Caspase-11 Contributes to Oviduct Pathology during Genital Chlamydia Infection in Mice

It has been shown that caspase-1, but not its upstream activator, ASC, contributes to oviduct pathology during mouse genital Chlamydia muridarum infection. We hypothesized that this dichotomy is due to the inadvertent absence of caspase-11 in previously used caspase-1-deficient mice. To address this, we studied the independent contributions of caspase-1 and -11 during genital Chlamydia infection.

ABSTRACT

It has been shown that caspase-1, but not its upstream activator, ASC, contributes to oviduct pathology during mouse genital Chlamydia muridarum infection. We hypothesized that this dichotomy is due to the inadvertent absence of caspase-11 in previously used caspase-1-deficient mice. To address this, we studied the independent contributions of caspase-1 and -11 during genital Chlamydia infection. Our results show that caspase-11 deficiency was sufficient to recapitulate the effect of the combined absence of both caspase-1 and caspase-11 on oviduct pathology. Further, mice that were deficient for both caspase-1 and -11 but that expressed caspase-11 as a transgene (essentially, caspase-1-deficient mice) had no significant difference in oviduct pathology from control mice. Caspase-11-deficient mice showed reduced dilation in both the oviducts and uterus. To determine the mechanism by which caspase-11-deficient mice developed reduced pathology, the chlamydial burden and immune cell infiltration were determined in the oviducts. In the caspase-11-deficient mice, we observed increased chlamydial burdens in the upper genital tract, which correlated with increased CD4 T cell recruitment, suggesting a contribution of caspase-11 in infection control. Additionally, there were significantly fewer neutrophils in the oviducts of caspase-11-deficient mice, supporting the observed decrease in the incidence of oviduct pathology. Therefore, caspase-11 activation contributes to pathogen control and oviduct disease independently of caspase-1 activation.

INTRODUCTION

In women, ascension of Chlamydia trachomatis infection to the upper genital tract can have dramatic negative impacts on reproductive health (reviewed in reference 1). Female mice infected genitally with C. muridarum exhibit similar irreversible damage to the upper reproductive tract, with significant dilation and scarring in the uterine horns and oviducts. This mouse model has proved to be an invaluable tool to determine the inflammatory drivers of irreversible scarring in the oviduct. It has been shown that a number of immune-signaling receptors, including Toll-like receptor 2 (2), tumor necrosis factor receptor I/II (3, 4), interferon alpha/beta receptor (IFNAR) (5), and interleukin-1 (IL-1) receptor (IL-1R) (6), all contribute at differential levels to oviduct pathology during infection. The cytokines that bind these receptors, the downstream signaling pathways, and their interactions with cell types involved in persistent inflammation have been ongoing areas of investigation.

Activation of canonical inflammasome caspase-1 is essential for secretion of cytokines, including IL-1β and IL-18, which are associated with inflammation. Infection with C. muridarum or C. trachomatis leads to activation of caspase-1 in epithelial and immune cells in vitro, leading to IL-1β/IL-18 secretion (7–9) through activation of inflammasome constituents NLRP3 (6, 10) and AIM2 (11), both of which then activate ASC. However, the contribution of caspase-1 and inflammasome activators to oviduct pathology in vivo is contradictory. It has been shown that caspase-1-deficient mice exhibit reduced oviduct pathology (8). However, mice deficient in the upstream sensors NLRP3 or NLRC4 or in ASC, the downstream adaptor of caspase-1 activation, show no difference in the incidence of hydrosalpinx, oviduct dilation, and neutrophil infiltration during C. muridarum infection (6), suggesting a limited role for canonical inflammasome signaling in oviduct pathology. Interestingly, Dixit and colleagues showed that the caspase-1-deficient mice used in all earlier studies conducted before 2011 were actually doubly deficient in both caspase-1 and caspase-11 (12). This occurred because caspase-1 gene deletions were generated in 129 mice which already carried a mutation in the adjacent Casp11 gene. Caspase-11 is known as the noncanonical inflammasome and is activated through mechanisms independent of the canonical inflammasome signaling cascade (13). Activation of caspase-11 has been shown to occur in IFN-γ-primed mouse macrophages during C. muridarum infection through the recruitment of cytosolic guanylate-binding proteins (GBPs) (11).

