Old Herborn University Seminar Monograph

16. HOST MICROFLORA CROSSTALK

EDITORS: PETER J. HEIDT TORE MIDTVEDT VOLKER RUSCH DIRK VAN DER WAAIJ Old Herborn University Seminar Monograph 16

ISBN 3-923022-27-1 ISSN 1431-6579

COPYRIGHT © 2003 BY HERBORN LITTERAE ALL RIGHTS RESERVED NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER

EDITORS: Peter J. Heidt, Ph.D., B.M. Department of Animal Science Biomedical Primate Research Centre (BPRC) Lange Kleiweg 139 2288 GJ - Rijswijk The Netherlands

Tore Midtvedt, M.D., Ph.D. Department of Medical Microbial Ecology Karolinska Insttute von Eulers väg 5 S 171 77 Stockholm Sweden

Volker Rusch, Dr. rer. nat. Institute for Integrative Biology Kornmarkt 2 D-35745 Herborn-Dill Germany

Dirk van der Waaij, M.D., Ph.D. Professor emeritus, University of Groningen Hoge Hereweg 50 9756 TJ - Glimmen The Netherlands

Verlag wissenschaftlicher Schriften und Bücher Am Kornmarkt 2 Postfach 1664 D-35745 Herborn-Dill Germany Telephone: +49 - 2772 - 921100 Telefax: +49 - 2772 - 921101 Contents ———————————————————————————————————————

Participating authors V

I. THE GUT IMMUNE SYSTEM AND THE MUCOSAL BACTERIA (Agnes E. Wold) 1 Summary ……………………………………………………………….. 1 IgA ……………………………………………………………………… 1 T cells…………………………………………………………………... 2 Induction of mucosal immune responses……………………………….. 2 Importance of gut flora on the specific immune system ………………… 3 The transient nature of the response to gut bacteria …………………….. 4 Immune response to food proteins ……………………………………… 5 Oral tolerance …………………………………………………………… 5 Mechanisms for oral tolerance ………………………………………….. 6 The normal microflora and oral tolerance……………………………….. 7 Influence of the commensal flora on innate immunity…………………... 7 Different bacteria elicit different innate responses ……………………… 8 Are we intolerant because of a faulty intestinal microflora? …………….. 9 Literature ……………………………………………………………….. 10 II. DEFENSINS AND DEFENSIN-LIKE MOLECULES: ANTIBACTERIAL MODE OF ACTION (Hilde Ulvatne, Ørjan Samuelsen, and Lars Harry Vorland) 17 Summary ……………………………………………………………….. 17 Introduction …………………………………………………………….. 17 Discussion………………………………………………..…………….. 21 Conclusion ……………………………………………….…………….. 25 Literature ……………………………………………………………….. 26 III. DEFENSINS AND BACTERIA, A QUESTION OF "LIVE OR LET DIE"? (Lars-Göran Axelsson) 33 Summary ……………………………………………………………….. 33 Introduction …………………………………………………………….. 33 Innate immunity of the gastrointestinal tract ……………………………. 34 Antimicrobial peptides, molecular properties and mode of action ………. 35 Small intestinal microbial peptides ……………………………………… 36 Pharmaceutical applications …………………………………………….. 38 Literature ……………………………………………………………….. 38

I Contents (continued) ———————————————————————————————————————

IV. PHENOTYPIC EXPRESSIONS IN THE SMALL INTESTINE (Elisabeth Norin)4 3 Introduction …………………………………………………………….. 43 Cell kinetics …………………………………………………………….. 44 Intestinal motility ……………………………………………………….. 45 Short chain fatty acids ………………………………………………….. 45 Intestinal tryptic activity ………………………………………………… 46 Bilirubin and urobilins………………………………………………….. 47 Concluding remarks…………………………………………………….. 47 Literature ……………………………………………………………….. 47 V. MICROBIAL P450: DOES IT EXIST, AND WHAT CAN IT MEAN? (Tore Midtvedt) 51 Summary ……………………………………………………………….. 51 Nomenclature …………………………………………………………… 51 P450s and metabolism of xenobiotics ……………………..…………... 51 P450s and evolution ……………………………………………………. 52 Xenobiotic metabolism and substrate specificity ……………………….. 53 Localisation of P450s…………………………………………………… 54 P450s and microorganisms ……………………………………………. 54 P450s and specific members of the intestinal flora …………………….. 54 Literature ……………………………………………………………….. 55 VI. A LINK BETWEEN MUCOSAL REGULATORY LYMPHOCYTES AND CHILDHOOD FOOD ALLERGY (Simon H. Murch) 57 Introduction …………………………………………………………….. 57 The increase in the incidence of food allergies………………………….. 57 Does IgE or IgA determine sensitisation? ………………………………. 57 Genetic predisposition to food allergy ………………………………….. 58 Regulation of IgE responses ……………………………………………. 58 Multiple food intolerance, food-allergic dysmotility and the eosinophil response ………………………………………………………………… 59 Demographics of allergic sensitisation: The role of enteric challenges….. 59 Mucosal challenges in children born in developing countries…………… 60 Contrasting changes in early-life gut flora in the developed-world child . . 61 The development of enteric tolerance …………………………………… 62 Mechanisms of oral tolerance: A central role for NF-κB ……………….. 63 Is food allergy regulated to defective generation of regulatory lymphocytes? …………………………………………………………… 64 Literature ……………………………………………………………….. 65

II Contents (continued) ———————————————————————————————————————

VII. SYNBIOTIC TREATMENT IN CLINICAL PRAXIS (Stig Bengmark)6 9 Introduction …………………………………………………………….. 69 Fermented food has outstanding quality………………………………… 69 Flora boosts the immune system of the host ……………………………. 71 Prebiotic fibres are essential ……………………………………………. 72 Probiotics in diarrhoea in children ……………………………………… 73 Probiotics - and prebiotics - in inflammatory bowel disease (IBD)……... 74 Probiotics in Helicobacter pylori infections …………………………….. 75 Synbiotics in ICU patients ……………………………………………… 76 Flora important on all body surfaces……………………………………. 77 Gut ecology and health - future aspects…………………………………. 78 Literature………………………………………………………………… 79 VIII. DEVELOPING AN IN VITRO MODEL ON THE INVESTIGATION OF THE CROSSTALK AMONG BACTERIA, ENTEROCYTES AND LEUKOCYTES NEAR THE INTESTINAL MUCOSA (Alexandr Parlesak and Christiane Bode) 83 Summary …………………………………………………………..…... 83 Introduction …………………………………………………………….. 83 Material and methods …………………………………………………… 87 First results and conclusion …………………………………………….. 87 Literature ……………………………………………………………….. 89 IX. HOW MUCOSAL IMMUNITY IS CONTROLLED BY LOCAL FACTORS (Stefan C. Meuer)9 1 The author's literature concerning this field…………………………….. 93 X. OLD HERBORN UNIVERSITY SEMINAR ON HOST MICROFLORA CROSSTALK: MINUTES AND OVERVIEW OF THE DISCUSSIONS (Dirk van der Waaij)9 5 Elaine Vaughan: Approaches to investigate the diversity and functionality of intestinal microbes……………………………………... 95 Vanya E. Grant: Flow cytometry - can it help to analyse complex biosystems? …………………………………………………………….. 96 Hilde Ulvatne: Defensins and defensin-like molecules: Antibacterial mode of action ………………………………………………………….. 96 Lars-Göran Axelsson: Defensins and bacteria, a question of "live or let die?"…………………………………………………………. 98 Agnes Wold: Mucosal immunology ……………………………………. 99

III Contents (continued) ———————————————————————————————————————

Simon Murch: A link between mucosal regulatory lymphocytes and childhood food allergy ………………………………………………… 100 Stig Bengmark: Synbiotic treatment in clinical praxis…………………. 100 Elisabeth Norin: Phenotypic expressions in the small intestine ……….. 100 Barbara H. Iglewski: Quorum sensing in Pseudomonas aeruginosa …. 101 Alexander Parlesak: Developing an in vitro model on the investigation of the crosstalk among bacteria, enterocytes and leukocytes near the intestinal mucosa ……………………………………………………… 101

IV Participating authors ———————————————————————————————————————

Lars-Göran Axelsson, Ph.D., Microbiology and Tumorbiology Center, Gas­ trobiology Unit, Division of Molecular Pathology, Karolinska Institute, Box 280, Theorells väg 3, S-171 77 Stockholm, Sweden. Stig Bengmark, M.D., Ph.D., Lund University, Ideon Research Park, Schelevägen 18, S-22 370 Lund, Sweden. Stefan C. Meuer, M.D., Ph.D., Institute for Immunology, Ruprecht-Karls- University, Im Neuenheimer Feld 305, D-69210 Heidelberg, Germany. Tore Midtvedt, M.D., Ph.D., Department of Medical Microbial Ecology, Karolinska Institute, von Eulers väg 5, Box 60 400, S-17177 Stockholm, Sweden. Simon H. Murch, B.Sc. Ph.D., Royal Free and University College School of Medicine, Royal Free Campus, Rowland Hill Street, London NW3 2PF, UK. Elisabeth Norin, Ph.D., Microbiology and Tumorbiology Center, Karolinska Institute, von Eulers väg 5, S-171 77 Stockholm, Sweden. Alexandr Parlesak, Ph.D., Institute for Biological Chemistry and Nutrition, Department Physiology of Nutrition, Hohenheim University, Garbenstraße 28, D-70599 Stuttgart, Germany. Hilde Ulvatne Marthinsen, M.D., Ph.D., Department of Medical Micro­ biology, University Hospital of North Norway, N-9038 Tromsø, Norway. Dirk van der Waaij, M.D., Ph.D., Hoge Hereweg 50, 9756 TJ Glimmen, The Netherlands. Agnes E. Wold, M.D., Ph.D., Department of Clinical Bacteriology, Göteborg University, Guldhedsgatan 10a, S-413 46 Göteborg, Sweden.

V VI THE GUT IMMUNE SYSTEM AND THE MUCOSAL BACTERIA AGNES E. WOLD Department of Clinical Bacteriology, Göteborg University, Göteborg, Sweden.

SUMMARY The gut associated immune system harbours the vast majority of all lymphoid cells in the human body. This corresponds to the fact that most antigens reach us via the gut. The mucosa of the upper airways and the gastro-intestinal tracts also host a rich normal bacterial flora, which serves as a stimulus to the immune system. There are many indi­ cations that the normal intestinal microflora affects the way in which other antigens, e.g. food antigens and other innocuous antigens, are handled. Thus, alterations in intestinal colonisation pattern might have predisposed for allergies and other hypersensitivity reactions. In this review, the immune system of the gut and its relation to the normal in­ testinal microflora will be discussed.

IgA The lamina propria is seeded with as al., 1970; Underdown and Dorrington, many as 1010 antibody producing 1974). Secretory IgA efficiently pre­ cells/m of small intestine. Most of these vents microbial attachment to host mu­ are plasma cells that produce dimeric cosal structures (Freter, 1969; Williams IgA, i.e. two IgA monomers held to­ and Gibbons, 1972; Svanborg-Edén gether by the polypeptide “joining and Svennerholm, 1978) and, hence, chain” (Brandtzaeg, 1994). These IgA strongly reduces translocation (Albanese dimers bind to secretory component, et al., 1994; Maxson et al., 1995; also termed “polymeric immunoglobulin Dickinson et al., 1998). Secretory IgA receptor”, a trans-membrane protein that does not activate complement (Russell et is exposed on the basolateral aspect of al., 1997) and bacteria coated by intestinal epithelial cells. The entire secretory IgA are not killed. In contrast, complex is transported through the epi­ IgA can counteract the inflammatory thelium to the lumenal side, where the potential of IgG or IgM antibodies that transmembrane part of secretory com­ are bound to the same target. ponent is cleaved off. The largest part of It is known that many of the com­ secretory component remains bound to mensal microbes in the large intestine IgA, together forming the secretory IgA are coated by IgA, which does not seem complex (Mestecky et al., 1999). to affect them negatively (van Saene and The secretory IgA molecule is spe­ van der Waaij, 1979; van der Waaij, cially designed to afford protection on 1996). In fact, IgA coating might even surfaces populated by microbes. It is be advantageous to commensal bacteria. hydrophilic and highly resistant to pro­ The carbohydrate chains of secretory teolysis, much due to its rich substitu­ IgA function as receptors for the man­ tion with carbohydrate chains (Brown et nose-specific adhesin of type 1 fim­

1 briae, the most common adhesin of E. type 1-fimbriated E. coli, compared to coli and other enterobacteria (Wold et age-matched controls, and the E. coli al., 1990). A complete lack of IgA in strains retrieved from IgA-deficient in­ serum and secretions constitutes is the dividuals also expressed less of the most common primary immunodefi­ mannose-specific adhesin compared to ciency, afflicting some 1/500 individu­ E. coli from the control individuals als, two thirds of whom are healthy. We (Friman et al., 1996; Friman et al., have investigated the effect of absence 2002). This suggests that the interaction of IgA in secretions on the commensal between bacteria and secretory IgA is E. coli flora. In fact, individuals who advantageous to the bacteria. lack IgA have a lower proportion of

T CELLS T cells are found in two compart­ specificity produce predominantly IL-2 ments in the intestinal mucosa: in the in the spleen and Peyer’s patches, but lamina propria and between the epithe­ IFN-γ or IL-10 in the gut lamina propria lial cells (so-called intra-epithelial lym­ (Saparov et al., 1997). The function of phocytes). lamina propria T cells under physiologic The T cells found in the villus lamina conditions is unknown. propria are mainly CD4-positive and The intra-epithelial cells are mainly display signs of activation (deMaria et of the CD8+ phenotype. They carry ei­ al., 1993; Schieferdecker et al., 1992). ther the αβ or γδ type of antigen recep­ They do not respond with proliferation tor. The exact function of the intraepi­ to mitogens and other stimuli, indicating thelial cells is unknown, but they can that they are terminally differentiated. synthesise IL-2 and IFN-γ (Lundqvist et They secrete cytokines spontaneously, al., 1996) and lyse virus-infected cells especially interferon-γ (Hauer et al., (Cebra et al., 1989). Their T cell 1997, 1998). It appears as the micromi­ receptors are oligoclonal suggesting that lieu in the mucosa favours development the entire population derives from a of a certain T cell phenotype. Trans­ limited set of T cell clones (Blumberg et genic T cells with the same antigen al., 1993).

INDUCTION OF MUCOSAL IMMUNE RESPONSES Intestinal immune responses are in­ M-cells lack brush borders and the en­ duced in the Peyer' s patches, which are zymatic machinery of the absorptive mucosal lymphoid nodules situated in epithelial cell (Neutra et al., 1996). In the wall of the small intestine (Craig and the patches, particulate and soluble anti­ Cebra, 1971). The patches are covered gens are degraded and presented by by a specialised epithelium, the folli­ macrophages and dendritic cells. T cells cle-associated epithelium. This epithe­ and B cells with the appropriate specifi­ lium contains specialised epithelial cells, cities proliferate, mature and leave the termed M-cells, which are specialised in patches via the efferent lymph. After transporting material from the lumen circulating in the blood for a few days, into the patches, without degrading it. they return to the intestine, but not to the

2 Peyer’s patches, but to the lamina termed "homing" (Craig and Cebra, propria of the intestine, and, to a lesser 1971). extent, to other mucosa. This process is

IMPORTANCE OF GUT FLORA ON THE SPECIFIC IMMUNE SYSTEM The majority of all lymphoid cells in antibodies" directed against blood group the gut are there because of the normal antigens, that were thought by Land­ intestinal microflora. Germfree animals steiner to occur spontaneously as part of have only one tenth as many IgA-pro­ the normal physiologic development, ducing cells and T cells in the intestinal have been shown to result from immu­ lamina propria as conventional animals nisation by gut microbes in the normal (Crabbé et al., 1968, 1970; Hashimoto flora (Wiener, 1951; Springer and et al., 1978). Horton, 1969; Scheffel and Kim, Mucosa not regularly colonised by 1979). microbes, for example the respiratory Both live and dead bacteria can in­ and urinary tracts, have comparatively duce mucosal immune responses, but more IgG in their secretions compared live, colonising bacteria are better than to the upper respiratory and gastrointes­ dead ones, supposedly because more tinal tracts (Reynolds, 1988; Svanborg antigen is delivered to the immune sys­ Edén et al., 1985). IgG activates com­ tem with a bacteria which replicates in plement, thereby lysing bacteria but also the intestinal tract (Hohmann et al., eliciting inflammation. Thus, the con­ 1979). Provided, however, that high tinuous presence of a normal intestinal enough doses were given over long microflora in the upper respiratory tract enough periods of time, killed E. coli or and gut seems to promote development Bacteroides can induce equally large of immune effector functions that are amounts of IgA-containing plasma cells non-inflammatogenic and mainly pre­ as colonisation by live bacteria of the vent too close contact between the same species (Moreau et al., 1978). The commensal microbes and the host. How best inducers of antibody responses are this regulation occurs is not known. In bacteria, which are able to invade the mice, it appears as if switch from IgM mucosa, because larger doses of bacte­ can occur in the lamina propria in the rial antigens will come into contact with absence of T cell help (Fagarasan et al., the immune system (Hohmann et al., 2001), but whether this is true in hu­ 1979). This raises the question whether mans is unclear (Brandtzaeg et al., bacteria in the intestinal lumen which are 2001). not coated by IgA avoid this type of When germfree animals are colo­ immune response simply because they nised by a normal flora, lymph nodes are not able to translocate and, hence, and Peyer' s patches increase in weight they may not be “seen” by the immune and germinal centres develop, the serum system. Anaerobes usually do not IgG concentration rises and antibodies translocate (Berg, 1983), while live appear towards the colonising micro­ facultative bacteria can regularly be organisms (Carter and Pollard, 1971). found in the mesenteric lymph nodes These antibodies may often cross-react provided that their population levels in with other bacteria, and non-bacterial the large intestinal microflora reaches a structures. For example, the "natural certain level (Wells et al., 1987; Berg et

3 al., 1988; Herías et al., 1995, 1997). response to bacterial colonisation, while Moreau and co-workers attempted to the γδ type does not (Kawaguchi et al., determine whether certain types of in­ 1993). However, the γδ type of lym­ testinal bacteria were better than others phocytes produces IFN-γ in response to in triggering the mucosal immune sys­ lumenal bacteria, which in turn upregu­ tem. She colonised germfree mice with lates the expression of MHC class II a range of Gram-positive and Gram- molecules on intestinal epithelial cells negative bacteria and measured the den­ (Matsumoto et al., 1999). The oligo­ sity of IgA-containing plasma cells in clonality of intraepithelial cells does, the intestinal lamina propria. The best however, not seem to be determined by inducers of IgA plasma cells were E. distinct bacterial antigens, since both coli and Bacteroides, while all tested germfree and conventional mice exhibit Gram-positive species were inferior such restricted clonality (Regnault et al., (Moreau et al., 1978). Quite to the con­ 1996). trary, Cebra and co-workers found most People ingesting probiotic bacteria IgA to be produced in response to exhibit activation of cell-mediated im­ colonisation by Gram-positive Listeria mune effector functions, such as en­ monocytogenes or segmented filamen­ hanced phagocytosis and secretion of tous bacteria, while the Gram-negative IFN-γ by blood lymphocytes and IFN-α Morganella morganii, Ochrobacterium by blood mononuclear cells (Wold, atrophii and Helicobacter muridium all 2001). In addition, immunoglobulin­ gave less IgA stimulation (Cebra, 1999 producing cells with specificity against and personal communication). the administered strain appear in the Intraepithelial cells carrying the αβ blood (Wold, 2001). type of receptor increase in numbers in

THE TRANSIENT NATURE OF THE RESPONSE TO GUT BACTERIA When a bacterial strain successfully Because of this phenomenon, a per­ colonises the intestine and reaches sistent activation of the mucosal immune numbers high enough to permit translo­ system requires a high turnover of cation, germinal centres are formed in bacterial strains in the microflora. In the Peyer' s patches, B cells committed accordance, Pakistani infants, who are to IgA production seed the mucosa and colonised by a never-ending succession secretory IgA is produced into the in­ of new enterobacteria (Adlerberth et al., testinal lumen. However, this immune 1991, 1998) have higher secretory IgA response is self-limiting, in that the se­ levels in their saliva, and higher anti-E. cretory IgA so produced, coats the bac­ coli antibody levels than Swedish in­ teria in the intestinal lumen, preventing fants of the same age (Mellander et al., further translocation and, hence, stimu­ 1995). Bottle-fed infants, who have a lation of the gut lymphoid tissue (Shroff more varied and less stable microflora et al, 1995). Despite the continued pres­ than breast-fed infants and probably en­ ence of the microbe in the gut flora, counter more translocated bacteria, dis­ there will be no, or only minimal, fur­ play signs of increased immune respon­ ther stimulation of the gut associated siveness (Wold and Adlerberth, 2000). lymphoid tissue (Shroff et al, 1995).

4 IMMU N E RES PON S E TO FOOD PROTEIN S Food antigens provide little stimula­ hydrolysed liquid diet, but much lower tion to the immune apparatus. Although than in conventional mice (Hashimoto et an estimated 0.01 to 0.1% of ingested al., 1978). One must also bear in mind food proteins are taken up into the cir­ that even the sterilised feed given to culation in an intact, theoretically fully germfree animals is contaminated by immunogenic form (Husby et al., endotoxin and other bacterial compo­ 1985a), the immune response to food nents (Midtvedt and Gustafsson, 1981), proteins in humans is limited to low which may contribute to this low-grade levels of serum antibodies of the IgG4 immune stimulation. Rats fed very high and IgG2 subclasses (Husby et al., doses of protein antigens form serum 1985b). These subclasses are poor in IgG antibodies, but no secretory IgA fixing complement and interacting with response (Peri et al., 1982; Wold et al., phagocytes. Development of antibody 1987, 1989). Hence, food proteins lack responses dominated by IgG1 and the features that enable strong immune IgG3, antibody isotypes with strong responses to develop. When a plasmid inflammatogenic properties, may result encoding production of the food protein in food intolerance reactions (Saalman et ovalbumin was cloned into E. coli and al., 1995, 2001). this strain is used to colonise germfree Accordingly, serum IgA and IgG are rats, secretory IgA antibodies against only slightly increased in germfree mice ovalbumin were produced (Dahlgren et fed a commercial rat diet compared with al., 1991). those fed an "antigen-free" extensively

ORAL TOLERANCE Exposure of the mucosal immune sponse to the antigen after systemic ad­ system to food antigens normally results ministration than in non-fed individuals. in development of specific immu­ The original observations of oral toler­ nological tolerance to these proteins. ance were made in guinea pigs (Wells, This means that if the food antigen is 1911; Chase, 1946), but most of the later administered systemically, it will work on tolerance has thereafter been evoke less of an immune response than done in rats and mice (Thomas and Par- it would in an individual who had not rot, 1974; Hanson et al., 1977). More been fed the protein. Thus, local swel­ recently it was also shown that humans ling will not be elicited by intradermal who ingest KLH develop T cell toler­ injection of the antigen (so called de­ ance to this protein, although the anti­ layed-type hypersensitivity, which is a body response was intact (Husby et al., sign of the presence of memory T cells 1994). Oral tolerance may be seen as a to the antigen), and T cells taken from way to economise the resources of the blood or lymph nodes will fail to prolif­ immune system by avoiding to react to a erate when stimulated with the same an­ wealth of innocuous antigens entering tigen in vitro. IgE-mediated hypersensi­ via the mucosal membranes. Oral toler­ tivity reactions do not develop. There ance also protects us from dangerous may also be a weaker and/or more inflammatogenic responses that would short-lived serum IgG antibody re­ destroy mucosal architecture and func­

5 tion. Allergies and hypersensitivity re­ tolerance to environmental, innocuous actions may be seen as a failure of the antigens. individual to develop and/or maintain

MECHANISMS FOR ORAL TOLERANCE

Despite decades of research, it is still IL-10 and/or TGF-β, as well as contact­ very unclear where and how oral toler­ mediated mechanisms have been de­ ance is induced and by which mecha­ scribed (Smith et al., 2000). Suppressor nisms immune responses are sup­ cells induced in an animal fed a protein pressed in the tolerant animal or human antigen may be transferred into a naïve being (Smith et al., 2000). Oral toler­ recipient and suppress immune ance may operate through at least two responses in this animal. Nothing is different mechanisms: Anergy and sup­ known about the requirements for in­ pressor (regulatory) T cells. ducing T-suppressor cells, but one Anergy means that T helper cells are might speculate that the gut mucosa paralysed when they encounter their an­ provides a suitable environment for the tigen on an antigen-presenting cell maturation of T-suppressor cells. which does not simultaneously deliver A very interesting new model sug­ the activating signals that the T cell gests that the intestinal epithelial cell needs in order to proliferate and mature plays a key role in oral tolerance (Karls­ to an effector cell. These so-called co­ son et al, 2001). It has been known for stimulatory signals may be T cell acti­ a long time that serum from an animal vating cytokines (for example IL-1 and fed a tolerogenic dose of a protein anti- IL-12), but also a direct binding be­ gen can be transferred to a naïve animal, tween the antigen-presenting cell and the which will become tolerant to that pro­ T cell via so called accessory molecules. tein without having eaten it. The serum Both T cell activating cytokines and factor appears a few hours after feeding. accessory molecules are produced when Telemo has suggested that the serum antigen-presenting cells are exposed to factor consist of membrane fragments, microbial products. Thereby, all so called “tolerosomes” produced by the antigens in or on microbes will be intestinal epithelial cell. According to presented in a highly immunogenic this model, intact protein is taken up by fashion. Food proteins, on the other intestinal epithelial cells, processed to hand, lack the capacity to elicit such peptide fragments and loaded onto signals in the antigen-presenting cell, MHC class II molecules within the because they possess no “danger sig­ epithelial cell. Membrane vesicles are nals”. T cells that encounter their anti­ then budded off from the baso-lateral gens in the absence of activating signals facet of the enterocyte which on their during antigen-presentation may be surface carry MHC class II molecules paralysed or even receive a death signal. with loaded peptides. These membrane Suppressor, or regulatory, T cells vesicles may disperse themselves in the are formed somewhere in the gut-asso­ whole body via the lymph and blood ciated immune system. They are anti­ stream and merge with membranes of gen-specific and become activated by antigen-presenting cells in the mucosa, feeding the specific antigen. However, liver, or lymph nodes. Supposedly, instead of helping other T cells, they these vesicles contain information that will suppress their function by mecha­ ensures that the antigen is presented in a nisms yet to be defined – secretion of non-immunogenic, tolerogenic, fashion.

6 In addition to the above two mecha­ be counteracted. Many individuals, es­ nisms, antigen non-specific anti-in­ pecially those living in developing flammatory signals may down-regulate countries with a high infectious burden, immune effector functions. For exam­ have high levels of specific IgE anti­ ple, whether a delayed-type hypersensi­ bodies to environmental antigens, such tivity reaction develops or not depends as mites, and mount a wheal-and-flare not only on the presence of memory T reaction if mite antigen is injected in the cells, but also on the local conditions in skin. However, they do not have clini­ the skin where the antigen is injected. In cal symptoms of allergy, probably due the presence of anti-inflammatory cyto­ to the existence of active anti-inflam­ kines or other mediators, recruitment of matory mechanisms (Yazdanbakhsh et T cells to the site of antigen deposit may al, 2002).

THE NORMAL MICROFLORA AND ORAL TOLERANCE Oral tolerance is more short-lived in ventional animals, and the greatest T cell germfree compared to conventional activation is seen when the antigen animals (Moreau and Courthier, 1988). presenting cells derive from germfree In conventional animals, administration animals which are fed an antigen free of cholera toxin or E. coli heat labile liquid sterile diet (Hooper et al., 1995). toxin breaks oral tolerance to food anti­ A range of products secreted by macro­ gens (Elson and Ealding, 1984; Gabo­ phages in response to bacterial products reau-Routhiau and Moreau, 1996). have been shown to decrease the T cell Thus, it is clear that bacteria or their stimulating capacity of dendritic cells, products profoundly interfere with re­ for example the cytokines TNF-α (Holt sponses to food and other environ­ et al., 1993) and IL-10 (Koch et al., mental antigens. It is possible that the 1996), the prostaglandin E2 (Chouiab et gut microenvironment, especially in the al., 1985), and nitrous oxide (Holt et presence of the correct intestinal micro­ al., 1993). Thus, animals from which flora, provides a milieu where it is diffi­ alveolar macrophages have been re­ cult to activate T cells because the anti­ moved display greatly enhanced im­ gen-presenting cells in the gut do pro­ mune responses to inhaled antigens vide much of co-stimulatory signals. (Holt et al., 1993). Similarly, depletion This might have to do with antigen of macrophages from a preparation of processing. Accordingly, antigen pre­ dendritic cells from gut lamina propria senting cells from germfree mice are also enhances their antigen presenting stronger stimulators of naïve T cells ability (Pavli et al., 1990). than antigen presenting cells from con­

INFLUENCE OF THE COMMENSAL FLORA ON INNATE IMMUNITY Antigen presentation represents the lecular constellations that present “dan­ crossroads between the ancient innate ger” to the host, i.e. molecular patterns and the more modern acquired immune that are only found in prokaryotes. The system. The innate immune system re­ acquired immune system can be acti­ acts in a fixed fashion to certain mo­ vated by any structure, because the

7 variable regions of antibodies and T cell prime their macrophages for such cyto­ receptors have endless variability. kine production (Nicaise et al., 1993). However, because antigen-presenting The long ranging effects of the mi­ cells, which are monocytes, macro­ croflora on sites which will not be colo­ phages or dendritic cells, belong to the nised by bacteria, such as the sterile innate immune system, they have kept peritoneal cavity, spleen or bone mar­ the tendency to become activated by mi­ row, can be explained by the fact that crobial products. When they digest mi­ bacterial products are taken up by mu­ crobes and simultaneously present the cosal macrophages, which then leave antigens contained therein to T cells, the mucosa carrying their microbial they will convey activating signals to the components with them. Bacterial LPS T cells. In this way, bacteria function as may persist within macrophages in a their own adjuvans. bioactive form for very long periods of There is no question that the normal time (Duncan and Morrison, 1984). intestinal microflora substantially influ­ Peptidoglycans from bacterial cell walls, ences the entire innate immune system probably deriving from the intestinal (Wold and Adlerberth, 2000). Perito­ microflora, have been detected inside neal macrophages from conventional macrophages in the red pulp of the animals have increased levels of cyclic spleen in rats and humans (Kool et al., AMP and lysosomal enzymes, phago­ 1994; Hoijer et al., 1995). Breakdown cytose more avidly, secrete more of products of peptidoglycans have been oxygen radicals, and display enhanced detected in the urine of healthy people, cytotoxic activity, compared to macro­ indicating a constant uptake, degrada­ phages obtained from germfree animals tion and excretion of bacteria or their (Meltzer, 1976; Johnson and Balish, components from the intestine (Johans­ 1980; Podroprigora et al., 1980; Mor­ sen and Kreuger, 1988). land and Midtvedt, 1984; Mitsuyama et The pyrogenic and sleep-inducing al., 1986). Spleen and bone marrow properties of microbial break-down macrophages from conventional animals products (Johanssen et al., 1991; Martin produce more IL-1, IL-6 and TNF-α et al., 1984) may account for the fact when stimulated by LPS than macro­ that germfree animals have lower body phages from germfree animals (Nicaise temperature than conventional ones, and et al., 1993, 1995). Mono-colonisation are more difficult to anaesthetise of germfree animals with E. coli, but (Midtvedt, personal communication). not bifidobacteria, was reported to

DIFFERENT BACTERIA ELICIT DIFFERENT INNATE RESPONSES We have recently found that Gram- 2000). These two cytokines have positive and Gram-negative bacteria af­ largely opposing properties. IL-12 is a fect human monocytes very differently. T cell activating cytokine, which stimu­ Whereas Gram-positive bacteria stimu­ lates cell-mediated effector functions, late secretion of very large quantities of such as production of IFN-γ in T cells IL-12 from human monocytes, this is and NK cells. IL-10 instead dampens T not seen with Gram-negative bacteria, cell activation and IFN-γ production and which instead stimulate production of reduces antigen-presentation. large amounts of IL-10 (Hessle et al.,

8 Further, the two groups of bacteria preferentially induced by Gram-negative elicit different patterns of pro-inflam­ bacteria. Complement factors are acute matory cytokines and inflammatory me­ phase proteins, whose synthesis is diators in human monocytes (Hessle et stimulated above all by IL-6. Lastly, the al., in manuscript). Gram-positive bac­ strong PGE2 response seen to Gram- teria induce more TNF-α than Gram- negative, but not Gram-positive, bacte­ negative bacteria, whereas the latter in­ ria may facilitate leakage of plasma duce more of IL-6, IL-8 and PGE2. proteins, including antibody and com­ Probably, the two response patterns are plement, out of the microvasculature. optimally suited to facilitate the killing Gram-positive and Gram-negative and removal of the two types of bacte­ bacteria may, considering the above in­ ria. Gram-positive bacteria, with their dications, affect antigen presentation very thick and sturdy cell wall, may not rather differently. Whereas Gram-posi­ be digested efficiently enough by un­ tive bacteria induce IL-12 production in primed monocytes/macrophages. If the antigen-presenting cells that will en­ phagocyte is primed by TNF-α and hance so called Th1 reactions, the large IFN-γ, their lytic capacity is enhanced. amounts of PGE2, and the compara­ Gram-negative bacteria may preferen­ tively low amounts of IL-12, produced tially require soluble factors (antibody in monocytes stimulated by Gram- and complement) for their elimination. negative bacteria would instead favour Antibodies are produced by plasma cells differentiation of the T cells into the Th2 that mature under the influence of IL-4, pathway (Hessle et al., in manuscript). IL-6 and IL-10, the two latter being

ARE WE INTOLERANT BECAUSE OF A FAULTY INTESTINAL MICROFLORA? There are several diseases that in­ tions (Matricardi et al., 1997, 2000) crease in societies where the standard of have all been linked to high risk of de­ living and level of hygiene are high. veloping allergies, while exposure to This has been most clearly demonstrated early day-care (Kramer et al., 1999), to for allergies (Williams et al., 1994; von pets (Hesselmar et al., 1999), or to a Mutius, 1994), but is also seen with life-stock farm environment (Braun- inflammatory bowel disease (Langholtz Farlander et al., 1999) all protect against et al., 1991; Munkholm et al., 1992), allergy development. and perhaps some autoimmune The hygienic life-style of modern so­ disorders which appear at an earlier age ciety has also led to a change in the in­ today (Pundziute-Lycka et al, 2002). testinal colonisation pattern in infancy. These diseases have in common an Infants born in Sweden are later colo­ uncontrolled damaging immune re­ nised by E. coli and other enterobacteria sponse, which can be seen as a lack of than infants born in Pakistan are, and tolerance. they have a lower strain turnover in their Allergies are strongly linked to ex­ microflora over the first 6 months of life cessively hygienic life-styles, such as (Adlerberth et al., 1991, 1998). those characterising modern Western Colonisation by enterococci and lacto­ societies. Good housing standard, small bacilli are also delayed in Western, families (Strachan et al., 1989, compared to African, infants (Bennet et Strachan, 2000) and absence of infec­ al., 1991). This has led to the hypothe­

9 sis that a defect development of the colonisers suggests that there is reduced normal intestinal microflora in infancy is colonisation resistance afforded by the the cause of the allergy epidemics in the intestinal ecosystem, perhaps as a func­ Western world (Wold, 1988). tion of a poorly developed anaerobic The present composition of the nor­ flora. A low exposure of the Swedish mal flora may not be adequate in pro­ infant to faecal bacteria is suggested by moting the induction of oral tolerance, a delayed acquisition of E. coli and a through a number of potential mecha­ low E. coli strain turn-over (Nowrouz­ nisms: ian et al, 2002). E. coli is a bacterial 1) A delayed colonisation with certain species that is only found in the intesti­ key bacterial species might deprive nal tract of man and animals and its the developing immune system of presence is a sign of faecal contamina­ certain necessary signals, tion. Since it has no other reservoir in 2) Other bacteria that can expand in the nature, its spread is greatly hampered by microflora in the absence of certain hygiene. bacteria, may provide the wrong sig­ Thus, we have observed both a nals to the immune system, prevent­ change in which bacterial groups that ing oral tolerance to develop against are most numerous in the microflora, innocuous antigens, and the turnover rate of strains in the 3) The low overall antigenic stimulation intestinal microflora of infants. Inter­ afforded by an abnormally stable in­ estingly, infant who have low secretory testinal microflora may prevent suffi­ IgA levels in saliva during their first cient T cell activation to generate months of life are at increased risk of suppressor T cells. developing atopy (Payette et al., 1977; We have recently observed that in in­ van Asperen et al., 1985; Neffen et al., fants born in Sweden in the late 1990s, 1986). It remains to be discovered staphylococci have become the major whether the shift in microbial colonisa­ colonisers of the newborn infants. This tion pattern that seems to have occurred includes not only coagulase-negative as a result of a highly hygienic life­ staphylococci, but also S. aureus style, is the cause of the allergy epi­ (Lindberg et al, 2000). The emergence demic characterising affluent societies. of staphylococci as major intestinal

LITERATURE

Adlerberth, I., Carlson, B., de Man, P., Jalil, Development of the gastrointestinal tract F., Khan, S.R., Larsson, P., Mellander, L., (Eds.: Sanderson I.R., Walker, W.A.). B.C. Svanborg, C., Wold, A.E., and Hanson, Decker Publ. Inc., Hamilton, Ontario, 279­ L.Å.: Intestinal colonization with Entero­ 292 (1999). bacteriaceae in Pakistani and S wedish hos­ Albanese, C.T., Smith, C.D., Watkins, S., pital-delivered infants. Acta Paediatr. Scand. Kurkchubasche, A., Simmons, R.L., and 80, 602-610 (1991). Rowe, M.I.: Effect of secretory IgA on Adlerberth, I., Jalil, F., Hanson, L.Å., Carls­ transepithelial passage of bacteria across the son, B., Mellander, L., Larsson, P., Kalil, intact ileum in vitro. J. Am. Coll. Surg. K., and Wold, A.E.: High turn-over rate of 179, 679-688 (1994). Escherichia coli strains in the intestinal Bennet, R., Erikson, M., Tafari, N., and Nord, flora of infants in Pakistan. Epidemiol. In­ C.-E.: Intestinal bacteria of newborn ethio­ fect. 12, 587-598 (1998). pian infants in relation to antibiotic treat­ Adlerberth, I., Wold, A.E., and Hanson, L.Å.: ment and colonization with potentially The ontogeny of the intestinal flora. In: pathogenic Gram-negative bacteria. Scand.

