Lactobacillus reuteri can reduce Gardnerella induced bacterial vaginosis in mice and modulate immune markers

Background Healthy vaginal microbiome is dominated by Lactobacilli, which constitute a strong line of defense against vaginal diseases like Bacterial vaginosis (BV). Bacterial Vaginosis is a polymicrobial disease characterized by gradual replacement of predominant population of Lactobacillus with anaerobic uropathogens such as Gardnerella vaginalis (GV), Prevotella and Mobiluncus spp. Due to antibiotic resistance in these pathogens, Lactobacillus spp. have been given attention in the prophylaxis and prevention of infections such as, urinary tract infections, genital infections and BV. Current study describes role of L. reuteri in reducing BV in a GV induced BV murine model. In addition, immunomodulatory effects of L. reuteri were assessed by analyzing gene expression of in�ammatory and anti-inammatory markers by real time PCR in vaginal tissue. Methods Study was divided into


Background
Bacterial vaginosis (BV) is one of the most prevailing bacterial infections in human that results from overgrowth of microorganisms naturally present in the vagina.The clinical features include thick grey discharge, an amine or " shy" vaginal odor, an elevated pH and/or the presence of super cial squamous cells studded with bacteria (clue cells) in wet mounts (1).Women with BV are at increased risk of severe health problems such as pelvic in ammatory disease, increased susceptibility to sexually transmitted infections and preterm birth (2).The prevalence of BV is about 18-35.3% in women of 16-25 years of age in Pakistan (3,4).Although the etiology of bacterial vaginosis remained poorly understood, previous studies have declared that this disorder is a consortium of bacteria that live in symbiotic association and cause dramatic changes in vaginal microbiota by decline in level of Lactobacillus spp.and an overgrowth of several aerobic and anaerobic organisms (5).
Recent genomic data have advocated that microbial community in BV vary from one individual to another (6).Moreover, potential role of each species to the cellular, biochemical and clinical characteristics of BV remain di cult to track down.Gardnerella vaginalis (GV) was the rst bacterium incriminated in the pathogenesis of BV and claimed to be related with the disease (7).The isolation of GV from almost 100% cases of BV has been reported by Srinivasan et al. (8).Studies have con rmed the pathogenic potential of G. vaginalis by cell adhesion, penetration, bio lm formation and cytolytic toxin production (9,10,11).G. vaginalis has a symbiotic relationship with anaerobes present in vagina (12).It produces amino acids which are utilized by P. bivia that in return secretes ammonia which is utilized by GV.The mechanisms employed in symbiotic relationship support not only the growth of strict anaerobes normally present in low number in vagina but also cause a shift to more alkaline pH which is unsuitable for bene cial microbes e.g., Lactobacillus spp.(13).
Lactobacillus spp.are probiotics that antagonize interruptions of indigenous micro ora (8, 14) and/or produce secondary metabolites to ameliorate harmful effects of bacterial pathogens (15).Probiotic bacteria are also well known to modulate innate and adaptive responses of host (16).Bacterial vaginosis (BV) is treated by clindamycin or metronidazole but recurrence rate is escalating with the passage of time (8,17).Under this scenario, it would be imperative to check Lactobacillus spp.isolated from healthy humans as therapeutic probiotics to restore and maintain a healthy genital tract (18).Previous literature reports antagonistic effects of L. johnsonii, L. rhamnosus, L. acidophilus and L. fermentum against GV in mice, but to the best of our knowledge, no in vivo data exists on inhibitory activity of L. reuteri in GV induced BV model.In addition, most preventive and therapeutic mechanisms are typically species and strain speci c (19).Moreover, most the studies reported their bene cial e ciencies through in vitro assays.Animal models are obligatory to con rm bene cial effects of potential strains in in vivo settings (20,21,22,23).Furthermore, animal studies can help to study human infections more in detail and help to nd treatments for such infections.L. reuteri strain used in this study was isolated from healthy women in our previous study.The aim of the current study was to check colonization of Lactobacillus reuteri in mice and nd its role in reduction of Bacterial vaginosis (BV) induced by using its major etiological agent GV in murine model.Lactobacilli are also reported to provide indirect protection through modulation of immune system.The expression of in ammatory and anti-in ammatory cytokines was also compared among groups to probe the role of L. reuteri in immunomodulation.