As the mice used in the earlier C. muridarum genital infection study (8) were actually deficient for both caspase-1 and caspase-11, we chose to revisit the differential contribution of caspase-1 and caspase-11 in oviduct pathology during C. muridarum infection. Because no other inflammasome constituents play a significant role in pathology (6), we hypothesize that it is actually caspase-11 and not caspase-1 that contributes to the oviduct pathology during infection. To address this, we compared infection course and pathology in mice deficient for caspase-11 or caspase-1, or both caspase-1 and caspase-11, with results for wild-type (WT) mice. The results from our study reveal a greater role for caspase-11 over caspase-1 in oviduct pathology during C. muridarum infection. Our results demonstrate that noncanonical inflammasome activation by caspase-11 enhances chlamydial killing and neutrophil recruitment in the infected tissue, both of which could be mediated by enhancing inflammatory cell death.

RESULTS

Both Casp11^−/− and Casp1-11^−/− mice infected intravaginally with C. muridarum developed less oviduct pathology.

Casp11^−/− mice were genotyped (DartMouse) and found to be 99.9% similar to the C57BL/6J mouse controls (see Fig. S1 in the supplemental material). Peritoneal macrophages from Casp11^−/− and Casp1-11^−/− mice were infected ex vivo with C. muridarum under resting conditions (Fig. 1A), after a brief exposure to lipopolysaccharide (LPS) (Fig. 1B) or after IFN-γ prestimulation (Fig. 1C). Casp1-11^−/− mouse macrophages were found to be defective for IL-1β release following infection under all conditions, while infected Casp11^−/− mouse macrophages were partially competent for IL-1β release after LPS exposure (Fig. 1A and B) and fully competent for IL-1β release after IFN-γ exposure. ASC-deficient mouse macrophages were used as a control and were defective for IL-1β secretion upon chlamydial infection (12) (Fig. 1A, B, and C). Additionally, Chlamydia infection upregulated caspase-11 expression in wild-type (WT) mouse macrophages, which was significantly reduced in Stat1^−/− macrophages, suggesting a dependence on IFNAR signaling (Fig. 1D). Chlamydia-induced CXCL10 expression served as a control for IFN-β and STAT1-dependent expression in the same cells (Fig. 1E).

FIG 1.

IL-1β release and caspase-11 expression during Chlamydia infection of resting and prestimulated macrophages. Thioglycolate-elicited macrophages from WT, Casp11^−/−, Casp1-11^−/−, and Asc^−/− mice peritonea were rested for 48 to 72 h and infected with C. muridarum ex vivo at an MOI of 1. Supernatants were collected at 24 p.i. and analyzed for IL-1β release by ELISA. This was done under resting conditions (A) or following a 6-h prestimulation with E. coli LPS (B) or with IFN-γ (C). Macrophages elicited from WT and Stat1^−/− mice were infected with C. muridarum at an MOI of 0.3 and 1.0 in the presence or absence of LPS prestimulation. At 24 h p.i., RNA was prepared, the relative expression of the genes for caspase-11 (D) and CXCL10 (E) was analyzed by RT-qPCR, and the results are expressed as the relative fold change in expression normalized to the level of actin gene expression. UI, uninfected. Significance was determined by one-way ANOVA. **, P < 0.001; ***, P < 0.0001; ns, not significant.

To compare the contributions of caspase-1 or caspase-11 to the genital infection course and pathology, WT, Casp11^−/−, and Casp1-11^−/− mice were infected intravaginally with C. muridarum. The infection course, as assessed by lower genital tract bacterial shedding, was found to be comparable between WT, Casp11^−/−, and Casp1-11^−/− mice, with all three groups clearing infection between days 21 and 25 (Fig. 2A). Casp11^−/− and Casp1-11^−/− mice showed a moderate delay in infection clearance in the lower genital tract, but these differences were not statistically significant. Gross pathology was assessed by the presence of oviduct hydrosalpinx (HS) during sacrifice on day 45 postinfection (p.i.). Both Casp11^−/− and Casp1-11^−/− mice had a reduced incidence of HS relative to WT mice, but no significant differences were observed between the two groups of gene deletion mice (Fig. 2B). Histopathological assessment of hematoxylin-eosin (H&E)-stained genital tract sections showed that oviducts from WT mice had significant dilation and thinning of the walls, while oviducts from Casp11^−/− mice retained their architecture (Fig. 2C). Histopathological scoring of H&E-stained genital tract sections at day 45 p.i. using high-power magnification showed reduced dilation in both the oviducts and the uteri of Casp11^−/− and Casp1-11^−/− mice compared to WT mice (Fig. 3A and B). The uteri of Casp1-11^−/− mice were less dilated than those of Casp11^−/− mice, with no scores being above 1 (Fig. 3B). No significant differences in cells of the acute phase of infection (polymorphonuclear leukocytes [PMN]) were observed in the oviducts between the groups, though a slight increase was observed in the uteri of Casp11^−/− mice (Fig. 3C and D). Further, a small decrease in the number of cells of the chronic phase of infection (mononuclear cells) was observed in the oviducts of Casp11^−/− and Casp1-11^−/− mice compared to WT mice (Fig. 3E and F). Plasma cells were not different in the oviducts of the 3 groups, although they were significantly reduced in Casp1-11^−/− mouse uteri (Fig. 3G and H). Likewise, edema was not different in the oviducts of the WT and Casp11^−/− mice but was significantly reduced in Casp1-11^−/− mouse uteri (Fig. 3I and J). Overall, our results indicate that in 70% of mice, caspase-11 deficiency was sufficient to significantly protect from hydrosalpinx and dilation. Additionally, the significant reduction in plasma cells in Casp1-11^−/− mouse uterine horns suggests that the combined absence of both caspase-1 and caspase-11 is likely more protective.

