Gastrointestinal Chlamydia-Induced CD8⁺ T Cells Promote Chlamydial Pathogenicity in the Female Upper Genital Tract

ABSTRACT

Chlamydia is known to both ascend to the upper genital tract and spread to the gastrointestinal tract following intravaginal inoculation. Gastrointestinal Chlamydia was recently reported to promote chlamydial pathogenicity in the genital tract since mice intravaginally inoculated with an attenuated Chlamydia strain, which alone failed to develop pathology in the genital tract, were restored to develop hydrosalpinx by intragastric coinoculation with wild-type Chlamydia. Gastrointestinal Chlamydia promoted hydrosalpinx via an indirect mechanism since Chlamydia in the gut did not directly spread to the genital tract lumen. In the current study, we further investigated the role of CD8⁺ T cells in the promotion of hydrosalpinx by gastrointestinal Chlamydia. First, we confirmed that intragastric coinoculation with wild-type Chlamydia promoted hydrosalpinx in mice that were inoculated with an attenuated Chlamydia strain in the genital tract 1 week earlier. Second, the promotion of hydrosalpinx by intragastrically coinoculated Chlamydia was blocked by depleting CD8⁺ T cells. Third, adoptive transfer of gastrointestinal Chlamydia-induced CD8⁺ T cells was sufficient for promoting hydrosalpinx in mice that were intravaginally inoculated with an attenuated Chlamydia strain. These observations have demonstrated that CD8⁺ T cells induced by gastrointestinal Chlamydia are both necessary and sufficient for promoting hydrosalpinx in the genital tract. The study has laid a foundation for further revealing the mechanisms by which Chlamydia-induced T lymphocyte responses (as a 2nd hit) promote hydrosalpinx in mice with genital Chlamydia-triggered tubal injury (as a 1st hit), a continuing effort in testing the two-hit hypothesis as a chlamydial pathogenic mechanism.

INTRODUCTION

Chlamydia trachomatis is a leading infectious cause of infertility due to its ability to induce tubal inflammation/adhesion/fibrosis/hydrosalpinx in women (1–4). However, the mechanisms by which C. trachomatis induces long-term sequelae in the upper genital tract remain unclear. The mouse-adapted species Chlamydia muridarum has been used to investigate C. trachomatis pathogenesis because of its ability to induce long-lasting tubal fibrosis/hydrosalpinx in mice (5–9). Following intravaginal inoculation, C. muridarum ascends to the oviduct to induce tubal inflammation (10) that may both clear chlamydial infection and damage tubal epithelia, triggering tissue-repairing responses, including transient fibrosis. In many cases, mouse tubal fibrosis continues despite the clearance of the initial chlamydial infection, leading to oviduct lumen occlusion/hydrosalpinx/infertility (5, 7, 9). The mouse model with C. muridarum induction of hydrosalpinx has been useful for investigating chlamydial pathogenic mechanisms. For example, a chlamydial plasmid was found to be an important pathogenic determinant since C. muridarum depleted of the plasmid was no longer able to induce hydrosalpinx (11–14). It was also found that mouse CD8⁺ T cells promoted (15, 16) while CD4⁺ T cells prevented (17) C. muridarum induction of hydrosalpinx. However, it remains unknown why and how long-lasting tubal fibrosis is still maintained long after oviduct infection is cleared.

Chlamydia is also routinely detected in the gastrointestinal (GI) tracts of humans (18–22) and animals (23–27). In vivo imaging of mice (28) revealed that genital C. muridarum spread to and colonized the GI tract for long periods (24). Spreading might occur via blood circulation (24, 26, 29, 30). Systemically disseminated organisms were cleared within 2 to 3 weeks, and only those that reached the GI tract lumen could persist for long periods (26, 27). Although GI Chlamydia has been hypothesized to serve as a reservoir for autoinoculation of the genital tract lumen to promote chlamydial pathogenicity (25, 31), mice intragastrically (i.g.) inoculated with C. muridarum failed to autoinoculate their genital tract lumen (27) and did not develop any significant pathology in either the genital tract or the GI tract (32, 33). The question was whether Chlamydia in the mouse GI tract could affect the pathogenicity of genital Chlamydia.

