Distinctive probiotic features share common TLR2‐dependent signalling in intestinal epithelial cells

Abstract The underlying mechanisms of probiotics and postbiotics are not well understood, but it is known that both affect the adaptive and innate immune responses. In addition, there is a growing concept that some probiotic strains have common core mechanisms that provide certain health benefits. Here, we aimed to elucidate the signalization of the probiotic bacterial strains Lactobacillus paragasseri K7, Limosilactobacillus fermentum L930BB, Bifidobacterium animalis subsp. animalis IM386 and Lactiplantibacillus plantarum WCFS1. We showed in in vitro experiments that the tested probiotics exhibit common TLR2‐ and TLR10‐dependent downstream signalling cascades involving inhibition of NF‐κB signal transduction. Under inflammatory conditions, the probiotics activated phosphatidylinositol 3‐kinase (PI3K)/Akt anti‐apoptotic pathways and protein kinase C (PKC)‐dependent pathways, which led to regulation of the actin cytoskeleton and tight junctions. These pathways contribute to the regeneration of the intestinal epithelium and modulation of the mucosal immune system, which, together with the inhibition of canonical TLR signalling, promote general immune tolerance. With this study we identified shared probiotic mechanisms and were the first to pinpoint the role of anti‐inflammatory probiotic signalling through TLR10.

responses involving both innate and adaptive immunity (Allaire et al., 2018). Over the years, research studies have shown that gut microbiota has an important impact on intestinal homeostasis and the regulation of the intestinal epithelial barrier integrity. Perturbations in the composition and function of gut microbiota have been associated with chronic diseases ranging from irritable bowel disease and metabolic disorders to neurological, cardiovascular and respiratory conditions (Tlaskalová-Hogenová et al., 2011). New preventive and therapeutic approaches now include changing the composition of the gut microbiota with prebiotics, probiotics, postbiotics, reconstituting the bacterial population by faecal transplantation or adding antimicrobial agents to the diet (Shen et al., 2018).
The impact of probiotics is not limited to living bacteria but also includes the regulation of cellular homeostasis by microbial DNA, soluble proteins, cell wall components and metabolites (Adams, 2010;Aguilar-Toalá et al., 2018). These microbe-associated molecular patterns (MAMP) are recognised by the family of innate immune pattern recognition receptors (PRRs), such as Toll-like receptors (TLRs) (Abreu, 2010). TLR signalling influences innate and adaptive immune regulation, which promotes pathogen clearance on the one hand and host-microbe symbiosis on the other (Spiljar, Merkler, & Trajkovski, 2017). However, TLRs are not only expressed in immune cells, such as antigen-presenting cells (APCs), but also in intestinal epithelial cells (IEC; Price et al., 2018). On the apical surface of IECs, TLR2 recognises microbial membrane structures such as lipoteichoic acid, peptidoglycan and lipopeptides and is therefore involved in the detection of probiotics, which upon recognition trigger mitogenactivated protein kinase (MAPK), nuclear factor kappa B (NF-κB) and phosphatidylinositol 3-kinase (PI3K) signalling cascades (Jiménez-Dalmaroni, Gerswhin, & Adamopoulos, 2016;Maier, Anderson, Altermann, & Roy, 2018). TLR2 forms heterodimers with TLR1 and TLR6; however, its signalling cascades are also associated with TLR10 (Guan et al., 2010;Jin et al., 2007;Kang et al., 2009). According to recent studies, TLR10 is a negative regulator of TLR signalling and differs from other TLRs in the absence of a classical downstream signalling cascade leading to transcriptional activation of NF-κB and proinflammatory cytokine production (Oosting et al., 2014). By suppressing the conventional MAPK signalling pathways, TLR10 thus plays an important role in immune tolerance (Jiang, Li, Hess, Guan, & Tapping, 2016). Because of its inhibitory effect on TLR1/TLR2 and TLR2/TLR6 signalling, TLR10 can regulate the immune response to a broad spectrum of microbial antigens.
Due to the difficulties in approving probiotic efficacy in clinical trials, current research is more focused on identifying probiotic mechanisms and host cell signalling pathways. Despite intensive research, the mechanisms behind cell signalling are not yet well understood. Our previous study on DSS-induced colitic mouse model shows that a combination of two probiotic strains helped in the regeneration of the intestinal epithelium during pathogenesis via pathways that are known downstream targets of TLR2 (Paveljšek et al., 2018). In the present study, we aimed to reveal the TLR2-mediated signalling pathways that are probably responsible for the regulation of the intestinal epithelial barrier. By using different in vitro setups of IECs challenged with probiotics, we investigated these signalling cascades triggered by four probiotic strains Lactobacillus paragasseri K7 isolated from the faeces of a 7-day-old infant (Treven, Trmči c, Matijaši c, & Rogelj, 2014), Limosilactobacillus fermentum L930BB and Bifidobacterium animalis subsp. animalis IM386 isolated from the human intestinal mucosa (Čitar et al., 2015;Paveljšek et al., 2018), and Lactiplantibacillus plantarum WCFS1, human saliva isolate (Karczewski et al., 2010;Zheng et al., 2020). Our study confirmed former research findings that showed that activation of TLRs by commensal microorganisms is crucial for protection against intestinal damage (Bach, 2018). In addition, it also opened a new direction in mechanistic studies on the maintenance of intestinal homeostasis by probiotics.

