All the 15 Lactobacillus strains included in this study (Figure 1) were previously isolated from vaginal swabs of healthy premenopausal Caucasian women (Parolin et al., 2015). Lactobacilli were grown in de Man, Rogosa and Sharpe (MRS) broth supplemented with 0.05% L-cysteine, for 18 h at 37°C, in anaerobic atmosphere. Anaerobic conditions were achieved by using anaerobic jars supplemented with GazPack EZ (Becton Dickinson and Company, Sparks, MD, United States).

As specificity control, clinical isolates of Streptococcus agalactiae and Enterococcus faecalis were used. These strains were part of a broad collection including bacteria isolated from vaginal swabs submitted to the Microbiology Laboratory of Sant’Orsola-Malpighi University Hospital of Bologna (Italy) for routine diagnostic procedures. These two strains were grown in Brain Heart Infusion (BHI) broth, in 5% CO₂ for 18 h at 37°C. Bacillus subtilis ATCC 31324 was also included as rod-shaped bacterium control. It was grown aerobically in BHI broth for 18 h at 37°C. All culture media were purchased from Becton Dickinson and Company.

Bacterial cultures were centrifuged at 5,000 × g for 10 min at 4°C. Cell pellets were washed and re-suspended in sterile saline to obtain stock cell suspensions of 5 × 10⁸ CFU/mL.

In this study, CT strain GO/86, serotype D, was used (Marangoni et al., 2015; Nardini et al., 2016). This strain was clinically isolated from a urethral swab of a patient with non-gonococcal urethritis, submitted to the Microbiology Laboratory of Sant’Orsola-Malpighi Hospital of Bologna, for routine diagnostic procedures and belongs to the laboratory collection. HeLa cells (ATCC CCL-2), a cell line originated from a human cervix carcinoma, were grown to confluent monolayers in 5% CO₂ at 37°C, inside a Forma Series II 3110 Water-Jacketed CO₂ Incubator (Thermo Fisher Scientific, Waltham, MA, United States). CT was propagated in HeLa cells, cultured in Dulbecco’s minimal essential medium (DMEM) (EuroClone, Pero, Italy), supplemented with 10% fetal bovine serum, 1% L-glutamine 200 mM, and antibiotics (vancomycin 10 mg/L, gentamicin 10 mg/L, and amphotericin B 0.3 mg/L). For the preparation of EBs, confluent HeLa cells were infected with CT in DMEM medium supplemented with cycloheximide 1 μg/mL, centrifuged at 640 × g for 2 h to facilitate cell penetration, then incubated at 37°C with 5% CO₂ for 48 h (Mastromarino et al., 2014). HeLa cells were then detached and fragmented by sonication, by using a Bandelin sonicator at minimum power. Samples were centrifuged at 500 × g for 10 min at 4°C, and supernatants, which contain EBs, were further centrifuged at 40,000 × g at 4°C for 1 h. The resulting pellets, containing the purified EBs, were re-suspended in sucrose-phosphate-glutamate (SPG) 0.2 M, divided into small aliquots and stored at −80°C.

To study the capacity of lactobacilli cells to interfere with the entry of CT EBs into HeLa cells, exclusion experiments were performed (Parolin et al., 2015). HeLa cells were seeded at 2 × 10⁴ cells/cm² in individual tubes containing sterile glass coverslips in 1 mL of medium, and allowed to reach 95–100% of confluence (approximately 1.25 × 10⁵ cells). Lactobacillus cells (0.1 mL of the stock cell suspension, corresponding to 5 × 10⁷ CFU) were incubated for 1 h at 37°C with 5% CO₂ on HeLa cells. Afterward, 5 × 10³ CT EBs (0.01 mL of a stock solution of 5 × 10⁵ EBs/mL) were added and further incubated for 1 h [multiplicity of infection (MOI) = 0.04]. HeLa tubes infected with 5 × 10³ CT EBs without lactobacilli were used as controls. All the experiments were conducted in triplicate. At the end of the incubation, each HeLa tube was first washed three times with phosphate buffer saline (PBS), then 1 mL of complete medium, supplemented with 1 μg/mL cycloheximide, was added. Individual tubes were centrifuged at 640 × g for 2 h to facilitate cell penetration and then incubated at 37°C with 5% CO₂ for 48 h. CT infection was evaluated by counting chlamydia inclusion forming units (IFUs) by direct immunofluorescence, using a monoclonal antibody against the chlamydial membrane lipopolysaccharide antigen conjugated with fluorescein (Meridian, Cincinnati, OH, United States), as previously reported (Marangoni et al., 2015). Slides were observed under epi-fluorescence microscope (Eclipse E600, Nikon, Tokyo, Japan) equipped with a super high pressure mercury lamp and Plan Fluor DLL 20, 40, 100× lenses. The number of IFUs was counted in 30 randomly chosen 200× microscopic fields. Results were expressed as the percentage (median percentage ± median absolute deviation) of CT infectivity, comparing the number of IFUs of the single experiments with the control tubes.

