Investigation of Streptococcus salivarius-mediated inhibition of pneumococcal adherence to pharyngeal epithelial cells

Background Pneumococcal adherence to the nasopharyngeal epithelium is a critical step in colonisation and disease. The probiotic bacterium, Streptococcus salivarius, can inhibit pneumococcal adherence to epithelial cells in vitro. We investigated the mechanism(s) of inhibition using a human pharyngeal epithelial cell line (Detroit 562) following pre-administration of two different strains of S. salivarius. Results Whilst the bacteriocin-encoding megaplasmids of S. salivarius strains K12 and M18 were essential to prevent pneumococcal growth on solid media, they were not required to inhibit pneumococcal adherence. Experiments testing S. salivarius K12 and two pneumococcal isolates (serotypes 19F and 6A) showed that inhibition of 19F may involve S. salivarius-mediated blocking of pneumococcal binding sites: a negative correlation was observed between adherence of K12 and 19F, and no inhibition occurred when K12 was prevented from contacting epithelial cells. K12-mediated inhibition of adherence by 6A may involve additional mechanisms, since no correlation was observed between adherence of K12 and 6A, and K12 could inhibit 6A adherence in the absence of cell contact. Conclusions These results suggest that S. salivarius employs several mechanisms, including blocking pneumococcal binding sites, to reduce pneumococcal adherence to pharyngeal epithelial cells. These findings extend our understanding of how probiotics may inhibit pneumococcal adherence and could assist with the development of novel strategies to prevent pneumococcal colonisation in the future. Electronic supplementary material The online version of this article (doi:10.1186/s12866-016-0843-z) contains supplementary material, which is available to authorized users.


Background
Streptococcus pneumoniae (the pneumococcus) commonly colonises the nasopharynx of healthy humans, especially young children. Carriage is considered a prerequisite for pneumococcal disease and facilitates the transmission of pneumococci throughout communities [1]. Dissemination of pneumococci from the nasopharynx to other body sites can give rise to diseases such as meningitis, sepsis, pneumonia, and otitis media. An estimated 800,000 children under the age of five die from pneumococcal infections each year, with most deaths occurring in low-income countries where carriage rates are especially high [2]. Strategies targeting the reduction of pneumococcal colonisation could potentially reduce this burden of disease.
Current pneumococcal conjugate vaccines (PCVs) induce protection against 10-13 of the most common diseasecausing serotypes via the induction of anti-capsular antibodies. Although PCVs have successfully reduced carriage and disease caused by vaccine serotypes, they are expensive to produce and have led to an increase in colonisation by non-vaccine serotypes (serotype replacement) [3]. In recent years, there has been increased interest in the use of probiotics, which are defined as live microorganisms that can confer a health benefit to the host, to reduce pathogen colonisation and respiratory tract infections [4]. Proposed mechanisms of probiotic action include inhibition of pathogen colonisation via competition for binding, direct inhibition due to the activity of secreted antimicrobial molecules and the induction of immunomodulatory effects in the host [5][6][7][8][9].
Streptococcus salivarius is a member of the respiratory tract microbiota and has been commercially available as an oral probiotic for more than a decade [10]. Small clinical trials have shown that administration of S. salivarius strains K12 and M18 can reduce the occurrence of tonsillitis and otitis media [11] and reduce dental plaque levels in children [12], as well as treat halitosis in adults [13]. Several in vitro studies have found that S. salivarius can prevent the growth of a range of respiratory pathogens, including the pneumococcus, through production of megaplasmid-encoded bacteriocins and bacteriocin-like inhibitory substances (BLIS) [7,[14][15][16]. However, the mechanisms by which they inhibit pathogen adherence in vivo are unknown. We have previously shown that S. salivarius K12 can inhibit pneumococcal adherence to a human epithelial cell line (CCL-23) [17]. Here, we demonstrate that the same phenomenon is observed in Detroit 562 pharyngeal epithelial cells and investigate the inhibitory mechanisms involved, including the role of the S. salivarius megaplasmid. Our results suggest that S. salivarius K12 inhibits pneumococcal adherence by blocking pneumococcal binding sites, although other mechanisms such as direct interference through the action of secreted molecules may also contribute.

