The present invention is concerned with modulating intestinal responses to infection, in particular through use of the bacterium Bacteroides thetaiotaomicron.
The mammalian intestine is colonised by several hundred bacterial species whose numbers increase dramatically in the colon, an anatomical site prone to infection, inflammatory disease and cancer. However, the resident flora of the gastrointestinal tract of adult animals conveys a high degree of protection (colonisation resistance) against infection (van der Waaij, 1984; Salminen et al, 1998). As a result, most opportunistic pathogens encountered in the environment are unable to establish a foothold and are rapidly excluded (van der Waaij, 1984; Salminen et al, 1998). However, if the commensal flora is compromised, opportunistic pathogens may then persist in the gut. For example, Pseudomonas aeruginosa does not usually persist in the murine gut but will colonise and cause very severe infection after disruption of the commensal flora with antibiotics (Pier et al, 1992). The commensal bacteria also confer some degree of protection against virulent pathogens, such as Salmonella. However, these pathogens can overwhelm or circumvent the protective effects of the commensal flora and cause severe infection if they are present in significant numbers.
Colonisation resistance is due, in part, to competitive exclusion of harmful bacteria, through preferential use of nutrients and substrates or blocking of potential attachment sites on the gut by the commensal flora. However, commensal bacteria also modulate the gut cellular and immune systems of the host (Bry et al, 1996; Herias et al, 1998; 1999; Hooper et al, 2000; 2001; Cebra, 1999; Snel et al, 1998; Talham et al, 1999; Lopez-Boado et al, 2000; Shu et al, 2000; Campbell et al, 2001). The flora may alter the gastrointestinal tract to create niche microenvironments that are well suited to them but not to other bacteria. Alternatively, they may modify epithelial cell responses to harmful bacteria (Campbell et al, 2001) and thereby attenuate changes that would facilitate colonisation and invasion.
The healthy gut maintains a hyporesponsive tone towards its diverse bacterial load, yet the presence of a threshold level of pathogenic bacteria is sufficient to activate transcriptional systems that rapidly upregulate proinflammatory gene expression in gut tissues. These transcription products then trigger a cascade of responses including the chemotaxis of polymorphonuclear (PMN) cells into the lamina propria of infected intestinal sites. Although these events are essential for bacterial clearance, they also cause tissue pathology that can exacerbate the symptoms of disease. The first point of contact for lumenal bacteria is a continuous layer of ephithelial cells that both interfaces and segregates the gut immune system. The ability of intestinal epithelial cells to discriminate between pathogenic and non-pathogenic bacteria is crucial in averting harmful inappropriate responses to colonising bacteria and in maintaining gut health. This discriminatory function is imprinted in the systems of bacterial recognition and cell signalling. Recognition of bacterial cell surface structures is, in part, a function of Toll-like receptors expressed on apical and basolateral surfaces of epithelial cells, which trigger Nuclear Factor Kappa B (NF-κB) - mediated immune activation (Gerwirtz et al, 2001). Hitherto, receptor systems linked to immune-suppressor activities have not been identified.
Administration of the human commensal bacterium B. thetaiotaomicron to germ-free mice triggered expression of key genes linked to intestinal maturation and development of barrier function (Hooper et al, 2001). It also increased intestinal levels of Fucα1, 2Galβ-glycans (Bry et al, 1996; Hooper et al, 2001) and matrilysin (Lopez-Boado et al, 2000). B. thetaiotaomicron was also found to modify the responses of epithelial cells to challenge in vitro with Salmonella (Campbell et al, 2001). In particular, there was suppression of some pro-inflammatory pathways (Campbell et al, 2001). Despite these potentially protective properties, B. thetaiotaomicron did not however increase the resistance of ex-germ free mice to infection by Clostridium perfringens serotype A (Yurdusev et al, 1989). Nonetheless, the pathogen was cleared from the gut if the mice were treated with B. thetaiotaomicron in combination with Fusobacterium necrogenes (Yurdusev et al, 1989) and non-pathogenic Clostridia strain Cl (Yurdusev et al, 1986). This suggests that, although B. thetaiotaomicron alone can induce potentially protective changes in the gut, it may need to act in tandem with other commensal strains to significantly enhance overall resistance to infection.
Therapeutic methods are described in International patent publication No.
International patent publication No.
International patent publication No.
In this invention we define a novel mechanism whereby non-pathogenic bacteria alter both the nucleocytoplasmic distribution of PPARγ and the transcriptional activity of (NF-κB).
Further, this invention relates to the use of non-pathogenic bacteria for the modulation of inflammatory responses. Still further, the invention provides evidence for differential histone acetylation of p65 (RelA) affecting IκBα-mediated export of RelA and for a novel mode of action for peroxisome proliferator-activated receptors (PPARγ). In particular, the invention describes the attenuation of epithelial inflammatory gene expression by altering both the nucleocytoplasmic distribution of PPARγ and the transcriptional activity of NF-κB by use of non-pathogenic bacteria. Further the present invention provides a means of attenuating inflammation for treatment and prevention of disease and inflammatory disorders.
The novel pathway described in the present invention can be used to screen for novel methods and for novel products to modulate inflammatory cytokine production. Further, the present invention describes the use of a non-pathogenic bacteria in attenuating inflammatory cytokine production and returning the immune system to homeostasis.
The modulation of inflammatory cytokine production described herein differs from known publications in a number of aspects.
Neish et al, 2000 describe regulation of epithelial responses by inhibition of IκBα ubiquitination. The present invention differs from Neish et al, 2000 by using a different mode of action.
Evidence is provided for differential histone acetylation of p65 (RelA) affecting IκBα-mediated export of RelA and for a novel mode of action for peroxisome proliferator-activated receptors (PPARγ). This novel data provides the potential for devising novel means of, and novel products for, modulating inflammatory cytokine production. In vivo validation of the results are also described.
The present invention provides an assay to select a candidate drug for the treatment of inflammatory disease due to cytokine production. The assay comprises:
The assay may further comprise a step of selecting a candidate drug which demonstrates at least one effect selected from the group consisting of:
According to a preferred embodiment, the effect of the candidate drug which is analysed is the nuclear export of transcription factor from the NF-κB family or the variation of the amount of PPARγ/RelA complexes in the cytosol of the cells.
In another preferred embodiment, the assay further comprises the step of analysing the variation in the level of one or more cytokines selected from the group consisting of TNF-α, IL-8, MIP-2α and Cox-2; preferably IL-8 and/or MIP-2α.
Desirably the intestinal cells used in the assay may be in form of an intestinal cell line maintained in cell culture. Suitable cell lines include Caco-2 cell line. Alternatively the assay could be performed in vivo using a suitable animal, like a mammal (eg. mice).
Optionally the assay may be conducted in the presence of known pathogenic bacteria, for example Salmonella sp.
The present invention is also useful for the treatment of inflammatory cytokine production associated diseases which includes the step of administering a therapeutically effective dose of Bacteroides thetaiotaomicron capable of modulating:
Preferably the invention is useful for the treatment of an inflammatory disease, especially a disease which is caused, at least in part, by the inflammatory response of intestinal cells. Generally such a disease will involve the production of inflammatory cytokines. The agent is Bacteroides thetaiotaomicron.
The present invention provides the use of B. thetaiotaomicron as a medicament for the treatment of an inflammatory disease which is caused, at least in part, by the inflammatory response of intestinal cells. Thus the present invention provides use of B. thetaiotaomicron for attenuating or preventing or treating inflammatory cytokine production associated diseases or conditions including, but not limited to; inflammatory bowel disease and diseases (particular mention may be made of Crohn's disease and Irritable Bowel Syndrome); rheumatoid arthritis; immunodeficiency syndrome; cachexia; multiple sclerosis, and inhibition of proliferation of keratinocytes.