| Signalling of probiotic bacterial strains via TLR2-NF-κB
We used a dual luciferase reporter assay in the HEK293 cell line, transfected with TLRs and stimulated with the probiotic strains L. paragasseri K7 (K7), L. fermentum L930BB (L930BB), B. animalis subsp. animalis IM386 (IM386) and L. plantarum WCFS1 (WCFS1) at different concentrations. We found that all used probiotic strains activate the transcription factor NF-κB through TLR1/TLR2 and through TLR2/TLR6 (Figure 1). IM386 triggered the highest activation of NF-κB compared to other strains. The probiotic activation of NF-κB through TLR2/TLR6 was higher than activation through TLR1/TLR2. Since TLR2 recognises lipoproteins, usually components of the bacterial surface, we were interested in whether the supernatant of bacterial growth medium and heat-killed bacteria also trigger NF-κB signalling cascades. We discovered that supernatant and heat-killed bacteria trigger NF-κB activation through TLR2 heterodimers in a similar concentration-dependent manner as live bacteria. However, in the case of the heat-killed strain WCFS1 and its supernatant, the activation of NF-κB was lower compared to the activation with live bacteria ( Figure S1). To determine the components that trigger the activation of NF-κB, we further fractionated the supernatant with 10 kDa filters.
We found that the components, responsible for TLR2-induced NF-κB activation, have a molecular weight of more than 10 kDa ( Figure 2).

| Probiotic inhibition of conventional TLR-NF-κB signalling pathway via TLR10 and stimulation of anti-apoptotic pathways
To test how probiotic strains affect TLR10 signalling cascades, we performed a dual luciferase reporter assay in the HEK293 cell line with expressed TLR1/TLR2, TLR2/TLR6 and increasing levels of expressed TLR10. We discovered that probiotic strains reduced the activation of NF-κB in the presence of TLR10 via both TLR1/ TLR2 and TLR2/TLR6 (Figure 3). To test the possible involvement of PI3K/Akt pathway in the inhibitory effect of TLR10, we used a nonoverexpression system on the IEC line Caco-2. After addition of probiotic strains, phosphorylation of Akt, which is a measure of its activation by PI3K, was observed. In addition, the use of TLR2 and PI3K inhibitors reduced phosphorylation of Akt ( Figure 4).
Since one of the most important signalling cascades via PI3K/Akt leads to the inhibition of apoptotic pathways, we tested whether the selected probiotic strains can influence cytokine-induced apoptosis and necrosis. We used the IEC line HT-29, which also expresses surface TLRs and shows a better inflammatory response to cytokines than the Caco-2 cell line. The extent of apoptosis and necrosis was determined by analysing the live/dead HT-29 cell populations with flow cytometry. The addition of probiotic strains to HT-29 cell line reduced necrosis and late apoptosis caused by the addition of inflammatory cytokines (Figures 5 and S2). Furthermore, the TLR2 synthetic ligands and the probiotic strains had a similar protective effect on cell necrosis, proving TLR2-mediated mechanism.

