Dectin-2 mediates phagocytosis of Lactobacillus paracasei KW3110 and IL-10 production by macrophages

Lactic acid bacteria (LAB) are most generally used as probiotics and some strains of LAB are known to have anti-inflammatory effects. A specific strain of lactic acid bacteria, Lactobacillus paracasei KW3110 (KW3110), activates macrophages to produce interleukin-10 (IL-10), an anti-inflammatory cytokine; however, the biological mechanism remains unclear. In this study, we showed that the amount of incorporated KW3110 into a macrophage cell line, RAW 264.7, was higher than other genetically related strains using fluorescence microscopy. RNA-seq analysis indicated that treatment of macrophages with KW3110 induced Dectin-2 gene expression, which is a pattern recognition receptor, recognizing α-mannose. In addition, antibody treatment and knock down of Dectin-2, or factors downstream in the signaling pathway, decreased the amount of incorporated KW3110 and IL-10 production. Substantial lectin array analysis also revealed that KW3110 had higher binding affinities to lectins, which recognize the carbohydrate chains comprised of α-mannose, than two other LAB. In conclusion, KW3110 is readily incorporated into macrophages, leading to IL-10 production. Dectin-2 mediated the phagocytosis of KW3110 into macrophages and this may be involved with the characteristic carbohydrate chains of KW3110.

Lactic acid bacteria (LAB) are widely known to have suppressive effects on inflammatory diseases through the induction of anti-inflammatory cytokines, such as IL-10. We previously demonstrated that L. paracasei KW3110, a specific strain of LAB, significantly induced IL-10 production in macrophages through nucleotide oligomerization domain (NOD) 2 signaling pathway 1 . Many studies have suggested that the Toll-like receptor (TLR) 2 pathway is involved in IL-10 induction by LAB in macrophages; however, our experiments showed that blocking TLR2 using neutralizing antibodies does not affect IL-10 production by KW3110 stimulation. Our results suggest that TLR2 is not crucial for IL-10 production induced by KW3110 2 . Although we have shown that KW3110 has strong anti-inflammatory effects 3 , the molecular mechanism underlying increased IL-10 production remains unclear.
Macrophages play a significant part in the immune response to injury by engulfing and digesting particles, including pathogens or bacteria, in a process called phagocytosis. It is known that macrophages produce IL-10 during phagocytosis of apoptotic cells 4 . During the first step of phagocytosis, the pattern-recognition receptors (PRRs) on the surface of macrophages, such as TLRs and C-type lectin receptors (CLRs), adhere to the surface of microbes or other particles. Among these receptors, CLRs are frequently reported to have an essential role in taking up microbes or other particles into phagocytic cells 5,6 ; therefore, CLRs are often called endocytic/phagocytic receptors or antigen-uptake receptors. Many types of CLRs, such as Dectin-1, Dectin-2 and Mincle, have been proposed to individually participate in the immune response to microbes, through binding to carbohydrates. While Dectin-1 recognizes β-glucans, Dectin-2 and Mincle recognize mannosylated ligands 7 ; however, it is not known whether recognition of KW3110 by macrophages is related to these CLRs. Intracellular signaling pathways involved in phagocytosis are also important in inducing IL-10 production. After phagocytosis, the fractions which are engulfed and internalized by endocytosis are collectively digested or escape from the endosome via endosomal peptide transporters. Proton coupled peptide transporter (POT), a member of a major transporter family, are expressed in all organisms and use the proton electrochemical gradient to drive the uptake of peptides across cell membranes 8,9 . POTs are responsible for the egress of fragments containing muramyl dipeptides (MDP). It is known that bacterial MDP stimulated a type of POT in RAW 264.7 cells which, in turn, facilitated KW3110 is efficiently incorporated into RAW 264.7 cells. The phagocytosis of several ligands, including bacteria, into RAW 264.7 cells activated IL-10 production 14 ; therefore, we evaluated whether KW3110, LGG and NRIC1942 are incorporated into RAW 264.7 cells and induce IL-10 production.
