The Recall Response Induced by Genital Challenge with Chlamydia muridarum Protects the Oviduct from Pathology but Not from Reinfection

Abstract

The significant morbidities of ectopic pregnancy and infertility observed in women after Chlamydia trachomatis genital infection result from ascension of the bacteria from the endocervix to the oviduct, where an overly aggressive inflammatory response leads to chronic scarring and Fallopian tube obstruction. A vaccine to prevent chlamydia-induced disease is urgently needed. An important question for vaccine development is whether sterilizing immunity at the level of the oviduct is essential for protection because of the possibility that a chlamydial component drives a deleterious anamnestic T cell response upon oviduct reinfection. We show that mice inoculated with attenuated plasmid-cured strains of Chlamydia muridarum are protected from oviduct pathology upon challenge with wild-type C. muridarum Nigg despite induction of a response that did not prevent reinfection of the oviduct. Interestingly, repeated abbreviated infections with Nigg also elicited recall responses that protected the oviduct from pathology despite low-level reinfection of this vulnerable tissue site. Challenged mice displayed significant decreases in tissue infiltration of inflammatory leukocytes with marked reductions in frequencies of neutrophils but significant increases in frequencies of CD4 Th1 and CD8 T cells. An anamnestic antibody response was also detected. These data indicate that exposure to a live attenuated chlamydial vaccine or repeated abbreviated genital infection with virulent chlamydiae promotes anamnestic antibody and T cell responses that protect the oviduct from pathology despite a lack of sterilizing immunity at the site.

INTRODUCTION

Chlamydia trachomatis is the most prevalent sexually transmitted bacterial pathogen in the world and a leading cause of preventable blindness. Although treatable with antibiotics, it continues to cause long-term sequelae, including ectopic pregnancy and infertility in women and blindness, in areas where trachoma is endemic. A vaccine is needed to prevent disease caused by the pathogen. A significant barrier to its development has been concerns regarding the potential for tissue-damaging delayed-type hypersensitivity (DTH) responses to chlamydial antigens, because early human trachoma vaccine trials revealed enhanced disease in a subset of individuals who received low-dose whole-cell vaccines (38), and heightened conjunctival inflammation was observed in immunized macaques challenged with a heterologous serovar (36). The risk that a whole-cell chlamydial vaccine could elicit such deleterious DTH responses has narrowed the focus to development of subunit vaccines.

Gamma interferon (IFN-γ)-producing CD4 Th1 cells are important for resolution of chlamydial infection and protection from disease (5, 10, 27). CD8 T cells also contribute to protection via IFN-γ release (16), but a recent study indicated that CD8 T cell production of tumor necrosis factor alpha (TNF-α) promotes oviduct damage after primary murine genital tract infection (19). Additional evidence pointing to a deleterious role for T cells linked heightened CD4 and CD8 T cell frequencies with worsened disease in the oviducts of guinea pigs infected with Chlamydia caviae after challenge infection, despite detection of significantly reduced bacterial loads in the oviducts (26, 28).

Mice that naturally resolve infection with Chlamydia muridarum Nigg frequently sustain such severe oviduct pathology (6) that it is impossible to accurately assess if exacerbated tissue damage occurs after reinfection. In contrast, mice with primary infections with the plasmid-cured strains of C. muridarum, CM3.1 and CM972, do not develop oviduct damage, and prior infection with these strains protects from pathology upon challenge with Nigg (21). In this study, we sought to determine if infection with live attenuated chlamydiae prevents oviduct damage via induction of a recall response that prevents reinfection of the oviduct, thereby prohibiting induction of T cell-mediated collateral tissue damage. In addition, prior studies have revealed that mice sustaining antibiotic-abbreviated infection with wild-type C. muridarum Nigg are as resistant to challenge infection as mice that resolve infection naturally, but no evaluation of oviduct damage after challenge was reported (32). We investigated the possibility that repeated abbreviated infection with fully virulent C. muridarum Nigg promotes adaptive recall responses that prevent reinfection of the oviduct and protect from pathology.

MATERIALS AND METHODS

Mice.

Eight-week-old female C3H/HeOuJ mice were purchased from Jackson Laboratories and maintained in the animal facilities at Children's Hospital of Pittsburgh. All experimental procedures were performed in accordance with institutional policies for animal health.

Chlamydia infection.