The present invention yet further provides the use of B. thetaiotaomicron in the manufacture of a medicament for the treatment of an inflammatory disease caused, at least in part, by an inflammatory response of intestinal cells, as defined in the claims.
The present invention is also useful for the treatment of, or to attenuate, the inflammatory response of intestinal cells which comprises administering a therapeutically effective dose of B. thetaiotaomicron.
In more detail B. thetaiotaomicron may be used to disrupt the NF-κB pathway and/or to inhibit the p65 (RelA) response and/or act as an excitor or inhibitor of PPARγ. B. thetaiotaomicron may therefore be used for the treatment of inflammatory cytokine production associated diseases.
The B. thetaiotaomicron may be administered to a patient live by means of a foodstuff or suppository.
Thus, this invention is useful for healing the inflammatory response of intestinal cells by administering a therapeutically effective dose of B. thetaiotaomicron.
For the prevention and treatment of conditions described herein, B. thetaiotamicron is preferentially delivered to the site of action in the gastrointestinal tract by oral administration in any appropriate formulae or carrier or excipient or diluent or stabilizer. Such delivery mechanisms may be of any formulation including but not limited to solid formulations such as tablets or capsules; liquid solutions such as yoghurt or drinks or suspensions. A preferential delivery mechanism is one that delivers the bacteria orally, preferably without harm through the acid environment of the stomach to the site of action within the intestine. Such bacteria may be given in conjunction with a prebiotic.
The invention also provides a novel mode of action of PPARγ and differential histone acetylation of p65 (RelA) affecting IκBα-mediated export of RelA for screening for specific ligands with the potential to modulate, either agonistically or antagonistically, inflammatory cytokine production via the novel pathway described herein.
The invention is useful for the treatment of inflammatory cytokine production associated diseases by administering a therapeutically effective dose of Bacteroides thetaiotaomicron capable of altering the transcriptional activity of NF-κB or capable of increasing the amount of PPARγ/RelA complexes in the cell cytosol.
The invention is useful for returning and maintaining the immune system of mammals to homeostasis, by administering a therapeutically effective dose of Bacteroides thetaiotaomicron capable of altering the transcriptional activity of NF-κB or capable of increasing the amount of PPARγ/RelA complexes in the cell cytosol.
We have found that the anti-inflammatory activity of B. thetaiotaomicron involves a totally novel mode of action, defined by accelerated NF-κB nuclear export and PPARγ-mediated sequestration in the cytosol. The experimental work (described in detail in the examples) was undertaken in three phases:
The experimental work performed shows that:
The present invention will now be illustrated further by reference to the following non-limiting, examples and to the figures in which:
The evidence for bacterial regulation of host inflammatory response was derived from studies investigating inflammatory gene expression following short-term exposure of intestinal Caco-2 cells to Salmonella enteritidis co-cultured in the presence and absence of Bacteroides thetaiotaomicron. Using cDNA macro-array technology (CLONTECH Atlas human cytokine/receptor array system) we identified several genes including TNF-α, IL-8, MIP-2α and COX-2, whose induction following S. enteritidis exposure was attenuated by the presence of B. thetaiotaomicron. These results were confirmed using Northern hybridization and real-time PCR (
In setting the experimental conditions for these studies, many optimisation experiments (including time course, bacterial dose/growth phase, Caco-2 cell passage/confluence studies) were undertaken. Importantly we established that the growth, attachment and invasion of bacteria were unaffected by culture/co-culture conditions, thereby ruling out the possibility that the data could be attributed to differential attachment/invasion (see
Inflammatory gene expression was determined in non-infected Caco-2 cells (1), cells incubated with 108 S. enteritidis alone (2), cells incubated with 108 S. enteritidis and 109 B. thetaiotaomicron (3) or cells incubated with 109 B. thetaiotaomicron alone (4). Bacteria were applied for 2 hours. Cells were washed and harvested for mRNA isolation and 5µg of mRNA was analysed by Northern hybridisation with 32P-labelled probes specific for TNF-α, IL-8, MIP-2α, TGF-β, COX-2 and G3PDH.
Some of the results describing the effects of B. thetaiotaomicron are consistent with the recent molecular analysis reporting by Hooper et al, 2001. Our results were confirmed using Northern hybridisation and real-time PCR. Levels of TNF-α, IL-8, MIP-2α and COX-2 were significantly increased in response to S. enteritidis challenge (
The quantitative data shown in
To investigate whether the attenuation of inflammatory cytokine expression was specific to B. thetaiotaomicron a related aerotolerant strain, B. vulgatus, was also studied.
Treatments were as described above for (1) and (2), cells incubated with 108 S. enteritidis and 109 B. vulgatus (3a) or cells incubated with 109 B. vulgatus alone (4a). Bacteria were applied for 2 hours.
B. vulgatus was negative for the biological activity (see
In the results obtained the small induction of inflammatory gene expression triggered by both of the commensal strains, which may be related to the fact that late log phase bacteria were tested; was noteworthy. The application of early log phase bacteria devoid of bacterial cell debris, may completely abrogate this response. A further experiment was therefore conducted in which Caco-2 cells were incubated with 108 S. enteritidis and 109 B. thetaiotaomicron for either 2 or 4 hours. The inhibitory effect of B. thetaiotaomicron on IL-8 expression was sustained for prolonged periods (
The physiological relevance of the data showing the suppression of inflammatory cytokines by B. thetaiotaomicron was verified using both a functional in vitro model of PMN recruitment and an in vivo S. enteritidis rat infection model. PMN recruitment in vitro and in vivo was monitored using myeloperoxidase (MPO) activity.
Caco-2 cells were seeded onto inverted Transwells (Corning) and various combinations of bacteria applied to the apical surfaces of the cells. Freshly isolated human PMNs were applied to the basolateral compartment and their transepithelial migration determined by MPO assay (Neish et al, 2000). Treatment groups (1-4) were as described above. Cells were incubated for 2 hours and then bacteria were removed and fresh media applied. Cells were incubated for a further 2 hours and then washed in HBSS. Both cells, and media derived from the apical compartment, were solubilised in a 1% Triton C-100 and MPO determined.
The results derived from the in vitro model of polymorphonuclear leukocyte (PMN) recruitment assay confirmed the anti-inflammatory activity of B. thetaiotaomicron (
It is noteworthy that S. enteritidis infection of the rat intestine is mainly in the ileum, the natural habitat for B. thetaiotaomicron.
Newly weaned (21 d) minimal flora rats (fed on normal laboratory diets) were split into 2 groups and one group additionally fed with anaerobically prepared jelly (0.5 g/d) containing 108 cfu of B. thetaiotaomicron for 19 days. Half of the rats in each group were then orally challenged with 108 S. enteritidis. The severity of the inflammatory response in all animals was assessed by determining MPO levels in ileal mucosa at 6 days post S. enteritidis infection. Treatments were, no B. thetaiotaomicron and no S. enteritidis (1), no B. thetaiotaomicron then 108 S. enteritidis (2), B. thetaiotaomicron then 108 S. enteritidis (3), and B. thetaiotaomicron alone (4). Experiments were undertaken at least 3 times with similar results.