We analyzed KW3110-stimulated RAW 264.7 cells by Transmission electron microscopy (TEM). As shown in Fig. 2a, KW3110 were incorporated into RAW 264.7 cells at least 2 h after stimulation. We also evaluated the incorporation of FITC-labeled LAB into RAW 264.7 cells using fluorescence microscopy and flow cytometry (Fig. 2b,c). Microscopy analysis showed that the amount of KW3110 internalized into RAW 264.7 cells was greater than that of NRIC1942 and LGG internalization (Fig. 2b). In addition, flow cytometry analysis showed that the fluorescence intensity of FITC-labelled KW3110 was higher than the other LAB in RAW 264.7 cells (Fig. 2c). These results indicated that KW3110 was incorporated more efficiently into RAW 264.7 cells than NRIC1942 and LGG. Incorporation into RAW 264.7 cells occurred in a KW3110 dose-dependent manner (Fig. 2d). In addition, a strong correlation was found between the amount of incorporation and IL-10 production (Fig. 2e). On the other hand, IL-10 production in RAW 264.7 cells stimulated by KW3110 was suppressed by a phagocytosis inhibitor (Fig. 2f).
Cleac4n, gene symbol of Dectin-2, was up-regulated in response to KW3110. To identify the pathways potentially involved in phagocytosis of KW3110 by RAW 264.7 cells, we performed RNA-seq analysis to investigate gene expression in RAW 264.7 cells treated with KW3110, NRIC1942, or LGG. Although many genes are up-or down-regulated in response to LAB species stimulation (Fig. 3a), gene expression was more LGG for 24 h. IL-10 concentration in the supernatants was measured by ELISA. Significant differences were compared to the control group, *p < 0.05, **p < 0.01. Data are the means ± SD. Data are representative of at least three independent experiments. www.nature.com/scientificreports/ www.nature.com/scientificreports/ altered in KW3110-stimulated RAW 264.7 cells when compared with cells stimulated with the other two species. Among the genes whose expression level changed, we focused on CLRs. The category classification of genes, based on KEGG (Kyoto Encyclopedia of Genes and Genomes), showed that many C-type lectin receptor signaling pathway genes were up-or down-regulated when RAW 264.7 cells were stimulated by KW3110 (Fig. 3b). CLRs are known to recognize microbial cell walls and influence the innate immune system through phagocytosis 15 . Clec4n (C-type lectin domain family 4, member N), gene symbol of Dectin-2, showed the highest amount of up-regulation by KW3110 compared with NRIC1942 and LGG. Clec7a (C-type lectin domain family 7, member A), gene symbol of Dectin-1, which belongs to the same protein family as Dectin-2, was not up-regulated by KW3110 (Fig. 3c). Clec4n was up-regulated in RAW 264.7 cells after stimulation by KW3110 and LGG but not NRIC1942. Upregulation of Clec4n was much higher in KW3110-stimulated RAW 264.7 cells than LGG-stimulated cells.
To confirm the gene expression levels obtained in the RNA-seq analysis, expression of Clec4n was evaluated using real-time PCR (Fig. 3d). The mRNA expression levels of Clec4n were significantly up-regulated only when stimulated by KW3110: the mRNA expression levels were not changed when stimulated by NRIC1942 or LGG.

Dectin-2 involvement in KW3110 phagocytosis.
To clarify the role of Dectin-2 in the phagocytosis of KW3110 into RAW 264.7 cells, we conducted siRNA and ELISA analyses. Knockdown of Dectin-2 (si-Dectin-2) significantly suppressed the incorporation of KW3110 into RAW 264.7 cells and IL-10 production when compared with the control siRNA (si-Control) cells (Fig. 4a,b). On the other hand, knockdown of Dectin-1 (si-Dectin-1), which belong to the same family as Dectin-2, did not decrease the incorporation of KW-3110 and IL-10 production (Supplemental Fig. 2a,b). Consistent with the siRNA treatment, pretreatment with anti-Dectin-2 antibody decreased incorporation of KW3110 into RAW 264.7 cells and IL-10 production when compared with the control cells ( Fig. 4c,d). Dectin-2 activates the downstream protein-tyrosine kinase, Syk 16 , which plays an essential role in phagocytosis in macrophages 17 . The mRNA expression levels of Syk were significantly up-regulated after KW3110 stimulation of RAW 264.7 cells (Supplemental Fig. 2c). We evaluated the amount of KW3110 incorporated and the KW3110-induced IL-10 production from si-Syk RAW 264.7 cells. As expected, both the amount of KW3110-incorporation into si-Syk RAW 264.7 cells and IL-10 production from si-Syk RAW 264.7 were significantly decreased (Fig. 4e,f).
