Comparative analysis of virulence factors & biotypes of Gardnerella vaginalis isolated from the genital tract of women with & without bacterial vaginosis

Bacterial vaginosis (BV) is the most common genital tract infection in women of reproductive age group. BV is characterized by a marked reduction in counts of lactobacilli which is frequently present as normal flora of the healthy vagina and an increase in numbers of anaerobic bacteria, including Gardnerella vaginalis, Atopobium vaginae, Mobiluncus spp., Bacteroides spp., Prevotella spp. and various other anaerobes123. Being polymicrobial in nature, BV aetiology remains unclear. However, BV involves the presence of a thick vaginal multispecies biofilm, where G. vaginalis is the predominant species. Similar to what happens in many other biofilm-related infections, standard antibiotics like metronidazole are unable to fully eradicate the vaginal biofilm, which can explain the high recurrence rates of BV. BV is associated with a variety of obstetric and gynaecological complications such as preterm birth, low birth weight, postpartum endometritis and pelvic inflammatory disease. The factors which are associated with pathophysiology of BV are not completely known12345.

It is a known fact that the presence of virulence factors is one of the important determinants of the pathogenic potential of the organism. The various virulence factors of G. vaginalis are pili, microcapsule, surface hydrophobicity, adherence, vaginolysin, phospholipase C, protease and siderophores, sialidases and prolidases67. The reason for an increase in the number of G. vaginalis which are usually present as normal flora of the female genital tract in cases of BV is not known. In the present study, the virulence factors and biotypes of G. vaginalis isolated from the genital tract of women with and without BV were compared.

Material & Methods

This cross-sectional study was conducted in the department of Microbiology, Father Muller Medical College (FMMC), Mangaluru, India from January 2014 to December 2016. All consecutive women of 15-45 yr of age with complaints of vaginal discharge were included in the study. Women who were menstruating at the time of the specimen collection or were on medication for any bacterial, fungal, parasitic or viral infections for up to one month before the specimen collection were excluded. The study was approved by the Institutional Ethics Committee of FMMCH, Mangaluru, and written informed consents were obtained from the patients.

A detailed clinical history of each woman was taken, and their vaginal swabs were collected into sterile tube containing 0.5 ml normal saline and immediately transported to the microbiology laboratory. The vaginal swabs were subjected to wet mount preparation, Gram staining, pH determination and whiff test. BV cases were identified using Amsel's and Nugent's criteria789. Sample was inoculated onto human blood bilayer Tween 80 agar (Columbia blood agar base with 0.0075% Tween 80) with G. vaginalis selective supplement containing gentamicin sulphate 2 mg, nalidixic acid 15 mg, amphotericin B 1 mg (HiMedia Laboratories, Pvt. Ltd., Mumbai, India) and chocolate agar. The culture plates were incubated at 37°C for 48 h in candle jar which provided 5-10 per cent CO₂. Colonies of G. vaginalis appeared small, smooth, round and β-haemolytic. G. vaginalis was identified as Gram-variable coccobacilli by Gram stain, catalase and oxidase negative, fermented glucose, maltose and starch and did not ferment mannitol, lactose, xylose and sucrose variable, α glucosidase positive and β glucosidase negative1011. The isolates were also tested for various virulence factors such as adherence to vaginal epithelial cells, biofilm formation, surface hydrophobicity, haemagglutination, phospholipase and protease production. All the tests were performed in triplicate. G. vaginalis ATCC 14018 was included as control with each test.

Evaluation of virulence factors: Vaginal swabs were collected from healthy asymptomatic women in the reproductive age group, and the vaginal discharge was eluted into 2 ml of 0.85 per cent sterile saline. The adherence of the isolates to these vaginal epithelial cells was determined12. An average of >10 adherent bacilli/cell was considered as good degree of adherence712. Haemagglutination assay was performed according to Scott et al13. The suspension giving a carpet of erythrocytes over the bottom of the well was considered as positive. The same test was done using 2 per cent human, sheep and chick red blood cells (RBCs)13. Biofilm formation in microtiter plates was determined as done in previous studies1415. The biofilm formation was graded as, optical density (OD) <0.1 as weak or non-biofilm producers, OD 0.1-0.2 as moderate and an OD >0.2 as good biofilm producer. Pseudomonas aeruginosa ATCC 27853 was used as a positive control1415. For protease production G. vaginalis isolates were inoculated into skim milk agar with 5 per cent horse serum and incubated in the candle jar (5-10% CO₂) at 37°C for 48 h. Clear zone around the growth indicated protease enzyme activity16. Brain heart infusion (BHI) agar supplemented with 1 per cent gelatin egg yolk was used for phospholipase production. G. vaginalis isolates were inoculated and incubated in the candle jar (5-10% CO₂) at 37°C for 48 h. Pearly layer was seen around growth in lipase-positive organism. P. aeruginosa (ATCC 27853) and Staphylococcus aureus (25923) were used as negative reference strains17. Surface hydrophobicity of the isolates was determined as given in the previous studies618. Surface hydrophobicity index ≤20 was graded as low and >20 as high. Biotyping was done using three tests - hippurate hydrolysis, phospholipase and O-Nitrophenyl β-D-galactopyranoside (ONPG) as described previously719.

