Probiotic Lactobacillus Reuteri and Oral Microbiota

Ethics statements The study was approved by the Regional Ethical Review Board in Umeå, Sweden (Dnr 2011-380-31M) and was conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from all participants.

Subjects and study design Healthy adult volunteers, aged 20-66 years, among students and employees at the Faculty of Medicine, Umeå University, Sweden were recruited to a double-blind, randomized controlled trial (RCT) through advertisements. Inclusion criteria were a self-reported healthy status and no use of antibiotics or probiotic products during 3 months prior to the study. Based on previous studies regarding the persistence of probiotic strains [31,32], the recruitment goal was at least 15 people per study group.

Forty-four volunteers, none of whom used tobacco products, were recruited and randomly allocated to either a test (n = 22) or placebo group (n = 22; Figure S1). Participants were asked to allow 2 lozenges per day to slowly melt in the mouth and to circulate the dissolved tablet contents around their mouths. One lozenge was taken in the morning and 1 in the evening for 12 weeks. The test lozenges contained L. reuteri (DSM 17938 and PTA 5289; 108 CFU per strain; BioGaia AB, Stockholm, Sweden), isomalt, hydrogenated palm oil, peppermint and menthol flavoring, peppermint oil, and sucralose (http://www.biogaia.com/product/biogaia-prodentis-oral-lozenges). The placebo lozenges were identical to the test lozenges in appearance, taste, and composition except the lactobacilli. Compliance was monitored as the percentage of lozenges consumed of the total assigned number. The remaining lozenges were counted when the containers were returned for monthly refills to assess this factor. Compliance was considered acceptable if ≤15% of the lozenges remained. Participants were asked to abstain from oral hygiene for 48 h and to not consume any food for at least 4 h before sampling. Participants were also instructed to not eat probiotic products throughout the study period.

Saliva and biofilm sampling Saliva and tooth biofilm samples were obtained immediately before (baseline) and after 4, 8, and 12 weeks of L. reuteri supplementation (Figure S1). Follow-up samples were collected 1 and 6 months after supplementation was terminated. Furthermore, whole stimulated saliva (~5 mL) was generated by chewing 1 g of paraffin and collected into ice-chilled sterile test tubes. One milliliter of saliva was used for cultural analysis, and the remaining saliva was centrifuged at 3,500 × g for 10 min at 4°C. The pellets were stored at -80°C until DNA extraction for strain-specific PCR reactions. For the pyrosequencing analysis, pooled supragingival plaque was collected with sterilized toothpicks and transferred to Eppendorf tubes (Sarstedt, Nümbrecht, Germany) containing 200 µL of TE-buffer (10 mM Tris, 1 mM EDTA, pH 7.6). The samples were stored at -80°C until DNA extraction.

Identification of lactobacilli by culture and PCR Aliquots of saliva were plated onto Rogosa agar (Merck, Darmstadt, Germany) to obtain Lactobacillus counts and on selective agar for tentative identification of the L. reuteri (DSM 17938 and PTA 5289) strains [33]. All plates were anaerobically incubated at 37°C for 48-72 h, except L. reuteri PTA 5289, which was anaerobically incubated at 40°C for 72 h.

DNA was extracted from saliva pellets as described [34]. L. reuteri DSM 17938 and PTA 5289-specific PCR were identified using KAPA2G Robust HotStart PCR Ready Mix (2×) (Kapa Biosystems, Boston, MA, USA) and strain-specific primers [31]. Briefly, 2 µL of DNA extract was added to a total reaction volume of 25 µL (containing 12.5 µL of master mix and each primer pair at a concentration of 0.5 µM). PCR conditions were 95°C for 3 min; 40 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 30 s; and 72°C for 5 min. PCR products were then verified by electrophoresis on 2% agarose gels allowed to run for 80 min at 120 V in 0.5× TBE (Tris/Borate/EDTA) buffer, pH 8.3, followed by ethidium bromide (0.2 µg/µL) staining.

