Potentiation of Polarized Intestinal Caco-2 Cell Responsiveness to Probiotics Complexed with Secretory IgA

Abstract

The precise mechanisms underlying the interaction between intestinal bacteria and the host epithelium lead to multiple consequences that remain poorly understood at the molecular level. Deciphering such events can provide valuable information as to the mode of action of commensal and probiotic microorganisms in the gastrointestinal environment. Potential roles of such microorganisms along the privileged target represented by the mucosal immune system include maturation prior, during and after weaning, and the reduction of inflammatory reactions in pathogenic conditions. Using human intestinal epithelial Caco-2 cell grown as polarized monolayers, we found that association of a Lactobacillus or a Bifidobacterium with nonspecific secretory IgA (SIgA) enhanced probiotic adhesion by a factor of 3.4-fold or more. Bacteria alone or in complex with SIgA reinforced transepithelial electrical resistance, a phenomenon coupled with increased phosphorylation of tight junction proteins zonula occludens-1 and occludin. In contrast, association with SIgA resulted in both enhanced level of nuclear translocation of NF-κB and production of epithelial polymeric Ig receptor as compared with bacteria alone. Moreover, thymic stromal lymphopoietin production was increased upon exposure to bacteria and further enhanced with SIgA-based complexes, whereas the level of pro-inflammatory epithelial cell mediators remained unaffected. Interestingly, SIgA-mediated potentiation of the Caco-2 cell responsiveness to the two probiotics tested involved Fab-independent interaction with the bacteria. These findings add to the multiple functions of SIgA and underscore a novel role of the antibody in interaction with intestinal bacteria.

Introduction

The gastrointestinal lymphoid tissue plays an important role in controlling transepithelial passage of bacteria across the intestinal mucosa by synthesizing more antibody (Ab)² molecules than any other lymphoid tissue. This is principally initiated in organized lymphoid tissues referred to as Peyer's patches, where antigen sampling by microfold (M) cells, processing/presentation by dendritic cells, and subsequent T cell activation leads ultimately to the development of immunoglobulin A (IgA)-producing cells in local and distant effector mucosal sites and glands (1–3). Antigen sampling by lamina propria dendritic cells and T cell-independent B cell maturation also contribute to mucosal IgA production (4). This class of Ab is the principal immunoglobulin produced in the gastrointestinal tract, and hence a key player to efficient humoral mucosal immunity, in particular for the maintenance of the integrity of the epithelial barrier.

In mucosal secretions, SIgA exists as a complex made of polymeric IgA (pIgA) in association with secretory component (SC), a highly glycosylated protein resulting from the cleavage of the polymeric Ig receptor (pIgR) that ensures the transport of the immunoglobulin across the epithelium. In the gut, bound SC contributes to both the stability of the Ab toward proteases, as well as its proper anchoring to mucus, conferring optimal biologic activity (5, 6). In addition to its function in immune exclusion preventing pathogen translocation across the epithelial barrier, abundant maternal SIgA Ab (up to 12 g/liter in colostrum and 1 g/liter in human milk (7)), by coating gut commensal bacteria in developing neonatal intestine, is thought to modulate initial exposure to the immature immune system (8).

The postnatal colonization of the gastrointestinal tract with commensals provides bacterial stimuli that are crucial for the functional development and homeostasis of the major compartments of the gastro-intestinal lymphoid tissue (9, 10). Studies with various strains of mostly lactic acid probiotics, which are normal residents of the intestinal microbiota (11, 12), represent a valuable approach to decipher the elaborated mechanisms involved in their dynamic interaction with intestinal epithelial cells. The relationship between intestinal microorganisms and SIgA led us to postulate that complexes of the two might cooperate to communicate with the interface represented by epithelial cells.

To address this hypothesis, we used an in vitro system consisting of polarized Caco-2 cell monolayers mimicking the intestinal epithelium, and exposed this latter apically to two probiotic strains Lactobacillus rhamnosus (LPR) or Bifidobacterium lactis (BL), either alone or in association with SIgA. Important features including bacterial adhesion, NF-κB activation, pIgR and thymic stromal lymphopoietin production were all potentiated upon complex formation with SIgA in a Fab-independent fashion. Tightness of the epithelial monolayer was promoted identically by the bacteria alone or in complex with SIgA. This indicates that the combination of probiotics and SIgA has a qualitative impact on epithelial cell function, and that this effect cannot simply be attributed to quantitative differences in bacterial binding. Our studies underscore for the first time the involvement of the Ab in mechanisms associated with epithelium responsiveness to probiotics and maintenance of the integrity of the intestinal barrier.