PHT2 is involved in transport of phagocytosed KW3110. We evaluated the Dectin-2 downstream signaling pathway in RAW 264.7 cells. It is already known that IL-10 production in macrophages by KW3110 stimulation is dependent on phagocytosis and the NOD2 signaling pathway 2 . Consistent with the previous study, we confirmed that pretreatment with NOD2 signaling inhibitor (Gefitinib) suppressed the effect of KW3110 on IL-10 production in RAW 264.7 cells (Fig. 5a). NOD2 recognizes bacterial peptidoglycan fragments containing muramyl dipeptides (MDP) in the cytoplasm of macrophages 18 . As some members of the POT family are known to be endo-lysosomal peptide transporters 19 , we investigated whether POT family members were involved in the signaling pathway. The POT family is comprised of four protein members: PEPT1 (gene symbol: SLC15A1), PEPT2 (gene symbol: SLC15A2), PHT1 (gene symbol: SLC15A4) and PHT2 (gene symbol: SLC15A3). Quantitative PCR revealed that SLC15A1 and SLC15A2 were not expressed in RAW 264.7 cells at the mRNA level (data not shown). Both SLC15A3 and SLC15A4 were expressed in RAW 264.7 cells. Whilst SLC15A3 expression levels increased 23.0-fold after KW3110 stimulation (Fig. 5b), stimulation by NRIC1942 or LGG did not affect the expression of SLC15A3 in RAW 264.7 cells. SLC15A4 expression increased 1.9-fold when stimulated by NRIC1942, and 1.2-fold when stimulated by KW3110. Moreover, siRNA of SLC15A3 (si-PHT2) suppressed IL-10 production in KW3110-stimulated RAW 264.7 cells (Fig. 5c). SiRNA of SLC15A4 (si-PHT1) did not decrease IL-10 production (Supplemental Fig. 3a).
Lectin microarray showed that α-mannose-binding lectins recognize KW3110. To investigate whether the polysaccharides on the surface of KW3110 have high affinity for Dectin-2, we performed lectin microarray analysis. Glycan profiles of KW3110 were mostly occupied by α-mannose-type glycans, including GNA, HHA, and NPA (Fig. 6a,b). The profiles were distinct from those of NRIC1942 and LGG.

Discussion
LAB are widely known for their anti-inflammatory effects. KW3110 is one of the LAB which is reported to activate macrophages and induce IL-10 production; however, the biological mechanism remains unclear. In this study, we revealed that the phagocytosis process is essential for IL-10 production in macrophages stimulated by KW3110. Dectin-2, which recognizes α-mannose, contributes to phagocytosis of KW3110 in RAW 264.7 cells. Furthermore, lectin microarray experiments revealed that KW3110 had higher binding affinities to lectins, which recognize the carbohydrate chains comprised of α-mannose, than NRIC1942 and LGG. These results suggest that the characteristic carbohydrate chains of KW3110 might contribute to easy incorporation into RAW 264.7 cells.
