|Home | About | Journals | Submit | Contact Us | Français|
Although the probiotic Escherichia coli strain Nissle 1917 has been proven to be efficacious for the treatment of inflammatory bowel diseases, the underlying mechanisms of action still remain elusive. The aim of the present study was to analyze the effects of E. coli Nissle 1917 on cell cycling and apoptosis of peripheral blood and lamina propria T cells (PBT and LPT, respectively). Anti-CD3-stimulated PBT and LPT were treated with E. coli Nissle 1917-conditioned medium (E. coli Nissle 1917-CM) or heat-inactivated E. coli Nissle 1917. Cyclin B1, DNA content, and caspase 3 expression were measured by flow cytometry to assess cell cycle kinetics and apoptosis. Protein levels of several cell cycle and apoptosis modulators were determined by immunoblotting, and cytokine profiles were determined by cytometric bead array. E. coli Nissle 1917-CM inhibits cell cycling and expansion of peripheral blood but not mucosal T cells. Bacterial lipoproteins mimicked the effect of E. coli Nissle 1917-CM; in contrast, heat-inactivated E. coli Nissle 1917, lipopolysaccharide, or CpG DNA did not alter PBT cell cycling. E. coli Nissle 1917-CM decreased cyclin D2, B1, and retinoblastoma protein expression, contributing to the reduction of T-cell proliferation. E. coli Nissle 1917 significantly inhibited the expression of interleukin-2 (IL-2), tumor necrosis factor α, and gamma interferon but increased IL-10 production in PBT. Using Toll-like receptor 2 (TLR-2) knockout mice, we further demonstrate that the inhibition of PBT proliferation by E. coli Nissle 1917-CM is TLR-2 dependent. The differential reaction of circulating and tissue-bound T cells towards E. coli Nissle 1917 may explain the beneficial effect of E. coli Nissle 1917 in intestinal inflammation. E. coli Nissle 1917 may downregulate the expansion of newly recruited T cells into the mucosa and limit intestinal inflammation, while already activated tissue-bound T cells may eliminate deleterious antigens in order to maintain immunological homeostasis.
Although the pathogenesis of inflammatory bowel disease (IBD) is not completely understood, the contributions of genetic and environmental factors are increasingly evident. Recently, a major role in the initiation and perpetuation of chronic IBD has been attributed to the luminal bacterial flora (40, 56). Interestingly, recent studies have demonstrated that bacterial products can directly activate T cells via Toll-like receptors (TLRs) even in the absence of antigen-presenting cells or costimulatory molecules (7, 32, 50), suggesting a link between the luminal bacterial flora and the immune system (19). Furthermore, in IBD, tolerance of the intestinal immune system against the physiological microflora is lost (14), and this breakdown is supposed to contribute to intestinal inflammation and may explain the beneficial effects of antibiotic therapy in IBD patients.
Escherichia coli strains play a pivotal role within the unique intestinal microecological system, which consists of an enormous variety and quantity of different microorganisms. These microorganisms can be found as physiological constituents of the intestinal microflora both in healthy individuals and under pathological conditions in the course of various gastrointestinal diseases. Several lines of evidence suggest that E. coli is involved in the pathogenesis of IBD (23). E. coli strains isolated from Crohn's disease lesions have been demonstrated to adhere to and to disrupt the intestinal barrier (12, 22). Furthermore, distinct E. coli genotypes are associated with chronic or early recurrent ileal lesions (43). In addition, E. coli-specific antibodies and the number of E. coli organisms in areas of intestinal inflammation are significantly elevated in IBD patients (70).
The probiotic E. coli strain Nissle 1917 has been reported to maintain remission of ulcerative colitis and pouchitis (35-37, 52) and to prevent colitis in different murine models of colitis (57). Although E. coli Nissle 1917 was originally isolated in 1917, the underlying mechanism of its beneficial effect in various intestinal diseases still remains elusive.
T cells play a major role in the pathogenesis of IBD (18). They must respond to antigens in a selective and balanced fashion that allows them to mount an effective response by progressing through the cell cycle, expanding, and finally undergoing apoptosis once the antigen has been cleared (4, 38, 39). This task is particularly challenging for T cells exposed to antigens that may be present in a variety of types and in large numbers, as in the gastrointestinal tract. In the intestinal mucosa T cells are required to maintain a state of immunological tolerance toward a myriad of dietary and bacterial antigens (15, 31, 49). To accomplish this critical assignment, intestinal T cells cycle differently from systemically circulating T cells, like peripheral blood T cells (PBT) (68), which encounter a completely different antigen repertoire (55). Whereas cell cycling and death are usually tightly balanced (63), many diseases are based on either uncontrolled cell cycling or impaired apoptosis, leading to unrestrained cell proliferation and, thus, cancer or autoimmune disorders such as IBD.
The aim of our study was to analyze the functional biological activities of E. coli Nissle 1917 on peripheral and mucosal T cells and thus to characterize the underlying mechanism of action. This was achieved by analyzing the effect of E. coli Nissle 1917 on peripheral blood and mucosal T-cell cycling, expansion, and apoptosis and by evaluating the expression of cell cycle regulatory molecules and cytokine secretion. Our results reveal that E. coli Nissle 1917 inhibits expansion of naive and memory PBT but not lamina propria T cells (LPT). This distinct reaction of circulating and tissue-bound T cells towards E. coli Nissle 1917 may explain the beneficial effect of probiotics in intestinal inflammation, where PBT are recruited from the circulation into the intestinal mucosa. Thus, E. coli Nissle 1917 may downregulate the expansion of newly recruited T cells into the mucosa and limit intestinal inflammation, whereas the insensitivity of LPT toward probiotic stimulation seems to avoid destabilization of the mucosal immune defense and may, therefore, contribute to maintenance of intestinal immunological homeostasis.