Plaque-purified C. muridarum strains Nigg and plasmid-cured CM972 and CM3.1 were propagated and titrated in L929 cells as described previously (21). Two repeat infection regimens were followed (Fig. 1A and B). In the genital tract challenge infection model (Fig. 1A), groups of anesthetized medroxyprogesterone-treated female C3H/HeOuJ mice were vaginally inoculated with 10⁵ inclusion-forming units (IFU) of plasmid-cured CM3.1 or CM972 or wild-type C. muridarum Nigg. A control group received sucrose-sodium phosphate-glutamic acid buffer (SPG) as a mock infection. Sixty-four days after primary inoculation, the mice were again treated with medroxyprogesterone before being vaginally inoculated with 10⁵ IFU of C. muridarum Nigg 7 days later (day 72 after primary inoculation). Groups of mice (n = 3 to 6/group) were euthanized on interval days postchallenge, and their cervices and oviducts were harvested for flow cytometric analyses and determination of bacterial burden. This procedure was repeated one or two times on each day examined, providing a total of 6 to 9 mice per time point. Additional groups of similarly treated mice (n = 5 to 7/group) were euthanized 42 or 56 days postchallenge, and their genital tract tissues were harvested for gross and histopathological assessment. In the genital tract multiple-infection model (Fig. 1B), groups of anesthetized medroxyprogesterone-treated female C3H/HeOuJ mice were vaginally inoculated with 10⁵ IFU CM3.1 or SPG. These groups were also retreated with medroxyprogesterone 49 days after primary inoculation and vaginally inoculated with 10⁵ IFU C. muridarum Nigg 7 days later. Infection was allowed to proceed for 7 days before treatment intraperitoneally with 0.3 mg doxycycline in 100 μl for 5 days, followed by 2 days of rest. This cycle of infection and treatment was repeated twice more, after which the mice were euthanized on interval days, and their cervices and oviducts were harvested for flow cytometric analyses (n = 3/group). Additional groups of similarly treated mice (n = 5/group) were euthanized 42 days after their last challenge inoculation, and their genital tract tissues were harvested for histopathological assessment.

Fig 1.

Murine genital tract repeated infection protocols. (A) Genital tract challenge infection model. Two experimental groups, CM3.1/Nigg and CM972/Nigg, were vaccinated with plasmid-cured C. muridarum and challenged with Nigg. Two control groups included a group infected twice with Nigg, Nigg/Nigg, and a group that received a single Nigg infection, SPG/Nigg. (B) Genital tract repeated infection model. Two experimental groups included a group vaccinated with CM3.1 and then challenged three times with Nigg, CM3.1/Nigg^d/Nigg^d/Nigg, and a group that received two abbreviated Nigg infections followed by a 3rd Nigg challenge, Nigg^d/Nigg^d/Nigg. A control group, SPG/Nigg, received a single Nigg infection.

Determination of bacterial load.

Lower genital tract infection was monitored by swabbing the cervix with a calcium alginate swab (Spectrum Industries), followed by enumeration of IFU (14). The oviduct bacterial load was determined by culture of homogenates on L929 cells; 1/10 volume of homogenate was blind passaged once before being titrated by plaque assay (22).

Histopathology.

Gross pathology was recorded, and the entire genital tract was removed en bloc and fixed in 10% formalin. Longitudinal 4-μm sections were cut, stained with hematoxylin and eosin, and evaluated by a pathologist blinded to the experimental design. Oviduct dilation was graded from 0 to 4, with 0 being normal, and each oviduct was individually scored (6).

Evaluation of antibody titers in vaginal fluid and sera.

Vaginal secretions were obtained by washing the vaginal vault twice with 50 μl phosphate-buffered saline (PBS) containing protease inhibitor (Roche Diagnostic). Sera were collected by retro-orbital bleeds or via cardiac puncture at the time of euthanasia, and samples were analyzed by enzyme-linked immunosorbent assay (ELISA) with X-irradiated Nigg elementary bodies used as the antigen (6). Goat anti-mouse IgG2a (1:5,000 dilution; BD Biosciences) conjugated to horseradish peroxidase was used as a detection antibody. Sera and lavage fluid from uninfected mice were used as negative controls. Pooled sera or lavage fluid from immune mice served as positive controls. Antibody titers for individual mice were determined as the highest dilution with an optical density greater than that of control wells.

Flow cytometry and detection of cytokines.

