Introduction

Probiotics are, according to the FAO and WHO’s definition, “live microorganisms which when administered in adequate amounts confer a health benefit on the host” [1]. Several health benefits of probiotics are known: their administration is beneficial in preventing and curing different types of diarrhea [2]; they are used in the treatment of inflammatory intestinal diseases [3] and it has been demonstrated that they help preventing from allergies [4]. Usually, the gut is the target organ for probiotic formulations, but recently other organs and tissues have been related to probiotics, such as skin, hair [5], the oral cavity [6] and the vagina [7], by showing their positive effects for human health. In the specific case of vagina, the microbial species play an important role in the maintenance of health and prevention from infections throughout several mechanisms: occupation of specific adhesion sites of the uro-vaginal epithelium, maintenance of a low pH and production of anti-microbial metabolites such as acids, bacteriocins, hydrogen peroxides and anti-adhesive polysaccharides [8]. In physiological conditions, the vagina hosts mainly Lactobacillus spp. which concurs to a healthy microbiota, hence preventing from colonization of pathogenic bacteria and fungi [9, 10]. On the contrary, the depletion of vaginal lactobacilli facilitates the overgrowth of diverse species such as Gardnerella vaginalis, Atopobium vaginae, Candida albicans, Escherichia coli, responsible for vaginal dysbiosis and uro-genital infections [9–12]. Different probiotics were reported [13–15] to restore the normal vaginal homeostasis by colonization of lactobacilli, when administered topically.

Recent works [16, 17] also suggest that oral administration of lactobacilli and bifidobacteria would colonize both the intestinal and vaginal mucosal surfaces. This opens a new prospect for probiotic therapy: the challenge is to develop probiotics formulation to treat the uro-genital infections and to reduce the recurrences of disease in time, thanks to the capacity to control the pathogenicity of other microbes and to restore the normal ecological balance in vagina. However, these studies did not describe the efficacy of oral administration in term of presence and persistence of probiotic at the vaginal mucosa.

In this context, the aim of the present work is to evaluate in the vagina of 60 pre-menopausal healthy women the detection of orally administered multispecies probiotic formulations showing anti-microbial properties in test in vitro. A formulation F_1 containing L. acidophilus PBS066 and L. reuteri PBS072 and the other F_2 composed by L. plantarum PBS067, L. rhamnosus PBS070 and B. animalis subsp. lactis PBS075 were compared with the placebo (F_3). Data on these two mixtures supported the anti-microbial activity exerted by the above-mentioned single strains by assays in vitro using cell-free supernatants against microorganisms as E. coli and C. albicans.

The pilot study developed in this paper represents an example to investigate how oral consumption of probiotics formulations can lead to increased levels of the consumed species in the vagina of the women recruited in the study that showed anti-microbial activity against microorganisms potentially involved in uro-genital infections.

Materials and methods

Strains, probiotic formulations and culture conditions

This study comprised five strains of Lactobacillus spp. and Bifidobacterium spp. supplied from a private collection (Principium Europe Srl) (Table 1).

Table 1 List of the strains used in this study, deposit number and the most relevant antimicrobial activities describe in Presti et al. [18], 2015
Full size table

Unless otherwise specified, Lactobacillus spp. strains were cultured in deMan, Rogosa and Sharpe (MRS) medium. For Bifidobacterium spp. strains the MRS medium was supplemented with 0.3 g/L l-cysteine hydrochloride monohydrate (cMRS) (Sigma-Aldrich). The cultures were incubated at 37 °C under microaerophilic or anaerobic conditions using anaerobic atmosphere generation bags (Anaerogen, Oxoid).

Two different formulations containing lactobacilli and bifidobacteria of the study (mix F_1 and mix F_2) or placebo (mix F_3) were prepared (Table 2). The composition of the probiotic mix F_1 was as follows: 5 × 109 CFU L. acidophilus PBS066 (40 mg as lyophilized), 5 × 109 CFU L. reuteri PBS072 (30 mg as lyophilized), 320 mg inulin, 5 mg silica, 5 mg talc. The F_2 composition was as follows: 5 × 109 CFU L. plantarum PBS067 (12 mg as lyophilized), 5 × 109 CFU L. rhamnosus PBS070 (20 mg as lyophilized), 5 × 109 CFU B. animalis subsp. lactis PBS075 (60 mg as lyophilized), 298 mg inulin, 5 mg silica, 5 mg talc. Placebo (F_3) composition was as follows: 390 mg inulin, 5 mg silica, 5 mg talc.