FIG 2.

Casp11^−/− and Casp1-11^−/− mice infected genitally with C. muridarum develop reduced oviduct pathology. Mice were infected intravaginally with 3 × 10⁵ IFUs of C. muridarum. (A) The infection course was determined by measurement of the number of IFUs in vaginal swabs, and the results of a representative experiment of three independent experiments are shown for WT (n = 10), Casp11^−/− (n = 10), and Casp1-11^−/− (n = 10) mice. Data are represented as the mean ± standard error of the mean (SEM), and significance was determined by two-way ANOVA. (B) The percentages of oviducts from WT (n = 21), Casp11^−/− (n = 21), and Casp1-11^−/− (n = 8) mice forming hydrosalpinx (HS) at day 45 postinfection are shown. Data are combined from 3 independent experiments for WT and Casp11^−/− mice and 2 independent experiments for Casp1-11^−/− mice. The number of oviducts with HS/total number of mice tested are represented over the bars, and significance was determined by Fisher’s exact test. (C) H&E images of oviducts from 5 WT and 5 Casp11^−/− mice on day 45 postinfection. Oviducts (OD) next to ovaries (OV) are indicated. *, dilations.

FIG 3.

Histopathology scores for oviducts and uterine horns from infected WT, Casp11^−/−, and Casp1-11^−/− mice at day 45 postinfection. Oviducts and uteri were scored for dilation (A, B), acute-phase cells (neutrophils) (C, D), chronic-phase cells (mononuclear cells) (E, F), plasma cells (G, H), and edema (I, J) by use of the grading system described in Materials and Methods. Each point represents the score (indicated on the y axis) for a single oviduct or uterine horn. Data are presented for scores obtained from three independent experiments. Significance was determined by the Kruskal-Wallis multiple-comparison test.

Oviduct pathology in Casp1-11^−/− mice was restored with expression of the Casp11 transgene.

To further assess the independent role of caspase-1, Casp1-11^−/− mice expressing Casp11 as a transgene (Casp1-11^−/− Casp11^Tg mice) (12) were used, as mice deficient for Casp1 alone were unavailable. Casp1-11^−/− Casp11^Tg mice have caspase-11 expression restored by the insertion of a Casp11 transgene and are functionally expected to behave like Casp1^−/− mice, and they are referred to as such from this point forward. Casp1^−/− mouse macrophages expressed caspase-11 but were defective for caspase-1-mediated IL-1β release upon Chlamydia infection (data not shown). The infection course in these mice was comparable to that in WT mice (Fig. 4A). Further, the mice showed no reduction in their incidence of hydrosalpinx compared to WT mice (Fig. 4B). Histological scoring of oviducts showed some reduction in oviduct dilation in the Casp1^−/− mice, but it was not significant (Fig. 4C). However, none of the Casp1^−/− oviducts had a score of grade 4, indicating some contribution of the caspase-1 pathway to pathology.

FIG 4.