Mice were euthanized, and their cervices and oviducts were harvested and processed for flow cytometric analysis (24). For detection of intracellular cytokines, single-cell suspensions obtained from the cervix and oviducts were stimulated for 4 h in the presence of Golgi plug (BD Biosciences), phorbol myristate acetate (25 ng/ml), and ionomycin (250 μg/ml). Various combinations of the following antibodies were used: CD45-peridinin chlorophyll protein (PerCP)-Cy5.5 (clone 30-F11), CD3-V450 (clone 500A2), CD4-phycoerythrin (PE) (clone RM4-5) or CD4-V450 (clone RM4-5), CD8-fluorescein isothiocyanate (FITC) (clone 53-6.7), Ly6G/C-FITC (clone RB6-8C5), and IFN-γ–allophycocyanin (APC) (clone XMG1.2) from BD Biosciences, in addition to F4/80-PE-Cy7 (clone BM8) from eBioscience (San Diego, CA). For detection of memory markers, the following antibodies were used: CD4-V450 (clone RM4-5), CD27-APC (clone LG:3A10), CD44-APC-Cy7 (clone IM7), and CD62L-PerCP-Cy5.5 (clone MEL-14) from BD Biosciences and interleukin-7Rα (IL-7Rα)-PE (clone A7R34) from eBioscience. Cell viability was determined using LIVE/DEAD Near-IR (Invitrogen). Data were acquired in a FACSAria (BD Biosciences) and analyzed using FlowJo software (Tree Star); 50,000 to 100,000 cells were collected for each data point. Absolute numbers of leukocytes in cervices and oviducts were calculated as follows: (number of cells counted by hemocytometer/cervix or 2 oviducts) × (number CD45⁺ cells/total number of cells analyzed by flow cytometry). Cytokines and chemokines in oviduct homogenates were quantified using the Multiplex Cytometric Bead Array (Millipore, Billerica, MA).

Statistics.

Statistical comparisons between the groups for bacterial burden analyzed on days after infection were made by two-factor repeated-measures analysis of variance (ANOVA). The Kruskal-Wallis one-way ANOVA on ranks was used to determine significant differences in oviduct dilatation scores between groups. Differences in serum antibody, oviduct burden, and flow cytometric data were determined by two-factor ANOVA with a Bonferroni posttest. Analyses were performed using GraphPad Prism. P values of <0.05 were considered significant.

RESULTS

Nonsterilizing immunity at the level of the oviduct effectively protects from oviduct pathology.

We confirmed our previous findings that negligible oviduct damage occurs in mice challenged with Nigg if they had primary infections with CM972 (Fig. 2B and E) or CM3.1 (Fig. 2C and F). At 42 and 56 days postinfection, oviducts from mice that received two Nigg infections were dilated with a loss of plicae and flattened epithelia (Fig. 2A and D). Oviducts from mice sacrificed 56 days following a second Nigg infection showed obvious fibrosis and scarring (Fig. 2D) consistent with previous observations that oviduct pathology evolves over time with increased evidence of fibrosis from day 42 to day 56 (21, 30), while protection against Nigg-induced pathology was observed in mice that had resolved primary infection with the plasmid-cured strains (Fig. 2E and F). Oviduct dilation scores were 0 for both oviducts in 5 of 15 mice that had primary inoculation with the plasmid-cured strains, whereas a score of 0 occurred in only 1 of 15 mice that had primary infections with Nigg. The median oviduct dilation score was 3 for mice with primary infections with Nigg and 1 for each of the groups with primary infections with the plasmid-cured strains, which did not differ from a control group of mice that were mock infected but were statistically significantly different from mice infected twice with Nigg.

Fig 2.

Oviduct histology after challenge infection. Mice were challenged with fully virulent wild-type C. muridarum Nigg after they had resolved primary infection with wild-type Nigg (A and D), plasmid-cured CM972 (B and E), or CM3.1 (C and F). The images show representative hematoxylin-and-eosin-stained sections of oviducts 42 days (A to C), or 56 days (D to F) after challenge infection. (D) The arrows indicate fibrosis in oviduct sections. Magnification (all sections), 40×.

The chlamydial burden determined in cervicovaginal swabs obtained from these mice was lowest in mice receiving Nigg as both the primary and challenge inocula compared to mice with primary infection (Fig. 3A). Challenge Nigg infections in mice that had previously resolved infection with either CM972 or CM3.1 were not statistically different from each other, but both showed significantly reduced chlamydial loads after challenge with Nigg compared to the primary-infection group (Fig. 3A).

Fig 3.

Bacterial burden and anti-chlamydial IgG2a titers. (A) Lower genital tract bacterial burden after challenge infection with Nigg in mice that had previously resolved a primary infection with CM972, CM3.1, or Nigg or received SPG buffer. The data are means ± SD, with 12 animals examined each day; P < 0.05 by two-way repeated-measures (RM) ANOVA for the SPG/Nigg group versus each of the other groups. (B) Oviduct bacterial burden after a single infection with Nigg or a challenge infection with Nigg in mice that had previously resolved infection with CM972, CM3.1, or Nigg. The data are means and SD of oviducts from 6 to 9 mice per group analyzed at each time point; *, P < 0.05; #, P < 0.001 by two-way ANOVA. (C) Serum anti-chlamydial IgG2a titers. The bars represent mean log₂ titers and SD from 3 mice per group per day; *, P < 0.05 for the SPG/Nigg (primary-infection) group versus each of the challenge groups. (D) Anti-chlamydial IgG2a titers in genital tract secretions; the bars are mean log₂ titers and SD from three mice per group per day; #, P < 0.001 for the SPG/Nigg group versus each of the challenge groups on the day indicated. The antibody titer data are representative of a single experiment that was repeated once.