Methods
Bacterial strains, cell lines and culture conditions Bacterial isolates are described in Tables 1 and 2. Pneumococcal isolates were selected to represent a range of serotypes and based on their ability to adhere to human epithelial cells. S. salivarius isolates were sourced from Blis Technologies Ltd, New Zealand. All isolates were cultured at 37°C in 5 % CO 2 on horse blood agar (HBA; Thermo Fisher Scientific) plates, in Todd-Hewitt broth (THB; Oxoid), or THB supplemented with 0.5 % (w/v) yeast extract (THY; Becton Dickinson). Deferred antagonism testing was carried out on BaCa (Columbia agar base; Life Technologies Ltd.) plates supplemented with human blood (5 %, v/v) and CaCO 3 (0.1 %, w/v) except where noted.

Deferred antagonism testing
Testing for deferred antagonism was performed as described previously [18]. Briefly, overnight THB cultures of S. salivarius (producer strain) were used to inoculate a 1-cm-wide line down the middle of a human blood agar plate (as described above). Plates were incubated for 18 h at 37°C in 5 % CO 2 , S. salivarius growth was removed, and the plate inverted over a chloroformsoaked cloth for 30 min to kill any remaining bacteria. Residual chloroform was then evaporated by exposure of the agar surface to air for 30 min. Overnight THB cultures of pneumococcal indicator strains were then streaked at

Adherence assays
Adherence assays were performed as described previously [17] with several modifications. D562 cells seeded into 24-well trays were washed twice with Hanks buffered salt solution (HBSS; Gibco®, Life Technologies) before the addition of 500 μl/well RPMI without serum. Bacteria were grown to log phase in THY (S. salivarius: 1.5 h, pneumococci: 3 h) and resuspended in 0.85 % (w/v) NaCl (Merck) to~1.5 × 10 9 CFU/ml. A 10-fold dilution series of S. salivarius was prepared in 0.85 % NaCl and 10 μl aliquots of high (~1.5 × 10 7 CFU/well), medium (~1.5 × 10 6 CFU/well), and low (~1.5 × 10 5 CFU/well) concentrations of bacteria were administered to duplicate wells before centrifugation at 114 × g for 3 min to promote bacterial adherence to the cell monolayer. Plates were incubated at 37°C in 5 % CO 2 for 1 h. Pneumococci were prepared as above and added to all wells at a concentration of~1.5 × 10 6 CFU/well (MOI of 10 pneumococci: 1 D562 cell) 1 h after S. salivarius. Centrifugation was performed as above and plates were incubated for a further 1 h. Cells (and adherent bacteria) were then washed three times with HBSS and harvested after the addition of 200 μl 0.25 % trypsin/EDTA to each well, incubation for 10 min at 37°C in 5 % CO 2, and addition of 800 μl of THY to each well with mechanical disruption. Lysates were stored at −20°C until tested. Pneumococcal adherence (genome copies/ well) was determined by quantitative real-time PCR (qPCR) targeting lytA (see below). The effect of bacterial infection on D562 