Human CD3 monoclonal antibody (MAb) (clone OKT3; Janssen-Cilag, Neuss, Germany) and CD2 MAb (clones T112 and T113; generously provided by Ellis Reinherz, Boston, Mass.) were used for T-cell activation. For stimulation of mouse PBT, functional grade purified anti-mouse CD3e (clone 145-2C11; eBioscience, San Diego, Calif.) was used. Fluorescein isothiocyanate (FITC)-conjugated anti-cyclin B1, bromodeoxyuridine (BrdU), CD28-FITC, α/β T-cell receptor (TCR α/β)-phycoerythrin (PE), TCR γ/δ-FITC, CD69-FITC, and PE-labeled anti-active caspase 3 were purchased from BD Pharmingen (Heidelberg, Germany). CD3-PE, CD45RO-PE, and CD45RA-FITC-labeled MAbs were obtained from DAKO (Hamburg, Germany). Secondary FITC-labeled goat anti-mouse was purchased from Biosource (Solingen, Germany). A Vybrant CFDA SE (carboxyfluorescein diacetate succinimidyl ester) cell tracer kit was obtained from Molecular Probes (Eugene, Oreg.). Propidium iodide (PI) was purchased from Calbiochem (Schwalbach, Germany). All protease and phosphatase inhibitors used for Western blotting were purchased from Sigma-Aldrich (Taufkirchen, Germany). The antibodies against human caspase 3, DNA fragmentation factor, poly(ADP-ribose) polymerase, retinoblastoma (Rb) protein, cyclin A, cyclin D2, p21, p27, and p53 were purchased from BD Pharmingen. Antibodies against human Bax, Bcl-2, cytochrome c, and E2F-1 were obtained from Santa Cruz Biotechnologies (Heidelberg, Germany). The cytometric bead array kit was purchased from BD Pharmingen. Highly purified lipopolysaccharide (LPS), generated from E. coli Nissle 1917, was provided by U. Zähringer (Forschungszentrum Borstel, Germany); synthetic bacterial lipoproteins (BLPs; Pam3-Cys-Ser-Lys-Lys-Lys-Lys-OH) and human serum albumin were generously provided by A. Zychlinsky (Max Planck Institute for Infection Biology, Department of Cellular Microbiology) (1); and immunostimulatory DNA sequences (5′-TsCsgsTsCsgsTsTsTsTsgsTsCsgsTsTsTsTsgsTsCsgsTsT-3′) were purchased from TIB MolBiol (Berlin, Germany). The Limulus amebocyte lysate for detecting gram-negative bacterial endotoxin was obtained from Cambrex (East Rutherford, N.J.).
Peripheral blood mononuclear cells (PBMC) from healthy volunteers were isolated from heparinized venous blood by using Ficoll-Hypaque density gradients. For isolation of PBT, PBMC were incubated for 30 min at 4°C with magnetically labeled CD19, CD14, and CD16 antibodies directed against B lymphocytes, monocytes, and neutrophils, respectively (Miltenyi Biotec Inc., Bergisch-Gladbach, Germany). T cells were then collected by using a magnetic cell sorting system (MACS; Miltenyi Biotec Inc.). LPT were isolated from surgical specimens obtained from patients admitted for bowel resection for malignant and nonmalignant conditions of the large bowel, including colon cancer and benign polyps as previously described (69). Briefly, the dissected intestinal mucosa was freed of mucus and epithelial cells in sequential washing steps with dithiothreitol and EDTA and digested overnight at 37°C with collagenase and DNase. Mononuclear cells were separated from the crude cell suspension by layering on a Ficoll-Hypaque density gradient. For LPT purification, macrophage-depleted lamina propria mononuclear cells were incubated for 30 min at 4°C with magnetically labeled beads as described above and collected by negative selection by using the MACS system. As assessed by flow cytometry, the purified PBT and LPT populations contained >99 and >92% CD3+ cells, respectively. For isolation of naive (CD45RA+) and memory (CD45RO+) PBT, PBMC were incubated for 30 min at 4°C with magnetically labeled CD14, CD16, and CD19 antibodies in combination with CD45RO antibodies to deplete for CD45RA+ PBT or with CD45RA antibodies to negatively select the CD45RO+ population, respectively. LPT were 91% CD45-RO+, and PBT were 54% naive T cells (CD45-RA+). The CD45RO-depleted PBT population was >95% CD45RA+ and less than 5% CD45RO+, while the CD45RA-depleted PBT population was >95% CD45RO+ and less than 5% CD45RA+. Further characterization of the cell populations revealed that freshly isolated PBT were 48% CD4+, 30% CD8+, and 0% CD69+. Freshly isolated LPT were 54% CD4+, 26% CD8+, and also 0% CD69+.
Cells were cultured in complete medium (RPMI 1640, 10% fetal calf serum [FCS], 1.5% HEPES buffer; Biowhittaker, Taufkirchen, Germany) containing 0, 2.5, 5, 10, 25, and 50% (vol/vol) E. coli Nissle 1917-conditioned medium (E. coli Nissle 1917-CM) in a humidified incubator containing 5% CO2, alone or in the presence of cross-linked plate-bound anti-CD3 MAb (OKT3; 10 μg/ml) or soluble anti-CD2 MAb pairs (T112 and T113; 1:1000).
