Estrogen Receptor-Alpha (ESR1) Governs the Lower Female Reproductive Tract Vulnerability to Candida albicans

Estradiol-based therapies predispose women to vaginal infections. Moreover, it has long been known that neutrophils are absent from the vaginal lumen during the ovulatory phase (high estradiol). However, the mechanisms that regulate neutrophil influx to the vagina remain unknown. We investigated the neutrophil transepithelial migration (TEM) into the vaginal lumen. We revealed that estradiol reduces the CD44 and CD47 epithelial expression in the vaginal ectocervix and fornix, which retain neutrophils at the apical epithelium through the estradiol receptor-alpha. In contrast, luteal progesterone increases epithelial expression of CD44 and CD47 to promote neutrophil migration into the vaginal lumen and Candida albicans destruction. Distinctive to vaginal mucosa, neutrophil infiltration is contingent to sex hormones to prevent sperm from neutrophil attack; although it may compromise immunity during ovulation. Thus, sex hormones orchestrate tolerance and immunity in the vaginal lumen by regulating neutrophil TEM.

inTrODUcTiOn Neutrophils maintain mucosal immunity since they are very efficient at phagocytosing and killing invading microbes (1). However, neutrophil migration into the tissues must be carefully regulated because of the potential bystander tissue damage resulting from uncontrolled release of toxic metabolites (2). Neutrophils develop in the bone marrow (1) until they are mature to circulate in the blood (3). They are mobilized into the microcirculation by CXCR2-ligand gradients (4). Once extravasated, resident sentinel cells release proinflammatory mediators and chemoattractants that activate and guide neutrophils through the interstitium toward sites of infection/inflammation (5)(6)(7). In mucosal tissues, such as bladder, lung, and gut, epithelial cells produce chemoattractants that guide neutrophils through the interstitium until they reach the subepithelial space (8). Next, neutrophils cross the epithelial barrier in sequential steps: epithelial basement membrane crossing, initial adhesion to the epithelium, migration between epithelial cells, and detachment from the apical epithelial surface into the luminal space where they can combat invading microbes (9). Although neutrophil transepithelial migration (TEM) to the intestinal lumen has been somewhat characterized (10), how they migrate into the female reproductive tract (FRT) remains obscure (11).
In order to maintain a compromise between commensal and opportunistic microbiota, exogenous spermatozoa, and the threat of sexually transmitted pathogens, the vaginal mucosa is controlled by cyclic hormonal levels. Cervico-vaginal mucus and neutrophils protect epithelial cells from pathogens and create a microenvironment that fosters endogenous vaginal flora while deterring invasive microbial species that also target these regions, including Human Papiloma Viruses, Candida albicans, T. vaginalis, C. trachomatis and N. gonorrhea (12)(13)(14). Furthermore, sperm cells are pooled in the ectocervix and their survival must be facilitated if fertilization is to occur (15). To allow this, neutrophils disappear from the vaginal lumen during the ovulatory phase (16)(17)(18)(19). However, these changes also create vulnerability to infection during ovulation (14). Although the leukocyte levels variation in the vaginal lumen during the ovarian cycle is well described in humans and mice (11,17) and has important clinical implications (20), the mechanisms that regulate neutrophil influx to the vaginal lumen is an unsolved and long-standing question.
To study the role of hormones in regulating neutrophil migration in the vaginal lumen, we used a mouse model of candida infection, in the presence of estradiol (E2 follicular and ovulatory phase) or progesterone (P4 luteal phase) and analyzed the role of these hormones in each step of neutrophil migration in the FRT. We discovered that E2-treatment, through the estradiol receptoralpha (ESR1), induces epithelial downregulation of CD47 and CD44, and retains neutrophils at the mucosa. We demonstrated how important molecules for neutrophil TEM in the gut (9) have a different regulation in the FRT to coordinate tolerance during ovulation to reconcile immunity and reproduction. Thus, the lack of neutrophils in the vaginal lumen during ovulation or following the use of E2-based therapies induces a vulnerability period that predisposes women to vaginal infections.  (21). Mice were maintained under specific pathogen-free conditions in the Animal Facility of IiSGM. Four-week-old mice (13-15 g) were bilaterally ovariectomized under anesthesia (22) and given 2 weeks to recover. Females were randomly assigned to the groups and then injected subcutaneously with 0.02 mg of 17β-E2 or 0.2 mg of P4 (Calbiochem, Germany) dissolved in 100 µl of sesame oil (Sigma-Aldrich, USA) every 4 days. Hormone treatment was based on a previous report, and the dose was sufficient to maintain hormone concentrations in female mice (23). Seventy-two hours after first hormonal treatment, 2 × 10 6 C. albicans blastoconidia in 20 µl of PBS were inoculated into the vagina. Vaginal secretions were collected by gently flushing the vaginal vault four times with 50 µl of sterile PBS using a 100 µl pipette tip inserted 5 mm in the vagina.