Oviducts from mice in each of the challenged groups had evidence of bacterial infection on at least 1 day examined, but the bacterial burden was extremely low, requiring use of the sensitive method of blind passage for detection and enumeration of viable bacteria (Table 1 and Fig. 3B). Resistance to reinfection of the oviduct was most frequent in mice with primary Nigg infection, as evidenced by significant reductions in the number of oviducts that were culture positive on days 13 and 15 compared to the other groups and the detection of lower titers of bacteria in animals that were culture positive. The bacterial burden in oviducts harvested from mice sustaining a primary infection was high by day 7, and this load was significantly (10²- to 10⁴-fold) higher than in oviducts from mice undergoing secondary infection (Fig. 3B).

Table 1.

Percentages of mice with culture-positive oviducts after blind passage in L929 cells

Group	No./total (%) of mice culture positive after challenge infection with Nigg
	Day 3	Day 7	Day 11	Day 13	Day 15
CM972/Nigg	1/6 (17)	4/6 (67)	5/9 (56)	6/9 (67)a	6/9 (67)a
CM3.1/Nigg	1/6 (17)	5/6 (83)	5/9 (56)	7/9 (78)a	6/9 (67)a
Nigg/Nigg	0/6 (0)	5/6 (83)	4/9 (44)	1/9 (11)	2/9 (29)
SPG/Nigg	2/6 (33)	6/6 (100)	9/9 (100)	9/9 (100)	9/9 (100)

^aP < 0.05 vs. Nigg/Nigg group by Fisher's exact test.

High titers of anti-chlamydial antibody are present in the sera and vaginal-lavage fluids of challenged mice.

In mice with primary Nigg infection, serum anti-chlamydial IgG2a titers gradually increased between days 3 and 15 postinfection (Fig. 3C). In contrast, in all three groups of mice that had previously resolved an infection, high titers of serum IgG2a were detected pre- and postchallenge, and these titers were consistently higher than those observed during primary infection. High titers of IgG2a were detected in vaginal-lavage fluids on the day prior to challenge in all groups of mice that had previously resolved an infection, and the titers remained high after challenge (Fig. 3D). No IgG2a was detected in lavage fluids from mice sustaining a primary infection with Nigg up to 10 days postinfection, suggesting a lag time for development of this local response.

Leukocyte infiltrates, percentages of neutrophils, and proinflammatory molecules are significantly reduced in genital tract tissues of challenged mice.

High numbers of CD45⁺ leukocytes were detected in the cervices (Fig. 4A) and oviducts (Fig. 4B) of mice with primary infections with Nigg by day 7, and the levels were sustained through day 15. In contrast, the numbers of leukocytes infiltrating the cervices and oviducts of challenged mice were low at almost every time point examined (Fig. 4A and B).

Fig 4.

Influx of inflammatory cells into the cervices and oviducts of mice after primary and challenge infections. (A and B) Absolute numbers of leukocytes (CD45⁺ cells) detected by flow cytometric analysis in the cervices (A) or oviducts (B) of mice on interval days after a single Nigg infection or after challenge infection with Nigg in mice that had previously resolved a primary infection with CM972, CM3.1, or Nigg. The bars are the means and SD of CD45⁺ cells detected per cervix or per two oviducts with three mice per group per time point. (C and D) Percentages (means and SD) of neutrophils (CD45⁺ Ly-6G/C^high F4/80⁻) determined in live CD45⁺ cells in the cervices (C) or oviducts (D) from the groups described in the legends to panels A and B. *, P < 0.05; **, P < 0.01; #, P < 0.001 by two-way ANOVA. (E and F) Levels of CXCL2 (E) or IL-1β (F) in oviduct homogenates of mice from the groups described in the legends to panels A and B. In all panels, the data are representative of one of two independent experiments. *, P < 0.05; #, P < 0.001 by two-way ANOVA.

A large percentage of the leukocytes infiltrating the cervices and oviducts after primary infection were neutrophils (Fig. 4C and D), and high percentages of neutrophils remained in both the lower and upper genital tract 15 days postinfection. Regardless of the primary infecting strain, the frequency of neutrophil detection in the cervices and oviducts after secondary challenge was significantly reduced with respect to primary infection on the majority of days examined (Fig. 4C and D). The significant decrease in absolute leukocyte numbers and percentages of neutrophils infiltrating the genital tract tissues paralleled the dramatic decrease in bacterial burden observed (Fig. 3A and B). Additionally, levels of the neutrophil chemokines CXCL2 (Fig. 4E), IL-17, and TNF-α (data not shown) and the proinflammatory and oviduct tissue-damaging cytokine IL-1β (9) (Fig. 4F) were significantly reduced in oviduct homogenates of mice sustaining challenge infection. The percentages of macrophages detected in the cervices did not differ between primary and challenge infections on any day examined and were higher in the oviducts of mice after primary infection only on day 3 (data not shown). These data indicate a strong positive association between bacterial burden and neutrophil, but not macrophage, influx.