Table 2 Contents of formulations of probiotics used in this study
Full size table

Both single strains and formulations were tested for their anti-microbial activity [18]. As antagonistic microorganisms for anti-microbial activity assays, E. coli ATCC 25922 and C. albicans ATCC 10231 were employed. With the exception of C. albicans, maintained in Sabouraud (Oxoid) agar (1.5% w/v Agar Technical, Oxoid) medium, the antagonists were cultured in Tryptic Soy Agar (TSA; Oxoid) at 37 °C in aerobiosis.

Assessment of the anti-microbial activity

Inhibitory activity of formulates not-neutralized cell-free supernatant on the growth of antagonistic microorganisms

Over-night MRS (or cMRS) cultures were centrifuged at 12,000×g at 4 °C for 10 min. The pHs of the supernatants were recorded and measured. An aliquot of the not-neutralized (NN) supernatant fractions was filtered with 0.22 µm pore filter membranes to remove any residual bacterial cell.

The antimicrobial activities were detected by measuring the growth inhibition of E. coli ATCC 25922 and C. albicans ATCC 10231 in liquid cultures in the presence of the NN cell culture supernatants. Single colonies from freshly streaked plates of the antagonists were resuspended in saline solution at a concentration of 106 CFU/mL. The assay was performed in 96-well plates. Each well contained: double concentrated TSB, 10% (v/v) cell suspension of antagonistic strains, 25% (v/v) of NN cell culture supernatants and distilled water up to 150 µL as final volume.

Positive controls were prepared by substituting to NN cell culture supernatants an equal volume of not-inoculated MRS medium. Bifidobacterium sp. strain NN supernatant was used as Ref. [18]. Plates were incubated at 37 °C in aerobiosis for 24 h. At regular times, the cultures were sampled and the growth was evaluated by count plate technique. Experiments were performed three times with a standard error around ±10%.

Anti-microbial activity of formulates vs antagonistic microorganisms by using living cells

The antimicrobial activity of the probiotic mixtures containing lactobacilli and bifidobacteria living cells against the antagonists was evaluated by the overlay method, using the protocol described by Presti et al. [18]. A suspension of each mix F_1 and mix F_2 or placebo (mix F_3) was cultured into MRS broth until the O.D.600nm was 0.2. Then, 50 µL of each culture was spread by forming a stripe 2 cm wide across the MRS (or cMRS) agar plates. The plates were incubated in anaerobiosis at 37 °C for 24 h. After strain growth, plates were overlaid with 10 mL of melted TSA or MYPG (3 g/L Malt extract, Difco; 3 g/L Yeast extract, Biolife; 3 g/L Bacto Peptone, Difco; 2 g/L glucose, Sigma-Aldrich; pH, 6.2) soft agar (8 g/L) media. E. coli ATCC 25922 and C. albicans ATCC 10231 colonies from freshly streaked plates were resuspended into 0.9% (w/v) saline solution at a concentration of 108 CFU/mL. Cell suspensions were streaked over TSA (or MYPG in the case of C. albicans) surface with a cotton swab. The plates were incubated at 37 °C for 24 h in aerobic conditions. The anti-microbial ability of the examined lactobacilli and bifidobacteria strains was semi-quantitatively evaluated in terms of absent (−), moderate (+) and strong (++) growth inhibition of the antagonist, depending on the dimension of the inhibition halos.

DNA extraction and manipulation

DNA from microbial cultures was extracted by the “InstaGene Matrix” (Biorad): 1 mL of culture (109 CFU/mL) was centrifuged at 4000 rpm and 12 °C for 15 min. DNA was extracted from the pellet following the protocol provided by the manufacturer. DNA was also extracted from tenfold dilutions of the 109 CFU/mL culture up to 102 CFU/mL to set up qPCR analysis.

