L. plantarum WCFS1 enhances Treg frequencies by activating DCs even in absence of sampling of bacteria in the Peyer Patches

Probiotics such as L. plantarum WCFS1 can modulate immune responses in healthy subjects but how this occurs is still largely unknown. Immune-sampling in the Peyer Patches has been suggested to be one of the mechanisms. Here we studied the systemic and intestinal immune effects in combination with a trafficking study through the intestine of a well-established immunomodulating probiotic, i.e. L. plantarum WCFS1. We demonstrate that not more than 2–3 bacteria were sampled and in many animals not any bacterium could be found in the PP. Despite this, L. plantarum was associated with a strong increase in infiltration of regulatory CD103+ DCs and generation of regulatory T cells in the spleen. Also, a reduced splenic T helper cell cytokine response was observed after ex vivo restimulation. L. plantarum enhanced Treg cells and attenuated the T helper 2 response in healthy mice. We demonstrate that, in healthy mice, immune sampling is a rare phenomenon and not required for immunomodulation. Also in absence of any sampling immune activation was found illustrating that host-microbe interaction on the Peyer Patches was enough to induce immunomodulation of DCs and T-cells.

In vivo bioluminescence imaging. Bioluminescence imaging was performed using a multimodal IVIS Lumina XR (Caliper, PerkinElmer) as described before 19 , which consists of a cooled charge-coupled-device camera mounted on a light-tight specimen chamber. Prior to bioluminescent imaging, mice were anesthetized with 2% isoflurane. d-Luciferin potassium salt (Caliper, PerkinElmer) at 30 mg/ml was then administered to animals inoculated with CBRluc-expressing strains by intragastric inoculation (200 μl/mouse). Anesthetised mice (induction with isoflurane) were placed into the camera chamber of the IVIS, where a controlled flow of 1.5% isoflurane in air was administered through a nose cone via a gas anesthesia system. A grayscale reference image under low illumination was taken as an overlay prior to quantification of emitted photons over 1 s to 5 min, depending on signal intensity and using the software Living Image (Caliper, PerkinElmer). For anatomical localization, a pseudocolor image representing light intensity (blue, least intense, to red, most intense) was generated using the Living Image software and superimposed over the grayscale reference image. For each individual mouse, there was only one ROI corresponding to the mouse digestive tract, and this ROI was determined manually. Bioluminescence was quantified using the Living Image software (given as p/s).
Cell isolation and restimulation. Following sacrifice of the mice, spleens or PPs were removed. Single cell suspensions were made by mechanical disruption of the tissue between two glass slides in 1 mL of ice cold RPMI containing 10% heat inactivated fetal calf serum (FCS). Subsequently a cell strainer was used to remove remaining clumps. The cells were washed, counted, and used for staining.
Part of the cells of the spleen were stimulated, the remaining cells were left unstimulated. An amount of 7 × 10 6 spleen cells were stimulated in RPMI with 10% FCS containing 40 nM Phorbol 12-myristate 13-acetate (PMA) (Sigma Aldrich) and 2 nM calcium ionophore (Ca 2+ ) (Sigma Aldrich). Monensin (3 µM) (Sigma Aldrich) was added to allow cytokine accumulation in the cellular cytoplasm. Cells were stimulated for four hours at 37 °C, after which they were washed twice in ice cold PBS containing 2% heat inactivated FCS (FACS buffer), and used for staining.
Dead cells were removed from the cell suspensions. For this, 1 mL of cell suspension was loaded on 1 mL of 1-step Polymorphs (Accurate Chemical and Scientific Corporation, Westbury, NY) with a density of 1.068 ± 0.001 g/ml, and centrifuged for 30 minutes at 300 × g at 4 °C. The interface was washed twice in ice cold FACS buffer and used for staining. After density gradient centrifugation, more than 90% of the cells was vital, which was confirmed by propidium iodide staining.
