Estradiol-Treated Female Mice as Surrogate Hosts for Neisseria gonorrhoeae Genital Tract Infections

Abstract

Historically, animal modeling of gonorrhea has been hampered by the exclusive adaptation of Neisseria gonorrhoeae to humans. Genital tract infection can be established in female mice that are treated with 17β-estradiol, however, and many features of experimental murine infection mimic human infection. Here we review the colonization kinetics and host response to experimental murine gonococcal infection, including mouse strain differences and evidence that IL-17 responses, toll-like receptor 4, and T regulatory cells play a role in infection. We also discuss the strengths and limitations of the mouse system and the potential of transgenic mice to circumvent host restrictions. Additionally, we review studies with genetically defined mutants that demonstrated a role for sialyltransferase and the MtrC-MtrD-MtrE active efflux pump in evading innate defenses in vivo, but not for factors hypothesized to protect against the phagocytic respiratory burst and H(2)O(2)-producing lactobacilli. Studies using estradiol-treated mice have also revealed the existence of non-host-restricted iron sources in the female genital tract and the influence of hormonal factors on colonization kinetics and selection for opacity (Opa) protein expression. Recent work by others with estradiol-treated mice that are transgenic for human carcinoembryonic adhesion molecules (CEACAMs) supports a role for Opa proteins in enhancing cellular attachment and thus reduced shedding of N. gonorrhoeae. Finally we discuss the use of the mouse model in product testing and a recently developed gonorrhea chlamydia coinfection model.

Keywords: Neisseria gonorrhoeae; Opa proteins; antimicrobial peptides; hormones; immune response; lactobacilli; mouse; neutrophils.

Figure 1

Schematic of mouse infection protocol and characteristics of infection. (A) In our laboratory, mice are treated with 17β-estradiol to promote Gc infection by implantation of a slow-release pellet under their skin (Jerse, ; not shown) or subcutaneous administration of water-soluble estradiol (estradiol_ws; 0.5 mg) on days −2, 0, and 2 (Song et al., 2008). The benefit of using estradiol_ws is that serum estradiol concentrations return to physiological levels within 24 h after administration, and in our treatment protocol, Gc is not exposed to abnormal levels by day 3 of infection as shown. The dotted line corresponds to estradiol levels in normal proestrus stage mice (Dalal et al., 2001) Mice are inoculated with bacteria on day 0. Not shown is the antibiotic treatment regimen, which has been adjusted over time due to changes in breeders and husbandry practices. Currently, streptomycin sulfate (Sm; 2.4 mg) and vancomycin hydrochloride (0.4 mg) are administered via intraperitoneal (i.p.) inoculation twice daily until day 2 of the infection period. Drinking water with trimethoprim sulfate (0.04 g/100 ml water) is used over the course of the experiment, with Sm (0.5 g/100 ml water) added on day 2 of the infection period. (B) An example of the average number of CFU recovered from a single vaginal swab suspended in 100 μl of PBS (left axis) and percent of PMNs in stained vaginal smears (right axis) is shown. Mice (n = 6) in this experiment were inoculated with 10⁶ CFU of strain FA1090. In published (Packiam et al., 2010) and unpublished data from our laboratory, increased levels of proinflammatory cytokines and chemokines are detected on day 5 of infection as shown, which corresponds to the influx of PMNs in infected mice. Asterisks denote time points at which the percentage of PMNs in infected mice (dashed lines, solid squares) was significantly higher than that in mice given PBS (dashed lines, open squares). Increased PMNs were observed in control mice on days 9 and 10, which is due to the effects of the estradiol wearing off and the resumption of the estrous cycle. Mice were colonized for an average of 9.9 days in this 12-day experiment. This figure was constructed with combined data reported in (Song et al., ; Packiam et al., 2010).

Figure 2

Gc interactions with innate defenses in the murine genital tract. Gc encounters a variety of innate defenses in the female genital tract including hydrophobic antimicrobial substances that bathe the mucosal surface, complement, and commensal flora (Boris and Barbes, 2000). Pathogen-induced activation of innate receptors on epithelial and immune cells causes secretion of antimicrobial peptides and proinflammatory cytokines and chemokines (Kolls et al., 2008), which recruit phagocytes to the infection site. PMNs can take up bacteria that are opsonized with complement deposition products. Phagocyte-produced reactive oxygen species (ROS) kill bacteria by damaging DNA, protein, and other macromolecular structures (Storz et al., 1990). PMNs also kill microbes via the release of pre-formed enzymes and antimicrobial peptides and entrapment in neutrophil extracellular traps (NETs) where they are exposed to toxic molecules produced by both pathways (Kobayashi et al., ; Papayannopoulos and Zychlinsky, 2009). (A) Stained vaginal smear from a Gc-infected mouse reveals the presence of nucleated (NUC) and squamous (SQ) epithelial cells and PMNs with intracellular diplococci. Black arrows denote Gc and white arrows denote lactobacilli. Human strains of H₂O₂-producing lactobacilli can be added to this ecosystem as described (Muench et al., 2009). (B) Cartoon depicting the murine genital mucosa on the left and some of the bacterial factors that have been tested to measure their role in protection from host innate defenses on the right. Gc has many redundant anti-oxidant systems including Ccp (cytochrome c peroxidase), Kat (catalase), MsrA/B (a methionine sulfoxide reductase), and the MntA–MntB–MntC manganese transporter, which do not protect Gc from phagocyte-derived (Kat, Ccp, MntC; Wu et al., 2009), or Lactobacillus-derived (Kat, Ccp) ROS in vivo (Muench et al., 2009). Gc add host sialic acid to their LOS via the action of sialyltransferase. This modification reduces uptake of the bacteria by PMNs and increases survival in vivo (Wu and Jerse, 2006). Antimicrobial peptides (APs) are actively expelled from the cell via the MtrC–MtrD–MtrE efflux pump and inactivation of this system is highly attenuating for murine infection (Jerse et al., 2003).

Figure 3

The reproductive cycle influences Gc infection of women and mice. (A) Culture rates from women with gonorrhea and the opacity phenotype of cervical isolates from infected women with respect to stages of the menstrual cycle (Koch, ; Johnson et al., ; James and Swanson, ; McCormack and Reynolds, 1982). In one study, women with gonorrhea were hospitalized without treatment and in four of four women, positive cultures were followed by five to six consecutive negative cultures during the secretory phase; cultures became positive again at menses (Koch, 1947). Gonococcal PID and disseminated gonococcal infection most frequently occur at or shortly after menses (Holmes et al., 1971) and Gc from fallopian tubes from women with salpingitis were reported to be Opa⁻ (Draper et al., 1980). (B) Cartoon depicting the cyclical recovery pattern seen in intact mice following inoculation with mostly Opa⁻ variants of strain FA1090. Within a day after inoculation with mostly Opa⁻ Gc, Opa⁺ variants predominate (early phase). This phase is followed by a period in which mostly Opa⁻ variants are isolated (mid-phase) and then a second Opa⁺ phase (late phase). A high percentage of isolates express multiple Opa proteins in the late phase. In mice that remain infected for more than 8 days, a second mid-phase is observed. The rise and fall of the Opa⁺ population corresponds to fluctuations in the total number of Gc recovered. This pattern is not seen in Ov⁻ mice. Hostile factors may reduce colonization during the mid-phase or perhaps Opa⁺ variants are less accessible for culture due to tissue invasion (Simms and Jerse, ; Cole et al., 2010).