Study Design
Experiments was conducted using completely randomized design with control groups using female mice (Mus musculus) as test units.The study was divided in two phases.In phase one, colonization of GV and L. reuteri were checked in immunocompromised mice.In phase 2 effectiveness of prophylactic application of L. reuteri was studied by challenging animals with GV.

Study Samples
Clinical signs of the disease, reisolation of L. reuteri and GV from vaginal washes, bacterial load and expression of Immune markers in vaginal tissue and histological changes in vaginal sections were taken into consideration.
Characterization of microbes used in study L. reuteri and GV were isolated from healthy women and patients diagnosed with Bacterial vaginosis (BV) respectively in a previous study.In in vitro setting L. reuteri was con rmed for its probiotic and antagonistic ability through in vitro procedures.it displays good adherence ability, self-aggregation and co-aggregation, acid, bile salt, NaCl and lysozyme tolerance as well as antagonistic activity against G. vaginalis, E. faecalis, E. coli, S. aureus and Streptococcus spp.(Data not shown).Antibiotic pro le shows that strain was sensitive to erythromycin (E15) streptomycin (S10), and gentamicin (CN10) and resistant to ampicillin/sulbactum (SAM20), oxacillin (OX5) and Methicillin (MET5).
GV is Gram variable, pleomorphic rod shaped and show small, transparent colonies on GV selective media (Columbia agar containing nalidixic acid (5mg/L), gentamicin sulfate (10mg/L) and human blood (5%).It shows β-hemolysis on human blood agar but not on sheep's blood.GV was negative for oxidase, catalase, nitrate reduction and Voges Proskauer and ferment starch, sucrose, fructose and maltose but not mannitol.In addition, it was resistant to ampicillin/sulbectum (SAM20).
Phase 1: (Establishment of GV murine model) Animal housing and grouping Two-month-old female BALB/c mice (n=15) weighing 20-30g were purchased from stocks of Government College University Lahore.Mice were housed in standard sized cages (12″×18″) at 28-33°C temperature and 40% humidity and they were fed conventional balanced diet and water ad libitum.After initial weighing, mice were equally distributed into three groups (A-C).Group A was maintained as negative control, group B and C were exposed to GV and L. reuteri respectively.

Initial screening of mice
Vaginal endogenous ora of each mice was collected by ushing vagina with 50 µl sterile Phosphate buffer saline (PBS) using micropipettes (Dragon LAB, China).Each vaginal sample obtained was cultured on MRS and GV selective media.Each isolated colony grown on these agar plates was studied for colony morphology, Gram staining.oxidase catalase, nitrate reduction, Voges Proskauer and antibiotic pro le to distinguish vaginal ora of mice from probiotic L. reuteri and etiological agent GV.

Preparation of mice for Bacterial colonization
All mice were prepared for bacterial colonization (24).Brie y, mice were injected subcutaneously βestradiol 17-valerate weekly throughout the experiment.Dexamethasone sodium phosphate injection was also given daily 2 and 4 days prior to exposure to L. reuteri and GV respectively and continued up to 6 days post exposure to GV.A dose of 5 × 10 5 CFU ml -1 of G. vaginalis and 1 × 10 6 CFU ml -1 of L. reuteri was intravaginally inoculated in mice of groups B and C respectively.Details of experimental treatments are given in Table 1.

Monitoring of infection (a) Clinical signs
Animals were monitored for clinical signs of morbidity such as animal irritation, ri ed fur, vaginal itching, discharge and redness which may be associated with GV induced vaginosis.
(b) Re-isolation of reuteri and GV Animals were screened for the presence of L. reuteri and GV for 9 days post exposure to GV.Five vaginal washes were collected on alternative days and spread on MRS and GV selective agar plates for reisolation of inoculated L. reuteri and GV respectively.These plates were incubated for 24-48h under anaerobic condition.After incubation, plates were observed for presence of G. vaginalis and L. reuteri.Colonies on Columbia agar plates were recon rmed as GV using biochemical procedures (Negative oxidase, catalase, nitrate reduction and Voges Proskauer tests).L. reuteri was detected on basis of colony morphology, Gram staining and matching antibiotic susceptibility pro le.On 10 th dpi, mice were euthanized following ULAM guide lines of euthanization by intraperitoneal injection of ketamine (200 mg/kg of body weight).Vaginal tissues were removed to quantify bacterial load of L. reuteri and GV (CFU/g).
Phase:2 Protective effects of L. reuteri against GV Animals Two-month-old female BALB/c mice, weighing 20-30g were procured from stocks of Government College University Lahore.Mice were housed in cages as mentioned in phase 1.A total of 32 mice were randomly divided into four groups I-IV (n = 8).Experimental treatments to each group are summarized in Table 2.