In short, 1 × 10 6 cells were incubated in FACS buffer containing 10% normal mouse serum for 30 minutes to prevent non-specific antibody staining. Subsequently, the cells were incubated with a cocktail of primary antibodies for 30 minutes, in the dark, after which the cells were washed in ice cold FACS buffer twice. Tubes stained for T cells were subsequently fixed in ice cold 1 × FACS Lysing solution (BD Biosciences) for 30 minutes in the dark, washed twice in ice cold FACS buffer and resuspended in the same buffer until analysis. Also, the tubes for intracellular cytokine staining were fixed in ice cold 1 × FACS Lysing solution, for 30 minutes in the dark and washed twice in ice cold 1 × permeabilisation buffer (eBioscience). Subsequently the cells were incubated with the intracellular antibody cocktails containing 2% normal rat serum in permeabilisation buffer for 30 minutes in the dark. Then cells were washed twice in ice cold permeabilisation buffer and resuspended in FACS buffer until analysis. Tubes for DC-staining were washed twice in ice cold FACS buffer, after which cells were incubated with the secondary step for 30 minutes in the dark. Subsequently, cells were fixed in ice cold 1 × FACS Lysing solution for 30 minutes in the dark followed by washing twice in ice cold FACS buffer and resuspended in FACS buffer until analysis.
Flow cytometry. During flow cytometry, at least 5 × 10 5 cells were analyzed. Flow Cytometry was performed using the LSR II Flow Cytometer system (BD Pharmingen), using FACS Diva software. Analysis was performed using the FlowJo 7.6.2 software. Lymphocytes were gated in the forward side scatter plot and the frequency of CD3 + T cells was determined. Within the T cell population, the frequency of CD8 + T cells and CD4 + T cells was determined. Within both the CD4 and CD8 T cell population the isotype controls for CD69, CD25, or the cytokines, were used to set the gate to 99% negative cells. This gate was then copied to the sample stained for CD69, CD25, or cytokines and the frequency of positive cells was determined. Further, within the CD4 T cell population the FoxP3 isotype control was used to set the gate to 99% negative cells. This gate was copied to the sample stained for FoxP3 and the frequency of positive cells was determined (Fig. 1). DCs were gated in the forward side scatter plot, based on size and granularity, and the frequency of MHC II + CD11c + cells was determined. CD19 + B-cells were excluded from analysis. Within the DC populations CD80, CD86, or CD103 isotype controls were used to set the gate to 99% negative cells. This gate was copied to the sample stained for CD80, CD86, and CD103 and the frequency of positive cells was determined ( Fig. 2A).
Definition of immune cell populations. In this study, different immune cell populations are defined based on the expression of specific cell surface markers. Splenic and mesenteric regulatory T cells (Treg) are defined based on the expression of CD25 and the transcription factor FoxP3. For this, FoxP3 + cells were gated within the CD4 T cell population and the expression of CD25 was confirmed. All CD4 + FoxP3 + cells consistently demonstrated CD25 expression. Results are depicted as the frequency of CD25 + FoxP3 + cells within the total CD4 T cell population (CD4 + CD3 + cells). Activated T cells are defined as the frequency of CD25 + FoxP3 − or CD69 + cells within the CD4 + CD3 + (T helper cells; Th) or CD8 + CD3 + (cytotoxic T cells; CTLs) population. Effector cell populations are defined as the frequency of IFN-γ + IL5 − , IFN-γ − IL5 + , IL10 + , and IL17 + IL22 + cells within the CD4 + CD3 + or CD8 + CD3 + cell population. DCs were defined as CD11c + MHC II + cells. Regulatory, intestinal DCs are depicted as the frequency of CD103 + cells within the CD11c + MHC II + population. Also, the frequency of activated DCs was determined and depicted as the frequency of CD80 + or CD86 + cells within the CD11c + MHC II + cell population.
Immunohistochemistry staining. To stain bioluminescent L. plantarum in PPs, PPs were frozen in precooled iso-propane, sectioned at 5 mm, and processed for immuno-histochemical staining. As primary antibody, we applied Anti-Firefly Luciferase antibody ab181640 (Abcam) in a 1:250 dilution. As second antibody, we applied FITC-conjugated rabbit anti-goat antibody (1:100, Dako) and as third biotinylated swine anti-rabbit antibody (1:100, Dako). Tissue slides of 2 Å~ were then washed for 10 min with PBS and incubated with DAPI solution (30 nM) for 5 min to visualize nuclei of the epithelial cells. After washing with PBS, samples were mounted with mounting medium. Slices were studied with a SP8 Leica confocal microscope (Leica Microsystems, Son, the Netherlands).
