Dectin-1 stimulating β-glucans inhibit Chlamydia infections both in vitro and in vivo

Abstract

Chlamydia trachomatis and Candida albicans are common inhabitants of the female genital tract. Candida albicans can impact the viability and pathogenesis of some bacteria. Previously, we investigated physical interactions between Ch. trachomatis elementary bodies (EBs) and Ca. albicans. This work indicated that EBs bind to Ca. albicans and become noninfectious by 24 h post-binding. Here, we continue our investigation of these interkingdom, polymicrobial interactions. Candida albicans adheres to bacteria or host surfaces via agglutinin-like sequence or heat shock 70 (Ssa) proteins. Chlamydia trachomatis EBs did not bind Ca. albicans Ssa2 deficient strains as efficiently as wild-type or complemented strains, indicating a role for this protein in chlamydial adherence to Candida. Additionally, Ca. albicans β-glucans inhibit chlamydial infection when exposure occurs during EB adsorption onto cervical cells. Laminarin, a β-glucan agonist of the C-type lectin receptor Dectin-1, inhibited chlamydial infection in both cervical epithelial cells and mice when exposure occurred prior to, during, or immediately following EB inoculation. Conversely, a Dectin-1 antagonist laminarin did not inhibit infection in vitro, suggesting that β-glucan inhibition of Ch. trachomatis requires C-type lectin receptor signaling. Overall, our data demonstrate that β-glucans from multiple species, including Ca. albicans, inhibit Chlamydia via stimulation of host-signaling pathways.

Chlamydia trachomatis, Candida albicans, bacterial–fungal interactions, opportunistic infections, β-glucan, vaginal microbiome

Introduction

Chlamydia trachomatis is a Gram-negative, obligate intracellular bacterium (Wyrick 2000). Chlamydiae transition between two forms during their developmental cycle. The extracellular, infectious form of the bacterium, called an elementary body (EB), binds to one of the several receptors on the surface of host cells, initiating uptake. Once inside the host cell, EBs convert to the larger, noninfectious, metabolically active form, called the reticulate body (RB). RBs replicate within a modified vacuole, an inclusion, using host resources for nutrients and energy. Following several replicative cycles, RBs convert into EBs, which are released from the host (Elwell et al. 2016). Chlamydial infections are the leading cause of bacterial sexually transmitted infections in the USA and worldwide with ∼1.6 million Ch. trachomatis genital infections reported in the USA annually (CDC, 2024). Genital chlamydial infections most often present as cervicitis in females or urethritis in males, but many are asymptomatic. If left untreated, chlamydial infections can ascend the genital tract, increasing the risk for sequalae such as pelvic inflammatory disease, ectopic pregnancy, and tubal infertility (Menon et al. 2015, Workowski and Bolan 2015, Russell et al. 2016). Recent work indicates that disruptions in the vaginal microbiome correlate with altered rates of chlamydial infections in humans. Specifically, chlamydial infections appear to be enhanced in the presence of bacterial vaginosis related species (Dehon Mott et al. 2021, Bommana et al. 2022, Klasner et al. 2024). Chlamydial infections are also altered by Lactobacillus species and are linked to a reduction in Lactobacillus in vivo (Chen et al. 2022).

Candida species are some of the most common fungal members of the human microbiome, being found in the oropharyngeal, gastrointestinal, skin, and vaginal microbiota (Mayer et al. 2013). Given their ubiquitous nature, it is not surprising that Ca. albicans is a major opportunistic fungal pathogen. Candida albicans is a polymorphic fungus that exists as yeast or invasive hyphal forms within the body during infection (Mayer et al. 2013). Candida albicans causes a wide range of infections from life-threatening fungal sepsis to local mucosal infections such as vulvovaginitis. It is estimated that 75% of women will experience at least one episode of vulvovaginal candidiasis during their lifetime, and many will suffer recurrent episodes (Achkar and Fries 2010, Allison et al. 2016). Several reports document interkingdom communication and interactions between Candida and their bacterial neighbors (Peleg et al. 2010). Depending on the composition of the microbiota and the environment, Ca. albicans/bacteria interactions can be symbiotic or antagonistic (Shirtliff et al. 2009, Nogueira et al. 2019). Several bacteria inhibit Ca. albicans growth and/or virulence. Pseudomonas aeruginosa inhibits Ca. albicans growth via release of quorum-sensing molecules and other soluble products (Hogan et al. 2004, Morales et al. 2013). Likewise, Ca. albicans augments P. aeruginosa quorum-sensing signals to impact bacterial growth (Cugini et al. 2007). Candida albicans also prevents growth of Neisseria gonorrheae in vitro (Kaye and Levison 1977). Conversely, other bacterial pathogens, such as Staphylococcus, benefit from attachment to Ca. albicans (Peleg et al. 2010, Peters et al. 2012, Kong et al. 2015, Pidwill et al. 2018).