A coinoculation experimental approach was successfully developed to determine whether GI tract Chlamydia could affect the pathogenicity of genital Chlamydia in mice (34). In coinoculated mice, the contributions of Chlamydia in the genital tract versus the GI tract to the development of hydrosalpinx can be measured separately. This is because although C. muridarum mutants such as plasmid-free (Pf) C. muridarum can still ascend to the upper genital tract, they are attenuated in both inducing hydrosalpinx and spreading to the GI tract. Thus, Pf C. muridarum may be able to cause only genital damage without triggering significant responses in the GI tract. Furthermore, intragastrically inoculated wild-type (Wt) C. muridarum is restricted to the GI tract without contaminating the genital tract lumen. Thus, intragastrically inoculated Wt C. muridarum cannot directly cause tissue injury in the genital tract. It was found that although intravaginal inoculation with Pf C. muridarum alone failed to induce hydrosalpinx, intragastric coinoculation with Wt C. muridarum successfully rescued Pf C. muridarum to induce hydrosalpinx. It is worth noting that intragastric coinoculation was applied 1 week after intravaginal inoculation in order to best mimic the kinetics of natural spread and avoid interference of the gastrointestinal responses with the genital infection course (33). Gastrointestinal Wt C. muridarum was restricted to the GI tract without spreading to the genital tract lumen when both Pf and Wt C. muridarum organisms were tracked simultaneously but separately. Thus, gastrointestinal Chlamydia must use an indirect mechanism to promote hydrosalpinx development in the genital tract. For example, immune responses induced by gastrointestinal Chlamydia may be recruited to the upper genital tract tissue to promote pathology. These observations and analyses led to the proposal of a two-hit model for partially explaining chlamydial pathogenicity in the genital tract (35). Genital Pf C. muridarum may cause tubal epithelial damage as a 1st hit since Pf C. muridarum is still able to ascend to the upper genital tract and infect tubal epithelial cells (11), while gastrointestinal Wt C. muridarum organisms may induce profibrotic immune responses as a 2nd hit to promote hydrosalpinx. The goal of the current study is to test whether gastrointestinal Chlamydia-induced T lymphocyte responses can act as a 2nd hit in the two-hit model.

In the current study, we have focused on the role of CD8⁺ T cells as a 2nd hit in promoting chlamydial pathogenicity in the upper genital tract. This is because CD8⁺ T cells have been shown to promote chlamydial induction of hydrosalpinx following intravaginal inoculation with Wt C. muridarum (15, 16). In those previous studies, although it was shown that hydrosalpinx-promoting or pathogenic CD8⁺ T cells were specific to chlamydial antigens, it was unknown whether these CD8⁺ T cells were induced by Chlamydia in the genital tract or GI tract. Since Wt C. muridarum is known to spread from the genital tract to the GI tract (24), it is possible that the chlamydial organisms that have arrived in the GI tract may be responsible for inducing pathogenic CD8⁺ T cells. To determine whether CD8⁺ T cells induced by GI Chlamydia can promote genital tract pathology, we continued to use the coinoculation mouse model (34). We first validated that intragastric coinoculation with Wt Chlamydia indeed promoted hydrosalpinx in mice intravaginally inoculated with the attenuated Pf Chlamydia strain. Next, anti-CD8 antibody was applied to the coinoculated mice to deplete CD8⁺ T cells. It was found that the depletion of CD8⁺ T cells blocked the development of hydrosalpinx in the coinoculated mice, indicating the necessity of CD8⁺ T cells for gastrointestinal Chlamydia to promote hydrosalpinx. Furthermore, when CD8⁺ T cells induced by gastrointestinal Chlamydia were adoptively transferred to recipient mice that were intravaginally inoculated with only the attenuated Pf Chlamydia strain, the recipient mice regained the ability to develop hydrosalpinx. However, when CD8⁺ T cells harvested from mice intravaginally inoculated with only the attenuated Pf Chlamydia strain were used as donor cells, the recipient mice still failed to develop hydrosalpinx. These observations have indicated that CD8⁺ T cells are sufficient for gastrointestinal Chlamydia to promote the pathogenicity of genital Chlamydia. Thus, the current study has demonstrated both the necessity and sufficiency of CD8⁺ T cells induced by gastrointestinal Chlamydia for promoting hydrosalpinx in the mouse genital tract. These results have laid a foundation for further revealing the mechanisms by which gastrointestinal Chlamydia-induced T lymphocytes provide a 2nd hit for promoting hydrosalpinx in mice with tubal epithelial injury (as a 1st hit).

RESULTS

The promotion of hydrosalpinx by GI Chlamydia is dependent on CD8⁺ T cells in coinoculated mice.

Although intravaginal inoculation with Pf C. muridarum alone failed to induce any significant hydrosalpinx in the female mouse upper genital tract (11, 12), it was recently found that intragastric (i.g.) coinoculation with Wt C. muridarum enabled Pf C. muridarum-infected mice to develop hydrosalpinx (34). Since gastrointestinal Wt chlamydial organisms did not directly spread to the genital tract lumen in coinoculated mice, it was hypothesized that gastrointestinal chlamydial organisms might promote pathology in the genital tract by inducing a CD8⁺ T cell-mediated hydrosalpinx-promoting response (35), which is supported by the previous finding that CD8⁺ T cells are necessary for promoting chlamydial pathogenicity in the female upper genital tract (15, 16).

To determine whether gastrointestinal Chlamydia can also induce CD8⁺ T cells for promoting chlamydial induction of hydrosalpinx in the mouse genital tract, we first used an anti-CD8 antibody to deplete CD8⁺ T cells from the coinoculated mice. As shown in Fig. 1, after the anti-CD8 antibody was applied to mice both intravaginally inoculated with Pf C. muridarum and i.g. coinoculated with Wt C. muridarum, the shedding courses of either Pf or Wt C. muridarum from either vaginal swabs or rectal swabs were similar to those of mice treated with control rat IgG. Thus, the anti-CD8 antibody treatment did not significantly alter the infectivity of the inoculated chlamydial organisms in either the genital tract or the GI tract. Since the Pf and Wt C. muridarum organisms were detected separately, it was found that intravaginally inoculated Pf C. muridarum displayed delayed and reduced spread from the genital tract to the GI tract, which is consistent with what we observed previously (32). Furthermore, Wt C. muridarum failed to spread to the genital tract lumen, as we demonstrated previously (27, 34). Nevertheless, the anti-CD8 antibody treatment significantly reduced CD8⁺ T cells compared to those of mice treated with control rat IgG (Fig. 2). The reduction in CD8⁺ T cells in the coinoculated mice was maintained throughout the infection course.