EXPERIMENTAL PROCEDURES

Caco-2 Cell Culture Conditions

Human colonic adenocarcinoma epithelial Caco-2 cells (HTB 37, American Type Tissue Collection) were grown at 37 °C in complete DMEM (C-DMEM) consisting of DMEM-Glutamax (Sigma) supplemented with 10% FBS (Sigma), 1% non essential amino acids (Sigma), 10 mm HEPES (pH 7.0, Sigma), 0.1% transferrin (Invitrogen AG, Basel, Switzerland) and 1% streptomycin/penicillin (Sigma), and used between passages 23 and 37. Cells cultivated to 80% confluency were seeded on Transwell filters (diameter, 12 mm; pore size, 0.4 μm; Corning Costar, Cambridge, MA) at a density of 0.8 × 10⁵ cells/cm². The culture medium was changed every 2 days. The formation of a polarized Caco-2 cell monolayer at week 3 was established by morphology and monitoring of the transepithelial electrical resistance (TER; ohms × cm²) using a Millicell-ERS apparatus (Millipore, Bedford, MA). TER values of well-differentiated monolayers were in the range of 380–450 ohms × cm² (13).

Bacterial Strains

The bacterial strains L. rhamnosus CGMCC 1.3724 and B. lactis CNCM I-3446 were obtained from the China general microbiological culture collection and Pasteur Institute microbiological culture collection, respectively. Ready to use vials of freeze-dried powders of live bacteria coded L. rhamnosus NCC4007 (LPR) and B. lactis NCC2818 (BL) were provided by Nestlé Research Center (Lausanne, Switzerland). LPR was cultured from frozen stocks overnight at 37 °C in Mann-Rogosa-Sharpe broth (Difco Laboratories, Detroit, MI) without agitation, and BL was grown in the same conditions in medium complemented with 0.05% l-cysteine. The commensal strain Escherichia coli strains Nissle 1917 and the cloning strain TG1 was grown overnight at 37 °C in Luria-Bertani broth (Difco) with agitation. Assessment of colony-forming units (CFU) per milliliter resulting from such culture conditions was carried out by plating of successive dilutions of the overnight incubations, or by measurement of the optical density at 600 nm (14).

Proteins and Ab

Purified mouse SC was produced and prepared as described in Crottet et al. (15). Hybridoma IgAC5 cells were grown at 37 °C in Integra Biosciences Celline-350 cartridges (Vitaris, Baar, Switzerland) in RPMI 1640 medium supplemented with 10% FBS, 2 mm glutamine, 2 mm sodium pyruvate, 10 mm HEPES (pH 7.0), 0.1 mm folic acid, and 100 units/ml penicillin and 100 μg/ml streptomycin, and purified polymeric IgA was recovered from crude supernatant by sizing chromatography onto Sephacryl S-300 columns (GE Healthcare, Otelfingen, Switzerland (16)). SIgA Ab molecules were reconstituted from equimolar of polymeric IgA and SC as published (5).

Association of SIgA with Bacteria

Overnight bacterial cultures were washed twice in phosphate-buffered saline (PBS: 116.3 mm NaCl, 10.4 mm Na₂HPO₄, 3.2 mm KH₂PO₄ (pH 7.4)) resuspended in PBS, and the number of bacteria was determined as indicated above. 2 × 10⁷ bacteria were mixed with 1 μg of SIgA in a final volume of 50 μl of PBS and incubated at ambient temperature for 30 min prior to use. For visualization of the Fab-independent association between bacteria and SIgA, the proteins were labeled with indocyanin-3 (Cy3) as published (17). Following washes in PBS, mixtures were laid into 8-well multitest slides, fixed with 2% paraformaldehyde in PBS for 25 min, washed, and mounted postcoating with Vectashield (Vector Laboratories, Burlingame, CA). Observations were performed using a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss, Jena, Germany). Images were taken with a 63× objective and processed using the microscope-related software (Zeiss). Quantification of bacterial coating by antibodies was performed with ImageJ software 1.41 (NIH). We measured on each series of pictures the area covered by bacteria (bacterial area) using the differential interference contrast (DIC) channel. In parallel, in each area associated with bacteria, we quantify the area containing IgA linked pixels with a fluorescence intensity superior to 15 units (fluorescent area) defined in our experimental settings as the background signal (signal range extends from 0 to 256 units using transformation into 8 bit gray scale). The following ratio was used to quantify the percentage of bacterial surface covered with fluorescent Ab molecules: 100 × (fluorescent area)/(bacterial area).

Exposure of Polarized Caco-2 Cell Monolayers to Bacteria Alone or in Complexes with SIgA

The apical compartment of Caco-2 cell monolayers was washed twice with PBS, and C-DMEM was replaced with the same medium lacking FBS and antibiotics (DMEM-A⁻). After 2 h of incubation, bacteria, SIgA, or SIgA-bacteria mixes in DMEM-A⁻ were added the co-culture was kept overnight (usually 16 h) prior to be subjected to multiple analyses as described below. In control experiments, we found that bacterial growth rate was not affected by the Ab.