Oviduct pathology in infected Casp1-11^−/− mice is restored by expression of the Casp11 transgene. Mice were infected intravaginally with 3 × 10⁵ IFUs of C. muridarum. (A) The infection course was determined by measurement of the number of IFUs in vaginal swabs of WT (n = 6) and Casp1-11^−/− Casp11^Tg (n = 4) mice. Data are representative of those from 3 independent experiments and represented as the mean ± SEM. Significance was determined by two-way ANOVA. (B) The percentages of oviducts from WT (n = 11) and Casp1-11^−/− Casp11^Tg (n = 20) mice forming hydrosalpinx (HS) at day 45 postinfection are indicated. Data from 3 independent experiments were combined. The number of mice with HS/total number of mice tested are represented over the bars, and significance was determined by Fisher’s exact test. (C) Median oviduct dilation scores were determined following histopathological examination, and significance was determined by the Kruskal-Wallis multiple-comparison test.

Casp11^−/− mice have reduced PMN infiltration and increased CD4 T cell infiltration in their oviducts.

To address the potential mechanism by which Casp11^−/− mice could be protected from oviduct dilation, we first examined the levels of inflammatory cytokines (tumor necrosis factor alpha [TNF-α], IL-1β, IL-1α, and IL-18) in cervical genital secretions of WT, Casp11^−/−, and Casp1-11^−/− mice. In Casp1-11^−/− mice, we saw a significant increase in TNF-α by 4 days p.i. and decreased IL-18 by 6 days p.i. (Fig. S3). IL-1α is released from dying cells and could serve as an indicator for caspase-1- and caspase-11-mediated cell death. However, the IL-1α levels in genital secretions were not significantly different between the WT and Casp11^−/− mice (Fig. S3). As none of these changes were associated with the absence of caspase-11 alone (Fig. S3), they provide limited clues as to what could be causing the differences in pathology in the upper genital tract of Casp11^−/− mice. Next, we examined the oviducts on days 14 and 21 p.i. for immune cell infiltration to determine if the reduction in pathology was associated with a decrease in their recruitment (Fig. 5A). Ly6G⁺ neutrophils were significantly reduced in Casp11^−/− mouse oviducts on day 14. However, by day 21 p.i., when overall neutrophil numbers were reduced, this difference disappeared. These data suggest a role for caspase-11 in neutrophil recruitment to the oviduct earlier in infection, which in turn could contribute to long-term pathology. No significant differences in the number of F4/80⁺ macrophages were observed. In contrast to the neutrophils, CD4 T cell numbers were not different between WT and Casp11^−/− mice on day 14, and CD4 T cells made up roughly 10% of the overall cell count. However, while these levels dropped in WT mouse oviducts by day 21, they stayed close to 10% in the Casp11^−/− mouse oviducts, creating a significant difference in CD4 T cell numbers as infection was waning, with significantly more CD4 T cells being found in Casp11^−/− mouse oviducts. To determine if the chlamydial tissue burden corresponded to the observed differences in immune cell populations, a portion of the tissue homogenate was analyzed for chlamydial DNA. No differences in chlamydial burden were observed between the WT and Casp11^−/− mouse cervix, uterus, and oviduct at day 14 p.i. (data not shown). At day 21 p.i., the bacterial burden was similar in the cervix of WT and Casp11^−/− mice. However, Casp11^−/− mouse oviducts had significantly increased chlamydial DNA, as determined by PCR, compared to WT mouse oviducts on day 21 postinfection, indicating prolonged bacterial infection in the oviducts of Casp11^−/− mice. A similar trend of an increased chlamydial burden at day 21 was also observed in Casp11^−/− mouse uteri. These data suggest that caspase-11 helps control infection in the oviducts and that, in its absence, infection can persist longer.

FIG 5.

Neutrophil infiltration is reduced in the oviducts of Casp11^−/− mice, but CD4 T cell infiltration is increased and accompanied by increased infection. Oviducts were harvested from infected mice at day 14 and day 21 p.i., and single-cell suspensions were generated for flow cytometry analysis. (A) Cells were gated for CD45 cells, and the proportions of Ly6G⁺ neutrophils, F4/80⁺ macrophages, and CD4 T cells in the CD45⁺ population were determined. Data are presented as the percentage of total live cells. (B) One-tenth of the cell suspension was frozen and processed for quantitative PCR for the MOMP gene. Data for day 21 p.i. are represented as the number of log₁₀ copies of the MOMP gene per milliliter of tissue homogenate. Significance was calculated by Student's t test.