The level of MPO in the ileal mucosa of rats challenged with S. enteritidis increased (P<0.005), but was significantly attenuated by prior oral inoculation and stabilisation of B. thetaiotaomicron within the flora (P<0.001) (
In addition to MPO data discussed above
The NF-κB family of transcription factors plays a central role in regulating inflammatory responses. These proteins share a highly conserved NH2-terminal sequence referred to as the Rel homology domain, which is required for their subunit dimerisation, DNA binding and interaction with the inhibitory IκB proteins. Signals that induce NF-κB lead to phosphorylation, of the inhibitory IκB proteins (IκBα at Ser-32 and Ser-36) which are then targeted for ubiquitination and proteosome-meditated degradation. Rel proteins, including RelA (p65), are thereby released to translocate to the nucleus and bind DNA.
Non-infected Caco-2 cells (1) were incubated with 108 S. enteritidis alone (2), 108 S. enteritidis and 109 B. thetaiotaomicron (3), and 109 B. thetaiotaomicron alone (4) for 2 hours. Cells were fixed in 4% paraformaldehyde for 30 minutes at room temperature, permeabilised with 0.2% Triton-X 1000 at 4°C and examined by indirect immunofluorescence microscopy. Primary antibodies (Santa Cruz)(A-D) anti-RelA [NF-κB p65], (E-H) anti-PPARγ, (I-L) anti-IκBα and (M-P) anti-pκBα. Secondary antibodies: Alexa Fluor species-specific anti-IgG (Molecular Probes).
We found that exposure of Caco-2 cells to S. enteritidis triggered a cascade of events leading to enhanced translocation of RelA to the nucleus (
Significantly, B. thetaiotaomicron was found to abolish the translocation of RelA to the nucleus with virtually all RelA localised to the cytosol following 2 hours of exposure to this bacterium (
Phosphorylation of IκBα and, importantly, its degradation was observed following exposure to S. enteritidis in the presence and absence of B. thetaiotaomicron (
Supershift EMSA performed on nuclear extracts, incubated with RelA [NF-κB p65] specific antibody (Santa Cruz), from non-infected Caco-2 cells (1) following incubation with 108 S. enteritidis (2), 108 S. enteritidis and 109 B. thetaiotaomicron (3), and 109 B. thetaiotaomicron alone (4) for 2 hours. Where indicated B. thetaiotaomicron was heat inactivated at 70°C for 15 minutes prior to its addition to the Caco-2 cells.
Using electrophoretic mobility shift assays (EMSA) we confirmed that RelA was the major subunit activated by S. enteritidis (
A similar experiment is illustrated in
We also showed that viable but not heat-inactivated B. thetaiotaomicron inhibited the RelA response (
Using the same experimental groups (1-4) as described above, bacteria were applied for 2 hours before mRNA was analysed by Northern hybridisation and blots probed with specific 32P-labelled IκBα and G3PDH probes. Both the levels of IκBα mRNA (
A recent report of immune-suppression by non-pathogenic bacteria attributed the mechanism to inhibition of both ubiquitination and degradation of lκBα (Neish et al, 2000), which clearly does not apply to our model system. The predominant cytosolic location of NF-κB complexes in cells co-cultured with both S. enteritidis and B. thetaiotaomicron could then be explained by a greater efficiency of the IκBα-mediated nuclear export over nuclear localisation. The levels of IκBα protein are moderately higher in the nuclei of cells treated with both S. enteritidis and B. thetaiotaomicron than S. enteritidis alone, which could potentially accelerate the deactivation of NF-κB and contribute to the overall anti-inflammatory properties of B. thetaiotaomicron. However, a more likely explanation is the differential deacetylation of RelA. It has recently been reported that the duration of nuclear RelA activation is determined by reversible acetylation (Chen et al, 2001). The acetylated form of RelA has a low affinity for IκBα whereas the deacetylation of RelA by histone deacetylase 3 (HDAC3) promotes the binding to IκBα and the CRM-1-mediated nuclear export of RelA. We showed that B. thetaiotaomicron promoted HDAC3 association with RelA (
Furthermore, both RelA and peroxisome proliferator-activiated receptor gamma (PPARγ) proteins appeared to co-localise within the same cellular compartment (
AP-1 complexes play an important role in the regulation of inflammatory gene expression and are rapidly activated by a variety of extracellular stimuli including growth factors, cytokines and bacteria (Meyer-ter-Vehn et al, 2000). We investigated whether the inhibitory effects of B. thetaiotaomicron, extended to this signalling pathway. AP-1 activity is regulated by two main mechanisms, involving enhanced expression and phosphorylation of AP-1 subunits consisting of homo- and hetero-dimers of the protooncogene families Fos (c-Fos, FosB, Fra-1 and Fra-2), Jun (JunB, c-Jun and JunD) and ATF (ATF2, ATF3/LRF2 and B-ATF), all members of the leucine zipper family of DNA binding proteins. These proteins are mainly controlled by three related kinases, collectively called mitogen-activated protein (MAP) kinases. We investigated whether the level of protein phosphorylation of p42 and p44 (extracellular signal-regulated kinase; ERK) kinases, the c-Jun N-terminal kinases (JNK)/stress-activated protein (SAP) kinases and the p38 kinases were altered in response to S. enteritidis and B. thetaiotaomicron. We found that S. enteritidis activated p38 MAP kinase (
To further investigate the mechanism of immune suppression by non-pathogenic bacteria we initially studied the anti-inflammatory cytokines, IL-10 and TGF-β. IL-10 gene expression was not affected by treatment with B. thetaiotaomicron (data not shown) although this does not preclude the possibility that constitutive protein may be involved. There was some suggestion from the data that the contra-inflammatory cytokine TGF-β may be involved. However, if this cytokine is acting to down-regulate inflammatory responses, it is more likely to be involved in the longer-term anti-inflammatory effects, as de novo cytokine synthesis would not have been significant over the time course of these acute studies.
The PPARs are emerging as important modulators of inflammatory processes (Nakajima et al, 2001). PPARs are ligand activated transcription factors that regulate gene expression by binding with retinoid X receptor (RXR), as heterodimeric partners, to specific DNA sequence elements termed PPAR response elements (PPRE). Recent work however, suggests that ligand activation of PPARγ may be important in modulating AP-1 and NF-κB-mediated gene expression (Su et al, 1999). We investigated the role of PPARs in the regulation of NF-κB by B. thetaiotaomicron. Consistent with receptor activation we were able to shown that the mRNA for PPARα and PPARγ decreased in cells exposed to both S. enteritidis and B. thetaiotaomicron (
Although we observed effects with PPARγ and PPARα agonists, in subsequent experiments we concentrated on PPARγ as it has been shown to be expressed at much higher levels than PPARα in the colonic mucosa (Mansen et al, 1996; Fajas et al, 1997).
In addition to regulating gene transcription via PPRE, PPARs have recently been shown to inhibit gene transcription by interfering with other transcription factor pathways through a mechanism independent of DNA binding (Delerive et al, 1999). PPARγ2 can be found in both cytosolic and nuclear compartments (Thuillier et al, 1998), although the physiological relevance of cytosolic PPARγ is currently unknown. Using immunocytochemical localisation on fixed Caco-2 cells we found that S. enteritidis induced nuclear accumulation of PPARγ (
All published data relating to PPARγ have focused however, on its nuclear site of action. Using immunocytochemistry we found that S. enteritidis induced nuclear accumulation of PPARγ in Caco-2 cells (
The Salmonella-induced PPARγ protein formed a heterodimeric complex with RXRα (demonstrated by IP, results not shown) and partitioned to the detergent insoluble cell fraction whereas that induced by B. thetaiotaomicron was detergent soluble (
Furthermore, even though PPARγ and PPARα ligands attenuate the Salmonella mediated cytokine induction (results not shown), the PPAR ligands tested in these experiments did not mimic the B. thetaiotaomicron effect on PPARγ cellular re-localisation, suggesting a novel endogenous ligand or mechanism of regulation for PPARγ.