E. coli Nissle 1917-CM was generated as described by Yan and Polk (79). Briefly, E. coli Nissle 1917-layered beads (provided by Ardeypharm GmbH) were incubated for 16 h at 37°C in Luria-Bertani broth. The bacterial culture was then harvested by centrifugation at 1,000 × g for 15 min. The supernatant was discarded, and the bacterial pellet washed twice in phosphate-buffered saline (PBS) and then resuspended in T-cell medium (RPMI, 10% FCS, 1.5% HEPES) without antibiotics. After 2 h at 37°C and 5% CO2, the culture was centrifuged at 1,000 × g, and the supernatant was recovered and sterile filtered through a 0.22-μm-pore-size syringe-driven filter. The same method was used to generate conditioned media of E. coli strains PZ 840, PZ 873, PZ 915, and DSM 498. In some experiments, E. coli Nissle 1917-CM was boiled for 1 h or frozen for 1 week at −80°C, and 10% FCS and 1.5% HEPES were added after the medium was chilled on ice or thawed. Additionally, heat-inactivated E. coli Nissle 1917 was prepared by resuspending bacteria in PBS and inactivating them for 1 h at 65°C. The bacteria were then washed twice and resuspended in complete medium. Proper inactivation was controlled by inoculation of heat-inactivated E. coli Nissle 1917 into Luria-Bertani broth and incubation for 24 h, demonstrating no bacterial growth. As determined by using a Limulus amebocyte lysate, the E. coli Nissle 1917-CM contained 1,300 ± 109 endotoxin units/ml.
Cell fluorescence was measured with a FACSCalibur (Becton Dickinson, Heidelberg, Germany) flow cytometer at excitation wavelengths of 488 and 633 nm with band-pass filters optimized for individual fluorochromes. Flow cytometry data were analyzed by using the CellQuest software program (BD Pharmingen).
Flow cytometry was performed after staining for DNA content and cyclin B1 essentially as previously described (68). Briefly, cells were washed twice with PBS, adjusted to 106 cells/sample and fixed in 90% methanol at −20°C. After fixation, cells were washed and incubated for 45 min at 4°C with a cyclin B1 FITC-conjugated MAb. After the final wash, cells were resupended in PBS and 5 μl of RNase (0.6 μg/ml, 30 to 60 Kunitz units; Sigma-Aldrich), incubated at 37°C for 15 min, and then chilled on ice. A total of 125 ml of PI (200 μg/ml) was added prior to analysis by flow cytometry. Each analysis was performed on at least 25,000 events.
Cell phenotype was analyzed by using CD45RA-PE-, CD45RO-FITC-, and CD3-PE-conjugated MAbs (all from Dako). The background level of immunofluorescence was determined by incubating cells with FITC- or PE-conjugated mouse immunoglobulin G. After a 30-min incubation on ice, cells were washed twice in 1% bovine serum albumin-PBS and fixed in 1% paraformaldehyde. Each analysis was performed on at least 10,000 events.
Analysis of cell division by dye dilution was performed by using a Vybrant CFDA SE cell tracer kit (Molecular Probes). Cells were washed twice with cold PBS, resuspended in PBS with 5 μM CFDA SE per 106 cells, and incubated for 15 min at 4°C in the dark. The staining was quenched by adding 5× cell culture medium containing 10% FCS. After staining, cells were cultured alone (unstimulated) or with soluble anti-CD2 MAb pairs (T112 and T113; 1:1000) or cross-linked plate-bound anti-CD3 MAb (OKT3; 10 μg/ml), each with CD28 (5 μg/ml) and interleukin-2 (IL-2; 20 U/ml). After 4 days, cells were harvested, washed twice in cold PBS, fixed with 1% paraformaldehyde, and analyzed by flow cytometry.
For the determination of S phase duration and potential doubling times, cells were grown for 3 days with or without the respective stimuli and then incubated for 60 min with 20 μM BrdU, which was then replaced by thymidine. At designated time points, cells were harvested and fixed with 90% methanol. The BrdU-labeled nuclei were then stained with an FITC-conjugated MAb against BrdU (BD Pharmingen). The nuclei were also stained with PI following the protocols described above. The denaturation of DNA, allowing antibody binding to the incorporated BrdU, was achieved by an acid treatment according to a previously described protocol (68). Mathematical analysis was performed according to the method of Begg et al. (5) and White et al. (78). The movement of BrdU-labeled cells across S phase relative to the position of G1 and G2+M (relative movement [RM]) was calculated by the following equation:
where FG1 is unlabeled G1 mean red fluorescence, FG2+M is unlabeled G2+M mean red fluorescence, and FS is the mean red fluorescence of the BrdU-labeled cells at time t. S phase duration (TS) was calculated as the time for one unit of relative movement. The potential doubling time was computed by the following equation:
with v defined as ln[1 + flu(t)/1 − fld(t)/2], where flu(t) is the fraction of labeled, undivided cells at time t and fld(t) is the fraction of labeled, divided cells at time t.
To determine the number of apoptotic and necrotic cells, cells were stained with MAb against annexin V (BD Pharmingen), to detect externalization of phosphatidylserine, and PI, to detect necrotic cells. Cells were cultured as described, harvested at the respective time points, stained with FITC-labeled annexin V and PI, and analyzed by flow cytometry by using the CellQuest software program (BD Pharmingen). A minimum of 15,000 cells was analyzed in each case.
For immunoblotting cells were washed in PBS and lysed in cell lysis buffer (1% Triton X-100, 0.5% NP-40, 0.1% sodium dodecyl sulfate, 0.5% sodium deoxycholate, 5 mM EDTA, 50 mM protease and phosphatase inhibitor cocktails, 1 mM phenylmethylsulfonyl fluoride, 100 μg of trypsin-chymotrypsin inhibitor per ml, 100 μg of chymostatin per ml in PBS). The concentration of proteins in each lysate was measured by using a Bio-Rad protein assay (München, Germany). Equivalent amounts of protein (10 μg) were fractionated on a 10 to 20% Tris-glycine gel and electrotransferred to a 0.2-μm-pore-size nitrocellulose membrane (Invitrogen, Karlsruhe, Germany). Membranes were blocked with 5% milk in 0.1% Tween 20-PBS (Fisher Scientific, Schwerte, Germany), followed by incubation for 60 min at room temperature with the indicated primary antibody (all primary antibodies were used at a concentration of 2 mg per ml of 5% milk in 0.1% Tween 20-PBS). The membranes were washed six times with 0.1% Tween 20-PBS and then incubated for 1 h with the appropriate horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology), washed, and incubated with the chemiluminescent substrate (Perkin-Elmer Life Sciences, Rodgau-Jügesheim, Germany) for 5 min. The membranes were then exposed to X-ray film (Amersham, Freiburg, Germany).