Exposure to L. reuteri
After con rming the absence of L. reuteri and GV in experimental mice, Estrogen and dexamethasone injections were given to all mice as described in phase 1.Exposure to probiotic was given following (25) with slight modi cation.Freshly cultured L. reuteri was re-suspended in agarized peptone (0.1%, 0.5% w/v peptone and agar respectively) at a concentration of 1×10 6 CFU ml -1 .Two-days post β-estradiol administration, group III and IV were intravaginally inoculated with 50 µl of L. reuteri suspension with an average inter-inoculation period of twelve hours.

Exposure to G. vaginalis
On 5 th day of experiment, the mice were exposed to GV in 50 µl inoculum (5 × 10 5 CFU ml -1 ) as mentioned in Table 2.These doses of L. reuteri and GV were selected following (25) with slight modi cations and initial application in phase 1.

Monitoring of G. vaginalis infection (a) Recording of clinical signs/symptoms
Mice were observed for clinical signs of BV including irritation, ri ed fur, vaginal itching, discharge and redness.All symptoms were given equal weightage.Animals were scored from 0-1 depending upon mean number of symptoms present e.g.animal showing two of the ve symptoms was scored as 0.4.After scoring, each animal mean score of each group was used for comparison of general health condition of each group.
(b) Re-isolation of reuteri and G. vaginalis Vaginal washes were collected and L. reuteri and GV were con rmed on the basis of biochemical characteristics and antibiotic pro le (as mentioned in phase 1).

Sampling of tissue
On 17 th day of experiment, mice were euthanized by using over dose of ketamine (200 mg/kg), and vaginal tissues were harvested.A part of vaginal tissue was preserved by placing in liquid nitrogen later on these were stored at -80°C till further use.Similarly, 100 -150 mg of tissue was used for determination of bacterial load and rest of the tissue was xed in 10% formalin for histological studies.

Bacterial load
Vaginal tissue (bisected longitudinally) was weighed and homogenized in sterile saline and 100 µl of suspension was spread on MRS and GV selective agar plates at 37°C for 24-48h under anaerobic condition.After incubation period, colonies were enumerated and expressed as log 10 CFU/ g of vaginal tissue (26).

Histopathological examination
In brief, para n-embedded vaginal sections (5µm) were stained with hematoxylin and eosin (H &E) and observed under light microscope to assess histopathological changes.Images were captured and degree of thickness and exfoliation of epithelium were observed.The pathology of vaginal tissue was recorded following (7).

RNA extraction and cDNA preparation
Total RNA was extracted from stored tissues at -80°C using TRIZOL RNA extraction protocol (27).RNA extracted was quanti ed using Nanodrop spectrophotometer followed by complementary DNA (cDNA) synthesis using 500 ng of RNA with Superscript TM IV First-Strand cDNA Synthesis Kit according to manufacturer's protocol.Quantity of cDNA was measured using aforementioned Nanodrop spectrophotometer.

Real-Time PCR
The PCR reaction mixture having a nal volume of 25 µl containing 12.5 µl SYBR green Dye, 1 µl cDNA, 1 µl forward and reverse primer each of nal concentration of 10 Pmole and 9.5 DNase -and RNase-free water.Details of the primers along with annealing temperatures are given in Table 3. Real-Time PCR was carried out on a Bio-Rad CFX Real-Time PCR system for 40 cycles using the following conditions: denaturation at 95 °C for 3 min, annealing at 54-59 °C for 10s and elongation at 72°C for 30s.Average quanti cation cycle (Cq) values from duplicate measurements were used to determine mRNA expressions.Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) was used as Normalization control (28).It was normalized with the average Cq value of the control group.