FIG 1.

Effect of antibody depletion of CD8⁺ cells on chlamydial shedding courses in coinoculated mice. Two groups of female CBA1/J mice were intravaginally infected with plasmid-free Chlamydia muridarum (Pf CM) and intragastrically coinoculated with mCherry-expressing wild-type Chlamydia muridarum (Wt CM) 7 days after intravaginal inoculation. Mice were further treated with either normal rat IgG (a and b) or anti-CD8 mAb (c and d) starting on day 7 after intravaginal inoculation and twice weekly thereafter, as indicated at the top. All mice were monitored for live chlamydial organism shedding from both vaginal (a and c) and rectal (b and d) swabs, and the titers are expressed as log₁₀ IFU per swab (y axis) on different days after intravaginal infection (x axis). Pf C. muridarum and Wt C. muridarum titers were quantitated separately. Note that the antibody treatment did not significantly alter live-organism shedding courses of either Pf C. muridarum or Wt C. muridarum (P > 0.05 by a Wilcoxon rank sum test [area under the curve]) (n = 10). Data are from two independent experiments.

FIG 2.

Monitoring the efficacy of CD8⁺ T cell depletion in coinoculated mice. Both CD8⁺ T cells and CD4⁺ T cells were counted after gating for CD3⁺ cells in the peripheral blood of all coinoculated mice (as described in the Fig. 1 legend) on days 5 and 8 and weekly thereafter after intravaginal inoculation with Pf C. muridarum. (a to d) Representative flow cytometry plots for each group, with peripheral blood samples taken on day 5 (prior to depletion for the rat IgG-treated group [a] and for the anti-CD8 antibody-treated group [c]) and on day 8 (after the 1st but prior to the 2nd depletion for the rat IgG group [b] and the anti-CD8 group [d]). (e) Ratios of peripheral blood CD8⁺ over CD4⁺ T cells (y axis). Note that anti-CD8 antibody significantly reduced CD8⁺ T cells in the peripheral blood throughout the infection course.

Importantly, the depletion of CD8⁺ T cells significantly reduced hydrosalpinx in the Chlamydia-coinoculated mice (Fig. 3). The coinoculated mice that were treated with control rat IgG developed significant hydrosalpinx at a rate of 60% (6 out of 10 mice developed hydrosalpinx) and with a mean severity score of 2.4. The gross pathology was further validated by microscopy for oviduct dilation. Significant oviduct dilation was observed, with an incidence rate of 70% and a mean severity score of 3. Since the detection of oviduct dilation under a microscope is more sensitive than naked-eye observation of hydrosalpinx, 1 more mouse was detected with oviduct dilation in addition to the 6 mice with hydrosalpinx. This observation is consistent with our recent finding that intragastric coinoculation with Wt C. muridarum is able to promote Pf C. muridarum to induce hydrosalpinx (34). However, after anti-CD8 antibody treatment, the coinoculated mice no longer developed any significant hydrosalpinx. Neither hydrosalpinx nor oviduct dilation was detected in any mice in the group. These observations demonstrate that induction of hydrosalpinx in the coinoculated mice is dependent on CD8⁺ T cells. Since genital Pf C. muridarum alone was unable to induce hydrosalpinx and it was the intragastrically coinoculated with Wt C. muridarum that was responsible for promoting hydrosalpinx, the above-described observations led us to conclude that CD8⁺ T cells were necessary for GI Chlamydia to promote chlamydial pathogenicity in the genital tract. The next question is whether GI Chlamydia-induced CD8⁺ T cells are sufficient for promoting hydrosalpinx in mice with genital inoculation with only Pf C. muridarum.

FIG 3.

Effect of CD8⁺ T cell depletion on hydrosalpinx development in coinoculated mice. The same two groups of coinoculated mice as the ones described in the legends of Fig. 1 and 2 were sacrificed for observing genital pathology both macroscopically for hydrosalpinx (left) and microscopically for oviduct dilation (right) on day 56 after intravaginal inoculation. (a and c) Only one gross image of the genital tract tissue is presented for each group, with vagina on the left and oviduct/ovary on the right. Hydrosalpinx is marked with white arrows. The magnified oviduct/ovary images are marked with the corresponding hydrosalpinx scores. (b and d) Only one microscopic image is presented for each group, with ovary and dilated oviduct (DO) (double arrowhead) (b) or normal oviduct (NO) (arrows) (d) indicated. Both hydrosalpinx (macroscopic observation) and tubal dilation (under microscopy) incidence rates and severity scores are listed below the corresponding group images. Note that anti-CD8 antibody treatment prevented all mice from developing any significant hydrosalpinx and oviduct dilation. *, P < 0.05 by Fisher’s exact test for pathology rates/incidences and a Wilcoxon rank sum test for pathology scores. Data were acquired from 2 independent experiments (n = 10).

