Substances released from probiotic Lactobacillus rhamnosus GR-1 potentiate NF-κB activity in Escherichia coli-stimulated urinary bladder cells (2024)

Abstract

Introduction

Lactobacillus is a highly versatile genus of lactic acid bacteria comprising of several strains with probiotic properties, used for a variety of ailments including gastrointestinal and urogenital diseases (Reid & Bruce, 2006; Collado et al., 2009). Probiotic mechanisms are diverse, ranging from direct action against microbial pathogens through the antimicrobial action of bacteriocins and competitive exclusion of pathogens, to interactions with host cells by altering their responses to invading pathogens. Modulation of cellular immune responses has emerged as a new and important field that contributes to human health, and certain probiotic lactobacilli have proven to be effective immunomodulators both in vitro and in vivo (Wells, 2011). One important group of transcription factors is the nuclear factor-κB (NF-κB) dimers, which regulate many immunologically important genes (Hoffmann et al., 2002). Bacterial structures readily induce NF-κB activity, which in turn can regulate the transcription of a number of genes involved in inflammation, such as the pro-inflammatory cytokines tumor necrosis factor, interleukin-6 (IL-6), and the CXCL8 chemokine, important in immune cell activation (Collart et al., 1990; Libermann & Baltimore, 1990; Kunsch & Rosen, 1993). Up- and down-regulation of NF-κB activity by lactobacilli in immune cells and epithelial cells have previously been established (van Baarlen et al., 2009).

Urinary tract infections (UTI) are one of the most common bacterial infections worldwide, and most of them are caused by uropathogenic Escherichia coli (UPEC) (Foxman, 2002). Epithelial cells recognize UPEC structures such as lipopolysaccharide and flagellin and respond by the production of inflammatory mediators, in large controlled by NF-κB (Bäckhed et al., 2001). UPEC can, however, interfere with normal NF-κB responses, by the production of substances that inhibit early activation of NF-κB (Cirl et al., 2008). Modulation of NF-κB activity by lactobacilli in urothelial cells is largely unexplored, although finding bioactive compounds produced by probiotic lactobacilli could be of scientific and therapeutic value in UTI prevention and therapy.

Although the urinary bladder is considered sterile, the urethra harbors a large number of microorganisms including lactobacilli (Dong et al., 2011) that are believed to act as a barrier, protecting from ascension of pathogens into the upper urinary tract. Lactobacillus rhamnosus GR-1 is a urethral isolate that has been used in combination with another Lactobacillus strain, Lactobacillus reuteri RC-14, to treat or prevent several urogenital conditions. The strain combination is able to normalize the vagin* in women with bacterial vaginosis (Anukam et al., 2006) and reduce UTI recurrences (Reid et al., 1995). While L. reuteri RC-14 is believed to secrete products that inhibit adhesion of Gram-positive pathogens to host tissue (Velraeds et al., 1996), L. rhamnosus GR-1 can kill E. coli and can disrupt biofilms produced by these microbes (McMillan et al., 2011). Moreover, this strain can modulate aspects of host immunity, including NF-κB and mitogen-activated protein kinases (Kim et al., 2006; Karlsson et al., 2012). However, the substances responsible for immune-modulating effects remain unknown. The aim of this study was to identify and characterize components of L. rhamnosus GR-1 that influence modulation of NF-κB activity in urothelial cells.

Materials and methods

Cell culture

The T24 cell line (ATCC HTB-4) is a human bladder epithelial cell line purchased from ATCC (LGC standards). Cells were cultured in RPMI 1640 medium supplemented with 2.05 mM l-glutamine and 10% fetal bovine serum (Hyclone), at 37 °C in a humidified environment. Cell passage number did not exceed 15 in cell challenge experiments.