While Ca. albicans is generally considered an opportunistic vaginal pathogen and Ch. trachomatis initially infects the cervix, EB is likely to encounter Candida present in the vaginal canal during transmission. Studies suggest that Ca. albicans resides in the cervix as well (Boeke et al. 1993, Afiuni-Zadeh et al. 2018). In fact, Og et al. (2010) found Candida in 30.1% of cervical smears collected using Sabouraud dextrose agar compared to 7% of traditional cervical Papanicolaou smears, suggesting that cervical Candida is more common than previously thought (Og et al. 2010). Candida albicans has also been detected in amniotic fluids of pregnant patients who had an intrauterine device or premature delivery (Marelli et al. 1996, Ito et al. 2013, Kusanovic et al. 2018). Thus, it is likely that Ch. trachomatis and Ca. albicans can interact at multiple points—the vaginal canal, cervix, or endometrium—during transmission. Despite the common nature of these two microbes, very few studies have examined the potential for direct interactions between them (Kelly et al. 2001, Kruppa et al. 2018, Filardo et al. 2019). In our previous study, we incubated Ch. trachomatis serovar E EB (CtE) with yeast or hyphal cultures of Ca. albicans strain SC5314. Data from these studies indicate that EBs bind to both Ca. albicans yeast and hyphae (Kruppa et al. 2018). Initially, EBs bound to Candida remain infectious, but they lose their infectivity when bound to Candida for extended periods of time. Specifically, EBs lose infectivity by 6 h post-binding (hpb) and are rendered completely noninfectious by 24 hpb (Kruppa et al. 2018). Herein, we expand upon our previous work to determine the Candida structures that Ch. trachomatis EBs bind. We also further examined the impact of EBs exposure to β-glucans, including Ca. albicans surface exposed β-glucans.

Materials and methods

Preparation of biological reagents and culture of Candida, Chlamydia, and cervical cells

Candida albicans β-glucans (CaBGs) were isolated from the wild-type (WT) strain SC5314 (Fonzi and Irwin 1993) as previously described (Kruppa et al. 2011). Dectin-1 agonist laminarin (Lam) and Dectin-1 antagonist laminarin (Lam-Ant) were purchased from Carbosynth and Sigma, respectively. Fluorescently labeled Lam and Lam-Ant were prepared by 1,3-diaminopropane derivatization and incubation with AlexaFluor 488 as described (Borriello et al. 2022). Purification and activity validation of these biomolecules have been previously described (Williams et al. 1991, Adams et al. 2008, Kruppa et al. 2011, Smith et al. 2018). All Candida strains were cultured overnight at 30°C in 1% yeast extract, 2% peptone, and 2% dextrose. Candida glabrata strain BG2 was a kind gift from Brendan Cormack (Cormack and Falkow 1999). Agglutinin-like sequence (Als) and Ssa-deficient mutant and complemented Ca. albicans strains (Table S1) were kind gifts from Dr. Lois Hoyer and Dr. Mira Edgerton, respectively (Li et al. 2006, Zhao et al. 2004, 2005, 2006, 2007a, 2007b, Pidwill et al. 2018, Oh et al. 2022). Chlamydia trachomatis serovar E and Chlamydia muridarum Weiss stocks were prepared in Hec-1B cells as previously described (Guseva et al. 2007). HeLa cells were maintained in Modified Eagle's Medium (Gibco) supplemented with 10% Fetal Bovine Serum (FBS), and gentamicin (Gibco). Short tandem repeat profiling was performed by the American Type Culture Collection to authenticate the identity and/or origin of all cell lines used in this study. All cell lines were tested for Mycoplasma by polymerase chain reaction (PCR) and found to be free of contamination (data not shown).

Candida/Chlamydia binding assay

Water, or Ca. albicans SC5314 (Ca, CaWT), Als, or Ssa deficient mutant (M) and complemented (C) strain yeast (1 × 10⁵ cells sample⁻¹) from overnight cultures were cultured on FBS-coated tissue culture plates for 3 h in HeLa culture medium at 37°C to promote adherence and hyphal formation. Following adherence and hyphal growth, Candida cultures were overlaid with 3 × 10⁵  Ch. trachomatis serovar E EB (CtE) or chlamydial inoculum diluent, 0.2 M sucrose, 0.02 M phosphate buffer, and 5 mM glutamine (2SPG) for 1 h. The cultures were then washed with phosphate buffered saline (PBS) to remove unbound EB and harvested for infectivity analysis by EB titer assays. In separate experiments, H₂O, Ca, or Ca. glabrata (Cg) yeast were plated in FBS-coated wells in HeLa culture medium and allowed to adhere for 1 h at 37°C. Wells were then overlaid with 2SPG or EB in 2SPG for 1 h, washed and collected for DNA isolation.