Bacterial Adhesion Assay

Transwell supports carrying polarized Caco-2 cell monolayers were washed five times with PBS. The Caco-2 cells were then detached by a 5-min incubation at 37 °C in the presence of trypsin/EDTA (Sigma) added to the apical (0.5 ml) and the basolateral (1.5 ml) surfaces. The adherent bacteria were dispersed by vigorous pipeting, serial dilutions (10⁻² to 10⁻⁵) were applied onto triplicate agar plates cast in Mann-Rogosa-Sharpe or Luria-Bertani broths, and CFU were determined after overnight incubation at 37 °C.

Whole Cell Lysates and Analysis of Caco-2 Cell Proteins

At the times indicated, Caco-2 cells grown in Transwell membranes were washed twice with PBS, prior to incubation for 5 min in 1 ml of hypotonic buffer (20 mm Tris-HCl, pH 7.5, 2 mm MgCl₂). Swollen cells were lysed by up-and-down pipetting in 150 μl of hypotonic buffer complemented with 1% (w/v) Triton X-100 (Pierce Chemical), Complete Protease Inhibitor Mixture (Roche Applied Science, Rotkreuz, Switzerland), 0.15 m NaCl, 0.1 mm EDTA, 0.1 mm sodium orthovanadate, 0.1 mm sodium fluoride, 2 μg/ml DNase 1, and 1 mm leupeptin. The cell lysate was transferred to a fresh siliconized 1.5-ml tube, and cleared by centrifugation at 13,000 × g. An aliquot was removed for protein determination using the bicinchoninic acid procedure (Pierce Chemical). Aliquots of the clarified lysate were stored at −20 °C prior to use.

For immunoprecipitation, whole Caco-2 cell extracts were pre-cleared by incubation with protein G-Sepharose beads (GE Healthcare) for 2 h at 4 °C. 50 μl of cleared whole Caco-2 cell extracts were incubated with 1/333 dilution of the specific Ab (anti-ZO-1 and anti-occludin from Santa Cruz Biotechnology) overnight at 4 °C. Protein G-Sepharose beads were added and incubated for an additional 1 h of incubation at 4 °C, and the immunocomplexes were washed with TENT buffer (50 mm Tris-HCl, pH 7.5, 5 mm EDTA, pH 8.0, 150 mm NaCl, 1% Triton X-100). SDS-PAGE was performed according to the Laemmli procedure using 10% polyacrylamide gels run in Mini Protean II gel apparatus (Bio-Rad). Separated proteins from whole Caco-2 cell lysates were electrotransferred onto polyvinylenedifluoride membranes that were saturated with 5% nonfat dry milk in PBS-Tween 20 (PBS-T). Tight junction proteins ZO-1 and occludin were probed with 1/500 dilution of anti-ZO-1 and anti-occludin, or 1/1000 dilution of anti-phosphotyrosine and anti-phosphoserine (Transduction Laboratories, Lexington, Kentucky) in 0.5% milk in PBS-T for 1 h at ambient temperature. After washing, the membranes were incubated with corresponding horseradish peroxidase-linked secondary Ab (anti-mouse IgG, anti-rabbit IgG, as appropriate) for 1 h at room temperature. Following final washes, membranes were incubated with enhanced chemiluminescence reagents (Interchim, Montluçon, France) prior to exposure to photographic films.

Polymeric Ig receptor in whole Caco-2 cell extracts was detected by immunoblotting using rabbit antiserum against the human protein (1/2000 (18)). The level of pIgR was normalized to β-actin in the same sample incubated with 1/2000 dilution of a specific antiserum (Alpha Diagnostic, San Antonio, TX). Quantification of human SC was carried out by ELISA as described (18).

Analysis of NF-κB Nuclear Translocation and IκBα Evaluation

Preparation of Caco-2 cell small-scale nuclear extracts and use in EMSA for the detection of NF-κB was carried out as described in Cottet et al. (13). The radioactivity associated with retarded oligonucleotide-NF-κB complexes was quantified on an Instant Imager reader (Packard, Palo Alto, CA). Members of the NF-κB family present in the nucleus from Caco-2 cells were identified by immunoblotting with rabbit antisera (Santa Cruz Biotechnology, Santa Cruz, CA; 1/500 dilution) directed against the p50 or p65 subunits. Cytoplasmic extracts containing IκBα were obtained according to Cottet et al. (13), and the presence of the protein was similarly assessed by Western blot with a specific antiserum (Santa Cruz Biotechnology). After binding of appropriate secondary Ab, membranes were processed for detection by chemiluminescence.