In order to exclude a potential effect of an acidic environment on CT EBs infectivity, the pH values of the culture medium were measured before the addition of lactobacilli cells and at the end of the incubation.

Exclusion experiments were also performed with S. agalactiae, E. faecalis, and B. subtilis cells instead of lactobacilli, applying the same conditions described above, to assess the specificity of the anti-CT activity.

Lactobacillus crispatus BC4, L. crispatus BC5 and L. gasseri BC14, selected among the most active strains in counteracting chlamydial infectivity, were used in dose-effect experiments. Specifically, total amounts of 5 × 10⁶ and 5 × 10⁵ lactobacilli cells were tested.

Lactobacillus crispatus BC5 was chosen as a model strain to study the molecular mechanisms underlining the interactions among lactobacilli, CT EBs and HeLa cells plasma membrane. HeLa cells were grown at 70–80% of confluence on glass coverslips, treated with L. crispatus BC5, S. agalactiae, E. faecalis, or B. subtilis cells (5 × 10⁷ CFU) for 1 h at 37°C with 5% CO₂, and then washed three times with PBS. A stock solution of Nile Red (NR, 9-diethylamino-5H-benzo[alpha]phenoxazine-5-one; Sigma-Aldrich, Milan, Italy) was prepared in DMSO at the concentration of 1 mM and stored protected from light. NR staining was performed for 5 min at 37°C at a final dye concentration of 5 μM. Cells were then fixed with 3% paraformaldehyde for 10 min, and repeatedly washed with 0.1 M glycine/PBS and 1% Bovine Serum Albumin (BSA)/PBS. Specimens were embedded in Mowiol and analyzed by a Nikon Coolscope II equipped with an Eclipse 90i microscope, using sequential laser excitations at 568 nm to reduce spectral bleedthrough artifacts. A Nikon Plan Apo 60× oil objective was used. Fluorescence intensity was quantified on at least eight randomly chosen microscopic fields by using Image J; results were expressed as average fluorescence (A.U.)/field ± SD.

The plasma membrane fluidity of HeLa cells was estimated by means of the fluorescence anisotropy of the hydrophobic probe TMA-DPH (1-4-trimethylammoniumphenil-6-phenil-1,3,5-hexatriene; Thermo Fisher Scientific). HeLa cells (70–80% confluence) were incubated with L. crispatus BC5, S. agalactiae, E. faecalis, or B. subtilis cells (5 × 10⁷ CFU) for 1 h at 37°C with 5% CO₂, then washed three times with PBS and re-suspended at a final concentration of 3 × 10⁵ cells/mL. The absorbance of the cell suspension was kept lower than 0.15 at the excitation wavelength of TMA-DPH. A few microliters of TMA-DPH stock solution were added to the cell suspension in order to obtain a final probe concentration of 1 μM. Fluorescence anisotropy measurements were performed by using a PTI QuantaMaster fluorimeter (Photon Technology International, North Edison, NJ, United States) equipped with a temperature-controlled cell holder and Polaroid HNP’B polarizers. Temperature was kept at 25°C. Excitation and emission wavelengths were set at 360 and 430 nm, respectively.

HeLa cells were grown on glass coverslips at 70–80% of confluence and treated with L. crispatus BC5, S. agalactiae, E. faecalis, or B. subtilis cells (5 × 10⁷ CFU) for 1 h as described above, then washed three times with PBS and fixed in paraformaldehyde. Samples were stained with anti-human CD49e (i.e., anti-human α5 integrin subunit) primary antibody (1:500 in 1% BSA/PBS, BioLegend, San Diego, CA, United States) overnight at 4°C, followed by incubation with Alexa 568-conjugated secondary antibody (1:1000 in 1% BSA/PBS, Thermo Fisher Scientific) for 1 h at room temperature. As a specificity control for α5 integrin subunit signal, untreated Hela cells were also stained with IgG isotype (1:500 in 1% BSA/PBS, Sigma-Aldrich), followed by Alexa 568-conjugated secondary antibody. Specimens were embedded in Mowiol and analyzed in confocal microscopy as described above. Fluorescence intensity was quantified on at least eight randomly chosen microscopic fields by using Image J; results were expressed as average fluorescence (A.U.)/field ± SD.