GI Chlamydia-induced CD8⁺ T cells are sufficient for promoting genital pathology in mice intravaginally inoculated with plasmid-free Chlamydia.

Having demonstrated the necessity of CD8⁺ T cells for the promotion of hydrosalpinx by GI Chlamydia, we next tested whether CD8⁺ T cells induced by GI Chlamydia can promote hydrosalpinx in mice with intravaginal inoculation with Pf C. muridarum. Since intragastrically inoculated Wt C. muridarum is restricted only to the GI tract without spreading to the genital tract lumen (27, 33, 34), we purified CD8⁺ T cells from the spleens of mice intragastrically inoculated with Wt C. muridarum for 10 days and used these cells as donor CD8⁺ T cells. These donor cells were designated GI Wt C. muridarum-induced CD8⁺ T cells. As a control, CD8⁺ T cells were also purified from mice intravaginally inoculated with Pf C. muridarum for 10 days, and these donor cells were designated genital Pf C. muridarum-induced CD8⁺ T cells. Mice with genital infection with Pf C. muridarum alone failed to develop hydrosalpinx (11, 12), suggesting that these mice might lack the ability to induce hydrosalpinx-promoting or pathogenic CD8⁺ T cells. Thus, we expected that genital Pf C. muridarum-induced CD8⁺ T cells would fail to promote chlamydial pathogenicity in the genital tract. As shown in Fig. 4, both GI Wt C. muridarum-induced CD8⁺ T cells and genital Pf C. muridarum-induced CD8⁺ T cells were sorter purified to >93% purity before transferring the cells to the recipient mice. After adoptive transfer, all recipient mice inoculated with only Pf C. muridarum in the genital tract developed similar shedding courses of live Pf C. muridarum from either the genital tract (vaginal swabs) or the GI tract (rectal swabs) regardless of the types of CD8⁺ T cells that were transferred (Fig. 5). Live Pf C. muridarum was detected in the genital tracts of all mice for ∼6 weeks, which is consistent with the knowledge that extended shedding courses are often detected in the female genital tract of CBA1/J mice (9). Although live Pf C. muridarum spread to the GI tract, the spread was significantly delayed and reduced, as demonstrated previously (32).

FIG 4.

Isolation of CD8⁺ T cells from donor mice intravaginally inoculated with plasmid-free Chlamydia or intragastrically inoculated with wild-type Chlamydia. CD8⁺ T cells were purified from the spleen of donor CBA1/J mice intravaginally (i.vag.) inoculated with plasmid-free Chlamydia muridarum (Pf CM) (a to c) or intragastrically (i.g.) inoculated with wild-type Chlamydia (Wt CM) (d to f) for 10 days. Example flow cytometry images from each donor group are shown. Splenocytes before (total) (a and d) and after (b and e) negative selection followed by fluorescence-activated cell sorting (FACS) (c and f) as indicated at the top were counted for CD8⁺ cells (x axis). Note that CD8⁺ cells were enriched by 3-fold or more after negative selection and further purified using FACS to achieve >93% purity after gating for CD3⁺ T cells. The sorter-purified CD8⁺ T cells from both donor groups were adoptively transferred to the corresponding recipient mice. FSC, forward scatter.

FIG 5.

Effect of CD8⁺ T cell transfer on live chlamydial shedding courses in mice intravaginally inoculated with plasmid-free Chlamydia only. Two groups of female CBA1/J mice were intravaginally (i.vag.) inoculated with plasmid-free Chlamydia muridarum (Pf CM). One group of mice was adoptively transferred with CD8⁺ T cells from donor mice intravaginally inoculated with plasmid-free Chlamydia muridarum [Pf CM (i.vag.)] (a and b), and the other group was adoptively transferred with donor CD8⁺ T cells induced by intragastric (i.g.) wild-type Chlamydia [Wt CM (i.g.)] (c and d). The preparation of donor cells is described in the Fig. 4 legend. Transfer was carried three times on days 14, 21, and 28 after the recipient mice were intravaginally inoculated with Pf C. muridarum, as indicated by arrows at the top. Both vaginal (a and c) and rectal (b and d) swabs were taken on days 3 and 7 and weekly thereafter (x axis) for titrating live chlamydial organisms (log₁₀ IFU per swab) (y axis). Note that intragastrically inoculated Wt C. muridarum-induced CD8⁺ T cells failed to alter the courses of chlamydial shedding from either the genital or gastrointestinal tract of the recipient mice compared to those of recipient mice receiving intravaginal inoculation with Pf C. muridarum-induced donor CD8⁺ T cells (Wilcoxon rank sum test [area under the curve]) (n = 15). Data are from three independent experiments.

