Dietary supplementation with probiotics during late pregnancy: outcome on vaginal microbiota and cytokine secretion

Background The vaginal microbiota of healthy women consists of a wide variety of anaerobic and aerobic bacterial genera and species dominated by the genus Lactobacillus. The activity of lactobacilli helps to maintain the natural healthy balance of the vaginal microbiota. This role is particularly important during pregnancy because vaginal dismicrobism is one of the most important mechanisms for preterm birth and perinatal complications. In the present study, we characterized the impact of a dietary supplementation with the probiotic VSL#3, a mixture of Lactobacillus, Bifidobacterium and Streptococcus strains, on the vaginal microbiota and immunological profiles of healthy women during late pregnancy. Results An association between the oral intake of the probiotic VSL#3 and changes in the composition of the vaginal microbiota of pregnant women was revealed by PCR-DGGE population profiling. Despite no significant changes were found in the amounts of the principal vaginal bacterial populations in women administered with VSL#3, qPCR results suggested a potential role of the probiotic product in counteracting the decrease of Bifidobacterium and the increase of Atopobium, that occurred in control women during late pregnancy. The modulation of the vaginal microbiota was associated with significant changes in some vaginal cytokines. In particular, the decrease of the anti-inflammatory cytokines IL-4 and IL-10 was observed only in control women but not in women supplemented with VSL#3. In addition, the probiotic consumption induced the decrease of the pro-inflammatory chemokine Eotaxin, suggesting a potential anti-inflammatory effect on the vaginal immunity. Conclusion Dietary supplementation with the probiotic VSL#3 during the last trimester of pregnancy was associated to a modulation of the vaginal microbiota and cytokine secretion, with potential implications in preventing preterm birth. Trial registration ClinicalTrials.gov NCT01367470


Background
The vaginal microbiota of healthy women consists of a wide variety of anaerobic and aerobic bacterial genera and species dominated by the facultative, microaerophilic anaerobic genus Lactobacillus [1]. The activity of lactobacilli helps to maintain the natural healthy balance of the vaginal microbiota. This role is particularly important during pregnancy because abnormalities in vaginal communities, such as bacterial vaginosis (BV) and aerobic vaginitis (AV), have been claimed as important mechanisms responsible for preterm birth and perinatal complications [2].
The association of lower genital tract infection with an increased risk of preterm delivery and preterm rupture of the fetal membranes has recently attracted great interest in the pathogenesis of such infection-related mechanisms [3,4]. Earlier studies showed an increased rate of prematurity in women with BV, an alteration of the endogenous vaginal microbiota associated with decreased levels of hydrogen peroxide-producing Lactobacillus species [4][5][6]. The mechanisms linking BV with preterm delivery have not been fully identified, but local immune response is hypothesized to be crucial. Despite the notion that BV is a non-inflammatory condition, evidence exists that demonstrates altered levels of certain pro-inflammatory cytokines in women with BV [7,8].
Parturition is characterized by cervical ripening and myometrial maturation with subsequent uterine contractions leading to cervical dilatation and birth [9]. The process of labor displays many of the hallmarks of inflammation. Acute inflammatory features, such as increased influx of leucocytes and elevated expression of pro-inflammatory cytokines, have been observed in cervical tissues and fetal membranes during both term and preterm labor [10][11][12].
A potentially novel way to protect against infectionmediated preterm birth is to use probiotic bacteria, especially lactobacilli. Probiotics, defined as "live microorganisms which, when administered in adequate amounts, confer a health benefit on the host" [13], are being studied for their ability to replenish vaginal lactobacilli and modulate immunity [14][15][16]. In addition, administration of probiotics to the mother during pregnancy and breast-feeding has been described by some studies as a safe and effective mode of enhancing the immunoprotective potential of the breast milk and preventing atopic eczema in the infant [17,18].
In recent years, culture-independent techniques based on the analysis of rRNA gene sequences have been developed, providing powerful tools to reveal the phylogenetic diversity of the microorganisms found within vaginal microbiota and to understand community dynamics [19][20][21][22][23][24]. In particular, PCR-denaturing gradient gel electrophoresis (PCR-DGGE) has been successfully used to identify the bacterial composition of different ecological niches, including the vaginal microbiota [22,25,26]. Realtime PCR is a powerful technique for the quantitative analysis of specific microbial populations belonging to complex ecosystems [22,27,28]. Specific primers can be used to focus the quantitative analysis on a particular genus, species or strain of interest.
Several bacterial species are known to colonize both the gastrointestinal and the reproductive tract, and the rectum has been suggested to play an important role as a source or reservoir for organisms that colonize the vagina [15,29]. On this basis, the aim of the present study was to evaluate the impact of a dietary supplementation with the probiotic product VSL#3, a mixture of Lactobacillus, Bifidobacterium and Streptococcus strains, on the vaginal microbiota and immunological profiles of asymptomatic healthy women during late pregnancy. The dynamics of the vaginal bacterial communities prior and after probiotic ingestion were assessed by PCR-DGGE and real-time PCR, while the modulation of the cytokine secretion in vaginal fluids was measured by Luminex W Immunoassay. Although previous studies demonstrated the therapeutic efficacy of VSL#3 in the management of gastrointestinal disorders, especially inflammatory bowel disease [30], as well as the ability of the VSL#3 strains to colonize the gut environment [31] and to modulate the immune response of the colonic mucosa [32], this is the first study that investigates the indirect effects of this probiotic formula on the vaginal microbiota.