In experiments exploring the role of the α5 integrin subunit in CT infection, HeLa cells grown on glass coverslips at 60% of confluence were first pre-exposed to control (IgG isotype) or anti-CD49e antibody (10 μg/mL) (BioLegend) for 1 h and then incubated with 5 × 10³ CT EBs for 48 h at 37°C and 5% CO₂. At the end of the incubation, samples were extensively washed in PBS, fixed in paraformaldehyde, permeabilized in ethanol and stained with the fluorescein-conjugated chlamydial membrane lipopolysaccharide antibody as described above. Specimens were embedded in Mowiol and analyzed in confocal microscopy as described above, using a Nikon Plan Apo 40× oil objective. Number of IFUs/field were counted by using Image J.

siRNA silencing was applied on HeLa cells after 48 h of adhesion, when cells were 60% confluent. The specific siRNA against ITGA5 gene expression and the scramble siRNA were transfected using lipofectamine according to the manufacturer’s specifications (Thermo Fisher Scientific). Control siRNA or ITGA5 siRNA (10 nM) were used to treat cells in phenol red-, serum-, and antibiotic-free RPMI. After 48 h of siRNA treatment, cells were infected with 5 × 10³ CT EBs for additional 48 h at 37°C and 5% CO_2. Samples were then washed three times with PBS, fixed in paraformaldehyde, and permeabilized in ethanol. Cells were stained for chlamydial membrane lipopolysaccharide and analyzed by confocal microscopy as described above.

Control, siRNA, and scramble HeLa cells were washed twice with PBS and lysed for 1 h in lysis buffer (HEPES, pH 7.4 40 mM, glycerophosphate 60 mM, p-nitrophenylphosphate 20 mM, Na₃PO₄ 0.5 mM NaCl 250 mM, Triton X-100 1%, PMSF 0.5 mM, and 10 mg/mL each of aprotinin, leupeptin, pepstatin and antipain, Sigma-Aldrich) at 0°C. Cell lysates were centrifuged at 12,000 × g for 20 min. Supernatants were collected and protein concentration determined by using the Bio-Rad protein assay method (Bio-Rad, Hercules, CA, United States). The proteins were resolved on a 7.5% polyacrylamide gel and immunoblotted with a rabbit anti-human α5 integrin subunit (1:1000 in PBS Tween, Cell Signaling Technology, Danvers, MA, United States), or a rabbit anti- β-actin (1:2000 in PBS Tween, Sigma-Aldrich) antibodies. Detection of immunoreactive bands was performed by using a HRP-conjugated secondary antibody (1:20,000 in PBS Tween, GE Healthcare, Milan, Italy) followed by WESTAR EtaC 2.0 (Cyanagen, Bologna, Italy). Densitometry analysis of immunoreactive bands was done by Fluor-S Max MultiImager (Bio-Rad). Relative quantification of α5 integrin subunit was performed by using β-actin signal as control.

Infectivity data were analyzed by using R computational software ¹, applying the non-parametric one-tailed Wilcoxon rank test. Fluorescence intensities were analyzed by Prism software (GraphPad Software, La Jolla, CA, United States), applying the Student’s t-test. Differences were deemed significant for P-values ≤ 0.01.

In the present work we investigated the capability of vaginal lactobacilli to counteract the infection process of C. trachomatis in HeLa cells, chosen as an in vitro epithelial model of cervical infection. Specifically, we tested the inhibitory activity of 15 Lactobacillus strains, previously isolated from vaginal swabs of healthy premenopausal women (Parolin et al., 2015) through an exclusion assay. The effects of lactobacilli cells against C. trachomatis infection are reported in Figure 1 and the raw data are listed in Supplementary Table S1. All the Lactobacillus strains significantly reduced the chlamydial infection. The most active strains were L. crispatus BC1, BC3, BC4, BC5, BC6, L. gasseri BC9, BC11, BC14, and L. vaginalis BC16.

We excluded that the observed anti-CT activity was related to an acidic environment, since the pH values of the culture media were not modified by the addition and incubation with lactobacilli cells: in particular, pH of DMEM without lactobacilli ranged between 8.0 and 8.1, whereas the pH values at 1 and 2 h post-incubation with lactobacilli ranged between 8.0 and 8.3.

In order to verify if the observed anti-Chlamydia activity was restricted to Lactobacillus strains, exclusion experiments were also performed with other Gram-positive microorganisms, namely S. agalactiae, E. faecalis, and B. subtilis. No reduction of Chlamydia infectivity was noticed for these control strains (Figure 1). In particular, as B. subtilis has similar shape and size of lactobacilli, we can suppose that the inhibition exerted by lactobacilli is not due to a physical barrier created by the large rod-shaped bacteria, but to a specific interaction with HeLa cells membrane.

We sought to investigate the effect of different doses of lactobacilli cells on the level of inhibition of CT infection. For this purpose, we selected three strains among the most active ones (L. crispatus BC4, L. crispatus BC5, and L. gasseri BC14) and we evaluated the inhibitory effects of two dilutions, corresponding to the doses of 5 × 10⁶ and 5 × 10⁵ cells (Figure 2). All Lactobacillus strains significantly reduced Chlamydia infectivity, confirming the strong antagonistic effect of Lactobacillus cells toward the infectious process of Chlamydia. Despite the antagonist activity was maintained even at lower doses, the level of inhibition was clearly dose dependent for all the lactobacilli.