Statistics. All data are expressed as the mean ± standard error of the mean (SEM). Normal distribution of the data-sets was confirmed by the Kolmogorov-Smirnov test. The two-sided Students t-test was used to determine changes in immune cell populations after probiotic treatment in vivo. P-values < 0.05 were considered statistically significant.

Figure 1.
Lymphocytes were gated in the forward side scatter plot and the frequency of CD3 + T cells was determined. Within the T cell population, the frequency of CD8 + T cells and CD4 + T cells was determined. Within both the CD4 and CD8 T cell population the isotype controls for CD69, CD25, or the cytokines, were used to set the gate to 99% negative cells. This gate was then copied to the sample stained for CD69, CD25, or cytokines and the frequency of positive cells was determined. Further, within the CD4 T cell population the FoxP3 isotype control was used to set the gate to 99% negative cells. This gate was copied to the sample stained for FoxP3 and the frequency of positive cells was determined.

Results
Bioluminescent L. plantarum trafficking through the GI tract of healthy mice. Intra-gastric gavage of 1-2 × 10 8 CFU bacteria was performed and followed by IVIS. L. plantarum had a transit time of two hours ( Fig. 3A,B). After this period, we did not detect any bioluminescent signal. After the fifth day, the intestine was removed and studied again by IVIS. In the isolated intestines, without interference of the skin, we observed no accumulation of the bacteria in immune sampling area, around PPs areas (Fig. 3C). Next, we excised the PPs to study uptake of the bacteria in this immune sampling organ. In only two of the eight animals we observed remnants or intact bacteria in the PPs (Fig. 4). All PPs were completely sliced and stained but we only found 2-4 bacteria in the two animals and nothing in the remaining 6 animals.

L. plantarum treatment affects the frequency of regulatory T cells and T-cell polarization in the spleen.
To determine whether, despite the absence of true sampling of bacteria in the PP, immunomodulation occurred, in a follow-up experiment we studied changes in T-cell frequency in the spleen in mice following L. plantarum WCFS1 administration (doses of 1-2 × 10 8 CFU). L. plantarum WCFS1 induced (n = 6 mice) increased frequencies of Treg cells in the spleen as compared to medium treated animals. L. plantarum WCFS1 had no effect on the frequency of pro-inflammatory activated T cells such as CD25 + CD69 + FoxP3 − and CD69 + cells in the CD4 + (T helper; Th) and CD69+ CD8 + CTL populations (Fig. 5).
In order to determine the consequences of L. plantarum WCFS1 administration on T cell responses, we performed another series of experiments. Cells from the spleen were restimulated with PMA/Ca 2+ ex vivo, after which the IFN-γ, IL5, IL10, and IL17 response of CD4 + and CD8 + T cells was determined.
The L. plantarum WCFS1 treatment had an attenuating effect on Th2 responses (IL5 producing CD4+ T cells), and demonstrated a trend towards increased IL10 + Th cell frequencies in the spleen (p < 0.1). Other polarized CD4 Th subsets were not affected (Fig. 6). L. plantarum had a more pronounced effect on CD8+ T cells. A 2-fold increase in frequencies of IFN-γ producing cells was observed in the CD8+ T cells as well as an increase in IL10 producing CD8+ T cells and a 6-fold increase in IL17-producing CD8+ T cells. L. plantarum had no effect on IL5-producing CD8+ T cells (Fig. 7).

L. plantarum does induce dendritic cell activation in the Peyer Patches even without sampling of the bacteria.
As immunomodulation was observed in the spleen, even in mice where no bacteria were found in the PPs, we questioned whether despite absence of immune sampling some immune upregulation had occurred in the PPs. To this end, we isolated PPs of mice treated for 5 days with L. plantarum and determined the frequency of CD11c+ MHC II+ DCs, CD103+ DCs, CD80+ DCs, or CD86+ DCs. These DCs are the first immune cells to respond to transcytosed antigens in the dome area, which is just below the follicular epithelium 23 . The percentage of CD103+ intestinal DCs was increased, while the percentage of CD103-intestinal DCs was decreased although did this not reach statistical significance (Fig. 2B). Although no effect of probiotic treatment was observed on % CD80+ DCs in the PP (Fig. 2C), the treatment with L. plantarum WCFS1 did increase the activation status of the DCs in the PPs as demonstrated by increased frequencies of CD86+ DCs (Fig. 2D).