DISCUSSION

Previous studies showed that mice deficient for caspase-1 were inadvertently also deficient for caspase-11. These doubly deficient mice were protected from oviduct pathology from Chlamydia infection, and this protection was attributed to the absence of caspase-1 (8). In the study described in this report, we used mice lacking caspase-11 and compared them to mice lacking both caspase-1 and -11 to reveal that caspase-11, rather than caspase-1, plays a role in oviduct pathology during Chlamydia infection. Further, Casp1-11^−/− mice expressing a caspase-11 transgene reverted to WT oviduct pathology rates, demonstrating that caspase-11 activation plays a larger role in pathology than caspase-1.

To determine if the reduction in oviduct pathology in Casp11^−/− mice was a result of reduced chlamydial ascension to the oviduct, the chlamydial burden was assessed in the lower and upper genital tract. No difference in chlamydial shedding was observed in the cervical swabs from Casp11^−/− mice and WT mice, but we did observe increased chlamydial burdens in the upper genital tract of Casp11^−/− mice on day 21 postinfection (p.i.). These data suggest a surprising role for caspase-11 activation in reducing the chlamydial burden in the oviduct, which was not evident in the cervix. Caspase-11 likely contributes to intracellular bacterial infection control through multiple mechanisms. Caspase-11 activation is highly protective against cytosolic bacteria, including Burkholderia and cytosolic invading mutants of Salmonella and Legionella, by direct sensing of LPS (13). Additionally, caspase-11 promoted lysosomal fusion of the vacuolar pathogen Legionella pneumophila via actin-cytoskeletal rearrangement (14). Outer membrane vesicles from Gram-negative bacteria containing LPS can also activate caspase-11 (15). IFN-γ-inducible guanylate-binding proteins (GBPs) enhanced caspase-11 activation and pyroptosis during infection with cytosol-invasive mutants of L. pneumophila (16). The model in which GBP-dependent caspase-11 activation occurs independently of vacuolar lysis is supported by several studies (reviewed in reference 17). To determine which of these mechanisms contributes to chlamydial control, Finethy et al. investigated the contribution of GBPs in the lysis of chlamydial inclusions and found that GBPs failed to bind to C. muridarum inclusions in infected macrophages, although they promoted pyroptosis by caspase-1 and caspase-11 activation (11). Therefore, caspase-11 is less likely to control chlamydial infection in the oviduct by direct intracellular bacterial killing. Instead, caspase-11 could act via pyroptotic cell death as an innate mechanism to reduce host cell availability. The contribution of caspase-11 in controlling ascending infection in the upper reproductive tract was significant, though marginal, as infection was not observed in Casp11^−/− mouse oviducts by day 45 (data not shown). Caspase-11 expression can be induced in both epithelial and recruited immune cells during type I and type II IFN exposure (18, 19), and STING activation can activate inflammasome and pyroptotic cell death (20). C. muridarum infection of macrophages induced caspase-11 expression, likely via IFN-β production and IFNAR signaling, as caspase-11 induction was inhibited in STAT1-deficient macrophages. In our study, it remains unclear which cell type contributes to caspase-11-mediated bacterial control. As C. muridarum can survive in macrophages and epithelial cells, we expect that both cell types can potentially contribute to caspase-11-mediated chlamydial control.

Oviduct pathology during Chlamydia infection has consistently been shown to be associated with neutrophil recruitment (6, 21–24). We observed that the reduced pathology in Casp11^−/− mice was associated with reduced neutrophil recruitment to the oviduct at day 14 p.i. Given that infection leads to epithelial chemokine production to bring neutrophils to the site of infection (25), one would expect increased infection in the oviduct to be associated with increased neutrophil recruitment. This is not what we have seen in Casp11^−/− mice, where we observed a delayed clearance of bacteria in the oviduct and reduced neutrophil recruitment. This disconnect between reduced pathology and delayed bacterial clearance has been previously described in IL-1R1-knockout (KO) mice (6), demonstrating that once an immune response is initiated by infection, inflammation can perpetuate in a feed-forward loop to lead to more inflammation and is a driving force in pathology. As such, it is also possible that we see increased bacterial burdens in Casp11^−/− oviducts because the epithelium is left intact in the absence of damage-causing neutrophils. Infection was also accompanied by increased CD4 T cells in the genital tracts of Casp11^−/− mice, likely a result of prolonged infection in the upper genital tract of Casp11^−/− mice. However, it is also possible CD4 T cell populations are increased because of unregulated expansion, as caspase-11 has been shown to be a negative regulator of T cell receptor activation and can limit T cell expansion (26).

