Free lipoproteins from Bifidobacterium longum alleviate osteoarthritis through modulation of the gut microbiome.

Aim: The "gut-joint" axis is suspected to be involved in the pathophysiology of osteoarthritis (OA). The present study aims at investigating the potential of lipoproteins (Lpps) secreted by Bifidobacterium longum to alleviate OA progression in the rat. Methods: Experimental OA was induced in rats harbouring Schaedler Flora maintained in SPF conditions. Two weeks post-injection, 20 rats were randomized to water (n = 10) or 0.3 mg/L Lpps solution (n = 10). Weight and food intake were monitored for 6 weeks. At sacrifice, joints were scored using macroscopic and histological criteria. Serum LPS, Schaedler flora as well as selected intestinal bacteria were analyzed. Results: Lpps intake prevents OA progression. The protected rats showed a significant increase in lactobacilli along the intestine as well as in Mucispirillum schaedleri in the colon and a significant decrease in Parabacteroides goldsteini and Akkermansia in caecum and colon, respectively. There was no significant difference in serum lipopolysaccharide or bacteria translocating in Peyer's patches. Labelled Lpps were not detected in bone marrow of the OA joint. The principal component analysis points out that OA prevention is primarily associated with bacteria involved in the tryptophane degradation pathway and SCFA formation. Conclusion: In rats deprived of bifidobacteria, intake of B.longum Lpps prevented OA development and modulated the intestinal microbiome with a possible impact on the bacterial end-products. The link between Lpps and the gut microbial metabolome warrants further investigation.

Gut-Joint Axis: Impact of Bifidobacterial Cell Wall Lipoproteins on Arthritis Development.

Gut microbiota affect progression of rheumatoid arthritis (RA). The present study aims at investigating the protective potential of Bifidobacterium longum cell wall lipoproteins (Lpps) shown to modulate the intestinal microbiome and prevent osteoarthritis. Arthritis was induced by collagen (CIA) or anti-collagen antibodies (CAIA) injection. Intake of 0.5 mg of Lpps/L, but not 0.25 and 1 mg of Lpps/L, significantly alleviated RA symptoms in CIA DBA/1OOaHsd mice. The arthritis index (AI) was also reduced in CAIA mice. In the CIA-protected group, colon Ligilactobacillus murinus, caecal Lactobacillus johnsonii and spleen weight correlated with AI, whereas the reverse was observed with splenic CD11c+ dendritic cells (cDCs). The unprotected CIA Lpps group harbored higher cecal and colon E. coli and lower caecal L. murinus. Lpps administration to CAIA mice after arthritis induction led to lower colon E. plexicaudatum counts. Splenocytes from CIA-protected mice triggered by LPS secreted higher Il-10 than control ones. However, a higher IL-10 response was not elicited in gnotobiotic RA mice splenocytes with lower cDCs' recruitment. Labeled bacteria with the Lpps signal were detected in CIA mice bone marrow (BM) cDCs 5 and 16 h post-gavage but not in Peyer's patches and the spleen. In vitro uptake of Lpps by primary BM and thymus cells was observed within 24 h. An FACS analysis detected the Lpps signal in the plasmacytoid cell compartment but not in cDCs. In conclusion, Lpps dosing is critical for preventing arthritis progression and appropriately modulating the microbiome. Our results also highlight the possible triggering of the immune system by Lpps.

[Coxsackievirus B infection and pathogenesis of type 1 diabetes]. / Infection à coxsackievirus B et pathogenèse du diabète de type 1.

Epidemiological and experimental studies suggest that enteroviruses (EV) and particularly coxsackieviruses B (CVB) are likely to trigger or accelerate the onset of islet autoimmunity and the development of type 1 diabetes (T1D) in genetically susceptible individuals. Several mutually non-exclusive mechanisms have been proposed to explain the involvement of CVB in the pathogenesis of T1D. CVB can infect and persist in the intestine, thymic cells, monocytes/macrophages, ductal cells and pancreatic ß-cells, which leads to structural or functional alterations of these cells. A chronic inflammatory response and disruption of tolerance towards ß-cells due to CVB infections are able to promote the recruitment and activation of pre-existing autoreactive T-cells and the destruction of ß-cells. Vaccine or therapeutic strategies to control EV infections have been developed and open perspectives for the prevention or treatment of T1D.