DNA from microbial cultures was used as template to set up primers and PCR conditions to apply for the detection of bacteria of the formulates by qPCR. Primers were designed as follows. The two DNA regions, pre-16S rRNA and IS 16S/23S rRNA sequences [19, 20] were used to identify species-specific DNA markers. For each probiotic, the obtained sequence served as a query to perform BLAST searches against publically available nucleotide databases (http://blast.ncbi.nlm.nih.gov). We collected the top BLAST hits, with a query cover >70% and identity >80%, and aligned them with the query sequence by using Clustal Omega [21] at http://www.ebi.ac.uk/Tools/msa/clustalo/. Forward and reverse primers were manually designed with the least conserved nucleotide sequences. Selected primers were firstly tested in silico on “Primer-BLAST” at the NCBI BLAST web site (http://www.ncbi.nlm.nih.gov/tools/primer-blast) running a default BLAST search against the entire database with the designed primers as queries.

PCR analyses were performed by using the identified primers and by using 10 ng of DNA extracted from each microbial culture. PCRs were performed with by puReTaq Ready-To-Go PCR beads (GE HealthCare Biosciences, Buckinghamshire, UK) in a 25 μL reaction according to the manufacturer’s instructions. PCR cycles consisted of an initial denaturation step for 7 min at 94 °C, 35 cycles of denaturation (45 s at 94 °C), annealing (30 s at 55 °C), extension (1 min at 72 °C) and a final extension at 72 °C for 7 min.

The efficacy of amplification was verified by agarose gel electrophoresis and ethidium bromide staining. In addition, each PCR product was sequenced to verify the correspondence of the selected region. Primer synthesis and DNA sequencing were supplied by Primm, Milan, Italy.

qPCR: primers validation and standard curves

Semi-quantitative analyses of probiotics were set up using qPCR. The 10 μL qPCR mix was assembled with 5 μL of “SsoFast EvaGreen Supermix with Low ROX” (BIO-RAD) and 0.5 μL of 10 μM primer forward and 10 μM primer reverse and 4 μL of DNA template. qPCR was performed in triplicate in the ABI 7500 Real-Time PCR system (Applied Biosystems) using standard cycling conditions: 2 min at 50 °C, 10 min at 95 °C, 40 cycles of 15 s at 95 °C and 1 min at 60 °C. This qPCR program was followed by a dissociation step to verify specificity.

Tenfold dilutions of DNA extracted from microbial cultures were tested in order to verify the detection limit of the qPCR reaction and hence select an optimum number of cycles for the reaction.

Standard curves were constructed for each strains using DNA extracted from microbial cells.

Cells were prepared using tenfold dilutions ranging from 109 to 102 CFU/mL.

Pilot study design

This study consisted in a randomized, double-blind, three-arm parallel group, and placebo-controlled pilot study. It involved 60 pre-menopausal women aged between 18 and 50 years old not suffering from vaginal or urinary tract infections during the 12 months prior to the date of their enrollment in the study. All the study procedures were approved by the Independent Ethics Committee for Non-Pharmacological Clinical Investigations and were carried out according to World Medical Association’s (WMA) Helsinki Declaration and its amendments (Ethical Principles for Medical Research Involving Human Subjects, adopted by the 18th WMA General Assembly Helsinki, Finland, June 1964).

All subjects provided a written informed consent before initiation of any study-related procedures. The study took place at Farcoderm Srl facilities. Farcoderm Srl is an independent testing laboratory for in vitro and in vivo safety and efficacy assessment of cosmetics, food supplements and medical devices.

Depending on a randomization plan, the 60 volunteers were divided into three groups of treatment; each subject was treated, for 14 days, respectively, with a dietary supplement containing F_1 or F_2 or placebo F_3 (Table 2). Vaginal swabs of the enrolled subjects were collected within first 4–5 cm of vagina, to take a sample of vaginal secretion, at four experimental times: the day before first intake of probiotics, after 7, 14 and 21 days from the first intake (day 21 represents seven days after the end of the treatment). Vaginal swabs were stored at −20 °C until total DNA isolation.

The tested products consisted of food supplements (capsules) containing lactobacilli and bifidobacteria (Principium Europe Srl, Solaro, MI, Italy) (Table 1). The composition of the probiotic mix F_1, F_2, and F_3 is reported above (Table 2).