To determine cytokine secretion, 105 PBT were cultured for 48 h with or without anti-CD3 MAb and incubated in the presence or absence of 0, 10, or 25% (vol/vol) E. coli Nissle 1917-CM. The supernatant was then collected, and cytokine secretion was determined by a cytometric bead array, performed according to the manufacturer's instructions (BD Pharmingen). Briefly, bead populations with distinct fluorescence intensities, coated with capture antibodies specific for tumor necrosis factor α (TNF-α), gamma interferon (IFN-γ), IL-10, and IL-2 proteins, were mixed with PE-conjugated detection antibodies and incubated with recombinant standards or test sample to form sandwich complexes. Following the acquisition of sample data by flow cytometry, the cytokine concentrations were calculated by using cytometric bead array analysis software (BD Pharmingen).
Mice were bred under specific pathogen-free conditions at the Max Planck Institute for Infection Biology, Department of Cellular Microbiology. Mice were housed in filter-top cages and provided with sterile water and food ad libitum. TLR-2−/− mice were described previously (71, 77).
Statistical analysis was performed by using a paired Student's t test. Results are expressed as mean ± standard error of the mean (SEM), and significance was inferred with P values <0.05.
E. coli Nissle 1917 has beneficial effects in T-cell-mediated intestinal diseases such as ulcerative colitis or pouchitis (35-37, 52). However, the underlying mechanisms still remain unclear. We were therefore interested in the effects of E. coli Nissle 1917 on T-cell function. Culture of T cells in the presence of living E. coli Nissle 1917 without antibiotics resulted in significant bacterial overgrowth and complete T-cell death within 12 h (data not shown). We therefore generated E. coli Nissle 1917-CM as described in Materials and Methods to examine the effect of factors secreted by E. coli Nissle 1917 on cell cycle progression of PBT. Immediately after isolation, PBT contained >98% cells in G0/G1 phase. When PBT were stimulated by anti-CD3 MAb for 3 days, the number of cells in the G0/G1 phase dropped to 71% ± 3.5%, while those in the S and G2/M phases increased to 29% ± 2.1% (Fig. (Fig.1A).1A). When 10, 25, or 50% (vol/vol) E. coli Nissle 1917-CM was added to the cell culture, the number of cycling cells was significantly reduced with only 21% ± 1.2%, 14% ± 1.6%, and 9% ± 1.1%, respectively, of the cells in S and G2/M phase (Fig. (Fig.1A).1A). The inhibitory effect of E. coli Nissle 1917-CM on PBT cell cycle progression was observed within 48 h of cell culture and started to be detectable in concentrations as low as 2.5% E. coli Nissle 1917-CM (data not shown). To test if this effect is specific for E. coli Nissle 1917, conditioned media from the nonprobiotic E. coli strains DSM 498, PZ 840, PZ 873, and PZ 915 were generated as described in Material and Methods and added to anti-CD3-activated T-cell cultures. As shown in Fig. Fig.1B,1B, conditioned media from other E. coli strains also inhibited PBT cell cycle progression. Suppression of cell cycling by E. coli Nissle 1917-CM was reversed when the conditioned medium was washed out and replaced by medium that had not been in contact with E. coli Nissle 1917 (data not shown). In order to confirm that the inhibitory effect of E. coli Nissle 1917-CM on PBT cell cycle progression was mediated by soluble factors secreted by E. coli Nissle 1917 into the culture medium, PBT were activated in the presence of heat-inactivated E. coli Nissle 1917 microorganisms. Under these conditions, neither cell cycle entry nor progression through the cell cycle was affected (data not shown).
To assess the functional relevance of the cell cycle inhibitory effect of E. coli Nissle 1917-CM, we performed a stathmokinetic analysis to determine the effect of E. coli Nissle 1917 on PBT cell cycle traverse times. This was accomplished by using BrdU incorporation to model and measure cell cycle kinetics. The S phase traverse times of PBT activated by CD3 signaling and cocultured with E. coli Nissle 1917-CM was substantially slower than that of PBT activated in the absence of E. coli Nissle 1917 (Table (Table1).1). Calculation of potential doubling times (Tpot) showed that in the absence of E. coli Nissle 1917-CM, PBT required 24.7 ± 2.4 h to double when activated by anti-CD3 MAb, whereas in the presence of 10 or 25% E. coli Nissle 1917-CM, anti-CD3-activated PBT needed 37.0 ± 4.9 and 88.2 ± 9.6 h, respectively, to double their cell populations (Table (Table11).