10 J. Infect. Dis. 23, 63-69 (1991). allergy by prior feeding of the sensitising Berg R.D.: Translocation of indigenous bacteria agens. Proc. Soc. Exp. Biol. 61, 257-259 from the intestinal tract. In: Intestinal (1946). microflora in health and disease (Ed. : Hent­ Chouaib, S., Welte, K., Mertelsmann, R., and ges, D.J.). Acad. Press, New York, Dupont, B.: Prostaglandin E2 acts at two 333-352 (1983). distinct pathways of lymphocyte activation: Berg, R.D., Wommack, E., and Deitch, E.A.: Inhibition of interleukin 2 production and Immunosuppression and intestinal bacterial downregulation of transferrin receptor ex­ overgrowth synergistically promote bacte­ pression. J. Immunol. 135, 1172-1179 rial translocation. Arch. Surg. 123, (1985). 1359-1364 (1988). Crabbé P.A., Bazin, H., Eyssen, H., and Blumberg R.S., Yockey, C.E., Gross, G.G., Heremans, J.F.: The normal microbial flora Ebert, E.C., and Balk, S.P.: Human intes­ as a major stimulus for proliferation of tinal intraepithelial lymphocytes are derived plasma cells synthesizing IgA in the gut. from a limited number of T cell clones that Int. Arch. Allergy 34, 362-375 (1968). utilize multiple V beta T cell receptor Crabbé P.A., Nash, D.R., Bazin, H., Eysseb, genes. J. Immunol. 150, 5144-5153 H., and Heremans, J.F.: Immunohisto­ (1993). chemical observations on lymphoid tissues Brandtzaeg, P.: Distribution and characteriza­ from conventional and germ-free mice. Lab. tion of mucosal immunoglobulin-producing Invest. 22, 448-457 (1970). cells. In: Handbook of Mucosal Immunol­ Craig S., and Cebra, J.: Peyer's patches: An ogy (Ed.: Ogra, P.L.). Acad. press, San Di­ enriched source of precursors for ego, 251-262 (1994). IgA-producing immunocytes in the rabbit. Brandtzaeg, P., Baekkevold, E.S., and Morton, J. Exp. Med. 134, 188-200 (1971). H.C.: From B to A the mucosal way. Nat. Dahlgren, U.I.H., Wold, A.E., Hanson, L.Å., Immunol. 2, 1093-1094 (2001). and Midtvedt, T. : Expression of a dietary Braun-Farlander, C., Gassner, M., Grize, L., protein in E. coli renders it strongly anti­ Neu, U., Sennhauser, F.H., Varonier, H.S., genic to gut lymphoid tissue. Immunology Vuille, J.C., and Wütrich, B.: Prevalence of 73, 394-397 (1991). hay fever and allergic sensitization in deMaria R ., Fais, S., Silvestri, M., Frati, L., farmer's children and their peers living in Pallone, F., Santoni, A.., and Testi, R.: the same rural community. SCARPOL Continuous in-vivo activation and transient team. Swiss Study on Childhood Allergy hyporesponsiveness to TcR/CD3 triggering and Respiratory Symptoms with Respect to of human gut lamina propria lymphocytes. Air Pollution. Clin. Exp. Allergy 29, Eur. J. Immunol. 23, 3104-3108 (1993). 28-34 (1999). Dickinson, E.C., Gorga, J.C., Garrett, M., Brown, W.R., Newcomb, R.W., and Ishizaka, Tuncer, R., Boyle, P., Watkins, S.C., Al­ I.K.: Proteolytic degradation of exocrine and ber, S.M., Parizhskaya, M., Trucco, M., serum immunoglobulins. J. Clin. Invest. Rowe, M.I., and Ford, H.R.: Immuno­ 49, 1374-1380 (1970). globulin A supplementation abrogates bac­ Carter P.B., and Pollard, M.: Host responses to terial translocation and preserves the archi­ "normal" microbial flora in germ-free mice. tecture of the intestinal epithelium. Surgery J. Reticuloendothel. Soc. 9, 580-587 124, 284-290 (1998). (1971). Duncan, R.L.J., and Morrison, D.C.: The fate Cebra, J.J., Cebra-Thomas, J.A., Cuff, C.F., of E. coli lipopolysaccharide after the up­ George, A., Kost, S.I., London, S.D., and take of E. coli by murine macrophages in Rubin, D.H.: Reovirus as probes of the gut vitro. J. Immunol. 132, 1416-1424 (1984). mucosal T cell population. Immunol. In­ Elson, C.O., and Ealding, W.: Cholera toxin vest. 18, 545-558 (1989). feeding did not induce oral tolerance in mice Cebra J.J.: Influences of microbiota on intesti­ and abrogated oral tolerance to an unrelated nal immune system development. Am. J. protein antigen. J. Immunol. 133, Clin. Nutr. 69 (Suppl.):1046S-1051S 2892-2897 (1984). (1999). Freter, R.: Studies on the mechanism of action Chase, M.W.: Inhibition of experimental drug of intestinal antibody in experimental chol­

11 era. Texas Rep. Biol. Med. 27 (Suppl. 1), Wold, A.E.: Escherichia coli K5 capsule 299-316 (1969). expression enhances colonization of the Fagarasan, S., Kinoshita, K., Muramatsu, M., large intestine in the gnotobiotic rat. Infect. Ikuta, K., and Honjo, T.: In situ class Immun. 65, 531-536 (1997). switching and differentiation to IgA-produc­ Hesslemar, B., Åberg, N., Åberg, B., Eriksson, ing cells in the gut lamina propria. Nature B., and Björkstén, B.: Does early exposure 413, 639-643 (2001). to cat or dog protect against later allergy Friman, V., Hanson, L.Å., and Wold, A.E.: development? Clin. Exp. Allergy 29, 611­ Decreased expression of type 1 fimbriae by 617 (1999). E. coli in the colonic microflora of IgA-de­ Hessle, C., Andersson, B., and Wold, A.E.: ficient individuals. Infect. Immun. 64, Gram-positive bacteria are potent inducers 2794-2798 (1996). of IL-12 from human monocytes while Friman, V., Nowrouzian, F., Adlerberth, I., Gram-negative bacteria preferentially induce and Wold, A.E.: Increased carriage rate of IL-10 production. Infect. Immun. 68, 3581­ genes for S fimbriae and hemolysin in in­ 3586 (2000). testinal Escherichia coli from from IgA-de­ Hessle, C., Andersson, B., and Wold, A.E.: ficient individuals. Microb. Pathogen. 32, Gram-positive and Gram-negative bacteria 35-42 (2002). induce different patterns of pro-inflamma­ Gaborieau-Routhiau, V. and Moreau, M.-C.: tory cytokines. In manuscript. Gut flora allows recovery of oral tolerance Hessle, C., Andersson, B., and Wold, A.E.: to ovalbumin in mice after transient break­ Gram-negative, but not Gram-positive bac­ down mediated by cholera toxin or Es­ teria are strong inducers of PGE2 produc­ cherichia coli heat-labile enterotoxin. Ped. tion in human monocytes. In manuscript. Res. 39, 625-629 (1996). Hohmann A., Schmidt, G., and Rowley, D.: Hanson, D.G., Vaz, N.M., Maia, L.C., Horn­ Intestinal and serum antibody responses in brook, M.M., Lynch, J.M., and Roy, mice after oral immunization with Salmo­ C.A.: Inhibition of specific immune re­ nella, Escherichia coli and Salmo­ sponses by feeding protein antigens. Int. nella-Escherichia coli hybrid strains. Infect. Arch. Allergy Appl. Immunol. 55, 526-532 Immun. 25, 27-33 (1979). (1977). Hoijer, M.A., Melief, M.-J., van Hashimoto, K., Handa, H., Umehara, K., and Helden-Meeuwsen, C.G., Eulderink, F., and Sasaki, S.: Germfree mice reared on an Hazenberg, M.M.: Detection of muramic "antigen free" diet. Lab. Anim. Sci. 28, acid in a carbohydrate fraction of human 38-45 (1978). spleen. Infect. Immun. 63, 1652-1657 Hauer A.C., Bajaj-Elliot, M., Williams, C.B., (1995). Walker-Smith, J.A., and MacDonald, T.T.: Holt, P.G., Oliver, J., Bilyk, N., McMenamin, An analysis of interferon gamma, IL-4, C., McMenamin, P.G., Kraal, G., and IL-5 and IL-10 production by ELISPOT and Thepen, T.: Downregulation of the antigen quantitative reverse transcriptase-PCR in presenting cell function(s) of pulmonary human Peyer's patches. Cytokine 10, dendritic cells in vivo by resident alveolar 627-634 (1998). macrophages. J. Exp. Med. 177, 397-407 Hauer A.C., Breese, E.J., Walker-Smith, J.A., (1993). and MacDonald, T.T.: The frequency of Hooper, D.C., Molowitz, E.H., Bos, N.A., cells secreting interferon-gamma and inter­ Ploplis, V.A., and Cebra, J.J.: Spleen cells leukin-4, -5, and -10 in the blood and duo­ from antigen-minimized mice are superior denal mucosa of children with cow's milk to spleen cells from germ-free and conven­ hypersensitivity. Ped. Res. 42, 629-638 tional mice in the stimulation of primary in (1997). vitro proliferative responses to nominal an­ Herías, M.V., Midtvedt, T., Hanson, L.Å., and tigens. Eur. J. Immunol. 25, 212-217 Wold, A.E.: Role of Escherichia coli P (1995). fimbriae in intestinal colonization in gno­ Husby, S., Jensenius, J.C., and, J.C. Svehag, tobiotic rats. Infect. Immun. 63, 4781-4789 J.C.: Passage of undegraded dietary antigen (1995). into the blood of healthy adults. Quantifica­ Herías, V., Midtvedt, T., Hanson, L.Å., and tion, estimation of size distribution and re­

12 lation of uptake to levels of specific anti­ sen, O., Kreiner, S., and Binder, V.: Inci­ bodies. Scand. J. Immunol. 22, 83-92 dence and prevalence of ulcerative colitis in (1985a). Copenhagen County from 1962 to 1987. Husby S., Jensenius, J.C., and Svehag, S.-E.: Scand. J. Gastroenterol. 26, 1247-1256 ELISA quantitation of IgG subclass anti­ (1991). bodies to dietary antigens. J. Immunol. Lindberg, E., Nowrouzian, F., Adlerberth, I., Meth. 82, 321-331 (1985b). and Wold, A.E.: Long-term persistence of Husby, S., Mestecky, J., Moldovenau, Z., enterotoxin-producing Staphylococcus Holland, S., and Elson, C.O.: Oral toler­ aureus in the intestinal microflora of ance in humans. T cell but not B cell toler­ healthy Swedish infants. Ped. Res. 48, 741­ ance after antigen feeding. J. Immunol. 747 (2000). 152, 4663-4670 (1994). Lundqvist C., Melgar, S., Yeung, M.M., Johannsen, L. and Krueger J. M.: Quantitation Hammarström, S., and Hammarström, of diaminopimelic acid in human urine. M.-L.: Intraepithelial lymphocytes in hu­ Adv. Biosci. 68, 445-449 (1988). man gut have lytic potential and a cytokine Johannsen, L., Wecke, J., Obal, F., and profile that suggest T helper 1 and cyto­ Krueger J.: Macrophages produce somno­ toxic functions. J. Immunol. 157, genic and pyrogenic muramyl peptides dur­ 1926-1934 (1996). ing digestion of staphylococci. Am. J. Martin, S.A., Karnovsky, J.L., Krueger, J.M., Physiol. 260, R126-R133 (1991). Pappenheimer, J.R., and Biemann, K.: Pep­ Johnson, W. J. and Balish, E.: Macrophage tidoglycans as promoters of slow wave function in germ-free, athymic (nu/nu) and sleep. 1. Structure of the sleep-promoting conventional flora (nu+) mice. J. Reticu­ factor isolated from human urine. J. Biol. loendothel. Soc. 28, 55-66 (1980). Chem. 259, 7514-7522 (1984). Kawaguchi, M., Nanno, M., Umesaki, Y., Matricardi, P.M., Rosmini, F., Ferrigno, L., Matsumoto, S., Okada, Y., Cai, Z., Shi­ Nisini, R., Rapicetta, M., Chionne, P., mamura, T., Matsuoka, Y., Ohwaki, M., Stroffolini, T., Pasquini, P., and D'Amelio, and Ishikawa, H.: Cytolytic activity of in­ R.: Cross sectional retrospective study of testinal intraepithelial lymphocytes in prevalence of atopy among Italian military germ-free mice is strain dependent and de­ students with antibodies against hepatitis A termined by T cells expressing gamma delta virus. BMJ 314, 999-1003 (1997). T-cell antigen receptors. Proc. Natl. Acad. Matsumoto, S., Nanno, M., Watanabe, N., Sci. 90, 8591-8594 (1993). Miyashita, M., Amasaki, H., Suzuki, K., Koch, F., Stanzl, U., Jennewein, P., Janke, and Umesaki, Y.: Physiological roles of γδ K., Heufler, C., Kampgen, E., Romani, T-cell receptor intraepithelial lymphocytes N., and Schuler, G.: High level IL-12 pro­ in cytoproliferation and differentiation of duction by murine dendritic cells: Upregula­ mouse intestinal epithelial cells. Immunol­ tion via MHC class II and CD40 molecules ogy 97, 18-25 (1999). and downregulation by IL-4 and IL-10. J. Maxson, R.T., Jackson, R.J., and S mith, Exp. Med. 184, 741-746 (1996). S.D.: The protective role of enteral IgA Kool, J., de Visser, H., Gerritsboeye, M.Y., supplementation in neonatal gut origin sep­ Klasen, I.S., Melief, M.J., van sis. J. Pediatr. Surg. 30, 231-233 (1995). Helden-Meeuwsen, C.G., van Lieshout, Mellander, L., Carlsson, B., Jalil, F., Söder­ L.M., Ruseler-van Embden, W.B., van den ström, T., and Hanson, L.Å.: Secretory IgA Berg, W.B., Bahr, G.M., and Hazenberg, antibody response against Escherichia coli M.P.: Detection of intestinal flora-derived antigens in infants in relation to exposure. bacterial-antigen complexes in splenic J. Pediatr. 107, 430-433 (1985). macrophages of rats. J. Histochem. Cyto­ Meltzer, M.S.: Tumoricidal resonses in vitro of chem. 11, 1435-1441 (1994). peritoneal macrophages from conven­ Kramer, U., Heinrich, J., Wijst, M., and tionally housed and germ-free nude mice. Wichman, H.E.: Age of entry to day nurs­ Cell. Immunol. 22, 176-181 (1976). ery and allergy in later childhood. Lancet Mestecky J., Moro, I., and Underdown, B.J.: 353, 450-454 (1999). Mucosal immunoglobulins. In: Mucosal Langholtz, E., Munkholm, P., Haagen Niel­ Immunology (Eds.: Ogra, P.L., Mestecky,

13 J., Lamm, M.E., Strober, W., Bienenstock, and bone marrow-derived macrophages. Cy­ J., and McGhee, J.R.). Academic Press, tokine 7, 713-719 (1995). San Diego, 133-152 (1999). Nowrouzian, F., Hesselmar, B., Saalman, R., Midtvedt, T. and Gustafsson, B.E.: Digestion Strannegård, I.-L., Åberg, N., Wold, A.E., of dead bacteria by germ-free rats. Curr. Mi­ and Adlerberth, I.: Escherichia coli in in­ crobiol. 6, 13-15 (1981). fants' intestinal microflora: Colonization Mitsuyama, M., Ohara, R., Amako, K., No­ rate, strain turn-over and virulence gene car­ moto, K., and Yokokura, T.: Ontogeny of riage. Pediatr. Res, in press. macrophage function to relase superoxide Pavli, P., Woodhams, C.E., Doe, W.F., and anion in conventional and germfree mice. Hume, D.A.: Isolation and characterization Infect. Immun. 52, 236-239 (1986). of antigen-presenting dendritic cells from Moreau, M.C. and Corthier, G.: Effect of gas­ the mouse intestinal lamina propria. Im­ trointestinal microflora on induction and munology 70, 40-47 (1990). maintenance of oral tolerance to ovalbumin Payette, K. and Weiss, N,S.: Salivary IgA lev­ in C3H/HeJ mice. Infect. Immun. 56, els in atopic children. Ann. Allergy 39, 2766-2768 (1988). 328-331 (1977). Moreau, M.C., Ducluzeau, R., Guy-Grand, D., Peri, B.A., Theodore, C.M., Losonsky, G.A., and Muller, M.C.: Increase in the popula­ Fishaut, J.M., Rothberg, R.M., and Ogra, tion of duodenal immunoglobulin A plas­ P.L.: Antibody content of rabbit milk and mocytes in axenic mice associated with dif­ serum following inhalation or ingestion of ferent living or dead bacterial strains of in­ respiratory symcytial virus and bovine se­ testinal origin. Infect. Immun. 21, 532-539 rum albumin. Clin. Exp. Immunol. 48, (1978). 91-101 (1982). Morland B. and Midtvedt, T. : P hagocytosis, Podroprigora, G.I., Hoffman, J., Janecek, J., peritoneal influx and enzyme activities in and Naprstka, J.: Cyclic 3',5'-adenosine peritoneal macrophages from germ-free, monophosphate in macrophages, intestinal conventional and ex-germ-free mice. Infect. mucosa, and blood plasma of germ-free and Immun. 44, 750-752 (1984). ordinary animals. Bull. Eksp. Biol. Med. Munkholm, P., Langholtz, E., Haagen Niel­ 89, 17-19 (1980). sen, O., Kreiner, S., and Binder, V.: Inci­ Pundziute-Lycka, A., Dahlquist, G., Nyström, dence and prevalence of Crohn's disease in L., Arnqvist, H., Björk, E., Blohme, G., the County of Copenhagen, 1962-1987: A Bolinder, J., Eriksson, J.W., Sundkvist, sixfold increase in incidence. Scand. J. Gas­ G., Östman, J., and the Swedish childhood troenterol. 27, 609-614 (1992). diabetes study group: Type 1 diabetes in the Neffen, H., Crisci, C.D., Busaniche, H., and 0-34 years group in Sweden. Diabetologia Yanez, A.: Correlation between serum IgA, 45, 783-791 (2002). secretory IgA and total serum IgE in asth­ Regnault, A., Levraud, J.P., Lim, A. Six, A., matic and rhinitic affected patients. Aller­ Moreau, C., Cumano, A., and Kourilsky, gol. Immunopathol. (Madr.) 14, 413-418 P.: The expansion and selection of T cell (1986). receptor αβ intestinal intraepithelial T cell Neutra M.R., Frey, A., and Kraehenbuhl, J.P.: clones. Eur. J. Immunol. 26, 914-921 M-cells: Gateways for mucosal infection (1996). and immunization. Cell 86, 345-348 Reynolds, H.Y.: Immunoglobulin G and its (1996). function in the human respiratory tract. Nicaise, P., Gleizes, A., Forestier, F., Quero, Mayo Clin. Proc. 63, 161-174 (1988). A.M., and Labarre, C.: Influence of intesti­ Russell, M.W., Sibley, D.A., Nikolova, E.B., nal bacterial flora on cytokine (IL1, IL-6 Tomana, M., and Mestecky, J.: IgA anti­ and TNF-α) production by mouse perito­ body as a non-inflammatory regulator of neal macrophages. Eur. Cytokine Netw. 4, immunity. Biochem. Soc. Trans. 25, 133-139 (1993). 466-470 (1997). Nicaise, P., Gleizes, A., Forestier, F., Sandre, Saalman, R., Dahlgren, U.I., Ahlstedt, S., C., Quero, A.M., and Labarre, C.: The in­ Fällström, S.P., Hanson, L.Å., and Wold, fluence of E. coli implantation in axenic A.E.: IgG subclass profile of serum anti­ mice on cytokine production by peritoneal gliadin antibodies and antibody-dependent

14 cell-mediating cytotoxicity in young chil­ fect. Immun. 22, 790-797 (1978). dren with coeliac disease. Scand. J. Immu­ Thomas, H.C. and Parrot, D.M.V.: The induc­ nol. 53, 92-98 (2001). tion of toelrance to a soluble protein anti- Saalman, R., Dahlgren, U.I., Fällström, S.P., gen by oral administration. Immunology Hanson, L.Å., Wold, A.E., and Ahlstedt, 27, 631-639 (1974). S.: ADCC-mediating capacity in children Underdown, B.J. and Dorrington, K.J.: Studies with cow's milk protein intolerance in rela­ on the structural and conformation al basis tion to IgG subclass profile of serum anti­ for the relative resistance of serum and se­ bodies to β-lactoglobulin. Scand. J. Immu­ cretory immunoglobulin A to proteolysis. nol. 42, 140-146 (1995). J. Immunol. 112, 949-959 (1974). Saparov, A., Elson, C.O., Devore-Carter, D., van Asperen, P.P., Gleeson, M., Kemp, A.S., Bucy, R.P, and Weaver, C.T.: Single-cell Cripps, A.W., Geraghty, S.B., Mellis, analyses of CD4+ T cells from alpha beta T C.M., Clancy, R.L.: The relationship be­ cell receptor-transgenic mice: A distinct tween atopy and salivary IgA deficiency in mucosal cytokine phenotype in the absence infancy. Clin. Exp. Immunol. 62, 753-757 of transgene-specific antigen. Eur. J. Im­ (1985). munol. 27, 1774-1781 (1997). van Saene, H.K. and van der Waaij, D.: A Scheffel, J.W. and Kim, Y.: B. Role of envi­ novel technique for detecting IgA coated po­ ronment in the development of "natural" tentially pathogenic microorganisms in the hemagglutinins in Minnesota miniature human intestine. J. Immunol. 30, 87-96 swine. Infect. Immun. 26, 202-210 (1979). (1979). Schieferdecker H.L., Ullrich, R., Hirseland, H., van der Waaij, L.A., Limburg, P.C., Mesander, and Zeitz, M.: T cell differentiation anti­ G., and van der Waaij, D.: In vivo IgA gens on lymphocytes in the human intesti­ coating of anaerobic bacteria in human nal lammina propria. J. Immunol. 149, faeces. Gut 38, 348-354 (1996). 2816-2822 (1992). von Mutius, E., Martinez, F.D., Fritzsch, C., Shroff, K.E., Meslin, K., and Cebra, J.J.: Nicolai, T., Roell, G., and Thiemann, Commensal enteric bacteria engender a H.H.: Prevalence of asthma and atopy in self-limiting humoral mucosal immune re­ two areas of West and East Germany. Am. sponse while permanently colonizing the J. Respir. Crit. Care Med. 149 (2 Pt 1), gut. Infect. Immun. 63, 3904-3913 (1995). 358-364 (1994). Smith, K.M., Eaton, A.D., Finlayson, L.M., Wells, C.L., Maddaus, M.A., Reynolds, C.M., and Garside, P.: Oral tolerance. Am. J. Jechorek, R.P., and Simmons, R.L.: Role Respir. Crit. Care Med. 162, S175-S178 of anaerobic flora in the translocation of (2000). aerobic and facultatively anaerobic intestinal Springer, G.F. and Horton, R.E.: Blood group bacteria. Infect. Immun. 55, 2689-2694 isoantibody stimulation in man by feeding (1987). blood group-reactive bacteria. J. Clin. In­ Wells, H.: Studies on the chemistry of anaphy­ vest. 48, 1280-1291 (1969). laxis. III. Experiments with isolated pro­ Strachan, D.P.: Hay fever, hygiene and house­ teins, especially those of hen’s eggs. J. In­ hold size. BMJ 289, 1259-1260 (1989). fect. Dis. 9, 147-151 (1911). Strachan, D.P.: Family size, infection and Wiener, A.S.: Origin of naturally occurring atopy: The first decade of the “hygiene hy­ hemagglutinins and hemolysins: A review. pothesis”. Thorax 55 (Suppl 1), S2-S10 J. Immunol. 66, 287-295 (1951). (2000). Williams, H.C., Strachan, D.P., and Hay, Svanborg Edén, C., Kulhavy, R., Mårild, S., R.J.: Childhood eczema: Disease of the ad­ Prince, S.J., and Mestecky, J.: Urinary vantaged? BMJ 308, 1132-1135 (1994). immunoglobulins in healthy individuals Williams, R.C. and Gibbons, R.J.: Inhibition and children with acute pyelonephritis. of bacterial adherence by secretory immu­ Scand. J. Immunol. 21, 305-313 (1985). noglobulin A: a mechanism of antigen dis­ Svanborg-Edén, C. and Svennerholm, A.-M.: posal. Science 177, 697-699 (1972). Secretory immunoglobulin A and G anti­ Wold, A.E., Mestecky, J., Tomana, M., Ko­ bodies prevent adhesion of Escherichia coli bata, A., Ohbayashi, H., Endo, T., and to human urinary tract epithelial cells. In­ Svanborg Edén, C.: Secretory immuno­

15 globulin A carries oligosaccharide receptors changes in the intestinal flora? Allergy 53 for Escherichia coli type 1 fimbrial lectin. (46 Suppl), 20-25 (1998). Infect. Immun. 58, 3073-3077 (1990). Wold, A.E. and Adlerberth, I.: Pathologic ef­ Wold, A.E., Dahlgren, U.I.H., Ahlstedt, S., fects of commensalism. In: Persistent Bac­ and Hanson, L.Å.: Lack of IgA antibody re­ terial Infections (Eds.: Nataro, J.P., Blaser, sponse in secretions of rat dams during M.J., and C unningham-Rundles, S.). ASM long-term ovalbumin feeding. Induction of Publishing Co., 115-144 (2000). systemic tolerance in pups but not in adult Wold, A.E. and Adlerberth, I.: Breast feeding rats. Int. Arch. Allergy Appl. Immunol. and the intestinal microflora of the infant - 84, 332-338, 1987. Implications for protection against infec­ Wold, A.E., Dahlgren, U.I.H., Hanson, L.Å., tious diseases. Adv. Exp. Med. Biol. 478, Mattsby-Baltzer, I., and Midtvedt, T. : Dif­ 77-93 (2000). ference between bacterial and food antigens Wold, A.E.: Immune effects of probiotics. in mucosal immunogenicity. Infect. Im­ Scand. J. Nutr. 45, 76-85 (2001). mun. 57, 2666-2673 (1989). Yazdanbakhsh, M., Kremsner, P.G., and van Wold, A.E.: The hygiene hypothesis revised: Is Ree, R.: Allergy, parasites, and the hygiene the rising frequency of allergy due to hypothesis. Science 296, 490-494 (2002)

16 DEFENSINS AND DEFENSIN-LIKE MOLECULES: ANTIBACTERIAL MODE OF ACTION HILDE ULVATNE1,2, ØRJAN SAMUELSEN2, and LARS HARRY VORLAND1,2 1Department of Medical Microbiology, University Hospital of North Norway, and 2Department of Microbiology and Virology, University of Tromsø, Tromsø, Norway

SUMMARY Antimicrobial peptides are important in the innate immunity and de­ fence mechanisms of all organisms. Several models have been pro­ posed in order to explain their antibacterial mode of action. Most an­ timicrobial peptides are amphipatic and cationic, and thus an effect on the cytoplasmic membrane of susceptible bacteria has been postulated as the main mode of action. The peptides may either form a channel, hereby inducing leakage of cytoplasmic content, or the peptide may in­ duce permeability changes in a detergent-like manner. Both modes of action may lead to the death of the bacterial cell. Intracellular targets have also been identified for some antimicrobial peptides, and include binding to macromolecules, inhibition of macromolecular biosynthesis, and inhibition of bacterial enzymes. Some peptides have also been shown to have more than one target. This review addresses the models describing the antibacterial mode of action of human defensins present in the gut. In addition, the antibacterial mode of action of related an­ timicrobial peptides is discussed.

INTRODUCTION The microbial load in the intestines rapid, non-oxidative, no-memory first­ of mammals is enormous (Moore and line defence system, the innate immu- Holdeman, 1974). Some of these mi- nity system, involving peptides and crobes are involved in the digestion and proteins with antimicrobial activity pro­ uptake of nutrients, and hence benefits tect the host against possible pathogenic the host. The presence of pathogenic bacteria. Such antimicrobial peptides are bacteria may however not benefit the present in the gastrointestinal tract host, and several mechanisms are in- across phyla. Magainins are found in volved in the protection of the intestine the stomach and intestine of the African from these bacteria. The mechanisms frog Xenopus laevis (Zasloff, 1992; include the presence of a normal bacte- Reilly et al., 1994), and the midgut of rial flora, volatile fatty acids, peristaltic some insects contains cells that produce movements, mucus, shedding of intes- antimicrobial peptides (Nicolas et al., tinal cells, and the presence of secretory 1996). Cecropin P1, an antimicrobial IgA antibodies (Mahida et al., 1997; peptide related to the cecropins found in Israel and Walker, 1988). In addition, a insects, has been isolated from the por­

17 cine proximal small intestine (Lee et al., distinct (Tables 2, 3 and 4). Most pep­ 1989). tides are active against both Gram- In the human digestive tract, many negative and Gram-positive bacteria, proteins and peptides with antimicrobial some also against fungi and protozoa, activity are present (summarised by Le­ while others are also active against vi­ hrer, 2001). Some peptides are confined ruses and mycobacterium. The minimal to the epithelial cells and protect them inhibitory concentrations of the peptides from invasion by microbes (e.g. β­ are in the range of 0.1-100 Hg/ml. They defensins, hCAP18/LL-37), thus cre­ show synergistic activity between them­ ating a barrier against microbes. Others selves and with other host defence enter the digestive system through sali­ molecules such as lactoferrin and vary glands (e.g. histatins), from the lysozyme (Bals et al., 1998a, 1998b; Paneth cells in the small intestine (e.g. Nagaoka et al, 2000; Singh et al., 2000; α-defensins), and from pancreas (e.g. Garcia et al., 2001a). β-defensins). Further, some antimicro­ Several inhibitors of antimicrobial bial proteins and peptides, i.e. lactofer­ activity have been described and inhibi­ rin and lactoferricin, can enter the GI­ tion of activity by NaCl has been impli­ tract either through food (Kuwata et al., cated in cystic fibrosis (Smith et al., 1998, 2001) or from endogenous 1996; Goldman et al., 1997). HBD-3 is sources (Kayazawa et al., 2002). An­ the only β-defensin that is salt-insensi­ timicrobial peptides are also found in the tive (Harder et al., 2001). The inhibition gut due to their presence in migrating by NaCl is also dependent on the polymorphonuclear cells (Handy et al., microbe, as high NaCl concentrations 1995). inhibits the activity of HNP-1 against Gram-positive and Gram-negative bac­ The defensins teria, but have no effect on the activity The defensins comprise the largest of HNP-1 against mycobacterium or group of mammalian peptides (Risso, Herpes simplex virus-1 (Daher et al., 2000), and are present throughout the 1986; Miyasaki et al., 1990; Ogata et digestive tract in all mammals, including al., 1992; Miyakawa et al., 1996). For humans (Table 1). There are two sub­ LL-37, NaCl inhibits the activity against families of human defensins: (i) α-de­ methicillin resistant S. aureus, but does fensins and (ii) β-defensins, differing not influence the activity against van­ from each other in the position of the comycin resistant Enterococcus faecium cysteine residues and in the bridge for­ (Turner et al., 1998). Divalent cations mation. The mature α-defensins com­ such as Ca2+ and Mg2+, serum and al­ prises 29-35 amino acids (Lehrer et al., bumin have also been reported to inhibit 1993), and the β-defensins 34-42 resi­ the activity of some peptides (see Tables dues (Selsted et al., 1993). In their 2, 3 and 4). mature form, all defensins share a The other GI-tract peptides covered similar structural conformation; they are in this review, lactoferricin B, magainin all β-sheets, cycled and stabilised by 2, cecropin P1 and LL-37, also show a three disulphide-bridges (Risso, 2000). broad spectrum of activity covering Gram-positive and Gram-negative bac­ Antimicrobial spectrum and ac­ teria, fungi, viruses, and protozoa (Ta­ tivity ble 2). Among these peptides, lactofer­ Although the antimicrobial peptides ricin B is the only peptide with antiviral in the GI-tract possess several similari­ activity (Andersen et al., 2001). ties, their antimicrobial properties are

18 19 20 GI-tract peptides are thus active fungi, protozoa, viruses, and myco­ against a wide range of Gram-positive bacterium. and Gram-negative bacteria, as well as

DISCUSSION Mode of action of antimicrobial forming peptides, the models include peptides the barrel-stave model (Shai, 1999; Due to the amphipatic, cationic Bechinger, 1999), the wormhole model structure of most antimicrobial peptides, (Matsuzaki et al., 1996; Ludtke et al., an effect on the cytoplasmic membrane 1996), and the two-state model (Huang, of susceptible bacteria has been postu­ 2000). Dependent upon the character of lated as the main mode of action (Ganz the pore, the formation of pores may and Lehrer, 1998). After an initial inter­ lead to leakage of ions and cytoplasmic action between peptide and bacterial cell content, influx of water, or both. surface, the peptide will traverse to the Despite the focus on bacterial mem­ outer leaflet of the cytoplasmic mem­ branes as targets for antimicrobial pep­ brane and cause an increased perme­ tides, several antimicrobial peptides ability, which eventually leads to cell have been shown to have intracellular death. targets. These include binding to DNA, In Gram-negative bacteria, the anti­ RNA and/or proteins (Park et al., 1998; bacterial peptides are thought to cross Otvos et al., 2000; Kragol et al., 2001), the outer membrane through a mecha­ inhibition of macromolecular biosynthe­ nism called “self-promoted-uptake” sis (Boman et al., 1993; Subbalakshmi (Hancock and Bell, 1988). Divalent and Sitaram, 1998; Castle et al., 1999; cations in the LPS are replaced by the Patrzykat et al., 2002) and inhibition of peptide, causing an increased perme­ bacterial enzymes (Nishikata et al., ability of the outer membrane, which 1991; Couto et al., 1993; Andreu and allows more peptide molecules to cross Rivas, 1998). the outer barrier. For Gram-positive bacteria, the initial interaction is shown Mode of action of defensins to be with the (L)TA (Vorland et al., Several lines of evidence argue for a 1999), yet there are no good explanation hypothesis involving the cytoplasmic for the subsequent crossing of the thick membrane as the bactericidal target for peptidoglycan layer present in Gram- defensins: positive organisms. (i) Defensins (HNP-1) sequentially Several models have been proposed permeabilise the outer and inner in order to explain the effect antimicro­ membrane of E. coli (Lehrer et al., bial peptides have on the cytoplasmic 1989), membrane (Table 5). In general, the (ii) Defensins (HNP-1) form voltage­ peptides may act by destabilising and dependent channels in artificial hereby permeabilising the membrane, or membranes (Kagan et al., 1990), by forming distinct pores/channels in (iii) Defensins induce leakage of cyto­ the membrane. For the former effect, plasmic content (Lehrer et al., the most known models include the 1989; Cociancich et al., 1993), formation of a peptide carpet (Gazit et (iv) Defensins induce leakage of vesicle al., 1995) and thinning of the membrane content from negatively charged (Ludtke et al., 1995; Berneche et al., liposomes (Wimley et al., 1994), 1998; Heller et al., 2000). For pore­

21 22 (v) Defensins (HNP 1-3) are active of a basket (Hill et al., 1991). This bas­ against enveloped viruses, but not ket has a hydrophobic bottom and a against non-enveloped viruses (Da­ polar top. The model pore is composed her et al., 1986), of 6 defensin dimers arranged with the (vi) The effect of defensins is abolished polar basket tops lining a ~20 Å pore. by membrane-depolarising agents The hydrophobic basket bottom face (Lehrer et al., 1988), and outwards towards the bilayer of the (vii) Metabolically active microbes are membrane. This channel allows the more susceptible to human α-de­ leakage of rather large molecules (up to fensins than resting microbes (Le­ ~4,400 Da). hrer et al., 1989). However, NMR studies have shown Non-specific electrostatic interactions that it is not possible to use the same Aley et al. (1994) reports of cell ag­ model to describe the mode of action for gregation and dramatic changes in mor­ all defensins (Hoover et al., 2001). phology of Giardia lamblia trophozoites HNP-3 dimers cannot be modelled us­ after exposure to HNP-1. The mode of ing HBD-2 monomers, and HBD-1 action was interpreted to involve bind­ monomers cannot be arranged into ing and lysis, an event that appeared to HBD-2 or HNP-3 type dimers. Hence, involve charge interactions. Further, the the exact mechanism of all defensins is high-resolution crystal structure of not known, and there are currently two HBD-2 show that peptide monomers are models describing the mode of action of capable of forming an octameric defensins. One model describes the structure with a uniform positively formation of multimeric pores in the charged outer surface (Hoover et al., cytoplasmic membrane (Wimley et al., 2000). However, the structural and 1994), and the other involves non-spe­ electrostatic properties of the HBD-2 cific electrostatic interactions between octamer support an electrostatic charge­ negatively charged moieties in the based mechanism of membrane perme­ membranes and the positive charges of abilisation by beta-defensins, rather than the side chains of defensin molecules a mechanism based on formation of (Hill et al., 1991). Both mechanisms bilayer-spanning pores. may lead to permeability changes, cell Electrostatic interactions may lead to rupture/lysis and death. cell death through a detergent like effect, where the formation of a carpet of The multimeric pore peptide molecules in the membrane re­ Wimley et al. (1994) have published sults in membrane disruption at a critical the results of an extensive study per­ ratio of lipid:peptide (Shai, 1999). The formed on HNP-2. HNP-2 binds to interactions may also cause separation negatively charged vesicles through of the polar lipid head groups of the electrostatic interactions, induce fusion phospholipids in the cytoplasmic mem­ of the outer monolayer of vesicles, and brane, as they are pushed aside by the cause leakage of vesicle content through hydrophobic residues of the membrane pores with a maximum diameter of ap­ associated peptide molecules (Ludtke et proximately 25 Å. The authors further al., 1995). As a result, gaps will be present a multimeric model of such a formed between the head groups, in­ pore made from HNP-2 molecules, ducing physical stress on the bacterial based upon the crystal structure of de­ cytoplasmic membrane, and result in the fensins showing dimers with the form collapse of the membrane.