DNA analysis of vaginal samples

Total DNA was extracted from the vaginal swabs by the “InstaGene Matrix” (Biorad) according to the following protocol: 120 μL were taken from each swab and centrifuged at 13,000 rpm at 10 °C for 10 min; 200 μL of “InstaGene Matrix” were added to the pellet; after an incubation at 56 °C and 1000 rpm for 30 min and 8 min at 100 °C, the solution was centrifuged at 12,000 rpm for 3 min. The supernatant containing DNA was collected and stored at −20 °C. DNA quality was verified by agarose gel electrophoresis.

DNA extracted from swabs was analyzed in triplicate, with each one of the five primer sets, with qPCR mix and qPCR protocol described before. We chose to analyze 4 μL of 1:10 dilutions of isolated DNA in order to reduce the effect of possible PCR inhibitors.

Each qPCR run contained multiple Non-Template Controls (NTC).

Statistical methods

To examine the effect of the administration of different probiotic formulations, we used a linear mixed model (LM). The different probiotics, the capsule (treatment: F_1, F_2, F_3) and the time point at four different experimental times (timepoint: t0, t7, t14, t21 days) were used as the explanatory variables. Moreover, we considered volunteers (sample) as random effect.

To apply generalized linear mixed model (GLMM) under Poisson-lognormal error to account for higher variation at the lower end of target abundance, MCMC.qpcr R package was used to convert Ct data in bacterial counts. The conversion to approximate counts uses the following formula:

$$ {\text{Count}}: \, E^{{({\text{Ct1}} - {\text{Ct}})}} $$

where E is the efficiency of amplification and Ct1 is the number of qPCR cycles required to detect a single target molecule.

Markov Chain Monte Carlo (MCMC) algorithm implemented in the package is used to sample from the joint posterior distribution over all model parameters, in order to estimate the effects of all experimental factors on the levels of specific microbial species. GLMM was used to test whether the levels of the different microbial species in different formulation groups (F_1, F_2, F_3) differed between the baseline (t0) and the subsequent time points (t7, t14, t21 days).

The experimental design is incorporated into the following model:

$$ { \ln }\left( {\text{counts}} \right) \, \sim {\text{species }} + {\text{ species}}:{\text{Formulation }} + {\text{species}}:{\text{Time }} + {\text{sample}} + {\text{species}}:{\text{sample}} + {\text{species}}:{\text{residual}} $$

where the logarithm of bacterial counting rate is the response variable and the fixed factors are Formulation and Time (baseline and subsequent time points). The three remaining factors: sample (different subjects of the study), species:sample and species: residual are defined as random factors, accounting for the variation in quality and quantity of biological material among samples.

In particular, lmer package [22] lmerTest package [23] was used to perform linear mixed models and do statistical tests. Plots are performed using ggplot2 package [24].

Results

Anti-microbial activity of the probiotic formulates

The anti-microbial activity of formulates F_1 and F_2 was tested against C. albicans and E. coli, as an example of microorganisms responsible for uro-genital infections, by growth inhibition with non-neutralized cell-free supernatants and by overlay assay on living cells. Both probiotics mixtures showed a strong inhibition rate against E. coli mediated by microbial culture supernatants during 24 h, while no influence was detected vs C. albicans. The inhibition degree of each probiotic formulation compared to the corresponding single strain is reported in Fig. 1. The curve of activity of formulates is in line with the expected profile from single strain performance. The inhibition capacity of the mixtures F_1 and F_2 was further investigated by direct contact of their cultures against the same microorganisms. The growth inhibition halos around the multiple colonies were coherent with the anti-microbial inhibition profiles detected for the single strains. F_1 showed a moderate activity against both C. albicans and E. coli according to a better performance of L. acidophilus contrary to L. reuteri, while a stronger inhibition capacity vs both pathogens was observed for F_2. The results are showed in Fig. 2.