Pretreatment with cytochalasin D suppressed IL-10 production in KW3110-stimulated RAW 264.7 cells. As cytochalasin D is a phagocytosis blocker, phagocytosis of KW3110 is suggested to be essential for IL-10 production from macrophages. We investigated the carbohydrate chains contained within the cell wall of LAB, which are known determining factors of phagocytosis in macrophages 20 . Carbohydrate chains are inserted into lipid bilayers in cells, as glycoproteins and glycolipids, and are known to be involved in cell-cell or cell-microbe interactions. Some PRRs, such as Mincle and Dectin-2, are known to impact on the clearance of pathogens. PRRs on macrophages interact with the surface carbohydrate of pathogens, leading to promotion of phagocytosis by macrophages 21  www.nature.com/scientificreports/   25 , whereas Dectin-2 recognizes α-mannose 7 . Both receptors are important in the recognition and phagocytosis of microbes, and contribute to inducing cytokine and ROS production via the Syk-CARD9-NF-kB pathway 16,21 . We showed that KW3110 significantly increased the expression levels of Clec4n in RAW 264.7 cells when compared with NRIC1942 and LGG. The uptake amount of KW3110 by RAW 264.7 cells was significantly suppressed by knock-down of Clec4n. Furthermore, transcription analysis by RNAseq showed that Clec4n was also the most up-regulated gene in the category of C-type lectin receptor signaling pathway genes in response to KW3110 stimulation. Whereas Clec4n gene expression was not up-regulated in Although previous reports showed that Dectin-2 mediates the phagocytosis of cancer cells by Kupffer cells and the phagocytosis of mycobacterium by dendritic cells (DC) 26,27 , to our knowledge, there are no reports that Dectin-2 is involved in the uptake of LAB. A previous study reported that the uptake of L. reuteri ATCC53608 was not affected when Dectin-2 was blocked in DC 28 . The results suggest that the involvement of Dectin-2 in phagocytosis depends on the LAB species and differences might be regulated by the structure of carbohydrate chains. Mincle is also reported to be a lectin receptor, recognizing α-mannose; a recent study reported that Malassezia is cooperatively recognized by Mincle and Dectin-2 through distinct ligands 29 . In this study, the uptake amount of KW3110 by si-Mincle cells did not significantly differ from si-control cells; suggesting that Mincle is not involved in at least phagocytosis of KW3110 by RAW 264.7 cells.
The engulfment of microbes by phagocytosis results in formation of an intracellular vacuole, termed the endosome. The fragment of LAB engulfed into the endosome is either collectively digested 30 or transferred into the cell via a transporter. The POT family of proteins are peptide transporters 19 , comprised of four protein members: the peptide transporters PEPT1, PEPT2, and the di/tripeptides transporters PHT1, PHT2, which are expressed in various cells. SLC15A1 (gene name of PEPT1) is abundantly expressed in the small intestine and SLC15A2 (gene name of PEPT2) is expressed in various organs such as lung, brain and kidney, and various cell lines 31 . In vitro studies showed that PEPT1 is involved in facilitating the immune response 32 . PEPT2 is known to be a high-affinity and low-capacity transporter. Both SLC15A4 (gene name of PHT1) and SLC15A3 (gene name of PHT2) are expressed in various organs, including brain, placenta, lung, and thymus. PHT1 is www.nature.com/scientificreports/ reported to be associated with some diseases, including diabetes 33 . PHT2 is reported to play a role in antiviral innate immune responses 34 . A recent study also indicated that both PHT1 and PHT2 localized to endosomes and were responsible for the egress of fragments 35 . We found that SLC15A4 and SLC15A3 were expressed in RAW 264.7 cells, whereas SLC15A1 and SLC15A2 were not. PEPT1 and PEPT2 proteins were not expressed in CD11b+ macrophages 10 . These results suggest that PEPT1 and PEPT2 are not involved in phagocytosis in some macrophages, such as RAW264.7 and CD11b+ cells. Interestingly, we demonstrated that only SLC15A3 expression was increased in RAW 264.7 cells when stimulated by KW3110, whereas SLC15A4 expression was not changed. We also showed that inhibition of SLC15A3 resulted in suppression of IL-10 production in RAW 264.7 cells.