Candida albicans
The SC5314 C. albicans (ATCC MYA-2876) strain was grown on Sabouraud dextrose chloramphenicol agar plates (Conda, Spain) overnight at 30°C prior to the experiments.

Vaginal epithelial cells Ex Vivo analysis
Lower FRT was excised and dissociated using the Tumor Dissociation Kit (Miltenyi Biotec, Germany) following manufactures instructions. Samples were washed and the cells were stained (see below) for phenotypic analysis by flow cytometry. Epithelial cells were analyzed by gating on the Cdh1 + cells.

confocal Microscopy
Whole FRT tissues were embedded in Tissue-Tek OCT (Sakura, Netherlands). Eight-micrometer sections were fixed with acetone, blocked (50 µg/ml mouse IgG and 10% BSA) and stained with (1-5 µg/ml) the antibodies: LY6G (ebioscence, USA), MMP9 (R&D Systems, USA), CD44 (Miltenyi Biotec, Germany), Sirpα (Biolegend, USA), CD47 (Miltenyi Biotec, Germany), CDH1 (Biolegend, USA) and CD55 (Miltenyi Biotec, Germany). For FRT screening and in vivo protein expression quantification, tissues were usually triple-stained and imaged using the glycerol ACS APO 20× NA 0.60 immersion objective of a confocal fluorescence microscope (SPE, Leica Microsystems), maintaining the acquisition settings all over the process for each sample and among samples, as previously described (23,24). Mean fluorescence intensities (MFI) were assessed at multiple regions of interest (10-15 by field) randomly depicted at specific areas of the FRT epithelium. Namely, the supraepithelial layer for CD44 and the middle layer of the stratified epithelium for CD47. For single-cell quantification, LY6G+ stained neutrophils were segmented and MFI of different markers quantified at matched cells. All quantifications, including line-profiles, were performed using the FIJI software (NIH).

Transepithelial Migration
VK2/E6E7 (ATCC ® CRL-2616TM) cells were grown on collagencoated permeable 0.33-cm 2 polycarbonate filters 5 µm pore size (Costar, USA). Confluent inverted VK2 monolayers were pretreated apically with 10 µg/ml of GM35 (27) or isotype control binding antibody (ebioscience, USA) for 20 min at room temperature. 1 × 10 6 PMNs were added to the upper part of the transwell and allowed to migrate for 1 h at 37°C in the presence of 100 nM fMLF (Abcam, USA) in the physiologically relevant basolateral-to-apical direction (27,28). Migrated PMNs were assessed at the lower chamber by flow cytometry.

statistical analysis
The test used to determine significance between the treatments in each experiment can be found in the figure legends. GraphPad Prism 5 (GraphPad Software, Inc., USA) was used to determine statistical significance. A p-value <0.05 was considered significant.

resUlTs estradiol Prevents neutrophil recruitment into the Vaginal lumen
To study the role of sex hormones in neutrophil influx into the vaginal lumen, we set up a model of C. albicans vaginal infection in hormone-treated mice. We previously detected a significant increase of neutrophils in the vaginal lavage of P4-treated mice (~25-fold) 6 h after infection compared with non-infected mice; in contrast, neutrophil infiltration was residual in E2-treated infected mice ( Figure 1A) (23). Inhibition of neutrophil influx to the vaginal lumen in E2-treated mice could be due to a diminished recruitment from the vascular beds into the stromal layer or an impeded TEM. To discriminate between these possibilities, we excised the lower FRT after vaginal lavage, dissociated the tissue, and quantified neutrophils by flow cytometry. We detected neutrophils in the vaginal lavage but not in the tissue of P4-treated infected mice; however, E2-treated mice showed minimal infiltration in the vaginal lavage and considerable neutrophil accumulation in the lower FRT tissue ( Figure 1A). Moreover, E2-treated ESR1 −/− mice showed higher neutrophil numbers (~15-fold) in the vaginal lavage than the Wt, indicating that ESR1 plays a key role in the downregulation of neutrophil migration across the vaginal epithelium ( Figure 1B).