Bacterial strains and growth conditions

Lactobacillus rhamnosus GR-1 was grown on deMan Rogosa Sharp (MRS) agar (Difco) at 37 °C for 24 h in a jar containing an oxygen-free environment (Anaerobic pouch system EZ, BD). Fresh cultures were inoculated into MRS broth (Difco) and grown statically for 24 h at 37 °C, anaerobically. This culture was used as starter culture for further experiments and growth of L. rhamnosus GR-1. Viable L. rhamnosus GR-1 was grown from a 1% inoculum for 24 h at 37 °C under anaerobic and static growth conditions. UPEC GR-12 was cultured using LB broth or agar (Difco) at 37 °C. Colonies of E. coli were inoculated into LB broth and grown overnight at 37 °C with constant shaking. Heat-killed L. rhamnosus GR-1 and E. coli GR-12 were prepared by washing bacteria in phosphate-buffered saline (PBS; pH 7.4) and heating cells for 1 h at 70 °C. Lack of viability was confirmed by plating heat-killed cells on their corresponding growth media.

Isolation of released products

A colony from a fresh plate of L. rhamnosus GR-1 was inoculated into 15 mL of MRS medium and grown for 24 h statically at 37 °C. The culture was then transferred to 600 mL of preheated MRS and grown for an additional 24 h. The lactobacilli culture was centrifuged (10 000 g) for 10 min at 4 °C and washed three times with sterile MilliQ water (Millipore). The pellet was resuspended in 200 mL of PBS and stirred gently for 2 h. The PBS–bacterial suspension was thereafter centrifuged (10 000 g) for 10 min at 4 °C to remove the bacterial cells, filtered through a 0.2-µm filter, and frozen in aliquots at −20 °C until use.

Molecular mass fractionation of isolated products

Lactobacillus isolated products were fractionated using centrifugal concentrators with 100-, 50-, 30-, 10-, and 5-kDa molecular-mass-cutoff (MMCO) polyethersulfone membranes (Vivaspin, Sartorius). Fractionation was performed sequentially, starting with the biggest MMCO membrane. Concentrator tubes were centrifuged at 5000 g for 40 min (5- to 50-kDa MMCO membranes) or 3000 g for 1 h (100-kDa MMCO membrane). After each fractionation, the eluent was transferred to another fractionation tube, with a smaller molecular MMCO membrane. All steps during fractionation was carried out at 4 °C, and the concentrate from each round of fractionation was saved at −20 °C. Protein content was measured with a micro-BCA assay (Pierce) against a bovine serum albumin standard to determine total protein in the solution before fractionation. A second BCA assay was performed on all fractions to determine the contribution of each fraction to total protein amount of released products from L. rhamnosus GR-1.

Trypsinization of surface proteins

Lactobacilli were grown in 150 mL of MRS at 37 °C until an OD_{600 nm} of 0.6. Culture supernatant was removed after 5-min centrifugation at 7000 g. The remaining pellet was washed three times in PBS and resuspended in diluted 1 : 1 (v/v) PBS with 40% (w/v) sucrose and 10 µg of sequencing grade modified trypsin (Promega). After 30 min at 37 °C, the digestion mixture was centrifuged at 3500 g for 10 min at 4 °C. The soluble fraction was filtered using 0.2-µm filter and stored at −20 °C until further use. Cell fraction was washed three times with PBS and kept at 70 °C for 1 h and frozen at −20 °C until use. Heat killing of bacteria was confirmed by plating on MRS agar.

Surface protein detachment by LiCl

Bacterial pellet was prepared as described previously and diluted 2 : 1 (v/v) in PBS containing 1 M LiCl. After 3 h of constant shaking at 20 °C, the protein-containing soluble fraction was collected by centrifugation for 10 min at 3500 g, 4 °C and stored at −20 °C. The remaining cell fraction was washed and heat-killed as described previously.

SDS-polyacrylamide gel electrophoresis

Unfractionated and fractionated L. rhamnosus GR-1-released products were heated to 95 °C for 5 min in sample buffer with 5% (v/v) 2-mercaptoethanol and separated by SDS-PAGE (10% resolving gel; Bio-Rad). Total protein amount of 5 µg was added for each fraction, and 0.5–1 µg was added of unfractionated released products from L. rhamnosus GR-1. The gel was stained using the plus one silver staining kit (GE Healthcare) according to the manufacturer's instructions.