Normalized expression ratio (NER) calculation
Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) was used as normalization control (Eissa et al., 2016).The degree of expression of all markers was evaluated by difference in Ct value (ΔCt value).The Ct values were normalized by both the "Ct value of GAPDH" and "Ct value of control group" to get ΔΔCt values.Mean of ΔΔCt values were compared between groups II, III and IV.

Statistical analysis
Data of bacterial load and in ammatory markers was parametric while Clinical symptoms and degree of epithelial exfoliation data was non parametric.Parametric data was analyzed by Independent sample Ttest and ANOVA following post-Hoc Tukey's test.Kruskel wallis H-test was applied to non-parametric data.The p value < 0.05 was considered signi cant.Data was analyzed using software IBM SPSS version 21 Chicago IL, USA.

Results
Phase 1: (Establishment of GV murine model)

Screening for the absence of L. reuteri and GV
There was no growth on GV selective agar plates.On MRS agar plates, only acid producing cocci were observed.Furthermore, normal microbial fauna of mice was susceptible to antibiotic ampicillin/sulbactum (SAM 20).

Clinical symptoms
Mice exposed with GV (group B) showed clinical signs of morbidity including irritation, ri ed fur, turbidity of vaginal discharge, itching and redness beginning at the 3 rd dpi.L. reuteri exposed mice (group C) seemed active except irritation at the beginning that might be due to inoculation procedures.As expected, Mice from negative control (group A) remained healthy throughout the experiment.

Re-isolation of L. reuteri and G. vaginalis
After intravaginal inoculation of L. reuteri and G. vaginalis, vaginal washes taken from 8 th to 16 th days of experiment were detected for presence of these strains.Distinguishing features of G. vaginalis including Gram variable, negative for oxidase, catalase, nitrate reduction, Voges Proskauer tests and resistance to ampicillin/sulbactum (SAM20) were observed in bacterial colonies grown on GV selective agar plates.While growth of L. reuteri on MRS plates was con rmed by characteristics like Gram positive, rod-shaped and antibiotic sensitivity to erythromycin (E15) streptomycin (S10), and gentamicin (CN10) while resistance to ampicillin/sulbactum (SAM 20), oxacillin (OX5) and Methicillin (MET5).Based on these assays, their re-isolation was con rmed.L. reuteri and GV were found to account for 4.8 log CFU/g and 4.39 log CFU/g respectively (Table 4).
Phase:2 Protective effects of L. reuteri against GV Clinical signs/ symptoms of infection Mice in groups I and III remained healthy and their average clinical score was 0 and 0.04 respectively throughout sampling period while mice challenged with GV only were less irritated at the beginning of experiment and scored 0.08 at 2 nd dpi but other clinical symptoms appeared in succeeding days and at end of the experiment and average score recorded was 0.84.Mice in prophylactic group showed clinical signs during initial period, scoring 0.24 that was reduced to 0.16 at the end of study (Fig 1).

Re-isolation frequency of L. reuteri and G. vaginalis
Vaginal washes obtained at sampling intervals were detected for presence of L. reuteri and G. vaginalis to con rm their colonization.All mice in group III were positive for L. reuteri while it was observed in 7 out of 8 animals in group IV (prophylactic group).G. vaginalis colonized successfully in all GV exposed group II and was continuously observed in vaginal washes throughout experiment but there was signi cant reduction (2 out of 8) in prophylactic group (Fig 2).

Bacterial load
Bacterial load of L. reuteri was 4.86 ±.16 and 4.91±.13CFU/g in group III and IV respectively.GV count was 4.71±.27 in group II and reduced signi cantly 1.5 log 10 CFU/g in prophylactic group (Table 5).

Histopathological examination
Vaginal sections were observed for epithelial thickness and exfoliation in response to GV infection as compared to negative control group.Epithelial exfoliation was higher in GV infection group which correlated positively with increased thickness of transitional epithelium (Fig 3).Both these factors were observed normal in I, III and IV groups.