estradiol induces neutrophil accumulation at the Fornix and ectocervix
To determine at which area neutrophil migration was being blocked, we excised the lower FRT after vaginal lavage and quantified neutrophil density at the subepithelial stroma and at the epithelium along the FRT tissue by confocal microscopy. The lower FRT can be divided in two histological regions: (1) the introitus, whose epithelium resembles skin and, (2) the vagina, ectocervix, and fornix (the transition between the vagina and the ectocervix), which are covered by a stratified squamous epithelium. In contrast, the upper FRT (endocervix, uterus, and oviduct) consists of a simple columnar epithelium (Figure 2A) (29). We detected a higher density of neutrophils in E2-treated compared with vehicle (Vh) and P4-treated mice both in the fornix and ectocervix. In addition, neutrophil accumulation was observed in the stroma and attached to the apical surface of the epithelium following vaginal lavage. We detected less neutrophil accumulation in the vagina, endocervix, and uterus than in the fornix and ectocervix (Figures 2B,C; Figures S2A,B in Supplementary Material). Furthermore, E2-treated ESR1 −/− mice showed lower neutrophil numbers in the fornix (~6-fold) and ectocervix (~36-fold) epithelium than the Wt (Figures 2D,E), in accordance with the vaginal lavage in the Figure 1B. Thus, we demonstrate that E2 arrests neutrophils in the ectocervix and fornix, which may explain the absence of neutrophils in the vaginal lumen of E2-treated mice ( Figure 1A). Therefore, we aimed to investigate how E2 regulates neutrophil migration across the vaginal epithelial barrier in vivo.

effect of Female sex hormones on neutrophil expression of Mediators of TeM
Neutrophil TEM is a multistep process; it starts with the initial crossing of the basement membrane and engagement of epithelial cells. Neutrophils secrete MMP9 to degrade collagen IV at the basement membrane (7). Next, the neutrophil integrin CD18/CD11b mediates the initial contact with epithelial cells, as shown in the lung, intestine, and bladder (9). Although E2-treatment resulted in accumulation of most neutrophils along the supra-epithelium, we detected accumulations in the subepithelium of the fornix and ectocervix. We, therefore, hypothesized that E2-treatment delays crossing of the basement membrane and/or engagement of the epithelial cells. We detected similar levels of MMP9 and CD11b in LY6G+ neutrophils of P4-, Vh, and E2-treated mice in the vagina, fornix, and ectocervix by confocal microscopy, as well as in vivo and ex vivo treatment by flow cytometry (Figures 3A-F). Suggesting that changes in these proteins are unrelated to the observed accumulation of neutrophils under the epithelium in the hormonetreated mice.
After the initial contact of neutrophils with the epithelium, key protein-protein interactions facilitate movement through the paracellular space between adjacent epithelial cells. One of the epithelial receptors important for this process is CD47 (30). Disruption of the interaction between neutrophil SIRPA (signalregulatory protein α or CD172a) and epithelial CD47 delays neutrophil TEM into the intestine (31). Analysis of SIRPA expression by vaginal neutrophils, however, revealed no differences in SIRPA expression in neutrophils in hormone-treated mice by confocal microscopy as well as in vivo and ex vivo treatment by flow cytometry (Figures 4A-C). This again suggests that neutrophil SIRPA is not subjected to regulation in E2-treated mice. In summary, our data suggest that E2 does not regulate the neutrophil TEM by acting on the neutrophils because E2-treatment did not affect neutrophil expression of the major receptors, CD11b, MMP9, and SIRPA.
To better understand these findings, we looked for estradiol and progesterone nuclear receptor expression in neutrophils from the lower FRT. Although it has been previously reported that circulating neutrophils express estradiol and progesterone receptors (23, 32), we could not detect any protein receptor expression in neutrophils from the ectocervix and fornix, whole FTR, or vaginal lavage; in contrast, we detected expression of both nuclear receptors in the epithelial cells and neutrophils from blood, bone marrow, and ex vivo hormone treated (Figures 5A-C). This could explain why hormones unaffected neutrophils and suggested that Female sex hormones act on the epithelium to regulate neutrophil TeM To study the role of sex hormones on epithelial cells and the regulation of the neutrophil passage through the lateral intercellular space between epithelial cells, we focused on CD47 expression (30,33). We detected lower surface expression of CD47 in ex vivo FRT epithelial cells isolated from E2 compared to P4-treated mice, as assessed by flow cytometry (Figure 6A). To determine its exact anatomical distribution, we examined CD47 expression in epithelial cells of the lower FRT. We detected stronger expression of CD47 in P4-and Vh-treated compared with E2-treated epithelium in the vagina (~4-fold), ectocervix (~2-fold), and fornix (~2-fold). This differential expression was observed with and without fungal infection (Figures 6B,C; Figure S3A in Supplementary Material), indicating an E2-dependent and candida independent effect. Likewise, E2-teated ESR1 −/− mice expressed elevated levels of CD47 in the ectocervix and fornix; thus indicating that the effect was fully ESR1-dependent (Figures 6D,E). In conclusion, E2 downregulates CD47 expression along the ectocervix and fornix epithelial cells in E2-treated mouse, which coincides with E2-dependent subepithelial accumulation of neutrophils. Hence, lower levels of epithelial CD47 expression might contribute to the negative regulation of neutrophil migration into the vaginal lumen induced by E2, and their retention in the ectocervix and fornix epithelium.