Using dual label immunocytochemistry we found that much of the PPARγ protein co-localized with RelA (
It has been suggested that nuclear PPARγ and NF-κB can form an inactive complex (Ricote et al, 1999). Furthermore, glutathione S-transferase pull-down experiments demonstrate that PPARα physically interacts with c-Jun and RelA p65 (Delerive et al, 1999). In our study, using immuno-purification (IP) protocols we show that isolation of PPARγ from cells treated with S. enteritidis in the presence of B. thetaiotaomicron resulted in the co-purification of RelA (
To further investigate the importance of PPARγ in the regulation of RelA we utilised a dominant negative (DN) form of the receptor (gifted by Professor Chatterjee, University of Cambridge, UK) (Gurnell et al, 2000). Within the PPAR receptors there is a striking conservation of both leucine and glutamic acid. These residues are essential for ligand binding and recruitment of nuclear co-activators. Mutation of these residues generates a DN form of the receptor, which impairs its ability to recruit co-activators and release two co-repressors, silencing mediator of retinoid and thyroid receptors (SMRT) and nuclear co-repressor (NcoR) (Gurnell et al, 2000). Co-immunoprecipitation experiments with DN PPARγ indicate that the SMRT interacts with the receptor in vivo and that the mutated PPARγ is a potent transcriptional repressor. Chimeric fluorescent protein constructs of human PPARγ and the dominant negative form of PPARγ containing a cyan fluorescent protein (CFP) on the carboxyl-terminal domain were prepared by PCR amplification from previously reported clones (Gurnell et al.2000). Chimeric RelA linked to YFP was gifted from Dr J Schmid (Schmid et al. 2000). Confirmation of successful expression was performed by Western blot analysis of transiently transfected Hela cells using anti-human PPARγ and anti-human RelA antibodies.
Double transfection studies with PPARγ DN and an NF-κB luciferase reporter (gifted by Dr Israel, Pasteur Institute, Paris, France) were performed. Consistent with a recent report we found that cells exposed for long periods to pathogenic S. enteritidis inactivated luciferase activity (Savkovic et al, 2000). To avoid this we stimulated the cells for 2 hours and then removed the bacteria by washing and determined luciferase protein production after 8 hours. We found both in the control and PPARγ DN transfected cells, that S. enteritidis induced NF-κB activation and luciferase protein synthesis (P<0.001) (
We further investigated the cellular distribution of PPARγ and RelA using fluorescent microscopy performed in Caco-2 cells and Hela cells transfected with chimeric constructs of WT and DN PPARγ and CFP and chimeric YFP and RelA. Much of the expressed PPARγ/RelA protein was localized within the nuclei of transfected cells as previously reported (Schmid et al. 2000). Similarly, in cells transfected with PPARγ WT and RelA the predominant expression was nuclear but there was also very clear evidence of punctate cytosolic labeling and co-localization of PPARγ and RelA following incubation with B. thetaiotaomicron (
The existence of bacteria within the gastrointestinal tract that actively suppress inflammation seems certain. A shift in the balance between enteric microrganisms that suppress inflammation and those that support inflammatory responses may be directly relevant to the etiology of inflammatory diseases of the gastrointestinal tract. Hence it is plausible that inflammatory bowel disease is related to the loss of bacterial strains that actively contribute to immune homeostasis and this may also explain why some patients respond to probiotic therapy (Madsen et al, 2001). Equally, dysfunction of PPARs, due to inadequate levels of receptor protein or due to expression of specific allelic variants that enhance susceptibility to inflammatory conditions in the gut, may be highly relevant. This view is further supported by recent published data indicating that heterozygous PPARy-deficient mice are significantly more susceptible to ischemia-reperfusion-induced intestinal inflammation and injury (Nakajima et al, 2001).
Tissue culture reagents were obtained from Sigma and Invitrogen, Antibodies obtained from Santa Cruz, Molecular probes, New England Biolabs and, molecular reagents obtained from Promega, Invitrogen and Amersham.
Caco-2 and Hela cells were routinely cultured in 35 mm culture dishes. Typically experiments required incubation of cells with the following four treatments: cells in medium alone (1); cells incubated with 108 S. enteritidis (2); cells incubated with 108 S. Enteritidis and 109 Bacteriodes thetaiotaomicron (3); and cells incubated with 109 B. thetaiotaomicron (4). Bacteria were applied for either 2 h or 4 h and removed by extensive washing. Other bacteria and ligands tested included E. coli 0157 H7, B. vulgatus, PMA, IL-1α/β and TNFα. All experiments were optimised and protocols were based on detailed dose responses and time courses. Transepithelial migration of PMN cells through Caco-2 cell monolayers was determined by MPO assays (Parkos et al.1991) The Caco-2 cells were incubated for 2 h before the bacteria were removed and replaced with fresh media. Thereafter cells were then incubated for a further 2 h. Both cells and media derived from the apical compartment were then solubilised in a 1% Triton X-100 and MPO determined.
Cytokine analyses were undertaken using Clontech macroarray, Northern hybridizations and real-time PCR of RNA isolated from Caco-2 cells. Total RNA and mRNA were isolated, cDNA produced and PCRs were performed under standard conditions. IL-8 protein concentration was determined by ELISA.
Nuclear extracts were incubated with single stranded 32P-Iabelled oligonucleotide probes containing consensus binding sequences for NF-κB (5'-AGT TGA GGG GAC TTT CCC AGG C-3') or AP-1 (5'-CGC TTG ATG AGT CAG CCG GAA-3'), separated by electrophoresis and visualized by autoradiography. EMSA supershifts were performed using specific NF-κB subunit antibodies. Effects of B. thetaiotaomicron on NFκB and AP-1 signaling were determined by EMSA, western blotting (protein and phosphoprotein), promoter-specific reporter analysis and target gene expression (eg. fos and jun).
Following experimental treatments Caco2 or Hela cells grown on 35mm culture dishes were fixed in 4% paraformaldehyde and permeabilised in 0.2% Triton X-100/PBS. Cells were incubated for 1 h at room temperature with primary antibodies (1µg/ml) diluted in PBS containing 1% serum from the species in which the secondary antibody was raised. Secondary antibodies 1µ/ml) were either Alexa Fluor 594 donkey anti-goat or Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes) as appropriate. Labelled cells were mounted with Vectorshield (Vector) and examined on a Zeiss Axioskop 50 widefield fluorescence microscope or on a Bio-Rad Radiance2100 laser scanning microscope. Representative digital images were imported into Adobe Photoshop 6.0 for final arrangement.
The coding sequence of PPARγ, and a dominant negative mutant of PPARγ (Gurnell et al.2000), were modified during PCR amplifications to add a Xho I recognition sequence at the 5' end and a Sac II recognition sequence at the 5' end and a Sac II sequence at the 3' end of the products. The amplifications were performed using Pfu DNA polymerase. Products were restricted and cloned into both pECFP-C1 and pEYFP-C1 (Promega). Successful construction was verified by DNA sequencing. Human p65 cloned into pEYFP-C1 was as previously reported (Schmid et al.2000).