Results
Bacterial population profiling with PCR-DGGE PCR-DGGE analysis with universal primers for bacteria (HDA1-GC/HDA2) was used to investigate: (i) the stability of the predominant vaginal bacterial communities over a period of 4 weeks in the last trimester of pregnancy, from the 33 rd (W33) to the 37 th (W37) week of gestation, and (ii) the influence of the oral consumption of the probiotic VSL#3 from W33 to W37 on the predominant vaginal microbiota ( Figure 1). DGGE band profiles displayed a relatively low complexity for both probiotic (P) and control (C) groups, as assessed by the richness index. Mean values of the richness index were 6.6 at both W33 and W37 for C group and shifted from 8.4 (W33) to 7.4 (W37) for P group without significant variations between W33 and W37.
Pearson correlation was used to calculate the similarity index (SI) between DGGE patterns related to the time points W33 and W37 for each pregnant woman ( Table 1). The SI median values of P group and C group were 73% and 79%, respectively. In particular, 3 women belonging to P group (N. 2, 9 and 10) and only one woman belonging to C group (N. 24) showed SI values lower that 50%. For each woman, significant differences between DGGE profiles related to W33 and W37 were searched by Wilcoxon Signed Rank Test. No significant variations were detected between W33 and W37 in control women. Significant differences (P < 0.05) were found for 5/15 (33%) women belonging to P group (N. 4,5,9,10,11). Interestingly, women N. 9 and 10 were the same presenting SIs < 50%. These data suggested a potential role of the probiotic formula in modulating the vaginal bacterial communities. The peak heights of the DGGE densitometric curves were analyzed using the Wilcoxon Signed Rank Test in order to search for significant differences in single species abundances between W33 and W37. No significant changes in species abundance were found for both P and C groups, even in women N. 4, 5, 9, 10, 11. Cluster analysis showed that the DGGE profiles related to the time points W33 and W37 clustered together for all the control women, except for the woman N. 24 ( Figure 1). Four supplemented women (N. 2, 9, 10 and 15) showed W33 and W37 DGGE profiles not closely related. However, the DGGE patterns of the majority of the women administered with VSL#3 grouped according to the subject and not to the time point, revealing that the inter-individual variability was higher than the variability induced by the probiotic supplementation.
Because of the importance of lactobacilli in the establishment of a healthy vaginal environment [2], DGGE analysis with Lactobacillus-specific primer set (Lac1/ Lac2-GC) was also carried out. This analysis allowed us to investigate the variations in lactobacilli population occurring physiologically from W33 and W37 and potentially associated with the VSL#3 intake ( Figure 2). Richness indexes ranged from 5.7 (W33) to 5.4 (W37) for P group and from 6.3 (W33) to 6.8 (W37) for C group. Mean values of SI were 79% and 80% for P and C groups, respectively (Table 1). Only 2 women included in P group showed SIs < 50% (N. 1 and 15). Wilcoxon Signed Rank Test highlighted significant differences between DGGE profiles related to W33 and W37 for women N. 7 and 10, accounting for 13% of women included in P group. Comparing this percentage with the 33% obtained by DGGE analysis with HDA1-GC/HDA2 primer set, the probiotic intake seemed to have a more extended impact on total bacteria than lactobacilli. Notably, only for woman N. 10, significant differences were found between W33-and W37-related DGGE patterns for both HDA1-GC/HDA2 and Lac1/Lac2-GC primer sets.
The peak height analysis by Wilcoxon Signed Rank Test allowed us to identify a band, denominated L16 ( Figure 2), which significantly changed after probiotic supplementation. Sequencing of the DNA extracted from this band revealed 100% homology with L. helveticus strains. The nucleotide sequence of this DGGE fragment was deposited in DDBJ Nucleotide Sequence Database under the accession number AB571603. L. helveticus was found to be a representative species within lactobacilli population since it was detected in 9 women supplemented with VSL#3 and 2 control women, corresponding to a frequency of occurrence of 40.7%. Notably, a general decrease in the intensity of L. helveticus band was observed in P group while no variations were appreciable in C group.
Cluster analysis showed that Lactobacillus-specific DGGE profiles related to the time points W33 and W37 were closely related for all control women and for the majority of women administered with VSL#3, except for the subjects N. 1 and 15 ( Figure 2).