23

Figure 1: Electron micrographs of bacteria exposed to antimicrobial peptides, immunolabelled with polyclonal antibodies towards the respective peptide, and further visualised with gold-marked protein A. Panel A; Negative control (E. coli not exposed to any peptide). Panel B; E. coli ex­ posed to magainin 2 for 30 minutes. Panel C; S. aureus exposed to lactoferricin B for one hour. Panel D; E. coli exposed to cecropin P1 for 30 minutes. The micrographs have previously been published by Haukland et al. (2001).

Targets other than the membrane membrane (see Table 5). Despite these The idea of antimicrobial peptides as effects, Haukland et al. (2001) have multi-target substances is growing. Re­ shown that magainin 2 are capable of sults involving other effects than those residing in the bacterial cytoplasm (Fig­ of the cytoplasmic membrane have been ure 1b). Cecropin P1 does not exhibit published. For example, in addition to this feature, and are confined to the its permeabilising effects, HNP-1 also bacterial cell wall (Figure 1d), consis­ causes a reduction in bacterial macro­ tent with the carpet model proposed for molecular biosynthesis and a drop in the the mode of action of this peptide (Gazit colony count (Lehrer et al., 1989). et al., 1995). Further, the magainins have been exten­ sively studied as pore-forming peptides While proposing the model of the (Matsuzaki, 1998), and several models multimeric pore, Wimley et al. (1994) have been used to explain the interaction also points out that the actual in vivo between magainin and the cytoplasmic mechanism for cell leakage may involve

24 at least three steps. These included the the proposal of two targets for defens­ initial interaction of monomeric defens­ ins. Working on tumour cells, they re­ ins and cell surface through electrostatic port that initial effects on the plasma interactions, the oligomerisation of de­ membrane were not sufficient for sub­ fensins, and at last pore formation. sequent lysis. A second phase was re­ Translocation of the peptide via pore quired which involved the continued formation is possible as a fourth step, presence of defensin. They conclude allowing the peptide to interfere with that there is two phases of interaction any intracellular process. between defensins and tumour cells, Evidence for the translocation of where the initial effect is on the cell several antimicrobial peptides are accu­ membrane, and the second phase is me­ mulating, and involve magainin 2 diated intracellularly by defensin inter­ (Haukland et al., 2001), lactoferricin B nalised through a permeabilised mem­ (Haukland et al., 2001), and buforin brane. A two-phased bactericidal activ­ (Park et al., 1998). For defensins, ity is also proposed for HBD-2 and E. Sharma and Khuller (2001) showed that coli (Tomita et al., 2000). Lactoferricin HNP-1 is an efficient inhibitor of DNA­ B also interacts with membranes (Ul­ synthesis in Mycobacterium tuberculo­ vatne et al. 2001), and can be traced into sis. They suggest that the cytoplasmic the cytoplasm at sub-inhibitory membrane is the primary target for concentrations (Figure 1c) (Haukland et HNP-1. Binding to this target causes al., 2001). Unpublished results show permeabilising of the membrane, and that lactoferricin B have an effect of thus enhanced access to the secondary, macromolecular biosynthesis (Ulvatne intracellular target. et al., in prep.). Lichtenstein (1991) has also made

CONCLUSION At this point, there is no doubt that efficient killing by pores must therefore most antimicrobial peptides, including be swift, rapid and sudden, to ensure the defensins, are membrane active that the bacteria do not initiate a defence molecules. Through their interaction response. Antimicrobial peptides may, with the cytoplasmic membrane, they by utilising another secondary target, be may cause severe damage to the bacte­ even more efficient in the battle against rial cell and cell death. It is likely that pathogenic bacteria. bacteria may compensate for the forma­ Due to the fact that most antimicro­ tion of pores/channels in the mem­ bial peptides also exhibit other effects, branes, while a detergent like effect is the in vivo effect and exact mode of ac­ irreversible since it involves a complete tion of antimicrobial peptides are hard to rupture of the bacterial integrity. The elucidate (Scott and Hancock, 2000). bacterial cell is not a closed system, and The other effects involve interactions the cell is in some kind of equilibrium with host cells to stimulate gene-expres­ with its surroundings through sensing sion from genes encoding transcription systems. Transport of nutrients, waste factors, chemokines, chemokine recep­ products, and other extracellular prod­ tors, integrins etc, products that also are ucts are constantly crossing the cyto­ part of the innate immunity (Hancock plasmic membrane, and some of this and Rozek, 2002). Antimicrobial pep­ transport happens through pores. An tides are therefore multi-functioning ef­

25 fector molecules involved in the delicate Therefore, further studies on defensins balance between microbes and host, and and other antimicrobial peptides must be their in vivo role must be regarded as performed in order to understand the in the whole interplay between the differ- vivo antibacterial mode of action. ent functions the peptides may have.

LITERATURE

Agerberth, B., Gunne, H., Odeberg, J., Kogner, fensin 1 (HNP-1) inhibits adenoviral infec­ P., Boman, H.G., and Gudmundsson, G.H.: tion in vitro. Regul. Pept. 101, 157-161 FALL-39, a putative human peptide (2001). antibiotic, is cysteine-free and expressed in Bechinger, B.: The structure, dynamics and bone marrow and testis. Proc. Natl. Acad. orientation of antimicrobial peptides in Sci. USA 92, 195-199 (1995). membranes by multidimensional solid-state Agerberth, B., Lee, J.Y., Bergman, T., Car­ NMR spectroscopy. Biochim. Biophys. lquist, M., Boman, H.G., Mutt, V., and Acta 1462, 157-183 (1999). Jornvall, H.: Amino acid sequence of PR­ Bellamy, W., Takase, M., Wakabayashi, H., 39. Isolation from pig intestine of a new Kawase, K., and Tomita, M.: Antibacterial member of the family of proline-arginine­ spectrum of lactoferricin B, a potent bacte­ rich antibacterial peptides. Eur. J. Biochem. ricidal peptide derived from the N-terminal 202, 849-854 (1991). region of bovine lactoferrin. J. Appl. Bacte­ Aley, S.B., Zimmerman, M., Hetsko, M., riol. 73, 472-479 (1992). Selsted, M.E., and Gillin, F.D.: Killing of Berneche, S., Nina, M., and Roux, B.: Mo­ Giardia lamblia by cryptdins and cationic lecular dynamics simulation of melittin in a neutrophil peptides. Infect. Immun. 62, dimyristoylphosphatidylcholine bilayer 5397-5403 (1994). membrane. Biophys. J. 75, 1603-1618 Andersen, J.H., Osbakk, S.A., Vorland, L.H., (1998). Traavik, T., and Gutteberg, T.J.: Lactofer­ Bessalle, R., Kapitkovsky, A., Gorea, A., rin and cyclic lactoferricin inhibit the entry Shalit, I., and Fridkin, M.: All-D-ma­ of human cytomegalovirus into human fi­ gainin: chirality, antimicrobial activity and broblasts. Antiviral Res. 51, 141-149 proteolytic resistance. FEBS Lett. 274, (2001). 151-155 (1990). Andra, J., Berninghausen, O., and Leippe, M.: Boman, H.G., Agerberth, B., and Boman, A.: Cecropins, antibacterial peptides from in­ Mechanisms of action on Escherichia coli sects and mammals, are potently fungicidal of cecropin P1 and PR-39, two antibacterial against Candida albicans. Med. Microbiol. peptides from pig intestine. Infect. Immun. Immunol. (Berl.). 189, 169-173 (2001). 61, 2978-2984 (1993). Andreu, D. and Rivas, L.: Animal antimicro­ Castle, M., Nazarian, A., Yi, S.S., and bial peptides: An overview. Biopolymers Tempst, P.: Lethal effects of apidaecin on 47, 415-433 (1998). Escherichia coli involve sequential molecu­ Bals, R., Wang, X., Wu, Z., Freeman, T., lar interactions with diverse targets. J. Biol. Bafna, V., Zasloff, M., and Wilson, J.M.: Chem. 274, 32555-32564 (1999). Human β-defensin 2 is a salt-sensitive pep­ Cirioni, O., Giacometti, A., Barchiesi, F., and tide antibiotic expressed in human lung. J. Scalise, G.: In-vitro activity of lytic pep­ Clin. Invest. 102, 874-880 (1998b). tides alone and in combination with macrol­ Bals, R., Wang, X., Zasloff, M., and Wilson, ides and inhibitors of dihydrofolate reductase J.M.: The peptide antibiotic LL-37/hCAP­ against Pneumocystis carinii. J. Antimi­ 18 is expressed in epithelia of the human crob. Chemother. 42, 445-451 (1998). lung where it has broad antimicrobial activ­ Cociancich, S., Ghazi, A., Hetru, C., ity at the airway surface. Proc. Natl. Acad. Hoffmann, J.A., and Letellier, L.: Insect de­ Sci. USA 95, 9541-9546 (1998a). fensin, an inducible antibacterial peptide, Bastian, A. and Schafer, H.: Human alpha-de­ forms voltage-dependent channels in Micro­

26 coccus luteus. J. Biol. Chem. 268, 19239­ Barchiesi, F., Paggi, A.M., Petrelli, E., 19245 (1993). and Scalise, G.: Comparative activities of Couto, M.A., Harwig, S.S., and Lehrer, R.I.: polycationic peptides and clinically used an­ Selective inhibition of microbial serine pro­ timicrobial agents against multidrug-resis­ teases by eNAP-2, an antimicrobial peptide tant nosocomial isolates of Acinetobacter from equine neutrophils. Infect. Immun. baumannii. J. Antimicrob. Chemother. 46, 61, 2991-2994 (1993). 807-810 (2000). Daher, K.A., Selsted, M.E., and Lehrer, R.I.: Giacometti, A., Cirioni O., Greganti G., Direct inactivation of viruses by human Quarta M., and Scalise G.: In vitro activi­ granulocyte defensins. J. Virol. 60, 1068­ ties of membrane-active peptides against 1074 (1986). Gram-positive and Gram-negative aerobic Gabay, J.E., Scott, R.W., Campanelli, D., bacteria. Antimicr. Agents Chemother. 42, Griffith, J., Wilde, C., Marra, M.N., See­ 3320-3324 (1998). ger, M., and Nathan, C.F.: Antibiotic pro­ Goldman, M.J., Anderson, G.M., Stolzenberg, teins of human polymorphonuclear leuko­ E.D., Kari, U.P., Zasloff, M., and Wilson, cytes. Proc. Natl. Acad. Sci. USA 86, J.M.: Human beta-defensin-1 is a salt-sensi­ 5610-5614 (1989). tive antibiotic in lung that is inactivated in Ganz, T. and Lehrer, R.I.: Antimicrobial pep­ cystic fibrosis. Cell 88, 553-560 (1997). tides of vertebrates. Curr. Opin. Immunol. Gropp, R., Frye, M., Wagner, T.O., and Bar­ 10, 41-44 (1998). gon, J.: Epithelial defensins impair adeno­ Ganz, T., Selsted, M.E., Szklarek, D., Harwig, viral infection: Implication for adenovirus­ S.S., Daher, K., Bainton, D.F., and Lehrer, mediated gene therapy. Hum. Gene Ther. R.I.: Defensins. Natural peptide antibiotics 10, 957-964 (1999). of human neutrophils. J. Clin. Invest. 76, Hamanaka, Y., Nakashima, M., Wada, A., Ito, 1427-1435 (1985). M., Kurazono, H., Hojo, H., Nakahara, Y., Garcia, J.R., Krause, A., Schulz, S., Rodri­ Kohno, S., Hirayama, T., and Sekine, I.: guez-Jimenez, F.J., Kluver, E., Adermann, Expression of human beta-defensin 2 (hBD­ K., Forssmann, U., Frimpong-Boateng, A., 2). in Helicobacter pylori induced gastritis: Bals, R., and Forssmann, W.G.: Human Antibacterial effect of hBD-2 against Heli­ beta-defensin 4: a novel inducible peptide cobacter pylori. Gut 49, 481-487 (2001). with a specific salt-sensitive spectrum of Hancock, R.E. and Bell, A.: Antibiotic uptake antimicrobial activity. FASEB J. 15, 1819­ into gram-negative bacteria. Eur. J. Clin. 1821 (2001a). Microbiol. Infect. Dis. 7, 713-720 (1988). Garcia, J.R., Jaumann, F., Schulz, S., Krause, Hancock, R.E. and Rozek, A.: Role of mem­ A., Rodriguez-Jimenez, J., Forssmann, U., branes in the activities of antimicrobial Adermann, K., Kluver, E., Vogelmeier, C., cationic peptides. FEMS Microbiol. Lett. Becker, D., Hedrich, R., Forssmann, W.G., 206, 143-149 (2002). and Bals, R.: Identification of a novel, mul­ Handy, L.M., Ghosh, S., and Ferguson, A.: tifunctional beta-defensin (human beta-de­ Investigation of neutrophil migration into fensin 3) with specific antimicrobial activ­ the gut by cytology of whole gut lavage ity. Its interaction with plasma membranes fluid. Eur. J. Gastroenterol. Hepatol. 7, 53­ of Xenopus oocytes and the induction of 58 (1995). macrophage chemoattraction. Cell Tissue Harder, J., Bartels, J., Christophers, E., and Res. 306, 257-264 (2001b). Schroder, J.M.: A peptide antibiotic from Gazit, E., Boman, A., Boman, H.G., and Shai, human skin. Nature 387, 861 (1997). Y.: Interaction of the mammalian antibacte­ Harder, J., Bartels, J., Christophers, E., and rial peptide cecropin P1 with phospholipid Schroder, J.M.: Isolation and characteriza­ vesicles. Biochemistry 34, 11479-11488 tion of human beta-defensin-3, a novel hu­ (1995). man inducible peptide antibiotic. J. Biol. Giacometti, A., Cirioni, O., Barchiesi, F., Del Chem. 276, 5707-5713 (2001). Prete, M.S., and Scalise, G.: Antimicrobial Hase, K., Eckmann, L., Leopard, J.D., Varki, activity of polycationic peptides. Peptides N., and Kagnoff, M.F.: Cell differentiation 20, 1265-1273 (1999). is a key determinant of cathelicidin LL- Giacometti, A., Cirioni, O., Del Prete, M.S., 37/human cationic antimicrobial protein 18

27 expression by human colon epithelium. In­ Proc. Natl. Acad. Sci. USA 87, 210-214 fect.Immun. 70, 953-963 (2002). (1990). Haukland, H.H., Ulvatne, H., Sandvik, K., and Kayazawa, M., Saitoh, O., Kojima, K., Naka­ Vorland, L.H.: The antimicrobial peptides gawa, K., Tanaka, S., Tabata, K., Matsuse, lactoferricin B and magainin 2 cross over R., Uchida, K., Hoshimoto, M., Hirata, I., the bacterial cytoplasmic membrane and re­ and Katsu, K.: Lactoferrin in whole gut side in the cytoplasm. FEBS Lett. 508, lavage fluid as a marker for disease activity 389-393 (2001). in inflammatory bowel disease: Compari­ Heller, W.T., Waring, A.J., Lehrer, R.I., Har­ son with other neutrophil-derived proteins. roun, T.A., Weiss, T.M., Yang, L., and Am. J. Gastroenterol. 97, 360-369 (2002). Huang, H.W.: Membrane thinning effect of Kragol, G., Lovas, S., Varadi, G., Condie, the beta-sheet antimicrobial protegrin. Bio­ B.A., Hoffmann, R., and Otvos, L. Jr.: The chemistry 39, 139-145 (2000). antibacterial peptide pyrrhocoricin inhibits Hill, C.P., Yee, J., Selsted, M.E., and Eisen­ the ATPase actions of DnaK and prevents berg, D.: Crystal structure of defensin chaperone-assisted protein folding. Bio­ HNP-3, an amphiphilic dimer: Mechanisms chemistry 40, 3016-3126 (2001). of membrane permeabilization. Science Kuwata, H., Yamauchi, K., Teraguchi, S., 251, 1481-1485 (1991). Ushida, Y., Shimokawa, Y., Toida, T., and Hoover, D.M., Chertov, O., and Lubkowski, Hayasawa, H.: Functional fragments of in­ J.: The structure of human beta-defensin-1: gested lactoferrin are resistant to proteolytic New insights into structural properties of degradation in the gastrointestinal tract of beta-defensins. J. Biol. Chem. 276, 39021­ adult rats. J. Nutr. 131, 2121-2127 (2001). 39026 (2001). Kuwata, H., Yip, T.T., Tomita, M., and Hoover, D.M., Rajashankar, K.R., Blumen­ Hutchens, T.W.: Direct evidence of the thal, R., Puri, A., Oppenheim, J.J., Cher­ generation in human stomach of an antimi­ tov, O., and Lubkowski, J.: The structure crobial peptide domain (lactoferricin) from of human beta-defensin-2 shows evidence of ingested lactoferrin. Biochim. Biophys. higher order oligomerization. J. Biol. Acta 1429, 129-141 (1998). Chem. 275, 32911-32918 (2000). Larrick, J.W., Hirata, M., Zhong, J., and Huang, H.W.: Action of antimicrobial pep­ Wright, S.C.: Anti-microbial activity of tides: two-state model. Biochemistry 39, human CAP18 peptides. Immunotechnol­ 8347-8352 (2000). ogy 1, 65-72 (1995). Israel, E.J. and Walker, W.A.: Host defense Lee, J.Y., Boman, A., Sun, C.X., Andersson, development in gut and related disorders. M., Jornvall, H., Mutt, V., and Boman, Pediatr. Clin. North Am. 35, 1-15 (1988). H.G.: Antibacterial peptides from pig intes­ Johansson, J., Gudmundsson, G.H., Rotten­ tine: Isolation of a mammalian cecropin. berg, M.E., Berndt, K.D., and Agerberth, Proc. Natl. Acad. Sci. USA 86, 9159-9162 B.: Conformation-dependent antibacterial (1989). activity of the naturally occurring human Lehrer, R.: Shigellae control the Gates of LL. peptide LL-37. J. Biol. Chem. 273, 3718­ Nat. Med. 7, 158-159 (2001). 3724 (1998). Lehrer, R.I., Barton, A., Daher, K.A., Harwig, Jones, D.E. and Bevins, C.L.: Paneth cells of S.S., Ganz, T., and Selsted, M.E.: Interac­ the human small intestine express an an­ tion of human defensins with Escherichia timicrobial peptide gene. J. Biol. Chem. coli. Mechanism of bactericidal activity. J. 267, 23216-23225 (1992). Clin. Invest. 84, 553-561 (1989). Jones, D.E., and Bevins, C.L.: Defensin-6 Lehrer, R.I., Ganz, T., Szklarek, D., and Sel­ mRNA in human Paneth cells: implica­ sted, M.E.: Modulation of the in vitro can­ tions for antimicrobial peptides in host de­ didacidal activity of human neutrophil de­ fense of the human bowel. FEBS Lett. 315, fensins by target cell metabolism and diva­ 187-192 (1993). lent cations. J. Clin. Invest. 81, 1829-1835 Kagan, B.L., Selsted, M.E., Ganz, T., and Le­ (1988). hrer, R.I.: Antimicrobial defensin peptides Lehrer, R.I., Lichtenstein, A.K., and Ganz, T.: form voltage-dependent ion-permeable Defensins: antimicrobial and cytotoxic pep­ channels in planar lipid bilayer membranes. tides of mammalian cells. Annu. Rev. Im­

28 munol. 11, 105-128 (1993). of human neutrophil azurophil granules that Lichtenstein, A.: Mechanism of mammalian mediate fungistasis against Histoplasma cell lysis mediated by peptide defensins. capsulatum. Infect. Immun. 68, 5668-5672 Evidence for an initial alteration of the (2000). plasma membrane. J. Clin. Invest. 88, 93­ Nicolas, E., Nappi, A.J., and Lemaitre, B.: 100 (1991). Expression of antimicrobial peptide genes Ludtke, S., He, K., and Huang, H.: Membrane after infection by parasitoid wasps in Dro­ thinning caused by magainin 2. Biochemis­ sophila. Dev. Comp. Immunol. 20, 175­ try 34, 16764-16769 (1995). 181 (1996). Ludtke, S.J., He, K., Heller, W.T., Harroun, Nishikata, M., Kanehira, T., Oh, H., Tani, H., T.A., Yang, L., and Huang, H.W.: Mem­ Tazaki, M., and Kuboki, Y.: Salivary brane pores induced by magainin. Biochem­ histatin as an inhibitor of a protease pro­ istry 35, 13723-13728 (1996). duced by the oral bacterium Bacteroides Mahida, Y.R., Rose, F., and Chan, W.C.: An­ gingivalis. Biochem. Biophys. Res. Com­ timicrobial peptides in the gastrointestinal mun. 174, 625-630 (1991). tract. Gut 40, 161-163 (1997). Ogata, K., Linzer, B.A., Zuberi, R.I., Ganz, Matsuzaki, K.: Magainins as paradigm for the T., Lehrer, R.I., and Catanzaro, A.: Activ­ mode of action of pore forming polypep­ ity of defensins from human neutrophilic tides Biochim. Biophys. Acta 1376, 391­ granulocytes against Mycobacterium 400 (1998). avium-Mycobacterium intracellulare. Infect. Matsuzaki, K., Murase, O., Fujii, N., and Mi­ Immun. 60, 4720-4725 (1992). yajima, K.: An antimicrobial peptide, ma­ O'Neil, D.A., Porter, E.M., Elewaut, D., An­ gainin 2, induced rapid flip-flop of phos­ derson, G.M., Eckmann, L., Ganz, T., and pholipids coupled with pore formation and Kagnoff, M.F.: Expression and regulation peptide translocation. Biochemistry 35, of the human beta-defensins hBD-1 and 11361-11368 (1996). hBD-2 in intestinal epithelium. J. Immu­ Miyakawa, Y., Ratnakar, P., Rao, A.G., Cos­ nol. 163, 6718-6724 (1999). tello, M.L., Mathieu-Costello, O., Lehrer, Oren, Z., Lerman, J.C., Gudmundsson, G.H., R.I., and Catanzaro, A.: In vitro activity of Agerberth, B., and Shai, Y.: Structure and the antimicrobial peptides human and rabbit organization of the human antimicrobial defensins and porcine leukocyte protegrin peptide LL-37 in phospholipid membranes: against Mycobacterium tuberculosis. Infect. Relevance to the molecular basis for its Immun. 64, 926-932 (1996). non-cell-selective activity. Biochem. J. Miyasaki, K.T., Bodeau, A.L., Ganz, T., Sel­ 341, 501-513 (1999). sted, M.E., and Lehrer, R.I.: In vitro sensi­ Otvos, L., Jr, Rogers, M.E., Consolvo, P.J., tivity of oral, gram-negative, facultative Condie, B.A., Lovas, S., Bulet, P., and bacteria to the bactericidal activity of hu­ Blaszczyk-Thurin, M.: Interaction between man neutrophil defensins. Infect. Immun. heat shock proteins and antimicrobial pep­ 58, 3934-3940 (1990). tides Biochemistry 39, 14150-14159 Moore, A.J., Beazley, W.D., Bibby, M.C., and (2000). Devine, D.A.: Antimicrobial activity of ce­ Park, C.B., Kim, M.S., and Kim, S.C.: A cropins. J. Antimicrob. Chemother. 37, novel antimicrobial peptide from Bufo bufo 1077-1089 (1996). gargarizans. Biochem. Biophys. Res. Moore, W.E. and Holdeman, L.V.: Human Commun. 218, 408-413 (1996). fecal flora: The normal flora of 20 Japanese- Park, C.B., Kim, H.S., and Kim, S.C.: Hawaiians. Appl. Microbiol. 27, 961-979 Mechanism of action of the antimicrobial (1974). peptide buforin II: Buforin II kills microor­ Nagaoka, I., Hirota, S., Yomogida, S., Oh­ ganisms by penetrating the cell membrane wada, A., and Hirata, M.: Synergistic ac­ and inhibiting cellular functions. Biochem. tions of antibacterial neutrophil defensins Biophys. Re.s Commun. 244, 253-257 and cathelicidins. Inflamm. Res. 49, 73-79 (1998). (2000). Patrzykat, A., Friedrich, C.L., Zhang, L., Newman, S.L., Gootee, L., Gabay, J.E., and Mendoza, V., and Hancock, R.E.: Sublethal Selsted, M.E.: Identification of constituents concentrations of pleurocidin-derived antim­

29 icrobial peptides inhibit macromolecular and Welsh, M.J.: Synergistic and additive synthesis in Escherichia coli. Antimicrob. killing by antimicrobial factors found in Agents Chemother. 46, 605-614 (2002). human airway surface liquid. Am. J. Porter, E.M., van Dam, E., Valore, E.V., and Physiol. Lung Cell. Mol. Physiol. 279, Ganz, T.: Broad-spectrum antimicrobial ac­ L799-L805 (2000). tivity of human intestinal defensin 5. In­ Smeianov, V., Scott, K., and Reid, G.: Activ­ fect. Immun. 65, 2396-2401 (1997). ity of cecropin P1 and FA-LL-37 against Reilly, D.S., Tomassini, N., Bevins, C.L., and urogenital microflora. Microbes Infect. 2, Zasloff, M.: A Paneth cell analogue in 773-777 (2000). Xenopus small intestine expresses antimi­ Smith, J.J., Travis, S.M., Greenberg, E.P., crobial peptide genes: conservation of an in­ and Welsh, M.J.: Cystic fibrosis airway testinal host-defense system. J. Histochem. epithelia fail to kill bacteria because of ab­ Cytochem. 42, 697-704 (1994). normal airway surface fluid. Cell 85, 229­ Risso, A.: Leukocyte antimicrobial peptides: 236 (1996). Multifunctional effector molecules of innate Subbalakshmi and C., Sitaram, N.: Mechanism immunity. J. Leukoc. Biol. 68, 785-792 of antimicrobial action of indolicidin. (2000). FEMS Microbiol. Lett. 160, 91-96 (1998). Scott, M.G. and Hancock, R.E.: Cationic an­ Tomita, T., Hitomi, S., Nagase, T., Matsui, timicrobial peptides and their multifunc­ H., Matsuse, T., Kimura, S., and Ouchi, tional role in the immune system. Crit. Y.: Effect of ions on antibacterial activity Rev. Immunol. 20, 407-431 (2000). of human beta defensin 2. Microbiol. Im­ Selsted, M.E., Tang, Y.Q., Morris, W.L., munol. 44, 749-754 (2000). McGuire, P.A., Novotny, M.J., Smith, Travis, S.M., Anderson, N.N., Forsyth, W.R., W., Henschen, A.H., and Cullor, J.S.: Pu­ Espiritu, C., Conway, B.D., Greenberg, rification, primary structures, and antibacte­ E.P., McCray, P.B., Jr, Lehrer, R.I., rial activities of beta-defensins, a new fam­ Welsh, M.J., and Tack, B.F.: Bactericidal ily of antimicrobial peptides from bovine activity of mammalian cathelicidin-derived neutrophils. J. Biol. Chem. 268, 6641­ peptides. Infect. Immun. 68, 2748-2755 6648 (1993). (2000). Shafer, W.M., Engle, S.A., Martin, L.E., and Turchany, J.M., Aley, S.B., and Gillin, F.D.: Spitznagel, J.K.: Killing of Proteus mirabi­ Giardicidal activity of lactoferrin and N-ter­ lis by polymorphonuclear leukocyte granule minal peptides. Infect. Immun. 63, 4550­ proteins: Evidence for species specificity by 4552 (1995). antimicrobial proteins. Infect. Immun. 56, Turner, J., Cho, Y., Dinh, N.N., Waring, A.J., 51-53 (1988). and Lehrer, R.I.: Activities of LL-37, a Shai, Y.: Mechanism of the binding, insertion cathelin-associated antimicrobial peptide of and destabilization of phospholipid bilayer human neutrophils. Antimicrob. Agents. membranes by alpha-helical antimicrobial Chemother. 42, 2206-2214 (1998). and cell non-selective membrane-lytic pep­ Ulvatne, H. and Vorland, L.H.: Bactericidal tides. Biochim. Biophys. Acta 1462, 55-70 kinetics of 3 lactoferricins against Staphy­ (1999). lococcus aureus and Escherichia coli. Scand. Sharma, S. and Khuller, G.: DNA as the intra­ J. Infect. Dis. 33, 507-511 (2001). cellular secondary target for antibacterial ac­ Ulvatne, H., Haukland, H.H., Samuelsen, Ø., tion of human neutrophil peptide-I against Kraemer, M., and Vorland, L.H.: Proteases Mycobacterium tuberculosis H37Ra. Curr. in Escherichia coli and Staphylococcus Microbiol. 43, 74-76 (2001). aureus confer reduced susceptibility to lac­ Singh, P.K., Jia, H.P., Wiles, K., Hesselberth, toferricin B. J. Antimicrob. Chemother. 50, J., Liu, L., Conway, B.A., Greenberg, 461-467 (2002). E.P., Valore, E.V., Welsh, M.J., Ganz, T., Ulvatne, H., Haukland, H.H., Olsvik, O., and Tack, B.F., and McCray, P.B. Jr.: Vorland, L.H.: Lactoferricin B causes depo­ Production of beta-defensins by human larization of the cytoplasmic membrane of airway epithelia. Proc. Natl. Acad. Sci. Escherichia coli ATCC 25922 and fusion of USA 95, 14961-14966 (1998). negatively charged liposomes. FEBS Lett. Singh, P.K., Tack, B.F., McCray, P.B., Jr, 492, 62-65 (2001).

30 Ulvatne, H., Samuelsen, Ø., Haukland, H., gar, E.A.: Susceptibility of Chlamydia tra­ Kraemer, M., and Vorland, L.H.: Lactofer­ chomatis to protegrins and defensins. Infect. ricin B inhibits macromolecular synthesis Immun. 64, 709-713 (1996). in Escherichia coli and Bacillus subtilis (ti­ Yasin, B., Pang, M., Turner, J.S., Cho, Y., tle pending). Manuscript in preparation. Dinh, N.N., Waring, A.J., Lehrer, R.I., and Valore E.V., Park C.H., Quayle A.J., Wiles Wagar, E.A.: Evaluation of the inactivation K.R., McCray P.B., and Ganz T.: Human of infectious Herpes simplex virus by host­ β-defensin 1: An antimicrobial peptide of defense peptides. Eur. J. Clin. Microbiol. urogenital tissues. J. Clin. Invest. 101, Infect. Dis. 19, 187-194 (2000). 1633-1642 (1998). Zasloff, M.: Magainins, a class of antimicro­ Vorland, L.H., Ulvatne, H., Rekdal, O., and bial peptides from Xenopus skin: Isolation, Svendsen, J.S.: Initial binding sites of an­ characterization of two active forms, and timicrobial peptides in Staphylococcus partial cDNA sequence of a precursor. Proc. aureus and Escherichia coli. Scand. J. In­ Natl. Acad. Sci. USA 84, 5449-5453 fect. Dis. 31, 467-473 (1999). (1987). Wilde, C.G., Griffith, J.E., Marra, M.N., Zasloff, M.: Antibiotic peptides as mediators of Snable, J.L., and Scott, R.W.: Purification innate immunity. Curr. Opin. Immunol. 4, and characterization of human neutrophil 3-7 (1992). peptide 4, a novel member of the defensin Zasloff, M., Martin, B., and Chen, H.: Antim­ family. J. Biol. Chem. 264, 11200-11203 icrobial activity of synthethic magainin (1989). peptides and several analogues. Proc. Natl. Wimley, W.C., Selsted, M.E., and White, Acad. Sci. USA 85, 910-913 (1988). S.H.: Interactions between human defensins Zhao, C., Wang, I., and Lehrer, R.I.: Wide­ and lipid bilayers: evidence for formation of spread expression of beta-defensin hBD-1 in multimeric pores. Protein Sci. 3, 1362­ human secretory glands and epithelial cells. 1373 (1994). FEBS Lett. 396, 319-322 (1996). Yasin, B., Harwig, S.S., Lehrer, R.I., and Wa­

31 32 DEFENSINS AND BACTERIA, A QUESTION OF “LIVE OR LET DIE”? LARS-GÖRAN AXELSSON Microbiology and Tumorbiology Center, Gastrobiology Unit, Division of Molecular Pathology, Karolinska Institute, Stockholm, Sweden

SUMMARY Antibacterial peptides have been found in many organs of the body. Defensins are a family of cationic antimicrobial peptides that are found in mammals, insects and plants. The innate immune system consists of the body’s own defence mechanisms which can be activated upon ex­ posure to foreign microorganisms without earlier exposure or priming. One important part of this system seems to be the production of en­ dogenous antimicrobial peptides. Such peptides, e.g., produced in the gastrointestinal tract of humans and animals have been suggested to modulate the acquirement of a bacterial microflora and play an impor­ tant part in to protect against infections in general. Antimicrobial peptides have broad-spectrum antibiotic activity against e.g. Gram-positive and Gram-negative bacteria, mycobacteria, fungi, parasites and viruses. Antimicrobial peptides seem to be less prone to be, or to induce re­ sistance in bacteria, which make them an interesting alternative in the treatment of infections with multiresistant bacteria. There seem to be hope for the emergence of effective synthetic an­ timicrobial peptides as drugs for human use in the near future.

INTRODUCTION Defensins are a family of cationic or priming (Boman, 1995; Zasloff, antimicrobial peptides that are found in 2002a). mammals, insects and plants (Del Pero Much of the early knowledge of an- et al., 2002; Hancock and Diamond, timicrobial peptides was gained from 2000; Hancock, 2001; Schutte et al., studies of the fly Drosophila that has 2002; Zasloff, 2002a) and as a compo- been utilised as a model for genetic nent in venoms (Corzo et al., 2001). studies. Drosophila lacks adaptive im- The defensins are members of the in- mune system an relies on the inborn, born, innate immune system (Axelsson innate, immune system and have been and Mahida, 2000; Fellerman and used as a model for the elucidation of Stange, 2001; Parkin and Cohen, the different pathways of the innate im­ 2001). The innate immune system con- munity (Hoffman and Reichhart, 2002) sists of the body’s own defence mecha- especially the Toll-pathway (Schwartz, nisms (Medzhitov, 2000) which can be 2002) and the importance of NFκB activated upon exposure to foreign mi- transactivators (Luster, 2002; Mahida croorganisms without earlier exposure and Johal, 2001).

