Immunomodulation of Lactobacillus rhamnosus GG (LGG)-derived soluble factors on antigen-presenting cells of healthy blood donors

Lactobacillus rhamnosus GG (LGG) cells have been shown to promote type-1 immune responsiveness; however knowledge of immunomodulation of soluble factors secreted by LGG is limited. This is the first study to investigate whether LGG soluble factors promote a comparable immune responsiveness as the bacterial cells. Both treatments − LGG conditioned medium with (CM + LGG) or without (CM) LGG cells, in this study increased expression of several toll-like receptors (TLRs) in all studied cell types and antigen presentation-associated receptor HLA-DR in macrophages and “intermediate” monocytes; but decreased that of activation markers on monocytes and macrophages and production of IL-10, IL-12 and TNFα in macrophages. In co-culture with mononuclear cells, CM increased Th1-type cytokine profile but not as pronounced as CM + LGG. This study suggests that LGG soluble factors exert similar immunomodulatory effects as the intact cells, but cells may be required for optimal type-1 immune responsiveness polarizing capacity of this probiotic strain.


Expression of activation markers and intracellular cytokines in macrophages.
MFIs of activation markers and intracellular cytokines of CM-treated or CM + LGG-treated macrophages are shown in Fig. 3. Both treatments increased HLA-DR MFI but decreased MFIs of activation markers CD80, CD86, CD163, CD206, CD64 and CD209. They also reduced MFIs of M1-type cytokines IL-12, TNFα , Th1 polarization-associated transcription factor T-bet; and M2-type cytokine IL-10. CD80 MFI of CM + LGG-treated macrophages was significantly higher than that of CM-treated macrophages.
Cytokine secretion profiles of DC-PBMC co-cultures. Cytokine secretion profiles of co-cultures of treated DCs and mononuclear cells are shown in Fig. 4. Both treatments increased type-1 immune response-associated cytokines (IL-1α , IFNγ and TNFα ) and Th17-associated cytokines (GM-CSF, IL-17F and IL-6). They decreased Th2 signature cytokine (IL-4) but increased IL-25. A pronounced increase was seen in the production of regulatory cytokine IL-10 with CM + LGG treated DCs but not with CM-treated DCs. Levels of IL-1α , IFNγ , TNFα , IL-12(p70), IL-23, GM-CSF, IL-17F and IL-10 were significantly higher in cultures with CM + LGG-treated DCs than in CM -treated DCs.

Discussion
To the best of our knowledge, this is the first study to demonstrate the immunomodulatory effects of LGG soluble factors on APCs of human blood donors. The results indicate that the soluble factors released during LGG growth LGG conditioned medium (CM); or LGG conditioned medium with LGG cells (CM + LGG); or 1000 U/ml recombinant human (rh) IFNγ , 1000 ng/ml lipopolysaccharides (LPS) and 1 μg/ml R848 (positive control; Pos); or alone (negative control; Neg) were determined by flow cytometer after 24-hour incubation. (a) "classical" CD14hiCD16-, (b) "intermediate" CD14hiCD16lo, and (c) "non-classical" CD14loCD16lo monocytic subsets were studied. Results are presented as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 in comparisons between treated monocytes and negative control. # p < 0.05, ## p < 0.01 and ### p < 0.001 in comparisons between CM-treated and CM + LGG-treated monocytes.
have a similar pattern but a relatively modest influence on APC activation and immunoresponse polarization as the viable bacterial cells per se.
In our previous study, we showed that LGG cells reduced TLR 2 expression of DCs, macrophages and monocytes and TLR 8 expression of macrophages of healthy blood donors 15 . In the present study, LGG soluble factors seemed to exert an opposite effect that they significantly increased TLRs 1, 4, 5, 6, 7, 8 and 9 expressions of DCs and macrophages, and TLRs 1 and 9 expressions of monocytes. Down-regulation of TLR mRNA levels has been indicated to be related to the ligation of those particular receptors 19 . In this case, the up-regulation of certain TLR expressions could be a general response due to failure of ligation of several, but not all, of the corresponding TLRs on the cell surface. The alteration of TLR expressions on the key APCs could influence the subsequent innate responsivenss to pathogen-associated molecular patterns (PAMPs) 20,21 . Intracellular antiviral pathways and production of antiviral effectors such as interferons (IFNs) and pro-inflammatory cytokines may be affected. Further investigations are, however, required.
One specifically interesting finding in this study is the significant reduction in the expression of several activation and inflammation markers on monocytes and macrophages, particularly the significant decrease in the expression of the adhesion molecule CD11b on all studied monocytic subsets. Increase in CD11b expression on circulating monocytes has been previously demonstrated to be associated with atherogenic inflammation and LGG conditioned medium (CM); or LGG conditioned medium with LGG cells (CM + LGG); or 1000 U/ml recombinant human (rh) IFNγ , 1000 ng/ml lipopolysaccharides (LPS) and 1 μg/ml R848 (positive control; Pos); or alone (negative control; Neg) were analyzed by flow cytometer after 24-hour incubation. Results were presented as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 in comparisons between treated macrophages and negative control. # p < 0.05, ## p < 0.01 and ### p < 0.001 in comparisons between CM-treated and CM + LGG-treated macrophages.
reflect the activity level of the disease 22,23 . Our unpublished data show that viable LGG cells (1 × 10 8 CFU/day) singificantly suppressed the atherogenesis in ApoE− /− mice with Western-diet induced atherosclerosis (Chan et al.,  unpublished). The treatments also consistently reduced other maturation and activation markers on the APCs but increased the expression of HLA-DR. Increase in HLA-DR expression indicates an increase in antigen presentation capacity. Antigen presentation by immature APCs theoretically promotes the formation of tolerogenic responses although no evidence has been found in the co-culture experiments [24][25][26] with autologous mononuclear cells. Moreover, both treatments decreased the expression of Th1 polarization-associated transcription factor T-bet in macrophages, implying a decrease in the expression of another signature M1-type cytokine IFNγ 27 , which was not measured in the present study.
Instead, DCs pretreated with CM alone or CM with LGG cells significantly increased the secretions of pro-inflammatory type-1 cytokines (IL-1α , IFNγ and TNFα ) in the PBMC co-cultures. While the effect was

