CD8+CXCR5+ T Cells Regulate Pathology in the Genital Tract

We have identified a CD8+CXCR5+ T cell that prevents the development of oviduct dilation following C. muridarum genital infection. Phenotypic studies show that CD8+CXCR5+ cells express markers of T regulatory cells (FoxP3, CD25, and GITR) but do not express a necessary component of cytotoxic cells (perforin). Cxcr5−/− mice have significantly lower numbers of CD8+ cells and lack the CD8+CXCR5+ population while the total number of CD4+ cells is equivalent between mouse strains. The transfer of CD8+ splenocytes from WT mice reduces the oviduct dilation seen in Cxcr5−/− mice following C. muridarum infection. Future studies will investigate the mechanism by which this cell type regulates genital tract pathology.


Introduction
Pelvic in�ammatory disease (PID) is de�ned as in�ammation of the uterus and/or fallopian tubes and is induced by a number of organisms following sexual transmission. Chlamydia trachomatis is the most common reportable sexually transmitted infection (STI) and is responsible for >1 million cases in the US and approximately 92 million cases worldwide each year [1,2]. Genital infection can lead to immune-mediated damage of the female reproductive organs and serious reproductive disability, including PID that can result in chronic pelvic pain, ectopic pregnancy, and infertility [3]. e risk of developing infertility increases by 40%-70% following reinfection [4]. e reinfection rate is approximately 13% and occurs within 6 months [5]. Delivery of treatments designed to reduce the local in�ammation and prevent �brotic disease to infected individuals may be a viable approach for further reducing PID and the costs associated with its treatment.
Regulatory T cells (Tregs) are comprised of multiple subsets of T cells that suppress other T cells from engaging in detrimental immune responses [6]. Tregs are broadly categorized as natural or inducible. Natural Tregs induce tolerance, delete autoreactive T cells, and dampen in�ammation during an autoimmune reaction [7][8][9][10][11][12][13]. Inducible Tregs arise during infections in response to the infectious process to restore the homeostatic environment. In some cases, Tregs can be actively induced by the pathogen and promote pathogen survival by preventing elimination [14]. Tregs have also been shown to protect mucosal surfaces of the intestine from in�ammation [15]. e linage can be phenotypically identi�ed by the Foxp3 transcription factor [16]. e most widely studied subset is phenotypically de�ned as CD4 + CD25 + FoxP3 + . is subset has been shown to indirectly prolong microbial growth by interfering with the priming of naive or unstimulated T cells [17].

Animals, Chlamydia, and Infection of Mice. A breeding colony was established with
−/− mice (8 generations in C57BL/6) obtained from Martin Lipp, Delbrück-Center for Molecular Medicine, Berlin, Germany. Female C57BL/6 mice, 5-6 weeks old (Harlan Sprague-Dawley, Indianapolis, IN, USA) were housed according to American Association of Accreditation of Laboratory Animal Care guidelines. Animal experimental procedures were approved by the UCLA Institutional Animal Care and Use Committee. Chlamydia muridarum was grown on con�uent McCoy cell monolayers, puri-�ed on Renogra�n gradients and stored at −80 ∘ C in sucrosephosphate-glutamine buffer (SPG) as previously described [21]. Mice were hormonally synchronized by subcutaneous injection with 2.5 mg of medroxyprogesterone acetate (Depo Provera, Upjohn, Kalamazoo, MI, USA) in 100 L saline 7 days prior to a vaginal challenge with 1.5 × 10 infection forming units (IFUs) of C. muridarum under anesthetization. Depo Provera drives mice into a state of anestrous and eliminates the variability in the rate and severity of infection due to the estrus cycle. Infection was monitored by measuring IFUs from cervical-vaginal swabs (Dacroswab Type 1, Spectrum Laboratories, Rancho Dominguez, CA, USA) as described [21].

2.2.
Histology. e genital tracts (GTs) were removed and, �xed in 10% formalin overnight, followed by 70% ethanol. Tissues were embedded en bloc in paraffin, sectioned (5 mm), and stained with hematoxylin and eosin. Tissue blocks were cut transversally from the ovary, and sections were collected at the beginning of the transitional region between ovary and oviduct. A veterinarian pathologist scored 2 sections from the right and le oviducts of each mouse for luminal dilation; 0 = luminal oviduct size of naïve mice, 1+ = mildly increased luminal oviduct size, 2+ = moderately increased luminal oviduct size, 3+ = severely increased luminal oviduct size, and 4+ = severely increased luminal oviduct size in greater than 75% of oviducts.