cells was monitored microscopically and no significant cytopathic effects were observed during the adherence assay conditions. The effect of S. salivarius on pneumococcal adherence was calculated by normalising all treatments against wells containing pneumococci alone. The ability of the pneumococcal isolates to invade cells was measured by incubating cells with media containing 10 μg/ml penicillin and 200 μg/ml gentamicin for 15 min after incubation with pneumococci for 1 h (Additional file 1), confirming the validity of this model to investigate adherence. The adherence of S. salivarius to D562 cells at the time of pneumococcal administration was calculated following incubation of~1.5 × 10 6 CFU/well bacteria for 1 h at 37°C in 5 % CO 2 . Cells were washed three times with HBSS to remove non-adherent bacteria and harvested as above. CFU/well was determined by viable count on HBA plates and expressed as mean % adherence ± standard deviation compared to the inoculum added. We performed preliminary experiments to determine whether heparin, which can inhibit pneumococcal adherence to epithelial cells in vitro by competitively binding to glycosylaminoglycans [19], could also inhibit S. salivarius adherence to D562 cells. Cells were incubated with 100 U/well of heparin (Pfizer) for 1 h before the addition of S. salivarius as above and results expressed as % adherence compared to S. salivarius without heparin. S. salivarius adherence to D562 cells over time was measured by qPCR targeting dex (see below).
We also investigated whether non-adherent S. salivarius had an effect on pneumococcal adherence. To do this, adherence assays were performed as described above, except that wells were washed three times with HBSS to remove unbound S. salivarius and fresh medium was added after S. salivarius pre-administration and before the addition of pneumococci. To determine the inhibitory effect of S. salivarius in the absence of cell contact, S. salivarius was preadministered to D562 cells via permeable transwell inserts with a pore size of 0.4 μm (Corning, In vitro Technologies). To further investigate the mechanisms involved in the inhibition of pneumococcal adherence, preparations of K12 were: a) heat-killed by incubating at 70°C for 30 min with their lack of viability confirmed by the absence of growth on HBA; b) enzymatically treated to remove outer surface proteins and carbohydrates of S. salivarius K12 by (i) resuspending K12 inocula in a final concentration of 5 mg/ml pronase E [20], or (ii) by the addition of 5 μl of 1 mg/ml sodium periodate [21] to K12 inocula prepared in 0.85 % NaCl for 15 min at 37°C in 5 % CO 2 , and c) reducing K12 protein synthesis by resuspending K12 inocula in 2 mg/ml spectinomycin and incubating for 30 min at 37°C in 5 % CO 2 . Following all treatments, K12 inocula were washed twice in 0.85 % NaCl prior to use in adherence assays.