. Cytokine secretion profiles of DC-peripheral blood mononuclear cell (PBMC) co-cultures.
Cytokine production levels of co-cultures of autologous PBMCs and monocyte-derived DCs (n ≥ 4) prestimulated 24 hours with LGG conditioned medium (CM); or LGG conditioned medium with LGG cells (CM + LGG); or 1000 U/ml recombinant human (rh) IFNγ , 1000 ng/ml lipopolysaccharides (LPS) and 1 μg/ml R848 (positive control; Pos); or alone (negative control; Neg) were analyzed by multiplex assay. Results are shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 in comparisons between treated cells and negative control. # p < 0.05, ## p < 0.01 and ### p < 0.001 in comparisons between CM-treated and CM + LGG-treated cells. evident with both pretreatments it was significantly greater with DCs pretreated with the cocktail of CM and viable LGG cells. This implies the importance of the cellular structures of LGG over its secreted metabolic products for pro-inflammatory effects, which is supported by previous studies 28,29 . These studies have suggested the importance of, for example, fimbriae or pilus adhesin on LGG for adhesion, close interaction between LGG and host cells, and modulation of inflammatory responsiveness and innate immune gene expression; lipoteochoic acid (LTA) for interaction with various pattern recognition receptors (PRRs); and exopolysaccharides (EPS) for modulation of adaptation of the bacterial cells.
In summary, the present study demonstrates in an in vitro model of APCs that the immunomodulation of LGG conditioned growth media shows a similar pattern as that of the viable probiotic strain per se. Both treatments decreased the constitutive expressions of several TLRs in all studied APCs, and activation markers and cytokine expressions of monocytes and macrophages. Based on the cytokine secretion profile of autologous co-cultures, the treated DCs appeared to promote type-1 pro-inflammatory responsiveness; and the influence was significantly higher with than without the viable cells. All these findings are of significant relevance to in vivo physiological responses, for example, manipulating intestinal microbial communities; suppressing adherence of pathogens to intestines; activating anti-apoptotic genes in intestinal epithelial cells; fortifying intestinal barrier; regulating intestinal homeostasis; and modulating systemic immunity [30][31][32][33] .

Conclusion
In conclusion, the soluble factors released by LGG during its growth have similar immunomodulatory effects as the intact cells, but the latter seem to be required for the optimal type-1 immune responsiveness polarizing capacity of this probiotic strain. These results can provide an evidence-based insight for further clinical development of this strain, such as drug synthesis and vaccine development, for various diseases.

Derivation of dendritic cells and macrophages from monocytes.
Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy blood donors (n ≥ 4) (Hong Kong Red Cross Blood Transfusion Service, Hong Kong, China) by density gradient centrifugation over Ficoll-Plaque TM Plus (GE Healthcare Life Sciences, Piscataway, NJ, USA) and seeded in complete RPMI-1640 medium (LONZA, Basel, Switzerland) for 2 hours at 37 °C to isolate monocytes. Extra PBMCs were frozen for later experiments. Isolated monocytes were derived to immature DCs in the presence of 40 ng/ml recombinant human (rh) granulocyte-macrophage colony-stimulating factor (GM-CSF) and 40 ng/ml rh interleukin (IL)-4 (both from PeproTech EC. Ltd, London, UK); or to immature macrophages in the presence of 40 ng/ml rh GM-CSF. Cells were incubated for 7 days with fresh supplemented medium added every 2 days. Protocol was modified from the studies of Lacey (2012) and Mohamadzadeh (2005) 16,17 , and approved by the University of Hong Kong. All methods were carried out in accordance with the approved guidelines.

Bacterial strain and stimulation of antigen-presenting cells (APCs).
LGG (ATCC 53103) was cul- Detection of toll-like receptor (TLR) expression by quantitative real-time PCR (qPCR). TLRs 1, 2, 4, 5, 6, 7, 8 and 9 mRNA levels of DCs, macrophages and monocytes were determined by qPCR with modified protocol and sequences of primers and probes from the study of Flacher et al. 18 . Data were acquired on ABI StepOne Plus Real-Time PCR System (Life Technologies, NY, USA).

Statistical Analysis.
Results were shown as mean ± standard deviation (SD) and analyzed with Kruskal-Wallis test, Mann-Whitney test, analysis of variance and two-tailed Student's t test. CM-treated and CM + LGG-treated cells were compared with the negative control and with each other. P-values below 0.05 were considered as significant. Statistical calculations were performed using GraphPad Software Prism Version 6.04 (San Diego, CA, USA) and SPSS Version 19 for Windows (Chicago, IL, USA).