33 One important part of this system teria, mycobacteria, fungi, parasites and seems to be the production of endoge­ viruses (Boman, 1995). nous antimicrobial peptides (Boman, Antimicrobial peptides have been 1995; Ganz, 1994). Such peptides, shown to be produced by many organs e.g., produced in the gastrointestinal and cells through the body (Boman, tract of humans and animals (Schon­ 1995; Ganz, 2000) but this paper will wetter, 1995), have been suggested to concentrate on the innate immunity of modulate the acquirement of a bacterial the gastrointestinal tract since it is ex­ microflora in neonates (Sepp, 1998) and posed to an enormous load of micro­ play an important part in the protection organisms, some causing disease but against infections in general. An­ others are necessary for our well-being timicrobial peptides have broad-spec­ and antimicrobial peptides seem to play trum antibiotic activity against e.g. an important control function on these Gram-positive and Gram-negative bac­ microorganisms.

INNATE IMMUNITY OF THE GASTROINTESTINAL TRACT The innate immunity of the gastroin­ large intestines. testinal system is in place when we are The small intestine harbours a rela­ born (Bry et al., 1994) and will play an tive small number of bacteria compared important role in defending the body to the large intestine (Hooper and against unwanted microbes, but also in Gordon, 2001; Skar et al., 1986, the acquirement and maintenance of a 1989). The large intestine, which in­ normal and healthy microflora, micro­ cludes the caecum and colon, is the part biota (Bevins et al., 1999; Boman, of the gastrointestinal tract where a large 2000; Hooper and Gordon, 2001). number of bacteria, as well as many dif­ The first sentinels at the gate to our ferent species are found (Cunliffe et al., gastrointestinal system are found in the 2001). These bacteria can be commen­ oral cavity. Here we find that mucus is sals, invading pathogenic bacteria or, as produced, layering the physical border recently suggested, opportunistic semi­ consisting of epithelial cells. This mu­ pathogens (Gillespie, 2002; Medzhitov cus layer is found throughout the gas­ and Janeway, 2002). One important cell trointestinal system and has many func­ type producing antimicrobial peptides tions such as being a physical barrier, and other antimicrobial substances is the protecting the epithelial cells, helping in Paneth cell (Porter et al., 2002; Zasloff, eliminating and transporting unwanted 2002b). Apart from antimicrobial pep­ substances out of the body, and also tides there are several other defence harbouring endogenous protecting bio­ mechanisms in the small intestine (Fig­ active molecules (Deplanke and Gas­ ure 1). kins, 2001). There are, for example, antimicrobial The antimicrobial peptides are also at phospholipase A2, lysozyme and tryp­ place already in the oral cavity (Bevins, sin. The trypsin seems to have multiple et al. 1999) and also in the airways functions since recently it was shown (Moser, et al. 2002). that trypsin is the enzyme that activate The stomach functions as a physical HD-5 (Gosh et al., 2002), one of the gatekeeper and a producer of chemicals, enzymes that are able to activate prode­ which will stop and eliminate many mi­ fensins. Matrilysin has been advocated crobes to further enter the small and as another activating enzyme (Wilson et

34 Figure 1: Defence mechanisms in the small intestine. al., 1999), however this enzyme could fensins to mature bioactive antimicrobial not be detected in germfree animals peptides in germfree animals (Axelsson (López-Boado et al., 2000) while others et al., 1999; Pütsep et al. 2000). have shown actual activation of prode-

ANTIMICROBIAL PEPTIDES, MOLECULAR PROPERTIES AND MODE OF ACTION The antimicrobial effect of antimi- peptides kill microbes is not known. crobial peptides on bacteria is com- Several models have been proposed and monly tested and expressed as for con- generally some kind of permeabilisation ventional antibiotics. The effective con- of the bacterial membrane is depicted centration range lies in the micro- to (van ‘t Hof et al., 2001). This is nanomolar range (Zasloff, 2002a). The accomplished by utilising electrostatic exact mechanism by which antimicrobial binding and the different hy­

35 Figure 2: Sections of small intestine stained with antibody to HD-5. Left: Positive Paneth cells in the very bottom of the crypt. Right: Positive cells are found higher up in the crypt and also positive material secreted into the lumen of the crypt. drophilic and hydrophobic properties of The range of microbes that are sensi­ the peptide (van ‘t Hof et al., 2001, tive to the antimicrobial peptides is quite Zasloff, 2002a). Interactions with lipids broad and the term broad-spectrum an­ have also been advocated and could play tibiotic applies to many of the peptides in concert with scavenger receptors (Periathamby and Dento, 2002; Porter et (Peiser et al., 2002). Certain antimicro­ al., 1997). Some strains show resis­ bial peptides, lactoferricin B and ma­ tance but this is due to the membrane gainin 2, have also been found to cross structure as such (Zasloff, 2002a) and over the bacterial membrane into the acquired resistance has been postulated cytoplasm (Haukland et al., 2001). as unlikely to occur (Peschel, 2002).

SMALL INTESTINAL MICROBIAL PEPTIDES Antimicrobial peptides seem to be mucus-producing goblet cells (Kamal et important to maintain a relatively mi­ al., 2001). This infection also lead to an crobe-free small intestine (Ganz, 2000; increase of Paneth cell-number, a more Ouelette et al., 2000; Ouelette and widespread presence of Paneth cells and Bevins, 2001). There is differential intermediate cells expressing mouse an­ presence of antimicrobial peptides timicrobial α-defensins, cryptdins (Ay­ throughout the gastrointestinal channel abe et al., 2002a). The modulation of (Frye et al, 2000). In mouse small in­ the mouse cryptdins has been shown to testine, the nematode Trichinella spiralis be dependent of Ca2+-activated potas­ induces atrophy of the villi, hyperplasia sium channels (Ayabe et al., 2002b). of the crypts of Lieberkühn and of the

36 A similar effect on Paneth cells has flora or being exposed to bacterial been seen in humans after Roux-en-y products, and in germfree animals. It Gastric bypass surgery (Sundbom et has been shown that germfree mice, al., 2002). This is a standard surgical which are bred for many generations in procedure for morbid obesity where an sterile environment after they have food and oral-nasal-pharyngeal secre­ been born under aseptic conditions and tion pass directly into the small bowel then maintained germfree, generate the without passing through the acid envi­ same products from enteric prodefens­ ronment of the normal stomach. Immu­ ins (Pütsep et al., 2000). So, animals nostaining of human intestinal antimi­ which are naive to microbes have pre­ crobial α-defensins, defensin-5 (HD-5), cursor forms of antimicrobial peptides showed an up-regulation of the anti­ which can be activated and momentarily bacterial peptide in the Paneth cells and exert their antimicrobial functions (Ay­ a spread of anti-defensin positive mate­ abe et al., 2002c). However, there is a rial upwards in the crypt wall and also a possibility that these animals are ex­ release inside the crypt lumen (Figure posed to bacterial products which have 2). been left unaffected by sterilisation in Since the acid environment of the their animal feed. For example there can stomach is by-passed, there is a possi­ be bacterial LPS originating in the raw bility for microbes to invade the other­ material or from the manufacturing wise protected small intestine. This process. could lead to overgrowth of bacteria in Experiments which compare germ­ these patients. However, these patients free animals with animals having a con­ display an almost normal microflora, ventional commensal microflora or with which could be the result of an activa­ animals mono-associated with a specific tion of anti-microbial peptides in re­ bacterial strain or species, can give sponse to an increased load of ingested valuable information of the interplay bacteria. between intestinal microflora and the HD-5 is normally stored in precursor individual or animal (Pütsep et al, form and is activated upon stimulation, 2000). The intestinal microflora has as discussed above, by bacteria but also profound effects on the development by inflammation (Axelsson, 1999; and maintenance of a healthy intestinal Cunliffe et al., 2001). However, in in­ mucosa (Falk et al., 1998). In the new­ flammatory bowel disease the epithelial born there is a succession of microbial barrier is defect and the intestinal tissue habitants building up this “normal” is exposed to bacteria and bacterial flora. Starting with bacteria acquired products which could be part of the from the mother during labour, neonates mechanism of this activation. Similar acquire for example Clostridium spp. changes in Paneth cell distribution as in and Bifidobacterium spp. and the the patients having gastric by-pass sur­ resultant flora is in part determined by gery could also be detected in patients environmental factors such as food and with active inflammatory bowel diseases eating habits, country of living, socio­ pointing to that similar mechanisms logical factors and level of sanitation could be at play, possibly involving (Falk et al., 1998). Postnatal studies bacterial interference with the mucosa have shown that the Paneth cells are (Cunliffe et al., 2002) found early and differentiate to mature This precursor form can be found in cells around postnatal day 14-28 (Bry et both individuals having a normal micro­ al., 1994).

37 PHARMACEUTICAL APPLICATIONS During the last years there has been a 2002). The most effective polymer was mounting problem with bacterial strains additionally tested with good results that have become resistant to commonly against ampicillin and streptomycin-re­ used antibiotics. Some strains have even sistant E. coli and tetracycline resistant acquired resistance to the antibiotics that Salmonella typhimurium. The ability to are used as a last means to treat life­ interact and disrupt phospholipid bi-lay­ threatening infections. Biotechnological ers was also confirmed (Tew et al., companies have seen the potential in 2002). using antimicrobial peptides for Another approach was used by self­ treatment of multi-resistant bacteria. For assembly of amino acids in synthetic example, Magainin Pharmaceuticals membranes into tubular structures (Fer­ Inc., Philadelphia, USA, was founded nandez-Lopez et al., 2001). Antibiotic in 1997 with one of the early peptide activity was established against several researchers, Dr. Michael Zasloff, as bacteria including methicillin resistant S. Executive Vice President. Clinical test­ aureus (MRSA) and the membrane ing has been done of some investiga­ disrupting ability was confirmed in a tional drug but no real break-through membrane depolarisation assay (Fer­ has been seen so far. One problem has nandez-Lopez et al., 2001). been the handling and administration of Another antimicrobial peptide has the drug, another problem, the very been identified in the mouse, cathelin­ high cost of manufacturing synthetic related antimicrobial peptide, CRAMP antimicrobial peptides. (Gallo et al., 1997). Analogues of this One approach has been to develop structure have now been designed and manufacturing processes to construct shown to have strong antibacterial ac­ the complex structures of biologically tivity, but without the endogenous effective antimicrobial peptides. Re­ CRAMPs haemolytic properties (Shin et cently, reports of successful synthesis al., 2000). of bioactive polymers displaying an­ With these developments there seem timicrobial activity have been publi­ to be hope for the emergence of effec­ cised. One approach has been to use tive synthetic antimicrobial peptides for amphiphilic acrylamide polymers re­ human use in the near future. These de­ sembling some of the properties of the velopments give the possibility to syn­ β-peptide class. These acrylamide thesise peptides that are suitable for polymers showed a bactericidal activity production at an industrial scale and at against E. coli and minimal inhibitory reasonable prices. However, clinical concentration (MIC) values for the testing is now needed to find out Gram-negative Klebsiella pneumoniae, whether they have adverse effects that ampicillin and streptomycin-resistant E. make them not suitable for use in man coli, and the Gram-positive Bacillus or animals. subtilis were established (Tew et al.,

LITERATURE

Axelsson, L.-G., Boman, A., Midtvedt, T., and tional antibacterial peptides, a response that Boman, H.G.: The small intestine of germ- is blocked by a cortisone pretreatment. Mi­ free mice contains antibacterial factors: A crob. Ecol. Health Dis. 11, 116-117 bacterial monocontamination induces addi- (1999).

38 Axelsson, L.-G. and Mahida. Y.: Antimicrobial 298-304 (2002). peptides in health and disease. Microb. Cunliffe, R.N., Rose, F.R.A.J., Keyte, J., Ecol. Health. Dis., Suppl. 2, 216-222 Abberley, L., Chan, W.C., and Mahida, (2000). Y.R.: Human defensin 5 is stored in precur­ Axelsson, L.-G.: Xenobiotics, intestinal mi­ sor form in normal Paneth cells and is ex­ croflora and innate gastrointestinal immu­ pressed by some villous epithelial cells and nity: What can clinical observations and by metaplastic Paneth cells in the colon in experimental animal models tell us about inflammatory bowel disease. Gut 48, 176­ environmentally related diseases. Microecol. 185 (2001). Ther. 28, 327-335 (1999). Del Pero, M., Boniotto, M., Zuccon, D., Ayabe, T., Satchell, D.P., Pesendorfer, P., Cervella, P., Spanò, A., Amoroso, A., and Tanabe, H., Wilson, C.L., Hagen, S.J., and Crovella, S.: β-defensin 1 gene variability Ouelette, A.J.: Activation of paneth cell α­ among non-human primates. Immunogenet­ defensins in mouse small intestine. J. Biol. ics 53, 907-913 (2002). Chem. 277, 5219-5228 (2002a). Deplancke, B. and Gaskins, H.R.: Microbial Ayabe, T., Wulff, H., Darmoul, D., Cahalan, modulation of innate defense: Goblet cells M.D., Chandy, K.G., and Ouelette, A.J.: and the intestinal mucus layer. Am. J. Modulation of mouse paneth cell α-de­ Clin. Nutr. 73 (Suppl.) 1131S-1141S fensin secretion by mIKCa1, a Ca2+-acti­ (2001). vated, intermediate conductance potassium Falk, P.G., Hooper, L.V., Midtvedt, T., and channel. J. Biol. Chem. 277, 3793-3800 Gordon, J.I.: Creating and maintaining the (2002b). gastrointestinal ecosystem: What we know Ayabe, T., Satchell, D.P., Wilson, C.L., and need to know from gnotobiology. Mi­ Parks, W.C., Selsted, M.E., and Ouelette, crobiol. Molec. Biol. Rev. 62, 1157-1170 A.J.: Secretion of microbicidal α-defensins (1998). by intestinal Paneth cells in response to Fellerman, K. and Stange, E.: Defensins - In­ bacteria. Nature Immunol. 1, 113-118 nate immunity at the epithelial frontier. (2000c). Eur. J. Gastroenterol. Hepatol. 13, 771-776 Bevins, C.L., Martin-Porter, E., and Ganz, T.: (2001). Defensins and innate host defence of the Fernandez-Lopez, S., Kim, H.S., Choi, E.C., gastrointestinal tract. Gut 45, 911-915 Delgado, M., Granja, J.R., Khasanov, A., (1999). Kraehenbuehl, K., Long, G., Weinberger, Boman HG.: Peptide antibiotics and their role D.A., Wilcoxen, K.M., and Ghadiri, M.R.: in innate immunity. Ann. Rev. Immunol. Antibacterial agents based on the cyclic 13, 61-92 (1995). D,L-alpha-peptide architecture. Nature 412, Boman, H.G.: Innate immunity and the normal 452-455 (2001). microflora. Immunol. Rev. 173, 5-16 Frye, M., Bargon, J., Lembcke, B., Wagner, (2000). T.O.F., and Gropp, R.: Differential expres­ Bry, L., Falk, P., Huttner, K., Ouellette, A., sion of human α- and β-defensins mRNA Midtvedt, T., and Gordon, J.I.: Paneth cell in gastrointestinal epithelia. Eur. J. Clin. differentiation in the developing intestine of Invest. 30, 695-701 (2000). normal and transgenic mice. Proc. Natl. Gallo, R.L., Kim, K.J., Bernfield, M., Kozak, Acad. Sci. USA 91, 10335-10339 (1994). C.A., Zanetti, M., Merluzzi, L., and Gen­ Corzo, G., Escoubas, P., Villegas, E., naro, R.: Identification of CRAMP, a Barnham, K.J., He, W., Norton, R.S., and cathelin-related antimicrobial peptide ex­ Nakajima, T.: Characterization of the pressed in the embryonic and adult mouse. unique amphipatic antimicrobial peptides J. Biol. Chem. 272, 13088-13093 (1997). from venom of the scorpion Pandinus im­ Ganz, T. and Lehrer, R.I.: Defensins. Curr. perator. Biochem. J. 359, 35-45 (2001). Opin. Immunol. 6, 584-589 (1994). Cunliffe, R.N. Kamal, M., Rose, F.R.A.J., Ganz, T.: Paneth cells- Guardians of the gut James, P.D., and Mahida, Y.R.: Expression cell hatchery. Nature Immunol. 1, 99-100 of antimicrobial nutrophil Defensins in (2000). epithelial cells of active inflammatory Ghosh, D., Porter, E., Shen, B., Lee, S.K., bowel disease mucosa. J. Clin. Pathol. 55, Wilk, D., Drazba, J., Yadav, S.P., Crabb,

39 J.Ganz, T., and Bevins, C.L.: Paneth cell immunity in mice. Infect. Immun. 70, trypsin is the processing enzyme for human 3068-3072 (2002). defensin-5. Nature Immunol. 3, 583-590 Ouelette, A.J. and Bevins, C.L.: Paneth cell (2002). defensins and innate immunity of the small Gillespie, S.H.: Microbes sans frontières. Lan­ bowel. Inflamm. Bowel Dis. 7, 43-50 cet 359, 93-95 (2002). (2001). Hancock, R.E.W. and Diamond, G.: The role Ouelette, A.J., Satchell, D.P., Hsieh, M.M., of cationic antimicrobial peptides in innate Hagen, S.J., and Selsted, M.E.: Characteri­ host defences. Trends Microbiol. 8, 402­ zation of luminal Paneth cell α-defensins in 410 (2000). mouse small intestine. Attenuated antimi­ Hancock, R.E.W.: Cationic peptides: Effectors crobial activities of peptides with truncated in innate immunity and novel antimicrobi­ amino termini. J. Biol. Chem. 275, 33969­ als. Lancet i, 156-164 (2001). 33973 (2000). Haukland, H.H., Ulvatne, H., Sandvik, K., and Parkin, J. and Cohen B.: An overview of the Vorland, L.H.: The antimicrobial peptides immune system. Lancet 357, 1777-1789 lactoferricin B and magainin 2 cross over (2001). the bacterial cytoplasmic membrane and re­ Peiser, L., Mukhopadhyay, S., and Gordon, S.: side in the cytoplasm. FEBS Lett. 508, Scavenger receptors in innate immunity. 389-393 (2001). Curr. Opin. Immunol. 14, 123-128 (2002). Hoffman, J.A. and Reichhart, J-M.: Drosophila Peschel, A.: How do bacteria resist human an­ innate immunity: An evolutionary perspec­ timicrobial peptides? Trends Microbiol 10, tive. Nature Immunol. 3, 121-126 (2002). 179-186 (2002). Hooper, L.V. and Gordon J.I.: Commensal Periathamby, A.R. and Dentino, A.R.: Current host-bacterial relationships in the gut. Sci­ status of defensins and their role in innate ence 292, 1115-1118 (2001). and adaptive immunity. FEMS Microbiol. Kamal, M., Wakelin, D., Ouelette, A.J., Lett. 206, 9-18 (2002). Smith, A., Podolsky, D.K., and Mahida, Porter, E.M., vanDam, E., Valore, E.V., and Y.R.: Mucosal T cells regulate paneth and Ganz, T.: Broad-spectrum antimicrobial ac­ intermdiate cell numbers in the small intes­ tivity of human intestinal defensin 5. In­ tine of T. spiralis-infected mice. Clin. Exp. fect. Immun. 65, 2396-2401 (1997). Pathol. 126, 117-125 (2001). Porter, E.M., Bevins, C.L, Gosh, D., and López-Boado, Y.S., Wilson, C.L., Hooper, Ganz, T.: The multifaceted Paneth cell. L.V., Gordon, J.I., Hultgren, S.J., and Cell. Mol. Life Sci. 59, 156-170 (2002). Parks, W.C.: Bacterial exposure induces and Pütsep, K., Axelsson, L-G, Boman, A., activates matrilysin in mucosal epithelial Midtvedt, T., Normark, S., Boman, H.G., cells. J. Cell. Biol. 148, 1305-1315 (2000). and Andersson, M.: Germ-free and colonized Luster, A.D.: The role of chemokines in link­ mice generate the same products from en­ ing innate and adaptive immunity. Curr. teric prodefensins. J. Biol. Chem. 275, Opin. Immunol. 14, 129-135 (2002). 40478-40482 (2000). Mahida, Y.R. and Johal, S.: NF-κB may de­ Schonwetter, B.S., Stolzenberg, E.D., and termine whether epithelial cell-microbial in­ Zasloff, M.A.: Epithelial antibiotics in­ teractions in the intestine are hostile or duced at sites of inflammation. Science friendly. Clin. Exp. Immunol. 123, 347­ 267, 1645-1648 (1995). 349 (2001). Schutte, B.C., Mitros, J.P., Bartlett, J.A., Medzhitov, R. and Janeway, C. Jr.: Innate Walters, J.D., Jia, H.P., Welsh, M.J., immunity. New. Eng. J. Med. 343, 338­ Casavant, T.L., and McCray, P.B.Jr.: Dis­ 344 (2000). covery of five conserved β-defensin gene Medzhitov, R. and Janeway, C.A. Jr.: Decod­ clusters using a computational search strat­ ing the patterns of self and nonself by the egy. PNAS 99, 2129-2133 (2002). innate immune system. Science 296, 298­ Schwartz, D.A.: The genetics of innate immu­ 300 (2002). nity. Chest 121, 62S-68S (2002). Moser, C., Weiner, D.J., Lysenko, E., Bals, Sepp E.: Formation of intestinal microbial R., Weiser, J.N., and Wilson, J.M.: β-de­ ecosystem in children. Dissertationes Medi­ fensin 1 contributes to pulmonary innate cinae Universitatis Tartunensis 43, Tartu

40 University Press, Tartu 1998. tric by-pass surgery. Microecol. Ther. 29, Shin, S.Y, Kang, S.-W., Lee, D.G., Eom, 191-193 (2002). S.H., Song, W.K., and Kim, J.I.: CRAMP Tew, G.N., Liu, D., Chen, B., Doerksen R.J. analogues having potent antibiotic activity Kaplan, J., Carroll, P.J., Klein, M.L., and against bacterial, fungal and tumor cells DeGrado, W.F.: De novo design of without hemolytic activity. Biochem. Bio­ biomimetic antimicrobial polymers. PNAS phys. Res. Commun. 275, 904-909 (2000). 99, 5110-5114 (2002). Skar, V., Skar, A.G., Midtvedt, T., and Osnes, van ‘t Hof, W., Veerman, E.C.I., Helmerhorst, M.: Bacterial growth in the duodenum and E.J., and Nieuw Amerongen, A.V.: Antim­ in the bile of patients with gallstone disease icrobial peptides: Properties and applicabil­ treated with endoscopic papillotomy. Endo­ ity. Biol. Chem. 382, 597-619 (2001). scopy 18, 10-13 (1986). Wilson, C.L., Ouelette, A.J., Satchell, D.P., Skar, V., Skar, A.G., and Osnes, M.: The duo­ Ayabe, T., López-Boado, Y.S., Stratman, denal bacterial flora in the region of papilla J.L., Hultgren, S.J. Matrisian, L.M., and of Vater in patients with and without duo­ Parks, W.C.: Regulation of intestinal al­ denal diverticula. Scand. J. Gastroenterol. pha-defensin activation by the metalloprote­ 24, 649-656 (1989). inase matrilysin in innate host defense. Sci­ Sundbom, M., Cunliffe, R.N., Mahida, Y.R., ence 286, 113-117 (1999). Rubio, C., Midtvedt, T., Gustavsson, S., Zasloff, M.: Antimicrobial peptides of mul­ and Axelsson, L.-G.: Differential crosstalk ticellular organisms. Nature 415, 389-395 between intestinal flora and Panethcells (2002a). regulates expression of lysozyme and hu­ Zasloff, M.: Trypsin, for the defense. Nature man defensin 5 in obese patients after gas­ Immunol. 3, 583-590 (2002b).

41 42 PHENOTYPIC EXPRESSIONS IN THE SMALL INTESTINE ELISABETH NORIN Microbiology and Tumorbiology Center, Karolinska Institute, Stockholm, Sweden

INTRODUCTION Every man and animal is born germ­ overcome today by the use of po­ free, i.e., without any microorganisms, lymerase chain reaction and direct se­ and the colonisation of all surfaces such quencing of 16S ribosomal DNA analy­ as the respiratory, urinary and alimen­ sis of the flora. However, there are still tary tract starts immediately after birth. limitations and the detection level is ap­ Initially, when space is not limited, proximately the same as in previously bacteria with a high multiplication rate well established microbiological evalua­ may dominate, but as the number of tions. A complementary method is to bacteria increases and accessible nutrient evaluate what the flora has done, i.e., to pools becomes limited, habitats will be evaluate the products – the outcome - of filled up by more specialised bacterial the crosstalk between the host and its species and the complexity of the flora microflora. Using this technique, sub­ increases. stances produced in, e.g., the small in­ When trying to investigate the com­ testine can be detected in faecal samples position of samples from any site the and thus reflect microbiological prod­ alimentary tract other than the mouth or ucts from areas normally difficult to lower part of large intestine from vol­ obtain samples from. unteers, it is difficult to obtain proper Comparisons of conventionally samples. Endoscopic sampling for mi­ raised (Conv) organisms and germfree crobiological evaluations has both ad­ (GF) counterparts have revealed series vantages and disadvantages, and some of anatomic, biochemical, immunologi­ investigators have also pointed on some cal and physiological phenotypes, col­ difficulties in the sampling depending lectively known as Microflora-Associ­ on from where it is taken in the stool ated Characteristics or MACs. When the specimens. Porcine faecal material was functionally active flora is absent, as in investigated with regard to microbes GF animals, healthy new-borns and present in materials from the inner and sometimes in relation to antimicrobial the outer content, and differences were treatment, a MAC is termed GAC found due to atmospheric conditions. (Germfree Animal Characteristic) Another problem arises for instance due (Midtvedt et al., 1985). Some pheno­ to storage and/or freezing and transport typic expressions are presented in Table conditions from the sampling to analysis 1. In the following, some of these phe­ – these variables will also influence the notypic expressions, occurring in the final findings (Rall et al., 1970). Some small intestine, will be presented and of these problems can however be discussed.

43 Table 1: Some anatomical structures, physiological and biochemical functions influenced by the microflora, and microorganisms involved (Modified from Midtvedt, 1999) ——————————————————————————————————————— Parameter MAC1 GAC2 Microorganism ——————————————————————————————————————— Anatomical/Physiological Caecum size (rodents) Normal Enlarged Partly known Cell kinetics Normal Slower Unknown Colloid osmotic pressure Normal Increased Unknown Electro-potential Eh, mV Low (<-100) High (>-100) Unknown Intestinal wall Thick Thin Unknown Migration motor complexes Normal Fewer Unknown Osmolality Normal Reduced Unknown Oxygen tension Low High Several species Production of peptides Normal Altered Unknown Biochemical β-aspartylglycine Absent Present Species in concert β-glucuronidase Low activity High activity Several species Bile acid metabolism Deconjugation No deconjugation Many species Dehydrogenation No dehydrogenation Many species Dehydroxylation No dehydroxylation A few species Bilirubin metabolism Deconjugation Little deconjugation Many species Urobilinogen No urobilinogen A few species Cholesterol metabolism Coprostanol No coprostanol A few species Faecal tryptic activity Little or absent High activity A few species Intestinal gasses Carbon dioxide Some CO2 Many species Hydrogen No hydrogen Some species Methane No methane A few species Mucin Degradation No degradation Several species Short-chain fatty acids Large amounts Far less Many species Several acids Few acids ——————————————————————————————————————— 1 Microflora-Associated Characteristic; 2 Germfree Animal Characteristic.

CELL KINETICS Since long time it has been assumed we have been able to show that there are that the rate of crypt epithelial cell pro­ different phenotypic expressions in dif­ liferation in the intestine represents a ferent compartments of the intestine in major defence mechanism against in­ GF and Conv rats and mice with regard vading intestinal microorganisms. The to age, gender and microbial status. In cell renewal system involves prolifera­ these studies, also diet and fasting time tion of undifferentiated epithelial cells was standardised. Cell kinetic and mor­ followed by differentiation and migra­ phological parameters reflecting the tion from the site of production to the crosstalk between the host and its flora functional site and finally elimination was investigated (Banasaz et al., 2000, from the mucosa. Under normal condi­ 2001). tions in mice, the cell turn over rate has We found a great similarity between been estimated to be about 108 cells per the mitotic index in young and old rats day (Hageman et al., 1970). By intro­ and mice, being high in the upper part ducing strict standardisation techniques, of the small intestine and lower in the

44 colon, and we also found higher mitotic young rats, there was initially an in­ indexes in males as compared to fe­ crease of the mitotic index. However, males. A second similarity was that the within some few days there was a crypt/villus ratio was on the same level marked decrease of the mitotic index throughout the small intestine. Thus, we and also occurrence of some epithelial found that in those areas, where the border disruptions without leading to number of microbes was high, microbes any disease or other signs of discomfort triggered the mitotic index. We also in the animals was seen. Lack of disease found that, when establishing a mi­ signs does not exclude that Clostridium crobial strain as a mono-contaminant, an difficile is a pathogenic microbe. By immediate triggering of the mitotic index opening up for other microbes to cause was seen, irrespectively of whether the disease, as we observed some few mono-contaminant was a probiotic patchy morphological alterations in the microbe (Banasaz et al., 2002), a intestinal mucosa of the rats, the disease pathogen or a commensal microbe signs could be caused by secondary (Banasaz, 2002). infections. This study needs to be When a toxin producing Clostridium expanded to animals harbouring a difficile strain was mono-inoculated into normal flora.

INTESTINAL MOTILITY It is well known that several micro­ time – some species influencing the bial species may cause increased intesti­ small intestine (Salmonella spp.) and nal motility, expressed as cramps and others the large intestine (Shigella spp.). diarrhoea, due to many different The microbes responsible for induc­ mechanisms. GF animals are known for ing motility, contractions and increased having far less spontaneous muscular transit time, found under normal contractions than their Conv counter­ physiological conditions, are not parts. Already in the sixties, a slower known. However, intestinal movements transit time in GF animals was reported are of a paramount importance for the (Abrams and Bishop, 1967; Gustafsson regulatory and protective role of the mi­ and Norman, 1969). Obviously, the croflora, and for the host (Midtvedt, intestinal flora plays an important role 1989). related to intestinal motility and transit

SHORT CHAIN FATTY ACIDS The intestinal microflora ferments the these acids are representing the net sum dietary and endogenous large carbo­ of production, absorption and secretion hydrates into mono- and di-saccharides of the acids throughout the whole intes­ in the small intestine and these appear to tinal tract. be the main contributors to the energy In the mouth of man, there are quite requirement in mammals after an an­ high amounts of these acids present, aerobic fermentation. The origin of in­ representing products of an anaerobic testinal short chain fatty acids has been metabolism in the gingival pockets. In substantiated in studies of GF and Conv the stomach and upper small intestine, rats and mice, and the faecal content of the amount of these acids is quite low

45 under normal, healthy conditions – in accounted for almost 11%, and less than contrary to, e.g., patients with microbial 2% of iso- and n-butyric and i-valeric small intestinal bacterial overgrowth. acid (Høverstad et al., 1984). A very Findings from these patients indicate similar relative distribution was earlier that there are overgrowth symptoms ­ found in the saliva, gastric juice and the patients seem to have a colon-like duodenal aspirates, however, essential flora in the lower small intestine, and different from the faeces content. This the main part of these acids are depends on the fact that, as the number produced by the altered flora in the jeju­ of microbes increases in the lower small num (Høverstad et al., 1885). As an intestine and in the colon, also the short example can be mentioned that in 6 chain fatty acid content is altered in the healthy volunteers, the short chain fatty composition, and totally 24-243 acid content in saliva was quite high mmol/kg faeces – with significantly (2780-9940 µmol/l), decreased succes­ higher amounts in men as compared to sively to 185-1470 µmol/l in jejunal women, have been found (Siigur et al., juice, and the main part was acetic acid 1994). (approximately 85%), propionic acid

INTESTINAL TRYPTIC ACTIVITY Trypsin is chosen as a model sub­ It is also shown that new-born chil­ stance for studying endogenous derived dren excrete faeces/meconium without digestive enzymes. The precursor – any tryptic activity, although immu­ trypsinogen - is excreted from the pan­ nological studies detect the molecule – creas and activated in the upper part of not yet activated (Norin, 1985). There­ the small intestine, mainly by brush after, the enzyme activity increases suc­ border enzymes. cessively during the first year of life In total, faecal tryptic activity in­ followed by a decrease down to adult volves the net sum of processes such as values some years later (Norin et al., secretion of trypsinogen from the pan­ 1985). creas, activation of the pro-enzyme in In Conv animals it is found that the the small intestine by enterokinase and enzymatic activity disappears mainly in presence of host-, microbial-, and diet­ the caecum. Only very seldom, a low derived compounds that inactivate or level of tryptic activity is detected in the otherwise degrades the trypsin molecule lower intestine and in faecal samples during the passage through the intestine. from rats and mice. It is found that the In GF rats, tryptic activity is detected in amount of this enzyme activity is vari­ the upper small intestine already at two able depending on species investigated days of age, and as the animals grow and on which diet the animals is given. older, tryptic activity is detected in Horses and pigs show the same low increasing amounts all the way down of levels of faecal tryptic activity (Collinder the intestine and in faecal samples, et al., 2000, 2002), as found in e.g. where high amounts of the enzyme rats, mice and man. activity is detected. This in contrast to Obviously, intestinal microbes are what has been found in materials from responsible for the inactivation of tryp­ Conv animals, which are more or less sin, and at least one human strain of devoid of tryptic activity in the lower Bacteroides distasonis (Ramare et al., intestine. 1996) has been isolated and found ca­

46 pable to inactivate pancreas derived nal tryptic activity could possibly con­ trypsin in both rats and mice. Previ­ tribute to the pathological symptoms ously, in Crohn´s disease patients, there observed, mainly in the lower intestine were found high levels of tryptic activity of these patients. These microbes are in faecal samples as compared to also found to be influenced by several samples from healthy volunteers (van antimicrobial drugs, when given to rats der Merwe and Mol, 1982), and intesti­ (Norin, 1997).

BILIRUBIN AND UROBILINS The bile pigments consist partly of dases and further transformed to series bilirubin, a toxic and water insoluble of metabolites, most often termed uro­ end-product after catabolism of haemo­ bilins. Most of the ß-glucuronidases are globin and some other haem-containing derived from the microbes, only a minor substances. The reduction of bilirubin to part is produced by the host. This urobilins by intestinal microbes rep­ detoxification process will be discussed resents one natural detoxification proc­ in a deeper setting later during this ess of toxic intestinal substances such as meeting. Enhancement of microbial xenobiotics, drugs, hormones and conversion of bilirubin to urobilins de­ certain dyes. Bilirubin as well as other creases the intestinal concentration of toxic substances is conjugated in the bilirubin, which is a potential risk factor liver with glucuronate to a less toxic and for hyperbilirubinaemia, which can lead water-soluble molecule, which are se­ to extrapyramidal disturbances, hearing creted with the bile into the intestine. loss, delay in motor development and In the intestine, the bilirubin conju­ less often, also to intellectual deficits gates are de-conjugated by ß-glucuroni­ (Saxerholt, 1990; Vitek et al., 2000).

CONCLUDING REMARKS The normal intestinal flora plays an troduction of live microbes when e.g., important regulatory and protective role the intestinal flora is still not fully settled in all organisms, but external distur­ or otherwise strongly disturbed, could bances could be harmful for the host alter the succession of the establishment and its flora. It is, e.g., known that an­ of the normal flora. One could thus timicrobial treatments causing marked speculate that, when the intestinal flora alterations of the flora-composition is immature and continuously de­ could induce altered intestinal functions. veloping, an altered succession of the Supplementation of the intestinal micro­ “normal” establishment pattern of the flora with live microbial species, used flora could cause unexpected conse­ for many years to protect and maintain quences related to functions both in the the balance in the intestine, could even­ small and in the large intestine. tually also be a potential risk factor. In­

LITERATURE

Abrams, G.D. and Bishop, J.E.: Effect of the motility. Proc. Soc. Exp. Biol. Med. 126, normal microbial flora on gastrointestinal 301-304 (1967).