Cell Sorting and Adoptive
Transfer. Lymphocytes were isolated from the spleens of infected and uninfected mice as described above. Cells were stained for CD3 and CD8 and sorted for the CD3 + CD8 + population using FACSAria cell sorter (BD Bioscience) in the UCLA FACS Core Labaratory. e purity of the sorted population was 99%. Cells were resuspended in saline and 1 × 10 6 cells were intravenously injected into the recipient through the tail vein. At the time of transfer, the mice were infected with C. muridarum as described above. ese mice were also synchronized with medroxyprogesterone acetate 7 days prior to infection as described above.

Phenotypic Characterization of CD8
We identi�ed a population of CD8 + cells that expressed CXCR5 and comprised 2%-3% of CD8 T cells in the MLN (Figure 1(a)) and spleen (data not shown). e percentage of CD8 + CXCR5 + cells in the MLN and spleen did not change in a naive mice or following C. muridarum genital infection (data not shown). Phenotypic analysis found that these cells were memory cells by high expression of CD44. Analysis of cytotoxic markers, granzyme B and perforin, showed that these cells did not express perforin. Although granzyme B was expressed, these cells are not considered cytotoxic since perforin is required for entry of granzymes into the cell cytoplasm as shown in mice de�cient in perforin [23][24][25]. In addition, cytotoxic memory cells express high levels of both granzyme B and perforin (Figure 1(b)) [26]. Examination of markers of Tregs showed a subpopulation of CD8 + CXCR5 + cells that expressed CD25, Foxp3 and GITR (glucocorticoid-induced tumor necrosis factor receptor family related gene) upon stimulation from mice during a genital infection with C. muridarum (Figure 1(c)). Analysis of CD122 or CD127 was negative (data not shown). Interestingly, induction of this population could not be found on stimulated cells from naive mice (Figure 1(d)). is �nding indicates that a subpopulation of CD8 + CXCR5 + cells express phenotypic markers associated with Tregs and not cytotoxic cells.
−/− mice had slightly less CD3 + T cells (data not shown) but had similar numbers of CD4 cells. In addition, the mice had signi�cantly less CD8 cells compared to WT mice ( Figure  2(b)). Further analysis showed that these cells were CD8 T cells by expression of the TCR and CD8 chain and not a T cell or NKT cell that could express CD8 (Figure 2(c)).

Adoptive Transfer of CD8 + Cells Reversed Oviduct Dilation aer C. muridarum Infection in
−/− Mice. WT and −/− mice were intravaginally infected with C. muridarum and splenic lymphocytes were puri�ed 7 days aer infection. e lymphocytes or puri�ed CD8 + cells were adoptively transferred into WT or −/− recipients. e genital tracts were harvested seven weeks aer infection, hematoxylin and eosin stained, and scored by a veterinarian pathologist. We found that the adoptive transfer of WT lymphocytes into WT recipients had signi�cantly reduced oviduct dilation compared to −/− mice given Infectious Diseases in Obstetrics and Gynecology and WT mice were infected with C. muridarum and the spleens were harvested 7 days later and lymphocytes were obtained by dissociating cells within the organs. Donor C57BL/6 WT mice were infected with C. muridarum and the spleens harvested 7 days aer infection. CD8 + cells were puri�ed by FACS sorting for expression of CD8. −/− lymphocytes and WT-CD8 + were adoptively transferred (1 × 10 6 ) into −/− mice as indicated. Genital tracts were harvested 7 weeks later and hemetoxylin and eosin staining was performed and scored. Oviducts were scored individually and were compared between recipient groups using Kruskal-Wallis test. Transfer groups are, WT donor into WT recipient; 6 mice (12 oviducts), −/− donor into −/− recipient, 6 mice (12 oviducts), and WT-CD8 donor into −/− recipient, mice (4 oviducts). e bars indicate the mean ± SEM and * . (b) Recipient mice were vaginally swabbed 9 days aer infection and all groups were compared by one-way ANOVA. ere were no differences between groups. e bars indicate the mean IFU ± SEM. −/− lymphocytes. Further, −/− mice given puri-�ed CD8 + cells from WT mice also showed signi�cantly reduced oviduct dilation (Figure 3(a)). It is possible that the transfer of CD8 effector cells could account for a reduction in oviduct dilation by perhaps reducing bacterial burden. However, there was no difference in bacterial burden in the GT on days 9 following infection between −/− mice given WT-CD8 cells or −/− lymphocytes or WT mice given WT lymphocytes (Figure 3(b)). Taken together, these data strongly suggest that CD8 + CXCR5 +/+ (WT mice) prevent or interfere with the development of oviduct dilation following C. muridarum genital infection.

