Twelve archived and previously characterized clinical isolates of Lactobacillus and pathogenic strains (Table 1) were tested to determine their adhesion properties and the inhibitory effects of Lactobacillus on pathogen adhesion. The Lactobacilli were isolated from vaginal samples obtained from healthy volunteers (>18 years of age) at the health checkup clinic of our hospital. These women of childbearing age did not suffer from any urogenital tract infections. They had not used antibacterial drugs within 3 months of admission, and they had not had sexual intercourse 1 week before admission. The pathogens were isolated from patients (>18 years of age) with urogenital tract infections. All isolates were identified by standard methods used in clinical microbiology laboratories. Every strain to be tested was recovered and purified before the experiment to ensure bacterial viability and purity. The strains were incubated at 37°C in an atmosphere of 5% CO₂ for 48 h. Each bacterial suspension was adjusted to a concentration of 1.0 × 10⁸ CFU/mL.

Two kinds of VECs, including the VK2/E6E7-ATCC-CRL-2616 (VK2/E6E7) and primary VECs, were used in our study (Table 1). The VK2/E6E7 VECs were obtained from the American Type Culture Collection (ATCC; Rockville, MD, USA). This cell line from the normal vaginal mucosal tissue of a premenopausal woman, who underwent anterior-posterior vaginal repair surgery, was established in 1996 (31). Primary VECs were collected from a healthy volunteer during the 17th and 18th days of her menstrual cycle. The volunteer had not used antibiotics, spermicidal products, or oral contraceptives, and had no known vaginal pathology. The VECs were digested with 0.15% trypsin and 0.01% EDTA for 5 min, and the digestion was terminated using Dulbecco's Modified Eagle's Medium-nutrient mixture F-12 (DMEM-F12) medium that contained 10% fetal bovine serum. Next, the cells were centrifuged for 10 min (13,400 × g), and the supernatant was discarded. The VEC concentration was adjusted to 10⁵ cells/mL using keratinocyte serum-free medium (K-SFM) containing antibiotics (counted by the cell-counting plate method and then diluted to the appropriate concentration). VEC suspensions (2 mL) were pipetted into a 6-well-culture plate with a built-in coverslip and kept for 18 h until adhesion to the cells. The VECs that did not adhere were removed by washing thrice with DMEM-F12 medium (without antibiotics) containing 10% fetal calf serum. Both kinds of adherent VECs in the RPMI1640 medium in each well were used to evaluate the adhesion of the five Lactobacillus strains. Only the adherent VK2/E6E7 VECs in each well were used for the adhesion inhibition experiments.

VK2/E6E7 and primary VECs were used in this experiment. A Lactobacillus suspension (2 mL), at a concentration of 1.0 × 10⁸ CFU/mL, was added to the adherent VECs in each well. They were then cultivated at 37°C in an atmosphere of 5% CO₂ for 1 h. The coverslip was removed, fixed in methanol for 15 min, and subjected to Gram staining. Lactobacillus strains were identified based on their color and morphology and counted. Lactobacillus adhesion was assessed using the adhesion index, which was calculated as follows: Adhesion index = adhering bacteria/cell number. The adherent Lactobacillus number is the total quantity of Lactobacilli adherent to 50 intact VECs, which were selected randomly. We observed these intact VECs and counted the Lactobacillus number using a light microscope at 1000 × magnification under oil immersion (32).

The results of the Lactobacillus adhesion evaluation experiments indicated that the source of the VECs might not affect the selection of the most adhesive Lactobacillus strain. The VK2/E6E7 VECs, as a model cell line, ensured the reproducibility of the results of the current experiment. Therefore, it was chosen for evaluating the inhibitory effects of the five Lactobacillus strains on pathogen adhesion.