Having shown that E. coli Nissle 1917-CM modulates PBT cell cycle kinetics, we next investigated how key regulatory molecules responsible for not only initiating and advancing but also inhibiting the cell cycle are modulated by E. coli Nissle 1917-CM. Western blot analysis showed that in PBT cyclin D2, the main regulator of the G1 phase, was upregulated by anti-CD3 activation and that this upregulation was profoundly reduced by E. coli Nissle 1917-CM (Fig. (Fig.2).2). The analysis of the Rb protein, which is essential for G1/S phase transition, showed that an increase in Rb protein phosphorylation was marked in CD3-activated PBT but, again, strongly inhibited by E. coli Nissle 1917-CM (Fig. (Fig.2).2). The functional relevance of this effect was confirmed when we showed that E2F-1 and cyclin A, both downstream and dependent on Rb protein phosphorylation, were also downregulated by E. coli Nissle 1917-CM (Fig. (Fig.2).2). We also analyzed cyclin B1, responsible for leading the cells to mitosis after they passed the restriction point at the G1/S interphase. To exactly localize the cyclin B1 increase within the cell cycle, DNA content was also determined by PI staining. As seen for early cell cycle regulators, E. coli Nissle 1917-CM dose dependently inhibited anti-CD3-induced expression of B1 (Fig. (Fig.3).3). To complement the analysis of cell cycle promoters, we determined the expression levels of key cell cycle inhibitors. Interestingly, even high doses of E. coli Nissle 1917-CM failed to increase protein expression levels of p21 or p53 (Fig. (Fig.33).
Since cell cycling and apoptosis are intimately linked (29, 75) and both features determine the life span of a cell, we next analyzed whether E. coli Nissle 1917-CM induces or modifies T-cell death. When PBT were stimulated with anti-CD3 MAb for 3 days, the proportion of apoptotic cells increased from 2% ± 0.9% to 5.8 ± 1.6% (data not shown). When purified and anti-CD3 MAb-activated PBT were cultured for 2 days in the presence of 0, 10, 25 or 50% (vol/vol) E. coli Nissle 1917-CM, the rates of apoptosis and necrosis did not increase, as determined by annexin V and PI staining (Fig. (Fig.4).4). Cell death may be initiated by the triggering of various death pathways by E. coli Nissle 1917 and then discontinued for unknown reasons. We therefore analyzed the effect of E. coli Nissle 1917-CM on key regulators of the intrinsic and extrinsic death pathways. However, even high doses of E. coli Nissle 1917-CM did not alter Bcl-2, Bax, or cytochrome c expression or induce caspase 3, caspase 8, or FLIP cleavage (data not shown).
The integration of all the cell cycle parameters studied above determines the intrinsic capacity of a T-cell population to divide and expand in response to receptor-mediated activation. To address the question of whether E. coli Nissle 1917 influences this capacity, we determined the number of cell divisions in CFDA SE-labeled PBT, activated in the presence or absence of E. coli Nissle 1917-CM. After 4 days of stimulation with CD3, CD28, and IL-2, four cell divisions containing 48% of the original population were seen in PBT (Fig. (Fig.5).5). In contrast, when 10% (vol/vol) E. coli Nissle 1917-CM was added, only 3% of equally activated PBT expanded and generated at most three daughter cell populations. This effect was augmented when the percentage of E. coli Nissle 1917-CM was increased to 25 or 50% (vol/vol), resulting in an almost complete depression of PBT expansion (Fig. (Fig.55).
Our data clearly demonstrate that E. coli Nissle 1917 substantially inhibits cell cycle entry and progression of anti-CD3-activated PBT in a dose-dependent manner. Thus, we were interested to evaluate the effects of E. coli Nissle 1917 on freshly isolated LPT, a population of effector-memory cells (55) that shows unique T-cell features with respect to cell cycle and apoptosis (47, 68). When LPT were stimulated by anti-CD3 or anti-CD2 MAb for 3 days, 87% ± 3.8% and 85% ± 2.9% of the cells remained in G0/G1 phase, respectively, whereas 13% ± 1.2% and 15% ± 2.0% of the cells entered the cell cycle, respectively (Fig. (Fig.6A).6A). The number of cells resting in G0/G1 or cycling in S/G2/M phase remained essentially identical when 10, 25 or 50% (vol/vol) E. coli Nissle 1917-CM was added to the cell culture (Fig. (Fig.6).6). The unresponsiveness of LPT toward E. coli Nissle 1917-CM was confirmed when cyclin B1 expression was measured in anti-CD3- and anti-CD-2-activated LPT (data not shown). We then determined the effect of E. coli Nissle 1917 on LPT apoptosis by using the same methods as described for PBT. When LPT were stimulated with anti-CD3 or anti-CD2 MAb for 3 days, the proportion of apoptotic cells markedly increased from 8% ± 2.5% to 28% ± 5.1% and 24% ± 4.7%, respectively (data not shown). Comparable to results with PBT, the number of apoptotic or necrotic cells was not altered by adding 10, 25, or 50% (vol/vol) E. coli Nissle 1917-CM to the cell culture (data not shown). To complete the analysis of E. coli Nissle 1917-CM-induced effects on LPT cell cycling and apoptosis, we assessed the effect of E. coli Nissle 1917-CM on LPT expansion. As determined by CFDA SE, when LPT were stimulated with CD3, CD28, and IL-2, 30 to 35% of the cells expanded, generating one daughter cell population (Fig. (Fig.7).7). In contrast to results with PBT, this number remained constant in the presence of 0, 10, 25 and 50% (vol/vol) E. coli Nissle 1917-CM (Fig. (Fig.77).
The failure of LPT to show different patterns to progress, expand, or undergo apoptosis following E. coli Nissle 1917 treatment may be due to either their residing in the interstitial matrix of the lamina propria or their effector-memory phenotype. To address this point, PBT were fractionated into naïve CD45RA+ and effector-memory CD45RO+ populations and stimulated with anti-CD3 in the absence or presence of 25% (vol/vol) E. coli Nissle 1917-CM. As shown in Fig. Fig.8,8, ,4.1%4.1% of CD45RA+ and 3.2% of CD45RO+ T cells reached the G2/M phase at day 3. When 25% (vol/vol) E. coli Nissle 1917-CM was added to the cell culture, the number of cells in the G2/M phase dropped significantly in both groups (Fig. (Fig.88).