Diabetes progression and alterations in gut bacterial translocation: prevention by diet supplementation with human milk in NOD mice.

Impaired intestinal barrier function occurs before type 1 diabetes (T1D) onset with a possible contribution of microbial translocation. Breastfeeding is associated with enhanced mucosal intestinal integrity and T1D protection. Our aim was to study the potential of human milk (HM) to prevent diabetes onset and modulate the translocation of gut bacteria susceptible to breastfeeding or associated to diabetes onset. We show that HM intake can prevent T1D in nonobese diabetic mice independently of bifidobacteria colonization. Prior to diabetes onset, HM mice harbored splenic bacterial counts and plasma lipopolysaccharides level similar to control mice but exhibited a reduced expansion of Anaerotruncus sp. in pancreas and Lactobacillus johnsonii and Barnesiella in Peyer's patches (PP). Surprisingly, pancreas and PP bacterial expansion did not correlate with their own gut localization but with ileal Escherichia coli and cecal HM-susceptible bacteria (the promoted L. murinus and Bacteroides vulgatus, and the repressed B. fragilis and E. coli), respectively. Besides, higher colonic B. vulgatus counts induced by HM intake were associated with low islet infiltration and pancreatic E. coli expansion. On another hand, splenic dendritic cells (DCs) were identified as negative covariate of PP Barnesiella, suggesting a possible HM contribution to preserving splenic DCs through the reduction of Barnesiella translocation. Fecal B. vulgatus also negatively correlated with PP Barnesiella expansion, indicating that the mouse coprophagic behavior likely added to HM effect. Our findings provide evidence that HM has a multilevel impact and cooperates with some gut bacteria for controlling bacterial translocation at the earliest stage of insulitis.

Bifidobacteria-derived lipoproteins inhibit infection with coxsackievirus B4 in vitro.

The aim of the present study was to investigate the potential of bifidobacteria in protecting cells from coxsackievirus B4 (CV-B4) infection. Bifidobacterial screening identified two of five strains that protected human epithelial type 2 (HEp-2) cell viability when bifidobacteria were incubated with viral particles prior to inoculation. In contrast, no effect was shown by incubating HEp-2 cells with bifidobacteria prior to CV-B4 inoculation. Cell wall lipoprotein aggregates (LpAs) secreted by the selected strains were assayed for their antiviral activity. The two LpAs exhibited antiviral activity when they were incubated with viral particles prior to inoculation of HEp-2 cells. Recombinant LpA-derived protein exhibited identical antiviral activity. To identify the peptide sequences interacting with the virus particles, LpA proteins were aligned with the peptide sequences of the north canyon rim and puff footprint onto coxsackievirus and adenovirus receptor (CAR). The in silico molecular docking study using CV-B3 as template showed low-energy binding, indicating a stable system for the selected peptides and consequently a likely binding interaction with CV-B. Bifidobacterium longum and Bifidobacterium breve peptides homologous to the viral north rim footprint onto CAR sequence formed hydrogen bonds with several viral residues in the north rim of the canyon, which were already predicted as interacting with CAR. In conclusion, proteins from bifidobacterial LpAs can inhibit infection with CV-B4, likely through binding to the capsid amino acids that interact with CAR.

In silico mining and characterization of bifidobacterial lipoprotein with CHAP domain secreted in an aggregated form.

Bifidobacterium breve C50 secretes a lipoprotein associated with glucose, acting in an aggregating form (>600kDa) as an agonist of TLR2/6. Similar lipoproteins were sought for in bifidobacteria. In silico, the closest homology was shown with a Bifidobacterium longum protein containing CHAP and lipobox domains. Two strains secreted aggregates whose peptides sequences aligned with the mined protein. C16:0 and C18:0 fatty acids detected in the aggregates further supported a lipoprotein structure. Glucose and mannose detected by gas chromatography were likely ligands of the lipoprotein. The binding of aggregates to galectin-1 indicated that hexosamines and galactose surrounded them. However, unlike B. breve C50, aggregate secreted by B. longum CBi0703 was unable to bind TLR2/6 likely because of a more hydrophobic structure. In gnotobiotic mice, the intake of B. longum aggregate induced, in splenic dendritic cells, the expression of genes involved in antigen presentation. A positive correlation between the number of dendritic cells and CD4(+)CD25(+) cells was observed in mice receiving these aggregates. In conclusion, B. longum secretes a lipoprotein forming aggregates which may influence dendritic and CD4(+)CD25(+) cell interactions independently of the TLR2/6 pathway.