Cytokine/Chemokine ELISA

CXCL-8 (interleukin (IL)-8), thymic stromal lymphopoietin (TSLP), CCL-5 (regulated on activation normal T cell expressed and secreted: RANTES), CCL-2 (monocyte chemoattractant protein (MCP)-1) possibly released in the basolateral compartment of polarized Caco-2 cells were measured by ELISA using commercial kits for human proteins (Biolegend, San Diego, CA), and expressed as picograms per ml of culture medium. Data are duplicates of 2–3 independent experiments.

Statistical Analysis

Statistical significance was determined using the two-tailed nonparametric Mann-Whitney U test. Standard error means and p values were calculated using the Prism5 application (GraphPad, San Diego, CA), and the limit of significance was set at p ≤ 0.05.

RESULTS

Adhesion of LPR Alone, BL Alone, or in Complexes with SIgA to Polarized Caco-2 Cell Monolayers

When the number of adherent LPR and BL bacteria per 100 Caco-2 cells was plotted against the concentration of added bacteria, a plateau-type binding pattern was obtained. A similar adhesion curve was produced with the E. coli commensal strain Nissle 1917 whereas virtually no binding was observed with non-commensal E. coli TG1 at all concentrations tested (Fig. 1A). In addition, differences in adhesion properties were detected between LPR, BL, and Nissle 1917 indicating that selective interaction with the apical surface of Caco-2 cells mimicking the intestinal barrier was indeed occurring in this simplified system. Time course experiments showed that there is no dependence of the final concentration of bacteria on adhesion.³ Because intestinal bacteria have been shown to be coated with a mixture of specific and nonspecific, “natural” SIgA (19, 20), we sought to examine whether the combination of LPR and BL with SIgA might have influenced the adhesion properties of the bacteria. We found that association with SIgA enhanced the capacity to adhere to polarized Caco-2 monolayers by a factor of 3.4- for LPR, and 3.9-fold for BL, respectively (Fig. 1B). Noteworthy, incubation of 2 × 10⁷ bacteria in the presence of SIgA leads to more adhesion to Caco-2 cells than when 1 × 10⁸ bacteria are added alone. We concluded that SIgA potentiated the capacity of LPR and BL to adhere to polarized Caco-2 cell monolayers.

FIGURE 1.

A, adhesion of L. rhamnosus LPR, B. lactis BL, and E. coli strains Nissle 1917 and TG1 to polarized intestinal Caco-2 cell monolayers. The number of bound bacteria per 100 Caco-2 cells is presented as a function of the concentration of freshly cultured bacteria added (CFU per ml). Bacterial counts were determined by plating of serial dilutions. Black bars, LPR; gray bars, BL; striped bars, Nissle 1917; white bars, E. coli TG 1. B, same as in A using 2 × 10⁷ bacteria alone (white bars) or associated with SIgA (black bars) prior to incubation with Caco-2 cell monolayers. Data were obtained from four independent experiments performed in triplicates. Significant statistical differences are indicated above the lanes.

Fab-independent Interaction of SIgA with LPR and BL

Based on previous reports showing of intestinal bacteria with endogenous SIgA (19, 20), it is conceivable that the SIgA Ab molecule we used, whose specificity is directed toward S. flexneri LPS serotype 5a, can however associate with both LPR and BL in a Fab-independent fashion, thus contributing to improved adhesion. We examined this possibility by combining bacteria with red Cy3-labeled SIgA. Laser-scanning confocal microscopy pictures show abundant red staining covering strings of bacteria typical of the two strains assessed, hence indicating Fab-independent association of the Ab and bacterium partners (Fig. 2). Using ImageJ software (see Experimental procedures), quantification of fluorescent SIgA associated with bacteria indicated that the surface of all bacteria were coated with up to 95% (LPR) and 78% (BL) by Ab molecules. This demonstrated that this was in the form of a complex with SIgA that cellular adhesion of the bacterium was improved.

FIGURE 2.

Laser-scanning confocal microscope imaging of the Fab-independent association of SIgAC5:Cy3 with LPR and BL. Bacteria are visualized by differential interference contrast (DIC, Nomarski), and bound Ab shows as co-localizing red fluorescent spots on the surface of bacteria. Formation of small strings is typical of the morphology of the microorganisms tested. One representative field obtained from 10 different observations after analysis of five different slides is shown. Bars: 5 μm.