Th1 cell infiltrates are significantly increased in the cervices and oviducts of mice after challenge infection.

In contrast to the marked decrease in frequencies of neutrophils observed after challenge in the cervices and oviducts (Fig. 4C and D), significant increases in frequencies of CD4 T cells were observed at these sites in challenged mice compared to mice with primary infections (Fig. 5A and B). High percentages of CD4 T cells were detected as early as 3 days postchallenge regardless of the chlamydial strain used for primary infection. Although the percentages of CD4 T cells detected in the oviducts during primary infection increased over time, they did not approach the levels observed during challenge until day 15 (Fig. 5B). Many of the CD4 T cells were of the Th1 phenotype, as determined by intracellular cytokine staining for IFN-γ (the average percentages of IFN-γ⁺ CD4 T cells in the cervices and oviducts of challenged mice were 56 and 48%, respectively) (Fig. 5C and D). A prominent Th1 phenotype was also observed with primary infection (the average percentages of IFN-γ⁺ CD4 T cells in the cervices and oviducts were 47 and 48%). Remarkably, despite the low absolute numbers of leukocytes in the cervices and oviducts of challenged mice, the absolute numbers of CD4 T cells in these tissues were similar to those of mice with primary infections on day 7 (cervix, mean ± standard deviation [SD] for CM972/Nigg = 1.4 × 10⁴ ± 9.8 × 10³; CM3.1/Nigg = 6.4 × 10⁴ ± 5.3 × 10⁴; Nigg/Nigg = 4.1 × 10⁴ ± 2.6 × 10⁴; SPG/Nigg = 1.2 × 10⁴ ± 1.2 × 10⁴; oviduct, mean ± SD for CM972/Nigg = 7.6 × 10³ ± 9.1 × 10³; CM3.1/Nigg = 1.0 × 10⁴ ± 1.1 × 10⁴; Nigg/Nigg = 1.4 × 10⁴ ± 0.7 × 10⁴; SPG/Nigg = 2.6 × 10⁴ ± 4.8 × 10⁴). Considering the significant decrease in bacterial burden upon challenge, the heightened frequency and similar absolute numbers of CD4 T cells indicated that an anamnestic CD4 T cell response was occurring at both tissue sites. The absence of oviduct pathology in mice with primary infections with plasmid-cured strains that were then challenged with wild-type Nigg suggested that this accelerated T cell response was not deleterious, but rather protective in nature.

Fig 5.

Frequencies of CD4 Th1 cells in the cervices and oviducts of mice sustaining challenge infection. (A and B) Percentages (means and SD) of CD4⁺ T cells (CD3⁺/CD4⁺) determined among live CD45⁺ cells in the cervices (A) or oviducts (B) of mice on interval days after a single Nigg infection or after challenge infection with Nigg in mice that had previously resolved a primary infection with CM972, CM3.1, or Nigg. *, P < 0.05; **, P < 0.01; #, P < 0.001 by two-way ANOVA. (C and D) Representative flow diagrams of cervical cells (C) and oviduct cells (D) harvested from mice that resolved a primary infection with CM972 and then sustained a challenge infection with Nigg for 11 days. CD45⁺/CD3⁺ cells were gated and analyzed for CD4 and IFN-γ.

Protection from oviduct pathology is associated with enhanced CD4 and CD8 T cell responses in the oviducts of mice sustaining multiple abbreviated infections with Nigg.

The impetus for development of a subunit rather than a whole-cell chlamydial vaccine derives from concern that antigens expressed by chlamydiae may induce deleterious T cell responses that induce damage if sterilizing immunity is not achieved. Plasmid-deficient C. muridarum does not express plasmid-encoded antigens, and reduced transcription of chromosomally encoded plasmid-responsive proteins has also been observed (20). In order to ensure that the mice that had been infected with plasmid-cured chlamydiae were exposed to the complete repertoire of chlamydial antigens, we modified our experimental protocol to incorporate repeated antibiotic-attenuated infections with Nigg (Fig. 1B). Arresting infection via antibiotic treatment 1 week after infection with Nigg, which expresses all plasmid-encoded and plasmid-regulated antigens, was anticipated to accomplish this without compromising the upper genital tract, because our previous studies had revealed that oviduct infection is rarely established prior to 7 days postinoculation (21). As anticipated, a reduced bacterial load was detected in cervicovaginal swabs from mice during a challenge infection with Nigg after they had resolved a primary infection with CM3.1 and two abbreviated infections with Nigg (Fig. 6A). Gross and histologic examination of oviducts from mice treated similarly and sacrificed 42 days after Nigg challenge revealed significantly lower oviduct dilatation scores than for mice sacrificed 42 days after a primary Nigg infection (Fig. 6B). Interestingly, mice that had not resolved a prior infection with plasmid-cured chlamydiae but sustained two antibiotic-abbreviated infections with Nigg before being challenged with Nigg also exhibited reduced lower genital tract bacterial burdens (Fig. 6A) and reduced oviduct dilatation scores compared to mice sacrificed 42 days after primary Nigg infection (Fig. 6B).