Statistical analysis
Data were analysed using Prism 6.0d (GraphPad Software, Inc.) and Excel (Microsoft). Student's t-test was used to compare pneumococcal adherence in all treatment groups. Spearman's rank test was used to correlate pneumococcal and S. salivarius K12 adherence in adherence assays. A p value of < 0.05 was considered statistically significant for all assays. For all experiments n ≥ 3 unless otherwise stated.

S. salivarius reduces pneumococcal adherence to pharyngeal epithelial cells
Previous studies in our laboratory demonstrated that S. salivarius K12 inhibits pneumococcal adherence to CCL-23 human epithelial cells in vitro, with the strongest effect seen when K12 was added before pneumococci (pre-administration). In this study, we investigated the mechanisms underlying this inhibition using a cell line more relevant to the pneumococcal ecological niche of the nasopharynx, Detroit 562 human pharyngeal epithelial cells (D562), and two isolates representing serotypes that commonly colonise (6A [PMP1043] and 19F [PMP843]). Pneumococcal adherence was measured following pre-administration of commercial probiotic S. salivarius strains K12 or M18 at high, medium and low doses.
The K12 megaplasmid is essential to prevent pneumococcal growth but not adherence The megaplasmids of S. salivarius K12 and M18 encode bacteriocins that prevent the growth of a range of bacterial species on solid media [14,16,23,24], but their roles in inhibition of pneumococcal growth or adherence are unknown. We compared the ability of K12, M18, their megaplasmid-negative derivatives, and a panel of S. salivarius strains to inhibit the growth of 15 pneumococcal isolates by deferred antagonism. S. salivarius strains K12, M18, A234, NR, T18A and Min5 inhibited all of the pneumococcal isolates tested (Table 4). Neither megaplasmid-negative strain (K12 mp− or M18 mp− ) or strain T30A prevented the growth of any of the pneumococcal isolates tested (Table 4).
Since the S. salivarius megaplasmids were essential to prevent the growth of pneumococci, and K12 caused a greater reduction in pneumococcal adherence than M18 (Fig. 1), we next compared the ability of K12 and K12 mp− strains to inhibit pneumococcal adherence to D562 cells. Although the exact copy number of the megaplasmid is Fig. 1 Pneumococcal adherence of serotypes 6A (a) and 19F (b) to pharyngeal epithelial cells following pre-administration of S. salivarius. Approximately 1.4 × 10 6 CFU pneumococci were added to D562 monolayers at an MOI of 11:1. Pneumococcal adherence was determined when incubated with pneumococci alone (Pnc, normalised to 100 %), or pre-incubated for 1 h with S. salivarius K12 or M18 at high (~1.6 × 10 7 CFU), medium (~1.6 × 10 6 CFU), or low (~1.6 × 10 5 CFU) doses. Data are mean + SD; n ≥ 3. * indicates p < 0.05, ** indicates p < 0.001 when compared to Pnc alone (Student's t-test) unknown, preliminary qPCR analysis confirmed that the ratio of K12 chromosomal DNA to megaplasmid DNA remained stable during bacterial growth (1.43, n = 1) and throughout assay conditions (1.17 and 1.20 after 1 and 2 h incubation with D562 monolayer, respectively, n = 1). All doses of K12 mp− significantly inhibited 6A adherence (high: p < 0.001, medium: p < 0.001, low: p = 0.016, Fig. 3a). When compared to K12, there was some evidence that the K12 mp− strain resulted in less inhibition than the K12 strain, but this difference was only statistically significant at the high dose (p = 0.044, Fig. 3a). For the 19F strain, only the high and medium doses of K12 mp− significantly inhibited pneumococcal adherence. We found no differences in the ability of K12 and K12 mp− to adhere to D562 cells (532 ± 86.49 % vs. 511 ± 6.92 %, p = 0.695) or in their adherence to D562 cells over time (Fig. 4, p > 0.05 for all time points).
Cell contact is required for K12 to inhibit adherence by serotype 19F, but not serotype 6A To determine whether contact of S. salivarius with D562 cells was required to inhibit pneumococcal adherence, we performed adherence assays in which K12 was administered using transwell inserts, preventing direct contact with epithelial cells but allowing the passage of secreted molecules. Using this transwell system, no inhibition of 19F adherence was observed at any dose, while 6A adherence was inhibited at the medium (p = 0.049) and low (p < 0.001) doses (Table 3). Preliminary experiments showed that culture supernatants did not inhibit pneumococcal growth in the deferred antagonism test.

Protein synthesis may be required for K12 to inhibit 6A adherence
The above experiments indicated the involvement of multiple mechanisms for S. salivarius inhibiting pneumococcal adherence. We further investigated the molecular mechanism(s) of S. salivarius K12 inhibition of 6A adherence by performing different treatments on the K12 inoculum prior to its use in adherence assays. Denaturing proteins and killing K12 cells (heat-treatment), interrupting protein synthesis (spectinomycin treatment), and removing outer surface carbohydrates (sodium periodate treatment) and proteins (pronase E treatment) all reduced inhibition of 6A adherence compared with untreated K12 (Fig. 5). However, only  High:~1.5 × 10 7 CFU/ml; Medium:~1.5 × 10 6 CFU/ml and Low:~1.5 × 10 5 CFU/ml spectinomycin-treated K12 could no longer significantly inhibit 6A adherence (p = 0.402, Fig. 5).