Discussion
Understanding the mechanistic basis for host-bacteria interactions that regulate intestinal immune processes is crucial for the development of novel therapeutic strategies with probiotics 3 . Although many studies demonstrate beneficial immune responses induced by probiotics, a key question that still needs to be addressed is how they interact with the gut immune system 23,24 . Current theories can be divided into those that propose an active immune sampling, mediated by M cells in the PPs, modulating macrophages and DCs beneath the epithelium 25 , or those mechanisms in which specific DCs in the lamina propria or PP take up probiotic antigens 6,7,26-28 . DCs have also been proposed to sample translocated bacteria that reach the lamina propria 29 . The current study was designed to evaluate whether sampling of probiotic bacteria in the PP is an actual part of the mechanism by which probiotics induce immune changes in the mucosa or that the second proposed mechanism, i.e. sensing by DCs without sampling, is responsible for the immune effects. Our data suggest that true sampling by PP M-cells of L. plantarum is a rare phenomenon but that DCs in PPs are activated, suggesting that sensing of DCs with PRRs 3 without active sampling is responsible for immune effect observed in the mice.
Neutra et al. 30 and Hamada et al. 31 suggested that sampling of commensal bacteria is likely to occur through M-cells of the PPs and lymphoid follicles. Surprisingly, in their studies actual proof for active large-scale sampling was lacking. Our study demonstrates that sampling of the probiotic strain L. plantarum WCFS1, which is also a commensal, does occur in the PP, but is rather rare. Our observation that immune cell populations changed Figure 4. Uptake of the bacteria in the PPs. L. plantarum-CBRluc (1-2 × 10 8 CFU) was inoculated intragastrically into mice. After 5 days, mice were sacrificed and PPs were excised. All PPs were completely sliced and stained. PPs were frozen in precooled iso-propane, sectioned at 5 mm, and processed for immunohistochemical staining. As primary antibody, we applied Anti-Firefly Luciferase antibody ab181640 (Abcam) in a 1:250 dilution. As second antibody, we applied FITC-conjugated rabbit anti-goat antibody (1:100, Dako) and as third biotinylated swine anti-rabbit antibody (1:100, Dako). Representative image showing that in only two of the eight animals we observed remnants or intact bacteria in the PPs. L. plantarum WCFS1 induced increased frequencies of regulatory T cells in the spleen while had no effect on the frequency of pro-inflammatory activated T cells. Regulatory T cell frequencies in the spleen (N = 6) following oral treatment with medium (white bars), or L. plantarum WCFS1 (black bars). Regulatory T cell frequencies are depicted as the frequency of CD25 + FoxP3 + cells within the CD4 T cell compartment, CD69 + cells within the CD4 T cell compartment (C&D), and CD69 + cells within the CD8 T cell compartment. Results are depicted as the mean ± standard error of the mean (SEM). Statistical significance was calculated using the Students t-test. *Represents P-values < 0.05.
SCIeNTIFIC RepoRts | (2018) 8:1785 | DOI:10.1038/s41598-018-20243-1 despite rare sampling in the PP corroborates the findings of Macpherson et al. 32,33 demonstrating that only tiny proportions (approximately 0·0001%) of a challenge dose of intestinal bacteria actually reach DCs either in the villi or the PPs but are efficaciously inducing mucosal immune responses. As a proposed mechanism Macpherson et al. suggested that the DCs in the PPs and isolated lymphoid follicles promiscuously sample intestinal luminal commensals, whereas lamina propria DCs sample those bacteria located within the surface mucus layer 33 . However, in our experiments sensing via DCs in the PPs is by far the most probable mechanism of immune modulations by this strain as CD11c+ MHC II + DCs isolated from the PPs were increased as well the activation status as illustrated by enhanced CD86 expression. The CD103+ DCs that migrate to the mesenteric lymph nodes to generate T regs were decreased although this did not reach statistical significance. Macpherson et al. observed that commensals do not penetrate further than the mesenteric lymph nodes. The few bacteria that did penetrate were phagocytosed and eliminated to avoid inflammation and prevent destruction of their luminal habitat 32,33 .