Caco2 or Hela cells were seeded on 35mm culture dishes grown to 90% confluence and transfected using standard lipofectamine-mediated methods (Invitrogen). After 48 h, bacterial incubations were undertaken as previously described; cells were fixed for 30 min at room temperature in 4% paraformaldehyde in 0.1M sodium phosphate buffer pH 7.4, washed in PBS and examined on a Zeiss Axioskop 50 widefield fluorescence microscope equipped with custom filters (Omega Optical) for CFP and YFP. Representative images were recorded with a Zeiss AxioCam digital camera, processed with Zeiss AxioVision 3.0 software and imported into Adobe Photoshop 6.0 for final arrangement. In some experiments transfected cells were subjected to the immunofluorescence analyses described above.
The following example was conducted to establish if B. thetaiotaomicron could alter the gut function of specific pathogen-free rats and increase their resistance to a pathogen challenge. Bacteroides generally appear in the small intestine at the suckling-weaning transition (Chang et al, 1994; Hooper et al, 2001). The bacterium was therefore given daily to specific pathogen free Hooded Lister rats from weaning (19 d) through to maturity (34-40 days) to ensure that its presence in high numbers during this period of major gut and immune development. At 34 days of age, the rats were orally challenged with Salmonella enterica var. Enteritidis S1400.
B. thetaiotaomicron, from Deutsche Sammlung von Microorganismen und GmBH (Braunschweig, Germany) was maintained frozen in Wilkins-Chalgren anaerobic agar. It was subcultured into Wilkins-Chalgren anaerobic broth or M10 medium [10 ml in Hungate tubes (Bryant, 1972)] and grown at 37°C for 48 hours. A sample (0.5 ml) was transferred to fresh media [10 ml in Hungate tubes] and grown at 37°C for 24 hours. This culture contained around 108-109 CFU ml-1.
Ten ml of B. thetaiotaomicron culture was centrifuged (2000g, 12 min), the pellet was washed with 0.05M phosphate buffered saline pH 7.2 and then resuspended in commercial jelly [10 ml, 37°C, prepared under anaerobic conditions]. The jelly was poured into a sterile petri dish under CO2 and allowed to set at 4°C. Approximately 1 hour later, the jelly was cut into weighed amounts (0.5g, ∼ 108 CFU B. thetaiotaomicron) and fed to the rats. The bacterium remained viable for at least 6 hours under these conditions. However, usually all jelly was eaten within a few minutes of being placed in the cage.
Salmonella enterica serovar Enteritidis S 1400 was originally isolated from poultry infection and has been characterised (Allen-Vercoe and Woodward, 1999). Stocks, maintained on Dorset egg slopes at 4°C, were sub-cultured on to Luria-Bertani agar plates and grown at 37°C overnight. Five-ten colonies from the plate were inoculated into 10 mls of Luria-Bertani broth and incubated with agitation at 37°C overnight to give approximately 1 x 109 CFU ml-1.
Animal studies: The Rowett Research Institute is licensed under the UK Animals (Scientific Procedures) Act 1986. The ethical review committee and the animal welfare unit of the institute and the appropriate governmental inspectorate monitor and review all animal studies. The management and experimental procedures undertaken were approved by the ethical committee and done in strict accordance with the requirements of the Act by staff licensed to carry out such procedures.
Twenty-four male specific pathogen-free Hooded Lister rats (40 g), bred in the small animal unit of the Rowett Research Institute, were weaned at 19 days and immediately transferred to a Class II facility. They were housed individually in metabolism cages (Techniplast, Kettering, UK) within flexi-film isolators (Moredun, Animal Health, Penicuik, UK) for the duration of the study (21 days). Nesting tunnels were provided in each metabolism cage and cages were arranged within the isolators to allow sight of but not contact with other animals. Sterile distilled water was available at all times. Rats were weighed daily and faeces was collected throughout the study. Experiments were undertaken at least 3 times with similar results. Data are the means ISD (n=3). The rats were initially given free access to a high quality semi-synthetic (100 g protein kg-1) lactalbumin-based diet (Grant et al, 2000). When they reached approximately 80 g (around 30 days old) their food intake was gradually reduced over a 3-4 day period to 7g rat-1 day-1, given as two feeds over the day, and then maintained at this level for the remainder of the study. This was the average daily free intake observed for rats of the same age after oral infection with Salmonella enterica var. Enteritidis and around 70% of the free intake of this diet by non-infected animals. The rats were housed and managed in this manner from weaning to reduce environmental exposure to bacteria and cross-contamination post-weaning (Grant, 1996).
Twelve of the rats were given B. thetaiotaomicron (approximately 108 CFU) once daily from weaning (19d) through to completion of the study. At 34 days of age, six rats pre-treated with B. thetaiotaomicron and six controls were given, by gavage, a single dose of 0.8 ml of Salmonella serovar Enteritidis S1400 culture (approximately 109 CFU). The remaining rats were given the equivalent amount of culture medium. All rats were returned to their cages and fed lactalbumin-based diet (7g rat-1 day-1) for a further 6 days. On the final day, the rats were fed 2 g of diet and were killed by halothane (Rhone Merieux, Essex, UK) overdose and exsanguination exactly two hours later. The abdomen was opened and tissues were aseptically removed.
The stomach and the small intestine were flushed with phosphate buffered saline (PBS, pH 7.2) to remove contents and non-adherent bacteria. Ten cm of jejunum (5-15 cm from the pylorus) and 10 cm of ileum (5-15 cm from the ileo-caecal junction) were removed. These and the stomach tissue, caecum plus contents, colon plus contents, mesenteric lymph node, a representative proportion of the liver and spleen (200-400mg) and one kidney were processed for viable counts.
Further pieces of jejunum and ileum (15-25 cm from the pylorus and 15-25 cm from the ileo-caecal junction respectively) were collected and frozen in liquid nitrogen for biochemical analysis, as was the remaining small intestine tissue (∼ 40 cm), liver, spleen, kidney, thymus, lungs, heart and gastrocnemius hind-limb muscles. These were weighed, freeze-dried and reweighed.
Tissue samples were weighed and then homogenised in Maximum Recovery Diluent (MRD, Fisher Scientific, UK) using a Janke-Kunkel Ultra-Turrax T25 tissue homogeniser at 20,000 rev min-1 for 30 seconds. Up to six sequential dilutions (1:10 v/v) of the primary homogenate were made in MRD. Samples of each dilution were plated onto the surface of well-dried XLD agar (Fisher Scientific, UK) and MacConkey agar No. 3 (Fisher Scientific, UK) plates and incubated overnight at 37°C. Viable counts were estimated by the method of Miles and Misra (1938) or by a spread plate method (Collins and Lyme, 1989).
DNA was extracted and purified using a QIAamp DNA Stool Mini Kit (Qiagen Ltd, Crawley, UK). The PCR was based on the method of Teng et al. (2000). The primer pair: 5'-TGGAGTTTTACTTTGAATGGAC-3' (BTH-F) and 5'CTGCCCTTTTACAATGGG-3' (BTH-R), identified by Teng et al. (2000), were purchased from Sigma-Genosys Ltd (Cambridge, UK). The reaction mixture (50 µl), based on reagents from a Taq PCR Core Kit (Qiagen Ltd, Crawley, UK), contained 50 pmol of both primers, 10 nmol dNTP mixture, 5 µl of 10x Qiagen PCR buffer, 10 µl 5X Q-Solution, 10 µl sample and 1.25U Taq DNA Polymerase. A hot-start PCR program was used. Reaction mix without Taq DNA Polymerase was heated at 94°C for 15 seconds and through 35 cycles of 94°C for 10 seconds, 55°C for 30 seconds and 74°C for 1 minute, followed by 1 cycle of 74°C for 2 minutes and 45°C for 2 seconds. Amplicons were separated on an agarose gel (10g/l) containing ethidium bromide (1 µg/ml gel). A single amplicon (721 bp) was obtained with DNA from faeces and B. thetaiotaomicron. The sequences of these amplicons appeared identical to each other and to the published sequence (Teng et al, 2000).