Conclusions
Our study describes the identi�cation of CD8 + CXCR5 + cells that possess the ability to regulate oviduct dilation that occurs following the immune response to C. muridarum genital infection. is study highlights the �nding that the CD8 + population is comprised of multiple subsets with differing function. Murthy et al., showed that CD8 cells secreting TNF cause oviduct dilation and hydrosalpinx [27]. Individual differences in PID or infertility may be in�uenced by the proportions of CD8 + cell subsets. e mechanism whereby regulation of genital tract pathology occurs is not known. However, a small percentage of CD8 + CXCR5 + cells express markers of Tregs; FoxP3, CD25, and GITR but not perforin, a marker of cytotoxic cells, and suggest that these cells function as T regulatory cells.
Our data show that adoptive transfer of CD8 + CXCR5 + cells prevent oviduct dilation following genital infection with C. muridarum and suggests that this cell type is necessary for preventing oviduct dilation following genital infection. However, CD8 + cells with suppressive activity require IL-10 + APC for expansion [28]. IL-10 + APC have been shown to reduce the number of antichlamydial 1 cells that develop and this results in a prolonged infection [29]. Others have suggested that there are a number of immune responses operating in different phases of the immune response and this could possibly explain the con�icting role of IL-10 + APC during C. muridarum infection.

Infectious Diseases in Obstetrics and Gynecology
A population of CD8 cells that also expresses CXCR5 has been reported but their function has not been identi�ed [30]. ere are a number of subsets of CD8 + Tregs that have been identi�ed and include the following: CD8 + CD25 + FoxP3 + , CD8 + CD45RC low FoxP3 + , CD8 + CD28 − FoxP3 − , CD8 + CD122 + FoxP3 − , and CD8 FoxP3 − [18]. We found that a small percentage of CD8 + CXCR5 + cells expressed FoxP3 upon antigen stimulation but did not express CD122 or CD8 . e mechanism of suppression of the CD8 subset has been reported. is subset is restricted to nonclassical MHC class Ib molecules, Qa-1, and has been found to suppress autoantibody formation and development of systemic lupus erythematosus (SLE-) like disease through inhibition of T FH [31]. Although CXCR5 expression does not prevent entry into germinal centers, the −/− mice used in our study do not have T FH cells or germinal centers and have reduced levels of immunoglobulins [32,33]. Likewise we also �nd a reduction in antichlamydial IgM and IgG levels (data not shown). Additionally, regulation by the CD8 subset depended on perforin and perforin was not found on activated CD8 + CXCR5 + cells in this study con�rming that CD8 + CXCR5 + identi�ed in this study are not the CD8 expressing cell described previously.
T regulatory cells primarily effect T cells and/or dendritic cells and have four mechanisms by which they suppress immune responses. T regulatory cells secrete inhibitory cytokines such as IL-10 and TGF , induce apoptosis of target cells using the granzyme-perforin pathway, deprive effector T cells of IL-2 which leads to apoptosis of effector T cells, and inhibit dendritic cell function [19]. Our study has found that the granzyme-perforin pathway is unlikely to be used by CD8 + CXCR5 + based on a lack of expression of the effector molecule, perforin. It would appear unlikely that suppressive cytokines are secreted since the resolution of infection is similar between −/− and WT mice and IL-10 secretion has been shown to prolong the course of infection [29]. Future studies will determine whether CD8 + CXCR5 + function as T regulatory cells and act directly on T cells or on another cell type such as dendritic cells.
�on��ct of �nterests e authors do not have a commercial or other association that might pose a con�ict of interests.