Pyrosequencing analysis For the 454 pyrosequencing analysis, 16 subjects were randomly selected among the test (n = 8) and control (n = 8) subjects (Figure S1). DNA was extracted from the baseline, 12-week exposure, and 1-month follow-up tooth biofilm samples as previously described [34]. The V3-V4 hypervariable region of the 16S rRNA gene was amplified via PCR using the forward primer 347F and reverse primer 803R [35]. For sample identification, fusion primers were created from these primers and unique barcode sequences according to the Roche guidelines for experimental amplicon design (www.454.com). DNA was amplified under the following running conditions: initial denaturation at 94°C for 3 min; 30 cycles of 94ºC for 15 s, 58ºC for 15 s, and 72ºC for 30 s; and a final extension at 72ºC for 8 min.

Amplicon processing and 454 sequencing were conducted at the Lund University Sequencing Facility (Faculty of Science, Lund, Sweden). After amplicon cleaning to remove short fragments (Agencourt AMPure XP; Beckman Coulter, Brea, CA, USA) and inspection (DNA 1000 kit on a 2100 Bioanalyzer; Agilent Technologies, Palo Alto, CA, USA), amplicons were quantified (Quant-iT ds DNA assay kit; Invitrogen, Carlsbad, CA, USA and Quantifluor fluorometer; Promega, Madison, WI, USA) and diluted to obtain a total of 107 copies/μL. Titration and library production (target: 10%-15% enrichment) were performed using emulsion PCR and the Lib-A kit (Roche Diagnostics, Branford, CT, USA). DNA-positive beads were enriched, counted (Innovatis CASY particle counter; Roche Innovatis, Bielefeld, Germany), processed (XLR70 sequencing kit; Roche Diagnostics), and loaded onto a picotiter plate for pyrosequencing on a 454 Life Sciences Genome Sequencer FLX+ machine (Roche Applied Sciences; Penzberg, Germany).

Data processing Subject characteristic, compliance, and culture data were processed using SPSS (version 22.0; IBM Corporation, Armonk, NY, USA). Descriptive statistics, such as means [95% confidence intervals (CI)] and proportion distributions were calculated. Differences between groups were tested with parametric or non-parametric tests depending on the data measurement and distribution levels. P < 0.05 was considered statistically significant.

Sequences with a minimum length of 300 base pairs after primer sequence removal, correct barcode sequences, a maximum of 1 incorrect base pair in the primer sequences, and compliance with the default quality criteria for homopolymers and quality scores in the Quantitative Insights into Microbial Ecology (QIIME) [36] software package (version 1.8.0) were retained for analysis. Any sequence beyond the reversed primer were removed. Sequences beginning with the reverse primer were reverse complemented. Sequences were then clustered into operational taxonomic units (OTUs) at a sequence similarity of 97% in the 16S rRNA chimera-checked Greengene database (dated May 2013) [37] using USEARCH [38]. OTUs with a single sequence were removed. One representative sequence per remaining OTU was taxonomically assigned as a named or unnamed cultivable species or uncultivable phylotype at ≥98.5% identity using HOMD 16S rRNA RefSeq, version 12.0 (http://www.homd.org) [28].

Rarefaction curves were calculated to compare microbial richness [39]. Principal coordinate analysis (PCoA) was applied to evaluate the phylogenetic beta diversity [40] of the bacterial profiles at different time points. Multivariate partial least-square analysis (PLS) modeling (SIMCA P+, version 12.0; Umetrics AB, Umeå, Sweden) was performed to search if microbiota structures were related to L. reuteri consumption (y-variables) [41,42]. Tested models included those with pyrosequencing data only and those to which lactobacilli and streptococci culture and PCR data had been added. Variables were autoscaled to unit variance and cross-validated Y predictions were calculated. Subject clustering was displayed in score loading plots, and the importance of each x-variable was displayed in loading plots. Variables, where the 95% CI of the PLS correlation coefficient did not inlude zero were considered statistically significant [42]. The Q2 value yielded the capacities of the x-variables to predict the outcome (test or placebo group allocation). Univariate analyses of single taxa were not applied because of the combination of small groups and a high number of repeated tests.