Discussion
To our knowledge, this is the first study describing the effect of a probiotic mixture, orally consumed during the last trimester of pregnancy, on the vaginal microbiota and immune response. Although several healthpromoting activities of probiotics have been described in relation to the gut homeostasis [16,32], less information is available regarding the interactions between orally administered probiotic bacteria and the vaginal microbial habitat.
The first step in ascertaining the influence of the dietary supplementation with the probiotic VSL#3 on the vaginal microbiota of pregnant women was the characterization of vaginal bacterial communities by using an integrated approach based on PCR-DGGE and qPCR.  2.3 × 10 -1 ± 1.5 × 10 -2 0.0 ± 0.0 W37 6.8 × 10 1 ± 5.1 2.7 × 10 -2 ± 6.6 × 10 -3 1.9 × 10 -1 ± 2.0 × 10 -2 0.0 ± 0.0 DGGE population profiling, conducted with universal primers for bacteria and Lactobacillus-specific primers, allowed us to investigate the variations of the predominant vaginal bacterial communities and Lactobacillus species occurring both physiologically in the last trimester of pregnancy and potentially associated with VSL#3 intake. The influence of the probiotic intake in modulating the predominant bacterial populations and Lactobacillus species could be hypothesized since significant differences between DGGE profiles at W33 and W37 were found only in women belonging to P group. Notably, the lower percentage of women belonging to P group who displayed significant differences in Lactobacillus-specific DGGE profiles between W33 and W37, compared to the universal bacterial DGGE patterns, suggested a major stability of lactobacilli population and a more extended impact of the probiotic VSL#3 on total bacteria than lactobacilli. However, no significant changes in single species abundances were found between W33-and W37-related universal DGGE profiles. Differently, the statistical analysis of the peak heights of the Lactobacillus-specific DGGE densitometric curves allowed us to identify a band, corresponding to L. helveticus, which significantly decreased after probiotic supplementation. Strains belonging to L. helveticus are used as starter cultures in the manufacturing of a variety of fermented dairy products, to modulate flavor. The presence of L. helveticus in vagina, likely due to the migration from the gut, can be related to a diet rich in yogurt and cheese. This work is not the first describing L. helveticus in vaginal samples. Stoyancheva et al. [33] identified this species among several Lactobacillus isolates from vaginal fluids of healthy Bulgarian women in childbearing age by using three different molecular techniques, amplified ribosomal DNA restriction analysis, ribotyping and PCR with species-specific primers. The decrease of L. helveticus observed in our study could be due to a competition between the Lactobacillus strains present in VSL#3 formula and dairy L. helveticus strains in colonizing vaginal environment.
Cluster analysis showed that universal and Lactobacillusspecific DGGE profiles related to the time points W33 and W37 of the control women were closely related. Also the DGGE patterns of the majority of women administered with VSL#3 grouped according to the subject and not to the time point, revealing that the interindividual variability was higher than variability induced by the probiotic supplementation.
The hypothesis of a positive action of VSL#3 on the vaginal microbiota of pregnant women was further supported by qPCR results, which suggested a role of the probiotic product in counteracting the decrease of the health-promoting Bifidobacterium genus and the increase of the BV-related Atopobium genus, that occurred in control women during late pregnancy. Notably, group B Streptococcus, which was found in two women (N.1 and 10) before the probiotic intake, was no longer found after the dietary supplementation (data not shown).
The second step of the present research was the investigation of the vaginal immunological profiles of the pregnant women in order to search for correlations between the VSL#3 intake and changes in vaginal immune response. Pregnancy has been referred to as a state of relative immune compromise. This notion has been related to both demonstration of depression of certain aspects of cell-mediated immunity and clinical observations of an increased severity of numerous infectious conditions in pregnant women [7]. On the other hand, preterm cervical ripening can be likened to an inflammatory process with cytokines as important mediators [34].