EB β-glucan exposure

Chlamydia trachomatis or Ch. muridarum EBs (3 × 10⁵ sample⁻¹) were exposed to β-glucans prior to, during, or immediately after inoculation onto HeLa cultures. In preincubation experiments, EBs were incubated with H₂O, CaBG (100 ng ml⁻¹), Lam, or Lam-Ant (0.1–2 mg ml⁻¹) in 2SPG for 1 h at 37°C with gentle agitation. β-Glucan-exposed EBs were then inoculated onto HeLa monolayers for 1 h at 4°C (binding assays) or 37°C. In a subset of experiments, β-glucan-exposed EBs were centrifuged (8000× g), washed with PBS to remove unbound β-glucans, and resuspended in 2SPG before HeLa culture inoculation at one multiplicity of infection. For exposure during inoculation, β-glucans were added to the chlamydial inoculum for the 1 h EB adsorption period on HeLa cultures. The bacterial inoculum was then removed and cultures were replenished with HeLa culture medium. Alternatively, Chlamydia-infected cultures were exposed to β-glucans after inoculation by adding H₂O, CaBG, Lam, or Lam-Ant to the culture medium after removal of the EB inoculum (0 hpi). In separate experiments, H₂O, Lam, or Lam-Ant diluted in culture medium was added to HeLa monolayers for 1 h at 37°C. β-Glucans were then removed by aspiration and the monolayers were washed with PBS before chlamydial inoculation. The inoculum was removed after 1 h at 4°C. The cultures were washed with PBS and fixed or replenished with tissue culture medium. Samples for binding assays were formaldehyde fixed after 1 h incubation at 4°C. All other infected cultures were incubated for 24 (Ch. muridarum) or 48 h (Ch. trachomatis) at 37°C before harvest for various analyses.

EB titer assays

Ca/CtE binding and tissue culture samples were scraped into 2SPG, collected, and frozen at −80°C. Vaginal swabs collected from Ch. muridarum infected mice were placed in vials with 2SPG and glass beads and frozen at −80°C. EBs were released from infected cells by freeze/thaw and sonication. Samples were then diluted in culture medium with 0.5 µg ml⁻¹ cycloheximide, and gentamicin. Vancomycin and/or amphotericin B (5 µg ml⁻¹) were added to vaginal swab samples or tissue culture samples containing Candida. The diluted samples were used to infect HeLa monolayers grown on coverslips by spin infection (1 h, 1100× g). Replicate infected monolayers were incubated for 24 or 48 h at 37°C before methanol fixation and stained with Bio-Rad Pathfinder anti-Chlamydiamajor outer membrane protein (MOMP) (Ch. trachomatis) or LPS (Ch. muridarum) stain. The number of inclusions in 10 reticule fields was counted, averaged, and used to determine the number of inclusion forming units (IFU) present in 1 ml of in vitro sample or vaginal swab.

PCR and reverse transcription-PCR (rt-PCR)

DNA or RNA was isolated via Qiagen QIAamp or RNeasy Kits according to the manufacturer’s instructions. Chlamydial 16S rRNA and human gapdh genes were amplified by PCR (Invitrogen) as described previously (Deka et al. 2006). HeLa cell RNA was reverse transcribed to cDNA (Qiagen) before PCR amplification of clece7A (Qiagen QuantiTect Primer assay #QT00024059). PCR amplimers were visualized on an ethidium bromide-stained agarose gel and quantified with a BioRad G-box/Syngene software in the ETSU Molecular Biology Core Facility (RRID: SCR_021 106). A serially diluted set of positive controls were included in each PCR to ensure that amplification conditions were within the linear range of the reaction.

Immunofluorescence assays measuring EB binding to host cells or laminarin

Formaldehyde-fixed, β-glucan-exposed, infected tissue culture samples were stained with anti-MOMP primary antibody (Abcam) and a Donkey anti-Mouse Alexa Fluor 594 secondary antibody (Thermo Fisher) or Pathfinder anti-MOMP (Bio-Rad) stain. Images were captured with a Ziess Axiovert microscope at 40× magnification. The intensity of MOMP fluorescent staining was measured area⁻¹ (µm²) in 10 fields (40× magnification) and averaged using Ziess Zen Software. EBs were incubated with Lam488 or Lam-Ant488 as described above. Following incubation, unbound β-glucan was removed by centrifugation. The samples were fixed and stained with anti-MOMP primary antibody (Abcam) and a Donkey anti-Mouse Alexa Flour 594 secondary antibody. Fluorescently labeled Lam, Lam-Ant, and MOMP were visualized at 60× magnification using an Amnis ImageStream flow cytometer.

Pronase A digestion of Candida surface proteins

Candida albicans yeast (1 × 10⁶ sample⁻¹) were incubated with water or pronase A (Sigma, 0.1 mg ml⁻¹) for 1 h at 37°C. After exposure, pronase A was removed by centrifugation and PBS washes. The water or pronase A-exposed yeast samples were incubated with EBs as described above. Immediately after the binding period, the yeast with bound EBs were collected by centrifugation and used to infect HeLa cultures in EB titer assays. Replicate samples were methanol-fixed and immunostained with 4',6-diamidino-2-phenylindole (DAPI) (Thermo Fisher) and anti-chlamydial MOMP stain (Bio-Rad). Fluorescent staining intensity was visualized (60× magnification) and quantified using an Amnis ImageStream flow cytometer. Co-localization of DAPI and MOMP signals in 1000 events was quantified with IDEAS image analysis software.