estradiol Downregulates epithelial cell cD44 expression
Finally, after migrating across epithelial cells, neutrophils emerge in the apical epithelium, where they need to detach from the epithelial cells before they reach the lumen (9, 10). CD44v6 and CD55 are membrane glycoproteins expressed on the apical surface of gut epithelial cells that play an anti-adhesive role and promote neutrophil release to the lumen (27). We detected that E2-treated mice displayed retention of neutrophils after vaginal lavage, which remained attached to the supra-epithelium, facing the vaginal lumen ( Figure 2B). Therefore, we explored CD44 expression in VK2 vaginal epithelial cells. We found apical epithelial expression of CD44 displaying the glycan sialyl Lewis A (sLe a ) ( Figure 7A). Importantly, sLe a on CD44v6, recognized by the mAb GM35, has previously been shown to be required for neutrophil entry into the intestinal lumen (34). Furthermore, blockade with mAb GM35 resulted in a reduction (~90%) in PMN TEM across VK2 cells ( Figure 7B). Thus, CD44v6/sLe a is necessary for the release of neutrophils from the apical surface of the vaginal epithelial cells in migration assays. Similarly, we detected higher surface expression of CD44 in ex vivo FRT epithelial cells isolated from P4-compared to E2-treated mice ( Figure 7C). Next, we explored the anatomical localization of the CD44 in the mouse lower FRT. We detected increased expression of epithelial CD44 in P4-and Vh-treated compared with E2-treated mice in the vagina (~9-fold), ectocervix (~9fold), and fornix (~6-fold). In addition, CD44 expression was unaltered by candida infection (Figures 8A,B; Figure S3B in  Supplementary Material), indicating, a candida independent and E2-dependent effect. Indeed, E2-teated ESR1 −/− mice showed strong CD44 expression in the ectocervix and fornix, indicating the key role of E2 in the CD44 repression (Figures 8C,D). Notably, E2 treatment strongly reduced the levels of soluble CD44 present in the vaginal lavage and in the surface of epithelial cells from the mouse lower FRT. This effect was abolished in the E2-teated ESR1 −/− mice (Figures 8E,F). In contrast to CD44, expression of CD55 was unaffected by the hormones (Figure S3C in Supplementary Material). Hence, our data are in agreement with a role of CD44 in neutrophil TEM previously reported in the gut, in which shedding of the extracellular domain of CD44v6 is associated with the release of neutrophils from the apical epithelial surface (9,34). In summary, E2 downregulates CD44 expression and shedding in the apical fornix and ectocervix epithelial cells, which together with the role of this receptor in TEM suggests that estradiol impairs neutrophil detachment, thus reducing their entry into the vaginal lumen during ovulation.

effect of Female sex hormones on cD44 and cD47 expression in gut and Bladder epithelial cells
Because hormonal treatment has a systemic effect, we wondered if E2-treatment also lowered epithelial expression of CD47 and CD44 in other mucosal lined organs. We observed a similar constitutive expression of CD47 in both gut and bladder epithelial cells. However, although constitutive epithelial expression of CD44 was observed in the bladder, CD44v6 expression in the gut was only observed following an inflammatory stimulus (34,35). Significantly, E2 treatment had no effect on expression of CD44 and CD47 in gut and bladder epithelia (Figures 9A,B). Therefore, E2 specifically modulates expression of receptors required for TEM in the lower FRT epithelium.