To the best of our knowledge, this is the first report combining probiotics induced immune responses with a localization study in the intestine of a luciferase-expressing Lactobacillus, L. plantarum WCFS1. To do so we applied a recently developed engineered L. plantarum WCFS1 with reporter genes-based labelling system 19 . Although the development of such a system was far from easy 19,20 , it is a very instrumental system to study immune sampling, persistence and trafficking of the bacteria through the gastrointestinal tract 34 . This system has been helpful to demonstrate not only that immune sampling is a rare phenomenon but also that persistence of the probiotic is minimal, if not absent. Within 2 hours the strain can pass though the mice intestine and reach the cecum/colon. Similar kinetic results have been shown by Daniel et al. 19 and others 35,36 . We also found that the cecum and colon were the predominant sites for persistence of L. plantarum in mice 19,35 . Interestingly, by using transcript profiles of L. plantarum WCFS1 residing in the ceca of germ-free mice, Marco et al. 37 demonstrated that wild-type L. plantarum modifies its gene expression in ways designed to minimize the levels of D-alanylated lipoteichoic acids (LTA) present on the cell surface in vivo, and therefore limit its exposure to components of the host's innate and/or adaptive immune system.
It is assumed that the majority of lactobacilli are passengers in the GI tract and that rarely more than 1% of the total number of bacteria colonize 38 . Despite low colonization rates lactobacilli are able to exert pronounced effects on the physiological and immunological systems of the host 39 . This also applies to L. plantarum WCFS1 Figure 6. L. plantarum WCFS1 treatment had an attenuating effect on Th2 responses. Polarized CD4 T cell frequencies in the spleen (N = 6) following oral treatment with medium (white bars), or L. plantarum WCFS1 (black bars). Polarized CD4 T cell frequencies are depicted as the frequency of IFNγ + cells within the CD4 T cell compartment, IL5 + cells within the CD4 T cell compartment, IL10 + cells within the CD4 T cell compartment, and IL17 + cells within the CD4 T cell compartment. Results are depicted as the mean ± standard error of the mean (SEM). Statistical significance was calculated using the Students t-test. *Represents P-values < 0.05. where after 5 days of administration, despite absence of colonization, the effects of the probiotic were found to be profound, with many intestinal and systemic effects, including enhanced T reg numbers. Our data suggest that just contact of the luminal bacteria with the DCs is enough to induce these effects 40,41 .
Our data also demonstrate that L. plantarum induced immune effect remote from the mucosal to the systemic circulation. In the spleen, i.e. the site of measurement for systemic effects in mice, L. plantarum induced a strong increase in infiltration of regulatory CD103 + DCs and increased the number of regulatory T cells 17 . These CD103 + DCs are probably not only the DCs that migrated from the PP but also CD103 + DCs that patrol underneath and between enterocytes, while extending dendrites toward the lumen, likely using specific tight junction proteins to penetrate between the epithelium 28 . These intraepithelial CD103+ DCs sample bacterial antigens for presentation in the mesenteric lymph nodes and the spleen. In addition, the observed changes were accompanied by a reduced splenic Th cell cytokine response, suggesting the promotion of a regulatory systemic phenotype in addition to the observed enhanced Treg numbers.
The observation that L. plantarum enhances Treg cells and attenuates the Th2 responsiveness in the systemic circulation suggests that this strain might prevent or reduce the frequency of typical Th2-mediated disorders such as asthma or allergy 42 . L. plantarum induces high levels of IL-12 and IFNγ suppressing Th2 differentiation 42,43 . The present results corroborate our previous findings that L. plantarum WCFS1 consumption directly increased the frequency of T reg cells, while decreasing the responsiveness of Th2 cells and increasing the responsiveness of the CTL compartment 17 . This demonstrates that L. plantarum not only stimulates skewing towards immune regulation, but can also directly improve the responsiveness of cytotoxic T lymphocytes by leaving the Th1 compartment unaltered. Also, several reports have indirectly demonstrated improved CTL responsiveness following probiotic treatment by demonstrating improved immune responses towards viral infections in vivo [44][45][46][47][48] .
Overall, our results highlight the fact that in absence of any immune sampling as observed in the majority of mice sensing of probiotics and DC activation still occurred and was enough to induce immunomodulation. With this study, we demonstrate that L. plantarum WCFS1 induce Treg responses at the site of interaction 49 , but also profoundly affects the systemic immune system by skewing it to a more regulatory phenotype.