Fig. 1
figure 1

Effect of non-neutralized culture supernatants of L. acidophilus (square), L. plantarum (cross), L. rhamnosus (triangle), B. animalis subsp. lactis (filled square), L. reuteri (filled triangle), mixture F_1 (diamond), mixture F_2 (filled diamond) on the growth of E. coli and C. albicans: A1 mixture F_1 vs E. coli; A2 mixture F_1 vs C. albicans; B1 mixture F_2 vs E. coli; B2 mixture F_2 vs C. albicans

Full size image
Fig. 2
figure 2

Inhibition growth of E. coli and C. albicans over probiotics formulate cultures: A1 mixture F_1 vs E. coli; A2 mixture F_1 vs C. albicans; B1 mixture F_2 vs E. coli; B2 mixture F_2 vs C. albicans

Full size image

Set up of DNA molecular tools for probiotics identification

To identify each probiotic, polymorphic DNA regions were identified among the pre-16S rRNA and IS 16S/23S sequences [19, 20]. BLAST analysis performed against the entire default database (nr/nt), excluding species of interest, allowed to identify the most polymorphic regions (data not shown).

For L. rhamnosus and L. plantarum the DNA marker regions were found in the 16S/23S IS, while for the other three probiotics the pre-16S partial sequence resulted more polymorphic and suitable to identify marker regions. For each selected marker, species-specific primer pairs were identified (Table 3).

Table 3 List of primers designed and used in this study
Full size table

For each primer combination, standard qPCR curves were constructed by using DNA of pure culture of each strain. Results are reported in Table 4, together with the R 2 value, the slope and the efficiency of the amplification. The values obtained show how the qPCR reaction set up is optimal. The R 2 values are high (>0.99) and the efficiencies are between 80 and 115%, limits necessary for a reliable standard curve [25]. The analysis performed on different culture dilutions showed a linear dynamic range to be between 109 CFU/mL of culture and 102 CFU/mL.

Table 4 Standard curves equations: y corresponds to the Cq and x the bacterial dilution in respect to 109 CFU (maximum concentration)
Full size table

Evaluation of vaginal probiotic amount in the subjects of the study

The selected primers were used to estimate the presence and persistence of each probiotic species in vagina from swab samples of the enrolled subjects by qPCR. PCR analyses were performed on DNA extracted from vaginal swabs at time t0, t7, t14 and t21 days.

The qPCR analysis demonstrated that the species-specific sequences associated with the probiotics of the formulations were detected only in vaginal DNA from subjects treated with the formulations F_1 and F_2 and not with the formulation F_3. Figure 3 shows the abundance of the different probiotic species for each treatment group at the different experimental times.

Fig. 3
figure 3

Ratio of probiotics of formulations (F_1 and F_2, vs F_3) by qPCR of species-specific sequences at the different times of treatment vs the amount at the baseline time point, expressed as bacterial counts. Upon the bars is reported the statistical analysis between treatments (***p < 0.001; **p < 0.01; *p < 0.05)

Full size image

The subjects treated with F_1 showed an increase in the level of both L. acidophilus and L. reuteri compared with F_3. This has been observed at all the times compared with the t0, including t21, 7 days after the follow-up from the last probiotics administration. The same trend was observed for L. rhamnosus, L. plantarum and B. animalis subsp. lactis in women treated with F_2 formulation. Differently, the amount of all five strains remained constant throughout the study in the vaginal DNA of women treated with the placebo F_3.

The increase of microbial cell number and the permanence of probiotic species at higher levels after the end of the oral administration (t21) indicate that these species are actually more abundant in the vaginal microbiota. This increase was observed with statistical significance (p value <0.05) since 7 days after the beginning of treatment for L. reuteri and L. acidophilus of the F_1 group and for L. rhamnosus and B. animalis subsp. lactis of the F_2 group. Instead, the increase of L. plantarum resulted statistically significant (p value <0.05) since 14 days from the beginning of the treatment. The permanence of bacteria was observed for all the species studied until the t21, 7 days from the last probiotics intake. In fact, the differences of t14 and t21 from t7 are not statistically significant (p value >0.05) for all probiotics, with the exception of L. plantarum.

For F_3 formulation there were not so variations in the amount of selected probiotics for the entire 21-day period.