Although the two groups of recipient mice that were transferred with different types of CD8⁺ T cells developed similar courses of shedding of live Pf C. muridarum, the two groups developed significantly different levels of pathology (Fig. 6). The recipient group that received GI Wt C. muridarum-induced CD8⁺ T cells developed significant hydrosalpinx, while the other group receiving genital Pf C. muridarum-induced CD8⁺ T cells failed to develop any significant hydrosalpinx. GI Wt C. muridarum-induced CD8⁺ T cells promoted the development of hydrosalpinx in 8 of the 15 recipient mice, with a mean severity score of 2.3, while genital Pf C. muridarum-induced CD8⁺ T cells (as a control) failed to promote the development of hydrosalpinx in any mice. The above-described observations were validated at the microscopic level. Oviduct dilation was detected in 9 of 15 mice receiving GI Wt C. muridarum-induced CD8⁺ T cells, with a mean severity score of 2.7. However, none of the 15 mice receiving the control CD8⁺ T cells developed any significant oviduct dilation. Thus, CD8⁺ T cells induced by GI Wt C. muridarum but not genital Pf C. muridarum are sufficient for promoting the recipient mice to develop significant hydrosalpinx. The recipient mice that did not receive pathogenic CD8⁺ T cells would not develop hydrosalpinx by themselves since they were intravaginally inoculated with only Pf C. muridarum.

FIG 6.

Effect of CD8⁺ T cell transfer on hydrosalpinx development in mice intravaginally inoculated with plasmid-free Chlamydia only. Two groups of recipient mice receiving donor CD8⁺ T cells induced by either intravaginal Pf C. muridarum or i.g. Wt C. muridarum were sacrificed on day 56 after the recipient mice were intravaginally inoculated with Pf C. muridarum for observing genital pathology both macroscopically for hydrosalpinx (a and c) and microscopically for oviduct dilation (b and d). (a and c) Only one gross image of the genital tract tissue is presented for each group, with vagina on the left and oviduct/ovary on the right. Hydrosalpinx is marked with white arrows. The magnified oviduct/ovary images are marked with the corresponding hydrosalpinx scores. (b and d) Only one microscopic image is presented for each group, with ovary and dilated oviduct (DO) (double arrowhead) (b) and normal oviduct (NO) (arrows) (d) indicated. Both hydrosalpinx (macroscopic observation) and tubal dilation (under microscopy) incidence rates and severity scores are listed below the corresponding group images. Note that donor CD8⁺ T cells induced by i.g. Wt C. muridarum but not intravaginal Pf C. muridarum significantly promoted hydrosalpinx development in the recipients. *, P < 0.05 by Fisher’s exact test for comparing pathology rates/incidences and a Wilcoxon test for pathology scores. The data were acquired from 3 independent experiments (n = 15).

DISCUSSION

Despite extensive research efforts, the precise mechanisms by which Chlamydia induces long-lasting tubal pathology and infertility in either women (2–4) or female mice (5, 7, 9) remain to be elucidated. We have recently reported that gastrointestinal Chlamydia can promote chlamydial induction of hydrosalpinx in the female mouse genital tract (34), which has led to the proposal of a two-hit model for explaining the chlamydial pathogenic mechanism in the upper genital tract. Sexually transmitted Chlamydia may both ascend to the upper genital tract to cause tubal epithelial damage (as a 1st hit) and spread to the GI tract to induce pathogenic immune responses (as a 2nd hit) to convert the initial damage-triggered tissue-repairing responses into excessive and pathological fibrosis in the oviduct. Both hits are proposed to be required for establishing and maintaining long-lasting pathology in the upper genital tract. In the current study, we have demonstrated that GI Chlamydia may induce CD8⁺ T cell responses as a 2nd hit to promote chlamydial pathogenicity in the genital tract. First, we have validated that intragastric coinoculation with Wt C. muridarum rescued attenuated Pf C. muridarum to induce hydrosalpinx, as we reported recently (34). Second, we have also confirmed that GI Wt C. muridarum failed to directly spread to the mouse genital tract lumen since no mCherry-expressing C. muridarum was ever detected in the vaginal swabs, suggesting that GI Wt C. muridarum must use an indirect mechanism to promote chlamydial pathogenicity in the upper genital tract. Third, the promotion of hydrosalpinx by GI Wt C. muridarum was blocked by the depletion of CD8⁺ T cells, indicating that CD8⁺ T cells are necessary for GI C. muridarum to promote genital pathology. Finally, CD8⁺ T cells induced by GI Wt C. muridarum but not genital Pf C. muridarum promoted recipient mice that were intravaginally inoculated with Pf C. muridarum alone to develop significant hydrosalpinx. Thus, CD8⁺ T cells are both necessary and sufficient for GI Chlamydia to promote chlamydial pathogenicity in the upper genital tract.