Chlamydia muridarum laminarin exposure in vivo

Groups of 6-week-old, female Balb/c mice (Jackson Laboratories) were injected subcutaneously with Depo-Provera (2 mg kg⁻¹). One week later, the mice were exposed to Lam (2 mg ml⁻¹) or H₂O 24 h before (LamB), during (LamD), and after (LamA) or before, during, and after (LamBDA) Ch. muridarum vaginal infection. Ten microliters of chlamydial inoculum or Lam diluted in 2SPG was pipetted into the vaginal canal to expose and/or infect the mice. Bacterial shedding was monitored by collection of vaginal swabs every 3 dpi for 21 days (Gravitte et al. 2022). Dilutions of swab samples were used in EB titer assays as described. All animal research was by the ETSU Committee on Animal Care, accredited through the Assessment and Accreditation of Laboratory Animal Care.

Statistical analysis

Unless otherwise noted, all in vitro experiments were performed in triplicate, with each independent experiment containing three biological replicates. In vivo studies contained a minimum of eight animals per group across three repeated studies. Values from independent experiments were averaged and presented as the mean ± the standard error of the mean (SEM). Data were analyzed with Excel and GraphPad Prism 10 software. Means were compared by analysis of variance (ANOVA) and subsequent independent two-sample t-tests with Tukey or Dunnett's multiple comparison corrections where appropriate. Comparisons with P-values ≤ .05 were considered significantly different.

Results

Chlamydia trachomatis EB bind specific Ca. albicans surface proteins

Previously, we demonstrated that preincubation of EBs with fungal mannans or β-glucans did not prevent Ca/CtE binding, suggesting that EBs bind to a different structure on Ca. albicans (Kruppa et al. 2018). To determine whether Ch. trachomatis can bind to a specific protein on the surface of Candida, we digested the surface proteins of Ca. albicans yeast with Pronase A prior to incubation with Ch. trachomatis EBs. Protease digestion decreased Ca/CtE binding as evidenced by: (i) decreased co-localization of chlamydial MOMP immune labeling with C. albicans yeast (Fig. 1a, b) and (ii) decreased recovery of infectious EBs bound to Ca. albicans at 0 hpb (Fig. 1c). These data indicate that Ch. trachomatis EBs bind to a specific protein on the surface of Ca. albicans. Several reports have documented that Gram-positive and Gram-negative bacteria bind to Ca. albicans (Shirtliff et al. 2009, Peleg et al. 2010, Nogueira et al. 2019). Streptococcus gordonii, Group B Streptococcus, and Staphylococcus aureus all bind to Ca. albicans via Als proteins (Klotz et al. 2007, Peters et al. 2012, Hoyer et al. 2014, Pidwill et al. 2018). Another group of surface-exposed Candida proteins that are important for attachment/invasion are the heat shock 70 proteins, Ssa1, and Ssa2 (Li et al. 2006, Sun et al. 2010). Based on these observations, we measured EB binding to a series of Ca. albicans Als 1–9 or Ssa deficient mutant strains (Figs 1d and S1) (Zhao et al. 2004, Zhao et al. 2006, 2007a, 2007b, Pidwill et al. 2018). First, we examined EB binding to the background Ca. albicans strains used to generate the Als and Ssa deficient and complemented strains, CAI12 and CAF4-2, respectively (Table S1). There was no significant difference in EB binding to Ca. albicans CAI12 or CAF4-2 and our laboratory WT Ca. albicans strain SC5314 (Fig. S1A). Thus, we used strain SC5314 as the WT control in subsequent experiments. Data from these studies indicate that EB binding to the Ca. albicans Ssa2 deficient strain (CaSsa1M) was reduced by >40% compared to the WT or complemented (CaSsa1C) strains (Fig. 1d), suggesting that EBs bind to this specific protein. EB binding to an Als3 deficient Ca. albicans strain (CaAls3M) was also reduced, albeit not significantly compared to the WT (Fig. 1d). In fact, we observed a decreased trend in EB binding to several Ca. albicans Als deficient strains, including Als5 and Als6 deficient mutants. However, EB binding was not restored to WT levels in the corresponding complemented strains (Fig. S1B), possibly due to differences in protein expression between the WT and complemented strains. (Zhao et al. 2007, Hoyer and Cota 2016).

Figure 1.

EBs bind to specific Ca. albicans surface proteins. (a−c) Ca. albicans (Ca) yeast cultures were exposed to H₂O or Pronase A, washed, and then incubated with Ch. trachomatis EB (CtE) for 1 h. EBs bound to yeast were collected by centrifugation and processed for immunofluorescent labeling (a and b) or EB titer assays (c). (a) Ca/CtE samples immunostained for MOMP and Candida nuclei (DAPI) were examined using the Amnis ImageStream flow cytometer (60× magnification). Representative images of Ca/CtE samples exposed to H₂O or Pronase A are shown. (b) Percentage of cells with MOMP/DAPI co-localization. (c) Chlamydial EB titer assay from HeLa cultures infected with Ca, CtE alone, or bound to H₂O or Pronase A-exposed Ca. (d) Chlamydial EB titer assays from CtE bound to hyphal cultures of Ca WT (SC5314), Als3, or Ssa2 deficient mutants (M) or complemented (C) strains. Data shown are means ± SEM (n = 6–9). Group means were compared by ANOVA and/or t-tests. *P ≤ .05, **P ≤ .01, ***P ≤ .001.