Increased TER after Overnight Contact between Polarized Caco-2 Cell Monolayers and LPR Alone, BL Alone, or in Complexes with SIgA

TER measures the permeability attained by epithelial cell lines grown on synthetic polycarbonate membranes after they have differentiated as polarized monolayers mimicking the intestinal barrier. Live commensals and probiotics modulate TER (14, 21), and this was used in the present study to investigate in vitro the possible effects of LPR or BL alone, and in complexes with SIgA, on Caco-2 cell monolayers. Functional polarity, i.e. formation of tight junctions observed by ZO-1 staining in laser scanning confocal microscope images (data not shown), was established at >350 ohms × cm² in the model used. Exposure to either LPR or BL for 6 h or more led to a reproducible and significant 18–25% increase in TER (0.01 < p < 0.005), an effect that reached a plateau at 15 h (Fig. 3A). Association of SIgA with bacteria did not further modulate the TER (Fig. 3A), although binding to polarized Caco-2 monolayers was potentiated in the presence of the Ab (Fig. 1). The same increase in TER was obtained when using 5 times less or 5 times more bacteria (data not shown), indicating that despite of SIgA-mediated improved adhesion, the Ab had no inhibitory/activatory effect on the properties of the bacteria in acting on TER.

FIGURE 3.

A, TER of epithelial Caco-2 cell monolayers exposed to 2 × 10⁷ LPR alone, 2 × 10⁷ BL alone, and in complexes with SIgA, determined at five time points. Description of symbols is given in the inset. SIgA used alone serve as control of the stability of the Caco-2 cell monolayer TER. Compilation of data from four independent experiments performed in triplicates is shown. Codes for symbols: light blue, LPR; dark blue, LPR + SIgA; pink, BL; purple, BL + SIgA; black, SIgA alone. B, overnight exposure of polarized Caco-2 cell monolayers to 2 × 10⁷ LPR or LPR-SIgA complexes increases phosphorylation of the tight junction proteins occludin and ZO-1. Cell lysates were immunoprecipitated (IP) with specific Ab, and detected by Western blot (Wb) with anti-phosphotyrosine, anti-phosphoserine, and Ab to occludin and ZO-1. C, quantification of phosphorylated occludin (P-occludin) and phosphorylated ZO-1 (P-ZO-1) after exposure of Caco-2 cells to 2 × 10⁷ LPR, 2 × 10⁷ BL, and in complexes with SIgA. Data were obtained from three independent experiments performed in triplicates. Comparative statistical analysis with the bar marked plain medium yielded p values < 0.002 for all experimental groups.

Raise in TER Is Associated with Increased Occludin and ZO-1 Phosphorylation upon Incubation with LPR Alone, BL Alone, or in Complexes with SIgA

Increased TER implies associated improved tightness of the junction-connecting cells involved in monolayer formation. Analysis of Caco-2 cell whole extracts 16 h post-exposure to bacteria revealed no change in the content of tight junction proteins occludin and ZO-1, despite the 18–25% increase in TER previously observed at this time point. However, the degree of phosphorylation of the two proteins was consistently increased upon incubation with LPR (Fig. 3, B and C) and BL (Fig. 3C) compared with resting polarized Caco-2 cell monolayers. Immunodetection of occludin and ZO-1 showed that very similar amounts of protein in Caco-2 cell extract immunoprecipitates were loaded (Fig. 3B). Densitometric analysis of raw data presented in Fig. 3B indicated that the degree of phosphorylation increased 2-fold (occludin) and up to 3-fold (ZO-1) after exposure to the bacteria alone, and that this value remained the same when SIgA was combined in the assay (Fig. 3C). This correlated at the molecular level with the contribution of probiotics in the reinforcement of the epithelial barrier and the absence of involvement of SIgA bound to bacteria in the raise in TER (Fig. 3A).

Effect of LPR Alone, BL Alone, or in Complexes with SIgA on NF-κB Activation in Polarized Caco-2 Epithelial Cell Monolayers

Activation of the transcription factor NF-κB is a recognized marker of the onset of multiple signaling pathways in many cell lines and tissues (22). In comparison with resting cells, exposure of polarized Caco-2 cells to LPR or BL led to the appearance of more shifted complexes when nuclear extracts were mixed with a consensus NF-κB DNA probe and examined by electrophoretic mobility shift assay (Fig. 4A). However, in comparison with the pathogen S. flexneri inducing strong nuclear translocation of NF-κB, both bacteria maintained low level of inducible complex formation under all conditions tested. In association with SIgA, higher levels of NF-κB-probe complexes were found in Caco-2 cell nuclear extracts as compared with bacteria alone, although the apparent degree of absolute nuclear translocation was again much lower than with S. flexneri-exposed Caco-2 cells. Although SIgA alone displayed the capacity to upregulate the formation of NF-κB-based complexes in EMSA, quantification of fold-induction values indicated a synergic, rather than additive, effect of the Ab (Fig. 4A, top of gels). We conclude that in contrast to enteropathogenic S. flexneri, exposure to LPR and BL appears to induce limited DNA binding by the transcription factor in nuclear extracts, a fine-tune effect that can be further modulated upon association with SIgA.