Precisely, how is caspase-11 activated during Chlamydia infection, and how does its activation contribute to oviduct pathology? Caspase-11 can be directly activated by Escherichia coli LPS (27). Whether chlamydial LPS can activate caspase-11 is unclear, as purified chlamydial LPS is at least 10 times less bioactive than typical endotoxins (28). A recent study showed that chlamydial LPS was less effective in caspase-11 activation (29). During in vivo infection, caspase-11 activation could also be caused by endogenous damage-associated molecular pattern molecules (DAMPs), such as oxidized phospholipids (30). Further, the impact of caspase-11 activation on pathology could be a result of amplified inflammation from the release of DAMPs, like IL-1α, which can enhance neutrophil recruitment and lead to tissue damage (reviewed in reference 31). Caspase-11 has been shown to play a significant role during Klebsiella pneumoniae infection by recruiting neutrophils and controlling bacterial burdens (32), largely through the cleavage and release of IL-1α. A similar association of caspase-11 and IL-1α and their role in neutrophil recruitment were observed in Legionella pneumophila and Yersinia pseudotuberculosis infections (33). While we observed differences in oviduct neutrophil levels, we did not detect any differences in IL-1α levels in cervical secretions (see Fig. S2 in the supplemental material) or in supernatants obtained from the oviduct tissue homogenates (data not shown) between WT and Casp11^−/− mice. However, IL-1α is typically released and acts in a localized fashion, and it is possible that our assay was not specific for the measurement of concentrated changes in areas of infection. In addition to neutrophils, CD8 T cells are associated with oviduct pathology (4). We did not observe any changes in CD8 T cell numbers in the oviducts of Casp11^−/− mice. However, caspase-11 activation in CD8 T cells has been reported (26) and could alter their effector function to influence pathology. Further studies are required to confirm the contribution of CD8 T cells in caspase-11-mediated pathology.

Caspase-11 activation is traditionally associated with pyroptotic cell death and the release of neutrophil-recruiting cytokines. Though the specific mechanisms taking place during C. muridarum infection have yet to be fully elucidated, we propose that caspase-11 contributes to pathology by enhancing cell death, which recruits neutrophils. Further, type I IFNs can upregulate caspase-11 expression in both epithelial and recruited immune cells (18, 19). We have previously shown that IFNAR^−/− mice develop less pathology (5). We speculate that IFNAR signaling leads to caspase-11 induction, contributing to oviduct pathology. Additionally, caspase-11 activation and pyroptotic cell death, which result in DAMP release, could contribute to a feed-forward loop of inflammation and enhance oviduct pathology. Overall, we propose a unifying hypothesis that cell death is a major driver of oviduct pathology and that while increasing cell death in the oviduct aims to reduce infection, it also results in long-term collateral damage to the tissue to cause pathology.

MATERIALS AND METHODS

Chlamydial stocks.

Chlamydia muridarum strain Nigg was propagated in L929 cells, and the infectious titer was determined by calculating the number of inclusion-forming units (IFUs) from infection of a fresh McCoy cell monolayer as described earlier (6). Chlamydial staining was done using mouse antichlamydial immune sera, followed by anti-mouse immunoglobulin-Alexa Fluor 488 secondary antibody (Invitrogen).

Mouse strains.

Mice deficient for caspase-11 (Casp11^−/− mice) and both caspase-1 and -11 (Casp1-11^−/− mice) and Casp1-11^−/− mice expressing the Casp11 transgene (Casp1-11^−/− Casp11^Tg mice, which are essentially Casp1^−/− mice) (12) were obtained from Genentech Inc. following a material transfer agreement for transfer to the University of North Carolina, Chapel Hill. Casp11^−/− mice were generated in the C57BL/6J mouse background; Casp1-11^−/− mice were originally on the 129 mouse background but were backcrossed to C57BL/6J mice for over 10 generations. Casp1-11^−/− Casp11^Tg mice were generated by Genentech Inc. (principal investigator, Vishva Dixit) by expressing the caspase-11 transgene in Casp1-11^−/− mouse blastocysts (34, 35) and were used as a proxy for caspase-1-deficient mice. Stat1^−/− mouse macrophages were a kind gift from John Alcorn (University of Pittsburgh). Casp11^−/− mice were tested for their genetic background by genotyping (DartMouse). All control mice used in this study were C57BL/6J mice (stock number 000664; The Jackson Laboratory, Bar Harbor, ME) and are referred to as the wild type (WT). Female mice aged 8 to 12 weeks were used for all experiments, which were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of North Carolina, Chapel Hill.