Two-dimensional gel electrophoresis and proteomics

Lactobacillus rhamnosus GR-1 material, from the three largest fractions (100, 50, and 30 kDa), was further analyzed using a two-dimensional gel electrophoresis system (Multiphor II; GE Healthcare). Immobiline dry strips (18 cm, pH 3–11NL) were rehydrated for 16 h with 340 µL of rehydration solution [8 M urea, 2% triton x-100, 0.5% (v/v) Immobilized pH gradient (IPG) buffer, 0.002% (w/v) bromophenol blue and 20 mM DTT] before isoelectric focusing (500 V, 1 min; 3500 V 1.5 h; 3500 V, 6 h) at 2 mA and 5 W in gradient mode. Following isoelectric focusing, the dry strips were equilibrated twice in a buffer containing urea (6 M), Tris–HCl (75 mM, pH 8.8), glycerol (29.3%, v/v), SDS (2% w/v), and bromophenol blue (0.002%, w/v). Equilibration buffer was supplemented with DTT (1% w/v) during the first equilibration and iodoacetamide (2.5%, w/v) for the second equilibration step. IPG strips were placed on top of Excelgels (ExcelGel SDS 2-D hom*ogenous 12.5; GE Healthcare), and the Multiphor unit was assembled according to the manufacturer's instructions. Separation was carried out at 15 °C with the following settings: 120 V and 20 mA for 40 min followed by 600 V and 50 mA for 1 h. Proteins were stained using a silver staining protocol compatible with protein characterization using MALDI-TOF MS fingerprinting analysis (Mortz et al., 2001). Selection of proteins for further analysis was carried out by an overlay method in Photoshop CS5 (Adobe) with three layers, each corresponding to one gel. We selected candidate proteins by identifying spots predominantly present in the 50-kDa fraction, yet absent in the 30-kDa fraction. The candidate proteins were sent for identification to Alphalyse A/S (Odense, Denmark) using MALDI-TOF MS fingerprinting and subsequent matching to database entries. Identified proteins were analyzed for the presence of signal peptides using the signalp software (version 4.0; (Petersen et al., 2011) and secretion potential through secretomep (version 2.0; (Bendtsen et al., 2005). Subcellular localization was predicted using psortb (version 3.0.2) (Yu et al., 2010).

Transfection and luciferase reporter assay

One day prior to transfection, approximately 5 × 10⁴ T24 cells per well were seeded into a 24-well tissue culture plate. The following day, cells were transfected using 1.5 µL per well of Lipofectamine 2000 (Invitrogen) for 6 h, according to the manufacturer's protocol. Each well received 0.54 and 0.06 µg of endotoxin-free (endotoxin-free plasmid DNA purification kit; Macherey-Nagel) pNF-kB-Luc (Clontech) and pRL-CMV (Promega) vectors, respectively. Luciferase expression was measured using the dual-luciferase reporter assay (Promega) and a Turner TD20/20 luminometer (Turner biosystems) set to a 10-s measurement with an initial 2 s delay. NF-κB activity was expressed as ‘relative NF-κB activity’ in the figures and calculated as the ratio between firefly luciferase and Renilla luciferase activities. Ratios were normalized, either against the nonstimulated control cells or against the group stimulated with heat-killed E. coli.

Epithelial cell line challenge

To determine the level of NF-κB activity, transfected T24 cells were used for experiments within 48 h after transfection. Stimulation of T24 cells with viable or heat-killed lactobacilli, respectively, was performed by adding 2 × 10⁷ CFU mL⁻¹ of lactobacilli to the cells in 24-well plates. Released products from L. rhamnosus GR-1 were added at a concentration of 0.3 mg mL⁻¹. Initial protein content from each fraction was calculated back to the original composition in the isolated released products and was added proportionally to the cells during stimulation.

Statistical analyses

Statistical differences between means in the cell culture experiments were evaluated using the Student's t-test, with P < 0.05 used as a limit for statistical significance. All statistical analyses were performed using graphpad prism version 4.0 (Graphad Software).