β-Glucans inhibit chlamydial entry and infectivity in human cervical cells

We next wanted to examine interactions between EBs and a common fungal cell wall structure, β-glucans. β-Glucans are β-d-glucose polysaccharides, which can be linear β(1–3) or branched β(1–3, 1–6) and are found in numerous plant and microbial species, including Ca. albicans (Lowman et al. 2014, Camilli et al. 2018). Fungal β-glucans serve as pathogen-associated molecular patterns to activate the immune response by ligation of C-type lectin receptors (CTLRs) such as Dectin-1. While Dectin-1 is best characterized in immune cells, it is also expressed on epithelial cells, including HeLa cells (Fig. S2A, B) and responds to β-glucan signaling in epithelial cells (Cohen-Kedar et al. 2014, Li et al. 2014, Yeh et al. 2017). Both here (Fig. 3a) and in our previous study, we have observed that pre-exposure to a highly purified, water-soluble β-glucan, laminarin, inhibits Ch. trachomatis infectivity by >80% (Kruppa et al. 2018). To determine whether this phenomenon is common to β-glucans, we exposed EBs to CaBGs. Preincubation of EBs with CaBGs decreased Ch. trachomatis infectivity modestly, but it was not decreased significantly in HeLa cells compared to EBs incubated in medium alone (Fig. 2a). Centrifugation and washing of the EBs following incubation with medium or CaBGs removed the inhibitory effect, suggesting that CaBG may inhibit chlamydial infection via interactions with host receptors (Fig. 2a). Indeed, CaBG exposure during chlamydial adsorption onto host cells significantly reduced chlamydial infection by 57% compared to the diluent control (Fig. 2b-c). Conversely, exposure to CaBG after EB inoculation caused no change in chlamydial infection (Fig. 2d). Unfortunately, purified CaBG, unlike Lam, is not water-soluble; thus, it is added to samples as a suspension, potentially limiting the physical interaction between CaBG and EB or host cells.

Figure 2.

CaBGs inhibit chlamydial infection in cervical cells. (a) Chlamydial EB titer assays from EB exposed to H₂O or CaBG prior to HeLa culture infection. (b-c) HeLa cultures were infected with Ch. trachomatis inoculum + H₂O or CaBG. At 48 hpi, the cultures were harvested for immunostaining (b) or EB titer assays (c). (b) Fixed monolayers were stained with Bio-Rad Pathfinder stain, MOMP, host cells, and DAPI nuclear stain. Representative images were captured with a ZEISS Axiovert (40× magnification) and Zen software. (d) Infected HeLa monolayers were exposed to H₂O or CaBG following EB inoculation (0 hpi). Cultures were harvested at 48 hpi for EB titer assays. Data shown are means ± SEM (n = 6–9). Means were compared by ANOVA and t-tests. **P ≤ .01.

Figure 3.

Laminarin inhibits chlamydial entry and infectivity in cervical epithelial cells. (a) Chlamydial EB titer assays from EB exposed to H₂O or Lam prior to HeLa culture infection. (b and c) EB binding assays. Replicate aliquots of Ch. trachomatis EB were preincubated with H₂O, Lam, or Lam-Ant before inoculation onto HeLa monolayers at 4°C for 1 h. Following the binding period, cultures were washed, fixed, and immunostained for MOMP. (b) Representative images. (c) Average intensity of MOMP fluorescent staining measured/area. (d and e) Uninfected HeLa monolayers were exposed to H₂O, Lam or Lam-Ant for 1 h. The cultures were then washed with PBS and inoculated with Ch. trachomatis 4°C for 1 h. Following the binding period, duplicate cultures were fixed for immunostaining (d) or used in EB titer assays (e). (d) Representative images of fixed monolayers immunostained with BioRad anti-MOMP. (f and g) EB titer assays. HeLa cultures were infected with Ch. trachomatis inoculum + H₂O or Lam (f) or exposed to H₂O or Lam at 0 hpi (g). (b and d) Images were captured and analyzed with a ZEISS Axiovert (40× magnification) and ZEN software. Data shown are means ± SEM (n = 6–9). Means were compared by ANOVA and t-tests. *P ≤ .05, ***P ≤ .001.