| Probiotic impact on the intestinal epithelial barrier
We first investigated the possible protective effect of probiotic strains on the permeability of the H 2 O 2 -disrupted Caco-2 cell monolayer. We monitored the transition of FITC-dextran from the apical to the basolateral compartment of semi-permeable membranes. The addition of selected probiotic strains and the synthetic TLR2 ligand-Pam3Cys-SK4 (PAM3) to the Caco-2 cell line cultured on permeable supports did not alter the cell monolayer permeability ( Figure 6). However, their addition to the H 2 O 2 -disrupted Caco-2 cells led to a reduction in permeability. Given that probiotics had the same protective effect as PAM3, we next investigated whether the probiotic influence on epithelial permeability is related to the PKC signalling pathway, a known TLR2-dependent pathway that impacts on the paracellular sealing complex. We found that the addition of PKC inhibitor Gӧ6983 (PKCinh) reduced the protective effect of the probiotic strains, which  (Engle, Goetz, & Alpers, 1998), these models also offer advantages, including more controlled conditions and robustness. However, for more thorough conclusions, it is still necessary to start from more complex models that better reflect Besides the poorly understood biological function of TLR10, which in contrast to other TLRs mainly has an inhibitory effect, there are also no known ligands for this receptor (Jiang et al., 2016). We showed that the co-expression of TLR10 with TLR1/TLR2 as well as with TLR2/TLR6 had an inhibitory effect on NF-κB-mediated response to probiotics, which has not been shown in the literature before. The signalling of probiotic strains through TLR10 may therefore be considered as an additional immune-regulatory effect that promotes remission in autoimmune intestinal diseases. In addition, the In IECs, the absence of TLR2 causes inadequate PI3K/Akt signalling and disruption of the intracellular anchoring complex. This leads to increased apoptosis, detachment of colonocytes and consequently to a greater permeability of the epithelial barrier, resulting in mucosal inflammation (Cario, Gerken, & Podolsky, 2007). Indeed, we showed that probiotics induce signalling through PI3K/Akt and reduce necrosis in inflammatory conditions and thereby help to maintain the barrier integrity.
We also examined the probiotic influence in stress conditions on activation of PKC-another TLR2-mediated downstream signalling pathway that can prevent internalisation or translocation of the cytosolic tight junction protein ZO-1 from the plasma membrane and is also important for the epithelial barrier integrity (Corr et al., 2014;Gu et al., 2016). IECs express a number of PKCs that regulate the phosphorylation state and localization of several tight junction proteins (Zyrek et al., 2007). According to Cario, Gerken, and Podolsky (2004), stimulation of the TLR2 pathway leads to activation of PKCα and PKCδ, and subsequently to redistribution of ZO-1 and improvement of cell monolayer integrity. In our study, both probiotic strains and TLR2 ligands triggered the activation of specific PKC isoforms, which had an influence on the localization of ZO-1 and consequently on the reduction of cell monolayer permeability, suggesting and confirming that TLR2 signalization is responsible for this downstream effect.
Interestingly, the same mechanism of action and the same TLR2 signalling cascades were observed for all selected probiotics. This can be explained by the fact that the MAMPs that trigger signalization have a similar basic structure and are conserved among a larger group of probiotics (Sanders, Benson, Lebeer, Merenstein, & Klaenhammer, 2018). It is also important to keep in mind that the bacterial cell surface is a dynamic property and the expression of bacterial surface components may vary due to the host environment (Lebeer, Vanderleyden, & De Keersmaecker, 2010). Furthermore, recent research demonstrates that some of the molecular complexes responsible for immuno-modulation in host cells may be common to a larger taxonomic group (Lebeer

| Fluorometric analysis of cell monolayer permeability
Caco-2 cells (26th passage) were seeded on 12-well semi-permeable membranes (Transwell, Corning) at a density of 7 × 10 4 cells per well and cultured for 21 days until differentiation with media changes every other day. The designated group of cells was treated with PKC inhibitor Gӧ6983 (PKCinh) (5 μg/mL, Sigma) 1 hr before the experiment. Afterwards, the probiotic strains (10 7 CFU per well) or PAM3 (20 μg/mL) were added to the apical compartment and incubated for 1 hr. After incubation H 2 O 2 (750 μM) was added to both compartments and 3 kDa FITC-dextran (0.2 mg/mL, Thermo Fisher Scientific) was added to the apical compartment. Samples were taken from the basolateral compartment and the fluorescence intensity was measured on a Synergy MX microtiter plate reader (Tecan). The experiment was performed in two biological replicates. Statistical analysis F I G U R E 8 TLR2-and TLR10-mediated probiotic signalling. TLR2-dependent bacterial-cellular interactions presumably help to improve intestinal immune tolerance, reduce necrosis and enhance barrier integrity. Probiotic signalling through TLR10 probably blocks canonical NF-κB cascades and promotes PI3/Akt pathway that reduces necrosis. Probiotics also mediate TLR2-dependent PKC stabilisation of tight junctions and thus most likely contribute to the formation of the entire sealing complex and a stronger intestinal epithelial barrier. Akt, protein kinase B; IEC, intestinal epithelial cell; JAM, junction adhesion molecule; NF-κB, nuclear factor kappa B; PI3K, phosphatidylinositol 3-kinase; PKC, protein kinase C; TLR, Tolllike receptor; ZO-1, zonula occludens-1 was carried out using unpaired Student's t test. A p-value of ≤.05 and ≤.01 was considered statistically significant.

| Confocal microscopy
Caco-2 cells were seeded onto eight-well tissue culture chambers (Ibidi) at a density of 1.78 × 10 4 cells per well and cultured for 21 days with media changes every other day. The designated group of cells was treated with PKCinh (20 μM, Sigma) 1 hr before the experiment.