FIGURE 4.

Effect of LPR and BL818 along the NF-κB activation pathway. A, electrophoretic mobility shift assay performed with nuclear extracts from polarized Caco-2 cell monolayers incubated for 16 h with 2 × 10⁷ LPR alone, 2 × 10⁷ BL alone, or in association with SIgA, as indicated at the bottom of the lanes. Comp corresponds to a 20-fold molar excess of unlabeled NF-κB oligoNT probe to identify the specific retarded complex. NF-κB-probe complexes obtained when using nuclear extracts from cells stimulated with 2 × 10⁷ enteropathogic S. flexneri are shown for comparison. B, immunoblotting of IκBα in cytoplasmic extracts from Caco-2 cells incubated as in A with BL and Ab, or S. flexneri for comparison (upper panel). Nuclear translocation of NF-κB subunits p50 and p65/relA induced by the contact of Caco-2 cells with 2 × 10⁷ BL alone and in complex with SIgA, or 2 × 10⁷ S. flexneri for comparison (lower panels). Immunoblotting was carried out on Caco-2 cell nuclear extracts. Identical amounts of nuclear and cellular extracts based on protein concentration were used for each set of experiments. Panels are representative of one individual triplicate experiment performed three times.

Exposure to BL Prevents Degradation of the NF-κB Inhibitor IκBα in Polarized Caco-2 Epithelial Cell Monolayers

To investigate at which level in the pathway of NF-κB activation the microbial-epithelial interaction exerts its effect, the regulation of the NF-κB inhibitory molecule IκBα was examined in cytoplasmic cell extracts (Fig. 4B, upper panel). Constant amounts of IκBα were detected by immunodetection in lysates of Caco-2 cells incubated with BL, BL-SIgA, or SIgA. This contrasted with the substantially reduced intensity of IκBα signal resulting from the analysis of Caco-2 cell lysates exposed to S. flexneri used as a positive trigger of the NF-κB pathway in epithelial cells. The inverse correlation between the levels of IκBα and NF-κB-DNA probe complexes was indicative of different degrees of NF-κB activation that could be finely dissected as a function of the bacterium or complexes tested. Occurrence of limited nuclear translocation of NF-κB following incubation with BL was further demonstrated upon direct analysis of the NF-κB subunits p50 and p65/relA (Fig. 4B, compare with the lane S. flexneri). The sum of these data indicated that the SIgA-BL complex (and LPR-SIgA, data not shown) adhering to the apical surface of polarized Caco-2 cell monolayers was moderately activatory of the signaling pathway(s) leading to NF-κB nuclear translocation. Reinforced interaction mediated by SIgA increased further NF-κB nuclear translocation, suggesting that “sensing” by Caco-2 cells is partly controlled by the presence of the Ab.

LPR and BL in Complexes with SIgA Induce pIgR Up-regulation in Polarized Caco-2 Epithelial Cell Monolayers.

We then investigated the possible impact of the exposure of Caco-2 cells to LPR, BL, or as SIgA-based complexes, on pIgR production known to be involved in mucosal homeostasis (23). Cell extracts were prepared after overnight incubation and lysates were analyzed by immunodetection (Fig. 5A). The pIgR protein (120 kDa) was partly converted into SC (85 kDa), a feature that reflects the rapid cleavage of pIgR at the cell surface (24). For standardization purposes, detection of β-actin protein was carried out in parallel. Densitometric analysis relative to β-actin revealed that, in comparison with medium, BL alone showed a slight beneficial effect on pIgR production (p = 0.041), in contrast to LPR (p = 0.080) (Fig. 5B). Comparison with bacteria alone showed that both LPR-SIgA and BL-SIgA immune complexes increased production of the pIgR/SC protein in a significant manner (p = 0.0019 and p = 0.0011, respectively) (Fig. 5B). SIgA alone did not have any effect on pIgR production, whereas enteropathogenic S. flexneri used as a positive control promoted production of the pIgR protein by a factor of almost 4 (Fig. 5B). The absolute quantity of pIgR in cell lysates was measured by ELISA at different time points (Fig. 5C). We found that BL in combination with SIgA yielded 6.1 ng of cell-associated pIgR/mg of total protein, while SIgA-LPR complex led to the production of 2.9 ng of pIgR/mg of total protein. Production did not change in a significant manner between 16 and 24 h of incubation. S. flexneri used as a positive control led to production of up to 10 ng pIgR/mg of total protein. These results indicate that the interaction between a probiotic-SIgA complex and epithelial Caco-2 cells prompts these latter to synthesize more pIgR involved in mucosal defense.

FIGURE 5.