In ammatory and anti-in ammatory cytokines expression
In order to investigate the role of innate immune system in anti-BV mechanism of probiotic therapy, the transcript levels of Pro and anti-in ammatory cytokines were determined in all groups.The relative mRNA expression of these cytokines is presented in Fig 4 .All results are expressed as normalized expression ratio (NER).In immunocompromised mice, the exposure to L. reuteri resulted in 1.18-fold and 1.19-fold decrese expression of IL-6 and IL-1β in comparison to GV exposed (group II).On the other hand, GV exposure resulted in 4.43 and 4.37-fold increase IL-6 and IL-1β expression respectively.Pre-exposure with L. reuteri decreased IL-6 and IL-1β gene expressions to the levels of 2.14 and 2.08-fold in GV infected mice (group IV).Interferon-γ which is anti-in ammatory cytokine expressed 0.98-fold and 0.25-fold in L. reuteri and GV exposed groups respectively but L. reuteri increased its expression to 0.30-fold in prophylactic group than GV exposed group (group II).ROR-γt levels were not in uenced by L. reuteri in prophylactic group.

Discussion
Bacterial vaginosis (BV) is the most serious disorder in women of childbearing age, contributing to over 60 per cent of all vulvovaginal infections (32).It can cause serious health hazards, including acquisition and transmission of multiple sexually transmitted agents (33), endometritis (34), spontaneous abortion (35) and pre-term birth (36).BV is a growing international concern due to high prevalence rate worldwide.Antimicrobial therapy is generally used to eradicate these infections but their effectiveness is currently diminishing due to adverse sequel and recurrence of infections (37).Role of probiotic fauna in reducing BV are being explored as alternative approach for restoration and maintenance of healthy vaginal ora and the prevention of recurrent diseases.Several in vitro studies provide evidence for antimicrobial effects of Lactobacillus spp.such as L. rhamnosus, L. acidophilus, L. reuteri, L. crispatus, L. gasseri, L. salivarius against uropathogens G. vaginalis, E. coli, S. agalactiae and C. albicans (25) but their con rmation in in vivo setting is mandatory.The mechanism used by lactobacilli in promoting local immunity to pathogens are not completely understood (18) in particular the role of L. reuteri on cytokine levels in female genital tract is least reported.Current study focused on prophylactic application of L. reuteri as a mean of controlling GV induced BV infection in mice.Initially, endogenous ora was explored for the presence of L. reuteri and GV by analyzing vaginal spreads obtained from each of the mice.we observed the presence of Gram-positive cocci on MRS plates and there was no growth on GV selective media.GV were absent in vaginal washes in consistent with our results no previous study has reported GV isolation from laboratory mice.Enterobacteriaceae, streptococci, Staphylococci are predominant vaginal ora of mice.After nding negative results for both strains, study was preceded further.Pre-oestral stage is prerequisite to induce BV in mice.It is well known that estrogen production positively in uences colonization of lactobacilli in healthy human vagina (1).
In the present study when Dexamethasone and βestradiol-immunosuppressed mice were intravaginally inoculated with L. reuteri and GV.Successful colonization was observed at the dose of 1 × 10 6 CFU ml − 1 and 5 × 10 5 CFU ml − 1 respectively.Both strains colonized successfully evident from their re-isolation up to 10th dpi and log 10 CFU/g count in tissue homogenate.Animal model was declared established based on reisolation and bacterial load of both strains in phase 1.In order to determine inhibitory effect of L. reuteri against GV, mice were prepared and colonized at the same dose used in phase 1. Clinical signs might be used as a tool to assess progress of infection in mice.Turbidity of discharge, redness and itching were sharply observed in mice exposed with GV.The mice were more sensitive to holding or touching due to vaginal irrigation.These signs were improved (normal range) in prophylactic group.Animal irritation and fur texture are reported rst time in this study in addition to other clinical signs mentioned in a previous study (38).The culture of GV in vaginal lavage showed that L. reuteri has restricted colonization of GV as compared to GV positive group.In general, infected mice are not to control bacterial proliferation, showing increased bacterial load and/or delayed clearance compared to healthy mice.