Results

Released products from L. rhamnosus GR-1-potentiated NF-κB activity in E. coli-challenged cells

T24 bladder cells responded to stimulation with heat-killed E. coli GR-12 by more than 10-fold induction of NF-κB, which is consistent with earlier findings (Karlsson et al., 2012). However, stimulation with viable, heat-killed, or released products from L. rhamnosus GR-1 did not significantly increase NF-kB activity in T24 cells (Fig. 1A). Interestingly, viable L. rhamnosus GR-1 potentiated the NF-κB activity induced by E. coli (Fig. 1B). The NF-κB activity during co-stimulation was significantly increased compared with cells challenged with E. coli alone. Isolated released substances from L. rhamnosus GR-1 similarly potentiated NF-κB activity during co-stimulation, although not to the same extent as viable bacteria. Of the different Lactobacillus preparations, heat-killed bacteria showed lowest, yet significant, NF-κB augmentation.

L. rhamnosus GR-1 surface components did not induce NF-κB augmentation

Heat-killed L. rhamnosus GR-1 showed a small but significant effect on NF-κB activity in E. coli-stimulated bladder cells. In an attempt to determine the role of components bound through covalent or ionic interactions to the bacterial surface, viable L. rhamnosus GR-1 underwent trypsin treatment to digest surface proteins into active fragments, and LiCl treatment to remove lattice proteins on the lactobacilli surface with possible immunomodulatory effects. Stimulation of bladder cells with Lactobacillus-derived peptides from trypsin digestion or LiCl-released proteins showed no effect on NF-κB activity. The L. rhamnosus GR-1 cells were heat-killed after trypsin or LiCl treatment to prevent the restoration of the surface proteins, and these cells did not show any effect on NF-κB activity (Fig. 2A). Moreover, none of the different trypsin treatment preparations significantly altered NF-κB activity in E. coli-challenged cells. However, although the soluble fraction from LiCl treatment had no effect on NF-κB in E. coli-challenged cells, the cellular LiCl fraction inhibited the E. coli-dependent increase in NF-κB activity (Fig. 2B).

High-molecular-mass fractions potentiated NF-κB responses in bladder cells treated with E. coli and L. rhamnosus GR-1 in co-culture

Size fractionation of the released products from L. rhamnosus GR-1 was performed using centrifugal concentrators with different MMCO membranes. Size separation of the different fractions was validated by SDS-PAGE, showing a decrease in the molecular mass of proteins with the use of columns with lower MMCO membranes (Fig. 3). The contribution of each fraction to the total protein amount released by L. rhamnosus GR-1 was evaluated by determining the protein content in each isolated fraction. Most proteins were found within the fraction containing the large-sized proteins, contributing to almost half (49%) of all proteins of the released products (Table 1). The smallest fraction collected from the 5-kDa MMCO membrane also contributed substantially to the total protein (32%), whereas the other fractions had lower protein quantities.

Released L. rhamnosus GR-1 products in PBS, fractionated by centrifugal concentrators with different MMCO membranes, were used in co-culture experiments together with heat-killed E. coli. The isolated fractions from tubes with MMCO membranes of 100, 50, 30, 10, and 5 kDa showed a differential response from E. coli-stimulated bladder cells. Individual stimulation with different fractions had no significant effect on NF-κB activity in nonstimulated cells (Fig. 4A). However, some of them were able to potentiate NF-κB activity in cells challenged with E. coli. Stimulation with unfractionated released products from L. rhamnosus GR-1 demonstrated further induction of NF-κB activity, as did the fractions from 100- and 50-kDa MMCO membrane subsets (Fig. 4B). The remaining fractions showed no significant differences from E. coli-stimulated cells.

NF-κB-potentiating fractions showed accumulation of specific proteins

Although both the 100-kDa and 50-kDa fractions potentiated NF-κB activity, we hypothesized that the 50-kDa fraction would have a greater accumulation of the active product compared with the 100-kDa fraction. The 50-kDa fraction marginally contributed to the total protein amount of released products (Table 1), yet retained high biological activity (Fig. 4B).