Discussion

The microbial species that inhabit the vaginal tract play an important role in the maintenance of health and prevention of infections. In particular, the presence of high numbers of lactic acid bacteria in the vagina is often equated with an health status [26]. It appears evident that the balance between a healthy and diseased state involves an equilibrium which can depends on different factors, such as hormone levels, douching, sexual practices, as well as bacterial interactions, and host defenses [27, 28]. A way to increase the level of vaginal lactobacilli is through the use of probiotics; two ways are commonly applied: direct application through vagitories or indirect application by oral consumption of probiotics. Several probiotics have been found to both increase the overall level of vaginal lactobacilli and to aid in the treatment of bacterial vaginosis [26]. Moreover, the capacity of probiotics to exert beneficial effects in human health is recognized to be strain-specific and a multi-species formulation could take the advantage of combining a greater spectrum of activities.

In this context, the present work allowed us to assess in the vagina of 60 pre-menopausal healthy women of a pilot study, the detection of orally administered multispecies probiotic formulations showing anti-microbial properties in test in vitro. First, we tested the antagonistic capacity against E. coli and C. albicans of the formulation F_1 and F_2 through cell supernatants and through the overlay contact between probiotic formulations and pathogens. Both formulations showed a strong anti-microbial activity against E. coli in both the conditions, as expected by the average trend of the single strains included in the formulations, while inhibition against C. albicans occurred only in the overlay assay, likewise the tested single strains. This is probably due to the reliable inhibitory activity of lower pH upon the E. coli growth, while the same conditions are not effective against yeasts. Moreover, it is reported that bacteriocins produced by probiotics are active only at certain pH ranges but can be neutralized at different pHs; this could be the reason for not having observed any inhibitory effect on the pathogens growth by using the neutralized cell supernatants [18].

Probiotics can reach the target of vagina when orally consumed, as an alternative way to a direct local administration; so they can locally exert their effect by competition–displacement of pathogens, reducing the infection relapses and related symptoms [26, 29].

Although previous studies [30–32] had already suggested this possibility, it is not so expected and the mechanism is still unclear. The current pilot study was aimed to investigate the vaginal detection after oral consumption of mixtures of probiotics would lead to increased levels of the consumed multispecies in the vagina of the enrolled subjects.

Our analysis showed an increased levels of all probiotics species (p value <0.05) detected in the vaginal swabs of women consuming the formulates F_1 and F_2 in comparison to women of the placebo group. The only exception was for L. plantarum, which was higher starting from 7 to 21 days of administration. This might suggest that: (1) L. plantarum is not as efficient as other probiotics, in colonizing vaginal mucosa from the intestinal region, (2) L. plantarum is already present at high levels in vaginal mucosa of healthy women and its abundance is not easily perturbed by the oral administration of probiotics. For this reason, L. plantarum might need a longer treatment or may require a higher concentration in administered capsules.

The abundance of probiotic strains in the vaginal DNA was assessed for both formulations until the last day of the experiment (t21 days), 7 days after the last intake (p value <0.05). This indicates that probiotics can actually colonize the vaginal microbiota in a short time. It would be particularly interesting to demonstrate if these can persist also after the menses. Indeed, the set-up of this study was purposely planned to match menstrual cycle of volunteers, with the first swab collected after menses and the last 21 days later, before the next menses. This was done to prevent from having blood traces in the vaginal swabs, so possibly altering the results.

The DNA markers used allowed to monitor the assessment and persistence of each probiotic species in the vaginal microbiota of healthy pre-menopausal women during the treatment and after the follow-up period. Our findings suggested that the five selected DNA markers can detect the increasing level and persistence of the studied bacteria into the vagina by using qPCR methods.

As lactobacilli and bifidobacteria have a fundamental role on the vaginal well-being, and more specifically an anti-microbial activity, we would highlighted that the anti-microbial effect detected against several pathogen microorganisms suggests that the five selected strains could similarly exert the antagonistic activity in vivo.

Conclusions

In conclusion, this study reports that the oral intake of two probiotic multi-species mixtures leads to an evident colonization of vagina of 60 volunteers of probiotic bacteria showing in vitro anti-microbial activity against pathogens involved in uro-genital infections. The adopted molecular tool represents a valid instrument to be used in future clinical trials to correlate the eventual clinical outcomes with the effective colonization in the treatment or prevention of vaginal dismicrobism and uro-genital infections.