Conclusion
The impact of the oral intake of the probiotic VSL#3 on the vaginal microbiota and immune response of pregnant women was investigated by molecular fingerprinting techniques (PCR-DGGE and qPCR) and Luminex W immunoassay. The major findings of this study are the following: (i) VSL#3 intake seems to be associated with a modulation of the predominant vaginal bacterial communities; (ii) VSL#3 modulation of Lactobacillus population appears to be related to variations of L. helveticus species; (iii) a potential role of the probiotic product in counteracting the physiological decrease of Bifidobacterium and increase of Atopobium could be hypothesized; (iv) the probiotic supplementation can be associated with a global antiinflammatory effect on the vaginal immunity, with potential implications in preventing preterm birth.

Study design and sample collection
A pilot, not randomized, controlled and perspective study was conducted. The study protocol was approved by the ethical committee of the University of Bari, Italy. Written informed consent was obtained from all the participants in the study. A total of 27 healthy pregnant women (21 to 42 years of age; mean, 32) who had no symptoms of vaginal or urinary tract infection were included in the present study ( Table 3). None of the subjects had received oral or local antimicrobial therapy within the previous 2 weeks. The recruited subjects were divided into 2 groups: (i) probiotic group [P (n=15)]; (ii) control group [C (n=12)] on the basis of their availability to consume the probiotic product. Women of the P group consumed 1 sachet once/day of VSL#3 (VSL Pharmaceuticals, Inc.,Towson, MD, USA) for 4 weeks from the 33 rd (W33) to the 37 th (W37) week of gestation. Women of the C group did not receive any dietary supplementation. VSL#3 sachet contains 900 billion viable lyophilized bacteria consisting of 4 strains of Lactobacillus (L. paracasei, L. plantarum, L. acidophilus, L. delbrueckii subsp. bulgaricus), 3 strains of Bifidobacterium (B. longum, B. breve, B. infantis) and 1 strain of Streptococcus thermophilus. Mid-vaginal swabs were collected from women of both P and C groups at the time points W33 and W37. Samples were placed in 1 ml of sterile saline and stored immediately at −80°C until use.
The individual characteristics (age, type of delivery and gestational age at birth) of women enrolled in the present study are reported in Table 3. Gestational age was determined by utilizing the last menstrual period and earliest ultrasound. C P Figure 5 Women registering significant variations in total levels of immune-mediators. P, probiotic group; C, control group; W33, 33 rd gestational week (black colour); W37, 37 th gestational week (grey colour). Identification numbers of women registering significant variations are reported in x-axis. Data are expressed as pg of total immune-mediators per μg of total vaginal proteins (y-axis). The diagrams show means with error bars representing the standard deviations.