role of neutrophils attached to the apical Fornix and ectocervix epithelium
We found that E2 reduces the expression on the epithelial cells of key receptors involved in TEM and retains neutrophils attached at the fornix and ectocervix epithelium. To evaluate the neutrophil potential killing activity, we challenged them in the in vivo C. albicans-infected model. P4-treated neutrophils showed a robust candidacidal activity (~85% reduction) in vivo, and depletion of  neutrophils impaired the vaginal fungal clearance (Figure 10A). However, E2-treated mice were inefficient clearing the fungal infection when they were attached to the epithelium ( Figure 10B). Further, fungal burden was reduced (~6-fold) in ESR1 −/− mice suggesting that ESR1 is necessary for the E2-dependent impaired candidacidal activity; whereas it had no effect on the P4-treated Wt and ESR1 −/− mice ( Figure 10C). Therefore, P4-treatment induces neutrophil dependent protection, and E2-treatment reduces neutrophil killing potential during ovulation, probably because of E2 dependent vaginal epithelial cells secretions (36) and induces susceptibility to vaginal infection, which is ESR1 dependent. Thus, we suggest that neutrophil retention could avoid sperm damage during ovulation to favor reproduction, although it induces susceptibility to C. albicans.

DiscUssiOn
We present an ESR1-dependent mechanism that significantly decreases neutrophil protection in the vaginal lumen during ovulation or E2-based therapies. E2 treatment alters CD44 and CD47 expression in the FRT epithelium, which accumulate neutrophils in the sub-and supra-epithelial spaces of the ectocervix and vaginal fornix rather than at the vaginal lumen. In contrast, P4-treatment promotes neutrophil migration to the vaginal lumen and neutrophil killing (Figure 11). Unique to vaginal mucosa, neutrophil TEM is independent of the infection but dependent on sex hormones to prevent sperm from neutrophil attack, although it may compromise immunity during ovulation. These data help to explain why E2-based hormonal replacement therapies predispose women to vaginal infections and provide a new mechanism to explain the long-standing observation that vaginal neutrophil levels vary during the ovarian cycle (16).
Neutrophils protect the lower FRT from opportunistic microbiota/sexual transmitted diseases-causing pathogens. During the luteal phase, neutrophils continuously enter the vaginal lumen through the ectocervix (17). However, during the ovulatory phase (high E2 levels), neutrophils disappear from the vaginal lumen (16), probably because E2-treatment alters CXCL1 gradients (23) and CD47 expression in the lower FRT epithelial cells, which might retard neutrophil migration. In addition, E2-treatment downregulates apical expression of CD44v6, thereby impairing neutrophil detachment. Besides, neutrophils have abolished their killing potential, probably because of E2-dependent vaginal epithelial cells secretions like heparan sulfate, which could affect the neutrophils interaction with candida (36).
Here, we present evidences of a mechanism by which E2 treatment acting on the epithelium, but not on the neutrophils, accumulates them at the ectocervix and fornix, where they provide barrier protection, and could contribute to tissue repair (37). Neutrophils accumulated in these areas would be prepared to quickly invade the vaginal lumen during the early luteal phase (17), when P4 restores neutrophil infiltration (23) by favoring neutrophil detachment and release to the vaginal lumen. Moreover, P4 reestablishes TEM of the sub-epithelial neutrophils to promote a second wave of immune cells that control the vaginal microbiota and trap sperm or sperm-associated microbes after ovulation (38). We propose that this could be an ESR1-driven, multilayer safety mechanism to prevent unspecific neutrophil damage. Epithelial cells represent the first barrier of defense against pathogens and produce signals in response to aggressions that attract the innate immune cells from the blood. Several tissues such as the gut, bladder, and lungs, constitutively present CD47 expression (8,30,39). In contrast, tissue-specific epithelial expression of CD44v6 has been reported (34). In any case, inflammatory signals upregulate epithelial expression of CD47 (40) and CD44v6 (34) to favor neutrophil TEM. However, unique to the lower FRT mucosa, epithelial CD44 and CD47 expression was independent on the candida infection, because neutrophil infiltration and functions must be timed to the ovarian cycle. Our data support the notion that E2 acts on several steps of neutrophil TEM independently of infection with C. albicans to anticipate immune protection, while protecting sperm from neutrophil attack, even at the risk of a compromised immunity during ovulation.
The absence of neutrophils in the vaginal lumen during ovulation was first reported in 1917 (16), but its mechanisms have remained unknown. We present a unique model by which sex hormones regulate neutrophil influx to the vaginal lumen, which is in a coordinated fashion that allows for both containment of infections and reproductive functions. Our data reveal that ESR1 is necessary for the E2-dependent downregulation of neutrophil TEM and impairs neutrophil killing functions during ovulation to protect sperm. Clinically, these data imply that E2-based treatments could compromise vaginal immunity and favor opportunistic microbiota or STDs causing pathogen infections.