47 Gustafsson, B.E. and Norman, A.: Influence of B., Gustafsson, B.E., Høverstad, T., Lin­ the diet on the turnover of bile acids in gaas, E., Norin, K.E., Saxerholt, H., and germfree and conventional rats. Br. J. Nutr. Steinbakk, M.: The influence of antibiotics 1969; 23,429-442. upon micro-flora-associated characteristics Banasaz, M., Alam, M, Norin, E., and in man and mammals. In: Germfree re­ Midtvedt, T.: Gender, age and microbial search; Micro-flora control and its applica­ status influence upon intestinal cell kinetics tion to the biomedical sciences (Eds.: in a compartmentalised manner. An ex­ Wostmann, B.S., Pleasants, J.R., Pollard, perimental study in germfree and conven­ P., Thea, B.A., and Wagner, M.). Alan R. tional rats. Microb. Ecol. Health Dis. 12, Liss, New York, 241-244 (1985). 208-218 (2000). Midtvedt, T.: The normal microflora, intestinal Banasaz, M., Norin, E., and Midtvedt, T.: The motility and influence of antibiotics. An role of gender, age and microbial status on overview. In: The regulatory and protective cell kinetics in the gastrointestinal tract of role of the normal microflora (Eds.: Grubb, mice. An experimental study in germfree R., Midtvedt, T., and Norin, E.). The and conventional mice. Microb. Ecol. MacMillan Press Ltd., London, 147-167 Health Dis. 13, 135-142 (2001). (1989). Banasaz, M., Norin, E., Holma, R., and Midtvedt, T.: Microbial functional activities. Midtvedt, T.: Increased enterocyte produc­ In: Probiotics, other nutritional factors, and tion in gnotobiotic rats mono-associated intestinal microflora. Nestlé Nutrition with Lactobacillus rhamnosus GG. Appl. Workshop 1997, Vol 42 (Eds.: Hansson, Env. Microbiol. 68, 3031-3034 (2002). L.Å. and Yolken, R.H.). Lippincott-Raven Banasaz, M.: Intestinal Cell Kinetics. Modula­ Publishers, Philadelphia, 79-96 (1999). tion caused by age, gender and microbial Norin, K.E.: Germfree animal characteristics. A status in rats and mice. An experimental study of some functional aspects of intes­ study in germfree, conventional and Lacto­ tinal flora in rats and man. Thesis, Karolin­ bacillus rhamnosus GG or Clostridium dif­ ska Institute, Stockholm, Sweden, 1985. ficile mono-associated animals. Thesis, Norin, K.E., Gustafsson, B.E., Lindblad, B.S., Karolinska Institute, Stockholm, Sweden, and Midtvedt, T.: The establishment of 2002. some micro-flora associated biochemical Collinder, E., Lindholm, A., Midtvedt, T., and characteristics in faeces from children during Norin, E.: Six intestinal micro-flora-associ­ the first years of life. Acta Paediatr. Scand. ated characteristics in sport horses. Equine 74, 207-212 (1985). Vet. J. 32, 222-227 (2000). Norin, K.E.: Influence of antibiotics on some Collinder, E., Cardona, M., Kozakova, H., intestinal microflora associated characteris­ Norin, E., Stern, S., and Midtvedt, T.: Bio­ tics. Anaerobe 3, 145-148 (1997). chemical intestinal parameters in pigs reared Rall, G.D., Wood, A.J., Wescott, R.B., and outdoors and indoors, and in germ-free pigs. Dommert, A.R.: Distribution of bacteria in J. Vet. Med. A, 49, 203-209 (2002). feces of swine. Appl. Microbiol. 20, 789­ Hageman, R.F., Sigdestad, C.P., and Lesher, 792 (1970). S.: A quantitative description of the intesti­ Ramare F, Hautefort I, Verhe F, Raibaud P, nal epithelium of the mouse. Am. J. Anat. Iovanna J.: Inactivation of tryptic activity 1, 390-429 (1970). by a human derived strain of Bacteroides Høverstad T, Bjørneklett A, Midtvedt T, Fausa distasonis in the large intestines of gnoto­ O, 1984 and Bøhmer T. Short-chain fatty biotic rats and mice. Appl. Environm. Mi­ acids in the proximal gastrointestinal tract crobiol. 62, 1434-1436 (1996). of healthy subjects. Scand. J. Gastroenterol. Saxerholt, H.: Intestinal bilirubin metabolism. 19, 1053-1058 (1984). Thesis, Karolinska Institute, Stockholm, Høverstad T, Bjørneklett A, , Fausa O, and Sweden, 1990. Midtvedt, T.: Short-chain fatty acids in the Siigur, U., Norin, K.E., Allgood, G., Schla­ small-bowel bacterial overgrowth syn­ gheck, T., and Midtvedt, T.: Concentrations drome. Scand. J. Gastroenterol. 20, 492­ and correlations of faecal short-chain fatty 499 (1984). acids and faecal water content in man. Mi­ Midtvedt,T., Bjørneklett, A., Carlstedt-Duke, crob. Ecol. Health Dis. 7, 287-294 (1994).

48 van der Merwe, J.P. and Mol, G.J.J.: Levels of Chmelar D, Fevery J.: Intestinal colonisa­ trypsin and a-chymotrypsin in feces from tion leading to fecal urobilinoid excretion patients with Crohn´s disease. Digestion may play a role in the pathogenesis of neo­ 24, 1-4 (1982). natal jaundice. J. Pediatr. Gastroenterol. Vitek L, Kotal P, Jirsa M, Malina J, Cerna M, Nutr. 30, 294-298 (2000).

49 50 MICROBIAL P450: DOES IT EXIST, AND WHAT CAN IT MEAN? TORE MIDTVEDT Laboratory of Medical Microbial Ecology, Department of Cell and Molecular Biology, Karolinska Institute, Stockholm, Sweden

SUMMARY P450 enzymes play important physiological and patho-physiological roles in the complex interplay between a host and his intestinal micro­ flora. The need for more information is underlined.

NOMENCLATURE The cytochrome P450s (CYP) con- another family. The families are divided stitute a superfamily of haem-thiolate into sub-families and enzymes within a enzymes and their Fe-carbon complexes sub-family are more than 55% identical show an absorption spectrum with a in sequence. A P450 enzyme is desig­ maximum near 450. There are now over nated by the root symbol “CYP” (de­ 1000 different P450s that have been scribing cytochrome and P450), an identified from the whole biological Arabic number denoting the family, a kingdom. Due to this large number a letter designating the subfamily and a standardised nomenclature has been de- Arabic numerals representing individual veloped (Nelson et al., 1996). In short, enzymes. It should be kept in mind that the P450 superfamily is subdivided into this system is based only on sequence families. An individual P450 within a similarity among the P450s and - un­ family are defined as having less than fortunately does not indicate the func­ 40% sequence identity with a P450 in tion(s) of individual P450s.

P450s AND METABOLISM OF XENOBIOTICS Nowadays, it is generally recognised Also due to their lipophilicity – if the that multicellular organisms, including xenobiotics are not metabolised in the humans, are continuously exposed to body - they will be concentrated in the foreign chemicals, collectively known tissue and sooner ore later they might be as xenobiotics. They are found in our toxic for the host. Therefore, in order to environment and include a vast range of be eliminated, many xenobiotics have to compounds such as drugs, industrial be converted into more water-soluble chemical, pollutants, pesticides, plant compounds, thereby influencing upon products, alkaloids and toxins. Most of their excretion in the urine or faeces. them are rather lipophilic, and due to The enzymes catalysing these reac­ their lipophilicity, many xenobiotics can tions, i.e. the xenobiotic-metabolising be - and are - absorbed through our sur­ enzymes, are – for convenient reasons ­ faces (intestine, lungs, and the skin). often divided into two groups refereed

51 Table 1: Major reactions and groups of enzymes in xenobiotic biotransformation —————————————————————————————————— Reaction Type of enzyme —————————————————————————————————— Phase I Cytochrome P450 Alcohol dehydrogenase Aldehyde dehydrogenase Oxidation Xanthine oxidase Monoamino oxidase Flavin mono-oxidase Reduction Quinone reduction Reductive dehalogenation (P450) Hydrolysis Epoxide hydrolase

Phase II Glucuronide conjugation UDP-glucuronosyltransferase Glutathione conjugation Glutathione S-transferase Sulphate conjugation Sulphotransferase Acetylation N-acetyltransferase Methylation Methyltransferase —————————————————————————————————— to as phase I and phase II (Table 1) As haviour of many drugs. In fact, some summarised by McLellan (2000), en­ drugs have to undergo biotransforma­ zymes involved in phase I reactions, tion before exerting their effects. On the most of which represent P450s, expose other side, however, biotransformation or introduce a function group (-OH, ­ included in phase I may convert many NH2, -SH or -COOH) on the com­ xenobiotics to more reactive electro­ pounds by oxidation, reduction or hy­ philic metabolites that can form protein drolysis reactions among others. In and DNA adducts, thereby exerting their general, these alterations increase hy­ toxic or tumourigenic effect (Nebert et drophilicity to a minor extent. It is evi­ al., 1996) It should also be kept in mind dent from Table 1 that phase II mainly that important co-factors for phase II involves conjugation of the compound, reactions are functional groups that are with molecules such as gluthatione, either present on the xenobiotics or have glucuronic acid, sulphate, taurine, gly­ been introduced during phase I reaction. cine and other amino acids. Thus, phase II biotransformation may It has to be kept in mind that the or may not be preceded by phase I biotransformations included in phase I biotransformation may change the pharmacokinetic be­

P450s AND EVOLUTION It is generally believed that P450s are metabolism of cholesterol to steroids) is very old enzyme, probably occurring found in both plants and animals, indi­ before the divergence of prokaryote and cated that animals diverged from plants eukaryote. A early eukaryotic mito- around 1400 millions years ago (giving chondrial P450 (influencing upon the rise to P450s localised in the mitochon­

52 Table 2: Major human P450 families and primary function(s) ———————————————————————————— Family Catalytic function

———————————————————————————— CYP1 Xenobiotic metabolism CYP2 Xenobiotic catabolism CYP3 Xenobiotic catabolism CYP4 Fatty acid hydroxylation CYP5 Thromboxane A2 synthase CYP7A Cholesterol 7-alpha-hydroxylase CYP8A Prostacyclin synthase CYP8B Sterol 12-alpha-hydroxylase CYP11A1 Cholesterol side-chain cleavage CYP11B1 Steroid 11-beta-hydroxylase CYP11B2 Aldosterone synthase CYP21 Steroid 21-hydroxylase CYP27A1 Sterol 27-hydroxylase CYP46 Cholesterol 24-hydroxylase ———————————————————————————— dria and endoplasmatic reticulum. A (Schuler, 1996). Going back in history, major evolutionary steps seems to have it seems reasonable to assume that when taken place around 900 millions years animals started to consume plants, the ago, resulting in one lineage continuing plants responded by evolving new as endogenous P450s and the other be­ genes to synthesise toxic metabolites. In gan a new function, i.e. xenobiotic me­ order to defend themselves from these tabolism (Nelson and Strobel, 1997). plant toxins, animals developed new In this respect, it is a fascinating theory enzymes to cope with these new plant that the xenobiotic-metabolising en­ toxins (Gonzales and Nebert, 1990). zymes have involved due to a continu­ Thus, the P450s have – and have played ous interaction or “evolutionary fight” – a crucial role in the ecological balance between plants and animals (Gonzales in Mother Nature. and Nebert, 1990; Nebert, 1997). In this respect, it is neither surprising Plants are continuously evolving bio­ that many prokaryotes may contain synthetic pathways in order to synthe­ P450s nor that many of our currently sise secondary metabolites for their re­ used drugs, often derived from natural productive cycles and to defend them­ plant metabolites, are metabolised by the selves from insect and animal predators P450 superfamily of enzymes.

XENOBIOTIC METABOLISM AND SUBSTRATE SPECIFICITY

As shown in Table 2, the CYP fami- to varying degrees – by several different lies 1,2 and 3 are primarily associated P450s and that one single enzyme can with xenobiotic metabolism. Most of the metabolise numerous, structurally di­ enzymes belonging to these 3 families verse chemicals. Taken together, the have an extremely broad substrate enzymes, including in these three fami­ specificity. It has also to be mentioned lies have a collective capacity to me­ that many substances are metabolised – tabolise – most often detoxify - an

53 enormous number of substances that we more upon endogenous compounds may be exposed to. than xenobiotics. On the other hand, It should be mentioned that, in gen­ enzymes belonging to the three first eral, enzymes belonging to other fami­ families may also metabolise some en­ lies have a higher degree of substrate dogenous compounds, as steroids. specificity and they are usually acting LOCALISATION OF P450s By far, liver is the main site of ex­ nation in cases of severe liver cirrhosis pression of xenobiotic-metabolising (Krishna and Klotz, 1994). This can be P450s. However, some of them may be exemplified as follows. The CYP3A found in extrahepatic tissue. It is known enzymes are involved in the metabolism that that extrahepatic tissue may con­ of around 50% of clinically useful tribute to the xenobiotic-metabolising agents and have a very wide substrate capacity of the body. In turn, this might specificity. CYP3A4is one of the major results in a high local turnover of a enzymes within this family. It is found drug, thereby influencing upon the local at relatively high levels in enterocytes in effect of the drug. Such extra-hepatic the small intestine. Similar to what metabolism might even compensate to found in the liver, it can be induced by some extent for reduced hepatic elimi­ rifampin (Kolars et al., 1994).

P450s AND MICROORGANISMS Over the years, it has been found that conventional animals have shown that cytochrome P450s are not uncommon in presence of an intestinal microflora in­ prokaryotes and it has been found to be duce and/or repress certain isoforms of present in a number of bacterial strains hepatic P450s (Nugon-Baudon et al., (Fulco, 1991; Nelson et al., 1996). The 1998). However, the mechanism(s) be­ huge enzymatic capacity of the intestinal hind these modulating effects of the in­ flora also indicate metabolic reactions testinal microflora are not well under­ similar to those carried out by stood. Additionally, it should be men­ mammalian P450s It is indeed well tioned that several microbial enzymes known that the intestinal microflora is might act upon metabolites formed able to carry out enzymatic reactions during phase II reactions. Indeed, de­ involved in phase II, and previous in­ conjugation of bile acids (Midtvedt, vestigations also indicate that the flora 1974) and steroid are solely a bacterial may (Bakke and Midtvedt, 1970) – but event and so is nearly also deconjuga­ not always (Borud et al., 1971, 1973), tion of glucuronides (Roed and play a role in phase I type of reactions. Midtvedt, 1977) and some drugs (Pep­ Comparative studies in germfree and percorn and Goldman, 1972)

P450s AND SPECIFIC MEMBERS OF THE INTESTINAL FLORA This possible enzyme-modulating ef- ing presence of P450s in 18 bacterial fect of the intestinal microflora was the strains, selected among the major group background for a resent study concern- of species known to be present in the

54 human intestinal microflora (John et al., should be taken into consideration is 2001). As summarised by the authors, that a long-term intake of a probiotic in a “the amino acid identity, Southern blot child might influence upon the devel­ and CO difference spectrum data all opment of the normal spectrum of suggest the presence of a cytochrome P450s in the liver. To the best of my P450-like gene in Eubacterium aerofa­ knowledge, investigations along these ciens”. They claim that Eubacterium lines are not included in any of the long­ might be found in the human intestine in term studies of intake of probiotics in a density of more than log 10 organisms cohorts of children. Additionally, in per ml of content, and that their find­ groups of individuals with an increased ings, “demonstrating the presence of number of microbes in the small intes­ cytochrome P450 or P450-like proteins tine (elderly people, patients using ant­ in microflora will help us to gain a better acida, etc.), a microbial metabolism of understanding of the role specific drugs might easily take place before microbes, like E. aerofaciens, may play they are absorbed. in metabolising xenobiotics. In addition, On the other hand, however, by ac­ it will allow us to understand its cepting that microbes may have the pos­ influence on hepatic cytochrome P450 sibility of exerting these types of enzy­ expression and its overall association matic reactions it should be a future goal with tumour suppression and/or forma­ to select specific bacterial strains with tion.” specific action(s) on the (pro)drug Surely, work like this should be ex­ given, thereby creating an increased lo­ tended. It goes without saying that if cal concentration of active drug. A more probiotics express P450 activity, they distant goal might be to profiling the may influence upon the metabolism – levels of P450 iso-enzymes in the liver. and efficacy - of several drugs. It is in­ The answers of the initial questions are: deed easy to predict that intake of some Microbial P450 does exist, but we do probiotics may influence upon the effi­ not know what it really means. cacy of contraceptives. A scenario that

LITERATURE

Borud, O., Midtvedt, I., and Gjessing L.R.: netic differences in drug oxidation. Trends Urinary phenolic compounds in gnotobiotic Genet. 6, 182-186 (1990). and conventional rats on a free Diet, and be­ John, G.H., Walls, S, Keith, R. Goodfox- fore and after L-DOPA loading on a milk Jones, J., Tucker, K., and Abraham, K.J.: diet. Acta Pharmacol. Toxicol. 31, 540-549 The presence of a cytochrome P450-like (1971). protein in the human intestinal flora Eubac­ Borud, O., Midtvedt, T., and Gjessing, L.R.: terium aerofaciens. Microb. Ecol. Health Phenolic metabolites in urine and faeces Dis. 13, 3-8 (2001). from rats given radioactive 14C-L-DOPA. Kolars, J.C., Lown, K.S., Schmiedlin-Ren, P., Acta Pharmacol. Toxicol. 33, 308-316 Ghosh, M., Fang, C., Wrighton, S.A., (1973). Marion, R.M., and Watkins, P.B.: CYP3A Fulco, A.J.: P450Bm-3 and other inducible gene expression in human gut epihelium. bacterial P450 cytochromes: Biochemistry Pharmacokinetics 4, 247-259 (1994). and regulation. Ann. Rev. Pharmacol. Krishna, D.R. and Klotz, U.: Extrahepatic me­ Toxicol. 31,177-203 (1991). tabolism of drugs in humans. Clin. Phar­ Gonzalez, F.J. and Nebert, D.W.: Evolution of macokin. 26, 144-160 (1994). the P450 gene superfamily: Animal-plant McLellan, R.A.: Interindiviual differences in “warfare”, molecular drive and human ge­ xenobiotic-metabolising enzymes: The hu­

55 man genetic factor. Thesis, Karolinska In­ Nelson, D.R. and Strobel, H.W.: Evolution of stitute, Stockholm, Sweden, 2000. cytochrome P-450 proteins. Mol. Biol. Midtvedt, T.: Microbial bile acid transforma­ Evol. 4, 572-593 (1987). tion. Amer. J. Clin. Nutr. 27, 1341-1347 Nugon-Baudon, L., Robot, S., Flinois, J.P., (1974). Lory, S., Beaune, P.: Effects of the bacte­ Nebert, D.W., KcKinnon, R.A., and Puga, A.: rial status of rats on the changes in some Human drug-metabolizing enzyme poly­ liver cytochrome P450 (EC 1.14.14.1) morphisms: Effects on risk of toxicity and apoprotein consequent to a glucosinolate­ cancer. DNA Cell. Biol. 15, 273-280 rich diet. Brit. J. Nutr. 80, 231-234 (1998). (1996). Peppercorn, M.A. and Goldman, P.: The role Nebert, D.W.: Polymorphisms in drug-metabo­ of intestinal bacteria in the metabolism of lizing enzymes: What is their clinical rele­ salicylazosulfapyridine. J. Pharmacol. Exp. vance and why do they exist? Am. J. Hum. Ther. 18, 55-562 (1972). Genet. 60, 265-271 (1997). Roed, T.O. and Midtvedt, T.: Origin of intesti­ Nelson, D.R., Koymans, L., Kamataki, T., nal beta-glucuronidase in germfree, mono­ Stegeman, J.J., Feyereisen, R., Waxman, contaminated and conventional rats. Acta D.J., Waterman, M.R., Gotoh, O., Coon, Path. Microbiol. Scand. (B) 85, 271-276 M.J., Estabrook, R.W., Gunsals, I.C., and (1977). Nebert, D.W.: P450 superfamily: Update Schuler, M.A.: The role of cytochrome P450 on new sequences, gene mapping, accession mono-oxygenases in plant-insect interacts. numbers and nomenclature. Pharmacogenet­ Plant Physiol. 112, 1411-1419. ics 6, 1-42 (1996).

56 A LINK BETWEEN MUCOSAL REGULATORY LYMPHOCYTES AND CHILDHOOD FOOD ALLERGY SIMON H. MURCH Royal Free and University College School of Medicine, London, UK

INTRODUCTION There has been important recent ad­ response to sensitising antigen, rather vance in understanding basic concepts than as a prime mediator of sensitisation of intestinal food allergy, and the role of itself (Walker-Smith and Murch, 1999). infectious challenge in the prevention of IgE responses to dietary antigens do allergy. There has also been increasing occur in children in the tropics, but appreciation of the role of non-IgE-me­ without consequent disease in most diated pathology, and the basic concepts children. In allergic children of the de­ of food allergy have broadened as the veloped world, children may manifest mechanisms of oral tolerance have been immediate and obvious reactions or a unravelled. It has been suggested that complex of delayed symptoms including the traditional emphasis on IgE-medi­ diet-responsive eczema and a marked ated allergy has become less appropri­ disturbance of intestinal motility ate, as evidence mounts that the role of (Murch, 2000). IgE may be one of modulation of the

THE INCREASE IN THE INCIDENCE OF FOOD ALLERGIES Food allergies are not alone in sensitisation, there has been recognition showing increased incidence. It is well­ of novel patterns of food allergic disease recognised that there has been substan­ in children. Marked increase in the tial increase in incidence of all types of incidence of food allergies of all kinds childhood allergy. It is not simply a has occurred. In addition, multiple food matter of incidence. Previously rare al­ allergies and sensitisation of exclusively lergies, such as to peanuts, have be­ breast-fed infants to maternal dietary come common (Ewan, 1996; Houri­ antigens have become commonplace hane, 1997). Thus, in addition to ad­ (Walker-Smith and Murch, 1999; vance in the scientific basis of allergic Murch, 2000).

DOES IgE OR IgA DETERMINE SENSITISATION? Genetic predisposition is clearly im- the lower quartile of the normal range portant in allergy, and candidate genes was more strongly predictive of allergic for allergy susceptibility have been disease than elevated IgE (Ludviksson, identified, most concerned with IgE 1993). This concords with Soothill’s generation (Cookson, 1999). A popula- early report (Soothill, 1976) of transient tion study from Iceland however dem- IgA deficiency of infancy in the patho­ onstrated that an IgA concentration in genesis of allergic sensitisation. It is

57 probably not coincidental that the cyto­ infants born in the developing world kine most centrally involved in isotype compared to developed world infants shift to IgA is transforming growth (El Seed and Dafallah, 1983). There is factor-β (TGF-β), a molecule now rec­ as yet little study of specific placental ognised as central in oral tolerance mechanisms regulating cord blood IgA, mechanisms. but breast milk cytokine concentrations The role of infectious exposures in may differ in atopic and non-atopic the generation of TGF-β responses will mothers (Jones and Warner, 2000). It be discussed later. However it is strik­ will be intriguing to compare breast milk ing that early-life IgA concentrations, cytokines in developing and developed­ including in cord blood, are elevated in world mothers.

GENETIC PREDISPOSITION TO FOOD ALLERGY While a history of other atopic dis­ Several susceptibility regions are shared eases is common in food allergic indi­ with chronic inflammatory bowel dis­ viduals, the major genetic studies have ease, where tolerance is lost to the en­ so far been only been carried out in teric flora rather than dietary antigen. classic atopy, and thus may not be a true Epidemiological studies suggest a representation of susceptibility for food role for T cell responses and MHC type allergy itself. There are regions on in food allergy, with regional variation chromosomes 2q, 5q, 6q, 12q and 13q in patterns of sensitisation despite that are consistently linked with atopic broadly similar antigen exposures (Hill disorders, with candidate genes includ­ et al., 1999). Although peanut hyper­ ing the IL-1 cluster, the IL-4/IL-9/IL-13 sensitivity is common in Indonesia, it is cluster, the Major Histocompatibility uncommon in Malaysia, Japan and the Complex and the interferon-γ gene Philippines. (Rosenwasser, 1997; Cookson, 1999).

REGULATION OF IgE RESPONSES

While non-IgE-mediated food allergy γ and IL-2) limit IgE production, as do may be the most frequent cause of Th1-associated cytokines such as IL-12 chronic symptoms, IgE-mediated and Il-18. By contrast, Th2 cytokines, mechanisms account for the majority of particularly IL-4 and IL-13, directly immediate hypersensitive reactions to promote IgE synthesis. This Th1 effect foods. Transient IgE responses to foods may partly explain the protection against are found in many normal children and allergy provided by childhood within these may thus not be clinically relevant the developing world, but do not en­ (Sigurs et al., 1994). However exag­ compass the role of Th2 responses gerated IgE responses are clearly im­ against helminths (Yazdanbakhsh et al., portant in severe food allergies and ana­ 2002). phylaxis. Class-switching to IgE in response Isotype shift to IgE is regulated by to IL-4 and IL-13, whose receptors products of Th1 and Th2 T cells (re­ share a common α chain (IL-4Rα), is viewed by Corry and Kheradmand, mediated by a signal cascade involving 1999). Th1 cytokines (particularly IFN­ Stat-6 (signal transducer and activator of

58 transcription-6), and gain of function transportation of mucosally-produced mutations in this pathway are associated IgE into the gut lumen or airway by a with both murine and human allergic IL-4-dependent mechanism distinct sensitisation (Shimoda et al., 1996; from the poly-immunoglobulin trans­ Hershey et al., 1997). There is evidence porter that mediates IgA secretion of compartmentalised IgE responses (Ramaswamy et al., 1994). within both the intestine and lung, with

MULTIPLE FOOD INTOLERANCE, FOOD-ALLERGIC DYSMOTILITY AND THE EOSINOPHIL RESPONSE In addition to clear increase in IgE­ establish basic oral tolerance mecha­ mediated and non-IgE-mediated food nisms, rather than the loss of previously allergies, a remarkable alteration in dis­ acquired tolerance of classic food al­ ease presentation has been noted in sev­ lergy. eral countries, where increasing num­ There is now increasing evidence bers of infants are now sensitising to that both gastro-oesophageal reflux and multiple antigens despite exclusive constipation may be features of the breastfeeding, often within the first food-allergic dysmotility syndrome, weeks of life (Hill et al., 1999; Walker- which is characterised by local eosino­ Smith and Murch, 1999; Murch, 2000). philic infiltration. Epithelial expression This was rare a generation ago and is of the eosinophil chemokine eotaxin ap­ still almost unknown in the developing pears to distinguish allergy-associated world. These children represent a major gastro-oesophageal reflux from primary clinical challenge, and extensive dietary mechanical reflux in infants (Butt, exclusions are often required. There is a 2002). Eotaxin-deficient mice are pro­ clear association with eczema, food-al­ tected from the dysmotility associated lergic colitis or enteropathy, and these with mucosal allergy, suggesting that infants demonstrate a prominent disrup­ this is an important local response (Ho­ tion of intestinal motility. This pattern of gan et al., 2001). disease suggests a primary failure to

DEMOGRAPHICS OF ALLERGIC SENSITISATION: THE ROLE OF ENTERIC CHALLENGES Improvement in social conditions in exposure to food-borne and oro-faecal an individual country appears to cause pathogens in childhood was associated rapid increase in childhood allergies. with a reduced risk of allergic sensitisa­ Thus former East Germany, Estonia and tion (Matricardi et al., 2000). Both rural Singapore have seen an increased upbringing and exposure to animals of­ incidence of allergic diseases of all fers protection against later allergy in kinds (Goh et al., 1996; von Mutius et both developed-world and developing­ al., 1998). The particular exposures that world children (Braun-Farländer et al., reduce risk of sensitisation appear to be 1999; Lewis, 2000). gastro-enterological rather than respira­ There may thus be an important role tory. Serology performed in Italian for early environmental exposures in the military recruits demonstrated that past determination of immune tolerance,

59 modulating the effects of genetic pre­ for the gut flora and probably also a disposition. Interest now centres on the maturational role for bacterial pathogens specific links between the innate im­ in the establishment of immune toler­ mune system and bacterial exposures in ance (Fearon and Locksley, 1996; Sudo early life (Table 1). There is increasing et al., 1997; Rook and Stanford, 1998; evidence to suggest an obligatory role Sebra, 1999).

MUCOSAL CHALLENGES IN CHILDREN BORN IN DEVELOPING COUNTRIES Most children born within the tropics (Lunn et al., 1991). Over a 1-year pe­ have evidence of enteropathy, and mu­ riod, increased L:M ratio accounted for cosal biopsies would usually be consid­ 40% of growth faltering in Gambian ered abnormal by UK standards. Our children. Infection alone accounts for a study of regulatory lymphocyte re­ minority of cases: Bacterial pathogens sponses in the mucosa of developing were isolated in <12% in one Keneba world children has been based on biop­ study and viruses detected more fre­ sies obtained at the MRC Unit at quently in non-diarrhoeal controls Keneba in the Gambia. Gambian chil­ (Rowland et al., 1978). Small bowel dren show a pattern of growth faltering bacterial overgrowth and Giardia lam­ typical of deprived areas of the devel­ blia infection are found in >80% of rural oping world, with UK normal growth Gambian infants, but neither correlate velocity for the first 4 months, prior to with growth or gut permeability (Lunn weaning, followed by decline against et al., 1999). UK centiles. At age 2 the mean weight­ It is noteworthy that infant mortality for-age lies 2SD below UK standards rates in Gambia 2002 (c. 100/1000) are (z-score –2). Previous studies from similar to those that were seen in Lon­ Keneba confirmed biochemical and don, Paris or New York 1902, where dietary deficiencies in these infants. De­ gastro-enteritis and wasting was also the spite massive dietary supplementation major cause of infant death in un­ (twice recommended values for energy, derprivileged children. There is clear 1 2 /2 times for protein), there was some evidence that polymorphisms in cyto­ short-term catch-up growth in mal­ kine response genes may give survival nourished children following gastro­ advantage in tropical children, and that enteritis, which reversed as soon as the these vary from country to country (e.g. child was discharged (Rowland et al., high TNF producers do better against 1981; Sullivan et al., 1992). Because of intracellular pathogens but have higher the failure of dietary intervention to re­ mortality from cerebral malaria). It is store growth, other factors have been likely that similar selection pressures studied. The most important is small will have existed a century ago in Euro­ bowel enteropathy, with particular evi­ pean children, and that these may play a dence of a role for excess paracellular role in the development of atopy as in­ permeability on lactulose:mannitol fectious challenge decreases. (L:M) dual sugar permeability testing

60 Table 1: Some potential interactions between innate immunity and the gut flora (After: Murch, 2001) ——————————————————————————————————————— Recognition Distribution Bacterial Effect transduced element component ——————————————————————————————————————— Mannose Dendritic cells, High-mannose Increased efficiency of receptor macrophages, B cells carbohydrates antigen presentation —————————————————————————————————————————————————————————————————————————————————————————————————— Natural Secreted by peritoneal Surface glycans Modulation of T cell Antibody and intestinal B-1 cells activation —————————————————————————————————————————————————————————————————————————————————————————————————— Complement Synergy with natural O- and N-linked Opsonisation. Also Antibody glycans regulates T cell activation and B cell tolerance —————————————————————————————————————————————————————————————————————————————————————————————————— Toll-like Dendritic cells, TLR2 - NF-κB signaling pathway Receptors macrophages, T cells, peptidoglycans Enterocytes Increased surface TLR4 -LPS expression of Class II TLR’s 1-10 MHC and co-stimulatory identified, most TLR9 - ligands with as yet Unmethylated CpG undetermined repeats in bacterial ligands DNA —————————————————————————————————————————————————————————————————————————————————————————————————— Mannan- Serum-derived. Binds to Carbohydrates on Activates complement binding lectin macrophages, monocytes Gram-negative and directly via serine and B cells Gram positive bacteria proteases MASP-1 and MASP-2 —————————————————————————————————————————————————————————————————————————————————————————————————— Nod receptors Intracellular recognition Bacterial LPS NF-κB signaling molecules in innate pathway immune cells —————————————————————————————————————————————————————————————————————————————————————————————————— Vα24 NK Epithelial lymphocyte Conserved glycolipid Modulate enterocyte T cells, subsets. Invariant T cell sequences, presented responses, polarise Vδ1 γδ T cells receptor chains by non-classical towards Th1 and MHC (CD1d) mucosal IgA production ———————————————————————————————————————

CONTRASTING CHANGES IN EARLY-LIFE GUT FLORA IN THE DEVELOPED-WORLD CHILD Lack of appropriate early infectious developed-world neonate has altered exposure has been postulated for many dramatically compared to infants born in years as a cause of the overall increase the developing world, with reduced in allergies (Rook and Stanford, 1999). colonisation by previously dominant These changes may be occurring from species such as Bifidobacteria and fre­ very early in life. There is evidence that quent discordance between the flora of the initial intestinal colonisation of the the mother and her child (Grutte and

61 Muller-Beuthow, 1979; Simhon et al., (Montgomery et al., 2000). In addition, 1982). These changes were not found in early analysis suggests that this is a risk infants born in Nigeria, when compared factor for inflammatory bowel disease, to London-born infants (Simhon et al., but not diabetes mellitus (unpublished 1982). Infants born by caesarean data). Further support that early-life section show prolonged abnormality in colonisation is a determinant of later composition of the intestinal flora and sensitisation has been provided by distinct alterations in immune function studies of gut flora in Estonian and (Gronlund et al., 1999a,b). Swedish children, and allergic children We examined the hypothesis that from either country showed reduced changes in infant handling practices at lactobacilli and anaerobes but higher the time of initial gut colonisation may numbers of coliforms and Staphylococ­ be important, by study of the 1970 UK cus aureus (Björksten et al., 1999, national birth cohort, where every infant 2001). These changes are detectable born in one week in April 1970 has very early in life, before the develop­ been followed long-term. Those infants ment of clinical allergies. who spent the first night away from Neonatal administration of probiotics their mother in the communal nursery to infants at risk of later allergies in­ had a significantly increased incidence duced a remarkable reduction in later of hayfever at age 26, suggesting that eczema (Kalliomaki et al., 2001). How­ increased exposures to non-familial mi­ ever it was notable that systemic IgE croorganisms or reduced colonisation responses were unaffected, arguing for by family microorganisms, may be as­ compartmentalisation of mucosal and sociated with later allergic disease cutaneous responses (Murch, 2001).

THE DEVELOPMENT OF ENTERIC TOLERANCE Food allergy requires breakdown of dietary antigen by activated antigen pre­ oral tolerance, in which systemic im­ senting cells, as seen in infant sensitisa­ munological tolerance to an antigen is tion to cow’s milk formula following induced by its ingestion. The molecular gastro-enteritis. High dose dietary anti- mechanisms of oral tolerance to dietary gen induces apoptosis of antigen-spe­ antigens have recently been partially cific lymphocytes within Peyer’s elucidated. The dose of ingested antigen patches in mice, but this has not yet is important in determining the means been demonstrated in man (Chen et al., by which tolerance is maintained 1995). By contrast, tolerance to low­ (Weiner, 1997; Strobel and Mowat, dose antigen is an activation-dependent 1998; Strober and Kelsall, 1998). Tol­ process, where antigen-specific regula­ erance for high doses of antigen occurs tory lymphocytes producing trans­ through anergy of potentially responsive forming growth factor-β (TGF-β) are T cells. This may reflect antigen generated (TH3 cells). This activation­ presentation by gut epithelium in the dependent response to low doses may absence of co-stimulatory ligands, or be more difficult to establish in infancy active suppression of the lymphocytes than high dose tolerance, which is me­ by suppressor cells or regulatory cyto­ diated by T cell anergy, as neonatal kines (Mayer, 2000). These pathways lymphocytes are relatively difficult to may be abrogated by breakdown of the activate. Indeed, oral administration of epithelial barrier and presentation of low-dose myelin basic protein sensitised

62 neonatal rats, while similar amounts in­ cosa and release TGF-β or IL-10 upon duced protective tolerance in adults encountering antigen, thus suppress (Miller et al., 1994). potential reactivity of all surrounding Both TH3 cells and IL-10 producing lymphocytes (Groux and Powrie, lymphocytes (Tr1 cells) suppress im­ 1999). If this process breaks down, mune reactivity within the intestine by a immunological tolerance may be lost process termed “bystander tolerance“, in and allergy or gut inflammation the con­ which they home to the intestinal mu­ sequence.