Bifidobacterium breve C50 secretes lipoprotein with CHAP domain recognized in aggregated form by TLR2.

Extracellular components secreted by Bifidobacterium breve C50 can induce maturation, high IL-10 production and prolonged survival of dendritic cells via a TLR2 pathway. In this study, the components were isolated from the supernatant by gel filtration chromatography. Antibodies raised against the major compounds with molecular weight above 600 kDa (Bb C50BC) also recognized compounds of lower molecular weight (200­600 kDa). TLR2 and TLR6 bound to the components already recognized by the antibodies. Trypsin digestion of Bb C50BC released three major peptides whose sequences displayed close similarities to a putative secreted protein with a CHAP amidase domain from B. breve. The 1300-bp genomic region corresponding to the hypothetical protein was amplified by PCR. The deduced polypeptide started with an N-terminal signal sequence of 45 amino acids, containing the lipobox motif (LAAC) with the cysteine in position 25, and 2 positively charged residues within the first 14 residues of the signal sequence. Lipid detection in Bb C50BC by GC/MS further supported the implication of a lipoprotein. Sugars were also detected in Bb C50BC. Close similarity with the glucan-binding protein B from Bifidobacterium animalis of two released peptides from Bb C50BC protein suggested that glucose moieties, possibly in glucan form, could be bound to the lipoprotein. Finally, heating at 100 °C for 5 min led to the breakdown of Bb C50BC in compounds of molecular weight below 67 kDa, which suggested that Bb C50BC was an aggregate. One might assume that a basic unit was formed by the lipoprotein bound putatively to glucan. Besides the other sugars and hexosamines recognized by galectin 1 were localized at the surface of the Bb C50BC aggregate. In conclusion, the extracellular components secreted by B. breve C50 were constituted of a lipoprotein putatively associated with glucose moieties and acting in an aggregating form as an agonist of TLR2/TLR6.

Intestinal colonization with bifidobacteria affects the expression of galectins in extraintestinal organs.

This study aimed at determining the contribution of intestinal bifidobacteria to the immune system activation using widely distributed galectins as markers of immune cell homoeostasis. In human flora-associated mice, bacteria were enumerated in the gut, blood, spleen, liver and lungs, while the expression of galectin-1 (Gal-1) and galectin-3 (Gal-3) was estimated by PCR in the intestine and real-time quantitative PCR in the other organs. Gal-1 and -3 were rarely expressed in the intestine. In blood, only Gal-1 was expressed while both galectins were expressed in all other organs. A high prevalence of colonic bifidobacteria was associated with a lower expression of both pulmonary galectins, whose levels negatively correlated with bifidobacterial counts. Caecal bifidobacterial counts also negatively correlated with pulmonary Gal-3 mRNA levels. The spleen was the only organ showing an upregulation of Gal-1 expression related to its bacterial contamination. However, this upregulation was only observed when bifidobacteria were not detected in the colon. A putative mechanism explaining the reduced expression of galectins when bifidobacteria highly colonize the mouse intestine could be that, by reducing the bacterial translocation, bifidobacteria also lead to a decreased blood concentration of substances produced by intestinal bacteria.

Infection of primary cultures of murine splenic and thymic cells with coxsackievirus B4.

Infection of primary cultures of total splenic and thymic cells from BALB/c and C3H/HeN mice with CVB4 E2 and JVB strains has been investigated. The presence of positive-strand viral RNA within cells was determined by semi-nested RT-PCR, and viral replication was attested by detection of intracellular negative-strand viral RNA and by release of infectious particles in culture supernatants. Viral replication occurred with both CVB4 strains to an extent dependent on the genetic background of the host. No interferon-alpha production was detected in the supernatants of CVB4-infected cultures using biological titration. Together these results suggest that infection of splenic and thymic cells can play a role in virus dissemination, and therefore in the pathophysiology of CVB4 infections.