Genital infections of mice and collection of swabs and sponges.

At 7 days prior to infection, each mouse received 2.5 mg of medroxyprogesterone acetate (Depo-Provera) subcutaneously in 100 μl of phosphate-buffered saline (PBS). A week later, the mice were anesthetized with pentobarbital (Nembutal; 240 to 250 μl of a 5-mg/ml stock) and infected by administering 3 × 10⁵ IFUs of C. muridarum in 20 μl SPG buffer (250 mM sucrose, 10 mM sodium phosphate, 5 mM l-glutamate) into the vaginal vault. Chlamydial shedding was determined at various times postinfection, and the number of IFUs was determined as previously described (6). Genital secretions were collected for cytokine analysis using DeRoyal ear wicks from days 2 through 10 postinfection, as described previously (36).

Histopathology.

Mice were sacrificed on day 45 postinfection, and genital tract tissues were removed and assessed for hydrosalpinx incidence. Tissues were fixed in 10% formalin and embedded in paraffin. Longitudinal sections (4 μm) were stained with hematoxylin and eosin and evaluated by a pathologist blind to the experimental design. The cervix, uterine horn, oviduct, and mesosalpinx were assessed for the presence of neutrophils (indicating the acute phase of infection), plasma cells, mononuclear cells (indicating the chronic phase of infection and consisting of cells other than plasma cells), and fibrosis. The right and left oviducts, mesosalpinx, and uterine horns were each evaluated individually. A five-tier, semiquantitative scoring system was used to quantitate the inflammation and fibrosis by observing H&E-stained sections under high power, where a score of 0 was normal, a score of 1+ indicated rare foci (minimal presence) of the parameter, a score of 2+ indicated scattered (1 to 4) aggregates or a mild diffuse increase in the parameter, a score of 3+ indicated numerous aggregates (>4) or moderate diffuse or confluent areas of the parameter, and a score of 4+ indicated severe diffuse infiltration or a confluence of the parameter. Luminal distention of the uterine horns, granulomas, and dilation of the oviducts were graded from 1 to 4, with grade 1 representing minimal luminal distention above normal and grade 4 representing a peak severity or frequency of the parameter.

Cytokine quantification.

Protein levels of IL-1α and IL-1β in genital secretions and culture supernatants were determined using enzyme-linked immunosorbent assay (ELISA) kits (BioLegend). Genital secretion eluates were diluted 1:20 in PBS for the IL-1α and IL-1β ELISAs and tested as is for TNF-α and IL-18. Optical densities were measured using a BioTek plate reader at 450 nm for quantification, and the absolute concentrations of diluted samples were determined using a standard curve.

Macrophage isolation and in vitro infection.

Thioglycolate-induced peritoneal macrophages were obtained and cultured as described previously (37) for in vitro infections. Following isolation from the mouse and plating into wells of 24-well tissue culture plates at a density of 8E–5/well, the cells were rested in macrophage complete medium (37) for 48 to 72 h before any experiment. Alternatively, some macrophages were frozen (90% fetal bovine serum [FBS], 10% dimethyl sulfoxide [DMSO]), stored in liquid nitrogen, and treated as described above upon thawing. C. muridarum was added at a multiplicity of infection (MOI) of 1 to the cells in the tissue culture plates, and the cells were centrifuged at 1,690 × g at 37°C for 1 h. Macrophages were prestimulated with E. coli LPS at 100 ng/ml or IFN-γ (100 ng/ml) at 6 h before infection, and after 1 h of infection they were spun to wash away the LPS and IFN-γ, as described earlier (38). Medium was replaced with fresh complete medium after centrifugation and incubated at 37°C for 24 h. To confirm infection by the observation of inclusions, 8E−5 macrophages were seeded onto coverslips, infected in parallel, and fixed with methanol for 10 min at room temperature at 24 h p.i. Coverslips were stained for chlamydial inclusions as described earlier and mounted with the ProLong antifade reagent (Invitrogen). Supernatants were collected for ELISA, and RNA was prepared from cell lysates using Qiagen RNeasy kits. The reverse transcription (RT) reaction and quantitative PCR (qPCR) were carried out as described earlier (39). The primers used for amplifying caspase-11 cDNA were GGAGAAATGTGGATCAGAGAGTC (sense) and TGTAGAGTAGAAGGCAATGAAGTC (antisense), and those used for amplifying CXCL10 were CCAGCCGTGGTCACATCAG (sense) and ACCTCCACATAGCTTACAGTACAG (antisense).