For the adhesion inhibition experiments, the concentration of each bacterium was adjusted to 1.0 × 10⁸ CFU/mL with RPMI1640 broth using a microplate reader. The adhesion inhibition experiments included exclusion, displacement, and competition inhibition experiments. Briefly, the built-in coverslip was removed, and pathogen adhesion was evaluated upon completion of the experiment. Pathogen adhesion to the VK2/E6E7 VECs was assessed using the adhesion index, as described previously (32). The adherent pathogen number is the quantity of pathogens adherent to 50 intact VECs (VK2/E6E7) selected at random. The pathogens that adhered to the VK2/E6E7 VECs because adhesion inhibition by Lactobacillus was absent were defined as the control group. The pathogens that adhered to the VK2/E6E7 VECs after Lactobacillus antagonization were defined as the experimental group. The inhibitory effects of Lactobacillus on pathogen adhesion were negatively correlated with pathogen adhesion to the VK2/E6E7 VECs in the experimental group. In other words, the stronger the pathogen adhesion, the weaker the inhibitory effects of Lactobacillus on pathogen adhesion in the experimental group. The statistically significant difference in the pathogen adhesion indexes of the control and experimental groups indicated that Lactobacillus showed inhibitory effects on pathogen adhesion.

For the control group, the pathogen suspension (1 mL) was added to a well with adherent VK2/E6E7 VECs. For the experimental group, the Lactobacillus suspension (1 mL) was added to another well with adherent VK2/E6E7 VECs. They were incubated at 37°C in an atmosphere of 5% CO₂ for 1 h. The non-adherent Lactobacillus and pathogens were removed by washing thrice with sterile phosphate-buffered saline (PBS). The pathogen suspension (1 mL) was added to the wells of the experimental and control groups and then incubated at 37°C in an atmosphere of 5% CO₂ for an additional hour. Adhesion was then evaluated.

The pathogen suspension (1 mL) was added to the wells containing adhered VK2/E6E7 VECs of the experimental and control groups. The cells were then incubated at 37°C in an atmosphere of 5% CO₂ for 1 h. The non-adherent pathogens were removed by washing thrice with sterile PBS. Lactobacillus (1 mL) was added to the well of the experimental group, whereas the pathogens (1 mL) were added to that of the control group. The cells were then cultured at 37°C in an atmosphere of 5% CO₂ for another hour, and adhesion was evaluated.

For the experimental group, the Lactobacillus suspension (1 mL) and the pathogens were mixed and incubated in a well containing adhered VK2/E6E7 VECs. For the control group, the pathogen suspension (1 mL) and RPMI1640 broth (1 mL) were added to the well with the adherent VK2/E6E7 VECs. They were then incubated at 37°C in an atmosphere of 5% CO₂ for 1 h, and adhesion was evaluated.

All experimental steps were repeated in triplicate to ensure the reproducibility of the results. The adhesion index values are expressed as the mean ± standard deviation (SD). An independent sample t-test was used to statistically analyze the differences between the two groups. A P < 0.05 was considered statistically significant. The statistical software used was SPSS 20.0.

The Ethics Committee of Peking University First Hospital approved this study (V2.0/201504.20), and written informed consent was obtained from all participants.

The adhesion of every Lactobacillus strain was stronger in the primary VECs than in the VK2/E6E7 VECs (P < 0.05) (Figure 1). These results indicated that the source of the VECs could affect the adhesion of the same Lactobacillus. The results could explain the differences in the effectiveness of Lactobacillus preparations in different populations (9, 25, 33) and the necessity of clinical trials to verify the anti-pathogenic effects of Lactobacillus preparations. The descending order of the five Lactobacillus strains based on the strongest adhesion to both types of VECs was 4#, 3#, 5#, 2#, and 1# (Figure 1). The results indicated that the source of the VECs might not affect the selection of the most adhesive Lactobacillus strain. VK2/E6E7 VECs, as a model cell line, ensured the reproducibility of the current results. Therefore, it was used in the experiments on the inhibitory effects of Lactobacillus on pathogen adhesion.

The adhesion of the three L. crispatus strains was stronger than that of the two L. gasseri strains (P < 0.05) (Figure 1), and this result was consistent with that of the study by Mousavi et al. (24). Although Mousavi et al. (24) chose Vero and HeLa cells to investigate Lactobacillus adherence, their results suggested that L. crispatus adhesion was stronger than L. gasseri adhesion, irrespective of the type of cells used. The L. crispatus and L. gasseri adhesion observed in both our and Mousavi's studies (24) was weaker than the Lactobacillus fermentum and Lactobacillus rhamnosus adhesion, which was investigated in the study by Ortiz et al. (34). Among the five Lactobacillus strains, L. crispatus 4# showed the strongest adhesion, and its adhesion indexes for primary and VK2/E6E7 VECs were 46.4 ± 3.41 and 25.4 ± 4.2, respectively (Figure 2). The differences in the adhesion of the Lactobacillus strains could be attributed to the different sizes, species, and growth rates of Lactobacilli and the different adherent cell types (24, 34, 35). Additionally, Lactobacillus adhesion was positively related to the secretion of a protein that contained MucBP-like domains (N506_1778) and a putative novel adhesin (N506_1709) with rib/alpha-like domain repeats and negatively related to the production of exopolysaccharides (36).