We focused our investigation on L. crispatus BC5, chosen as model strain. We thus wondered if 1 h-treatment of Hela cells with L. crispatus BC5 cells prior to the infection with CT would be able to induce modifications at the HeLa plasma membrane level, in terms of lipid organization, fluidity, and modulation of protein exposure. HeLa cells were stained with the lipid dye NR and analyzed by confocal microscopy (Figure 3A). The interaction of L. crispatus BC5 with HeLa caused a decrease in NR emission fluorescence (Fluorescence intensity A.U. control: 37093 ± 8347, BC5: 16616 ± 6257, P = 0.001), indicating a reduced exposure of polar membrane lipids. Steady-state fluorescence anisotropy of TMA-DPH was used to evaluate the possible modifications of the physico-chemical characteristics of HeLa plasma membrane upon interaction with L. crispatus BC5. Under our experimental conditions, in the absence of L. crispatus BC5, an anisotropy value of 0.266 ± 0.002 was measured. The treatment with Lactobacillus cells induced a significant decrease in TMA-DPH anisotropy (0.257 ± 0.001; P = 0.006).

It is known that numerous pathogens, including C. trachomatis, exploit integrins expressed on target cell plasma membrane in order to mediate adhesion and cell penetration. For this reason, we investigated whether L. crispatus BC5 pre-treatment could alter integrin exposure on HeLa plasma membrane, by immunostaining of CD49e and confocal microscopy analysis (Figure 3B). The pre-incubation of HeLa cells with L. crispatus BC5 significantly reduced the exposure of α5 integrin subunit on the plasma membrane (Fluorescence intensity A.U. control: 41449 ± 4765, BC5: 27453 ± 3056, P = 0.001). Specificity of α5 integrin subunit staining was verified by using IgG isotype followed by fluorescent secondary antibody (Fluorescence intensity A.U. IgG: 3878 ± 337).

In order to verify if the modification of membrane properties are restricted to L. crispatus BC5, the same experiments were also performed with other Gram-positive microorganisms. HeLa cells were thus incubated with S. agalactiae, E. faecalis, and B. subtilis. The interaction of S. agalactiae and E. faecalis with HeLa cells caused a strong decrease in NR emission fluorescence (Figure 3A) (Fluorescence intensity A.U. S. agalactiae: 16484 ± 5048, E. faecalis: 12911 ± 3413, P = 0.001 vs. control, for both comparisons), but these effects on membrane lipid organization were not associated with changes in the membrane exposure of α5 integrin subunit (Figure 3B) (Fluorescence intensity A.U. S. agalactiae: 40549 ± 4986, P = 0.319; E. faecalis: 40059 ± 1333, P = 0.214). B. subtilis treatment did not affect HeLa plasma membrane lipid organization (Figure 3A) (NR Fluorescence intensity A.U. 42424 ± 7803, P = 0.338) nor α5 integrin subunit exposure (Figure 3B) (Fluorescence intensity A.U. 42655 ± 3073, P = 0.142). The measurement of anisotropy in Hela cells incubated with S. agalactiae, E. faecalis, and B. subtilis was not performed because these control microorganisms incorporated the TMA-DPH probe.

To examine the role of α5β1 in CT infectious process in HeLa cells, CT infectivity was evaluated after pre-incubation of HeLa cells with a function-blocking anti-α5 integrin subunit antibody or control IgG. Internalization of CT EBs by HeLa cells was reduced by approximately 60% in the presence of the anti-CD49e antibody (P < 0.01), while the pre-incubation with isotype IgG did not interfere with C. trachomatis infection (Figures 4A,B).

To further confirm the involvement of α5 integrin in CT infectious process in HeLa cells, we used a specific ITGA5 siRNA to knockdown endogenous α5 integrin subunit expression. Validated siRNAs were used for the analysis: a specific anti-ITGA5 and a correspondent negative siRNA were chosen (scramble). A silencing time course was assessed by quantification of α5 integrin subunit protein 48–120 h after silencing with siRNA 10 nM. Total protein lysates were resolved in SDS-PAGE and α5 integrin subunit protein was identified by a specific antibody in western blot. By normalization on the amount of β-actin, quantification of α5 integrin subunit in the specific siRNA sample was compared with the relative scramble sample (Figures 4C,D). Since α5 integrin subunit expression was already greatly reduced after 48 h of siRNA treatment, this time point was chosen as optimal. Following silencing, cells were infected with CT EBs. Transfection of cells with α5 integrin-specific siRNA reduced CT infection by 64%. In contrast, transfection of cells with non-specific control siRNA did not affect CT infection (Figures 4E,F).