Bacterial load is a simple and easy method to determine level of infection.Bacterial load was quanti ed in homogenate of vaginal tissue removed by euthanizing mice.GV count was 4.71 log 10 CFU/g in GV exposed group but reduced signi cantly 1.5 log units in prophylactic group (group IV).This nding clearly indicated prophylactic in uence of L. reuteri on GV count.It might be due to bene cial effects of different mechanisms such as antimicrobial compounds (hydrogen peroxide, organic acid, biosurfactants and bacteriocins) (39,40) host immune response (25) bio lm formation and colonization ability (41) suggested for probiotics.L. reuteri compete out to a signi cant extent that 80% mice were found free of GV strain in prophylactic effect (group IV).Our results are consistent with ndings (42) who reported anticandida activity of L. reuteri in in vivo model.The prophylactic effect of L. reuteri was further con rmed by histological study of hematoxylin and eosin (H &E) stained vaginal sections and measuring expression levels of pro-in ammatory IL-6, IL-1β and ROR γt and anti-in ammatory IFN-γ markers.GV resulted in robust epithelial exfoliation and increased thickness of transitional epithelium in GV exposed animal.These two features were comparable to normal range in prophylactic and L. reuteri exposed groups.Epithelial shedding exfoliation and epithelial thickness have long been reported with BV.Exfoliation removes adherent pathogens and appears one of the defense mechanisms against pathogens but excessive exfoliation increases the establishment of BV-associated bacteria and risk of secondary infection.Our results are consistent with well-known study of (7).It is reasonable to suspect that local immunity may contribute local protection against potential pathogens.The epithelial cells layer possesses "Toll-like receptors" (TLR) in their surface recognize molecular patterns associated with the pathogens (PAMPs) and trigger sequence of events to release pro-in ammatory cytokines (IL-6, TNF-α and IL-1β) and consequently, the activation of the acquired immune system, which is, the activation of lymphocytes T and B and release of anti-in ammatory cytokines (IFN-γ).IL-6 is a multifunctional cytokine produced by immune cells following infection.IL-6 production induces acute phase proteins such as serum amyloid A, Creative protein (CRP) and α1-antitrypsin as part of an in ammatory response (43).
IL-1 β is highly in ammatory, activates a range of cells including macrophages and T lymphocytes that may thus lead to production of other cytokines.RORγt belongs to nuclear hormone receptor superfamily and is required for differentiation of the Th17 lineage and also su cient to direct the expression of the hallmark cytokines of this lineage.Interferon-γ produced by Th1-cells, NK cells, NKT cells is an important cytokine in recognizing and eliminating pathogens.It increases the e ciency of immune system by enhancing its competence to deliver anti-microbial effector functions.In this study, L. reuteri signi cantly reduced the mRNA expressions of pro-in ammatory markers IL-6, IL-1β in prophylactic group in comparison ROR γt expression was not in uenced signi cantly.This might be due to fact that effects on cytokine production are strain dependent.IFN-γ which is anti-in ammatory expressed higher in probiotic group (group III) as compared to GV exposed mice (group II) but L. reuteri elevated IFN-γ expression in prophylactic group (group IV).These ndings suggest that L. reuteri regulate immune system of host to attenuate GV-induced vaginosis in mice.
Studies suggest that the expression levels of pro-in ammatory cytokines, such as IL-6, TNF-α, RORγt and IL-1β in patients with BV infection has been reported higher than normal values and such in ammatory cytokines can attract in ammatory cells chemotactically and further worsen the local mucosal injury (31,44,45).

Conclusions
It is concluded that L. reuteri can inhibit clinical symptoms of BV by reducing the growth of pathogen GV and can induce immunity response of host which is comparable to the role of antibiotics.These ndings suggest clinical applications of metabolites in of L. reuteri as pharmabiotics in treatment of Bacterial vaginosis in human beings.

Limitations The Study
Animal models have had limitations in modeling the complex clinical entity of BV.The genetic modi cation and uorescent labeling of strains would made these experiments more informative.GV dose was given at 5 × 10 5 CFU ml -1   d Vaginal washes were collected on each sampling day to confirm presence of L. reuteri and GV e Animals were sacrificed on 10 th dpi and vaginal tissue was removed to determine bacterial load

Figure 2 Comparison
Figure 2

Figure 3 Epithelial
Figure 3 Neveu WA, Allard JB, Dienz O, Wargo MJ, Ciliberto G, Whittaker LA, Rincon M. IL-6 is required for airway mucus production induced by inhaled fungal allergens.The Journal of Immunology.2009 # Group A: Negative control, B: GV (positive control), C: L. reuteri a Dexamethasone sodium phosphate b L. reuteri exposure at the dose of 1 × 10 6 CFU ml -1 c