Fig 6.

Infection, pathology, and T cell responses in mice receiving multiple challenge infections with Nigg. (A) Bacterial burdens obtained at intervals after infection with Nigg in mice that resolved a primary infection with CM3.1 and then sustained two abbreviated infections with Nigg with antibiotic treatment, or simply two abbreviated infections with Nigg with antibiotic treatment, or were naïve for infection. The data are means ± SD with 11 or 12 animals examined each day; *, P < 0.05 by two-way RM ANOVA for the naïve group versus each of the other groups for the infection course after final challenge. (B) Oviduct dilatation scores determined by histologic examination of tissues harvested 42 days postinfection with Nigg in mice that were treated similarly to those described in the legend to panel A; n = 10 to 12 oviducts per group; *, P < 0.05 by Wilcoxon rank sum test. (C and D) Percentages (means and SD) of CD4⁺ T cells (CD3⁺/CD4⁺) (C) and percentages of CD8⁺ T cells (CD3⁺/CD8⁺) (D) determined among live CD45⁺ cells in the oviducts of mice treated similarly to those described in the legend to panel A and euthanized 7, 11, or 17 days after final challenge with Nigg. *, P < 0.05; **, P < 0.01; #, P < 0.001 by two-way ANOVA.

Oviducts from similarly treated mice contained CD4 (Fig. 6C) and CD8 (Fig. 6D) T cells at frequencies significantly higher than those from mice that received a single Nigg infection, and frequencies of CD4 T cells were higher than those of CD8 T cells (Fig. 6C and D). In addition, although the absolute numbers of leukocytes infiltrating the oviducts of multiply challenged mice were significantly decreased (data not shown), we observed that the absolute numbers of CD4 and CD8 T cells infiltrating their oviducts 7 days after challenge infection were marginally higher than those of mice sustaining primary infection (CD4 cells, mean ± SD for CM3.1/Nigg^d/Nigg^d/Nigg = 1.1 × 10⁴ ± 1.4 × 10⁴; Nigg^d/Nigg^d/Nigg = 6.2 × 10³ ± 9.0 × 10³; and SPG/Nigg = 1.5 × 10³ ± 2.6 × 10³; CD8 cells, mean ± SD for CM3.1/Nigg^d/Nigg^d/Nigg = 5.6 × 10³ ± 6.9 × 10³; Nigg^d/Nigg^d/Nigg = 3.1 × 10³ ± 4.8 × 10³; and SPG/Nigg = 8.3 × 10² ± 1.5 × 10³). The earlier peak of CD4 T cells in the CM3.1/Nigg^d/Nigg^d/Nigg group implies this regimen may promote enhanced protection. Clearly, these data indicated that a heightened recall T cell response was occurring at the level of the oviduct in challenged mice that were protected from oviduct pathology.

Iliac nodes of mice receiving multiple challenge infections exhibit a significantly greater frequency of early effector memory CD4 T cells but no increase in central memory CD4 T cells compared to mice resolving primary infection.

Immunity to chlamydial infection is partial and short-lived in mice, guinea pigs, and humans (2, 3, 17, 25), indicating that natural infection elicits an ineffective memory response. We sought to determine if memory CD4 T cell populations were induced in the draining lymph nodes of mice sustaining multiple infections. On the basis of the cell surface phenotype, we classified CD4⁺ T cells into effector (CD44^hi/IL-7Rα⁻) and memory (CD44hi/IL-7Rα⁺) cells (Fig. 7A). We wished to determine if these CD4 memory cells also contained subpopulations suggestive of differential stimulation or maturation stages. We therefore examined subsets within the IL-7Rα⁺ memory population using CD62L and CD27, a costimulatory receptor required for generation and maintenance of T cell immunity that has been used to define early effector memory cells (31). The CD4 memory population was primarily made up of effector memory rather than central memory cells (Fig. 7A and B). The central memory population was not increased by multiple challenge infections (Fig. 7B), and the frequencies of effector CD4 T cells detected in the iliac nodes were similar in singly and multiply infected mice (Fig. 7B). However, multiple infections led to increased percentages of effector memory CD4⁺ T cells (Fig. 7B) that were positive for CD27 and therefore considered early effector memory cells (Fig. 7C). Although effector memory cells are not thought to be the main component of long-lived memory, the detection of increased percentages in mice with multiple infections is encouraging, since these cells can mediate potent recall responses (7, 29).