The finding that intragastric coinoculation with Wt C. muridarum successfully rescued intravaginally inoculated Pf C. muridarum to induce hydrosalpinx (34) has provided direct experimental evidence supporting the two-hit model (35). The two-hit model was initially proposed based on the correlation of chlamydial pathogenicity with chlamydial spread to the GI tract (32, 36, 37). The 1st hit consists of tubal epithelial damage and wound-healing responses such as transient fibrosis, which is caused by chlamydial ascending infection (10, 11, 38, 39). During the acute tubal inflammatory response, chlamydial antigens must be processed and presented in the infected oviduct tissues, which is required for Chlamydia-specific CD4⁺ Th1 cells to clear tubal infection (17, 40). Similarly, major histocompatibility complex (MHC) class I-restricted epitopes may also be produced and presented to CD8⁺ T cells. The question is whether antigen-presenting cells that present chlamydial epitopes can persist in the upper genital tract long after the tubal infection is cleared. The 2nd hit emphasizes the recruitment of pathogenic T lymphocytes to the oviduct for promoting tubal fibrosis and hydrosalpinx. We now report that CD8⁺ T cells induced by Chlamydia in the GI tract can act as a 2nd hit to promote chlamydial pathogenicity in the upper genital tract, which is consistent with previous observations that CD8⁺ T cells are required for chlamydial induction of hydrosalpinx in mice genitally infected with wild-type C. muridarum (15, 16, 41). Since it is now known that wild-type C. muridarum can spread from the genital tract to the GI tract (24), pathogenic CD8⁺ T cells were likely induced by C. muridarum that spread to the GI tract. The next step is to determine the nature of these pathogenic CD8⁺ T cells.

We have used the coinoculation mouse model approach for demonstrating the role of CD8⁺ T cells induced by gastrointestinal Chlamydia in chlamydial pathogenicity in the genital tract by taking advantage of the facts that both intravaginal infection with Pf Chlamydia alone can cause only the 1st hit and intragastric coinoculation with Wt C. muridarum can provide only the 2nd hit. Intravaginally inoculated Pf Chlamydia has been shown to ascend to the mouse oviduct (11, 39), which allows it to cause the 1st hit. Although genital Pf C. muridarum can also spread to the GI tract, spreading is significantly delayed and reduced (32). Live Pf C. muridarum can be detected only in rectal swabs on day 14 after its intravaginal inoculation. This delay in the shedding of live Pf C. muridarum in rectal swabs was later found to be caused by the inability of Pf C. muridarum to colonize the gastrointestinal tract since orally delivered Pf C. muridarum also displayed a similar delay. Interestingly, it was recently found that Pf C. muridarum efficiently entered persistent infection in the small intestine (42) and successfully avoided the activation of intestinal T cell responses (43, 44) in order to reach the large intestine for establishing long-lasting colonization (45). Thus, although intravaginally inoculated Pf C. muridarum can spread to the GI tract, it may not induce strong immune responses in the GI tract, including pathogenic CD8⁺ T cells. This assumption is consistent with the fact that genital inoculation with Pf C. muridarum failed to induce hydrosalpinx (11, 12), although Pf C. muridarum could productively infect the oviduct (11). The conclusion that Pf C. muridarum is unable to induce pathogenic CD8⁺ T cells is also supported by the result that CD8⁺ T cells harvested from the spleen of mice on day 10 after intravaginal inoculation with Pf C. muridarum failed to promote hydrosalpinx in the recipient mice. Clearly, intravaginal inoculation with Pf C. muridarum can cause only the 1st hit. Next, we used intragastric coinoculation with Wt C. muridarum to produce a 2nd hit. Since GI Wt C. muridarum is known to be restricted to the gut only without spreading to the genital lumen (27, 33, 34), the pathogenic CD8⁺ T cells in the coinfected mice are likely induced by GI C. muridarum. This conclusion is consistent with the finding that CD8⁺ T cells harvested from mice on day 10 after intragastric inoculation with Wt C. muridarum promoted hydrosalpinx in the recipient mice that were intravaginally inoculated with Pf C. muridarum. Since the recipient mice had never been exposed to any Wt C. muridarum in any tissues, the adoptive-transfer experimental data have convincingly demonstrated that GI C. muridarum-induced CD8⁺ T cells can act as a 2nd hit for converting the tissue-repairing response triggered by the 1st hit into a long-lasting profibrotic response.

Regardless of whether pathogenic CD8⁺ T cells are induced by Chlamydia in the genital tract or GI tract, the next question is to determine how pathogenic CD8⁺ T cells promote hydrosalpinx in the mouse oviduct. Mice deficient in either tumor necrosis factor receptor 1 (TNFR1) (46) or interleukin-13 (IL-13) (47) significantly reduced hydrosalpinx induction by Chlamydia. IL-13-positive (IL-13⁺) CD8⁺ T cells have been isolated from Chlamydia-infected mice (48). IL-13⁺ CD8⁺ T cells have been shown to promote pathogenic fibrosis in other systems (49–51). GI tract infections are known to induce fibrosis-promoting lymphocytes (52–54). These analyses suggest that GI Chlamydia-induced CD8⁺ T cells may promote tubal fibrosis and hydrosalpinx by producing profibrotic cytokines. More studies are required to investigate the mechanisms by which Chlamydia-induced CD8⁺ T cells promote the development of hydrosalpinx.