Discussion
In this study, we found that various bacteriocin-producing strains of S. salivarius can prevent pneumococcal growth on solid media and that the commercial probiotic strains, K12 and M18, inhibit pneumococcal 6A and 19F adherence to a pharyngeal epithelial cell line. Studies investigating the probiotic mechanisms of S. salivarius have largely focused on the role of bacteriocins and BLIS. The~190 kb megaplasmid of S. salivarius K12 harbours genes encoding the bacteriocins salivaricin A2 and B, which prevent the growth of some bacterial pathogens, including Streptococcus pyogenes, in vitro [7,14,15]. Our results showed that the K12 megaplasmid is required to prevent pneumococcal growth in vitro, but is not essential to inhibit pneumococcal adherence. Therefore the primary mechanism of inhibition of pneumococcal adherence in this model does not appear to be mediated by megaplasmid-encoded bacteriocins; although we note that lack of inhibition of 19F adherence by a low dose of K12 mpindicates a possible role for the megaplasmid at this concentration. Instead, our findings indicated that S. salivarius K12 can inhibit pneumococcal adherence by blocking pneumococcal binding sites on D562 cells. We observed a correlation between S. salivarius K12 adherence to D562 cells and inhibition of adherence by pneumococcal isolate PMP843 (19F). We also found that "+" indicates inhibition of pneumococcal growth and "−" indicates no inhibition, n = 2 NT not tested a performed on horse blood agar plates Fig. 3 Adherence of pneumococcal serotypes 6A (a) and 19F (b) to pharyngeal epithelial cells following pre-administration of megaplasmid positive (K12) and negative (K12 mp− ) S. salivarius K12 strains. Approximately 1.5 × 10 6 CFU pneumococci were added to D562 monolayers at an MOI of 10:1. Pneumococcal adherence was determined when incubated with pneumococci alone (Pnc, normalised to 100 %), or pre-incubated for 1 h with S. salivarius K12 or S. salivarius K12 mpat high (~2 × 10 7 CFU); medium (~2 × 10 6 CFU, med); or low (~2 × 10 5 CFU) doses. Data are mean + SD; n ≥ 6. * indicates p < 0.05, ** indicates p < 0.001 when compared to Pnc alone (Student's t test) K12 contact with cells was essential for this inhibition. K12 binding to D562 cells did not require megaplasmidencoded molecules, since a megaplasmid-negative strain of K12 displayed equal adherence. We observed differences in the inhibition of pneumococcal adherence depending on both the strain of S. salivarius used and the pneumococcal isolate tested. This was not unexpected, given that S. salivarius strains vary in their capacity to adhere to epithelial cells and inhibit other pathogens [7,15,16]. The pneumococcus is a diverse pathogen, with different strains and serotypes known to behave differently in vitro and in vivo [25,26]. Overall, S. salivarius K12 displayed stronger inhibition of pneumococcal adherence than the M18 strain, possibly due to its increased capacity to adhere to D562 cells. Interestingly, K12 adherence to D562 cells did not correlate with 6A adherence, suggesting that mechanisms other than S. salivarius-mediated blocking of pneumococcal binding sites on epithelial cells may be involved in the inhibition of 6A adherence. This was further supported by the inhibition of 6A adherence by all doses of M18 despite its low adherence capability. Although a high dose of megaplasmidnegative K12 could still inhibit 6A adherence, the effect was significantly less than that observed with megaplasmidpositive K12, indicating a possible contribution to inhibition by the K12 megaplasmid in this assay. Transwell experiments demonstrated that for 19F, contact of S. salivarius K12 with D562 cells was required to inhibit pneumococcal adherence, consistent with a previous report showing that contact with epithelial cells was needed for S. salivarius K12 to inhibit adherence by S. pyogenes to HEp-2 cells [27]. However, low doses of K12 administered in the absence of cell contact could inhibit adherence of 6A. This could be due to bacteriocins or other secreted products, which may not necessarily be encoded on the megaplasmid. The differences observed using varying doses of probiotic suggest that quorum sensing may play a role in the expression of S. salivarius genes relevant to pneumococcal inhibition.
Further transwell experiments that test the K12 megaplasmid-negative strain may determine whether the molecules involved are chromosomally or plasmid encoded. Preliminary attempts to detect if bacteriocins were secreted through transwell membranes by deferred antagonism were unsuccessful, however this approach may have been insufficiently sensitive.