Fig 7.

Effector and central memory CD4⁺ T cells in multiply challenged mice. (A) Gating strategy used to define effector (CD44^hi/IL-7Rα⁻) and memory (CD44^hi/IL-7Rα⁺) populations. CD4⁺ memory cells were subdivided using CD62L and CD27 to measure central memory (Tcm) (CD62L^hiCD27⁺) and early effector memory (TemE) (CD62L^loCD27⁺) T cells, as well as late effector memory T cells (TemL) (CD62L^lo CD27⁻). (B) Percentages (means and SD) of effector memory (Tem), central memory (Tcm), and effector (Teff) CD4 T cells in iliac nodes harvested 7 days after Nigg infection from mice that previously resolved a primary infection with CM3.1 and then sustained two abbreviated infections with Nigg or mice that previously sustained two abbreviated infections with Nigg or were naïve for infection. (C) Percentages (means and SD) of central, early, and late effector memory cells. *, P < 0.05; **, P < 0.01; #, P < 0.001 by one-way ANOVA with Tukey's posttest.

DISCUSSION

We have previously determined that mice that resolved infection with plasmid-cured strains of C. muridarum did not proceed to develop oviduct pathology when challenged with virulent wild-type Nigg, despite detection of reinfection at the level of the cervix. Our prior investigations also revealed induction of high titers of chlamydia-specific IgG2a in mice with primary infections with plasmid-cured C. muridarum, indicating that a prominent Th1 response had been induced in these mice that likely contributed to protection upon challenge. However, because of the concern that tissue-damaging T cell recall responses will occur in the absence of sterilizing immunity, we sought to determine if the complete protection from oviduct pathology observed in mice with primary infections with the plasmid-cured strains had resulted from induction of an adaptive immune response that prevented ascension of bacteria to the oviduct.

We determined that mice with primary infections with wild-type Nigg, as well as those with primary infections with plasmid-cured strains of C. muridarum, sustain lower bacterial loads in both the lower genital tract and the oviduct during challenge, but sterilizing immunity was not elicited at either site. Importantly, we observed significant decreases in the numbers of infiltrating leukocytes in challenged mice that were associated with significant decreases in the frequencies of neutrophils and in the levels of neutrophil-inducing chemokines and cytokines. Although levels of oviduct infection were extremely low, using the sensitive technique of blind passage, we were able to confirm the presence of live chlamydiae in the oviduct tissues after challenge in all groups. Thus, vaginal inoculation with plasmid-cured C. muridarum serves as a model of mucosal vaccination using a live attenuated vaccine, which although unsuccessful in inducing sterilizing immunity at the cervix or the oviduct, resulted in complete protection from oviduct disease. The cellular recall response in the cervix and the oviduct primarily consisted of IFN-γ-producing CD4 T cells. This raises the possibility that a vaccine that can induce anti-chlamydial antibody and CD4 Th1 memory responses that limit Fallopian tube infection may be sufficient to prevent disease in humans. The low bacterial burden and leukocyte infiltration detected in oviducts of mice that were challenged with Nigg after resolving a primary infection with Nigg suggests that the enhanced oviduct fibrosis observed in mice euthanized 56 days postchallenge compared to those euthanized 42 days postchallenge is the result of progressive oviduct tissue remodeling rather than a secondary inflammatory insult.

Data from trachoma vaccine trials in humans (38) and macaques (36) has raised concern that a whole-cell vaccine might elicit a more severe pathological response in vaccinees exposed to infection caused by stimulation of T cells that are hypersensitive to a specific antigen(s). This concern was extended to genital tract infection when peritubal adhesions were observed in four of four pigtailed macaques that had received five sequential cervical inoculations but in none of seven that received three or fewer inoculations (23). Nevertheless, repeated cervical inoculations in pigtailed macaques (37) and marmosets (12) resulted in protective immunity, as evidenced by shortened durations of repeat infections. Although repeated direct inoculation of subcutaneous oviduct transplants in pigtailed macaques results in fibrosis and scarring associated with prominent CD8 T cell infiltrates (34), this model prohibits immune protection associated with reduced ascension of bacteria to the oviduct. Human epidemiologic studies indicate increased risk for genital tract disease with repeated infection (1, 8, 15). If the adaptive response is insufficient to restrict the bacterial load in the oviduct to levels that do not elicit influx of tissue-damaging neutrophils, enhanced oviduct disease will likely occur with repeat infection due to repeated innate immune-cell-mediated tissue damage.