It is worth pointing out that the 2-hit model may represent only one of the many mechanisms by which Chlamydia induces hydrosalpinx. We have recently shown that in the absence of CD8⁺ T cells, CD4⁺ T cells can also contribute to C. muridarum induction of hydrosalpinx (55). Thus, under different infection conditions and host backgrounds, different subsets of lymphocytes may make different contributions to chlamydial infection and pathogenicity. Caution should be taken when applying the two-hit mouse model to human C. trachomatis pathogenesis. Although murine bacterial C. muridarum infection in the mouse genital tract can induce tubal fibrosis/hydrosalpinx similar to that observed in C. trachomatis-infected women (2–4), C. muridarum may be transmitted naturally among mice via the oral-fecal route instead of genital tract transmission. Indeed, many C. muridarum virulence-related factors are more important for C. muridarum to colonize the GI tract than to infect the genital tract (32, 36, 37), suggesting that the C. muridarum virulence factors have been selected by the GI tract environment for evading immunity in the GI tract. In contrast, C. trachomatis may have been transmitted sexually among humans, which may force C. trachomatis to adapt to the human genital mucosa. Nevertheless, C. trachomatis has been frequently detected in the GI tracts of humans but without any significant association with GI pathologies (18–22). It is likely that C. trachomatis may have experienced selection pressures from both the genital and GI tracts. Hopefully, the current mouse studies may provide information for investigating the impact of GI C. trachomatis on C. trachomatis pathogenicity in the genital tract of women.

MATERIALS AND METHODS

Chlamydial organism growth.

C. muridarum clone Nigg3 (GenBank accession number CP009760.1) was used to derive plasmid-free (Pf) C. muridarum (clone CMUT3.G5; GenBank accession number CP006974.1) as described previously (11). The CMUT3.G5 clone was further used for transformation with pmCherry:CM to create CMpmCherry as described previously (14, 56). The CMUT3.G5 clone is referred to as Pf C. muridarum, while the CMpmCherry clone is referred to as wild-type (Wt) C. muridarum. Both clones were propagated in HeLa cells (human cervical carcinoma epithelial cells; ATCC CCL2.1) and purified as elementary bodies (EBs) as reported previously (57). Aliquots of purified EBs were stored at −80°C until use. SPG (sucrose-phosphate-glutamate) storage buffer consists of 220 mM sucrose, 12.5 mM phosphate, and 4 mM l-glutamic acid (pH 7.5).

Mouse infection and coinfection.

Mouse experiments were carried out in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (58). The protocol was approved by the Committee on the Ethics of Laboratory Animal Experiments of the University of Texas Health Science Center at San Antonio.

C. muridarum EBs were used to inoculate 6- to 7-week-old female CBA1/J mice (stock number 000656; Jackson Laboratories, Inc., Bar Harbor, ME) intravaginally and/or intragastrically as described previously (24, 37). For intravaginal inoculation, stock EBs diluted in 10 μl SPG medium that contained 2 × 10⁵ inclusion-forming units (IFU) were delivered to the ectocervix using a 20-μl micropipette tip. Five days prior to inoculation, each mouse was injected subcutaneously with 2.5 mg Depo-Provera (Pharmacia Upjohn, Kalamazoo, MI) suspended in sterile phosphate-buffered saline (PBS). For intragastric inoculation, EBs diluted in 100 μl SPG medium that contained 2 × 10⁵ IFU were delivered to the stomach using a straight balled-end needle designed for mouse oral gavage (catalog number N-PK 020; Braintree Scientific, Inc., Braintree, MA). In some experiments, the same mice were coinoculated intragastrically with Wt C. muridarum 7 days after intravaginal inoculation with Pf C. muridarum. The 7-day delay of the coinoculation protocol was established based on temporal titration in pilot experiments to achieve maximal exacerbation of hydrosalpinx without affecting the infectivity of genital C. muridarum (34). An advanced or simultaneous coinoculation with Chlamydia in the GI tract could induce protective immunity against subsequent chlamydial infection in the genital tract, as described previously (33). Following intravaginal inoculation, both vaginal and rectal swabs were taken periodically for titrating viable organisms as described previously (24, 26). In some experiments, the coinoculated mice were treated with an anti-CD8 antibody, while in others, donor CD8⁺ T cells were transferred to mice with intravaginal inoculation with Pf C. muridarum as described below. On day 56 after intravaginal inoculation, all mice were sacrificed for observing genital tract pathology as described below.

Antibody depletion and adoptive transfer.

In some experiments, the coinoculated mice were treated with anti-CD8 antibody (clone 53-6.7, rat IgG2aκ, purified as described previously [59] or purchased from Bio X Cell, West Lebanon, NH). The antibody treatment began on day 7 after intravaginal inoculation (on the day of intragastric coinoculation) and then twice weekly throughout the experiment. Normal rat IgG was used as a control. The antibody or rat IgG was administered intraperitoneally with 500 μg IgG in 500 μl PBS as described previously (59, 60).