MECHANISMS OF ORAL TOLERANCE: A CENTRAL ROLE FOR NF-κκB The final pathway in the generation vents the normal establishment and of tolorogenic lymphocytes appears to maintenance of tolerance to dietary anti- be shared with Th1 responses, with the gen. In study of transgenic mice, whose nuclear transcription factor NF-κB in a only T cells recognised a class II MHC­ pivotal role. Impaired NF-κB responses restricted peptide in hen egg lysozyme, are clearly linked to mucosal sensitisa­ Newberry and colleagues (1999) dem­ tion, to both the intestinal flora and to onstrated an obligatory role for physio­ dietary antigen in experimental models, logical inflammation in the establish­ and indeed human IBD. ment of oral tolerance. These mice could Spontaneous IBD develops in re­ only respond to their luminal flora sponse to the normal flora in CE3Hej through innate immune cells, but were mice, who are genetically deficient in fully tolerant to dietary hen egg the LPD-sensing Toll-like receptor Tlr4 lysozyme in normal conditions. When (Poltorak et al., 1998). Targeted dele­ mucosal production of prostaglandin E2 tion of NF-κB sub-units in murine (PgE2) was prevented by cyclo-oxy­ knockouts also induces intestinal in­ genase 2 (COX-2) antagonists, toler­ flammation. In man, the primary genetic ance was abrogated and enteropathy de­ association in Crohn’s disease is a loss­ veloped in response to antigen inges­ of-function mutation in Nod2, an intra­ tion. Thus PgE2 appears to play a fun­ cellular LPS sensor which induces a damental role prevention of immune re­ NF-κB response to bacterial responses activity to dietary antigens. Constitutive (Ogura et al., 2001). More profound production of PgE2 occurs in lamina multi-system inflammation occurs in propria macrophages in response to the children with mutations in the NEMO enteric flora, and it functions as a potent molecule (IKKγ), which is an important inducer of IL-10 production in lympho­ regulator of NF-κB function (Courtois cytes (Newberry et al., 1999). In turn, et al., 2001). Thus the paradox is seen IL-10 is critical in the generation of that a sub-optimal inflammatory re­ regulatory lymphocytes, probably sponse to the normal flora leads to an through facilitating generation of TGF-β exaggerated pro-inflammatory response. producing cells (Groux and Powrie, Similar mechanisms appear to apply 1999). There is also clear evidence that in the generation of tolerance to dietary mucosal inflammation induces a com­ antigen. Again, blockade of a suffi­ pensatory TGF-β response (Xian et al., ciently pro-inflammatory response pre­ 1999).

63 Figure. Mean density of lymphocyte populations in the duodenal lamina propria of well-grown UK and Gambian infants. (Data from Campbell et al., 2003).

IS FOOD ALLERGY RELATED TO DEFECTIVE GENERATION OF REGULATORY LYMPHOCYTES?

The data presented above argue for a TGF-β+ lymphocytes in the Gambian specific role of early infectious expo­ children was also some 10 fold higher sures in the generation of enteric toler­ than in the UK controls. We noted pro­ ance, mediated specifically through gressive reduction of TGF-β producing regulatory lymphocytes. They would lymphocytes with worsening nutritional suggest that children brought up in un­ status in these children, while Th1 re­ derdeveloped countries with low allergy sponses were relatively maintained. would have higher numbers of regula­ In studies of UK children with food tory lymphocytes. In addition, those allergies, using flow cytometry, immu­ children within privileged countries who nohistochemistry and in situ hybridisa­ develop allergies may be less efficient in tion, we found the dominant abnormal­ generating regulatory lymphocytes than ity to be failure to generate mucosa those who do not. Our data suggest that TGF-β producing lymphocytes, rather this might be so. than simple deviation of Th1/Th2 re­ In our immunohistochemical analysis sponses (Pérez-Machado et al., 2000). of Gambian children, the mucosal den­ These data thus support early con­ sity of CD25+ lymphocytes was 50-200 tention (Murch, 1996) that impaired times that seen in UK normal controls generation of regulatory lymphocytes (Campbell et al., 2002). The density of might underlie the increase in food al­ IFN-γ and TNF-α expressing mononu­ lergies within the developed world. All clear cells was approximately 10 times the available evidence points toward a that see in UK infants (Figure 1). Im­ critical role for the gut flora in this proc­ portantly, the density of IL-10+ and ess.

64 LITERATURE

Björksten, B., Naaber, P., Sepp, E., and tures and associations. BMJ 312, 1074­ Mikelsaar, M.: The intestinal microflora in 1078 (1996). allergic Estonian and Swedish 2-year old Fearon, D.T. and Locksley, R.M.: The instruc­ children. Clin. Exp. Allergy 29, 342-346 tive role of innate immunity in the acquired (1999). immune response. Science 272, 50-54 Björksten, B., Sepp, E., Julge, K., Voor, T., (1996). and Mikelsaar, M.: Allergy development Goh, D.Y., Chew, F.T., Quek, S.C., and Lee, and the intestinal microflora during the first B.W.: Prevalence and severity of asthma, year of life. J. Allergy Clin. Immunol. rhinitis, and eczema in Singapore school­ 108, 516-520 (2001). children. Arch. Dis. Child. 74, 131-135 Braun-Fahrlander, C., Gassner, M., Grize, L., (1996). Neu, U., Sennhauser, F.H., Varonier, H.S., Gronlund, M.M., Lehtonen, O.P., Eerola, E., Vuille, J.C., and Wuthrich, B.: Prevalence and Kero, P.: Fecal microflora in healthy of hay fever and allergic sensitisation in infants born by different methods of deliv­ farmer’s children and their peers living in ery: Permanent changes in intestinal flora the same rural community. Clin. Exp. Al­ after cesarean section. J. Pediatr. Gastroen­ lergy 29, 28-34 (1999). terol. Nutr. 28, 19-25 (1999a). Butt, A.M., Murch, S.H., Ng, C.L., Kitching, Gronlund, M.M., Nuutila, J., Pelto, L., Lilius, P., Montgomery, S.M., Phillips, A.D., E.M., Isolauri, E., Salminen, S., Kero, P., Walker-Smith, J.A., and Thomson, M.A.: and Lehtonen, O.P.: Mode of delivery Upregulated eotaxin expression and T cell affects the phagocyte functions of infants infiltration in the basal oesophageal epithe­ for the first 6 months of life. Clin. Exp. lium in cow’s milk-associated reflux oe­ Immunol. 116, 521-526 (1999b). sophagitis. Arch. Dis. Child. 87, 124-130 Groux, H. and Powrie, F.: Regulatory T cells 2002. and inflammatory bowel disease. Immunol. Campbell, D.I., Murch, S.H., Lunn, P.G., Today 20, 442-446 (1999). Elia, M., Sullivan, P.B., Sanyang, M.S., Grutte, F.K., Muller-Beuthow, W.: Wandlung and Jobarteh, B.: Chronic T cell-mediated der normaken darmflora des menschlichen enteropathy in rural West African children: Sauglings innerhalb der letzen 20 Jahre. Relationship with nutritional status and Nahrung 23, 455-465 (1979). small bowel function. Pediatr. Res. 2003 Hershey, G.K., Friedrich, M.F., Esswein, (in press). L.A., Thomas, M.L., and Chatila, T.A.: Chen, Y., Inobe, J., Marks, R., Gonnella, P., The association of atopy with a gain-of­ Kuchroo, V.K., and Weiner, H.L.: Periph­ function mutation in the α subunit of the eral deletion of antigen-reactive T cells in interleukin-4 receptors. New Engl. J. Med. oral tolerance. Nature 376, 177-180 (1995). 337, 1720-1725 (1997). Cookson, W.: The alliance of genes and envi­ Hill, D.J., Hosking, C.S., and Heine, R.G.: ronment in asthma and allergy. Nature 402 Clinical spectrum of food allergy in chil­ (suppl.), B5-B11 (1999). dren in Australia and South-East Asia: Iden­ Corry, D.B. and Kheradmand, F.: Induction and tification and targets for treatment. Ann. regulation of the IgE response. Nature 402 Med. 31, 272-281 (1999). (suppl.), B18-23 (1999). Hogan, S.P., Mishra, A., Brandt, E.B., Roy­ Courtois, G., Smahi, A., Israël, A.: alty, M.P., Pope, S.M., Zimmermann, N., NEMO/IKKγ: linking NF-κB to human Foster, P.S., and Rothenberg, M.E.: A disease. Trends Molec. Med. 7, 427-430 pathological function for eotaxin and eosi­ (2001). nophils in eosinophilic gastrointestinal in­ El Seed, A.M. and Dafallah, A.A.: Serum im­ flammation. Nature Immunol. 2, 353-360 munoglobulin levels in normal Sudanese (2001). children. Ann. Trop. Paediatr. 3, 97-99 Hourihane, J.O.: Peanut allergy – current status (1983). and future challenges. Clin. Exp. Allergy Ewan, P.: Clinical study of peanut and nut 27, 1240-1246 (1997). allergy in 62 consecutive patients: New fea­ Jones, C.A. and Warner, J.O.: Breast milk as

65 an important source of cytokines for the R.G.: Cycloxygenase-2-dependent arachi­ offspring. Clin. Exp. Allergy 30, 599-601 donic acid metabolites are essential modula­ (2000). tors of the immune response to dietary an­ Kalliomaki, M., Salminen, S., Arvillomi, H., tigen. Nature Medicine 5, 900-906 (1999). Kero, P., Koskinen, P., and Isolauri, E.: Ogura, Y., Bonen, D.K., Inohara, N., Nicolae, Probiotics in primary prevention of atopic D.L., Chen, F.F., Ramos, R., Britton, H., disease: A randomised placebo-controlled Moran, T., Karaliuskas, R., Duerr, R.H., trial. Lancet 357, 1076-1079 (2001). Achkar, J.P., Brant, S.R., Bayless, T.M., Lewis, S.A.: Animals and allergy. Clin. Exp. Kirschner, B.S., Hanauer, S.B., Nunez, G., Allergy 30, 153-157 (2000). and Cho, J.H.: A frameshift mutation in Ludviksson, B.R., Elriksson, T.H., Ardal, B., Nod2 associated with susceptibility to Sigfusson, A., and Valdimarsson, H.: Cor­ Crohn’s disease. Nature 411, 603-606 relation between serum immunoglobulin A (2001). concentrations and allergic manifestations Pérez-Machado, M., Ashwood, P., Sim, R., in infants. J. Pediatr. 121, 23-27 (1992). Thomson, M.A., Walker-Smith, J.A., and Lunn, P.G., Northrop-Clewes, C.A., and Murch, S.H.: Evidence at protein and Downes, R.M.: Intestinal permeability, mRNA level of reduced TH3 lymphocytes in mucosal injury, and growth faltering in multiple food allergy (abstract). JPGN 31, Gambian infants. Lancet 338, 907-910 S127-S128 (2000). (1991). Poltorak, A., He, X., Smirnova, I., Liu, M.Y., Lunn, P.G., Erinoso, H.O., Northrop-Clewes, Van Huffel, C., Du, X., Birdwell, D., C.A., and Boyce, S.A.: Giardia intestinalis Alejos, E., Silva, M., Galanos, C., is unlikely to be a major cause of the poor Freudenberg, M., Ricciardi-Castagnoli, P., growth of rural Gambian infants. J. Nutr. Layton, B., and Beutler, B.: Defective LPS 129, 872-877 (1999). signalling in C3H/HeJ and C57BL/10ScCr Matricardi, P.M., Rosmini, F., Riondino, S., mice: Mutations in Tlr4 gene. Science 282, Fortini, M., Ferrigno, L., Rapicetta, M., 2085-2088 (1998). Bonini, S.: Exposure to foodborne and oro­ Ramaswamy, K., Hakimi, J., and Bell, R.G.: fecal microbes versus airborne viruses in re­ Evidence for an interleukin-4 inducible im­ lation to atopy and allergic asthma. BMJ munoglobulin E uptake and transport 320, 412-417 (2000). mechanism in the intestine. J. Exp. Med. Mayer, L.: Mucosal immunity and gastrointes­ 180, 1793-1803 (1994). tinal antigen processing. J. Pediatr. Gastro­ Rosenwasser, L.J.: Genetics of atopy and enterol. Nutr. 30, S4-S12 (2000). asthma: Promoter-based candidate gene stud­ Miller, A., Lider, O., Abramsky, O., and ies for IL-4. Int. Arch. Allergy Immunol. Weiner, H.L.: Orally administered myelin 113, 61-64 (1997). basic protein in neonates primes for im­ Rook, G.A.W. and Stanford, J.L.: Give us this mune responses and enhances experimental day our daily germs. Immunol. Today 19, autoimmune encephalomyelitis in adult 113-116 (1998). animals. Eur. J. Immunol. 24: 1026-1032 Rowland, M.G., Davies, H., Patterson, S., (1994). Dourmashkin, R.R., Tyrrell, D.A., Mat­ Montgomery, S.M., Wakefield, A.J., Morris, thews, T.H., Parry, J., Hall, J., and Larson, D.L., Pounder, R.E., and Murch, S.H.: The H.E.: Viruses and diarrhoea in West Africa initial care of newborn infants and subse­ and London: A collaborative study. Trans­ quent hayfever. Allergy 55, 916-922 act. Royal Soc. Med. Hyg. 72, 95-98 (2000). (1978). Murch, S.: Diabetes and cows’ milk. Lancet Rowland, M.G.M., Cole, T.J., and McCollum, 348, 1656 (1996). J.P.K.: Weanling diarrhoea in The Gambia: Murch, S.H.: The immunologic basis for in­ Implications of a jejunal intubation study. testinal food allergy. Curr. Opin. Gastroen­ Transact. Royal Soc. Med. Hyg. 75, 215­ terol. 16, 552-557 (2000). 218 (1981). Murch, S.H.: Toll of allergy reduced by probi­ Sebra, J.J.: Influences of microbiota on intes­ otics. Lancet 357, 1057-1059 (2001). tinal immune system development. Newberry, R.D., Stenson, W.F., and Lorenz, Am. J. Clin. Nutr. 69 (suppl), 1046S­

66 1051S (1999). velopment of an IgE production system Shimoda, K., van Deursen, J., Sangster, M.Y., susceptible to oral tolerance induction. J. Sarawar, S.R., Carson, R.T., Tripp, R.A., Immunol. 159, 1739-1745 (1997). Chu, C., Quelle, F.W., Nosaka, T., Sullivan, P.B., Lunn, P.G., Northrop-Clewes, Vignali, D.A., Doherty, P.C., Grosveld, C., Crowe, P.T., Marsh, M.N., and Neale, G., Paul, W.E., and Ihle, J.N.: Lack of Il­ G.: Persistent diarrhea and malnutrition ­ 4-induced T cell response and IgE class the impact of treatment on small bowel switching in mice with disrupted Stat6 structure and permeability. J. Pediatr. Gas­ gene. Nature 380, 630-633 (1996). troenterol. Nutr. 14, 208-215 (1992). Sigurs, N., Hattevig, G., Kjellman, B., Kjell­ von Mutius, E., Weiland, S.K., Fritzsch, C., man, N.I., Nilsson, L., Björksten, B.: Ap­ Duhme, H., and Keil, U.: Increasing preva­ pearance of atopic disease in relation to se­ lence of hay fever and atopy among children rum IgE antibodies in children followed in Leipzig, East Germany. Lancet 351, from birth for 4 to 15 years. J. Allergy 862-866 (1998). Clin. Immunol. 94, 757-763 (1994). Walker-Smith, J.A. and Murch, S.H.: Gastro­ Simhon, A., Douglas, J.R., Drasar, B.S., and intestinal food allergy. In: Diseases of the Soothill, J.F.: Effect of feeding on infants’ Small Intestine in Childhood, 4th Edition. faecal flora. Arch. Dis. Child. 57, 54-58 Isis Medical Media, Oxford, 205-234 (1982). (1999). Soothill, J.F., Stokes, C.R., Turner, M.W., Weiner, H.L.: Oral tolerance: Immune mecha­ Norman, A.P., and Taylor, B.: Predispos­ nisms and treatment of autoimmune dis­ ing factors and the development of reaginic eases. Immunol. Today 18, 335-343 allergy in infancy. Clin. Allergy 6, 305­ (1997). 319 (1976). Xian, C.J., Xu, X., Mardell, C.E, Howarth, Strobel, S. and Mowat, A.M.: Immune re­ G.S., Byard, R.W., Moore, D.J., Mietti­ sponses to dietary antigens: Oral tolerance. nen, P., and Read, L.C.: Site-specific Immunol. Today 19, 173-181 (1998). changes in transforming growth factor-α Strober, W. and Kelsall, B.: To be responsive and -β1 expression in colonic mucosa of or not to be responsive, that is the mucosal adolescents with inflammatory bowel dis­ question. Gastroenterology 114, 214-217 ease. Scand. J. Gastroenterol. 34, 591-600 (1998). (1999). Sudo, N., Sawamura, S., Tanaka, K., Aiba, Yazdanbakhsh, M., Kremsner, P.G., and van Y., Kubo, C., and Koga, Y.: The require­ Ree, R.: Allergy, parasites, and the hygiene ment of intestinal bacterial flora for the de­ hypothesis. Science 296, 490-494 (2002).

67 68 SYNBIOTIC TREATMENT IN CLINICAL PRAXIS STIG BENGMARK Lund University, Ideon Research Park, Lund, Sweden

“Health and well-being is more than merely absence of disease” Mark Twain

INTRODUCTION Health and well-being seems de­ milk, compared to about 30% in free­ pending on availability of some more living, grass-eating and fresh-plant eat­ than two million different molecules, all ing cows. It is also known that treat­ available in the body in rather exact ment of foods with high temperature not amounts. Most of the molecules are only destroys important nutrients, par­ supplied by foods and made available in ticularly antioxidants, but also adds to the large intestine by fermentative ac­ the food cancer-promoting chemicals, tions. It is a considerable problem that mutagens. the variability in the food supplied both Several observations suggest that to domestic animals and humans has health and well-being is the result of a dramatically decreased in the modern dynamic interplay and balance - homeo­ society. Our Palaeolithic forefathers are stasis – between numerous processes said to each year have consumed food that control energy balance, appetite, from some four to five hundred plants. cell proliferation, repair systems, Modern man has reduced this to a few apoptosis, metabolic rate, stress re­ dozens, and furthermore, many impor­ sponse, immune response and numer­ tant molecules in foods are destroyed by ous other processes on which we are modern methods to store and prepare depending for our well-being (Frame et the food. The nutritional content of key al., 1998). The attention increasingly farming products such as meat and milk given to the homeostasis between has with modern farming methods omega-3 and omega-6, to the balance changed dramatically. As an example, between pro- and anti-inflammatory due to the limited variation of diet to cytokines, and balance in Th1 and Th2 domestic cows is the content of omega­ immune response serves as examples. 3 fatty acids only about 2% in beef and

FERMENTED FOOD HAS OUTSTANDING QUALITY Our Palaeolithic forefathers used modern technologies as it not only fermentation as their main method to maintains the content of important nutri­ prepare and store food. This method, ents, especially antioxidants, but also unfortunately today abandoned in de- sometimes increases it. Microbial en­ veloped countries, but still in use in zymes are known to release numerous most developing countries is superior to nutrients from fruit and vegetable fibres

69 Figure 1: Correlation between percentage of calories as fat in various countries and age-adjusted mortality in breast cancer. (Reproduced with permission from: Carroll, 1994). and make them accessible to the me­ badly tolerate the dramatic changes, es­ tabolism of the body. As a matter of fact pecially in food habits, which have oc­ the majority of the more than two mil­ curred, during the recent few hundred lion molecules constituting our body are years (Eaton and Konner, 1985), and products of microbial digestion in the that this could be an explanation to the lower GI tract. The complexity of func­ epidemic in chronic diseases, which has tions by the flora is illustrated by the occurred during the last few decades – fact the intestinal microbes together see further Bengmark (2001). contain more than 300,000 different It is clear that people, who live in ru­ genes, compared to the about 65,000 in ral areas of developing countries, and the rest of the human body. consume large amounts of fruits, vege­ The Palaeolithic food, the food to tables and live microbes – but also much which are genes have been adjusted less of animal fat have a much richer GI during several millions of years, is said flora, a better immune response, a to have contained only half as much of compared to Westerners reduced ability proteins, 1/4 as much of saturated fat to form blood clots, and a significantly and 1/10 of sodium salts. Instead it better resistance to disease. They also contained at least 4-5 times as much of rarely suffer the endemic diseases so plant fibres, 10 times as much of anti­ frequently observed in Western oxidants, fifty times as much of omega­ countries as demonstrated for breast 3 fatty acids, and billion times or more cancer in Figure 1, similar associations of microbes. It is reasonable to assume being demonstrated for several forms of that the human genes, adapted during cancer such as colonic cancer and million of years to the lifestyle and food prostatic cancer, prostatic hyperplasia, habits of our prehistoric ancestors, diabetes, coronary heart disease, neuro­

70 Table 1: Probiotics-claimed molecular effects —————————————————————————————————— General: Produces nutrients and antioxidants Produces growth and coagulation factors Activates the MALT system Modulates Th1/Th2 response Promotes antioxidant actions Controls potentially pathogenic microorganisms (PPMs) Reduces production of endotoxins Reduces mutagenicity Humoral: Stimulates IgA production Inhibits IgE production Stimulates NO production Modulates cytokine response Cellular: Stimulates macrophage function Stimulates NK cell activity Promotes growth and regeneration Promotes apoptosis —————————————————————————————————— degenerative diseases and other endemic by microbial fermentation and absorbed diseases. It is not unrealistic to suggest by the mucosa are important contribut­ that the numerous compounds released ing factors.

FLORA BOOSTS THE IMMUNE SYSTEM OF THE HOST It is increasingly observed that the is said that there are greater genetic dif­ GI tract is a key organ in our immune ferences between on LAB and another defence. Here up to 80% of the immune than between a fish and a human being. cells are to be found, and here up 80% LAB with the strongest ability to fer­ of the immunoglobulins are produced ment fibre are found on fruits and (Brandzaeg et al., 1989). Commensal vegetable fibres, often semi-resistant to flora, some powerful supplemented lac­ fermentation/digestion by microbial en­ tic acid bacteria (LAB), often referred to zymes and found in ethnic foods such as probiotics, and bioactive fibres from as sauerkraut and sourdough. As an ex­ fruit and vegetables, often called prebi­ ample, oligofructans such as inulin and otics, as well as their fermented prod­ phleins, fibres rich in several fruit and ucts, synbiotics, are known to have vegetables, but claimed to have strong immuno-modulatory, anti-infectious, biological effects, are difficult to fer­ anti-inflammatory and antioxidant ef­ ment and only a few LAB are able to do fects. Table 1 summarises some the ef­ so (Müller and Lier, 1994). Only 16 of fects on the immune system described in studied 712 LAB were able to ferment the literature. But not all LAB and all phlein-type fibre and only 8/712 inulin fibres are equally effective. The LAB in type fibre. Lactobacillus plantarum was various yoghurts are chosen for their clearly the most effective and only three palatability and have most often rather other LAB species, Lactobacillus para­ weak immuno-modulatory effects. LAB casei subsp. paracasei, Lactobacillus do not constitute an authentic genus, it brevis and Pediococcus pentosaceus

71 demonstrated ability to ferment these Lactobacillus paracasei subsp. paracasei relatively resistant fibres. Kruszewska 19, chosen among 355 human strains. et al. (2002) did in a recent study isolate Interestingly all the strains were able to no less than 180 microbial strains from transcribe NF-κB, to induce pro-in­ growing rye. Several of these demon­ flammatory cytokines (IL-1β and IL-8) strated strong bioactivity including and anti-inflammatory (IL-10) and to strong adhesion to human mucus, in­ produce antioxidants. In all these proc­ duction of pro- and anti-inflammatory esses did Lactobacillus plantarum 2592 cytokines and antioxidant activity. As show superior ability compared to the some of the isolated LAB showed supe­ others. These four LAB are together rior bioactivity did they choose to spe­ bioactive fibres (inulin, beta-glucan, cifically study the effects of Leu­ resistant starch and pectin) chosen to conostoc mesenteroides 77:1, Lactoba­ constitute a new symbiotic composition cillus plantarum 2592, and Pediococcus – Synbiotic 2000 – which presently is pentosaceus 16:1 – one from each genus under clinical evaluation (see further of the family of Lactobacillus – plus below).

PREBIOTIC FIBRES ARE ESSENTIAL The human digestive tract is for its portant function of fibre is to block re­ growth and functions much depending ceptors and prevent colonisation by po­ on supply of prebiotics. In contrast to tentially pathogenic microbes. For ex­ cows milk is breast milk very rich in ample mixing 2.5% of D-mannose into fibres. Apart from elephant milk, no drinking water reduces significantly other mammalian milk analysed till to­ colonisation of newborn chicken with day contains as much of fibre as breast Salmonella (Oyofo et al., 1989). milk (Gnoth et al., 2000). The complex Up to 25% of the adult population in fucosylated oligosaccharides in human Western countries suffer today of meta­ milk, with structural similarities to im­ bolic syndrome, a condition in which munomodulating cell surface glyco­ insulin resistance is a significant char­ conjugates, are supposed to protect acteristics. However, increased intake breast-fed infants against inflammations of dietary fibre (celluloses, hemicellulo­ and infections (Gnoth et al., 2000). In ses, pectins and starches), the main addition, these fibres are likely to func­ substrate of SCFA production, in­ tion as prebiotics and stimulate growth creases insulin sensitivity in humans of the non-pathogenic gut microflora in (Randle et al., 1963). Malhotra, an In­ the breast-fed infants. dian physician, observed already in Fruit and vegetable fibres are known 1968, that men living in North India to also have strong influence on intesti­ (Udaipur), and consuming large quanti­ nal growth in most mammalians. As an ties of cellulose and vegetable fibres + example, six weeks of supply of fer­ live lactobacilli had a longer mean clot­ mentable fibres (beet pulp and oli­ ting time, and soft jelly-like clots com­ gofructose) to experimental animals in­ pared to men in urban Madras. It has creases the GI surface area by 28%, the also been shown that supplement of the mucosal mass by 37%, the mucosal fibre konjac-glucomannan to baboons weight by 35% and the capacity for car­ living on Western diet significantly rier-mediated glucose uptake by 95% lowers the plasma level of fibrinogen (Buddington et al., 1999). Another im­ and of factor X (Vorster et al., 1985)

72 and that plasma viscosity and fibrinogen fibres and in glutamin-rich proteins. levels decrease significantly in diabetic When supplied to experimental animals children on supplementation with the did it result in dramatic improvement of fibre guar gum (Koepp and Hegewisch, induced colitis, an effect further 1981). It is also reported in the literature attenuated by combining with LAB not only that the incidence of post-op­ (Fukuda et al., 2002). Also antibiotics erative thrombosis is significantly re­ (vancomycin, metronidazole) did in this duced in patients on a high fibre diet study significantly attenuate clinical and (Frohn, 1976; Latto, 1976), and that pathological scores – but in contrast to supply of fibre (glucans) will signifi­ treatment with pre- and probiotics did cantly reduce the mortality rate in hos­ treatment with antibiotics result in a pital infections in patients with severe significant decrease in caecal buturate trauma (DeFillipe et al., 1993). levels. Another group of fibres, fructo-oli­ Pectins are one of the several fibre gosacharides (FOS), are known both to groups known for their many strong increase the numbers of certain lactic bioactivities. It is a superior mucosa acid bacteria, particularly Bifidobacteria, protectant, strong antioxidant, vehicle but also to significantly decrease the for transport of LAB through the GI number of Enterobacteriaceae in healthy tract and a superior substrate for bacte­ humans. Mixing 10% of FOS in the diet rial fermentation. The unripe banana to experimental animals had both pro­ (green sweet banana as well as plantain) tective and therapeutic effects against is rich in both pectin and resistant sodium sulphate-induced colitis starch. 250 g/l of green banana (Umemoto et al., 1998). It is observed (equivalent to two fruits) or 2 g pec­ that supplementing patients with IBD tin/kg food was recently tried as a sup­ with 30 g fructo-oligosaccharides (FOS) plement to rice diet in children in Bang­ per day increases significantly the ladesh suffering from persistent diar­ amount of luminal SCFAs (Umemoto et rhoea. The amounts of and frequency of al., 1998). But also other fibres such as stools, the duration of diarrhoea, num­ Psyllium husk (Hallert et al., 1991) and bers of vomiting, and use of oral rehy­ Platago ovata seeds (pectins) dration or amounts i.v. fluid solutions (Fernandez-Banares et al., 1999) are given were all significantly reduced in reported to have similar effects. A most the two treatment groups (Rabbani et recent study is of considerable interest. al., 2002). Recovery on third day was Germinated barley, the aleurone and seen in 59% in the green banana group, scutellum fraction, of the grain is in 55% in the pectin group compared to known to be rich both in hemicellulose 15% in the only rice group.

PROBIOTICS IN DIARRHOEA IN CHILDREN A larger European multi-centre trial Clinical signs of diarrhoea lasted in children one month to three years of 58.3±27.6 hours in the LAB-treated age was undertaken: One-hundred-and­ group to be compared to 71.9±35.8 forty children were randomly allocated hours (p=0.03) in the placebo group. to oral rehydration and placebo, another Diarrhoea lasted in rotavirus-positive 147 children to oral rehydration and children treated with LAB 56.2±16.9 daily supply of 1010 CFU of Lactoba­ hours compared to 76.6±41.6 in the cillus GG (Gualdalini et al., 2000). control group (p=0.008).

73 Lactobacillus GG was also tried in half (19 vs. 38, p<0.03) in the patients order to prevent diarrhoea in a placebo­ transferred from other hospitals – pa­ controlled trial performed in 204 under­ tients which most likely were sicker and nourished Peruvian children, age 6 to came late under treatment. No compli­ 24 months (Oberhelman et al., 1999). cations could be attributed to the use of The treatment was given to all children probiotics, even when given to very during a period of 15 months. The lac­ sick new-born children with an average tobacillus-treated children had fewer weight of 2600 g (range <1000 to episodes if diarrhoea (5.21 epi­ >4000 g), and often suffering from se­ sodes/child and year compared to 6.02 vere conditions such as sepsis, pneu­ in the placebo group, p=0.028). The monia or meningitis. Incidentally it was therapeutic gain, as pointed out by du observed that the LAB-treated children Pont (1999) and others, must be re­ suffered significantly less diaper der­ garded as modest. Most likely use of matitis. other and more potent LAB, or combi­ Lactobacillus GG (LGG) was also nations of LAB, should lead to a more tried in order to prevent diarrhoea in a significant therapeutic success. series of 202 antibiotic-treated children. One thousand three-hundred-thirty­ Twenty-five of the placebo placebo­ six new-born Columbian children with treated (26%) and only 7 of the LGG­ risk of developing severe diarrhoea re­ treated children developed diarrhoea ceived prophylactically during one week (Vanderhof et al., 1999). The mean du­ (or until they were discharged) a daily ration of diarrhoea was 4.7 days in the supply 250 million live Lactobacillus LGG group vs. 5.88 days in the pla­ acidophilus and 250 million live Bifi­ cebo group. Again, the efficacy of the dobacterium infantis and the outcome treatment is not impressive, and as compared to outcome for similar chil­ pointed out by Saavendra (1999), “the dren treated during the year before reduction of 1 day of two liquid stools (Hoyos, 1999): The incidence of narco­ over a 10 day period in a child might be tising enterocolitis was reduced by two questioned”. third (18 vs. 47, p<0.0005), and by

PROBIOTICS – AND PREBIOTICS ­ IN INFLAMMATORY BOWEL DISEASE (IBD) We observed in the early nineties Another recent study found both quan­ that humans with inflammatory bowel titative and qualitative changes in the disease have a reduced LAB flora, but LAB flora, when studying colonic bi­ also that induced colitis in experimental opsies from patents with ulcerative coli­ animals could be significantly reduced tis (UC) (Pathmakanthan et al., 1999). by supply of a combination of pre- and A significant quantitative decrease in probiotics – synbiotics (Fabia et al., growth of Lactobacillus spp. in colitis 1993a, 1993b). Subsequently it has biopsies was observed, but also a re­ been convincingly demonstrated that the duction in total aerobic speciation: 18 concentrations of endogenous Lactoba­ subspecies being found in UC patients cillus and Bifidobacteria are signifi­ compared to 32 in controls. Further­ cantly reduced in patients with active more, anaerobic speciation revealed in Crohn´s disease, ulcerative colitis, average 4.7 subspecies in UC patients pouchitis as well as in experimental co­ compared to 6.7 in controls. Inciden­ litis (Favier et al., 1997; Sartor, 1999). tally it was observed that Bacteroides

74 thetaiotaomicron occurred more often in nation of several LAB might have UC patients: 8/10 biopsies vs. 4/10 in strong clinical effects in IBD, eventually controls - an observations, which sig­ stronger than the use of single-bacteria nificance remains to be explored. treatments. It is tempting to anticipate A LAB cocktail called VSL#3 con­ that a cocktail consisting in LAB, where sisting in four Lactobacillus strains, each of the bacteria has been chosen three Bifidobacterium strains plus with the regard to their documented Streptococcus salivarius ssp. thermo­ metabolic and immunological effects, philus (5x1011 cells/g) is presently tried should eventually be even more suc­ quite extensively around the world. This cessful. It also tempting to suggest that composition is most probably chosen at combination with strong bioactive fibres random without any further (prebiotics) might even more improve documentation of the molecu­ the efficacy of treatment. The ideal lar/immunological effects for each of the treatment remedy will probably be com­ LAB, nor any evidence of synergistic plex, and much remains before the most effects. When three gram a day was suitable prebiotics, and the most effec­ given during one year did 15/20 patients tive probiotics have been identified. remain in remission, one lost to follow A most recent study is of consider­ up and 4/20 showed signs of relapse able interest (Swidsinsky, 2002). These (Venturi et al., 1999). VSL#3 was also authors studied the flora in 305 IBD pa­ tried in a small controlled study in pa­ tients and 40 controls using the most tients with pouchitis. Only 3/20 patients modern techniques: Quantitative PCR, had relapse of the disease when sup­ cloning, sequencing fluorescence in situ plied with VSL#3 compared 20/20 con­ hybridisation and electron microscopy. trol patients (Gionchetti et al., 2000). They observed a high density of mu­ These results are most likely better than cosal bacteria in sick patients, but also what presently can be achieved by any that the microbial “close to mucosa”­ conventional treatment, an assumption density of microbes increased progres­ supported by a recent systematic review sively with increasing severity of dis­ of the literature suggesting that ” met­ ease. Patients with > 10,000 CFU/Hl ronidazole is an effective treatment for showed a pronounced “band” of bacte­ active chronic disease” (odds ratio ria attached to mucosa, and patients with 12.34) but “oral probiotic therapy with >50,000 CFU/Hl had also signs of VSL#3 for maintaining remission” inclusions of polymorphic bacteria (odds ratio 15.33) (Sandborn, 1999). within solitary enterocytes next to lam­ Although the scientific basis for ina propria. The authors speculate that treatment of IBD seems reasonable and healthy mucosa is sterile - capable of attractive, it must be emphasised that it holding back faecal bacteria and prevent is far too soon to recommend routine a close contact of the microflora to the use of probiotics in IBD. Further stud­ epithelial surface. It is likely that this ied are much warranted. The good re­ function can be supported by treatment sults obtained in the two small studies with a combination of pre- and probiot­ cited above seem to suggest that combi­ ics (synbiotics).