Data generated using the female guinea pig model with vaginal inoculation of C. caviae indicated that immunity that developed after primary infection was sufficient to significantly limit the bacterial burden upon challenge (25). A separate study indicated that more severe oviduct dilatation was observed in animals after reinfection (28). However, pathological assessment of challenged guinea pigs was performed 30 to 37 days after secondary infection (107 to 114 days after primary inoculation), whereas with primary infections, guinea pigs were examined at 75 to 85 days (28). Thus, the enhanced dilatation that was observed may reflect changes resulting from ongoing fibrotic processes in response to initial infection rather than the cumulative effect of episodes of tissue destruction triggered by each reoccurrence. A later study revealed enhanced frequencies of CD4 and CD8 T cells, but not Mac-1-positive cells, in the oviducts of guinea pigs 21 days after challenge infection compared to animals sustaining a primary infection (26). This finding parallels our observation that an increased frequency of CD4 T cells in the oviducts correlates with control of the chlamydial load and may be protective rather than pathological.

Since plasmid-encoded proteins are absent from plasmid-cured strains of C. muridarum, it was possible that mucosal inoculation with these attenuated strains failed to expose the mouse to sensitizing pathological ligands. However, we determined that mucosal vaccination with CM3.1 still protected mice challenged several times with virulent wild-type Nigg from oviduct disease, indicating that hypersensitivity after exposure to plasmid-encoded or regulated antigens was not induced during these reinfection episodes. Interestingly, repeated abbreviated infections with Nigg also resulted in substantial protection from oviduct disease. Protection was associated with increased frequencies of CD4 and CD8 T cells in challenged mice, indicating that a deleterious DTH response to specific chlamydial antigens associated with the whole organism was not induced.

We were unable to demonstrate any role for CD4 cells in pathology using these models, but a recent study has revealed that production of TNF-α by CD8 T cells contributes to oviduct damage during primary infection (19). Our results do not contradict this function of CD8 T cells, for although enhanced frequencies of CD8 T cells were observed after repeat infection, the phenotype of these cells remains to be investigated. Additionally, although the absolute numbers of CD8 T cells infiltrating the oviduct tissues early after infection were similar in primary and challenged mice, in mice sustaining primary infection, these cells increased over time and surpassed those in mice that sustained challenge infection.

For the first time, we determined that multiple abbreviated infections with fully virulent Nigg resulted in resistance to reinfection and substantial protection from oviduct disease. This not only rules out the development of deleterious DTH responses, but importantly, it confirms the ability of an infection attenuated via early treatment to induce resistance to infection and disease. A prior antibiotic treatment study in the mouse model revealed that a minimum duration of infection of 7 days was required for resistance to reinfection; earlier treatment compromised the immune response (33). Our data indicate that two abbreviated infections were not only sufficient to limit infection, but limited disease upon challenge. Analysis of memory CD4 T cell populations in the draining iliac nodes revealed increased frequencies of effector memory cells in mice that had sustained multiple infections. Effector memory cells can play a prominent role in recall responses and are likely contributors to the protection observed in this study.

These results have significant implications regarding screening and treatment programs in humans and induction of natural immunity. Such programs have resulted in reduced rates of complications (4, 18). At the same time, rates of infection in screened populations have risen, causing speculation that early case identification and treatment interfere with development of immunity (3). Others have suggested that the detection of increased rates of infection reflect a greater awareness of the infection, which has led to increasing testing being done and, consequently, a larger number of positive tests (35). Our data support the premise that the severity of inflammation resulting from chlamydial infection is more important in determining disease outcome than the frequency of infection and that rapid identification and treatment of incident infections may reduce disease without hampering development of natural immunity.

Our results are encouraging with respect to the prospects of the safety of a chlamydial vaccine that induces robust Th1 immunity. The induction of such a response in combination with anti-chlamydial antibody appears to be exclusively protective in the mouse model. Furthermore, given the lack of resident lymphoid tissue in the genital tract, sterilizing immunity may be impossible to induce. Our results suggest that complete protection from oviduct disease can be achieved despite ascension of infection to the oviduct, as long as the bacterial burden is reduced to a level that prevents a significant influx of neutrophils. Recent exciting work described protection against ocular pathology in a subset of nonhuman primates that were given multiple ocular infections with a plasmid-cured ocular strain of C. trachomatis prior to challenge with virulent A2497. Sterilizing immunity was required for protection from disease, for in the monkeys that became infected upon challenge, the intensity of ocular inflammation was equal to that in unvaccinated monkeys (13). A logical reason for the lack of a requirement for sterilizing immunity to protect from genital versus ocular disease may simply be related to the biological necessity for chlamydiae to ascend from the endocervix to the oviduct, giving time for the recall response to reduce the bacterial burden to levels that result in minimal inflammation. Although an attenuated chlamydial vaccine is unlikely to be marketable for prevention of genital tract disease, a combination of chlamydial antigens that elicit neutralizing antibody responses with Th1-inducing adjuvants holds promise for development of a safe and effective vaccine to prevent chlamydia-induced disease.