To evaluate the differences in protein patterns between the isolated fractions, we performed two-dimensional gel electrophoresis on the three largest fractions (100, 50, and 30 kDa). The 50-kDa fraction had a large number of protein spots with high intensity, although some of the spots were present in all fractions. Two of the spots were selected as reference proteins, while eight other proteins with exclusively high presence in the 50-kDa fraction were identified (Table 2, Fig. 5A). Phosphoglycerate kinase and the NLP/P60 family protein were predominantly found in the 50-kDa fraction (Fig. 5B). Chaperonin GroEL, elongation factor Tu, FKBP-type peptidyl-prolyl cis-trans isomerase, enolase, and L-lactate dehydrogenase were found in the 50-kDa fraction and to a lesser extent in the 100-kDa fractions, whereas the 30-kDa fraction exhibited a very different protein pattern (Fig. 5C). One protein that accumulated in the 100-kDa fraction (Fig. 5C, labeled with an asterisk) was identified as fructose/tagatose bisphosphate aldolase. Two of the identified proteins were predicted to be extracellular as determined by psortb, signalp, and secretomep: FKBP-type peptidyl-prolyl cis-trans isomerase and an NLP/P60 family protein (Table 2).

Discussion

We establish that released proteins from the probiotic L. rhamnosus GR-1 could potentiate E. coli-dependent NF-κB activity in T24 urothelial cells, although they did not affect NF-κB activity in nonstimulated cells. Immunomodulatory properties of lactobacilli are dependent on both cell wall-anchored and released products (Grangette et al., 2005; Kim et al., 2006). We isolated released products from L. rhamnosus GR-1 and separated them into five different enriched molecular mass fractions. The fractionation of secreted proteins from L. rhamnosus GR-1 suggested that the effect was dependent on one or multiple substances ranging from 50 kDa to larger than 100 kDa in size, whereas components from the Lactobacillus surface demonstrated no induction of NF-κB activity. Identification of the individual proteins present within the active fractions revealed several putative immunomodulatory factors that might be of importance in regulating NF-κB activity.

Although size separation on SDS-PAGE showed that the fractionations gave distinct groups of proteins, there was no clear cutoff size separating the different fractions. Fractionation by MMCO membranes was influenced by the tertiary and quaternary structures found in native proteins. Therefore, SDS-PAGE separation of denatured proteins most likely underestimated the effectiveness of the fractionation process. It cannot be ruled out that the largest fraction containing proteins from the 100-kDa MMCO membrane includes cell wall residues that were not completely removed during centrifugation and filtration of the released products in PBS. These residues might contribute to the large protein amount present in the 100-kDa MMCO membrane fraction (49% of total protein).

Cell wall components of lactobacilli such as lipoteichoic acid have previously been shown to stimulate activation of the NF-κB pathway (Matsuguchi et al., 2003). However, cells concomitantly stimulated with E. coli and heat-killed L. rhamnosus GR-1 exhibited a very low induction of NF-κB activity compared with cells stimulated with viable L. rhamnosus GR-1 or released products, although there is no difference in the amount of lipoteichoic acid between viable and heat-killed lactobacilli. Moreover, neither peptides from trypsin digestion nor proteins isolated by LiCl treatment showed any effect on bladder cells during co-stimulation with E. coli. LiCl is known to remove surface (S)-layer and other noncovalently-bound proteins from the Lactobacillus surface. Such S-layer proteins are able to modulate immune function in dendritic cells and T lymphocytes, without the involvement of Toll-like receptor 2 and Toll-like receptor 4 (Konstantinov et al., 2008). As the soluble fraction from LiCl-treated L. rhamnosus GR-1 had no effect on NF-κB activity, S-layer proteins were eliminated as a source of NF-κB potentiation. However, the L. rhamnosus GR-1 cell fraction following LiCl treatment considerably reduced NF-κB activity in E. coli-stimulated bladder cells, a yet unexplained result.