Flow cytometric analysis of genital tract cells from infected mice.

Cervical tissue, uterine horns, and both oviducts were excised from the genital tract of an individual mouse and minced. Cervical tissue and uterine tissue were incubated separately with 1 ml of type I collagenase (1 mg/ml; catalog number C0130; Sigma) for 20 min at 37°C with shaking. After addition of EDTA (0.5 M) to neutralize the reaction, 100 μl of cell suspension was diluted with an equal volume of SPG buffer and frozen at −80°C for later determination of the numbers of IFUs and for PCR, as described earlier. The remaining cells were passed through a mesh filter (mesh size, 70 μm) by grinding the tissue using the plunger of a 3-ml syringe. The cells were rinsed with Dulbecco modified Eagle medium with 1% FBS and collected in 10 ml. Cell suspensions were centrifuged at 300 × g at 4°C and then resuspended in 10 ml of PBS. Cells were centrifuged and resuspended (to 2 × 10⁵ to 5 × 10⁵ cells/25 μl) in florescence-activated cell sorting (FACS) buffer (2 mM 0.5 M EDTA, 1% bovine serum albumin, 1× PBS) containing Fc Block (25 μg/ml Fc Block; BD) for 10 min, washed twice in FACS buffer, and then stained for individual cell surface markers or isotype controls (5 μg/ml) in a 96-well V-bottom plate for 30 min on ice in the dark. The following monoclonal antibodies (BD Biosciences) were used: CD45 on peridinin chlorophyll protein-Cy5.5 (clone 30-11), Ly6G on fluorescein isothiocyanate (clone IA8), F4/80 on allophycocyanin (APC) (clone BMB), CD11b on phycoerythrin (PE)-Cy7, CD8a on APC-Cy7, CD3 on V450, and CD4 on PE (clone RM-4). A final mix with 25 μl of the diluted antibodies in FACS buffer was added to each well for staining. Ultra-Comp beads (eBioscience) were used for compensation. Cells were washed with 200 μl of cold FACS buffer and centrifuged at 300 × g for 5 min at 4°C. Cells were resuspended in 150 μl of FACS buffer, and data were acquired on a Cyan ADP (Beckman Coulter) or LSR II (BD Biosciences) flow cytometer. Data were analyzed using FlowJo software (TreeStar Inc.).

Genomic DNA extraction and quantitative PCR.

A Zymo Quick-DNA microprep kit was used to extract genomic DNA from the chlamydial stock to generate a standard curve or from 100 μl of tissue homogenate (1/10 of the FACS homogenate). Standard curves were generated by dilution of genomic DNA and converting the concentration to copy number, which was equal to (the amount [in nanograms] of genomic DNA · 6.022 × 10²³)/(length · 1 × 10⁹ [a conversion factor] · 650 [molecular weight, in base pairs]). Tenfold serial dilutions were made, and a standard curve was generated using major outer membrane protein (MOMP)-specific primers by plotting the number of log₁₀ DNA copies on the x axis versus the threshold cycle values on the y axis. An ompA (MOMP) quantitative PCR was performed using primers Nigg MOMP-1F (GAGCCATAATTGCTACAG) and Nigg MOMP-1R (CCTCTAATAAGATTGTATAATTAGC). The amplicon size of this PCR product is 107 bp. PCR cycling conditions involved initial denaturation at 95°C for 180 s, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Melt curve analyses were recorded from 65°C to 95°C with an increment of 0.5°C for every 5 s for each single peak. The MOMP DNA copy numbers in the DNA from tissue homogenate samples were determined from the standard curve. These values were log₁₀ transformed after factoring the corresponding dilutions to give values of the number of log₁₀ MOMP copies per milliliter.

Statistical analysis.

Unless otherwise indicated, experiments for each representative datum were performed at least three independent times. Statistical comparisons between the two mouse strains for IFU quantification or cytokine analysis over the course of infection were made by two-way analysis of variance (ANOVA) with the post hoc Tukey test as a multiple-comparison procedure. Significant differences in the frequency of hydrosalpinx between the two groups were determined by Fisher’s exact test. Pathology scores were analyzed by the Kruskal-Wallis one-way ANOVA on ranks. One-way ANOVA or an unpaired Student's t test was used for determining the significance between individual groups of the flow cytometry data.