PROBIOTICS IN HELICOBACTER PYLORI INFECTIONS It is now almost fifteen years since it the capacity to inhibit Helicobacter py­ was demonstrated that lactic acid pro- lori (Bhatia et al., 1989). The antibacte­ duced by Lactobacillus acidophilus has rial activity of seventeen strains of lac­

75 tobacilli against ten different strains of when live Lactobacillus GG was used H. pylori was recently studied (Lorca et (Armuzzi, 2001a, 2001b). al., 2001). All Lactobacillus strains Daily oral consumption of 4x50 ml were able to inhibit H. pylori, but the of the supernatant from a whey-based effect was lost if pH was adjusted to Lactobacillus acidophilus (La1) culture, 6.0. However, the effect of Lactobacil­ combined with either omeprazole or lus acidophilus CRL 639 remained even placebo, was reported to show a sig­ after pH was adjusted. The effect nificant reduction in breath test both seemed less related to pH and more to with and without supply of omeprazole, release of a proteinaceous compound, immediately as well as six weeks after with autolysin effects. the treatment episode (Michetti et al., One-hundred-and-twenty H. pylori 1999). It should be remembered that patients were randomised to, in addition whey is extraordinarily rich in immu­ to a 7-day triple therapy (Rabeprozole, nologically active and anti-infectious Chlarithromycin, Amoxicillin), receive substances such as lactoferrin, lysozym either placebo or a lyophilised and inac­ and many other antimicrobial peptides. tivated culture of Lactobacillus aci­ It is thus, this far not clear whether the dophilus. The eradication rate was sig­ observed effects are due to the Lactoba­ nificantly improved by supplementation cillus used, to the whey or a combina­ of the LAB: 52/59 patients (88%) vs. tion of both. 42/58 patients (72%) (p=0.03) (Can­ A recent study (Sakamoto et al., ducci et al., 2000). The effects of live 2001) reports considerable improve­ Lactobacillus GG was also investigated ment, both in urea breath test and serum but with less success: although the pepsinogen in 31 patients with Helico­ study reports improved tolerability (re­ bacter pylori infections treated during duced antibiotic-induced bloating, diar­ eight weeks with Lactobacillus Gasseri rhoea and taste disturbances), no im­ OLL 2716. provement in the rate of eradication was

SYNBIOTICS IN ICU PATIENTS There are good reasons to believe 1986). Pancreatic sepsis seems to be a that pre-, pro-, and synbiotics could strong determinant for complication dramatically change the outcome for such as multiple organ failure (MOF) critically ill patients, and be a good al­ and of death. It has rather recently been ternative to the use of antibiotics in ICU shown that infection of the pancreatic patients. It is regrettable that this far tissue is almost always preceded by only a handful of studies have been per­ about one week of colonisation the large formed in critically ill and postoperative intestine with non-coli Gram-negatives: patients, and, furthermore, and most of Pseudomonas, Klebsiella, Citrobacter, these studies are under publication. Enterobacter, Acinetobacter, Morgan­ ella, Serratia or Proteus (Luiten et al., Severe acute pancreatitis 1998). Prevention of such a colonisa­ Contamination of the pancreatic tis­ tion could be expected to have a dra­ sue occurs frequently in severe pan­ matic influence on outcome. creatitis, being reported to be 24% dur­ A prospective double-blind random­ ing the first week and amounting to ised study, comparing the influence of 72% during the third week (Beger et al., Lactobacillus plantarum 299 and oat fi­

76 bre with heat-killed Lactobacillus plan­ resection and intestinal by-pass (Rayes tarum 299 and oat fibre (control) was et al., 2002a). The groups treated with recently performed in severe pancreatitis either live or heat-killed LAB suffered (Oláh et al., 2002). The study was de­ less infections (3/30 in each group, e.g. signed to be concluded when repeat sta­ 10%) compared to 9/30 (30%) in the tistical analysis demonstrated statisti­ parenteral group (p>0.001). An even cally significant differences between the larger difference was observed when the two study groups. This occurred after subgroup of gastric and pancreatic all together 45 patients had entered the surgery patients was separately ana­ study. At that time 22 patients had dur­ lysed: None of eight patients receiving ing seven days received treatment with live LAB group, one of eight patients live LAB and 23 with heat-killed LAB (12%) receiving heat-killed LAB group during seven days. Infected necrosis and 3/6 (50%) conventionally treated and abscesses occurred in 1/22 patients with parenteral nutrition suffered infec­ (4.5%) in the live LAB group and in tions. 7/23 patients (30%) with heat-killed LAB. Abscesses occurred in 1/22 Liver transplantation patients (4.5%) in the treatment group vs. 7/23 A separate study was performed in (30%) (p=0.023) in the control group. human liver transplants by the same Although the length of stay was 13.7 group of clinicians in a study with a days in the treatment group vs. 21.4 similarly sized material of patients. days in the control group, the differ­ Comparison was made between selec­ ences had not reached statistical signifi­ tive bowel decontamination (SBD) + a cance at the time when the study was standard enteral formula, live Lactoba­ interrupted. The only patient who de­ cillus plantarum 299 + oat and inulin veloped sepsis in the treatment group fibres, and heat-killed Lactobacillus did that after fifteen days, e.g. eight plantarum 299 + oat and inulin fibres days after the treatment has been dis­ (Rayes et al., 2002b). The total amount continued. This seems to suggest that of fibres in the two last groups was treatment should be provided for a about 11 gram. The LAB were supple­ minimum of 14 days and most likely as mented during the first five days. The long as the patients are on antibiotics or sepsis rate was 48% in the selective have signs of GI colonisation. bowel decontamination group, 34% in the group treated with heat-inactivated Abdominal surgery patients LAB and 13% in the group receiving A prospective randomised study live LAB. Also the mean duration of compares the effect of live Lactobacillus antibiotic therapy, the mean total hospi­ plantarum 299 in a dosis of 109 with tal stay and the stay on ICU were heat-killed Lactobacillus plantarum 299 shorter compared to the groups with in the same dose and parenteral nutrition inactivated lactobacilli and fibre or with in 3x30 patients undergoing abdominal SBD. However, the size of the patient operations such as liver resection, pan­ material did not allow statistical signifi­ creas resection, gastric resection, colon cance to be reached.

FLORA IMPORTANT ON ALL BODY SURFACES Not only the gastro-intestinal tract, tion of infection and inflammation. Sec­ but also all body surfaces are coated by ond to the GI tract with its one to two a protective flora, essential for preven- kg of flora is the skin, calculated in the

77 adult human to be inhabited by ap­ recent reports suggest that consumption proximately 200 gram of bacteria. Other of LAB-containing drinks prevents important sites are the mouth and phar­ formation of biofilm and removes both ynx, the respiratory tract and the vagina, yeast and bacteria from silicon rubber each supposed to be inhabited by voice prostheses (Free et al., 2000; van approximately 20 gram of flora. Too der Mei et al., 2000). The flora is in­ much washing and cleaning will impair variably reduced at all these sites in sick this defence and open the door for op­ and hospitalised patients due to special portunistic infections. Animals have the hygienic requirements and large supply instinct to lick their wounds and hereby of antibiotics and other drugs. An over­ provide both protective flora and growth flow of probiotic bacteria from the GI factors, produced by the salivary tract to all the other sites seems normally glands. To apply topically lactic acid to occur, a function, which most likely bacteria on the skin and around all is severely reduced in the sick. It is not penetrations of the skin by foreign ma­ unlikely that in the future a dietary terials such as tubes, drains, tracheo­ supply of pre- and probiotics be com­ stomies etc is receiving an increasing plemented by spraying or applying gels interest. Such a treatment could also be of LAB on sensitive body surfaces, es­ of potential interest for treatment of pecially around the skin penetrations, burns. but also by using LAB-containing aero­ It has been observed that infants sols to promote flora of the respiratory treated with probiotics suffer much less tract, where such a protection layer is diaper dermatitis (Hoyos, 1999). Two much needed.

GUT ECOLOGY AND HEALTH – FUTURE ASPECTS Gut ecology is of the greatest im­ bodily functions such as appetite, sleep, portance for maintenance of health and mood and circadian rhythm, most likely prevention of diseases, increasingly through signal molecules but also seen in Western countries and increas­ through metabolites produced by mi­ ingly linked with deranged gut flora and crobial fermentation in the gut, known mucosal lesions. Not only has it been to influence lymphocyte function, pro­ observed that diseases such as rheuma­ duction of immunoglobulins and resis­ toid arthritis (Midtvedt, 1987; Zhang et tance to disease - see further Bengmark al., 2000; Nieuwenhuis et al., 2000) (2002a,b). and atopic diseases (Satomi, 1966; There are also indications in the lit­ Rock, 1998; Wold, 1998) are associated erature that intestinal microflora stimu­ with gut flora derangements but also lates myo-electric activity in the intestine diseases such as autism (Sandler et al., and hereby controls gastrointestinal 2000; Wakefield et al., 2000; Furlano et motility and transit of food under diges­ al., 2001; Lindsey, 2001; Torrente et tion (Husebye et al., 1994). Recent al., 2002), graft-versus-host disease studies performed in germfree animals (van Bekkum et al., 1974; Porrata et al., do also suggest (Hooper et al., 2001) 2001) and formation of serosal adhesion that some commensal bacteria modulates (Bothin et al., 2001) are intimately the genes involved in whole series of associated with gut flora and mucosa. important intestinal functions such as Some observations suggest that both nutrient absorption, mucosal barrier pre- and probiotics can modify basic fortification, xenobiotic metabolism,

78 angiogenesis and postnatal intestinal re-absorption of electrolytes and water maturation. to a complex organs, which holds im­ The dramatic change in recent years portant keys to health and well-being. in our knowledge and understanding of Further exploration of the large intestine the complex functions of the lower GI and its interaction with flora has the tract and its function has without ques­ prospect of helping us to understand tion contributed considerably to our un­ and prevent a whole series of diseases derstanding of health and disease. Dur­ including the endemic diseases so much ing my lifetime the view on the large plaguing the Western, and increasingly intestine and its functions has radically also the Eastern world. changed from being an organ mainly for

LITERATURE

Armuzzi, A., Cremonini, F., Ojetti, V., Bar­ synbiotics. Curr. Opin. Crit. Care 8, 145­ tolozzi, F., Canducci, F., Candelli, M., 151 (2002). Santarelli, L., Cammarota, G., De Lorenzo, Bhatia, S.J., Kochar, N., and Abraham, P.: A., Pola, P., Gasbarrini, G., and Gasbar­ Lactobacillus acidophilus inhibits growth rini, A.: Effect of Lactobacillus GG sup­ of Campylobacter pylori in vitro. J. Clin. plementation on antibiotic-associated gas­ Microbiol. 27, 2328-2330 (1989). trointestinal side effects during Helicobacter Bothin, C., Okada, M., Midtvedt, T., and Per­ pylori eradication therapy: A pilot study. beck, L.: The intestinal flora influences ad­ Digestion 63, 1-7 (2001a). hesion formation around surgical anasto­ Armuzzi, A., Cremonini, F., Bartolozzi, F., moses. Br. J. Surg. 88, 143-145 (2001). Canducci, F., Candelli, M., Ojetti, V., Brandtzaeg, P., Halstensen, T.S., Kett, K., Cammarota, G., Anti, M., De Lorenzo, A., Krajci, P., Kvale, D., Rognum, T.O., Pola, P., Gasbarrini, G., Gasbarrini, A.: Scott, H., and Sollid, L.M.: Immunobiol­ The effect of oral administration of Lacto­ ogy and immunopathology of human gut bacillus GG on associated gastrointestinal mucosa: Humoral immunity and intraepi­ side effects during Helicobacter eradication. thelial lymphocytes. Gastroenterology 97, Aliment. Pharmacol. Ther. 15, 163-169 1562-1584 (1989). (2001b). Buddington, R.K., Buddington, K.K., and Beger, H.G., Bittner, R., and Büchler, M.: Bac­ Sunvold, G.D.: Influence of fermentable fi­ terial contamination of pancreatic necrosis – ber on small ingtestinal dimensions and A prospective clinical study. Gastroenterol­ transport of glucose and proline in dogs. ogy 91, 433-438 (1986). Am. J. Vet. Res. 60, 354-358 (1999). Bengmark, S.: Nutrition and resistance to dis­ Carroll, K.K.: Lipids in cancer. In: Nutrition ease. In: Old Herborn University Seminar and disease update. Cancer (Eds.: Carroll, Monographs 14: Intestinal tranbslocation K.K. and Kritchevsky, D.). AOCS Press, (Eds.: Heidt, P.J., Nieuwenhuis, P., Rusch, Champaign, 235-296 (1994). V.D., and van der Waaij, D.). Herborn Lit­ de Felippe Jr., J., Da Rocha e Silva Jr., M., terae, Herborn-Dill, 117-133 (2001). Marciel, F.M., Soares Ade, M., and Men­ Bengmark, S.: Use of pro-, pre- and synbiotics des, N.F.: Infection prevention in patients in the ICU – Future options. In: Nutri­ with severe multiple trauma with the im­ tional considerations in the intensive care munomodulator beta 1-3 polyglucose. unit – Science, rationale and practice (Eds.: Surg. Gyn. Obst. 177, 383-388 (1993). Shikora, S.A., Martindale, R.G., and Du Pont, H.: Prevention of diarrhea by the Schwaitzberg, S.D.). Kendall/Hunt Pub­ probiotic Lactobacillus GG. J. Pediatr. 134, lishing Company, Dubuque, 381-399 1-2 (1999). (2002a). Eaton, B.S. and Konnor, M.: Paleolithic nu­ Bengmark, S.: Gut microbial ecology in criti­ trition: A consideration of its nature and cal illness: Is there a role for pre-, pro-, and current implications. N. Engl. J. Med. 312,

79 283-289 (1985). Wakefield, A.J., Walker-Smith, J.A., and Fabia, R., Ar´Rajab, A., Johansson, M.L., Murch, S.H.: Colonic CD8 and gamma Andersson, R., Willen, R., Jeppsson, B., delta T-cell infiltration with epithelial dam­ Molin, G., and Bengmark, S.: Impairment age in children with autism. J. Pediatr. 138, of bacterial flora in human ulcerative colitis 366-72 (2001). and in experimental colitis in the rat. Di­ Gionchetti, P., Rizello, F., Venturi, A., Bri­ gestion 54, 248-255 (1993a). gidi, P., Matteuzzi, D., Bazzocchi, G., Fabia, R., Ar´Rajab, A., Johansson, M.L., Poggioli, G., Miglioli, M., and Campieri, Willen, R., Andersson, R., Molin, G., and M.: Oral bacteriotherapy as maintenance Bengmark, S.: The effect of exogenous ad­ treatment in patients with chronic pouchi­ ministration of Lactobacillus reuteri R2LC tis: A double-blind, placebo-controlled trial. and oat fibre on acetic acid-induced colitis in Gastroenterology 119, 305-309 (2000). the rat. Scand. J. Gastroenterol. 28, 155­ Gnoth, M.J., Kunz, C., Kinne-Saffran, E., and 162 (1993b). Rudloff, S.: Human milk oligosacharides Favier, C., Neut, C., Mizon, C., Cortot, A., are minimally digested in vitro. J. Nutr. Colombel, J.F., and Mizon, J.: Fecal β-D­ 130, 3014-3020 (2000). galactosidase and bifidobacteria are decreased Guandalini, S., Pensabene, L., Zikri, M.A., in Crohn´s disease. Dig. Dis. Sci. 42, 817­ Dias, J.A., Casali, L.G., Hoekstra, H., 822 (1997). Kolacek, S., Massar, K., Micetic-Turk, D., Fernandez-Banares, F., Hinojosa, J., Sanchez- Papadopoulou, A., de Sousa, J.S., Sandhu, Lombrana, J.L., Navarro, E., Martinez- B., Szajewska, H., and Weizman, Z.: Lac­ Salmeron, J.F., Garcia-Puges, A., Gon­ tobacillus GG administered in oral rehydra­ zalez-Huix, F., Riera, J., Gonzalez-Lara, tion solution to children with acute diar­ V., Dominguez-Abascal, F., Gine, J.J., rhea: A multicenter European study. J. Pe­ Moles, J., Gomollon, F., and Gassull, diatr. Gastroenterol. Nutr. 30, 54-60 M.A.: Randomized clinical trial of Plantago (2000). ovata seeds (dietary fiber) as compared to Hallert, C., Kaldma, M., and Petersson, B.G.,: mesalamine in maintaining remission in Ispaghula husk may relieve gastrointestinal ulcerative colitis. Am. J. Gastroenterol. 94, symptoms in ulcerative colitis in remis­ 427-433 (1999). sion. Scand. J. Gastroenterol. 26, 747-750 Frame, L.T., Hart, R.W., and Leakey, J.E.A.: (1991). Caloric restriction as a mechanism mediat­ Hooper, L.V., Wong, M.H., Thelin, A., Hans­ ing resistance to environmental disease. son, L., Falk, P.G., Gordon, J.I.: Molecu­ Environ. Health Perspect. 106, 313-324 lar analysis of commensal host-microbial (1998). relationships in the intestine. Science 291, Free, R.H., van der Mei, H.C., Dijk, F., van 881-884 (2001). Weissenbruch, R., Busscher, H.J., and Al­ Hoyos, A.B.: Reduced incidence of necrotizing bers, F.W.: Biofilm formation on voice enterocolitis associated with enteral admini­ prostheses: Influence of dairy products in stration of Lactobacillus acidophilus and Bi­ vitro. Acta Otolaryngol. 120, 92­ fidobacterium infantis to neonates in an in­ 99 (2000). tensive care unit. Int. J. Infect. Dis. 3, 197­ Frohn, M.J.N.: Letter to the Editor. Lancet ii, 202 (1999). 1019-1020 (1976). Husebye, E., Hellström, P.M., and Midtvedt, Fukuda, M., Kanauchi, O., Araki, Y., Andoh, T.: Intestinal flora stimulates myoelectric A., Mitsuyama, K., Takagi, K., Toyonaga, activity of rat small intestine by promoting A., Sata, M., Fujiyama, Y., Fukuoka, M., cyclic initiation and aboral propagation of Matsumoto, Y., and Bamba, T.: Prebiotic migrating myoelectric complex. Dig. Dis. treatment of experimental colitis with ger­ Science 39, 946-956 (1994). minated barley food stuff: A comparison Koepp, O. and Hegewisch, S.: Effects of guar with probiotic and antibiotic treatment. Int. on plasma viscosity and related parameters. J. Molecul. Med. 9, 65-70 (2002). Europ. J. Pediatr. 137, 31-33 (1981). Furlano, R.I., Anthony, A., Day, R., Brown, Kruszewska, K., Lan, J., Lorca, G., Yanagi­ A., McGarvey, L., Thomson, M.A., Da­ sawa, N., Marklinder, I., and Ljungh, Å.: vies, S.E., Berelowitz, M., Forbes, A., Selection of lactic acid bacteria as probiotic

80 strains by in vitro tests. In: Proceedings of (1989). the XVI International Congress on micro­ Pathmakanthan, S., Thornley, J.P., and bial ecology and disease, October 3-6, Hawkey, C.J.: Mucosally associated bacte­ 2001, Noordwijkerhout, The Netherlands ria flora of the human colon: Quantitative (Eds.: Heidt, P.J., Midtvedt, T., Norin, E., and species specific differences between and Rusch, V.). Microecol. Dis. 29, 37-51 normal and inflamed colonic biopsies. Mi­ (2002). crob. Ecol. Health Dis. 11, 169-174 Latto, C.: Postoperative deep-vein thrombosis, (1999). pulmonary embolism, and high-fibre diet. Porrata, L.F., Gertz, M.A., Inwards, D.J., Lit­ Lancet ii, 1197 (1976). zow, M.R., Lacy, M.Q., Tefferi, A., Lindsay, .LA.: Saccharomyces boulardii: Po­ Gastineau, D.A., Dispenzieri, A., Ansell, tential adjunctive treatment for children S.M., Micallef, I.N., Geyer, S.M., and with autism and diarrhea. J. Child. Neurol. Markovic, S.N.: Early lymphocyte recovery 16, 387 (2001). predicts superior survival after autologous Lorca, G.L., Wadström, T., Fond de Valdez, hematopoietic stem cell transplantation in G., and Ljungh, Å.: Lactobacillus acidophi­ multiple myeloma or non-Hodgkin pa­ lus autolysins inhibit Helicobacter pylori in tients. Blood 98, 579-585 (2001). vitro. Curr. Microbiol. 42, 39-44 (2001). Rabbani, G.H., Teka, T., Zaman, B., Majid, Luiten, E.J.T., Hop, W.C.J., Endtz, H.P., and N., Khatun, M., and Fuchs, G.J.: Clinical Bruining, H.A.: Prognostic importance of studies in persistant diarrhea; dietary man­ gram-negative intestinal colonization pan­ agement swith green banana or pectin in creatic infection in severe acute pancreatitis. Bangladesh children. Gastroenterology 121, Intensive Care Med. 24, 438-445 (1998). 554-560 (2001). Malhotra, S.L.: Studies in blood coagulation, Randle, P.J., Hales, C.N., Garland, B.P., and diet and ischaemic heart disease in two Newsholme, E.A.: The glucose fatty-acid population groups in India. Br. Heart J. 30, cycle. Its role in insulin sensitivity and the 303-308 (1968). metabolic disturbances of diabetes mellitus. Midtvedt, T.: Intestinal bacteria and rheumatic Lancet i, 785-89 (1963). disease. Scand. J. Rheum. 64 (Suppl), 49­ Rayes, N., Hansen, S., Boucsein, K., Seehofer, 54 (1987). D., Müller, A.R., Serke, S., Brammer, N., Nieuwenhuis, E.E., Visser, M.R., Kavelaars, Bengmark, S., and Neuhaus, P.: A., Cobelens, P.M., Fleer, A., Harmsen, Comparison of parenteral and early enteral W., Verhoef, J., Akkermans, L.M., and nutrition with fibre and Lactobacilli after Heijnen, C.J.: Oral antibiotics as a novel major abdominal surgery – a prospective therapy for arthritis: Evidence for a benefi­ randomized trial. Nutrition, 18, 609-615 cial effect of intestinal Escherichia coli. Ar­ (2002a). thritis Rheum. 43, 2583-2589 (2000). Rayes, N., Hansen, S., Boucsein, K. , Müller, Oberhelman, R.A., Gilman, R.H., Sheen, P., A.R., Serke, S., Brammer, N., Bengmark, Taylor, D.N., Black, R.E., Cabrera, L., S., and Neuhaus, P.: Early enteral supply Lescano, A.G., Meza, R., and Madico, G.: of Lactobacillus and fibre vs. selective A placebo-controlled trial of Lactobacillus bowel decontamination (SBD) – a con­ GG to prevent diarrhea in undernourished trolled trial in liver transplant recipients. Peruvian children. J. Pediatr. 134, 15-20 Transplantation, 74, 123-127 (2002b). (1999). Rook, G.A. and Stanford, J.L.: Give us this Oláh, A., Belágyi, T., Issekutz, Á., Gamal, day our daily germs. Immunol. Today 19, M.E., and Bengmark, S.: Early enteral nu­ 113-116 (1998). trition with specific Lactobacillus and fibre Saavendra, J.M.: Probiotics plus antibiotics; reduces sepsis in severe acute pancreatitis. regulating our bacterial environment. J. Pe­ Brit. J. Surg., 89, 1103-1107 (2002). diatr. 135, 535-537 (1999). Oyofu, B.A., DeLoach, J.R., Corrier, D.E., Sakamoto, I., Igarashi, M., Kimura, K., Norman, J.O., Ziprin, R.L., and Mollen­ Takagi, A., Miwa, T., and Koga, Y.: Sup­ hauer, H.H.: Prevention of Salmonella ty­ pressive effect of Lactobacillus Gasseri phimurium colonization of broilers with d­ OLL2716(LG21) on Helicobacter pylori in­ mannose. Poult. Sci. 68, 1357-1360 fections in humans. J. Antimicrob. Che­

81 mother. 47, 709-710 (2001). Wakefield, A.J., Anthony, A., Murch, S.H., Sandborn, W.J., McLeod, R., and Jewell, D.P.: Thomson, M., Montgomery, S.M., Da­ Medical therapy for induction and vies, S., O'Leary, J.J., Berelowitz, M., and maintenance of remission in pouchitis: A Walker-Smith, J.A.: Enterocolitis in chil­ systemic review. Inflamm. Bowel Dis. 5, dren with development disorders. Am. J. 33-39 (1999). Gastroenterol. 95, 2285-2295 (2000). Sandler, R.H., Finegold, S.M., Bolte, E.R., van Bekkum, D.W., Rodenburg, J., Heidt, Buchanan, C.P., Maxwell, A.P., Vaisanen, P.J., and van der Waaij, D.: Mitigation of M.L., Nelson, M.N., and Wexler, H.M.: secondary disease of allogenic mouse radia­ Short-term benefit from oral vancomycin tion chimeras by modification of intestinal treatment of regressive-onset autism. J. flora. J. Natl. Cancer Inst. 52, 401-404 Child. Neurol. 15, 429-435 (2000). (1974). Sartor, R.B.: Microbial factors in the patho­ van der Mei, H.C., Free, R.H., Elving, G.J., genesis of Crohn´s disease, ulcerative coli­ van Weissenbruch, R., Albers, F.W., and tis and experimental intestinal inflamma­ Busscher, H.J.: Effect of probiotic bacteria tion. In: Inflammatory bowel diseases, 5th on prevalence of yeasts in oropharyngeal edition (Ed.: Kirsner, J.G.). Saunders, biofilms on silicone rubber voice prosthe­ Philadelphia, 153-178 (1999). ses in vitro. J. Med. Microbiol. 49, 713­ Satomi, T.: Study on allergy to intestinal flora. 718 (2000 1. Toxin activity of anticlostolysin theta in Venturi, A., Gionchetti, P., Rizzello, F., Jo­ various diseases. 2. Experimental renal hansson, R., Zucconi, E., Brigidi, P., Mat­ lesions induced by sensitization with teuzzi, D., and Campieri, M.: Impact on Clostridium welchii toxin. Arerugi 15, 48­ the composition of the faecal flora by a new 54 and 55-61 (1966) (in Japanese). probiotic preparation: Preliminary data on Torrente, F., Ashwood, P., Day, R., Machado, maintenance treatment of patients with ul­ N., Furlano, R.I., Anthony, A., Davies, cerative colitis. Aliment. Pharmacol. Ther. S.E., Wakefield, A.J., Thomson, M.A., 13, 1103-1108 (1999). Walker-Smith, J.A., and Murch, S.H.: Wold, A.: The hygiene hypothesis revised: Is Small intestinal enteropathy with epithelial the rising frequency of allergy due to IgG and complement deposition in children changes in the intestinal flora? Allergy 53 with regressive autism. Mol Psychiatry, 7, (Suppl), 20-25 (1998). 375-382 (2002). Vorster, H.H., Kruger, H.S., Fryllinck, S., Umemoto, Y., Tanimura, H., Ishimoto, K., Botha, B.J., Lombaard, W.A., and De Masaki, K., and Sahara, M.: Trial of a new Jager, J.: Physiological effects of the die­ fructo-oligosacharide therapy against IBD. tary fiber component konjac glucomannan AGA Abstract G4510. Gastroenterology in rats and baboons J. Plant Foods 6, 263­ 114, 1102 (1998). 274 (1985). Vanderhoof, J.A., Whitney, D.B., Antonson, Zhang, X., Rimpilainen, M., Simelyte, E., and D.L., Hanner, T.L., Lupo, J.V., and Toivanen, P.: What determines arthri­ Young, R.J.: Lactobacillus GG in the pre­ togenicity of bacterial cell wall? A study on vention of antibiotic-associated diarrhea in Eubacterium cell wall-induced arthritis. children. J. Pediatr. 135, 564-568 (1999). Rheumatology (Oxford) 39, 274-82 (2000).

82 DEVELOPING AN IN VITRO MODEL ON THE INVESTIGATION OF THE CROSSTALK AMONG BACTERIA, ENTEROCYTES AND LEUKOCYTES NEAR THE INTESTINAL MUCOSA ALEXANDR PARLESAK and CHRISTIANE BODE Department of Physiology of Nutrition, Hohenheim University, Stuttgart, Germany

SUMMARY In vitro models comprising human cells are potent tools to improve our understanding on the crosstalk between enterocytes and bacteria on the one hand and between leukocytes and enterocytes on the other. The present manuscript deals with the possibility to develop a co-culture model for the investigation of inflammatory processes near the intesti­ nal mucosa. One important demand on such a model is the spatial sepa­ ration of bacteria and leukocytes mimicking the morphologic conditions in the intestine. This is achieved by the separation of the apical com­ partment with non-pathogenic bacteria from the basolateral compart­ ment with human leukocytes by a confluent monolayer of a differenti­ ated cell line of intestinal origin (Caco-2). The gut-associated lymphoid tissue (GALT) comprises Peyer’s patches, mesenteric lymph nodes, intra-epithelial lymphocytes, and leukocytes of the lamina propria. The complex interaction of these four compartments cannot be simulated by an in vitro model, but inflamma­ tory processes are likely to be initiated and controlled by alpha-beta T cells in the lamina propria, the action of which is likely to be simulated most closely by the present model. Preliminary results on the kinetics of cytokines in the basolateral compartment revealed a temporary occurrence of tumour-necrosis fac­ tor alpha, a steady increase in the concentration of interleukin (IL)-8 and a stable concentration of IL-6 and IL-10 after 24 hours. Cytokines associated with the activation of Th1-cells (IL-2 and interferon gamma) occurred after 24 h and steadily increased from that time. Th2-type cy­ tokines (IL-4, IL-5) were not detectable in the basal medium. The present model offers chances in the investigation of drugs and food ingredients on the enhancement or impairment of inflammatory processes in the intestine avoiding differences between the immune re­ sponses of man and animals in e.g. feeding models with rodents.

INTRODUCTION Background inflammatory bowel disease (IBD), To improve our understanding of namely Crohn’s disease and ulcerative mechanisms inducing processes in the colitis, are not completely understood. intestine makes sense from different Second, the need for developing new points of view. First, events leading to strategies to take influence on the effi­

83 cacy of the immune response by com­ they have to be distinguished clearly due pounds or nutrients will increase. Espe­ to their immunological functionality. cially as the percentage of elderly people Developing an in vitro-model for the increases in industrialised countries, modulation of inflammatory processes problems associated with the impaired in the intestine, one has to be aware immune response in these people (‘im­ which parts of this immune system munosenescence’) have to be met. should and can be simulated. The systemic immune response in terms of resistance to pathogens can Which part of the GALT action highly be improved by the application of can be mimicked in an in vitro probiotic organisms. In mice, the ap­ system? plication of Lactobacillus casei prior to The initiation of the specific immu­ infection with Salmonella typhimurium nity acquired within the GALT takes (dose: 20 x LD50) prevented completely place in the Peyer’s patches (PP), which the occurrence of pathogens in the liver can be considered as classical secondary and spleen (Perdigon et al., 1991). As lymphoid organs. After controlled evident from these experiments, the ef­ translocation of single bacteria by fect of e.g. probiotics can be immense specialised M-cells, interdigitating but further investigations are necessary dendritic cells (IDC) present antigens to to understand the mechanisms standing T lymphocytes and macrophages mainly behind these effects. Although insights to B lymphocytes. During the matura­ can be achieved by using experiments tion process, the primed cells leave the with animals, considerable differences PP via the mesenteric lymph nodes to can occur in parameters of immune relocate after maturation in the lamina function between animals and man (Le­ propria or other sites such as the mam­ brec et al., 1995). Therefore, in vitro mary gland, the lacrimal gland and be­ models using cells of human origin have neath other mucosa surfaces of the become more important during the last body. The expression ‘homing’ termed two decades, but efforts in the de­ for this process is not completely exact, velopment of such models mimicking as the cells never return to the site of the immune response in the intestine their initiation but to places where they were very limited so far. Furthermore, can fulfil their efferent immunological at least at the primary stage of investi­ tasks. The main function of the cells gating immunomodulating properties of having undergone this fate is the pre­ foodstuff, screening methods will be dominant production of IgA. Evidently, needed enabling first statements on their this process involving differentiation effectiveness. processes in other tissues than near the An important mediator of the sys­ intestinal mucosa cannot be mimicked in temic immune response is the gut-asso­ an in vitro model dealing with the im­ ciated lymphoid tissue (GALT). This mune response in the intestine. There­ complex system, which is part of the fore, investigations within such a model mucosa-associated lymphoid tissue can only focus on actions of cells other (MALT), comprises four different parts: than naïve T/B cells in PP and activated The Peyer’s patches, mesenteric lymph memory B lymphocytes and nodes, intra-epithelial lymphocytes CD45RO+/CD45RBlow lymphocytes, (IEL), and various leukocytes located in which have lost their ability for being the lamina propria. Despite the fact that stimulated in terms of acute inflamma­ the latter two compartments are sepa­ tion (Abreu-Martin et Targan, 1996). rated only by the thin basal membrane,

84 Figure 1: Arrangement of bacteria, Caco-2 cells and peripheral blood mononuclear cells (PBMC) for the investigation of immune responses near the intestinal surface

Although an important part of the intestine, the following demands should GALT, intra-epithelial lymphocytes be met by cells representing the three (IEL), which comprise a considerable parts of the system mentioned above: percentage of CD8-positive cells Leukocytes should mainly comprise (>80%) expressing frequently the human naïve αβ-T cells and cells of the CD8αα+ receptor (Latthe et al., 1994), monocyte/macrophage lineage being are also considered as being practically susceptible for an inflammatory chal­ not being able to respond to TcR stimuli lenge. An acceptable source for these (Ebert, 1989). The low, but quite con­ leukocytes is peripheral blood mononu­ stant percentage of γδ-TcR cells (in hu­ clear cells (PBMC), which can easily be mans: ~10% of total T cells; Viney et obtained in large amounts from cell-en­ al., 1990) seems also to be of inferior riched plasma (‘buffy coats’). importance for intestinal inflammation, Enterocytes are only appropriate for an as αβ-TcR cell depleted mice failed to in vitro model of intestinal inflammation develop inflammation after Salmonella if they can maintain the separation of infection but not γδ-TcR cell depleted two compartments: the ‘outer’ (apical) animals (Weintraub et al., 1997). compartment with bacteria and the ‘in­ Hence, an in vitro model focussing on ner’ (basolateral) compartment with inflammatory processes within the in­ leukocytes. This will only succeed with testine should mainly be oriented on the cells that keep their proliferation only interaction among naïve LPL (and until confluence and are able to develop macrophages), enterocytes and the in­ a polar differentiation. Caco-2 cells meet testinal flora. these requirements and are able to form desmosomes, microvilli, and tight Which kind of cells can be used junctions (Hidalgo et al., 1989). Addi­ in the co-culture model? tionally, this extremely well-defined cell To achieve the best possible results line (Pinto et al., 1989) keeps its viabil­ in experiments concerning reproducibil­ ity also in media suitable for incubation ity and similarity to conditions in the of leukocytes such as DME medium.

85 The Caco-2 cells have to be kept for experiments with this model by Haller et growth until confluence and differentia­ al. (2000) revealed that practically all tion on a semi-permeable membrane cytokines are secreted into the baso­ until the end of incubation to enable lateral but not the apical compartment. signalling between leukocytes and the Out of the numerous cytoki­ enterocyte-like cell line by soluble fac­ nes/chemokines, those indicating acute tors. To guarantee that bacteria cannot inflammation [tumour necrosis factor transmigrate through the membrane, a alpha (TNF-α), Interleukin 1 beta (IL­ pore diameter below the average size of 1β), IL-8)], immunomodulation [IL-10, bacteria (e.g. 0.4 Hm) is suitable for this transforming growth factor beta (TGF­ purpose. β), IL-6], activation of Th1 cells (IL-2, Bacteria found in the intestine are very interferon gamma (IFN-γ)], and activa­ inhomogeneous and their composition tion of Th2 cells (IL-4, IL-5, IL-13) can differ significantly between single might be of special importance for the hosts. Apart from studies where differ­ assessment of changes in the immune ences among specific rods are to be in­ response. For a profound understanding vestigated, the application of one non­ of the interaction between these cy­ pathogenic, well-characterised bacteria tokines and the eukaryotic cells of the species that is able to evoke an immu­ system, the recording of a time kinetic nological response within the described of each cytokine is necessary. The in­ system seems appropriate. The immu­ volvement of the enterocyte-like cell line nomodulating effect of drugs or nutri­ can be estimated by (semi-)quantitative ents to be investigated can than be as­ measurement of mRNA expression for sessed by enhancement or impairment some of the cytokines mentioned above of the immunological response to the (i. E. IL-8 and TNF-α, Haller et al., standardised bacterial challenge. Evi­ 2000). dently, the germs have to be impaired in Further important information can be their growth to avoid destruction of the obtained by measuring the release of Caco-2 cell monolayer. A major prob­ reactive oxygen species (ROS). The ac­ lem in this context is the question in tivation of phagocytotic cells, which are how far the growth of the applied bacte­ also part of the PBMC (monocytes, ria can be reduced without changing the ~20-35%), results in the pentose phos­ induced immune response. Antibiotics phate shunt, NADPH production, and damaging the integrity of the bacterial the massive release of ROS. As ROS membrane should be avoided, as the are known as potent reagents being able bacterial lysate is likely to cause modifi­ to modify DNA, especially mechanisms cations of the induced immuno-activa­ leading to the high coincidence of tion. A sketch of the co-culture model is chronic inflammatory bowel disease given in Figure 1. (IBD) and malignant neoplasiae in the colon might be elucidated by investiga­ How to measure the extent of the tions with the present model. The re­ immune response? lease of ROS can easily be measured Concerning the readout of the ex­ online by the application of chemilumi­ periments, the middle-termed immune nescence-enhancing techniques (Parle­ response can be assessed by measure­ sak et al., 1998). ment of cytokines in the medium. First

86 MATERIAL AND METHODS Cell culture rod Escherichia coli K12 was used at The cell line Caco-2 was obtained the stationary growth phase. Gentamicin from the ‘Deutsche Sammlung von (120 Hg/ml; Gibco BRL/Life Tech­ Mikroorganismen und Zellkulturen’ nologies) was added to the medium to (Braunschweig, Germany: DSMZ No. avoid uncontrolled growth that would ACC169). Cells were cultured in inserts lead to destruction of the Caco-2 cell on a semi-permeable PET membrane monolayer (Haller et al., 2000). (Falcon, Becton-Dickinson, Le Pont De Mononuclear leukocytes from pe­ Claix, France). Inserts were placed into ripheral venous blood (PBMC) of cavities of six-well plates (Greiner, healthy donors (n=3) were isolated from Frickenhausen, Germany), which were buffy coats (cell-enriched plasma) by filled with cell culture medium that con­ density gradient centrifugation (Ficoll® sisted of Dulbecco’s Modified Eagle solution, Seromed/Biochrom, Berlin, Medium (DMEM) and heat-inactivated Germany). After centrifugation, mono­ (56°C for 30 min) foetal calf serum nuclear leukocytes were taken from the (FCS) (Gibco BRL/Life Technologies, interface and washed three times with Karlsruhe, Germany). FCS was con­ cell culture medium. trolled for endotoxin content (<0.2 Each experiment was performed as a ng/ml: LAL Test, Chromogenix, triplicate. Incubations with non-patho­ Mölndal, Sweden). Confluency was genic bacteria (E. coli K12, 2.0 x 107 checked by measurement of transepi­ CFU) were applied either apically or thelial electrical resistance (TEER; basolaterally with PBMC (4.0 x 106). Haller et al., 2000) and visual control of The release of cytokines into the baso­ cell layer integrity under the micro­ lateral compartment was measured by scope. After complete differentiation, commercially available ELISA kits (BD which lasted for further 10 days, co-in­ Pharmingen, Heidelberg, Germany). cubation experiments were performed. The resulting values were compared to As a stimulus, the non-pathogenic the corresponding control experiments.