The release of cytokines is crucial for the differentiation of T cells. Having demonstrated that E. coli Nissle 1917 substantially inhibits PBT cell cycling, we were interested whether E. coli Nissle 1917 modulates cytokine release patterns in T cells. PBT were isolated and activated with anti-CD3 MAb in the absence or presence of 10 and 25% (vol/vol) E. coli Nissle 1917-CM. After 3 days, the supernatants were collected, and cytokine expression was assessed by using a cytometric bead array assay. As depicted in Fig. Fig.9,9, E. coli Nissle 1917-CM significantly inhibited IL-2, TNF-α, and IFN-γ production (P < 0.01) in a dose-dependent manner. In contrast to the profound downregulation of proinflammatory cytokines, the expression of IL-10 was markedly upregulated by E. coli Nissle 1917-CM (P < 0.01) (Fig. (Fig.99).
Our data presented so far demonstrate that E. coli Nissle 1917 inhibits T-cell cycling early in the cell cycle. The response of a cell upon antigen presentation and its subsequent activation and cycling depends on cell activation via the TCR. In addition, costimulatory molecules such as CD2 and CD28 are crucial to fully stimulate a T cell and avoid anergy (59). To investigate whether E. coli Nissle 1917 alters T-cell activation and expression of costimulatory molecules, PBT were activated via the CD3 pathway in the absence or presence of 25 and 50% (vol/vol) E. coli Nissle 1917-CM, and the distributions of CD69, CD2, and CD28 as well as TCR α/β and TCR γ/δ were determined by flow cytometry. When PBT were activated by anti-CD3 MAbs, 35% ± 3.1% of the cells were CD69 positive after 24 h. When 25 or 50% (vol/vol) E. coli Nissle 1917-CM was added to the culture medium, the number of CD69-positive cells dropped significantly to 12% ± 2.9% and 7% ± 1.9%, respectively (P < 0.05). In addition, as assessed by flow cytometry, CD2 expression dropped by 46% ± 5.9% (P < 0.01), and CD28 expression decreased by 58% ± 4.7% (P < 0.01). We also determined the effect of E. coli Nissle 1917 on TCR α/β and TCR γ/δ cell populations. When PBT were activated by anti-CD3 MAb in the absence of E. coli Nissle 1917-CM, 94% of the cell population contained TCR α/β and 6% contained TCR γ/δ chains. Interestingly, when 50% (vol/vol) E. coli Nissle 1917-CM was added to the cell culture, the TCR γ/δ-positive cell fraction increased to 16% of the cells analyzed (data not shown).
Having demonstrated that E. coli Nissle 1917-CM inhibits cell cycle progression of activated PBT, we were next interested to determine which mechanism mediates this effect. We therefore first boiled and froze the E. coli Nissle 1917-CM at −80°C for a prolonged time. As depicted in Fig. Fig.10,10, the inhibitory effect of E. coli Nissle 1917-CM on T-cell cycling was preserved. This indicates that proteins are not likely responsible for the inhibitory effect of E. coli Nissle 1917 on the T-cell cycle. We therefore analyzed the effect of bacterial products such as LPS, BLPs, or immunostimulatory DNA (CpG DNA) on the T-cell cycle. These pathogen-associated molecular patterns (PAMP) have been demonstrated to activate the innate immune response and to mediate their effects through different TLRs (10). To investigate whether these PAMPs mimic the effect of E. coli Nissle 1917-CM, PBT were stimulated for 3 days with anti-CD3 MAbs in the presence or absence of LPS generated from E. coli Nissle 1917 (2 μg/ml), BLP (100 ng/ml), or CpG DNA (2 μg/ml), and the number of cycling cells was assessed by measuring their DNA content. As shown in Fig. Fig.10,10, whereas LPS and CpG DNA did not influence PBT cycling, BLPs mimicked the effect of E. coli Nissle 1917-CM and substantially decreased PBT cycling (P < 0.05) (Fig. (Fig.10).10). BLPs signals through TLR-2, and just recently TLR-2 expression has been described on activated T cells (32). We therefore isolated PBT from wild-type and TLR-2 knockout mice and activated the cells in the presence of 0, 10, and 25% (vol/vol) E. coli Nissle 1917-CM and BLPs (100 ng/ml). As depicted in Fig. Fig.11,11, comparable to results with human PBT, E. coli Nissle 1917-CM dose dependently inhibited cell progression into the S or G2/M phase. In contrast, when PBT from TLR-2-deficient mice were cultured in the presence of 0, 10, and 25% (vol/vol) E. coli Nissle 1917-CM, the number of cycling cells remained essentially unchanged (Fig. (Fig.11).11). In addition, by demonstrating that BLPs inhibit cell cycle progression of wild-type but not TLR-2-deficient PBT, we confirm the specificity of this pathway and provide evidence that BLPs play a crucial role in the modulation of T-cell cycle progression by binding to TLR-2 (Fig. (Fig.1111).
The host must meet the challenging task of avoiding an overly aggressive response to the 1014 microorganisms constituting the intestinal microflora while also mounting an effective immune response to intestinal pathogens and preventing bacterial spread within the organism (9). This generally tightly regulated balance between tolerance and immunity is disturbed in IBD, where insufficiently characterized environmental factors are assumed to induce and facilitate mucosal inflammation in a genetically susceptible host (18, 61, 62). Although the pathogenesis of IBD is not fully understood, it is widely accepted that a disturbed T-cell function plays a crucial role in the initiation and perpetuation of IBD (18). Whereas modulation of a deregulated immune response is a fundamental basis in the treatment of IBD, other contributors to disease initiation and perpetuation such as the intestinal microflora have been largely neglected (60). Recently, probiotic bacteria have been used in the treatment of IBD and have been shown to exert significant beneficial effects in ulcerative colitis and pouchitis (20, 21, 35, 36, 46, 52). The aim of this study was, therefore, to extensively investigate the effects of E. coli Nissle 1917 on peripheral and mucosal T cell function.
