the levels of CD4 T-cell activation correlated with the relative amounts of epithelial cell surface MHC-II induced by the different IFN pretreatments: untreated < IFN-β pretreated < IFN-β/γ pretreated IFN-γ pretreated. The relative cell surface levels of MHC-II correlate with the published C. muridarum clearance rates from IFN knockout mice: IFN-γ knockout (IFN-β pretreatment) < wild type (IFN-β/γ pretreatment) IFNAR1 knockout (IFN-γ pretreatment).
Derivation of biologically intact upper reproductive tract epithelial cell lines and Chlamydia-specific CD4 T-cell clones from mice that self-cleared primary C. muridarum genital tract infections gave us the opportunity to investigate basic Chlamydia pathogenesis questions related to CD4 T-cell interactions with infected epithelial cells. The first question addressed was whether Chlamydia-specific CD4 T cells could recognize infected reproductive tract epithelial cells. Compared with professional APCs, epithelial cells express low levels of MHC-II, lack critical costimulatory molecules, and lack an obvious antigen presentation pathway for Chlamydia antigens. There have been doubts about whether CD4 T cells directly mediate clearance through interactions with infected reproductive tract epithelial cells (29), though depletion studies show a dominant role for CD4 T cells in primary infection (23) and knockout mice show an absolute dependence on MHC-II for clearance of C. muridarum from the genital tract (22). Data presented here clearly show that Chlamydia-specific CD4 T cells can recognize infected epithelial cells in antigen-specific fashion, including engagement of the CD4 coreceptor by epithelial MHC-II molecules.
We investigated the timing of Chlamydia-specific CD4 T-cell recognition of infected epithelial cells over the course of infection. When T cells recognize infected targets is an important mechanistic consideration, as delayed recognition narrows the window for an effector mechanism to act before noninfectious reticulate bodies transition to infectious elementary bodies. We found that CD4 T-cell recognition of infected epithelial cells occurred roughly 12 to 18 h postinfection and that recognition was influenced by type 1 and 2 IFNs. This window of time is relatively late during the replication cycle of C. muridarum in oviduct epithelial cells (26) but is theoretically early enough to allow T-cell-mediated disruption of replication by either disabling the epithelial host cell (incubator) or by directly attacking the inclusion by, for example, production of nitric oxide (13).
We stringently tested whether our CD4 T-cell clones could control Chlamydia replication in vitro by using them at an effector-to-target ratio of <1 in flat-bottom tissue culture plates, which provide an additional surface area challenge for physically small lymphocytes. Eight of our 10 CD4 T-cell clones controlled replication of C. muridarum in the oviduct epithelial cell line even though T-cell activation by infected epithelial cells was clearly submaximal, as reflected by lower IFN-γ production (50- to 100-fold) and lack of proliferation compared with activation by antigen-pulsed irradiated splenocytes. Interestingly the CD4 T cells that could control replication in vitro had two phenotypes. One set of CD4 clones (uvmo-1, -2, and -3) was able to control C. muridarum replication with or without IFN-γ pretreatment of epithelial cells; the other set (LN4-10, -12, and -13 and Spl4-10 and -11) was dependent on IFN-γ pretreatment of the epithelial cells prior to infection. We propose that the existence of Chlamydia-specific CD4 T-cell clones that recognize and control C. muridarum replication in epithelial monolayers without IFN-γ in vitro explains the in vivo observation that 99.9% of C. muridarum is cleared from the genital tracts of IFN-γ knockout mice with nearly normal kinetics (6, 27). We hypothesize that this IFN-γ-independent subset of CD4 T cells recognize Chlamydia antigens that are efficiently processed, making it to the cell surface bound to MHC-II molecules at very low levels of MHC-II expression. Alternatively, this T-cell subset may have high-affinity TCRs that allow activation with relatively few antigen-MHC complexes on the epithelial cell surface. While it is not possible to draw conclusions based on the limited numbers of T-cell clones from each derivation strategy, it is interesting that the three most effective CD4 clones came from a culture system in which the APC was a naive splenocyte pulsed with UV-inactivated C. muridarum. Because the inoculum was low, ~2.5 IFU per naive splenocyte, this suggests either very efficient processing of select C. muridarum antigens or a bias of the culture system toward selection of T cells with high-affinity TCRs. Either scenario suggests that development of a successful subunit vaccine could depend on choosing a specific Chlamydia antigen(s).
The source of the Chlamydia-specific CD4 T-cell clones did not appear to be predictive of their ability to control in vitro replication of C. muridarum in epithelial cells. The most effective and least effective CD4 T-cell clones came from the spleens of immune mice. CD4 clones derived from lymph nodes draining the reproductive tract were not more effective that those from splenocytes. The finding that Chlamydia-specific CD4 T cells (uvmo-4 and LN4-11) could be activated yet ineffective for controlling replication raises the possibility that ineffective CD4 T-cell responses contribute to inflammatory damage without contributing to clearance.