FIRST RESULTS AND CONCLUSION First results from experiments with and L. sakei but not by L. johnsonii. this model revealed that findings ob­ Hence, the Caco-2 cells can distinguish tained in this model cannot be compared between single bacteria species only if with those achieved by bacteria-leuko­ mononuclear leukocytes are present in cyte or bacteria-enterocyte co-incuba­ the basolateral compartment. Further tions. Haller et al. (2000) stimulated the actual findings with this model differentiated Caco-2 cell monolayer in demonstrated that IL-2, which cannot be the described trans-membrane system produced by the enterocyte-like cell line with different bacteria (Lactobacillus (Jung et al., 1995), is produced in the johnsonii, L. sakei, and E. coli) with basolateral compartment only if the and without leukocytes in the basolateral Caco-2 cell monolayer separates bacteria compartment. Messenger RNA for IL-8 and leukocytes. If bacteria have cell-cell and monocyte chemoattractant protein 1 contact to the leukocytes, practically no (MCP-1) was expressed in Caco-2 cells IL-2 production occurs within the first only if leukocytes were present in the 68 h after challenge. basal compartment and only by E. coli

87 Figure 2: Kinetics of cytokine release into the basolateral compartment of the co-culture model after apical stimulation with E. coli K12 (4.0 x 107 bacteria); the absolute concentrations of the single cytokines are not identical.

These findings underscore the diver­ single cytokines differ significantly in sity of the present model compared to some orders of magnitude regarding models where the spatial separation their absolute concentration. As can be between bacteria and leukocytes by a expected, out of the measured cytokines cell monolayer comparable to the intes­ (TNF-α, IL-1β, IL-12, IL-8, IL-6, IL­ tinal mucosa does not occur. Further­ 2, IL-10, IFN-γ, IL-4, IL-5, and TGF­ more, from the results motioned above a β) those having a systemic effect (IL-6) three-step mechanism can be concluded or being important for chemotaxis (IL­ in the activation of the eucaryotic cell 8) contribute a major part of the pro­ ‘crosstalk’. First, the leukocytes secrete duced cytokine mass (ng/ml range; unidentified soluble factors sensitising >90%) while cytokines with local effi­ the Caco-2 cells for selective activation cacy occur only at clearly inferior con­ by bacteria. Second, leukocytes are centrations (pg/ml-range). Cytokines activated by soluble factors from Caco-2 associated to the activation of Th2 cells cells, enhancing the production of (IL-4 and IL-5) were not detected in the cytokines already produced by the basolateral medium at all. A schematic enterocyte-like cell line (TNF-α, IL-8, overview of the time kinetics of the in­ IL-6 and others) and initiating the vestigated cytokines is given in Figure production of cytokines that cannot be 2. synthesised by the Caco-2 cells (IL-2, To further elucidate the differences IL-10, IFN-γ and others). In the third of the present model as compared to step, changes in the concentration of models of leukocyte challenge by direct cytokines in the basolateral compartment contact with bacteria, further experi­ regulate the activation status of ments were performed. Keeping the enterocytes. number of leukocytes constant, the di­ Using a non-pathogenic rod as a rect cell-cell challenge with bacteria (E. standard, a characteristic pattern of cy­ coli K12) without Caco-2 cells resulted tokine concentrations that depends on in the production of about 12,000 to the incubation time will be evoked. The 20,000 pg/ml TNF-α after 16 h. The

88 presence of soluble factors produced by pg/ml). Evidently, investigations on ef­ the differentiated Caco-2 cells reduced fects of e.g. nutrients on immuno­ this production by about 45%. The spa­ modulation in the intestine with the pre­ tial separation of the bacteria from the sent model are performed in a com­ leukocytes by the semi-permeable pletely different inflammatory ‘range’ membrane only reduced the TNF-α than those where leukocytes are chal­ production by about 80%. Separation of lenged directly by bacteria or bacterial leukocytes and bacteria by the semi­ products. permeable membrane and the differenti­ In conclusion, from the primary ex­ ated, confluent Caco-2 cell monolayer perimental results obtained with the de­ reduced the TNF-α synthesis to about scribed co-culture model it can be seen 80 to 300 pg/ml (~1 to 2%), which was suitable for the assessment of immuno­ still significantly higher than the value modulation near the intestinal mucosa resulting from the incubation of Caco-2 by food ingredients, drugs, and differ­ cells and leukocytes solely (2-30 ent bacteria species.

LITERATURE

Abreu-Martin, M.T. and Targan, S.R.: Regula­ pharmaceutical drugs. In vitro evaluation of tion of immune responses of the intestinal immune effects using rodent or human im­ mucosa. Crit. Rev. Immunol. 16, 277-309 mune cells. Toxicology 96, 147-156 (1996). (1995). Ebert, E.C.: Proliferative responses of human Parlesak, A., Diedrich, J.P., Schäfer, C., and intraepithelial lymphocytes to various T­ Bode, C.: A low concentration of ethanol cell stimuli. Gastroenterology 97, 1372­ reduces the chemiluminescence of human 1381 (1989). granulocytes and monocytes but not the Haller, D., Bode, C., Hammes, W.P., Pfeifer, tumor necrosis factor alpha production by A.M., Schiffrin, E.J., and Blum, S.: Non­ monocytes after endotoxin stimulation. In­ pathogenic bacteria elicit a differential cyto­ fect. Immun. 66, 2809-2813 (1998). kine response by intestinal epithelial Perdigon, G., Alvarez, S., Pesce de Ruiz Hol­ cell/leucocyte co-cultures. Gut 47, 79-87 gado, A.: Immunoadjuvant activity of oral (2000). Lactobacillus casei: Influence of dose on the Hidalgo, I.J., Raub, T.J., and Borchardt, R.T.: secretory immune response and protective Characterization of the human colon carci­ capacity in intestinal infections. J. Dairy noma cell line (Caco-2) as a model system Res. 58, 485-496 (1991). for intestinal epithelial permeability. Gas­ Pinto, M., Robine-Leon, S., Appay, M.D., troenterology 96, 736-749 (1989). Kedinger, M., Triadou, N., Dussaulx, E., Jung, H.C., Eckmann, L., Yang, S.K., Panja, Lacroix, B., Simon-Assmann, P., Haffen, A., Fierer, J., Morzycka-Wroblewska, E., K., Fogh, J., and Zweibaum, A.: Entero­ and Kagnoff, M.F.: A distinct array of cyte-like differentiation and polarization of proinflammatory cytokines is expressed in the human colon carcinoma cell line Caco-2 human colon epithelial cells in response to in culture. Biol. Cell 47, 323-376 (1983). bacterial invasion. J. Clin. Invest. 95, 55­ Viney, J., MacDonald, T.T., and Spencer, J.: 65 (1995). Gamma/delta T cells in the gut epithelium. Latthe, M., Terry, L., and MacDonald, T.T.: Gut 31, 841-844 (1990). High frequency of CD8 alpha alpha ho­ Weintraub, B.C., Eckmann, L., Okamoto, S., modimer-bearing T cells in human fetal in­ Hense, M., Hedrick, S.M., and Fierer, J.: testine. Eur. J. Immunol. 24, 1703-1705 Role of alphabeta and gammadelta T cells (1994). in the host response to Salmonella infec­ Lebrec, H., Roger, R., Blot, C., Burleson, tion as demonstrated in T-cell-receptor-defi­ G.R., Bohuon, C., and Pallardy, M.: Im­ cient mice of defined Ity genotypes. Infect. munotoxicological investigation using Immun. 65, 2306-2312 (1997).

89 90 HOW MUCOSAL IMMUNITY IS CONTROLLED BY LOCAL FACTORS STEFAN C. MEUER Institute for Immunology, Ruprecht-Karls-University, Heidelberg, Germany

The mucosal immune system, the be explained at the level of T cells: body's largest immunological compart­ When LPT were mixed with peripheral ment, is constantly exposed to an enor­ blood macrophages and subsequently mous load of foreign antigens derived simulated through TCR/CD3 their pro­ from commensal bacteria and food. liferative behaviour was comparable to Under physiologic circumstances, how­ that of PBT. In a reciprocal fashion, ever, no systemic immunity can be gen­ PBT when incubated with mucosal erated at this interphase of the body with macrophages could not respond to such its environment. Given that the T stimulation. Therefore, one had to con­ lymphocyte population which exists in clude that central regulators of immune the lamina propria and organised tissues responses in the gut where derived from of the mucosa (LPT) is of polyclonal the monocyte/macrophage lineage. A nature one has to assume that luminal T phenotypic comparison between mu­ cells can interact with antigenic determi­ cosal monocytes and monocytes circu­ nants present in this location. Under lating in peripheral blood then pointed physiologic circumstances, however, towards substantial differences between "typical" immune responses are not these two cell-populations in that the generated. Since such reactions of the receptor for LPS (CD14) as well as a mucosal immune system toward luminal variety of additional adhesion molecules antigens result in inflammatory reactions known to be required for T cell co­ as observed in various types of stimulation such as CD58 and CD54 inflammatory bowel diseases. were down-regulated on LPMO. The An initial important clue to under­ functional consequences of this pheno­ standing why mucosal T cells, unlike typic alteration and the question how peripheral blood or lymphnode T cells, such a particular phenotype as observed do not respond to antigen encounter by in LPMO can be generated will be dealt their specific T cell receptors with sys­ with below. temic immune responses in vivo was the Earlier data had indicated that the finding that freshly isolated LPT when mucosal environment produces pro- compared with peripheral blood T cells oxidative substances of as yet unknown (PBT) do not undergo clonal ex­ nature. Thus, supernatants generated pansion/proliferation and cytokine pro­ from mucosal cells (epithelial cells and duction when stimulated in vitro monocytes) when added to peripheral through their CD3 antigen-receptor blood mononuclear cells (PBMC) dose complex. Interestingly, no phenotypic dependently inhibited their proliferative abnormalities could be detected in LPT response to antigen receptor stimulation. when analysed extensively in order to If one, however, added reducing understand their lack of proliferation to substances such as 2-mercapo-ethanol T cell receptor simulation. A second set (2-ME) inhibition was abolished and of experiments then made obvious that proliferation was normal. This sug­ this fundamental observation could not gested that the activity of pro-oxidative

91 locally secreted products could be parative analysis of monocytes from counteracted by reducing agents and blood and monocytes from the mucosa pointed towards the direction of a clearly demonstrated that mucosal physiologic pro-oxidative state of the monocytes, unlike their blood counter­ mucosal microenvironment. This find­ parts cannot produce cystein as a co­ ing was of particular interest because T factor for T cell activation. Importantly, lymphocytes are considerably sensitive cystein production by blood monocytes to inhibition by pro-oxidative sub­ does not occur spontaneously but has to stances. This is due to the fact that they be induced through engagement of par­ do not express the cystin/glutamate ticular cell-surface-receptors including transporter complex in their plasma CD14 (LTS-receptor) and CD58 (LFA­ membranes. Thus, therefore, they can­ 3) which are exactly those that are not not take up cystin which represents the expressed but LP/MO. precursor for the synthesis of Glu­ It was now very important to eluci­ tathion (GSH), one of the most potent date the mechanism how cell-surface­ intracellular reducing systems. Regular receptors such as CD14 or CD58, re­ GHS-levels are required for cell-cycle spectively, can be down-regulated on progression, transcriptional activity and monocytes. There exists one cytokine likely also for the stability of intracellu­ known to exert such an effect, namely lar proteins (particularly those which interleukin 10 (IL-10). It was to our exist as disulphide linked dimers). surprise to discover that IL-10 is pro­ Given the above finding of a pro-oxida­ duced in large quantities by mucosal tive state it was important to investigate epithelial cells, which are known to se­ the possibility whether GSH-levels in crete their products largely in a baso­ LPT would be significantly lower than lateral direction, i.e. towards the lamina in PBT which was indeed the case. propria. There, IL-10 could potentially Thus, LPT contain only 10% - 20% of down-regulate the above mentioned cell the GSH concentrations as PBT. surface receptors with the consequence Needless to say, the availability of re­ that monocytes can no longer be in­ ducing substances leads to an increase duced to produce cystein. in intracellular GSH-concentrations in Having worked up our way from the LPT with an alteration of their func­ T cell as the most distal effector element tional phenotype e.g. with regard to in an interactive compartment we identi­ proliferation. fied mucosal monocytes as central Physiologically, antigen encounter regulators of T cell reactivity. The latter, by T cells only leads to clonal T cell ex­ however, are controlled by mucosal pansion and an immune response spe­ factors such as pro-oxidative substances cifically directed at antigen when anti­ and epithelial cell derived interleukin 10. gens are presented on appropriate MHC Most probably, activities of epithelial molecules and when appropriate addi­ cells require induction or stimulation by tional stimuli are provided towards them luminal components. The intestinal mi­ (co-stimulation). croflora through interaction with epithe­ Given that T cells lack expression of lial cell components may play an im­ the cystin-transporter (see above) they portant role for the physiologic function are dependent for their GSH-synthesis of the latter, which dictate the functional pathway on cells which can produce a programs of immunocompetent cells secrete cysteine in their vicinity. This is homing to the lamina propria. There­ an important activity and contribution of fore, it will be of interest for an under­ monocytes to T cell responses. Com­ standing of homeostasis in the largest

92 immunological compartment of the hu­ competent cells are, despite ameliorating man body to elucidate precisely the acute symptoms are unlikely to affect mechanisms through which - in a sym­ the causal processes underlying un­ biotic fashion - commensal flora influ­ wanted immune responses. The latter ences functional phenotypes of the rather result as a consequence of a dis­ body's cellular elements. turbed microenvironment, which is In conclusion, the presented studies characterised by epithelial, cells inter­ of mucosal immunity highlight aspects acting with microbial components. of physiological and pathological im­ Therefore, eventually, therapies which mune regulation and point to the direc­ are directed at those elements will ad­ tion that therapies of mucosal inflam­ dress pathologic processes at their ba­ matory processes which target immuno­ sis.

THE AUTHOR'S LITERATURE CONCERNING THIS FIELD

Pirzer, U., Schürmann, G., Post, S., Betzler, Merz, H., Wallich, R., and Meuer, S.C.: M., and Meuer, S.C.: Differential Respon­ Cytokine messenger RNA expression and siveness to CD3-Ti versus CD2-dependent proliferation status of intestinal mononu­ activation of human intestinal T lympho­ clear cells in noninflamed gut and Crohn's cytes. Eur. J. Immunol. 20, 2339-2342 disease. Virchows Archiv B 426, 51-60 (1990). (1995). Qiao, L., Schürmann, G., Betzler, M., and Qiao, L., Braunstein, J., Golling, M., Schür­ Meuer, S.C.: Activation and signalling mann, G., Autschbach, F., Möller, P., and status of human lamina propria lympho­ Meuer, S.: Differential regulation of human cytes. Gastroenterology 101, 1529-1536 T cell responsiveness by mucosal versus (1991). blood monocytes. Eur. J. Immunol. 26, Qiao, L., Schürmann, G., Betzler, M., and 922-927 (1996). Meuer, S.C.: Down-regulation of protein Braunstein, J., Qiao, L., Autschbach, F., kinase C activation in human lamina pro­ Schürmann, G., and Meuer, S.: T cells of pria T lymphocytes: Influence of intestinal the human intestinal lamina propria are mucosa on T cell reactivity. Eur. J. Immu­ high producers of interleukin 10. Gut 41, nol. 21, 2385-2389 (1991). 215-220 (1997). Schürmann, G., Betzler, M., Post, S., Her­ Autschbach, F., Braunstein, J., Helmke, B., farth, Ch., and Meuer, S.C.: Soluble inter­ Zuna, I., Schürmann, G., Niemir, Z., Wal­ leukin-2 receptor, interleukin-6 and inter­ lich, R., Otto, H.F., and Meuer, S.C.: In leukin-l in patients with Crohn's disease situ expression of intereukin-10 in nonin­ and ulcerative colitis: Preoperative levels flamed human gut and in inflammatory and postoperative changes of serum concen­ bowel disease. Am. J. Pathol. 153, 121­ trations. Digestion 51, 51-59 (1992). 130 (1998). Qiao, L., Schürmann, G., Autschbach, F., Stallmach, A., Wittig, B., Giese, T., Pfister, Wallich, R., and Meuer, S.C.: Human in­ K., Hoffmann, J.C., Bulfone-Paus, S., testinal mucosa alters T cell reactivities. Kunzendorf, U., Meuer, S.C., and Zeitz, Gastroenterology 105, 814-819 (1993). M.: Protection of TNBS-induced colitis by Qiao, L., Golling, M., Autschbach, F., an IL-2IgG2b fusion protein in mice. Gas­ Schürmann, G., and Meuer, S.C.: T cell re­ troenterology 117, 866-876 (1999). ceptor repertoire and functional behavior of Sido, B., Braunstein, J., Breitkreuz, R., Her­ lamina propria T lymphocytes in inflamma­ farth, C., and Meuer, S.C.: Thiol-mediated tory bowel disease. Clin. Exp.Immunol. redox regulation of intestinal lamina propria 97, 303-308 (1994). T lymphocytes. J. Exp. Med. 192, 907-912 Autschbach, F., Schürmann, G., Qiao, L., (2000).

93 94 OLD HERBORN UNIVERSITY SEMINAR ON HOST MICROFLORA CROSSTALK MINUTES AND OVERVIEW OF THE DISCUSSIONS DIRK VAN DER WAAIJ Hoge Hereweg 50, 9756 TJ Glimmen, The Netherlands

DISCUSSION PARTICIPANTS (in alphabetical order): Lars-Göran Axelsson, Stig Bengmark, Vanya E. Grant, Peter J. Heidt, Barbara H. Iglewski, Tore Midtvedt, Stefan Meuerer, Simon Murch, Elisabeth Norin, Alexandr Parlesak, Volker Rusch, Hilde Uvatne, Dirk van der Waaij, Elaine Vaughan, Agnes Wold

Elaine Vaughan: Approaches to of the microflora for 16S ribosomal investigate the diversity and RNA fluorescent probe design. functionality of intestinal mi­ The question arose as to how the re­ crobes. trieval of mucosal biopsies affects the Some advantages and disadvantages microflora. (The best material for mu­ of various molecular methods for inves­ cosa-associated microflora comes from tigating the microflora diversity were patients with abdominal pain who are discussed. The microbial diversity and biopsied in the colon. The best biopsy community behaviour over time of GI­ material for mucosa-associated micro­ tract faecal samples has been studied flora comes from patients who do not using fingerprinting techniques, such as suffer infectious bowel diseases. Very DGGE or TGGE (denaturing or tem­ early in the disease of Chronic Infec­ perature gradient gel electrophoresis). tious Bowel Disease, changes in the TGGE/DGGE of 16S rRNA PCR histologic appearance and inflammation products are especially suited to study can be seen. The question was also diversity in samples with largely un­ raised as to how soon after onset of the known microbial content, like the GI­ disease does the microflora change? In tract, without culturing. The intensity of man this cannot be studied because the a band in D-TGGE is a semi-quantitative colon needs to be flushed prior to scopy measure for the relative abundance of and biopsy. The benefit of scopy how­ this sequence in the population. Bands ever, is the histology. Histologic ap­ can be excised from the D/TGGE gels, pearance is the same in all patients while and advantage of this technique over the microflora differs between individu­ other fingerprinting methods, and als. Bacteria adhere to specific places sequenced, and the identity determined (glycocalyx) which differ between indi­ by comparison to the databases. viduals. In the crypts there are bacteria, D/TGGE can also be done on gut mu­ which are perhaps not susceptible to cosa biopsies. Fluorescent in situ hy­ oxygen; the way to isolate them is to bridisation (FISH) with image or flow make microtome (delete?) sections of the cytometry (rapid) analysis are excellent mucosa from the lamina propria up to methods to identify and quantify bacte­ the villi until the crypts are reached. rial groups in microflora without cul­ Using high throughput genomics turing, but do require prior knowledge methods involving DNA microar­

95 rays/chips, the diversity as well as the 2. allows analysis of many intesti­ functionality of the microflora may be nal/faecal samples. studied. Besides sequencing genomes of specific microflora commensals, the Note: In case bacteria have been ex­ construction of metagenomic libraries posed to antibiotics such as penicillins, consisting of the microflora DNA allows cephalosporins, aminoglycosides etc. access to the genetic and functional they may have become “leaky”. diversity of the microflora in the absence of culturing. Samples from quite a number of people are required to study Hilde Ulvatne: Defensins and de­ the influence of factors such as age and fensin-like molecules: Antibacte­ diseases. Thanks to the computer, mo­ rial Mode of Action. lecular analyses as described above can Antimicrobial peptides (AMP) can be be considered these days. classified by structure, and include: 1. α -helices, 2. ß -sheets stabilised by disulphide Vanya E. Grant: Flow cytometry: bridges, Can it help to analyse complex 3. extended structures, and biosystems? 4. loop-structures. 1. Validation by hybridisation as only a AMP are present in every organism so variable percentage is culturable. far investigated, and are located in epi­ 2. Why do it? thelia, leukocytes, and mucosal secre­ 3. High dimensionally; data handling tions. Any organism may possess AMP i.e. an explosion of computer inten­ from more than one structural class of sive data are generated. AMP and many AMP within the same 4. Where is the money? group (e.g. frogs have many different The reason why hybridisation is often AMP of the α-helical group). AMP may difficult is because bacteria in the gut act in synergy with each other and with must be “stressed”. This is because of other antimicrobial effector molecules in lack of nutrients etc. The majority of the the host. AMP have different antimicro­ crypts are loaded with microbes, which bial spectra, and some are more efficient are as yet not determined, even with 10­ against Gram-positive than Gram-nega­ 12 probes. Crypts may form a site for tive bacteria, while others again are translocation. Gram-staining easily more efficient against Gram-negative misses crypt colonisation by bacteria. than Gram-positive bacteria. Most AMP Questions that can be answered: Per­ are gene-encoded, but some are gener­ centage of bacteria with intact barrier ated by proteolytic cleavage of a native function that can be identified with protein (e.g. lactoferricin and LL-37). probes for prokaryotics is relatively In humans, defensins are produced low, because only sick and dead bacteria throughout the GI-tract, including the can pick up the probe. Treatment with mouth, epithelia in the oesophagus, the “Dibac C4” opens up the barrier. Some stomach, and the small intestine and in fluorescent stains act on Gram-positive colon. In the human gut, there are at bacteria only. This makes differentiation least 9 different defensins (HNP 1- 4 in between Gram-positive and Gram- the neutrophils, HD-5 and 6 in the negative fractions in mixtures easy in Paneth cells, HBD-1 in the small intes­ flow-cytometry. tine and colon, HBD-2 in the gastric The ‘money‘ (benefit) is in: mucosa and colon, and HBD-3 in colon) 1. well cut biopsies (these should be least artificially changed).

96 Figure 1: The interactions of antimicrobial peptides and the bacterial cell

The role of defensins in modulation cationic antimicrobial peptide, is gener­ of the intestinal microflora is difficult to ated upon gastric pepsin cleavage of bo­ elucidate, because of the high variability vine lactoferrin. of bacteria in the intestine. There are The modes of action of AMP: Due to also differences in the susceptibility the amphipathic, cationic structure of pattern, and no clear-cut correlation most of these peptides, an effect on the between susceptibility and pathogenic­ cytoplasmic membrane of susceptible ity. Furthermore, bacterial proteases bacteria has traditionally been postulated may destroy AMPs. Only little research as the main mode of action. Although an has been done on development of resis­ effect on the cytoplasmic membrane has tance, but resistance mechanisms exist. been shown for most peptides, other Lactoferrin (LF) is a multifunctional, mechanisms have also been reported, iron-binding protein of 70 kDa. LF is including interference with intracellular present in secretions like milk and tears, processes. in saliva, semen and colustrum, in addi­ The mode(s) of action of AMP is il­ tion to being a part of the secondary lustrated in Figure 1. At several points, granules of polymorphonuclear leuko­ the interaction between peptide and cytes. Its strategic appearances on the bacteria may lead to the bacterial cell mucosal surfaces and in leukocytes in­ death, as indicated with arrows. dicate a role in the primary defence of Lactoferrin B causes only minor the organism. It has been shown to ex­ leakage from liposomes, no general ert a broad antimicrobial activity, and collapse of membrane integrity (meas­ the antimicrobial activity is not iron-de­ ured by uptake of propidium iodide), pendent. Lactoferricin B, a 25 residue minor effect on bacterial respiration

97 (measured by CTC), but causes con­ together with the secretion of antibacte­ centration dependent depolarisation of rial defensins. the cytoplasmic membrane in E. coli The importance of antimicrobial sub­ (measured by JC-1). However, even stances has been studied in patients though the membrane of E. coli gets having surgical by-pass treatment depolarised, it does not necessarily against morbid obesity. After Roux-en­ cause death of these bacteria. This leads y Gastric bypass (GBP), a standard to the conclusion that ‘membrane-effect’ surgical procedure for morbid obesity, is not responsible for death of E. coli. food and oral-nasal-pharyngeal secre­ The mode of action of lactoferrin B may tion pass directly into the small bowel be hypothesised as the following: without passing through the acid envi­ 1. Attachment to bacterial surface. ronment of the normal stomach. Post­ 2. Binding with the cytoplasmic mem­ operatively (3-6 weeks), complying pa­ brane: tients underwent gastroscopy with jeju­ * bacterial stress response? nal biopsies using a sterile forceps. * induction of SOS-response? Postoperatively Gram-negative cocci, * inhibition of growth? anaerobic bacteria and yeasts were 3. Intracellular action: found in a lower number than found in * shut down of metabolism? control individuals. The Paneth cell, * depolarisation? histochemically, specific phloxine-tar­ * bacterial action = MBC? trazine stain showed a striking reduction Preliminary results indicate that lactofer­ of stained granula in the postoperative ricin B affects the bacterial protein syn­ biopsies, a result also obtained with the thesis. At a concentration of 5 mg/ml specific antibody to lysozyme. How­ suppression is observed, and 10 and 25 ever, immunohistochemical staining of Hg/ml further increase the effect. These human defensin-5 showed an upregula­ in vitro findings do not necessarily re­ tion of the antibacterial peptide in flect what may occur in vivo at these granula in the entire lower crypt wall as concentration levels. Through the bacte­ well in the crypt lumen. rial response to defensin attack, bacteria This indicates that the high exposure may even be protected against the effect of the small intestine to environmental of other antibiotics, like aminoglyco­ bacterial flora after GBS differentially sides, which affect protein synthesis. regulates the secretions of antibacterial substances and thus controls bacterial colonisation and prevents deleterious Lars-Göran Axelsson: Defensins bacterial overgrowth. and bacteria, a question of "live Mice having a conventional intestinal or let die?" microbiota are known to produce at least The human small intestine responds 17 enteric antimicrobial peptides, i.e. to bacterial challenge of the small bowel defensins. by secreting antimicrobial substances as Germfree mice can be used to com­ in the mouse. Specific cells, Paneth pare innate and classical immune re­ cells, are strategically situated at the sponses to microbes. In the germfree bottom of the small intestinal crypts and mice, the basic activity of defensins can contain granula packed with bioactive be studied and by mono-association substances. Paneth cell granules are de­ with a single bacterial strain specific re­ granulated upon stimulation and secrete sponses can be detected. HPLC frac­ e.g. lysozyme and phopholipase-2 into tionation of small intestinal crude ex­ the lumen of the small intestinal crypts, tracts has shown that the sterile intestine

98 of germfree mice contains at least three 2. Maintaining a healthy intestinal flora antibacterial components. A mono-as­ under different environmental condi­ sociation with Aeromonas hydrophila tions: (Bo-3N) produced two additional com­ * maintaining the balance between ponents. A pre-treatment with cortisone commensals. abolished the two first peaks with anti­ 3. Protection against pathogens: bacterial activity, which implies a gene *eradicate before critical infectious control by NFkB/IkB. number of organisms is reached, These results points to the role de­ * maintain commensal flora. fensins can have in the new-born in 4. Treatment of disease: controlling the establishment of a func­ * maintain healthy flora vs. aller­ tional intestinal bacterial flora and later gies, in protecting against unwanted coloni­ * maintain healthy flora vs. arthritis. sation of pathogens and deleterious mi­ 5. Treatment of cancer: crobes. * synthetic drugs. Conditions in humans where Paneth cell hyperplasia occurs, e.g. in the stomach, shows that there are genetic or Agnes Wold: Mucosal immunol­ mediator driven control of Paneth cells ogy. and its secretions. Lately there has been Hygiene hypothesis by Strachan much interest in the pharmacological (1989): Microbial exposure reduces the regulation or the administration of syn­ risk of allergy development. thesised defensins to substitute for anti­ Children with elder siblings have biotics to which microbes has acquired fewer allergies. This could be due to a resistance against. Recently several lower degree of contamination in the methods to manufacture different an­ first child than in the siblings. Indeed, timicrobial peptides have been pub­ when "day care" is started early (<2 lished. months) it seems to protect to allergy In the near future there is the impor­ development. Also growing up on a tant and much promising possibility to "life-stock farm" reduces the risk for induce and regulate or use defensins as allergy. new pharmacological entities in the Swedish infants develop IgA later control of infections caused by resistant than Pakistan children and children who or multi-resistant microbes. develop allergy later on have a lower Defensins might be important in the: IgA in their saliva compared to infants 1. Regulation of the process of acquir­ who become not allergic. ing an functional healthy intestinal When oral tolerance is induced in flora in the new-born: conventional mice (and in GF controls * are there qualitative differences that do not become so easily tolerant between individuals, following the same treatment) and their * if so, can the defensin production serum is injected into conventional nude and secretion be up-, respectively (athymic) mice the latter become tolerant down-regulated on demand in but not following injection of serum certain individuals under patho­ from the GF. logical conditions. Strachan, D.P.: Hay fever, hygiene and house­ hold size. BMJ 289, 1259-1260 (1989).

99 Simon Murch: A link between duce high amounts of volatile fatty ac­ mucosal regulatory lymphocytes ids, TNF and PAI-1. These patients and childhood food allergy. may show Prostate hyperplasia, hyper­ A hypothesis linking parasites with tension, diabetes, arteriosclerosis, hy­ development of allergy is presented. perinsulinimia, hyperuraemia and obvi­ Low hygienic circumstances involving a ously obesity. Their risk for develop­ high burden of pathogens/adjuvans ment of cancer is increased. disbalance the Th1/Th2 ratio. Viruses, Cows milk that comes from cows bacteria and protozoa may stimulate that are fed on hay (not fresh grass) may DC1 cells and stimulate Th1 forming. carry: Helminths and allergies on the other * trophic hormones, hand stimulate DC2 cells which stimu­ * bovine growth factor (IGF-1), late Th2 formation. and Multiple Food Allergy: * xeno-estrogens. * Breast-feeding has no preventive According to Swidsinsky, intestinal effect on allergy. mucus is normally free of bacteria. In * Allergies are less severe in pa­ patients with inflammatory bowel dis­ tients with enteropathies. ease (IBD) on the other hand, the bacte­ * May be found in subjects with rial concentration in the mucus may be "normal" bowel function (their high. This may either be the cause or the villi however may be reduced) result of IBD. * Transient IgA deficiency. * IgG sub-clones deficiency like IgG2 and IgG4. Elisabeth Norin: Phenotypic ex­ *Lymphocyte subset abnormality pressions in the small intestine. (low CD8, Low NK cells and In groups of 5 to 7 mice and rats the CD19 cells). MACs were studied under several con­ Ovalbumin sensitisation can induce "al­ ditions: lergic bowel dysmobility" in children. * Fasting overnight. Eotaxin attracts eosinophils, while * Vincristine injection. IL-10 and TGF-β do the opposite as * Similac diet. they induce development of suppressor In the rats, samples were taken at stan­ cells. dard (same) places of the intestines. "Para-cellular leakage" next to Significant differences were seen in the Peyer’s patches attracts cells to infiltrate mitotic index of the crypt cells in the the lamina propria. In Gambian children germfree animals. In their caecum a dif­ this is seen at the age of one year and ference in mitotic index (MI) was seen they may be dead by their second year. between males and females. The MI It seems that T-cells are involved but it was highest in rats following feeding of is uncertain whether these are Sup­ Lactobacillus rhamnosis GG and Clos­ presser cells. It is at the moment that tridium difficile (in particular in toxin "flora repair" will be of help to stop the producing strains on the 3rd day). process. In mice, similar results were ob­ tained (young male animals had a higher mitotic index than females). Rats show Stig Bengmark: Synbiotic treat­ higher mitotic indexes than mice. ment in clinical praxis. Speculation: The effect of treatment Allergy is also important in human on the MI is immunological, mediated adults. People who grow fat and pro­ by γδ T cells.

100 Barbara H. Iglewski: Quorum 4. Siderophores (β-cepacia) to acquire sensing in Pseudomonas aerugi- iron. nosa. 5. Antibiotic sensitivity (efflux pumps) Quorum sensing is bacterial commu­ regulate siderophores which may act nication co-ordinated activity of cells as two-edged sword. involved in the crosstalk. In this respect 6. Evade host defence (alter membrane it is important to know whether bacteria proteins). of different species can "talk" with each 7. Plasmid transfer (transfers genetic other. info). Cell-Cell signalling: Pseudomonas gene chip has 5769 * Gram-positive bacteria: Post genes. In vitro 433 genes of which 259 trans-stationally modified peptides genes upregulate and 179 down regu­ are transported out of the cell to late. Efflux pump transport level in act on receptors on the cell mem­ PAO1 effect of nutrients shows that brane (outside the cell). Classic is three genes are involved at a very low two components systemic kinase. level of quorum sensing. * Gram-negative bacteria: Quorum Lessons for future experiments: sensing occurs by small molecular 1. There is a need for three repeats cyclates WSL. per experimental variable. AI-1 is an auto-inducer. These small 2. High quality RNA is essential molecules are diffusing (or pumped) out (free DNA and intact mRNA). of the cell and may then diffuse back in 3. Gene 2 experiment in variable re­ to bind specific regulatory receptors. sponses involving: AI-2 are structures like antibodies are a. growth stage, diffusing (or pumped) out to bind to b. media, specific receptors on the cell surface. c. +/- O2. This results in synthesis of two compo­ 4. Limitations: nent kinases. Among AI-molecules a. can not see regulation short much homology exists between differ­ lived messenger RNA, ent bacteria. b. can not see regulation if gene is In mice, when the following bacteria poorly expressed, are co-cultured in agar beads and put in c. genes with multiple regulators the mouse lung, Pseudomonas aerugi­ pose challenges. nosa produces 30-C12-HSα which can 5. Transcript analysis is a good activate has-R in E. coli containing complement to other approaches HasR+LacB-Gfp. (to mutagenesis, proteomics, gene The question rises whether bacteria can transmission etc. talk with host cells. Polarised lung tis­ sue cells show their normal cillilar movement and other activities. When P. Alexander Parlesak: Developing aeruginosa is added, the cells will pro­ an in vitro model on the investi­ duce AIs. gation of the crosstalk among What type of bacterial behaviour is bacteria, enterocytes and leuko­ regulated by crosstalk (=quorum sens­ cytes near the intestinal mucosa. ing)? Model: It was clearly to be avoided 1. Virulence (toxins, exo-enzymes). to get a Graft versus Host effect be­ 2. Invasion (swarming, chemotaxis, tween the monocyte layer underneath a proteasis). filter, which separated them from the 3. Antibiotic production (self-defence). enterocytes. In this model crosstalk was

101 studied between bacteria and cells of tions were tested). host origin. The costs of testing one 4. LPS (endotoxin). compound was high (fi 20,000.-). 5. No treatment. Selective stimulation of Caco-2 oc­ Both E. coli K12 and Lactobacillus curred only in the presence of leuko­ sakei stimulated expression of IL-8, cytes on the basal side caused produc­ while all bacteria stimulated expression tion of β-actin, IL-8, TNFα, IL-1β and of β-actin. TNF-γ. The process follows three subsequent Bacteria tested were human isolates: steps: 1. E. coli K12 (two concentrations were Step 1: leukocytes stimulate the bacte­ tested). ria, 2. Lactobacillus johnsonii (two concen­ Step 2: stimulate leukocytes stimulate trations were tested). Caco-2 cells (enterocytes), 3. Lactobacillus sakei (two concentra­ Step 3: enterocytes affect bacteria.

102