Proliferation of T cells in response to antigen stimulation is necessary to expand the T-cell pool, generate effector cells, and, thus, mount an effective immune response (25). We therefore assessed whether E. coli Nissle 1917 affects the initiation and progression of T cells through the cell cycle. Culture of activated PBT with E. coli Nissle 1917-conditioned T-cell medium strongly inhibited cell cycle entry and progression. Recently, it has been shown that the probiotic E. coli strain Laves 1913 ameliorates experimental colitis in mice and reduces the proliferative response of lymphocytes to specific bacterial antigens (33). This is in accordance with our finding that cell cycle progression was also inhibited by other E. coli strains and suggests that the capability to modulate T-cell cycling is shared by different E. coli strains.
Cell cycling is the result of traversing through individual cycle phases, and the time a cell spends in each phase determines how quickly a T cell reaches mitosis and finally divides (26). Activation of PBT with anti-CD3 MAb leads to a short S phase, resulting in a 24-h potential doubling time. When PBT were activated in the presence of E. coli Nissle 1917-CM, the time spent in the S phase increased significantly, resulting in a potential doubling time of up to 88 h when 25% (vol/vol) E. coli Nissle 1917-CM was added. These results demonstrate that E. coli Nissle 1917-CM significantly reduces the capacity of PBT to expand following antigen stimulation, resulting in an attenuated immune response.
To understand the molecular mechanisms underlying the effect of E. coli Nissle 1917 on cell cycle control, we measured expression levels of the key promoters responsible for the progression of distinct phases of the cell cycle (cyclin D2, Rb protein, E2F-1, cyclin A, and cyclin B1) (63), as well as relevant inhibitors of T-cell cycling (p21 and p53) (4, 64). The delayed cycling of PBT in the presence of E. coli Nissle 1917 after stimulation with CD3 MAb coincided with a lack of upregulation of cyclin D2. Furthermore, E. coli Nissle 1917 reduced phosphorylation of Rb protein and thus the release of pocket proteins such as E2F oncogene products (76). This step would propel the cells beyond the restriction point at the G1/S interphase, the so-called point of no return, where cell cycling is independent from the initial activation signal and can only be terminated by apoptosis (8, 67). Consequently, cell cycle promoters of later cell cycle phases that depend on the phosphorylation Rb protein were also downregulated by E. coli Nissle 1917. Although cell cycle progression was markedly decreased by E. coli Nissle 1917, p21 and p53 were not upregulated, indicating that no endogenous inhibitor can be assigned responsibility for this effect.
Apoptosis of T cells preserves immunological homeostasis and tolerance by countering the profound changes in the number of T cells stimulated by diverse antigens (75). To prevent autoimmunity or malignancy, the cell death program is initiated to terminate cell function once the antigen has been cleared or when cell proliferation is uncontrolled (39). Rheumatoid arthritis and IBD are characterized by an impaired apoptosis of T cells, and drugs that induce apoptosis of T cells or macrophages, like the anti-TNF-α antibody infliximab, play an important role in treating these diseases (17, 41, 72, 73). However, the immunosuppression induced by these drugs is sustained and causes an increased rate of opportunistic infections (11, 17, 27, 28, 53). E. coli Nissle 1917-CM did not induce apoptosis of T cells, demonstrating that E. coli Nissle 1917 selectively reduces the expansion of T cells by inhibiting their proliferating capacity but not by inducing their death. Since the full capacity to cycle was restored when E. coli Nissle 1917-CM was removed, our findings demonstrate that the immunosuppressive effect of E. coli Nissle 1917 can be eliminated and is reversible, unlike drugs that induce T-cell apoptosis.
LPT represent a highly specialized population of immune cells dedicated to maintaining immunological homeostasis by establishing a state of tolerance to dietary and bacterial antigens and eliminating dangerous non-self antigens (14, 44, 66). Local homeostasis requires a tight regulation of LPT responses, including proliferation, clonal expansion, cytokine production, and achieving a balance between cell death and survival (4, 42). Although E. coli Nissle 1917-CM significantly suppresses PBT expansion, cell cycling of mucosal T cells was not restrained, indicating that the effect of E. coli Nissle 1917 on T-cell cycling is distinct in different T-cell populations. E. coli Nissle 1917-CM caused a comparable reduction of cell cycle progression in CD3-activated CD45RA+ and CD45RO+ PBT, suggesting that the lack of LPT to respond to E. coli Nissle 1917-CM may not be solely caused by their memory status. Other factors like the mucosal microenvironment and the overall activation status may play an essential role. The unique response pattern of LPT and PBT may provide a possibility to limit the risk of an impaired mucosal immune response to potentially deleterious microbes and to downregulate PBT proliferation and expansion when PBT are actively recruited into the mucosa during intestinal inflammation (6). During intestinal inflammation, E. coli Nissle 1917 also translocates from the gastrointestinal lumen into the lamina propria and Peyer's patches (57, 58) and there may come in contact with PBT to decrease their expansion. However, migration of E. coli Nissle 1917 into the lamina propria in the course of intestinal inflammation is not a prerequisite for orally administrated probiotics to exert systemic immune-modulatory effects, since three orally administrated probiotics (Lactobacillus casei, Lactobacillus gasseri, and Lactobacillus rhamnosus) have been demonstrated to inhibit lymphoproliferation without inflammation in a mouse model (30).