Tissue samples were homogenised (1:80 w/v) in ice-cold 5mM potassium phosphate buffer pH 6.0 using a Janke-Kunkel Ultra-Turrax T25 tissue homogeniser at 20,000 rev min-1 for 30 seconds and centrifuged (3000g x 30 min, 4°C). The pellet (1:20 w/v) was sonicated (3 x 5 sec) in ice-cold 0.5M potassium phosphate pH 6.0 containing hexadecyltrimethyl-ammonium bromide (HETAB, 5 g/l) and ethylenediaminetetraacetic acid (EDTA, 3.72g/l), left on ice for 30 min and centrifuged (3000g x 30 min, 4°C) [Stucchi et al, 2000]. The supernatant was frozen until assayed. Myeloperoxidase (MPO) activity was determined by monitoring H2O2 dependent oxidation of 3,3',5,5'-tetramethylbenzidine (TMB, Dynex Technologies, Ashford, UK) in 50mM potassium phosphate buffer pH 6.0 (Zimmerman and Granger, 1990). Absorbance at 450 nm was measured after termination of the reaction with 0.18M H2S04. Human myeloperoxidase (Calbiochem, UK) was used as Human myeloperoxidase (Calbiochem, UK) was used as a standard and values were expressed as MPO equivalents.
Immunoreactive MPO in intestinal contents or faeces was determined by a competitive ELISA method. Microtiter plates (Immulon 4, Dynex Technologies, Ashford, UK) were coated (10ng/well] with human myeloperoxidase in 10mM phosphate buffered saline pH 7.4 [PBS] overnight at 4°C. After washing [PBS containing Tween-20 (1ml/l], plates were blocked with PBS containing bovine serum albumin (BSA, 10g/l) for 1 hour at room temperature. Plates were washed and samples or standards added [50µl/well] and serially diluted in PBS containing BSA (1g/l), Tween-20 (1ml/l) and leupeptin (1mg/l). Rabbit anti-human myeloperoxidase antibody [Calbiochem, UK, 50µl/well of 1:4000 dilution] was also added. The plates were incubated for 1 hour, washed and then reacted for 1 hour with biotinylated anti-rabbit IgG followed by Extravidin/peroxidase (EXTRA-3 kit, Sigma-Aldrich, Poole, UK). After washing, TMB reagent (7µl/well) was added and incubated in the dark for up to 2 hours. The reaction was stopped by addition (50µl/well) of 0.18M H2SO4 and the absorbance read at 450nm. Values were expressed as immunoreactive MPO equivalents.
IgG was determined by competitive ELISA. Microtiter plates were coated [1µg/well] with rat IgG (Sigma-Aldrich, Poole, UK) in PBS overnight at 4°C. After blocking and washing of the plates, samples or rat IgG standards [50µl/well] were added and serially diluted in PBS containing BSA (1g/l) and Tween-20 (1ml/l). Biotinylated anti-rat IgG antibodies (Sigma-Aldrich, Poole, UK; 50µl/well of 1:1000 dilution) were also added and the plates incubated for 1 hour. Plates were then reacted with Extravidin/peroxidase and TMB reagent as per MPO immunoassay. Values were expressed as IgG equivalents.
IgA was determined by capture ELISA. Microtiter plates were coated with goat anti-rat IgA (Sigma-Aldrich, Poole, UK; 100µl/well of a 1:100 dilution in PBS) overnight at 4°C. After blocking and washing of the plates, samples or standards [100µl/well] were added, diluted in PBS containing BSA (1g/l), Tween-20 (1ml/l) and leupeptin (1 mg/l) and incubated for 1 hour. After washing, plates were incubated with mouse anti-rat IgA antibodies (Sigma-Aldrich, Poole, UK; 1:500 dilution) for 1 hour, rewashed and incubated with biotinylated anti-mouse IgG antibodies (1:1000 dilution) for a further 1 hour. Plates were then processed as per MPO immunoassay. Values were expressed as IgA equivalents.
Microtiter plates were coated with lipopolysaccharide (LPS; 10µg/well) from Salmonella enteritidis (Sigma Aldrich, Poole, Dorset). After blocking and washing, samples or standards [100µl/well] were added. Plates were washed one-hour later, biotinylated anti-rat IgG or anti-rat IgA was added and the plates were then processed as above. One unit of LPS-specific antibodies was defined as that contained in the volume (µl) giving an absorbance of at least 0.2 in the ELISA. The positive control was sera collected from rats 22 days after infection with SE: LPS-IgA, 3.2 x 104 units/ml; LPS-IgG, 1.7 x 105 units/ml.
Freeze dried faeces or intestine contents were extracted (1:10 w/v) in ice-cold 0.5M potassium phosphate pH 6.0 containing HETAB, EDTA and NaN3 (1g/l), left on ice for 30 min, centrifuged (3000g, 30 min, 4°C) and the supernatants were frozen.
Sample or leukocyte-elastase (Sigma-Aldrich, Poole, UK) was serially diluted in 0.2M Tris HCl pH 8.0 containing 1 M NaCl and leupeptin (1 mg/l) on a microtiter plate. Substrate (N-succinyl-ala-ala-val-p-nitroanilide, 0.2g/l, Sigma-Aldrich, Poole, UK) was added and the absorbance at 405 nm monitored immediately and at intervals up to 20 hours during incubation at 37°C. Values were expressed as leukocyte-elastase equivalents.
DNA, RNA, protein were determined as before by the diphenylamine, orcinol and modified Lowry methods (Grant et al, 2000). Salmon testes DNA, yeast RNA and bovine serum albumin were used as standards.
Data were assessed by one way analysis of variance (ANOVA) in combination with the Tukey multiple comparison test using the Instat Statistical Package (GraphPad Software Inc., San Diego, USA).
Bacteroides thetaiotaomicron (BT) was detected in faeces samples collected from BT-treated rats. BT was also found in faeces of control rats. However, the levels (equivalent of approximately 104 CFU/g faeces) were below those found in faeces from rats treated daily with BT (equivalent of approximately 106 CFU/g faeces).
Body weight gains, major organ weights and small intestine and liver compositions for BT rats were similar to those for controls (Tables 1-4). The general distribution of lactose fermentors and non-lactose fermentors in the gastrointestinal tract seemed to be unaffected and, as with controls, no lactose fermentors or non-lactose fermentors bacteria were detected in mesenteric lymph node, liver, spleen or kidney samples. In addition, the levels of non-specific IgA, elastase, and immunoreactive MPO in intestinal contents and faeces, MPO in intestine tissue and non-specific IgG and IgA in serum were comparable to those samples collected from control animals.
S. enteritidis [SE] was found throughout the gastrointestinal tract six days after oral challenge (Table 1). In addition, they were detectable in the mesenteric lymph node, liver and spleen. Total non-lactose fermentors had a similar tissue distribution in these rats (Table 1). The numbers of lactose fermentors in the gut did not appear to be affected by salmonella infection. However, significant numbers of lactose fermentors did appear in the mesenteric lymph nodes, liver and spleen of SE-infected rats (Table 1).