Does the intestinal bifidobacterial colonisation affect bacterial translocation?

The aim of this work was to investigate the possible role of the intestinal anaerobic flora (especially bifidobacteria) in regulating bacterial translocation (BT) which can be defined as the passage of intestinal microbes through the mucosa to internal organs. Default in BT regulation concurs with pathogenesis of sepsis in various human conditions, such as acute pancreatitis, cirrhosis, necrotising enterocolitis or multiple organ failure. The intestinal flora was studied in human flora associated mice (HF mice) and BT was quantified in Peyer's patches (PP), blood, spleen, liver and lungs. HF mice displayed a heterogenic intestinal colonisation with bifidobacteria. High colonisation of both caecum and colon by bifidobacteria led to a poorer bacterial contamination of blood, liver and lungs. Moreover, ileal, caecal and colonic bifidobacterial counts negatively correlated with the bacterial dissemination (number of contaminated organs per mouse). In contrast, Bacteroides fragilis group counts positively correlated with bacteraemia, lungs contamination or bacterial dissemination. Additionally, clostridia localised in the colon affected bacterial uptake by PP and lungs contamination as indicated by positive correlations between bacterial populations in these respective locations. These results indicate that bifidobacteria, when established in high counts, reduced BT to liver, blood and lungs, whereas B. fragilis group favoured the bacterial passage. Clostridia established in the distal ileum also seemed to favour BT to lungs. The manipulation of the bacterial flora to optimise the regulatory effect on BT should therefore focus on the selective promotion of bifidobacteria and avoid an increase in potentially detrimental populations such as B. fragilis group and clostridia.

Increased poliovirus-specific intestinal antibody response coincides with promotion of Bifidobacterium longum-infantis and Bifidobacterium breve in infants: a randomized, double-blind, placebo-controlled trial.

To determine whether the size of the intestinal bifidobacterial population can influence the immune response to poliovirus vaccination in infants, we set up a randomized, placebo-controlled trial. From birth to 4 mo, infants were given a fermented infant formula (FIF) or a standard formula (placebo). Bifidobacteria were quantified monthly in infant stools. Antipoliovirus IgA response to Pentacoq was assessed before and 1 mo after the second vaccine injection. Thirty infants were randomized, and 20 completed the study (nine in the placebo group and 11 in the FIF group). Fecal bifidobacterial level was significantly higher with the FIF group at 4 mo of age (p=0.0498). Furthermore, B. longum/B. infantis carriage was higher at 4 mo in the FIF group (p=0.0399). Antipoliovirus IgA titers increased after Pentacoq challenge (p <0.001), and the rise was significantly higher in the FIF group (p <0.02). Antibody titers correlated with bifidobacteria, especially with B. longum/B. infantis and B. breve levels (p <0.002). Infants who harbored B. longum/B. infantis also exhibited higher levels of antipoliovirus IgAs (p <0.002). In conclusion, the present results indicate that antipoliovirus response can be triggered with a fermented formula that is able to favor intestinal bifidobacteria. Whether this effect on the immune system is achieved through the bifidogenic effect of the formula (mainly through B. longum/B. infantis and B. breve stimulation) or directly linked to compounds (i.e. peptides) produced by milk fermentation remains to be investigated.

Multiplex PCR using 16S rRNA gene-targeted primers for the identification of bifidobacteria from human origin.

Three multiplex polymerase chain reactions (PCRs) targeted on Bifidobacterium and related species were designed to identify human species. The selected primers yielded amplified products of various sizes, each specific for a species. Three to four pairs were gathered in one PCR reaction and their specificity under multiplex conditions was confirmed using DNA from 26 reference strains. Using this technique on unidentified faecal strains, B. bifidum, B. longum and B. breve species were commonly recovered in infants while B. adolescentis, B. catenulatum/B. pseudocatenulatum continuum and B. longum species were predominant in adults. Thus, a single PCR can provide the assignment of a strain to one these species, reducing the number of PCR reactions and hands-on time for the identification of human isolates of bifidobacteria. Moreover, this technique is also applicable for the in situ detection of bifidobacteria in DNA extracts from human stools.