Bacterial MDP stimulated PHT2 in RAW cells which, in turn, facilitated the NOD2-dependent immune response 10 . SLC15A3 is responsible for the egress of MDP 35 . IL-10 production after KW3110 treatment was abolished by the inhibition of NOD2 in bone marrow macrophages 2 . Taken together, it is suggested that phagocytosed KW3110 is digested in endosomes and the MDP of the phagocytosed KW3110 reaches the cytosol via PHT2 where it activates NOD2 signaling, leading to IL-10 production. Surprisingly, IL-10 production in KW3110-stimulated RAW 264.7 cells was increased with knock-down of SLC15A4. We also demonstrated that SLC15A3 expression was significantly increased in si-PHT1 cells when compared to si-Control cells. Increased expression of SLC15A4 may contribute to the activation of MDP transport, leading to IL-10 production (Supplemental Fig. 3b).
In conclusion, we showed that KW3110 is easily incorporated into RAW 264.7 cells, and leads to IL-10 production. Using si-RNA techniques, we revealed that Dectin-2 expression in RAW 264.7 cells has a key role in phagocytosis of KW3110. In addition, the expression levels of SLC15A3 were high in KW3110 stimulated RAW 264.7 cells. After phagocytosis, the engulfed KW3110 may reach the cytosol via PHT2 and activate the IL-10 signaling pathway (Fig. 7).

Flow cytometric analysis.
After pre-incubation of RAW 264.7 cells for 24 h, FITC-labeled LAB (10 μg/ ml) were added. After 24 h, cells were washed with DMEM and resuspended in 0.2% trypan blue for 5 min to quench the fluorescence of non-ingested, but membrane-associated bacteria. The macrophages were suspended in 4% paraformaldehyde (Wako, Osaka, Japan) for flow cytometry analysis. Flow cytometric analysis was performed on a FACS Canto II (BD Biosciences, San Jose, CA, USA) and analyzed using FlowJo software. Cell populations were gated by the software program in the scatter diagram (forward-scattered light versus side-scattered light). FITC-positive population was defined as over 10 3 of intensity. The uptake quantity was determined as Cell Number × MFI.
Microscopic analysis of phagocytosis. After pre-incubation of RAW 264.7 cells for 24 h, FITC-labeled LAB (10 μg/ml) were added. After 24 h, cells were washed with DMEM and suspended in 0.2% trypan blue for 5 min; cells were then washed in and resuspended in PBS. RAW 264.7 cells were observed by confocal laser scanning microscope FV1000 (OLYMPUS, Tokyo, Japan).
Lectin array. The lectin microarray was performed using RayBio Lectin Array 95 (RayBiotech, Inc. Norcross, GA). All LAB species were dissolved at a concentration of 0.2 mg/ml in PBS and used as samples. After their addition to labeling reagent for 1 h, the samples were dialyzed and centrifuged at 1000 rpm for 5 min at 4 °C. Hybridizations were performed on the supernatant according to the manufacturer's protocol.
Transmission electron microscopy. RAW 264.7 cells were grown for 24 h on 35 mm dishes (5.0 × 10 5 cells). Cells were washed with PBS and exposed to 1 μg/m1 KW3110 in DMEM for 24 h. After fixation, each  TGT GTC CGT CGT GGA TCT GA  TTG CTG TTG AAG TCG CAG GAG   CLEC7A  AAT CCT GTG CTT TGT GGT AG  GAC TGA GAA AAA CCT CCT GTAG   Clec4n  ATT TCA TCA CCC AGC AGC  AAA ACA TCA TTC CA GCCCC   Syk  CTG GTT CCA TGG CAA CAT CTC  TGG CCC TGA TCA GGA ATT TTC   SLC15A1  GCC GGA CCA GAT GCA GAC GG  GCG GGT ACA CCA CAG CGT CC   SLC15A2  ATG GTG CTG GCA CTT GTT GTGT  AAC GGT TCC TGA AGC GGT TGCA   SLC15A4  GCT GCC ACC TGC ATT ACT ACTTC CGT ACT TCA CAG ACA CAA TGAG   SLC15A 3  GCT GAA GCT TGC GTT CCA A  AAC AGG TGG GCA CTT TCA GAGT