The numbers of viable salmonella detected in the liver and spleen after challenge with salmonella were greatly reduced if rats had also been treated with B. thetaiotaomicron [SE+BT] (Table 1). Total non-lactose fermentors in these tissues were also lowered but lactose fermentor numbers were unchanged. The levels of salmonella, total non-lactose fermentors and lactose fermentors throughout the gastrointestinal tract and in the mesenteric lymph node were similar for SE+BT and SE (Table 1). In addition, faecal excretion of salmonella over the 6 day experiment occurred at comparable rates in pathogen-infected rats (SE+BT, 7.0±0.8 Log10CFU/g d-1, SE, 6.2±0.6 Log10CFU/g d-1).
*SE+BT differs significantly from SE (p≤0.05)
The weight of the small intestine was increased as a result of salmonella infection (Tables 2,3). This was most marked in the ileal tissue and appeared to be due to accumulation of water and increases in the protein, RNA and DNA content of the tissue (Table 3). Treatment with BT slightly modified the response to infection. Thus, the water content of small intestine from SE+BT [748±22 mg/g] was considerably lower than that in tissue from SE rats [799±17 mg/g]. Indeed, it was similar to the levels in tissue from non-infected rats [BT, 745±24 mg/g controls, 760±10 mg/g]. In contrast, there was nc significant difference between SE and SE+BT rats in their small intestine dry weights or DNA, RNA and protein contents.
Myeloperoxidase (MPO) activity was significantly elevated in ileal tissue collected from SE rats (Table 4). In addition, immunoreactive-MPO and elastase activity in intestinal contents and faeces were also greatly increased (Table 4). In contrast, the levels of these enzymes in equivalent samples from SE+BT rats were similar to or only slightly higher than control samples.
Immunoreactive non-specific IgA in intestine contents, faeces and serum was increased as a result of salmonella inflection (Table 4). There were however no differences between SE+BT and SE in these IgA responses. Small amounts of LPS-specific IgA were also detected in the blood and intestine contents. Again, there were no significant differences between SE+BT and SE, although there was a tendency for LPS-specific IgA levels in the intestine to be higher for SE+BT rats.
Gastrocnemius hind-limb muscle weights were significantly lower in SE rats than in controls (Table 2). This was not evident in SE+BT rats and appeared to be a result of a reduced muscle accretion. SE rats deposited approximately 6.6±5.7 mg wet weight (1.4±1.2 mg dry weight] of gastrocnemius muscle per day whereas with SE+BT the accretion rate was 17.1±4.1 mg wet weight [4.2±1.0 mg dry weight] per day. Deposition of this muscle in controls was 15.5±2.9 mg wet weight [3.8±0.7 mg dry weight] per day and 17.0±4.2 mg wet weight [4.2±1.0 mg dry weight] per day in BT rats. This suggests that the SE rats deposited an average of 0.3g (wet weight) total skeletal muscle daily whilst with SE+BT, BT or control rats the levels were around 0.7-0.8 g daily (
Liver, spleen and mesenteric lymph node wet weights were significantly elevated in salmonella-infected rates (Table 3). This was due primarily to accumulation of water in the liver and spleen and an increase in both water and dry matter in the mesenteric lymph node (Table 4). Peroxidase activity in the mesenteric lymph node of infected rats was also elevated [SE, 31.0±25.0; SE+BT, 37.7±23.8; BT, 0.9±0.9; control, 0.4±0.3 µg]. There were however no significant differences between the SE and SE+BT groups in these tissue parameters.
Salmonellosis: Salmonella enterica var. Enteritidis S1400 colonised the whole gastrointestinal tract of rats and translocated to the mesenteric lymph node, liver and spleen. In addition, the wet and dry weight of the small intestine was greatly increased as a result of infection. This was most marked in the ileum and was linked to accumulation of water and higher levels of protein, RNA and DNA in the tissue. The basic characteristics of the infection were therefore similar to those seen for other Enteritidis or S. enterica var. Typhimurium strains in the rat models (Naughton et al, 1996; 2001; Ewen et al, 1997; Bovee-Oudenhoven et al, 1999; Islam et al, 2000; Havelaar et al, 2001).
Enteritidis and Typhimurium cause a self-limiting infection in rats, that is localised primarily to the gastrointestinal tract with invasion occurring via the ileum and only limited systemic spread being evidence (Naughton et al, 1996; Bovee-Oudenhoven et al, 1997; 1999; Islam et al, 2000). Furthermore, severe bacteremia and death is rare, unless the health status or gut integrity of the rats has been compromised by other factors prior to infection. This contrasts starkly with salmonellosis in some mouse models in which these strains elicit a severe typhoid-like illness with a high incidence of mortality (Lu et al, 1999; Kingsley and Baumler, 2000; Schechter and Lee, 2000). Thus, salmonellosis in the rat has strong similarities to the self-limiting gastroenteritis-type infections common in humans and domesticated animals infected by Enteritidis or Typhimurium.
In the present study, salmonella infection was found to provoke a strong inflammatory response in the distal but not the proximal small intestine. Thus, the activity of myeloperoxidase (MPO, a neutrophil marker) in ileal tissue was significantly elevated. In addition, levels of immunoreactive-MPO and leukocyte elastase activity in intestinal contents and faeces were increased. This is compatible with infiltration of polymorphonuclear leukocytes and other inflammatory cells into ileal tissue and the exvasion of inflammatory cells into the intestinal lumen, as previously observed in animal models of salmonellosis (Naughton et al, 1995; 1996; Vassiloyanakopoulos et al, 1998; Darwin and Miller, 1999; Henderson et al, 1999). In addition, it is consistent with the ileum being the primary site of colonisation and invasion by salmonella (Carter and Collins, 1974; Naughton et al, 1996).
Non-specific IgA in serum, intestine contents and faeces was significantly increased as a result of salmonella infection. Furthermore, there were indications of development of an LPS-specific IgA response by 6 days post-infection. In addition, faecal dry matter, water and protein outputs were elevated, as were the levels of protein in the intestinal contents. Liver and spleen weights were also increased as a result of the infection.
Salmonellosis was also found to impair skeletal muscle metabolism in rats. The infected animals continued to deposit skeletal muscle but the daily accretion rates were about 40-50% of those observed in controls, despite both sets of rats having the same dry matter (7g/rat d-1) and protein (0.7g/rat d-1) intake. This may have been due to diversion and utilisation of nutrients to support defensive responses against infection (Klasing and Calvert, 1999) and/or the actions of endotoxin or other bacterial factors on muscle protein synthesis (Friman et al, 1984; Lang et al, 2000).
B. thetaiotaomicron: In the present study, rats were exposed to high levels of exogenous B. thetaiotaomicron throughout a period of rapid growth and maturation of the gut and development of the immune system. Some Bacteroides strains are opportunistic pathogens and would be likely to have deleterious effects during this developmental period. However, as in studies with mature ex-germ-free animals monocontaminated with B. thetaiotaomicron (Hooper et al, 1999; 2001; Noack et al, 2000), there were no indications that B. thetaiotaomicron adversely affected gut or systemic metabolism of the specific pathogen-free rats. Indeed, all the parameters monitored in rats treated with B. thetaiotaomicron alone, including inflammatory markers and immunoglobulin levels, were similar to those obtained for controls.