It is tempting to attribute T-cell-mediated control of Chlamydia replication in vitro to an indirect IFN-γ-based mechanism such as induction of nitric oxide because the three most effective CD4 T-cell clones (uvmo-1, -2, and -3) make the most IFN-γ in response to infected epithelial cells. However production of IFN-γ in response to infected epithelial cells only loosely correlated with the ability of individual CD4 T-cell clones to control C. muridarum replication, and none of the five CD4 T-cell clones dependent on IFN-γ pretreatment were able to control C. muridarum replication when exogenous IFN-γ was added to epithelial monolayers at the time of infection, arguing against a direct effect of IFN-γ. In addition, C. muridarum replicating in murine epithelial cells (this report; 26, 31) and mouse embryonic fibroblasts (1, 4) is largely indifferent to recombinant IFN-γ. This makes it highly unlikely that IFN-γ of T-cell origin, which would not be present until several hours after recognition (12 to 18 h postinfection), could directly contribute to the anti-Chlamydia effector mechanism of these clones. Inhibition of nitric oxide production does not interfere with the ability of CD4 T-cell clones (uvmo-1, -2, and -3) to control C. muridarum replication in vitro (our unpublished data). Our data support the hypothesis that the major role of IFN-γ at the epithelial interface during infection is to upregulate MHC-II on epithelial cells in order to efficiently activate Chlamydia-specific CD4 T cells.
We showed that infection modestly blunted upregulation of cell surface MHC-II when IFN-γ was added at the time of infection but not when added 14 h prior to infection, consistent with data published by Zhong et al. for C. trachomatis serovar L2 (38). We hypothesize that, for the CD4 T-cell clones dependent on IFN-γ pretreatment, IFN-γ is needed to upregulate MHC-II in order to increase the availability of MHC-II molecules to present Chlamydia antigens at the cell surface, though we cannot rule out IFN-γ effects on antigen processing. At the low levels of epithelial cell surface MHC-II seen with simultaneous IFN-γ addition and infection, the modest infection-associated blunting of MHC-II upregulation resulted in significant attenuation of CD4 T-cell-mediated control of C. muridarum replication for the five IFN-γ-dependent CD4 T-cell clones, presumably through decreasing T-cell recognition and/or activation. The in vivo importance of infection interference with IFN-γ-mediated upregulation of MHC-II, however, is unclear, as Darville et al. (7) have shown physiologic IFN-γ levels in genital secretions by day 3 post-C. muridarum infection and El-Asrar et al. have shown upregulated epithelial MHC-II in human trachoma clinical specimens (11, 12). These data combined with our in vitro IFN-γ pretreatment data suggest that the host may be able to overcome this evasion tactic by “pretreating” the vulnerable reproductive tract epithelium with IFN-γ prior to impending infection of individual epithelial cells.
Type 1 and 2 IFNs modulated T-cell activation by infected epithelial cells in this report, and C. muridarum infections of IFN knockout mice have yielded interesting results in the existing literature. In wild-type mice both type 1 (IFN-α/β) and type 2 (IFN-γ) IFNs are part of the inflammatory milieu during Chlamydia genital tract infections. Nagarajan et al. have shown that IFNAR1 knockout mice lacking the receptor for IFN-α/β clear C. muridarum from the genital tract more efficiently and with less oviduct pathology than wild-type mice (24). We showed here that IFN-γ pretreatment of epithelial cells dramatically improved CD4 T-cell activation by infected epithelial cells and that this effect was greatly attenuated when IFN-β was included during pretreatment. The rate of C. muridarum genital tract clearance by different mouse strains has been correlated with the timing and magnitude of IFN-γ responses (7, 8), which occur before demonstrable recall cellular immunity has developed (35). It is possible that the balance of IFN-β versus IFN-γ within an individual may be an important parameter affecting resolution of genital tract infections.
IFN-β attenuation of IFN-γ-enhanced T-cell activation correlated with marked attenuation of IFN-γ-induced MHC-II expression on C57epi.1 epithelial cells. This finding is consistent with previously described IFN-β blockade of IFN-γ-inducible MHC-II expression (18). IFNAR1 knockout mice likely clear C. muridarum from the genital tract faster because the reproductive tract epithelium expresses higher levels of MHC-II, and on that basis infected epithelial cells more efficiently activate the CD4 T cells, allowing a larger fraction of Chlamydia-specific CD4 T cells to participate in clearance. The MHC-II level on IFN-β/γ-pretreated C57epi.1 cells was only slightly higher than that on IFN-β-pretreated cells, and both were higher than levels on untreated cells. That is, IFN-γ knockout mice (type 1 only IFN milieu) likely have some epithelial MHC-II, and that level is sufficient for IFN-γ-independent CD4 T-cell clones like uvmo-1, -2, and -3, described in this report, to mediate clearance, albeit with a lower efficiency, consistent with the observed low-level residual genital tract shedding of C. muridarum in IFN-γ knockout mice.
Based on the demonstrated ability of C. muridarum-specific CD4 T-cell clones to recognize infected upper reproductive tract epithelial cells and control Chlamydia replication in them, it is reasonable to interpret the CD4 depletion data (23) and MHC-II knockout mouse study (22) as showing a critical role for CD4 T cells in directly clearing Chlamydia from the genital tract. These results do not address the controversial role of CD8 T cells and, in our view, do not rule out a redundant noncritical role for CD8 T cells in genital tract clearance. Our data suggest that not all Chlamydia-specific Th1 cells, even those generated during the course of a genital tract infection, are capable of contributing to clearance. Therefore, identification of the protective CD4 T-cell subsubset(s) and the antigens that they recognize may be critical for development of an effective Chlamydia vaccine.
We have no conflicts of interest related to this study.
We thank Wilbert Derbigny for his thoughtful critique of the manuscript as it was being developed.
This research was funded by NIH T32-AI060519, 1R01AI070514-01A1, and the Showalter Trust.
Editor: R. P. Morrison
Published ahead of print on 10 August 2009.