An effective inflammatory immune response initially requires recruitment of immunocompetent cells to the site of inflammation and subsequently appropriate activation and regulation (6). Cytokines play a critical role in this setting, since they regulate the proliferation and differentiation of T cells and determine the course of an inflammatory process by releasing pro- and anti-inflammatory cytokines. In accordance with the demonstrated suppression of T-cell cycling, E. coli Nissle 1917-CM reduced IL-2 secretion of activated PBT, inhibiting the capacity of PBT to cycle. The potent proinflammatory cytokines TNF-α and IFN-γ have been demonstrated to play a crucial role in the pathogenesis of IBD (18, 54). E. coli Nissle 1917-CM profoundly reduced the secretion of these cytokines, indicating that E. coli Nissle 1917-CM has immunomodulatory capacities beyond the inhibition of cell cycling. The beneficial effects of E. coli Nissle 1917 on inflammatory processes are further underlined by the upregulation of IL-10 secretion, a potent immunomodulatory cytokine with anti-inflammatory properties, which has beneficial effects in treating mucosal inflammation (13, 45).
The expansion of T cells depends on their activation, which is achieved by the presentation of antigenic peptides in the context of the major histocompatibility complex to the TCR (74). In addition, costimulatory molecules are needed to fully activate T cells and avoid anergy (42, 59). E. coli Nissle 1917-CM significantly downregulates the expression of CD2 and CD28 and therefore reduces activation of PBT when they are stimulated by anti-CD3 MAb. This finding indicates that E. coli Nissle 1917 exerts its immunomodulatory effects early in the activation process, an observation further supported by the fact that E. coli Nissle 1917 downregulates protein expression of early cell cycle regulators in PBT.
About 95% of PBT have TCRs that are composed of α and β polypeptide chains, and 5% of the PBT have a TCR composed of γ and δ chains (2). TCR α/β cells recognize antigens only when they are bound to major histocompatibility complex molecules (3). In contrast, TCR γ/δ cells can recognize antigens directly as intact proteins or nonpeptide compounds and therefore play a crucial role in the immune response to microbial pathogens (16). It has been shown that the number of γ/δ T cells is increased after infection with Escherichia coli in C3H/HeN mice (48). In this study, we demonstrate for the first time, that soluble factors released from E. coli Nissle 1917 increase the number of γ/δ T cells and therefore increase the capability of PBT populations to recognize microbial pathogens.
E. coli Nissle 1917-CM inhibited PBT cell cycling even when the medium was frozen or boiled, suggesting that this phenomenon is not caused by secreted or shed peptides but may be mediated through PAMPs. Interestingly, CM consisting of heat-inactivated bacteria did not modulate cell cycling of PBT. This may be explained by alterations of the bacterial wall or BLPs during the boiling process of living bacteria. Another explanation for this observation would be an interruption of putative BLP modifications in living bacteria that occur before shedding and that are essential for the binding of BLPs to TLRs. To determine whether PAMPs may be involved in E. coli Nissle 1917-mediated effects on T-cell function, the effects of LPS, BLPs, or immunostimulatory DNA (CpG DNA) on T-cell function were examined. Interestingly, only BLPs induced a significant downregulation of PBT cell cycling that was comparable to the effect of E. coli Nissle 1917-CM. BLPs act mainly through TLR-2 (1) and act as adjuvants for human T-cell responses (32, 65). Using TLR-2-deficient mice, we could demonstrate that the suppressing effect of E. coli Nissle 1917-CM on PBT cell cycling is mediated by TLR-2, identifying this pathway as a mediator of E. coli Nissle 1917 signaling in PBT. This finding is supported by the fact that TLR-2 is expressed on activated T cells (32). Rachmilewitz and coworkers just published a study that shows that genomic DNA isolated from VSL-3, a probiotic mixture containing viable lyophilized gram-positive bacteria, and E. coli DH5α ameliorate the severity of dextran sulfate sodium-induced colitis (51). This effect was mediated by TLR-9 signaling. Genomic DNA or CpG DNA is known to signal via TLR-9 (24, 34), which is expressed on antigen-presenting cells. T cells do not respond to CpG DNA in the absence of dendritic cells (24), which is in accordance with the results of our study, where CpG DNA did not modulate T-cell cycling in a purified T-cell system, thus demonstrating for the first time a direct interaction of probiotics with T cells via TLRs and suggesting different pathways by which probiotics can mediate their effects.
In conclusion, we demonstrate evidence that E. coli Nissle 1917-CM and BLPs inhibit PBT cell cycling via the TLR-2 receptor pathway. In contrast to PBT cell cycling, LPT cell cycling was not affected by E. coli Nissle 1917. E. coli Nissle 1917-CM inhibited activation of costimulatory molecules, reduced activation of PBT, and increased the proportion of γ/δ T cells. In addition, E. coli Nissle 1917-CM decreased the secretion of proinflammatory cytokines and increased the secretion of anti-inflammatory cytokines. Although a reduction of PBT cell cycling was also observed to a lesser extent when conditioned media were generated from other E. coli strains, E. coli Nissle 1917 is currently the only E. coli strain with clinically proven efficacy in the treatment of IBD while also fulfilling the safety requirements necessary in the treatment humans. The results of this study provide further evidence that probiotic bacteria broadly influence the human immune system reveal an underlying mechanism, and may explain the beneficial effects of probiotic bacteria like E. coli Nissle 1917 in intestinal inflammation.
This work was financially supported by the Deutsche Forschungsgemeinschaft (grant STU247/2-1 and 247/3-1 to A.S.), ArdeyPharm GmbH, and the Charité Medical School, Berlin, Germany.
We thank Arturo Zychlinsky for helpful discussion and critical reading of the manuscript, Johanna Harder-d'Heureuse and Diana Metzke for technical assistance, and U. Zaehringer for providing purified E. coli Nissle 1917 LPS. We also thank the Department of Surgery, DRK Clinics Westend, for providing tissue samples.
Editor: J. D. Clements