A, Western blot analysis of lysates recovered from polarized Caco-2 cell monolayers exposed to 2 × 10⁷ LPR alone, 2 × 10⁷ BL alone, and in the form of complexes with SIgA, as indicated on the top of the lanes. Detection was performed with rabbit anti-serum against SC recognizing both SC and the precursor pIgR. Signals caused by β-actin were obtained using a specific antiserum. B, densitometric analysis of two independent experiments was carried out with standardization based on the β-actin signal. The lane content is indicated below the plot. Significant statistical differences are indicated above the lanes. C, quantification of pIgR and converted SC was assessed by ELISA and is reported as a function of the amount measured expressed in ng per mg of protein in Caco-2 cell lysates. Data were gathered from three independent experiments performed in triplicates. Codes for symbols: light blue, LPR; dark blue, LPR + SIgA; pink, BL; purple, BL + SIgA; black, SIgA alone; circle, plain medium; diamond, S. flexneri.

LPR Alone, BL Alone, or in Complexes with SIgA Promote TSLP Up-regulation in Polarized Caco-2 Epithelial Cell Monolayers

Production of chemokines in the basolateral compartment of polarized Caco-2 cell monolayers was assessed by ELISA after overnight incubation with bacteria in the presence or absence of SIgA (Table 1). Upon incubation with either bacterium alone or in complex with SIgA, the level of measured CXCL-8 involved in recruitment of monocytes/neutrophils in epithelia was close to that obtained using plain cells, and substantially below that obtained after incubation with pro-inflammatory S. flexneri (p < 0.0002). Secretion of TSLP known to play an important role in maintaining an intestinal noninflammatory environment (25) favorable to the development of semi-mature dendritic cells (26–28) was increased after exposure to LPR (3-fold) and BL (4-fold) in comparison with control cells (p < 0.008). The release of TSLP was further enhanced (>2-fold) when the bacteria were added in combination with SIgA (p < 0.002), most likely a consequence of the better adhesion observed previously (Fig. 1). Production of both mediators of neutrophil and monocyte recruitment CCL-5 (RANTES) and CCL-2 (MCP-1) by cells exposed to LPR and BL, or SIgA-based complexes thereof, were below the level of detection, in contrast to high levels obtained after incubation with S. flexneri. Altogether, these results demonstrate a specific non-inflammatory type of response of Caco-2 cells in interaction with probiotics and SIgA-based complexes.

TABLE 1.

Analysis of chemokines released basolaterally by polarized Caco-2 cell monolayers exposed to LPR alone, BL alone, or in complexes with SlgA, with exposure to S. flexneri used as a positive control. Values are expressed in pg/ml ± S.D. of chemokine measured in the basolateral compartment of Transwell filters. CXCL-8, interleukin-8; TSLP, thymic stromal lymphopoietin; CCL-5, RANTES; CCL-2, MCP-1. Data were gathered from three independent experiments performed in triplicate.

	CXCL-8		TSLP		CCL-5		CCL-2
	−SlgA	+SlgA	−SlgA	+SlgA	−SlgA	+SlgA	−SlgA	+SlgA
Caco-2	12 ± 4	12 ± 3	19 ± 3	13 ± 6	<2	<2	<2	<2
Caco-2 + LPR	19 ± 1	20 ± 3	62 ± 6a	144 ± 26b	<2	<2	<2	<2
Caco-2 + BL	17 ± 9	18 ± 2	84 ± 7a	169 ± 28b	<2	<2	<2	<2
Caco-2 + S. flexneri	2000 ± 79	N.D.c	33 ± 11	N.D.	347 ± 23	N.D.	478 ± 57	N.D.

^a Significant statistical differences between cells exposed to medium or bacteria alone are highlighted in bold.

^b Significant statistical differences between cells exposed to the bacteria alone or as SlgA-based complexes are highlighted in bold.

^c N.D., not determined, as the SlgA used in the assays neutralizes S. flexneri.

DISCUSSION

The role of probiotic and commensal bacteria in the physiology of the gastrointestinal tract is incompletely understood, mostly because of the complex and intricate array of cellular and molecular partners involved. In particular, the mechanisms underlying the beneficial effects of probiotics, as well as the impact on epithelial maturation and communication with the mucosal immune system are still in need of investigation. Using polarized Caco-2 cell monolayers, we found that incubation with the probiotics LPR and BL led to modifications of several features including adhesion, permeability, and signaling events involved in NF-κB nuclear translocation, production of pIgR, and induction of immune mediators. The further novelty of our data resides in the demonstration that binding of nonspecific SIgA to bacteria potentiates their effect on selected events associated with adhesion and cell signaling, reflecting that different sensing pathways could be identified in the in vitro setting used inhere.