Overall, the data obtained in this study suggest that the primary mechanism by which S. salivarius K12 inhibits pneumococcal adherence to D562 cells is by blocking pneumococcal attachment to the epithelial cell surface. A number of pneumococcal surface molecules have been shown to play a role in adherence, such as the adhesion molecule PsaA, which binds to cell surface carbohydrates [28]. For S. salivarius, various cell surface molecules including fibrils and fimbriae have been implicated in adherence [29]. As heparin does not block S. salivarius adherence to D562 cells, glycosylaminoglycans do not seem to be a shared cell surface receptor. Since airway epithelial cells derived from carcinomas, such as D562 cells, can display altered expression of surface molecules compared to primary human epithelial cells [30], further investigation of common receptors used by these species should include testing of primary cells. Although inhibition of pneumococcal adherence could be due to competition for receptor-mediated binding, it could also result through a less specific mechanism, such as steric hindrance, or the ability of S. salivarius to mask pneumococcal binding sites. Other potential mechanisms by which S. salivarius may inhibit pneumococcal adherence, such as immune modulation of epithelial cells [31], were not examined in this study. Additionally, the in vitro model lacks the complexity of the nasopharynx, which can harbour a wide range of colonising bacterial species that interact with each other and with epithelial cells. Fig. 4 Time course of S. salivarius adherence to D562 cells. Cells were inoculated with either~1.1 × 10 5 CFU of S. salivarius K12 or~1.9 × 10 5 CFU of S. salivarius K12 mpand the number of adherent bacteria measured over three hours. Median ± IQR for both S. salivarius isolates are depicted (n ≥ 2) Fig. 5 Adherence of pneumococcal serotype 6A to pharyngeal epithelial cells following pre-administration of treated S. salivarius K12. Approximately 1.5 × 10 6 CFU pneumococci were added to D562 monolayers at an MOI of 10:1. Pneumococcal adherence was determined when incubated with pneumococci alone (Pnc, normalised to 100 %), or pre-incubated for 1 h with approximately 1.5 × 10 7 CFU S. salivarius K12 (untreated), or treated with sodium periodate, Pronase E, heat-killed (heat), or spectinomycin. Data are mean + SD; n ≥ 3. * indicates p < 0.05, ** indicates p < 0.001 when compared to Pnc alone (Student's t test) Whilst megaplasmid-encoded bacteriocins are required to inhibit the growth of pneumococci on solid media, our data suggest that it is the adherence of S. salivarius to epithelial cells that primarily blocks pneumococcal adherence in vitro. Therefore, strategies to use S. salivarius as a probiotic in the respiratory tract should optimise the ability of this species to colonise the target tissue. In vitro studies performed in our laboratory, and by others, have shown that the addition of probiotic bacteria to epithelial cells prior to pathogen administration is more effective than attempting to disrupt established pathogen colonisation [17,27]. In clinical studies, high doses of M18 achieve greater colonisation rates and density in the oral cavity compared to lower doses [15]. Clinical studies of S. salivarius K12 as an oral probiotic have involved daily administration of high doses of the bacterium in an effort to achieve long-term colonisation of the oral cavity [32][33][34]. S. salivarius may be better suited as a probiotic for the oral cavity, where it predominates, than for the nasopharynx, the preferred niche of the pneumococcus. Strategies that facilitate S. salivarius adherence to the nasopharyngeal epithelium, such as the nasal spray recently employed by Santagati and colleagues [35], would likely be needed for this probiotic to inhibit pneumococcal colonisation in vivo.

Conclusions
Bacteriocin-encoding megaplasmids of S. salivarius strains K12 and M18 were essential to prevent pneumococcal growth on solid media but were not required to inhibit pneumococcal adherence to pharyngeal epithelial cells. Our results suggest that S. salivarius K12 employs several mechanisms, including blocking pneumococcal binding sites, to reduce pneumococcal adherence to pharyngeal epithelial cells. Further research is needed to identify the specific molecules involved. These findings contribute to our understanding of how probiotics may inhibit pneumococcal adherence and could assist with the development of novel strategies to prevent pneumococcal colonisation in the future.

Additional files
Additional file 1: Supporting Information S1. Raw data used to generate figures and tables. (XLSX 52 kb) Additional file 2: Table S1. Specificity of dex and mp gene targets to detect Streptococcus salivarius. DNA was extracted from pure cultures grown on HBA agar and 0.1 ng DNA of each isolate listed was used as a template for the dex and mp qPCR reactions, as described in the methods section. Mean Ct from duplicate wells is shown.