To investigate the effects of β-glucans on chlamydial infection further, we employed a Dectin-1 agonist Lam and a Dectin-1 antagonist Lam preparation (Lam-Ant). Lam has a molecular weight of 34.4 kDa, whereas Lam-Ant is smaller at 4.8 kDa. Previously, Smith et al. found that both Lam preparations are water-soluble, chemically pure, structurally uniform, and comprised primarily of β(1–3)-d-glucose linkages. Although both bind to Dectin-1, only Lam stimulated cytokine production in response to Dectin-1 binding (Smith et al. 2018). Preincubation with Lam reduces chlamydial infection in HeLa cultures compared to diluent controls (Fig. 3a). However, unlike CaBG, centrifugation and washing the EB/Lam mixtures to remove unbound laminarin from the chlamydial inoculum did not reverse the inhibitory effect of Lam on chlamydial infection compared to water exposure, suggesting that laminarin binds to chlamydial EB (Fig. 3a). This observation is supported by immunofluorescent imaging of EB/laminarin mixtures, which confirmed that EB bind to both Lam and Lam-Ant (Fig. S2C). EB preincubation with Lam significantly decreased EB binding to HeLa cells compared to H₂O, whereas Lam-Ant did not block EB binding to HeLa cells (Fig. 3b-c). Likewise, both EB binding to HeLa monolayers (Fig. 3d) and bacterial progeny production (Fig. 3e) were reduced when uninfected cultures were exposed to Lam, but not Lam-Ant or H₂O prior to chlamydial infection. Depyrogenation of Lam did not prevent the decrease in EB attachment to host cells, indicating that our observations are due to Lam, not a reagent contaminant (Fig. S2D). Lastly, we found that Lam, but not Lam-Ant exposure during (Fig. 3f) or immediately after (Fig. 3g) EB adsorption onto epithelial cultures, caused a 70% reduction in EB progeny production compared to the diluent control. Together these data indicate that inhibition of chlamydial infection by β-glucans is not limited by the origin of the β-glucan as both CaBG and laminarin, isolated from the seaweed Cystoseira barbata, inhibited chlamydial infection. Furthermore, these results suggest that Dectin-1 signaling prior to or early during infection may be required for inhibition of Ch. trachomatis by β-glucans.

β-Glucans inhibit Ch. muridarum infection in vitro and in vivo

Given our observations that β-glucans, particularly Lam, inhibit chlamydial infection in vitro, we sought to determine whether β-glucans could inhibit chlamydial infection in vivo. We first tested the effect of Lam on Ch. muridarum as it is a common chlamydial species used in murine models. Like Ch. trachomatis, Ch. muridarum infectivity was decreased with Lam exposure before and/or during EB adsorption onto HeLa monolayers (Fig. 4a and b), indicating that the inhibitory action of β-glucans on chlamydiae is not species specific. Next, we vaginally infected mice with Ch. muridarum in the presence of Lam before, during, and/or after infection. Infected mice exposed to Lam shed significantly fewer EBs compared to water-exposed animals [Fig. 4c (two-way ANOVA P = .01) and Fig. S3]. Mice exposed to laminarin before infection showed a 66% reduction in EB shedding 3 dpi compared to H₂O exposure (Fig. S3A). EB shedding was also reduced by >50% on both 3 and 6 dpi in mice exposed to Lam before, during, and after Ch. muridarum infection compared to the H₂O controls (Fig. S3B).

Figure 4.

Laminarin inhibits Ch. muridarum infection in vitro and in vivo. (a-b) EB titer assays. (a) Ch. muridarum EB (Cm) were incubated with H₂O or Lam for 1 h prior to infection. (b) HeLa cultures were infected with Cm inoculum + H₂O or Lam during infection. Data shown are means ± SEM (n = 6–9). Group means were compared by ANOVA and t-tests. ***P ≤ .001, ^****P ≤ .0001. (c) EB titer assays. Balb/c mice were vaginally infected with Cm inoculum + H₂O or Lam. Replicate groups were treated with Lam 24 h before (B), during (D), after (A) or before, during, and after (BDA) infection. Vaginal swabs collected on 3–21 dpi for EB titration. Data shown are means ± SEM. (8–16 mice/group). Means were compared by ANOVA and subsequent t-tests. Significant P-values are described in text.

Discussion

Given the high incidence of chlamydial infections and the ubiquitous nature of Ca. albicans in the vaginal microbiome, there are ample opportunities for these organisms to encounter one another during chlamydial transmission and infection. However, very few studies have examined the interactions between these microbial neighbors. Kelly et al. (2001) examined the effects of Ca. albicans and Ch. trachomatis infection on each other, but found that neither pathogen significantly altered the immune response or shedding of the other microbe in mice. Importantly, this study did not investigate the impact of direct physical interaction between the two microbes nor simultaneous infection with both pathogens in vivo. Filardo et al. (2019) concluded that Candida and Gardnerella biofilms could act as a reservoir for EB infection. However, they did not assess the number of EBs recovered from these biofilms to determine whether infectivity was altered compared to the original inoculum. Data from our previous study indicate that direct Ca. albicans/Ch. trachomatis interactions abolish EB infectivity after EBs remain bound to Ca. albicans for a period of hours (Kruppa et al. 2018). Here we demonstrate that (i) EBs bind to specific proteins on the surface of Ca. albicans and (ii) exposure of EB to β-glucans prior to or early during infection inhibits chlamydial host cell entry and progeny production in vitro and in vivo.