Preferential adhesion of LPR, BL, and Nissle 1917 probiotics strains to Caco-2 cell monolayer in comparison with E. coli TG1 is in agreement with previous studies using similar in vitro systems. Our data allowed us to demonstrate that SIgA increased adhesion to Caco-2 epithelial cells grown as polarized monolayers. Such an effect can find an explanation in the surface expression of epithelial CD71 (the transferrin receptor), which exhibits SIgA binding properties (29), as anchoring of the Ab through mucin cannot occur in Caco-2 cells unable to produce mucus. Human SIgA purified from colostrum, when combined with bacteria, displayed the same potentiating effect on adhesion to polarized CaCo-2 cells monolayers.³ Unexpectedly, enhanced adhesion did not translate into further increased TER, which ranged in between 18–25% for all conditions tested. This supports the hypothesis that SIgA-mediated enhanced adhesion of bacteria to Caco-2 cells does not necessarily improve permeability, as confirmed by the stability in TER observed upon incubation of Caco-2 cells with 5 times less or 5 times more bacteria. Consistent with this, phosphorylation of ZO-1 and occludin linked to epithelial tightness was the same for bacteria and SIgA-based complexes, suggesting that improved adhesion to Caco-2 cells might have consequences in other pathways, pathways we identified in this work.

Infection of epithelial cells with pathogenic microbes including S. flexneri induces rapid degradation of IκBα, resulting in the release of the NF-κB complex that translocates to the nucleus where it triggers transcription of a variety of genes required for immune responses (30). In contrast, exposure to LPR or BL only partially induced nuclear translocation of the transcription factor, an effect potentiated by SIgA. This indicates that exposure of epithelial cells to certain probiotic microorganisms such as LPR and BL maintains a low, if not basal, degree of NF-κB activation that may be instrumental for the maintenance of homeostasis (31). This is with keeping in mind that although a central regulator in gene expression, NF-κB acts in concert with a plethora of other transcription factors to promote optimal and cell-specific transcriptional control of multiple target genes. Consistent with this, induction of NF-κB after exposure to BL or SIgA-probiotic complexes did not lead to the production by Caco-2 cells of chemokines involved in the recruitment of pro-inflammatory cells. However, production of TSLP involved in maintaining an intestinal noninflammatory environment (25) was promoted by Caco-2 cells incubated with either LPR or BL, and significantly further enhanced when complexed with SIgA.

Efficient transport of IgA from the lamina propria into mucosal secretions is mediated by pIgR produced by epithelial cells (23). Up-regulation of intestinal pIgR mRNA expression has been reported in formely germ-free mice colonized with the commensal Bacteroides thetaiotamicron (32). In comparison with LPR or BL alone, we found that exposure of Caco-2 cell monolayers to SIgA-probiotic complexes triggered production of the pIgR protein. Increased production of pIgR that in turn would augment IgA translocation can be seen as an additional step in controlling the microbiota through a regulatory loop mechanism implying both innate and adaptive immunity.

Moreover, the data presented provide a mechanistic explanation to the underestimated function of natural SIgA in the regulation of the endogenous microbiota, both preweaning through maternal Ab, and after weaning through early, low affinity, Ab responses induced in the neonates (7). Indeed, by intervening in the complex interactions governing host-microbe cross-talk, SIgA may contribute to regulate epithelial cell responses to their environment. Maternal SIgA, by coating gut commensal bacteria in the neonatal intestine, would control initial exposure to the developing immune system and thus subsequent maturation. This may keep the level of gut-colonizing bacteria at bay until sufficient amounts of endogeneous SIgA can be produced by the neonate at the time of weaning.

“Natural” existence of bacterium-SIgA complexes anytime may contribute to maintain commensals in close association with the epithelium, and guarantee self-limiting control of microorganisms permanently colonizing the gut (33). We postulate that such a monitoring mechanism would promote optimal gut microbial colonization and ensures a dynamic and plastic interplay with epithelial cells. This adds to the already documented role of SIgA in limiting dissemination of microorganisms in the gastrointestinal lymphoid tissue (17, 34), and the immunomodulatory properties of SIgA in the intestine (35–37). Together, this will result in the onset of modulatory pathways crucial to maturation of both the epithelium and innate and adaptive immunity.

Our results further unravel that intimate association of microorganisms with SIgA potentiates the communication with epithelial monolayers to different degrees, as reflected by the observation that not all features examined were subject to changes, and that differences between Lactobacillus LPR and Bifidobacterium BL were indeed identified. The observation that the mucus-binding protein of another Lactobacillus species exhibits IgG and IgA binding activity might suggest a track to explore to better understand the complexity of interactions that ultimately lead to gut and mucus adhesion (38). Maintenance of intestinal integrity and proper functioning requires that harmonious microbial-epithelial interactions occur, and our novel data reveal the functional importance of SIgA in participating to this complex homeostatic balance.