DNA extraction from vaginal samples
Frozen vaginal swabs were thawed, mixed by vortex shaker for 1 min and then removed from the liquid. The liquid was centrifuged at 10,000 × g for 15 min, and the pellet was washed 3 times in saline at 40°C. The pellet was resuspended in 180 μl of enzymatic lysis buffer (20 mM Tris-HCl, pH 8, 2 mM EDTA, 1.2% Triton X-100, 20 mg/ml lysozyme) and incubated at 37°C for 30 min. Glass beads (200 mg) were added and the sample was mixed by vortexing for 1 min. Total DNA was extracted by using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) following the protocol "Pretreatment for Gram-positive bacteria". A slight modification was introduced: a centrifugation step (8000 × g for 5 min) was carried out after incubation with proteinase K to remove glass beads. DNA amounts were quantified by using NanoDrop 1000 (Thermo Scientific, Wilmington, DE).

PCR-DGGE and cluster analysis
Amplification reactions were performed in a Biometra Thermal Cycler T Gradient (Biometra, Göttingen, Germany). GoTaq Flexi DNA Polymerase (Promega, Madison, WI) was used as thermostable DNA polymerase. The reaction mixture contained 0.5 μM of each primer, 200 μM of each dNTP, 2 mM MgCl 2 solution, 1.25 U of GoTaq Flexi DNA Polymerase, 5 μl of Green GoTaq Flexi buffer 5X, and 2 μl of the bacterial DNA template (30-40 ng) in a final volume of 25 μl. The universal primers HDA1-GCclamp and HDA2 for bacteria [39] were used to amplify a conserved region within the 16S rRNA gene. The thermocycle program consisted of the following time and temperature profile: 95°C for 5 min; 30 cycles of 95°C for 30 s, 56°C for 30 s, 72°C for 60 s; and 72°C for 8 min. The Lactobacillus genus-specific primers Lac1 and Lac2-GCclamp [40] were used to amplify a specific region of the 16S rRNA gene of lactobacilli. The amplification program was 95°C for 5 min; 35 cycles of 95°C for 30 s, 61°C for 30 s, 72°C for 60 s; and 72°C for 8 min. A volume of 8 μl of PCR samples was loaded on DGGE gels, containing 30-50% and 25-55% gradients of urea and formamide for universal bacteria and lactobacilli, respectively. DGGE analysis was performed by using the D-Code Universal Mutation System Apparatus (Bio-Rad, Los Angeles, CA), as previously described [22]. Following electrophoresis, gels were silver stained [41] and scanned using a Molecular Imager Gel Doc XR System (Bio-Rad). DGGE gel images were analyzed using the FPQuest software version 4.5 (Bio-Rad). In order to compensate for gel-to-gel differences and external distortion to electrophoresis, the DGGE patterns were aligned and normalized using an external reference marker. The marker for the DGGE analysis with the universal primers for bacteria contained PCR amplicons from Bacteroides, Coriobacterium, Enterococcus faecalis, Bifidobacterium bifidum, Lactobacillus casei, Acidaminococcus fermentas and Atopobium. The marker for the DGGE analysis with Lactobacillus-specific primers contained PCR amplicons from L. plantarum, L. paracasei, L. brevis, L. gasseri, L. acidophilus and L. delbrueckii subsp. bulgaricus. After normalization, bands were defined for each sample using the appropriate densitometric curve. The similarity in the profiles was calculated on the basis of the Pearson correlation coefficient with the Ward clustering algorithm. Cluster analysis of the DGGE patterns was performed using the FPQuest software.