For adoptive-transfer experiments, CD8⁺ T cells were prepared from donor CBA1/J mice either intravaginally infected with Pf C. muridarum or intragastrically infected with Wt C. muridarum. Both spleen and mesenteric lymph nodes harvested from donor mice on days 7 to 14 after chlamydial infection were used for preparing CD8⁺ T cells using a combination of negative selection and flow cytometry sorting. For negative selection, single-cell suspensions prepared from spleen and lymph nodes were treated with a cocktail of biotin-conjugated antibodies to remove CD4⁺ T cells (clone GK1.5, rat IgG2bκ, catalog number 100404; BioLegend, Inc., San Diego, CA), CD19⁺ B cells (clone 6D5, rat IgG2aκ, catalog number 115504; BioLegend), CD45R/B220⁺ cells (clone RA3-6B2, rat IgG2aκ, catalog number 103204; BioLegend), and MHC class II (I-A^k)-positive cells (clone M5/114.15.2, rat IgG2bκ, catalog number 107604; BioLegend) by using magnetic beads. The remaining cells were labeled with anti-CD8 antibody conjugated with Efluor 450 (clone eBioH35-17.2, catalog number 12-0083-81; Thermo Fisher Scientific, Inc., Waltham, MA) for flow cytometry sorting using the BD FACSAria II system (catalog number 642886; BD Biosciences, San Jose, CA). Portions of the cell samples prior to negative selection, after negative selection, and after sorting were saved for monitoring CD8⁺ T cell purity. The sorter-purified CD8⁺ T cells were injected into the recipient mice retro-orbitally with 2 × 10⁶ cells per injection three times on days 14, 21, and 35 after the recipient mice were intravaginally infected with Pf C. muridarum.

Titrating live chlamydial organisms recovered from swabs.

To quantitate live chlamydial organisms in vaginal or rectal swabs, each swab was soaked in 0.5 ml of SPG medium and vortexed with glass beads, and the chlamydial organisms released into the supernatants were titrated on HeLa cell monolayers in duplicate. The infected cultures were processed for an immunofluorescence assay as described previously (61) and below. Inclusions were counted in five random fields per coverslip under a fluorescence microscope. For coverslips with <1 IFU per field, entire coverslips were counted. Coverslips showing obvious cytotoxicity to HeLa cells were excluded. The total number of IFU per swab was calculated based on the mean IFU per view, the ratio of the view area to that of the well, the dilution factor, and inoculation volumes. Where possible, the mean IFU per swab were derived from the serially diluted and duplicate samples for any given swab. The total numbers of IFU per swab were converted into log₁₀ units, which were used to calculate the means and standard deviations across mice of the same group at each time point.

Immunofluorescence assay.

The immunofluorescence assay used for titration of live organisms was carried out as described previously (62). A rabbit antibody (designated R1604, raised with purified C. muridarum EBs) was used as a primary antibody to label all C. muridarum bacteria in HeLa cells, which was visualized with goat anti-rabbit IgG conjugated with Cy2 (green, catalog number 111-225-144; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). The DNA dye Hoechst 3328 (blue; Sigma-Aldrich, St. Louis, MO) was used to visualize nuclei. Doubly labeled samples were used for counting C. muridarum bacteria under a fluorescence microscope (AX70; Olympus) equipped with a charge-coupled-device (CCD) camera (Hamamatsu). For samples containing two different types of C. muridarum bacteria such as plasmid-free C. muridarum and wild-type C. muridarum that expresses mCherry, we used a rat anti-mCherry monoclonal antibody (mAb) (catalog number M11217; Thermo Fisher Scientific) together with R1604 as the primary antibody to differentiate them. mCherry-expressing wild-type C. muridarum cells were visualized with goat anti-rat IgG conjugated with Cy3 (red; Jackson ImmunoResearch Laboratories). Wild-type C. muridarum cells were labeled both green and red, while plasmid-free C. muridarum was green only. Distinct inclusions were counted separately using the AX70 microscope for calculating IFU recoveries.

Evaluating genital tract pathology macroscopically and microscopically.

On day 56 after intravaginal infection, mice were euthanized for evaluating genital tract pathology. The focus was on upper genital tract hydrosalpinx. Before the tissues were removed, an in situ gross examination was performed for evidence of oviduct hydrosalpinx or any other related abnormalities of oviducts. The severity of oviduct hydrosalpinx was scored based on the following criteria: no hydrosalpinx (score of 0); hydrosalpinx detectable only under a microscope (1); and hydrosalpinx clearly visible with the naked eye but with a size smaller than the ovary on the same side (2), equal to the ovary on the same side (3), or larger than the ovary on the same side (4). The excised tissues, after photographing, were fixed in 10% neutral formalin, embedded in paraffin, and serially sectioned longitudinally (with 5 μm/section). Efforts were made to include the cervix, both uterine horns, and oviducts as well as luminal structures of each tissue in each section. The sections were stained with hematoxylin and eosin (H&E) as described previously (61). The H&E-stained sections were assessed by a pathologist blinded to mouse treatment and scored for the severity of inflammation and pathologies based on modified schemes established previously (63). The oviducts were scored for both luminal dilation (0, no significant dilatation; 1, mild dilation of a single cross section; 2, one to three dilated cross sections; 3, more than three dilated cross sections; 4, confluent pronounced dilation) and inflammatory cell infiltration (0, no significant infiltration; 1, infiltration at a single focus; 2, infiltration at two to four foci; 3, infiltration at more than four foci; 4, confluent infiltration). Scores assigned to individual mice were calculated as means ± standard errors for each group of animals.

Statistical analyses.

All data, including the time courses of live organism shedding (IFU), genome copies, and pathology scores, were compared using the area under the curve (AUC) between two groups using a Wilcoxon rank sum test (an in-house Excel sheet), while category data, including the number of mice positive for live-organism shedding or hydrosalpinx, were analyzed using Fisher’s exact test (http://vassarstats.net/tab2x2.html).