Als proteins, named due to their similarity to Saccharomyces cerevisiae α-agglutinin, are important components of the Candida cell surface. They are involved in the attachment of yeast and/or hyphal forms to abiotic or host surfaces (Hoyer and Cota 2016). Interestingly, EBs did not significantly bind to Candida glabrata (Fig. S4), which does not express Als homologs (Klotz et al. 2007, Silva et al. 2012), suggesting that Ch. trachomatis binding to Ca. albicans is species specific. Other bacteria, including Streptococcus gordonii, Staphylococcus aureus, and Escherichia coli, are known to bind to Candida Als proteins (Peleg et al. 2010, Peters et al. 2012, Kong et al. 2015, Pidwill et al. 2018). Bacterial binding to these proteins can provide a niche for bacterial growth and pathogenesis. In fact, S. aureus binding to Als3 is known to enhance bacterial dissemination to distal organs in a murine model of oral candidiasis (Peters et al. 2012, Kong et al. 2015). Here, we observed a reduced trend in EB binding to Als3, Als5, and Als6 deficient strains compared to WT Ca. albicans (Fig. S1B). These observations suggest that EBs may bind multiple Als proteins. Of these, only the Als3C strain restored EB binding to the WT level (Fig. 1d). It is possible that we did not observe significant differences in EB binding to specific Als deficient mutant strains, because Als proteins have redundant functions and varying expression levels in Ca. albicans yeast and hyphal forms (Hoyer et al. 1998, Zhao et al. 2004). However, Als1–5 have been detected on Ca. albicans hyphae and early germ tubes, which were present in our experiments (Hoyer et al. 1998, Zhao et al. 2004). Thus, while we cannot definitively conclude that EBs bind to specific Als proteins, this study suggests that an obligate intracellular bacterial pathogen can bind to Ca. albicans via Als proteins. Ssa proteins are surface-exposed heat-shock-like proteins important for Ca. albicans invasion. There have been no previous reports of bacteria binding to Ssa proteins. However, proteins in host receptor complexes that are involved in clathrin-mediated endocytosis, including Hsp70 and protein disulfide isomerase, have been linked to successful chlamydial entry into epithelial cells (Chang et al. 2002, Davis et al. 2002, Hybiske and Stephens 2007, Abromaitis and Stephens 2009, Hall et al. 2011). Here, we show evidence that EBs bind to these surface exposed heat-shock proteins, suggesting that the homologous nature between Candida and human Hsp70 proteins may provide the basis for this interaction. Initially, these studies along with observations by Filardo et al. (2019) led us to hypothesize that Candida may provide a reservoir of chlamydial infection. However, we observed that Ca/CtE binding causes chlamydiae to lose infectivity in a relatively short period of time. This observation prompted us to explore mechanisms of Candida inhibition on chlamydial infection.

Fungal β-glucans are an integral component of the Ca. albicans cell wall. During infection, they activate the immune response by ligation of CTLRs. A myriad of positive or negative health-related outcomes have been linked to β-glucan stimulation of host receptors (Kurashige et al. 1997, Ikeda et al. 2007, Camilli et al. 2018). Interactions between fungal glucans and their primary receptor, Dectin-1, along with toll-like receptor (TLR) cross-talk, and the microbiome can positively or negatively impact colitis (Iliev et al. 2012, Heinsbroek et al. 2015, Tang et al. 2015). Exposure to glucans promoted antibiotic tolerance in fungal/bacterial co-species biofilms, and increased Salmonella infection in mice (de Brucker et al. 2015, Kuda et al. 2015). Conversely, a glucan-containing cream prepared had antibacterial activity against both Gram-positive and Gram-negative bacteria, including E. coli, and promoted healing in a rat wound model (Sellimi et al. 2018). Recent studies indicate that the innate immune system can be trained to respond more rapidly and effectively to pathogens by exposure to fungal β-glucans (Cheng et al. 2014, Ifrim et al. 2014, Saeed et al. 2014, Camilli et al. 2018). When considering the wide range of reported glucan effects on disease, it is important to note that β-glucan preparations vary between studies in their source, purity, and receptor-stimulating activity, potentially contributing to the positive or negative outcomes observed (Smith et al. 2018). Thus, it is important to employ a high-purity β-glucan with known structure and activity, as we have done here with CaBG and Lam.