Adhesion not only helps Lactobacillus protect the mucosal epithelium but is also an important virulence factor of pathogens, beneficial for colonization, and biofilm formation (37). It is difficult for single-use antibiotics to penetrate biofilms (38); thus, biofilms protect pathogens and promote disease recurrence (39–41).

When the Lactobacillus-mediated antagonization of Gardnerella adhesion to the VK2/E6E7 VECs was tested, the descending orders of the strains with the strongest inhibition were 4#, 1#, 3#, 5#, and 2# in the exclusion experiments; 4#, 1#, 5#, 2#, and 3# in the displacement experiments; and 4#, 1#, 5#, 3#, and 2# in the competition experiments (Figure 3A). These results suggested that, among the five Lactobacillus strains, L. crispatus 4# showed the strongest inhibition of Gardnerella adhesion. Similarly, L. crispatus 4# also showed the strongest inhibition of Mobiluncus (Figure 3B), E. coli (Figure 3F), and E. faecalis (Figure 3G) adhesion. L. crispatus 3# showed the strongest inhibition of S. aureus adhesion (Figure 3E). The descending orders of the strains with the strongest inhibitory effects on S. aureus were 3#, 1#, 2#, 5#, and 4# in the exclusion experiments; 1#, 3#, 4#, 2#, and 5# in the displacement experiments; and 3#, 5#, 4#, 1#, and 2# in the competition experiments (Figure 3E). L. gasseri 1# showed the strongest inhibition of C. albicans (Figure 3C) and S. agalactiae (Figure 3D) adhesion. The descending orders of the strains with the strongest inhibitory effects on S. agalactiae adhesion were 1#, 5#, 4#, 3#, and 2# in the exclusion experiments; 1#, 4#, 3#, 5#, and 2# in the displacement experiments; and 1#, 3#, 5#, 4#, and 2# in the competition experiments (Figure 3D). However, further investigation of the antibacterial properties of Lactobacillus and clinical trials are required to confirm these results. More details are shown in Figures 3, 4.

All five Lactobacillus strains could effectively antagonize pathogen adhesion to the VK2/E6E7 VECs through exclusion, displacement, and competition inhibition (P < 0.05, Figure 3). However, the degree of the inhibitory effects of these strains on pathogen adhesion was not consistent. The current observations partially elucidated the mechanisms by which Lactobacillus preparations successfully antagonized pathogenic infections in clinical trials and showed that these effects were different (9, 25, 33).

Lactobacillus effectively prevented pathogen adhesion to the VECs (P < 0.05, Figure 3), indicating that Lactobacillus adhesion to VECs was beneficial for the prevention of pathogenic infections. For each Lactobacillus strain, the displacement effect was stronger than the exclusion and competition effects on pathogen adhesion (Figure 3). This showed that Lactobacillus could effectively remove pathogens attached to the vaginal epithelium; thus, the results provide a basis for the preparation of Lactobacillus formulations for effective vaginal infection treatment. Multiple studies have shown that exopolysaccharides produced by Lactobacillus can shield adhesin and reduce pathogen adhesion to cell surfaces (42–44). Ortiz et al. (34) selected L. fermentum and L. rhamnosus to antagonize S. aureus and S. agalactiae adhesion to VECs; they found that the competition effect was far stronger than the exclusion and displacement effects. Abedi et al. (45) showed that the exclusion, displacement, and competition effects were not significantly different. These findings suggested that Lactobacillus could effectively prevent pathogen adhesion to VECs, but the effects of this adhesion inhibition varied; possible reasons for this include differences in the anti-adhesive substances produced when different Lactobacillus species compete for the same adhesion receptor (46). Anti-adhesive substances can degrade pathogenic carbohydrate receptors, construct biofilms, induce biosurfactant production, produce receptor analogs, and cause steric blockade or receptor inhibition (45, 47, 48). Other possible causes include the time and initial amount of the pathogens in the in vitro culture (49), the bacterial size, and the presence of the arcA gene in the bacteria (48).