B. thetaiotaomicron and salmonellosis: B. thetaiotaomicron modified the nature of the infection caused by Enteritidis S 1400 in rats and reduced its overall severity. In particular, inflammatory responses in the small intestine were limited, the numbers of viable salmonella found in the liver and spleen were greatly reduced and skeletal muscle accretion rates were restored to around control levels. B. thetaiotaomicron thus had effects both locally in the gut and at remote systemic tissues.
Some aspects of salmonellosis were however unaffected by B. thetaiotaomicron. Enteritidis numbers in the gastrointestinal tract and mesenteric lymph nodes and faecal excretion of the pathogen were unaltered. In addition, the enlargement of the small intestine (increase in dry weight and DNA, RNA and protein content) associated with Enteritidis infection was also evident in rats given B. thetaiotaomicron and Enteritidis. This suggests that B. thetaiotaomicron does not interfere directly with salmonella itself or block its general effects on host metabolism. Instead, the bacterium may selectively modulate host-responses against infection, possibly targetting those that are potentially detrimental to gut integrity.
A number of bacterial strains have recently been shown to give partial protection against Enteritidis or Typhimurium infection. Some outcompete the pathogen for attachment sites or nutrients in the gut or produce bactericidal compounds. As a result, they reduce the numbers of salmonella found in the intestine and reaching the mesenteric lymph nodes, liver and spleen (Bernet-Camard et al, 1997; Hudault et al, 1997; 2001; Hendriksson and Conway, 2001). In contrast, many have little or no effect on the numbers of salmonella in the gut. However, as with B. thetaiotaomicron, they still significantly ameliorate the pathogenic infection (Silva et al, 1999; Filho-Lima et al, 2000; Schu et al, 2000; Hendriksson and Conway, 2001; Maia et al, 2001). Protection given by the latter strains is thus by mechanisms other than competitive exclusion or secretion of bactericidal compounds.
Salmonella breach the intestine epithelium by interfering with cell metabolism, cell-cell interactions and host-response mechanisms (Darwin and Miller, 1999; Netea et al, 2000; Eaves-Pyles, 2001; Gewirtz et al, 2001; Lu and Walker, 2001; Ohl and Miller, 2001). In particular, they trigger rapid infiltration of polymorphonuclear leukocytes into the tissue and provoke acute inflammation and severe disruption of the gut (Madara, 1997; Ohl and Miller, 2001). Neutrophil infiltration into the gut is mediated by chemoattractant chemokines, such as IL-8, that are secreted by epithelial cells in response to pathogenic infection (McCormick et al, 1995; Madara, 1997; Darwin and Miller, 1999; Fleckstein and Kopecko, 2001). Release of this chemoattractant was high if epithelial cells were cultured with salmonella in vitro (McCormick et al, 1995; Campbell et al, 2001). In contrast, its output was greatly reduced if the epithelial cells were cultured with salmonella and B. thetaiotaomicron in combination (Campbell et al, 2001).
Myeloperoxidase levels in ileal tissue and intestine contents of rats given B. thetaiotaomicron and Enteritidis were much lower than in comparable samples from animals dosed with salmonella alone. Thus, the inflammatory responses normally triggered in the intestine by salmonella, were apparently attenuated if animals had also been treated with B. thetaiotaomicron. The bacterium may, as found in vitro (Campbell et al, 2001), block salmonella-linked production of chemoattractant chemokines by epithelial enterocytes and thus prevent the recruitment of neutrophils into the tissue. By modulating this host-response in vivo, B. thetaiotaomicron may limit the degree of gut inflammation and damage that occurs as a result of infection and thereby preserve gut integrity. This would also reduce the demands of the gut for nutrients to support repair (Klasing and Calvert, 1999) and allow more to go to other tissues, such as the skeletal muscle.
Salmonella drain to the mesenteric lymph nodes, once they pass through the intestinal epithelium (Kingsley and Baulmer, 2000). They may then be cleared from there by macrophages or alternatively break out from the mesenteric lymph nodes and spread to the blood, liver and spleen (Kingsley and Baulmer, 2000; Lu and Walker, 2001). Significant numbers of Enteritidis were detected in the mesenteric lymph nodes, liver and spleen of infected rats. Treatment of animals with B. thetaiotaomicron greatly reduced the levels of viable pathogen found in the liver and spleen. However, it did not affect the Enteritidis numbers in the mesenteric lymph nodes. Therefore, B. thetaiotaomicron did not appear to limit invasion and drainage of salmonella to this site, despite its major effects on intestinal metabolism. Since pathogen levels were reduced in the systemic tissues, this may indicate that the main protective effects of B. thetaiotaomicron arise through changes in systemic metabolism.
Commensal bacteria are potent immuno-modulators (Herias et al, 1999; Talham et al, 1999; Scharek et al, 2000; Isolauri et al, 2001; Lu and Walker, 2001). In particular, treatment of mice with Bifidobacterium lactis appeared to enhance their immune function and responsiveness against salmonella (Schu et al, 2000). Phagocytotic activity in blood and peritoneal cells, lymphocyte mitogenic responsiveness and secretion of Typhimurium-specific antibodies were elevated. B. thetaiotaomicron may have similar effects on systemic metabolism. The lower numbers of salmonella in the liver and spleen of B. thetaiotaomicron-treated rats might therefore be the result of enhanced clearance of the pathogen from the blood and internal organs.
Lactose fermentors (E. coli), albeit in low numbers, were also found in the liver and spleen of Enteritidis-infected rats. However, unlike salmonella, the levels of these bacteria were not affected by B. thetaiotaomicron. There may therefore be some selectivity in the bacterial clearance promoted by B. thetaiotaomicron. Alternatively, because the E. coli are derived from the normal flora of the rats, they may not be readily recognised as potentially harmful, even although they are present at inappropriate tissue sites.
B. thetaiotaomicron may however reduce systemic spread of salmonella by acting locally on the gut. It is generally accepted that salmonella attach to the gut epithelium, invade through the tissue, drain to the mesenteric lymph nodes and may then spread to the liver and spleen (Kingsley and Baulmer, 2000). However, recent work suggests that systemic spread of the pathogen can occur by an alternative route. Pathogens sampled luminally by dendritic cells or in the subepithelium by CD18-expressing phagocytes may be transferred directly to the liver and spleen, without drainage through the lymphatic system (Varquez-Torres et al, 1999; Isberg and Barnes, 2000; Rescigno et al, 2001). As a result, salmonella found in the liver and spleen are likely to be derived from the two uptake routes. Recruitment of these phagocytes will probably be mediated by chemoattractant chemokines produced by epithelial cells (Izadpanah et al, 2001; Kellennan and McEvoy, 2001). Since B. thetaiotaomicron appears to limit neutrophil recruitment and inflammation in the gut by preventing secretion of the necessary chemoattractant chemokines (Campbell et al 2001), it may also suppress the chemokines required to recruit dendritic cells or CD18-expressing phagocytes and as a result prevent or limit uptake of salmonella by this system.
Orally administered B. thetaiotaomicron reduced the severity of infection caused by Enteritidis in rats. The numbers of viable salmonella in the liver and spleen were greatly reduced, skeletal muscle accretion rates were restored to normal and inflammatory responses in the small intestine were attenuated. This was due possibly to enhanced immune responsiveness and rapid clearance of salmonella at remote sites, such as the liver and spleen. However, it was also, at least in part, a result of local effects of the bacterium on the gut. Neutrophil recruitment and inflammation in the intestine was prevented or delayed by the action of B. thetaiotaomicron. It may also have blocked direct uptake of salmonella in phagocytes to the liver and spleen.
This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.