Statistical analysis: Statistical analysis was done using SPSS version 16 (IBM, Chicago, USA). The various biotypes and virulence factors of G. vaginalis from women with or without BV were compared using Chi-square test.

Results & Discussion

A total of 176 isolates of G. vaginalis were obtained from 811 women with abnormal vaginal discharge. According to Nugent's criteria, vaginal samples were classified into three groups. Group A consisted of 75 women with Nugent score 4-6 (not consistent with BV by Amsel's criteria). These were considered as non-BV cases. Group B included 89 women (Nugent score ≥7 and BV case by Amsel). These were considered as BV cases. Group C included 12 women with Nugent score 0-3 (non-BV case by Amsel) having other vaginal infections such as vulvovaginitis (n=6), cervicitis (n=4) and HIV reactive (n=2). These were also considered as non-BV cases. There were 23 women with Candida in group A, 49 in group B, and eight in group C. As group C has only 12 women, all comparisons were made between groups A (non-BV) and B (BV). Table I shows the virulence factors expressed by the three groups.

A study proposed a conceptual model to explain the pathogenesis of BV20. In our study most isolates of G. vaginalis obtained from BV cases exhibited various virulence factors such as adherence to vaginal epithelial cells, biofilm formation, phospholipase C production, protease production, surface hydrophobicity and haemagglutination rather than the non-BV cases (Table I).

According to a previous study, G. vaginalis was the first isolate to adhere to the vaginal epithelial cells causing the scaffolding to which other organism adhere21. However, another study reported that adherence of G. vaginalis was inhibited by the lactobacilli and adherent lactobacilli were displaced by G. vaginalis22. Other groups studied the influence of biofilm formation by G. vaginalis and other anaerobes from the time of their initial adhesion until biofilm formation in BV cases. It was reported that the synergistic interaction occurred between G. vaginalis and anaerobes from commensal status till biofilm formation2324. G. vaginalis was shown to have the highest virulence potential due to higher initial adhesion, cytotoxicity and the ability to form a biofilm25.

Past studies have reported difference in genotype and in vitro expression of virulence factors between BV and non-BV isolates of G. vaginalis2627. The BV-associated isolate encoded a different variant of a biofilm-associated protein gene and demonstrated greater adherence, aggregation and biofilm formation than that isolated from healthy vagina26. In our study also BV-associated isolates showed better adherence, biofilm formation, haemagglutination, phospholipase and protease production in comparison to non-BV isolates. In our study, maximum number of isolates belonged to biotypes 6, 2 and 3. Biotype 3 was more associated with non-BV rather than BV (Table II); biotypes 6, 2 and 1 were more associated with cases of BV. Maximum virulence factors were expressed by biotypes 6, 2 and 1 (Table III).

Our earlier study showed no relationship between biotypes and virulence factors or biotypes from BV and non-BV cases7. Another study showed that phospholipase-producing biotypes 1, 2, 3 and 4 were frequently associated with cases of BV rather than non-BV and also the patient acquired a different biotype after treatment27. In a longitudinal study on biotypes of G. vaginalis biotypes 2, 3 and 7 were more frequently isolated from cases of BV and biotype 7 was more isolated from non-BV cases27. In our study biotypes 6, 2 and 1 showed association with various virulence factors.

In conclusion, majority of G. vaginalis isolates from cases of BV exhibited more number of virulence factors than those from healthy women. Biotype 3 was more prevalent in non-BV cases while biotypes 6, 2 and 1 were associated with cases of BV and expressed maximum virulence factors.