Sequencing of DGGE fragment
The DNA fragment of interest was excised from the denaturing gel with a sterile scalpel, washed once in 1X PCR buffer, and incubated in 20 μl of the same buffer overnight at 4°C. Two μl of the buffer solution were used as a template for PCR reaction. Reamplification of the 16S rRNA region was conducted as described above by employing the primers Lac1 and Lac2 (without the GC-clamp). The re-amplified fragment was purified using the Wizard SV Gel and PCR Clean-up system (Promega), and then subjected to automated sequence analysis of both DNA strands with Lac1 and Lac2. Big-Dye terminators (ABI-PerkinElmer, Foster City, CA) were used with a 377 sequencer (ABI). Sequence identity was determined by comparison with the rRNA gene sequences deposited in GenBank database using BLAST algorithm (www.ncbi.nlm.nih.gov/BLAST).

Bioplex immunoassay
Cytokine levels were determined using a multiplexed bead immunoassay. Prior to assay, vaginal samples were concentrated 10 times with Microcon spin devices (YM3, Millipore Corporation, Billerica, MA) and subsequently resuspended in Bio-Plex Assay Buffer. The levels of 27 immune-mediators, 15 cytokines (IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12(p70), IL-13, IL-15, IL-17, IFN-γ, TNFα), 7 chemokines (MCP-1, MIP-1α, MIP-1β, RANTES, Eotaxin, IL-8, IP-10) and 5 growth factors (PDGF-BB, FGF basic, G-CSF, GM-CSF, VEGF), were measured using the human ultrasensitive cytokine 27-plex antibody bead kit (Bio-Rad). Assays were performed in 96-well filter plates, as previously described [49]. Briefly, the filter plate was prewetted with washing buffer (Bio-Rad) and the solution was aspirated from the wells using a vacuum manifold (Millipore Corporation). Microsphere beads coated with monoclonal antibodies against the different target analytes were added to the wells. Samples and standards were pipetted into the wells and incubated for 30 min with the beads. Wells were washed using a vacuum manifold (Millipore Corporation) and biotinylated secondary antibodies were added. After incubation for 30 min, beads were washed then incubated for 10 min with streptavidin-PE conjugated to the fluorescent protein, R-phycoerythrin (streptavidin/R-phycoerythrin). After washing to remove the unbound streptavidin/R-phycoerythrin, the beads (a minimum of 100 per analyte) were analyzed in the Luminex 200 instrument (MiraiBio, Alameda, CA). The Luminex 200 monitors the spectral properties of the beads to distinguish the different analytes, while simultaneously measuring the amount of fluorescence associated with R-phycoerythrin, reported as median fluorescence intensity. The concentration of the samples was estimated from the standard curve using a fifth-order polynomial equation and expressed as pg/ml after adjusting for the dilution factor (Bio-Plex Manager software version 5.0). Samples below the detection limit of the assay were recorded as zero, while samples above the upper limit of quantification of the standard curves were assigned the highest value of the curve. The intra-assay CV including ultrafiltration and immunoassay averaged 19%. Concentrations of cytokines, chemokines and growth factors were then converted in pg of the target molecule per μg of total proteins present in the vaginal sample.

Statistical analysis
Statistical analysis was performed using SigmaStat (Systat Software, Point Richmond, CA). For each subject, variations of the DGGE profiles related to the time points W33 and W37 were analyzed by Pearson correlation. Significant differences in the intensity of each DGGE band among all vaginal samples and in the amounts of the bacterial genera and species determined by qPCR were searched by using Wilcoxon Signed Rank Test. This test was also used to analyze differences in cytokines, chemokines and growth factors. A P value below 0.05 was considered statistically significant.