Interestingly, both chlamydiae and β-glucans bind to host cells using a variety of CTLRs (Kuo et al. 2002, Puolakkainen et al. 2005, Campbell et al. 2012, Subbarayal et al. 2015). β-Glucans, including laminarin, bind to CTLR on the surface of a variety of mammalian cells. Swidergall et al. demonstrated that β-glucans stimulate cell signaling pathways via interaction with the Ephrin receptor A2 (EphA2) on the surface of lung epithelial cells (Swidergall et al. 2019). However, the best characterized β-glucan receptor is Dectin-1, which activates NFκB signaling via the spleen tyrosine kinase/caspase recruitment domain 9 pathway when stimulated by glucans. This process increases the expression of inflammatory cytokines, including interferons, which are inhibitory to chlamydial infection (Fadel and Eley 2007, Geijtenbeek and Gringhuis 2009, Blazer et al. 2010, Cohen-Kedar et al. 2014, Shiokawa et al. 2017, Swidergall et al. 2019). In response to β-glucan ligation, Dectin-1 also modulates immune responses by cross-talk with the TLR2 receptor (Gantner et al. 2003). Interestingly, Cheng et al. (2014) proposed an intriguing mechanism for immune training in which Dectin-1-ependent signaling facilitates training of the metabolic response to infection and augments antimicrobial immunity. Their findings are consistent with work demonstrating that glucan treatment augments the host response to infection with a variety of pathogens (Williams et al. 2004, Romero et al. 2011, Stark et al. 2015). While Dectin-1 is not a known Ch. trachomatis host receptor, the CTLR, macrophage mannose receptor, and EphA2 are known to aid in host entry (Kuo et al. 2002, Subbarayal et al. 2015). Our data suggest that β-glucans, particularly Lam, may block chlamydial host cell receptors such as EphA2 and possibly Dectin-1, as both are expressed on HeLa cells (Fig. S2B). Alternatively, given that Lam-Ant binds to EBs, but did not inhibit chlamydial infection, β-glucans may inhibit Ch. trachomatis by activating anti-chlamydial host signaling pathways. Furthermore, Lam reduced Ch. muridarum shedding in mice even when Lam exposure occurred after bacterial inoculation, suggesting that β-glucan activation of the immune response may limit chlamydial infection in vivo.

In this study, we focused our investigations on ways that Ca/CtE binding could impact early events during chlamydial infection in vitro. However, chlamydiae can also interact with CTLR at other times during their developmental cycle. Following entry, Ch. trachomatis recruits EphA2 to the surface of the growing inclusion and uses it to activate phosphoinositide 3-kinases, and mitogen-activated protein kinase signaling pathways. These pathways benefit chlamydial development by promoting antiapoptotic signals in the host cell (Subbarayal et al. 2015). Thus, while our in vivo data indicate that β-glucan exposure reduces chlamydial shedding, it is possible that direct stimulation of chlamydia infected cells by β-glucans could promote chlamydial development depending on the timing of the glucan-induced signal. Here, we also investigated chlamydial interactions with a single structural element in the Candida cell wall. In reality, it is likely that Candida-bound EBs will be exposed to multiple fungal structures and molecules, including secreted molecules such as quorum-sensing molecules and candidalysin. Studies have shown that these secreted products can inhibit bacterial growth (Hassan Abdel-Rhman et al. 2015, Arias et al. 2016, Liang et al. 2024), suggesting that Candida inhibition of chlamydial infection may occur through multiple mechanisms.

Overall, this work highlights the fact that interkingdom interactions between the human microbiome and bacterial pathogens can alter a pathogen’s ability to establish infection. There are numerous studies examining the effects of interactions between commensal and pathogenic bacteria in the body. However, there remains a lack of knowledge about interactions between pathogenic bacteria and the fungal microbiome. Thus, further research is needed to fully understand (i) the extent that fungal microbiome species interact with pathogenic microbes and (ii) the consequences of these interactions on infection and disease.

Acknowledgments

The authors would like to thank the ETSU Division of Laboratory Animal Resources, Molecular Biology Core Facility (RRID: SCR_021106), and the Center for Excellence in Inflammation, Infectious Disease and Immunity for their assistance in this project. We would also like to thank Kenton Hall and Dr. Regenia Campbell for their assistance and helpful conversations regarding this work. Lastly, we are especially grateful for the generosity of Dr. Cormack, Dr. Lois Hoyer, and Dr. Mira Edgerton in sharing their Candida strains with us. The featured image for this article was created with BioRender.com.

Author contributions

Jennifer Kintner (Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing), Morgan Callaghan (Data curation, Formal analysis, Investigation, Writing – review & editing), Lillith Bulawa (Data curation, Formal analysis, Investigation, Writing – review & editing), Angela Chu (Data curation, Formal analysis, Investigation, Writing – review & editing), Zuchao Ma (Methodology, Resources, Validation, Writing – review & editing), David L. Williams (Funding acquisition, Methodology, Resources, Validation, Writing – review & editing), Robert V. Schoborg (Conceptualization, Methodology, Resources, Writing – review & editing), Michael D. Kruppa (Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Writing – review & editing), and Jennifer V. Hall (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing)

Conflict of interest

None declared.

Funding

This work was supported by the National Institutes of Health (AI159877 to M.D.K., RO1GM119197 and R21AI173607 to D.L.W., and R15AI117632 to J.V.H.), the Carroll H. Long endowment to D.L.W., and ETSU Research Development Committee Major Grant and Quillen Research Enhancement Award to J.V.H.