Although released products or fractions alone were unable to affect NF-kB activity, they considerably elevated NF-κB activity in cells subjected to E. coli challenge. Released products from L. rhamnosus GR-1 and the two largest MMCO fractions showed comparably high NF-κB potentiation, whereas the three remaining fractions did not affect NF-κB potentiation. There are currently a limited number of published studies on secreted products from lactobacilli including L. rhamnosus. A report on released products from L. rhamnosus GG, a well-studied probiotic strain, showed that this bacterium secreted very few proteins during growth in MRS medium (Sanchez et al., 2009). By means of two-dimensional gel electrophoresis and subsequent identification of proteins present in the fractions, we could identify proteins that were putative candidates for modulating the activity of NF-κB. GroEL is a cytoplasmic protein that normally acts as a chaperone, aiding in the normal folding of proteins. GroEL was found in the 50-kDa fraction and has previously been reported to be released from lactobacilli and able to initiate production of CXCL8, an important chemokine regulated by NF-κB, in gastrointestinal HT-29 cells (Bergonzelli et al., 2006). A similar increase in pro-inflammatory CXCL8 production was seen by elongation factor Tu, isolated from the same strain (Granato et al., 2004). In our study, elongation factor Tu was absent in the 30-kDa fraction and found in low levels in the 100-kDa fraction, whereas a strong expression was seen in the immunoactive 50-kDa fraction. One of the identified proteins with predicted secretion as determined by both psortb and signalp was the NLP/P60 family protein. This protein contains an NLPC/P60 domain, which is also shared with the cytoprotective p40 and p75 proteins secreted by L. rhamnosus GG (Yan et al., 2007). The NLP/P60 family protein from L. rhamnosus GR‑1 shared 47% hom*ology with p40 (GI: 258507026) and 41% to p75 (GI: 259648676) from L. rhamnosus GG (determined by NCBI protein blast), enough to motivate characterization of the eventual cytoprotective effects of these proteins on urothelial cells as well. Moreover, the NLP/P60 protein is a peptidoglycan peptidase (Anantharaman & Aravind, 2003) that could facilitate the release of peptidoglycan from bacteria, thus increasing the immunological response. Although microorganism-derived nucleases are known to be immunostimulatory, through production of short immunostimulatory oligodeoxynucleotides from DNA and expressed by L. rhamnosus (Iliev et al., 2008), none of the identified proteins matched a protein with known nuclease activity. Additionally, it should be noted that not all immunomodulatory properties of lactobacilli are because of the activities of proteins. Apart from bacterial metabolites, polyphosphate from a probiotic Lactobacillus brevis strain was recently demonstrated to be responsible for activation of heat shock protein (hsp) 27 and mitogen-activated protein kinase p38 in mouse small intestine (Segawa et al., 2011).

Most studies on probiotic immunomodulation have explored the anti-inflammatory consequences of probiotics. It is, however, known that many UPEC are equipped with a number of different immune-evasive and immune-suppressive traits that aid them in colonization of the bladder and establishment of infection, including subversion of NF-κB activation (Cirl et al., 2008), enough to merit studies on pro-inflammatory actions of probiotics. Moreover, low levels of Toll-like receptor expression and the chemokine receptor CXCR1 are linked to high UTI susceptibility in humans (Frendéus et al., 2000; Yin et al., 2010), supporting the importance of an appropriately mounted immune response during UTI.

Lactobacilli are generally regarded as safe for consumption, although such administration limits their use to already colonized mucosal surfaces such as the gastrointestinal and genital tract, while they are not routinely administered to sterile body compartments such as the bladder. Unfortunately, there are currently no comprehensive studies on normal ascension of vagin*l lactobacilli into the ureter, a strategic locus for inhibiting entrance of uropathogens. However, a healthy urinary bladder is considered to be sterile and not suitable for administration of live probiotic microbes. Nonetheless, instillation of L. rhamnosus GG into murine bladders did not lead to persistent induction of cytokine or chemokine genes, although accompanied by a low influx of macrophages (Seow et al., 2008).

The proteins identified from L. rhamnosus GR‑1 in this study may prove valuable for in vivo treatment of a UTI, if using live bacteria is not an option because of adverse effects. We believe that by strengthening the normal immune response, the recognition of UPEC by epithelial cells is ensured, and proper activation of NF-κB facilitates the clearance of pathogens from the urinary system.

Acknowledgements

The authors would like to thank Prof Gregor Reid for providing the probiotic L. rhamnosus GR-1 and Prof Robert Brummer for critical reading of the manuscript. The work was supported by the Knowledge Foundation (PEO), Magnus Bergvalls Foundation (JJ), Sparbanksstiftelsen Nya (JJ), and Carl Tryggers Foundation (NS), Sweden.

References

Author notes

Present address: Hazem Khalaf, Division of Clinical Medicine, School of Health and Medical Sciences, OÖrebro University, Örebro, Sweden

© 2012 Federation of European Microbiological Societies. Published by Blackwell Publishing Ltd. All rights reserved